The Language of Fiction Emar Maier Full Chapter

The Language of Fiction Emar Maier
Visit to download the full and correct content document:

https://ebookmass.com/product/the-language-of-fiction-emar-maier/
OUP CORRECTED PROOF – FINAL, 11/9/2021, SPi
The Language of Fiction

The Language of Fiction

Edited by
EMAR MAIER AND ANDREAS STOKKE
1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Emar Maier and Andreas Stokke 2021
© The chapters their several authors 2021
The moral rights of the authors have been asserted
First Edition published in 2021
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2021939061
ISBN 978–0–19–884637–6
DOI: 10.1093/oso/9780198846376.001.0001
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents
List of contributors vii

1. Introduction 1
Emar Maier and Andreas Stokke
I. TRUTH, REFERENCE, AND IMAGINATION

2. Fictional reference as simulation 17
François Recanati
3. Sharing real and fictional reference 37
Hans Kamp
4. Fictional truth: In defense of the Reality Principle 88
Nils Franzén
5. On the generation of content 107
Sandro Zucchi
6. Do the imaginings that fictions invite have a direction of fit? 131
Manuel García-Carpintero
II. STORYTELLING
7. In search of the narrator 157
Regine Eckardt
8. Extracting fictional truth from unreliable sources 186
Emar Maier and Merel Semeijn
9. Narrative and point of view 221
Samuel Cumming
10. A puzzle about narrative progression and causal reasoning 255
Daniel Altshuler
11. Isomorphic mapping in fictional interpretation 277
Matthias Bauer and Sigrid Beck
vi 
III. PERSPECTIVE SHIFT

12. Metalinguistic acts in fiction 301
Nellie Wieland
13. Computing perspective shift in narratives 325
Márta Abrusán
14. Derogatory terms in free indirect discourse 349
Isidora Stojanovic
15. Protagonist projection, character focus, and mixed quotation 379
Andreas Stokke
Index 405
List of contributors
Márta Abrusán is a tenured research fellow at the Centre national de la recherche

scientifique (CNRS), based at the Institut Jean Nicod in Paris, France. She is interested in
natural language semantics and pragmatics, and currently works on issues related to
presuppositions, lexical semantics, and perspective shift. She also has an interest in the
question of how semantics might influence grammaticality judgments; her book related to
this topic, entitled Weak Island Semantics, was published by OUP in 2014.
Daniel Altshuler is an Associate Professor of Semantics at the University of Oxford. His
research explores how compositional semantics interacts with discourse coherence. His
monograph, Events, States and Times (2016, De Gruyter), investigates the meaning of tense
and the indexical now in narrative discourse. His monograph (with Robert Truswell),
Extraction from Coordinate Structures at the Syntax–Discourse Interface (OUP, forthcom-
ing), offers a focused case study of patterns of extraction involving and. Currently, he is
exploring how fictional discourse motivates extensions of dynamic-semantic frameworks,
looking at imaginative resistance and garden-pathing.
Matthias Bauer is Professor of English Literature at Eberhard Karls University, Tübingen.
With Sigrid Beck, he chairs a literary-linguistic research project on “Interpretability in
Context”, which is part of the Collaborative Research Center (CRC 833) “The Construction
of Meaning”. The group has recently published Linguistics Meets Literature: More on the
Grammar of Emily Dickinson with De Gruyter. Matthias Bauer also chairs the interdiscip-
linary Research Training Group “Ambiguity: Production and Perception”, funded like the
CRC by the German Research Foundation (DFG). Another DFG research project is
concerned with the aesthetics of co-creativity in early modern English literature.
Sigrid Beck is Professor of Linguistics in the English Department of the Eberhard Karls
University, Tübingen. She is co-chair with Matthias Bauer of the project “Interpretability in
Context” and chair of the Collaborative Research Center (CRC 833) “The Construction of
Meaning”. Her field of expertise is natural language semantics. In addition to interdiscip-
linary work at the literature–linguistics interface, her current research interests include the
diachronic development of universal quantification and the incremental processing of
semantic composition.
Samuel Cumming is Professor in the Department of Philosophy, University of California,
Los Angeles. He studies the details and foundations of conventional representational
systems (mainly natural language and film), and is interested in the roles coordination
and precedent play in the determination of reference, truth conditions, and propositional
content.
Regine Eckardt is Professor of German and General Linguistics at the University of
Konstanz. Her research interests are in formal semantics and pragmatics in synchrony
viii   
and diachrony. She has worked on focus constructions, particles, perspective taking, and
polarity phenomena, and is the author of Meaning Change in Grammaticalization (OUP,
2006) and The Semantics of Free Indirect Discourse (Brill, 2015). Her current interest in the
connections between fiction and formal semantics goes back to her collaboration in the
Research Unit Textstrukturen at the University of Göttingen, starting with its foundation in
2009 and ongoing.
Nils Franzén is a researcher in theoretical philosophy at the University of Uppsala. His PhD
thesis, “Sense and Sensibility: Four Essays on Evaluative Discourse”, concerns the nature of
moral and aesthetic language. Franzén has published on aesthetic evaluation, first-hand
experience, affective states of mind, and other topics.
Manuel García-Carpintero received his PhD at the University of Barcelona, and has taught
there since then. He works in the philosophy of language and mind, and related epistemo-
logical and metaphysical issues. Currently he is completing a book under contract with
OUP on the nature of speech acts in general and assertion in particular, entitled Tell Me
What You Know.
Hans Kamp is a linguist and philosopher at the University of Stuttgart and University of
Texas, Austin. He is best known for his work on two-dimensional semantics (Formal
Properties of “Now”, 1971) and dynamic semantics, in particular the founding of discourse
representation theory (A Theory of Truth and Semantic Representation, 1981 and From
Discourse to Logic, 1993 (with U. Reyle)). A collection of his articles appeared as Meaning
and the Dynamics of Interpretation (von Heusinger and ter Meulen, eds., 2014).
Emar Maier is Assistant Professor at the University of Groningen, affiliated with both the
philosophy and linguistics departments. After receiving his PhD in philosophy (Nijmegen,
2006), he led an ERC Starting Grant project (2011–16) on quotation. He is currently leading
an NWO VIDI project investigating the semantics of imagination and fiction, combining
topics and insights from philosophy (e.g. mental files, imaginative resistance) and linguis-
tics (e.g. dynamic semantics). His research interests include: fiction and imagination;
pictorial semantics and ‘superlinguistics’; attitude ascription and quotation; and reference
and indexicality.
François Recanati is a Professor at Collège de France. His publications in the philosophy of
language and mind include Meaning and Force (Cambridge University Press, 1987), Direct
Reference: From Language to Thought (Blackwell, 1993), Oratio Obliqua, Oratio Recta (MIT
Press, 2000), Literal Meaning (Cambridge University Press, 2004), Perspectival Thought
(Oxford University Press, 2007), Truth-Conditional Pragmatics (Oxford University Press,
2010), Mental Files (Oxford University Press, 2012), and Mental Files in Flux (Oxford
University Press, 2016). He is a member of Academia Europaea and a Foreign Honorary
Member of the American Academy of Arts and Sciences.
Merel Semeijn is currently finishing her PhD at the faculty of philosophy, University of
Groningen. She specializes in semantics and philosophy of fiction. In her dissertation she
develops an extension of the Stalnakerian common ground framework to model different
ways we use language to engage with fiction and how these uses relate to non-fictional
language uses such as assertion and lying. She has previously published on the role of
   ix
pretend play in theory of mind acquisition and on metafictional discourse. Current interests
include fiction, fiction operators, learning from fiction, (bald-faced) lies, and common
ground.
Isidora Stojanovic is a Director of Research at the Centre national de recherche scientifique
(CNRS) at the Institut Jean Nicod in Paris, France. She holds a PhD in philosophy from
Stanford University and a PhD in cognitive science from the École Polytechnique. Her
articles have appeared in journals such as American Philosophical Quarterly, Erkenntnis,
Linguistics and Philosophy, Inquiry, Southern Journal of Philosophy, and Synthese. She has
contributed chapters to many collective volumes, six of which are from Oxford University
Press. Her research lies at the interface between philosophy of language and formal
semantics and combines theoretical tools with empirical methods.
Andreas Stokke is a Docent and Senior Lecturer in Philosophy at the Department of
Philosophy, Uppsala University, and a Pro Futura Scientia Fellow at the Swedish
Collegium for Advanced Study. Previously he has been a visiting scholar at the
Department of Philosophy, New York University, an academic visitor of the Faculty of
Philosophy, University of Oxford, and a visiting fellow at Magdalen College, Oxford.
Stokke’s research is mainly in the fields of philosophy of language and epistemology, but
he has also worked on ethics and the philosophy of action. He has published extensively on
lying and insincerity, protagonist projection, dynamic semantics, and many other topics.
Nellie Wieland is a Professor in the Department of Philosophy at California State

University Long Beach. She received her PhD from UC San Diego in 2007. She works
primarily in the areas of philosophy of language and social metaphysics. She has written on
problems of agency, autonomy, and authority in both of those areas. She also works on
quotation and other metalinguistic phenomena as a way of theorizing about the nature of
language.
Sandro Zucchi is a Professor at the Università degli Studi di Milano. He has worked on
quantification, tense, aspect, truth in fiction, and sign languages. Recently, he has been
working on a variablist semantics for pronouns and the semantics of conditionals.
1
Introduction
Emar Maier and Andreas Stokke
Language is used to represent states of affairs in the world around us.

Paradigmatic language use aims at saying something true about how things
actually are. Yet we also use language for other purposes. It is common to note
that a key feature of human language and thought is our comprehensive capacity
for displacement—the ability to talk and think about other things than those in
our immediate environment.¹ One form of displacement is temporal, as when
talking about something in the past or future. Another is spatial, concerning how
things are at remote locations. And yet another form of displacement is modal
thought and talk about what might, could, should, or must be the case.
At least as central as these basic forms of displaced language is the way we use
language to create and sustain fictions. Fictional language is related to modal talk,
but it is wrong to equate the two. We do not just use fictional language to represent
what might, could, should, or must be the case. Our engagement with fictional
stories is a central practice of our thought and talk in its own right. It is hard to
think of any actual human culture that has not in some way engaged in fictional
discourse, and it is a cliché that our cognitive and social lives are permeated
by fiction.
While we depend on facts and knowledge about the actual world, we are just as
much bound up in works of the imagination. We live our lives in the company of
fictional stories and characters. We compare ourselves to them, and we are
emotionally involved with them. This ranges from the frivolous to the serious.
One among countless similar current websites offers to “Test Your Personality
And We’ll Tell You If You’re Sherlock Holmes Or Watson!”² But many people
have a profound feeling that certain novels or movies make life worth living, and
would be bemused if asked to name a nonfiction book or documentary film that
means the same to them. Much of what we do, say, and think is influenced by, if
not copied from, works of fiction that we treasure. Whole generations idolize and
imitate certain fictional characters, and we all know and make frequent reference
to a large stock of shared fictional stories.
¹ Hockett (1960) famously counted displacement among the “design features” of human language.
² www.thequiz.com/test-your-personality-and-well-tell-you-if-youre-sherlock-holmes-or-watson.
Accessed 22 July 2021.
Emar Maier and Andreas Stokke, Introduction In: The Language of Fiction. Edited by: Emar Maier and Andreas Stokke,
Oxford University Press. © Emar Maier and Andreas Stokke 2021. DOI: 10.1093/oso/9780198846376.003.0001
2     
Even though fiction is not always couched in language, it is undeniable that our
most valued fictions are linguistic. Most fiction is conveyed and understood
through linguistic utterances, spoken or in writing. Unsurprisingly, therefore,
the language of fiction has long been a subject of research in literary criticism
and philosophical aesthetics.
In broad terms this work grew out of movements in twentieth-century
literary theory such as Russian Formalism, Structuralism, and New Criticism.
While these research programs differed substantially, they shared the general view
that literary works merit study in their own right and should not just be subject to
normative appraisal. In one of the classics of this field, Seymour Chatman puts it
succinctly:
On this view literary theory is the study of the nature of literature. It is not
concerned with the evaluation or description of any particular literary work for
its own sake. It is not literary criticism but the study of the givens of criticism, the
nature of literary objects and their parts. (Chatman 1978, 18)
These twentieth-century trends were to a large extent influenced by linguistic

ideas of the time, in particular structural linguistics and semiotics.³
The development of other, more recent theoretical approaches in semantics and
pragmatics calls for new work on the linguistic aspects of fictional discourse.
Indeed, recently there has been a fresh surge of interest in studying fictional
discourse within the intersection between linguistics and philosophy of language.
Drawing on the longstanding and successful collaborative efforts between lin-
guists and philosophers of language on such topics as anaphora, attitude ascrip-
tions, indexicality, counterfactuals, presupposition, and quotation, work in this
field has taken a new look at the semantic and pragmatic features of fictional
language.
This volume contributes to this growing field of research by presenting fourteen
new essays on the language of fiction by both philosophers and linguists. The
contributions are divided into three parts:
Part I: Truth, reference, and imagination

Part II: Storytelling
Part III: Perspective shift
This division is intended to reflect key spheres of interest in the field. In this
introductory chapter we provide brief overviews of each of these areas of research.
³ This tradition was founded on the work of de Saussure (2011 [1916]), Hjelmslev (1953 [1943]),
Jakobson and Halle (1956), Greimas (1987 [1970]), and others.
 3
Summaries and descriptions of the contributions are given in the introductions to

the individual parts.
1.1 Truth, reference, and imagination
It is a well-established observation that the central semantic notions of truth and

reference are highly problematic in the context of fiction (cf. e.g. Lewis 1983
[1978]; Currie 1990; Walton 1990; Woodward 2011; Friend 2017). A familiar set
of traditional puzzles concerns the reference of fictional names. Why do we judge
sentences like those in (1) as true if the name “Frodo” does not refer?
(1) a. Frodo is a hobbit.

b. Frodo was created by Tolkien.
c. Frodo doesn’t exist.
Recent interdisciplinary work explores, for instance, the use of mental files,
presupposition theory, and dynamic semantics to deal with a cluster of funda-
mental philosophical puzzles of truth and reference in fiction.
Much philosophical work on fictional names and characters has been focused
on metaphysical issues. Realist views hold that there are fictional characters.
Among these, so-called Meinongian realists agree that there are fictional charac-
ters but insist that fictional characters do not exist.⁴ For the Meinongian, Frodo
Baggins is an object that is a hobbit, that was born in the Shire, that traveled to
Mount Doom, but that does not exist. This gives the Meinongian the advantage of
being able to claim that “Frodo” uniformly refers to this non-existent hobbit, even
in the problematic case of (1c).
Most realists, however, reject the Meinongian conclusion that there are things
that do not exist. One general complaint is a sense of perplexity at the suggestion
that there is a metaphysically significant difference between, for example, (2a)
and (2b).⁵
(2) a. There are hobbits.

b. Hobbits exist.
However, this metaphysically motivated rejection of Meinongianism leaves the

realist with the challenge of how to accommodate the clear sense that negative
⁴ Wolterstorff (1979), Parsons (1980, 1982), and Priest (2005). The relation of these theories, and
others in the vicinity, to Meinong’s (1960 [1904]) original view remains an open question. On this see
van Inwagen (1977, 299, fn. 1).
⁵ Cf. van Inwagen (1977).
4     
existentials like (1c) are true. On the other hand, if negative existentials are indeed
true, the realist’s rejection of the difference between being and existence prima
facie forces her to conclude that there are no fictional characters after all, thereby
disowning the fundamental realist tenet.
An alternative approach to these perennial issues stems from philosophical
theories of fiction that aim to understand the speech act involved in fictional
discourse (Searle 1975; Currie 1990; Walton 1990; Davies 1996; Kripke 2011). It is
uncontroversial that utterances used to tell fictional stories are not assertions.
Consider, for example, the first sentences of A.S. Byatt’s The Children’s Book:
(3) Two boys stood in the Prince Consort Gallery, and looked down on a third.
It was June 19th, 1895. (Byatt, 2009, 5)
It is a datum that, in writing (3), Byatt was not asserting that on 19 June 1895 two
boys were standing in the Prince Consort Gallery looking down on a third boy.
Even though (3) may be true or false of the actual world, it is not put forward as a
claim about what was actually the case.
One tradition holds that, instead, authors of fiction are asserting under pretense
(Searle 1975; Kripke 2011; Recanati 2000, 2018, this volume). So, on this view,
Byatt was pretending to assert something about what was the case on 19 June
1895. Yet it is fair to say that the most influential view—Matravers (2014) calls it
the “Consensus View”—has been that fictional utterances are prescriptions, or
invitations, to imagine things (Walton 1990; Currie 1990; Friend 2011; Stock 2017;
Maier 2017; García-Carpintero, this volume). While Byatt was neither asserting
something about the actual world nor merely pretending to do so, she was inviting
us to imagine that on 19 June 1895 two boys were standing in the Prince Consort
Gallery looking down on a third boy.
Building on these views, and others, philosophers and linguists have proposed
semantic and pragmatic frameworks for understanding fictional discourse in ways
that also provide accounts of puzzling uses of fictional names, like negative existen-
tials (Eckardt 2015; Recanati 2018, this volume; Bauer and Beck 2014; Maier 2017;
Kamp, this volume). In some of these frameworks uses of fictional names interact
with discourse referents that can be exported from information imagined as making
up story content to other areas of discourse figuring metafictional, counterfictional,
and existential uses of names originally introduced by fictional utterances.
Within this broad area of research, there has been work on the longstanding
problem of truth in fiction, and relatedly the generation of fictional content over
and above what is made explicit by a narrator (Bonomi and Zucchi 2003; Eckardt
2015; Zucchi, this volume, Maier and Semeijn, this volume; Franzén, this volume).
Some have proposed models that take fictional content as generated in ways that
parallel familiar ways of thinking of discourse in terms of common ground
information, and there have been efforts to understand fictional utterances as
 5
aimed at incrementing alternative kinds of contextual (Bonomi and Zucchi 2003;

Eckardt 2015, this volume; Stokke 2013a; Zucchi, this volume) or psychological
information states (Maier 2017; Kamp, this volume; Recanati 2018). Many issues
remain unresolved in relation to these recent linguistically oriented approaches to
fictional discourse. One central challenge concerns how to understand the kind of
imagining involved in entertaining fiction, particularly in relation to the trad-
itional problem of intensional identity—that is, roughly, our ability to share
attitudes toward things that may or may not exist, such as fictional characters
(Friend 2011). If you say (4a) and I reply with (4b), in what sense are we talking
about the same thing, or character, given that that thing does not exist?
(4) a. Hamlet is Danish.

b. Yes, and he lives at Elsinore.
A particularly challenging instance of this kind of problem arises for counter-

fictional statements such as those in (5) and interfictional statements, as in (6).
(5) a. Hamlet might not have returned from England.

b. Frodo could have died before reaching Mount Doom.
c. Sancho Panza could have murdered Don Quixote.
(6) a. Hamlet is younger than Sherlock Holmes.

b. Anna Karenina is smarter than Rodion Raskolnikov.
c. Sancho Panza would have liked Frodo Baggins.
Most agree that both counterfictional and interfictional statements can be true or
false. Much of our everyday talk about fiction as well as literary criticism will be
seen as defective if we cannot evaluate such uses of fictional names in terms of
truth and falsity. However, there is no work that prescribes imagining these things,
so what makes them true?
Frameworks designed to account for phenomena such as anaphora and pre-
supposition projection, as well as other tools from philosophy of language and
formal semantics, have provided fruitful new ways of approaching these and
surrounding issues (Pagin 2014; García-Carpintero 2018; Maier 2017).
1.2 Storytelling
An obvious fact about much fiction is that it is presented in the form of stories.
Most of the fiction that we encounter is narrative in its form, and extensive work
in literary criticism, philosophical aesthetics, and the philosophy of literature has
been concerned with narrativity. At the same time, traditionally, relatively little
6     
research in theoretical linguistics and philosophy of language has been concerned

with how to understand narrativity qua linguistic phenomenon.
Recent work has taken up these issues concerning the specifically linguistic
properties of narration (Fludernik 1993; Eckardt 2015; Zeman 2016, 2018; Bauer
and Beck, this volume). It has long been thought that a central linguistic feature of
storytelling concerns sequences of events and temporal progression. Yet linguis-
tically oriented work on narrative has questioned whether the temporal order of
events should be taken as a hallmark of narration, and has instead suggested that
the (or a) central feature of narratives is a specific kind of interplay between points
of view (Zeman 2018). Such approaches take the familiar sense in which narratives
involve a point of view, the narrator’s, from which events and things are repre-
sented as a defining characteristic of narrative.
Of particular significance here is the study of how the use of tense facilitates the
double-perspective involved in narrative, on these views. One suggestion is that
narrating events in the past tense allows the narrator (or speaker) to separate two
points of view corresponding to the present and the past time that is represented
(Zeman 2018). Studying these effects has provided promising new ways of
approaching the nature of narrative as a linguistic phenomenon.
Relatedly, linguistic approaches have been brought to bear on issues concerning
to what extent narrative progression is compatible with representation of a
character’s thoughts or attitudes from within their point of view (Caenepeel
1989; Cumming, this volume). This has involved re-examining the ancient dis-
tinction between diegesis (the representation of states of affairs or events) and
mimesis (the representation of the thoughts or speech of one or more protagon-
ists) (Fludernik 1993; Cumming, this volume).
According to a traditional view, the distinction between diegetic and mimetic
aspects of narratives is relatively sharp, the thought being that narrative progres-
sion through diegesis is interrupted once mimesis of a particular point of view
takes over (Caenepeel 1989). The motivation for this kind of view can be illus-
trated by toy examples like (7).
(7) [a] John and Mary were walking down the road. [b] It was a sunny day. [c]
“Nice weather,” [d] said Mary. [e] Then they saw a car coming toward them.
[f ] “I wonder who that could be,” [g] said John.
In (7), [a], [b], [d], [e], and [g] may be seen as diegetic narration of events, while
[c] and [f] present mimetic representations of what was said. As such, one might
think that [c] and [f] do not contribute to the progression of the narrative qua
sequence of events occurring in the story. That is, they are distinct elements that
do not add to our information about what is the case in the world of the story.
On the other hand, recent work on free indirect discourse and other modes of
representing things from the point of view of protagonists has questioned the
 7
sharp divide between diegetic and mimetic aspects of narratives (Fludernik 1993;
Wieland, this volume; see also later in this chapter). These ways of analyzing
narrative discourse from within linguistically motivated theories of tense and
aspect have likewise been brought to bear on the relation between narrators and
characters. Some have likened this kind of key characteristic feature of narration
to grammatical dichotomies such as between speaker and observer (Zeman 2018).
Similarly, formal frameworks for understanding discourse relations have tried
to tackle the nature of narrative relations (Asher and Lascarides 2003; Asher 2012;
Altshuler, this volume). Such theories have seen narration as a rhetorical relation
between discourse units. Yet there is still no consensus on what defines narration
qua linguistic phenomenon, or on whether to primarily think of narrative as a
mode of discourse, a genre, or in terms of its cognitive features (Zeman 2018).
One issue in this ongoing investigation is the role of a narrator, or, more
abstractly, an organizing point of view. A narrator can be seen as minimally
involved in any utterance, in the sense of being the speaker (Zucchi, this volume).
Yet there is a need for understanding the arguably more involved sense in which
fictional stories, in particular, are presented from the perspective of someone telling
the story. Narrators may be more or less present; they may be more or less explicit
about themselves; they may be reliable or unreliable (Eckardt, this volume). While
these issues are not new, there is interest in approaching them from the point of
view of semantic and pragmatic frameworks for understanding discourse.
A particular instance of this concerns various ways in which audiences to
fictions sometimes export information from fictional contexts to their own non-
fictional context. Philosophical work on fiction has been interested in various ways
in which audiences can often be said to learn from fiction (Lewis 1983 [1978];
Walton 1990; Currie 1990; Gendler 2000; Stock 2017). Building on this work,
linguistic theories of fiction are exploring mechanisms for extracting true infor-
mation from fictional discourse (Bauer and Beck 2014, this volume; Maier and
Semeijn, this volume; Kamp, this volume).
1.3 Perspective shift
In parallel with efforts to better understand the linguistic characteristics of

mimetic and diegetic features of narratives, some philosophers and linguists
have been interested in analyzing certain kinds of perspective shift in fictional
discourse. As a canonical instance of perspective shift in fictional discourse, free
indirect discourse, which blends the perspectives of narrator and protagonist, has
received significant attention from both philosophers and linguists in recent years
(Schlenker 2004; Sharvit 2008; Eckardt 2015; Maier 2015).
Like the more well-known forms of direct discourse and indirect discourse, free
indirect discourse is a metarepresentational device for representing (or reporting)
8     
thoughts or speech. According to the standard gloss, in free indirect discourse

tenses and person-features of pronouns are used as in indirect discourse; that is,
from the point of view of the speaker or narrator. By contrast, everything else
appears as in direct discourse; that is, from the point of view of the subject of the
report. Take this passage from Iris Murdoch’s The Sandcastle:
(8) Oh let the train come! Now at last the train was appearing, there it was after
all. [ . . . ] No sign of Rain. Perhaps she had been taken ill. Perhaps she had
decided not to come. Perhaps she had been offended at something he had
said yesterday. Perhaps she had decided to go away altogether. No, there she
was, thank God, so small he hadn’t seen her behind that porter. She had seen
him and was waving. She was almost at the barrier now. She was here.
(Murdoch 2003 [1957], 236)
Characteristically, (8) uses exclamations like oh, no, and thank God, and indexicals
like now, yesterday, and here, from the point of view of the protagonist—that is, as
in a direct discourse report—while using the past tense and the third-person
pronoun him to refer to the subject of the report from the point of view of the
narrator, as in indirect discourse. Accordingly, work on free indirect discourse has
been engaged with analyzing the mechanisms by which this style of report
preserves certain elements as in direct discourse such as exclamations, indexicals,
dialect words, and speaker-oriented items (Stojanovic, this volume).
More recent developments suggest a more general notion of point of view and
perspective shift in fictional narrative (Stokke 2013b, this volume; Hinterwimmer
2017; Abrusán 2020, this volume). One area of interest has been instances of so-
called protagonist projection, in which a single word or phrase is re-oriented to the
perspective of the narrator, as in this much-discussed example:
(9) He gave her a ring studded with diamonds, but they turned out to be glass.
(Holton 1997)
In this case “diamonds” is used from the perspective of someone other than the
speaker or narrator. On a natural reading, “diamonds” is used because she (the
female protagonist) thought the ring was studded with diamonds.
A pressing question in this area concerns to what extent this kind of perspective
shift should be understood as genuinely mimetic; that is, to what extent does it
convey that the relevant protagonist explicitly made an utterance or had a thought
involving the relevant words (Hinterwimmer 2017; Abrusán 2020, this volume)?
This issue can be illustrated by this example from Vladimir Nabokov’s Pnin (cited
and discussed in Wood 2008, 25):
 9
(10) Fastidious Pnin rinsed it, and was wiping it, when the leggy thing somehow
slipped out of the towel and fell like a man from a roof. (Nabokov 2010
[1957], 151)
Should we think that “leggy” is used in (10) because Pnin thought to himself,
“That leggy thing!” or something similar? It may be more accurate to say that
“leggy” is used here because from Pnin’s perspective the nutcracker could be
described as “leggy,” even if he did not explicitly think or say so.
Some researchers have argued that in cases of more general viewpoint shift or
focalization (Genette 1980; Bal 1997) things and events are described from the
point of view of a protagonist, and yet there is no representation of their explicit
thoughts or utterances (Hinterwimmer 2017). An adequate account of these
phenomena from a linguistic point of view can be expected to shed new light on
the interplay between mimetic and diegetic aspects of fictional discourse.
Similarly, semantic and pragmatic theories have proposed that in cases of so-
called represented perception the narrator or author describes not so much how
things are at the world of the story but how the relevant protagonist thinks they
are (Caenepeel 1989). Consider, as an example, (11).
(11) Mary thought the noise at the door was her cat, and went to open it.
Something awful was waiting on the other side. (Cumming, this volume)
The second clause in (11) can be read as conveying that Mary thought that something
awful was waiting on the other side of the door, rather than conveying that, according
to the story, something awful was waiting on the other side of the door.
It is still an open question to what extent these more general perspectival
phenomena like protagonist projection, viewpoint shift, and represented percep-
tion relate to more well-understood categories like free indirect discourse and
mixed quotation.
References
Abrusán, M. (2020). The spectrum of perspective shift: Protagonist projection vs. free
indirect discourse. Linguistics and Philosophy. doi: 10.1007/s10988-020-09300-z.
Asher, N. (2012). Implicatures in discourse. In G. Grewendorf and T.E. Zimmermann
(Eds.), Discourse and Grammar: From Sentence Types to Lexical Categories [Studies
in Generative Grammar 112] (pp. 11–42). Berlin / New York: de Gruyter.
Asher, N. and Lascarides, A. (2003). Logics of Conversation. Cambridge: Cambridge
University Press.
10     
Bal, M. (1997). Narratology: Introduction to the Theory of Narrative. Toronto:

University of Toronto Press.
Bauer, M. and Beck, S. (2014). On the meaning of fictional texts. In D. Gutzmann,
J. Köpping, and C. Meier (Eds.), Approaches to Meaning (pp. 250–275). London:
Brill.
Bonomi, A. and Zucchi, S. (2003). A pragmatic framework for truth in fiction.
Dialectica, 57(2), 103–120.
Byatt, A.S. (2009). The Children’s Book. New York: Knopf.
Caenepeel, M. (1989). Aspect, Temporal Ordering and Perspective in Narrative Fiction.
PhD thesis, University of Edinburgh.
Chatman, S. (1978). Story and Discourse: Narrative Structure in Fiction and Film.
Ithaca, NY: Cornell University Press.
Currie, G. (1990). The Nature of Fiction. Cambridge / New York: Cambridge
University Press.
Davies, D. (1996). Fictive utterance and the fictionality of narratives and works. British
Journal of Aesthetics, 55(1), 39–55.
de Saussure, F. (2011 [1916]). Course in General Linguistics, W. Baskin (trans.).
New York: Columbia University Press.
Eckardt, R. (2015). The Semantics of Free Indirect Discourse: How Texts Allow Us to
Mindread and Eavesdrop. Leiden / Boston, MA: Brill.
Fludernik, M. (1993). The Fictions of Language and the Languages of Fiction: The
Linguistic Representation of Speech and Consciousness. London: Routledge.
Friend, S. (2011). The great beetle debate: A study in imagining with names.
Philosophical Studies, 153(2), 183–211.
Friend, S. (2017). The real foundation of fictional worlds. Australasian Journal of
Philosophy, 95(1), 29–42.
García-Carpintero, M. (2018). Co-identification and fictional names. Philosophy and
Phenomenological Research, 101(1), 3–34.
Gendler, T.S. (2000). The puzzle of imaginative resistance. Journal of Philosophy, 97(2),
55–81.
Genette, G. (1980). Narrative Discourse: An Essay in Method. Oxford: Blackwell.
Greimas, A.J. (1987 [1970]). On Meaning, F. Collins and P. Perron (trans.).
Minneapolis, MN: University of Minnesota Press.
Hinterwimmer, S. (2017). Two kinds of perspective taking in narrative texts. In
Semantics and Linguistic Theory (Volume 27, pp. 282–301). Washington, D.C.:
Linguistic Society of America.
Hjelmslev, L. (1953 [1943]). Prolegomena to a Theory of Language. Baltimore, MD:
Indiana University Publications in Anthropology and Linguistics.
Hockett, C. (1960). The origin of speech. Scientific American, 89–96.
 11
Holton, R. (1997). Some telling examples: A reply to Tsohatzidis. Journal of

Pragmatics, 28, 625–628.
Jakobson, R. and Halle, M. (1956). Fundamentals of Language. The Hague: Mouton.
Kripke, S. A. (2011). Vacuous names and fictional entities. In Philosophical Troubles
(pp. 52–74). Oxford / New York: Oxford University Press.
Lewis, D. (1983 [1978]). Truth in fiction [with postscripts]. In Philosophical Papers
(Vol. 1, pp. 261–280). Oxford / New York: Oxford University Press.
Maier, E. (2015). Quotation and unquotation in free indirect discourse. Mind and
Language, 30(3), 345–373.
Maier, E. (2017). Fictional names in psychologistic semantics. Theoretical Linguistics,
43(1–2), 1–45.
Matravers, D. (2014). Fiction and Narrative. Oxford / New York: Oxford University
Press.
Meinong, A. (1960 [1904]). On the theory of objects. In R. Chisholm (Ed.), Realism
and the Background of Phenomenology (pp. 76–117). New York: Free Press.
Murdoch, I. (2003 [1957]). The Sandcastle. London: Vintage Books.
Nabokov, V. (2010 [1957]). Pnin. London: Penguin Books.
Pagin, P. (2014). Intersubjective intentional identity. In M. García-Carpintero and
G. Martí (Eds.), Empty Representation: Reference and Non-Existence (pp. 91–113).
Oxford: Oxford University Press.
Parsons, T. (1980). Nonexistent Objects. New Haven, CT: Yale University Press.
Parsons, T. (1982). Fregean theories of fictional objects. Topoi, 1, 81–87.
Priest, G. (2005). Towards Non-Being: The Logic and Metaphysics of Intentionality.
Oxford: Clarendon Press.
Recanati, F. (2000). Oratio Obliqua, Oratio Recta: An Essay on Metarepresentation.
Cambridge, MA: MIT Press.
Recanati, F. (2018). Fictional, metafictional, parafictional. Proceedings of the
Aristotelian Society, 118, 25–54.
Schlenker, P. (2004). Context of thought and context of utterance (a note on free
indirect discourse and the historical present). Mind and Language, 19(3), 279–304.
Searle, J.R. (1975). The logical status of fictional discourse. New Literary History, 6(2),
319–332.
Sharvit, Y. (2008). The puzzle of free indirect discourse. Linguistics and Philosophy,
31(3), 353–395.
Stock, K. (2017). Only Imagine: Fiction, Interpretation, and Imagination. Oxford and
New York: Oxford University Press.
Stokke, A. (2013a). Lying and asserting. Journal of Philosophy, 110(1), 33–60.
Stokke, A. (2013b). Protagonist projection. Mind and Language, 28(2), 204–232.
12     
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.
Wolterstorff, N. (1979). Characters and their names. Poetics, 8, 101–127.
Wood, J. (2008). How Fiction Works. New York: Farrar, Straus and Giroux.
Woodward, R. (2011). Truth in fiction. Philosophy Compass, 6(3), 158–167.
Zeman, S. (2016). Introduction: Perspectives on narrativity and narrative perspectivi-
zation. In N. Igl and S. Zeman (Eds.), Perspectives on Narrativity and Narrative
Perspectivization [Linguistic Approaches to Literature 21] (pp. 1–14). Amsterdam /
Philadelphia, PA: Benjamins.
Zeman, S. (2018). What is a Narration—and why does it matter? In A. Hübl and
M. Steinbach (Eds.), Linguistic Foundations of Narration in Spoken and Sign
Languages (pp. 173–206). Amsterdam / Philadelphia, PA: Benjamins.
PART I
TRUTH, REFERENCE, AND
IMAGINATION
The observation that fictional contexts are highly problematic for our most central
semantic notions, like truth and reference, runs through the history of semantics,
going back at least to Frege’s 1892 “Über Sinn und Bedeutung.” Since the Harry
Potter books are works of fiction, the name “Hermione,” as used in those books,
does not actually have a referent. Since, by compositionality, the truth value of a
sentence depends on the referents of its parts, it follows that any sentence
containing that name is neither true nor false. That includes fictional statements
taken directly from a work of fiction, but also so-called parafictional statements
that we might use in talking about fictions, as in (1).
(1) a. Hermione is kind of a wizard genius.

b. In the Harry Potter books, Hermione is kind of a wizard genius.
This is unsatisfactory: in the context of a discussion about the Harry Potter

universe, (1a) is arguably true, and (1b) is true regardless of context.
Lewis (1983 [1978]) famously provides a semantic account of the truth of
statements like these. The operator “in the Harry Potter books” in (1b) is analyzed
as a modal operator, quantifying over all possible worlds in which those books are
“told as known fact.” For Lewis, the unprefixed statement in (1a) is just an
abbreviation of the explicitly modalized statement in (1b). Lewis’s account is thus
a fully semantic account of the linguistic judgment that sentences like in (1) are true.
Concerning the kind of utterances that authors of fictions make, Searle (1975),
Currie (1990), and Walton (1990) argue for more pragmatic ways of understanding
our engagement with fictions and fictional characters. On their views, a statement in a
work of fiction is not an invisibly prefixed assertion but rather a pretend assertion
(Searle) or an instance of a sui generis speech act of fiction-making (Currie).
According to Walton, while assertions are intended to be believed, fictional state-
ments (and artworks more generally) are prescriptions to imagine.
Although the idea that fiction is fundamentally distinct from assertion in being
directed at imagination as opposed to knowledge or belief is still dominant in
14 , ,  
aesthetics and philosophy of language, it has recently been challenged by

Matravers (2014), who proposes to shift attention away from imagination and
the distinction between fiction and non-fiction toward a more fundamental
distinction between confrontation and representation.
All of the above accounts are so-called anti-realist accounts: fictional characters
like Hermione Granger do not actually exist, so fictional names do not actually
refer. Such accounts have a hard time dealing with metafictional statements like
(2) that are clearly not meant to be relativized to some fictional world, nor are they
straightforwardly understood as invitations to imagine:
(2) Hermione doesn’t really exist—she’s just a fictional character made up by

J.K. Rowling.
Instead, it seems that metafictional statements like (2) are unmodalized assertions.
This, however, reinstates Frege’s original problem: if the name “Hermione” is
fictional and does not actually refer, the sentence cannot be true. A Meinongian
solution would be to give up our ontological prejudice against fictional characters
and admit that Hermione does exist, perhaps in roughly the same way that the
number 37 exists, as an abstract object.
In Part I we collect five papers that address some combination of these
fundamental, semantic issues related to truth, reference, and imagination in
fictional, parafictional, and metafictional statements. In the following we briefly
introduce these contributions in order.
In “Fictional reference as simulation,” François Recanati defends a pragmatic,
Searle/Frege-style account of fiction as pretend assertion, and more specifically of
fictional reference as pretend or simulated reference. To make sense of metafic-
tional statements he also admits fictional characters into his ontology, leading to
what he calls an “oecumenical approach” to the semantics of fictional names: in
fictional works and parafictional statements fictional names pretend to refer to
flesh-and-blood entities, while in metafictional statements such names actually
refer, but to fictional characters. The former use is primary, and the fictional
characters merely supervene on a practice of fictional reference. Fictional refer-
ence in turn is parasitic on genuine reference, which it simulates. Recanati then
compares his hierarchical picture of reference to Matravers’s flat model where
fiction and non-fiction are treated on a par.
In “Sharing real and fictional reference,” Hans Kamp also takes up the problems
of the reference of fictional names. To formulate his proposal he offers an extension
of his Discourse Representation Theory (DRT) framework to model the dynamic
interpretation of fiction. A key innovation over earlier attempts in this direction is
the mechanism whereby fiction-internal anchors can be exported for metafictional
talk, which turns them into anchors referring to fictional characters as abstract
objects. This move makes Kamp’s proposal a close cousin of Recanati’s
, ,   15
oecumenical approach. Kamp partly motivates his account by demonstrating how

it can handle mixed parafictional and metafictional discourse, as in (3).
(3) At that point Gollum bites off Frodo’s finger. I think that was too much of a
Deus ex Machina. But I loved the end nonetheless.
In “Fictional truth: In defense of the Reality Principle,” Nils Franzén revisits

Lewis’s “Truth in Fiction” and examines its controversial reality principle, which
allows us to import facts from the actual world into our reading of a fictional work.
It is this principle that makes it true in The Lord of the Rings that humans breathe
oxygen, and that gold is insoluble in nitric acid, even though these facts are never
stated explicitly in the work. On Lewis’s original analysis, this principle is enforced
by analyzing fictional truth as counterfactual truth: φ is true in a fiction if and only
if: were the fiction told as known fact, φ would be true. The truth of that
counterfactual truth in turn is defined as truth in the fiction-compatible worlds
that are closest to the actual world. Over the years the principle has come under
attack from various sides (e.g. Badura and Berto 2018; Currie 1990), for instance
on the grounds that it seems to entail that it is true in The Lord of the Rings that
Trump is the forty-fifth U.S. president. Franzén offers a comprehensive overview
of the counterarguments and a refinement of Lewis’s formulation that avoids the
various pitfalls.
In “On the generation of content,” Sandro Zucchi likewise builds on Lewis’s
account of truth in fiction. Following Matravers’s thesis that fiction and non-
fiction are interpreted in essentially the same way, he extends the Lewisian
possible worlds account of fiction to a general account of the content of (fictional
and non-fictional) media. His leading example is the case of Robinson Crusoe,
originally published and read as a truthful memoir, but later revealed by the
author to be a work of pure fiction. Crucially, this apparent transition from
non-fiction to fiction did not in any way affect what was true according to the
story, supporting the idea that the same principles of content generation are at
play regardless of whether a given text is read as fiction or non-fiction. The
proposed semantic analysis of media content and the “according to” operator is
applied to a number of familiar puzzles including the reference of names and
indexicals in fiction, impossible fictions, and unreliable narration.
In “Do the imaginings that fictions invite have a direction of fit?,” Manuel
García-Carpintero details a normative, i.e. Austinian rather than Gricean, account
of fictionality, incorporating the basic Walton/Currie idea of fiction as a prescrip-
tion to imagine. The main selling point here is that this normative account allows
us to make sense of what he calls the “direction-of-fit asymmetry.” On the one
hand, fiction-making involves a speech act of prescribing or inviting certain
imaginings. As such, it has a telic or world-to-mind direction of fit, like commands
and desires (so called because whenever the content of a desire does not match the
16 , ,  
world, we are inclined to go change the world, not our desire). On the other hand,
the fiction-prompted imaginings themselves arguably have the opposite, a thetic
or mind-to-world direction of fit, much like assertions and beliefs (whenever the
content of a belief does not match the world, it’s the belief that should change, not
the world). García-Carpintero shows how a properly normative approach can
make sense of this asymmetry observation.
References
Badura, C. and Berto, F. (2018). Truth in fiction, impossible worlds, and belief revision.
Australasian Journal of Philosophy, 97(1), 1–16. https://doi.org/10.1080/00048402.
2018.1435698.
Currie, G. (1990). The Nature of Fiction. Cambridge: Cambridge University Press.
Lewis, D. (1983 [1978]). Truth in fiction [with postscripts]. In Philosophical Papers
(Vol. 1, pp. 261–280). Oxford and New York: Oxford University Press.
Matravers, D. (2014). Fiction and Narrative. Oxford: Oxford University Press.
www.oxfordscholarship.com/view/10.1093/acprof:oso/9780199647019.001.0001/
acprof-9780199647019.
Searle, J.R. (1975). The logical status of fictional discourse. New Literary History, 6(2),
319–332. https://doi.org/10.2307/468422.
Walton, K.L. (1990). Mimesis as Make-Believe: On the Foundations of the
Representational Arts. Cambridge, MA: Harvard University Press.
2
Fictional reference as simulation
François Recanati
2.1 Fictional names and their uses
In a text of fiction a variety of prima facie referential devices (e.g. proper names,
pronouns, demonstratives, etc.) are used to talk about the protagonists of the story.
In some cases a fictional text features a protagonist who exists in the actual world
(e.g. Napoleon in War and Peace). It is plausible that genuine reference occurs in
such cases: the name ‘Napoleon’ in War and Peace arguably refers to the real
Napoleon, but what is said of him is said in the fictional mode; that is, nonseriously.
The speaker makes a fictitious assertion, based on pretence, even though the ancillary
act of reference is genuine.¹ In other cases, however, the protagonists are themselves
‘fictional characters’ who do not exist in reality but only ‘in the world of the fiction’.
With respect to fictional characters, there are two options. Either we maintain
that genuine reference occurs, the target of the referential act being not a flesh-
and-blood individual (as in the case of Napoleon) but the ‘fictional character’
construed as some kind of abstract object: a merely possible individual, or an
intentional entity (an object of thought), or a cultural artefact created by the
author of the fiction. This is the realist option, which enriches the ontology with a
special kind of object, namely fictional characters, alongside ordinary individuals.
Or—second option, that which I favour—we deny that the referential act is
genuine in such cases: the fictitiousness affects not only the overall assertion, as
in the Napoleon case, but also the ancillary act of reference.
Both views have been developed in the literature. I myself have argued for an
oecumenical view according to which both types of reference may occur (genuine
reference to fictitious objects, and fictitious reference to ordinary objects), corres-
ponding to different uses of fictional names such as ‘Sherlock Holmes’ (Recanati
2018).² Fictional uses, i.e. the uses of such names in the fictional text itself, are
¹ Some authors think that even the act of reference is non-genuine in such cases. Thus Frege writes:
‘If Schiller’s Don Carlos were to be regarded as a piece of history, then to a large extent the drama would
be false. But a work of fiction is not meant to be taken seriously in this way at all: it’s all play. Even the
proper names in the drama, though they correspond to names of historical personages, are mock proper
names; they are not meant to be taken seriously in the work’ (‘Logic’, in Frege 1979, 130; emphasis mine).
² I am not the first one to put forward such an oecumenical view. See e.g. Kripke (2013),
Zalta (2000).
François Recanati, Fictional reference as simulation In: The Language of Fiction. Edited by: Emar Maier and Andreas Stokke,
Oxford University Press. © François Recanati 2021. DOI: 10.1093/oso/9780198846376.003.0002
18  
instances of fictitious (non-genuine) reference to an ordinary flesh-and-blood

individual. The individual in question (Sherlock Holmes) does not exist, and
there cannot be genuine reference to something that does not exist; but it is
possible to pretend that the object exists, and to ‘refer’ to it under that pretence.
This is what happens in fiction when the protagonist talked about is a fictional
rather than a real individual. But there are also instances of genuine reference to
fictional characters (construed as abstract entities rather than as ordinary flesh-
and-blood individuals). Such instances occur in talk about the fiction.
Consider the following statement:
(1) Sherlock Holmes is a fictional character created by Conan Doyle in 1887.
We can establish that the name ‘Sherlock Holmes’ in (1) refers to a cultural
artefact by means of the following argument:
• Statement (1) is true.

• If (1) is true, then the name ‘Sherlock Holmes’ in (1) refers to something
(otherwise the statement would be neither true nor false).
• What the name ‘Sherlock Holmes’ refers to in (1) is something which Conan
Doyle created in 1887, hence some kind of artefact (otherwise the statement
would be false rather than true).
• The artefact in question is not a concrete thing; it belongs to the domain of
ideas.
This argument can be blocked at various steps by denying one of the premises.
But all of the premises sound plausible enough, and the argument provides at least
prima facie justification for the claim that ‘Sherlock Holmes’ in (1) refers to a
cultural artefact, which I take to be a kind of abstract entity (Thomasson 1999).
This type of argument, however, only applies to so-called ‘metafictional’ uses of
fictional names, such as (1). It evidently does not apply to uses of fictional names
in fiction—fictional uses, such as (2) below—nor does it apply to so-called
‘parafictional’ uses, like (3):
(2) Sherlock Holmes shook his head and lit his pipe.
(3) In Conan Doyle’s novels, Sherlock Holmes is a detective who solves mys-
teries and whose adventures are narrated by his friend Dr Watson.
In the case of (2), the argument does not apply because the first premise (to the
effect that the statement is true) wouldn’t be plausible all. A fictional statement
involving a fictional name, like (2), does not count as true or false. So one cannot
argue from the fact that (2) is true to the conclusion that the name ‘Holmes’ in (2)
    19
must refer to something. In the case of (3), the argument does not apply either,
despite the fact that the first premise, in that case, would be plausible. (Statement
(3) seems to be true, just like statement (1).) That is because the third premise (to
the effect that what the name ‘Holmes’ refers to must be an artefact, in view of
what is truly predicated of it) would be false: what is truly predicated of Holmes in
(3) is that he is a detective who solves mysteries etc. It is not true that anything of
which these things are truly predicated must be an abstract artefact rather than a
flesh-and-blood individual.³
I conclude that a prima facie case can be made for the view that the name
‘Holmes’ in (1) refers to an abstract artefact, while no such case can be made for
either (2) or (3). In the case of (2), it is very tempting to go for the nonrealist
alternative: the name ‘Holmes’ there does not genuinely refer to anything; it only
pretends to refer to a flesh-and-blood individual. This is fictitious reference to an
ordinary object, rather than genuine reference to a fictitious object.⁴ The oecu-
menical view accepts that there are these two types of use of fictional names, and
considers the fictional use as basic. The practice of fiction, based on pretence, is
what gives birth to the abstract artefacts which supervene on it and can in turn be
referred to in metafictional sentences such as (1). As Schiffer writes, fictional
entities are ‘abstract entities whose existence supervenes on the pretending use of
words’ (2003, 52). Or Searle: ‘It is the pretended reference which creates the
fictional character’ (1975, 330). Or Kripke: ‘A fictional character [ . . . ] is an
abstract entity. It exists in virtue of more concrete activities of telling stories,
writing plays, writing novels, and so on [all of which involve pretence]’ (2013: 73).
The real difficulty for the oecumenical view comes from parafictional uses like
(3), which share features with both fictional and metafictional uses. I’ll return to
that issue in sections 2.4 and 2.5. Meanwhile, I want to discuss another, related
debate in the philosophy of fiction, concerning what differentiates fictional talk
from serious talk.
2.1.1 Matravers against the simulation view
Following a venerable tradition, I have claimed that fictional uses of referring

expressions rest on pretence. The author of the fiction pretends to make asser-
tions, i.e. to report facts of which s/he has knowledge (Lewis 1978). When the
³ As Merel Semeijn pointed out to me, one might argue that what is truly predicated of Holmes in
(3) is that in Conan Doyle’s stories he is a detective etc. Even so, the problem of the ‘wrong kind of
object’ (as Klauk 2014 calls it) remains: in the Holmes stories, it is a flesh-and-blood individual, not an
abstract object, who solves mysteries.
⁴ If there is such a difference between the use of the name ‘Sherlock Holmes’ in (1) and (3), why do
we have the sense that it refers to the same fictional character in both? For a tentative answer, based on
the idea of a two-sided mental file, see Recanati (2018).
20  
protagonists of the fiction are themselves fictional characters, the pretend asser-
tion rests on an ancillary act of pretend reference. One of the merits of the pretence
approach is that it is compatible with a demanding notion of reference, such as the
view that genuine reference requires some form of acquaintance with the referent
(Recanati 2012). On this view one cannot refer to something that does not exist, or
of which one has only descriptive knowledge; but one can pretend to refer to such
things, and that is what happens in fiction.
On the oecumenical view I have described, the fictional use of referring
expressions is more basic than the metafictional use, through which we refer to
the abstract entities which supervene on the fictional uses; but the regular use of
referring expressions in acts of genuine reference (to things with which we are
acquainted) is itself more basic than the fictional use. The fictional uses mimic
the regular uses, in such a way that fictional reference is parasitic on genuine
reference. This is exactly parallel to what happens in the case of assertions.
Pretend assertions (such as those made by the author of a fiction) mimic genuine
assertions. Whichever norms regulate genuine assertions are tacitly invoked in
pretend assertions: whoever pretends to assert pretends that the normative
conditions of genuine assertion are satisfied. Likewise, whoever pretends to
refer pretends that the normative conditions of genuine reference are satisfied.
Let us call the idea that fictional reference and fictional assertion are parasitic on
genuine reference and genuine assertion, which they simulate, the asymmetric
dependency thesis about fiction (ADT). The reason I need to give it a name is that
this idea has been called into question recently. That is the second debate I want
to say something about.
Some philosophers of fiction seem to repudiate ADT. Thus Matravers (2014)
rejects the Waltonian idea that it is characteristic of fiction to mandate a special
kind of attitude on the part of the consumer, viz. make-believe, pretence, or
imagination, with that attitude itself being characterized as the simulation of
ordinary credal attitudes such as belief. According to Gregory Currie, whom
Matravers takes to be a representative of the view he rejects, reading a non-
fictional text, say a newspaper article, normally yields belief (or possibly disbelief ),
but reading a fictional text has a different effect:
We do not acquire from them beliefs in the straightforward way that we acquire
beliefs from nonfiction. With fictions, our mental processes are engaged off-line,
and what we acquire instead of beliefs is imaginings which simulate belief.
(Currie 1995, 148)
Imaginings themselves are not on a par with beliefs. They are simulated beliefs,
where simulation is an operation (the operation of ‘running off-line’) which
‘sever[s] the connections between our mental states and their perceptual causes
and behavioural effects’ (Currie 1995, 149):
    21
Simulation transmutes beliefs into imaginings. Just as a belief and a desire may
have the same content but differ functionally, so may a belief and an imagining.
Believing that it will rain has certain connections to perception and behaviour
which, if they are severed, transmutes the belief into a case of imagining that it
will rain. (Currie 1995, 149)
Matravers rejects the whole picture. The alleged disconnection from perception
and behaviour is not characteristic of fiction, he points out, but of representation
more generally: ‘it is true that horror movie viewers do not typically flee the
cinema screaming, but neither do documentary viewers, television news viewers,
or any other kind of viewer’ (Matravers 2014, 38). Matravers contrasts represen-
tation with confrontation.⁵ There is confrontation whenever we directly experi-
ence a state of affairs, to which we can react online. Often, however, what we
encounter is not directly a state of affairs given in experience but the representa-
tion of a state of affairs. When we read a newspaper article, the situation it
describes is not directly experienced, but represented. Understanding a represen-
tation is a matter of building a mental model of the situation it describes, and that
is something we do whether the representation is fictional or not. If this building
of a mental model is called imagination, then imagination is involved both in
processing fiction and non-fiction:
The distinction between confrontations and representations is more funda-

mental than the distinction between non-fictions and fictions. Confrontations
do not require the imagination; I do not need to imagine being confronted by
a wolf if there is one in front of me. Something is needed to explain my
engagement with representations, [however]. If philosophy does need some
notion of [ . . . ] ‘make-believe’, it applies to this category rather than only to
fictions. (Matravers 2014, 53)
Matravers argues for ‘a “two-stage” model of engaging with representations’,

where ‘the first stage is neutral between non-fictional and fictional representa-
tions: we build a mental model of the representation that is compartmentalized
but not isolated from our pre-existing structure of belief ’ (Matravers 2014, 90).
Fiction and non-fiction are only differentiated in the second stage, when we relate
the representation to our pre-existing structure of belief:
⁵ ‘The difference between a confrontation relation and a representation relation aligns with the
difference between situations in which our mental states are online and situations in which our mental
states are offline. That is, in confrontation relations our mental states are caused by perceptual inputs
from the objects of those states, and cause actions towards objects in our egocentric space. In
representation relations our mental states are not caused by perceptions of the objects of those states,
and do not result in actions towards objects in our egocentric space (although, of course, they can still
cause actions)’ (Matravers 2014, 50).
22  
My objection to Currie is that his functional description [ . . . ]—non standard

inputs and an absence of motivation as an output—does not apply only to fiction
but to our engaging with representations generally. There is a further question of
the relation between this activity and our pre-existing structure of belief. That is
not to do with engaging in simulation, but with the result of engaging in
simulation. Put very roughly, simulating fiction scenarios does not result in our
forming beliefs and simulating non-fictional scenarios does result in our forming
beliefs. (Matravers 2014, 27)
The two-stage model advocated by Matravers is reminiscent of the traditional

Fregean picture, according to which an utterance, whether serious or fictional,
expresses a thought (first stage), which may or may not be asserted, that is, put
forward as true (second stage). Only in the case of a serious utterance is the
thought asserted and the hearer intended to believe it. On a Fregean interpretation
of Matravers’ two-stage model, the first stage corresponds to the process of
grasping the thought, entertaining it. It is common to serious and fictional
statements, and if it involves the imagination, then imagination is involved in
both cases. The second stage is what Frege calls ‘the step from sense to reference’.
(For Frege, the thought it expresses is the sense of an utterance, while its truth-
value—true or false, as the case may be—is its reference.) The step from sense to
reference is not taken in the case of fiction: the thought which is expressed is not
asserted; it is not taken to be true (or false), but is merely entertained. The same
thing happens at the level of the referring expression which is a constituent of the
fictional sentence. The referring expression has a sense (an individual concept) but
through that sense no reference is made to an actual individual falling under that
concept in the fictional case. Only when the reference is genuine rather than
fictional is the step from sense to reference taken.
On the Fregean interpretation, the two-stage model is indeed incompatible with
ADT. ADT rests on the following two theses, as we have seen:
(i) it is characteristic of fiction to mandate a special kind of attitude on the part of

the consumer, viz. imagination rather than belief (Walton 1990)
(ii) that attitude itself is the simulation of belief (Currie 1995).
Proponents of the two-stage model reject (i) because, as Matravers puts it, fiction
and non-fiction alike are representations, and to understand a representation one
has to imagine what it represents—one has to build a ‘mental model’ of its content.
This is not characteristic of fiction as opposed to non-fictional texts. As for (ii), it is
incompatible with the Fregean claim that grasping a thought, entertaining it (or,
as I have just put it, ‘imagining what it represents’), is a simpler act of the mind
than judging the thought to be true. To judge a thought to be true, one must grasp
it. Grasping/entertaining the thought is a simple act of the mind, while judging/
    23
asserting is more complex: it involves grasping/entertaining plus ‘endorsing’. This

point has been made quite explicitly by Scott Soames, a staunch advocate of the
Fregean picture:
Some [cognitive acts]—judging and asserting that o is red—involve further

cognitive acts in addition to predicating redness of o, whereas others—seeing
and imagining—do not. To judge or assert that o is red is to think of o as red and
to do something else. In the case of judging, this something else is endorsing, in
the sense of adopting that way of thinking—of o as red—as potential basis for
further thought or action. (Soames 2014, 228–229)
Acts of judgement or denial are, for Soames, more complex than simple acts of
entertaining, since they involve an additional act of endorsement or (in the case of
denial) rejection. In contrast, imagining consists of a single act of entertainment,
and so do, according to Soames, mental acts such as seeing and visualizing
(Soames 2014, 229). Fiction works the same way, according to the two-stage
model (on its ‘Fregean’ interpretation). Propositions are entertained, and the
subject imagines what it would be like for them to be true, but no endorsement
takes place: the content of the representation is not adopted as belief.
The contrast between the two views is reasonably clear. According to ADT,
belief (or judgement, or assertion) is primary, while fiction involves ‘make-believe’
(or imagination), which is the simulation of belief, hence a more complex cogni-
tive act. According to the Fregean picture, what is primary is grasping/entertain-
ing, which works in the same way for fictional and non-fictional discourse and
possibly involves the imagination (understood as the construction of a mental
model). Serious assertions involve an additional act of endorsement of the pro-
positions entertained. No such act takes place when the discourse is fictional.
2.2 Force cancellation
Recently, there has been a sustained attack on the Fregean picture and the force/
content distinction on which it is based. An act of assertion has a certain content
(what is asserted) but it also has ‘assertive force’. According to the Fregean picture,
these components can be separated: besides being asserted, the content can be
expressed or entertained in a forceless manner, i.e. without judging the propos-
ition to be true or asserting it. There is a sense in which this is fairly uncontro-
versial, but there is also a sense in which it isn’t. According to Peter Hanks (2007,
2015), propositions themselves, e.g. the proposition that o is red (Soames’
example), are inherently assertive, because what holds together the object talked
about (o) and the property ascribed to it (redness) is the fact that the property is
predicated of the object. Pace Soames (2010, 2015), this notion of predication is
24  
not neutral and forceless. If you predicate redness of o, you ascribe that property to
the object. To be sure, the proposition that o is F can be expressed without being
actually asserted, e.g. when it occurs in the antecedent of a conditional. But this
(the so-called ‘Frege point’) does not object to there being an assertive component
inherent in the proposition, contrary to what Peter Geach is often said to have
claimed (1960, 1965). When a proposition occurs unasserted, Hanks argues, the
assertive ingredient is cancelled:
Frege’s main reason for adopting the content–force distinction—the fact that we
do not assert the antecedent or the consequent in an utterance of a conditional—
is consistent with thinking that an assertoric element is included in the contents
of declarative sentences. Frege’s reaction to this fact about conditionals was to
hold that the contents of declarative sentences are devoid of any assertive
element, but this is not the only reaction one might have. An alternative is to
hold that in certain contexts, for example, when a sentence is used inside a
conditional, the assertive element is cancelled by the presence of the conditional.
(Hanks 2011, 15)
In fictional texts, clearly, propositions occur without assertive force. For example,
the actor who says that p on stage does not seriously assert that p: she merely
pretends to assert that p. But this can be analysed by saying that the assertive force
inherent in the proposition that p is cancelled because the utterance takes place in
a theatrical context. Theatrical contexts are, by convention, contexts in which
force is cancelled (Hanks 2015, 96n). As Dummett puts it, ‘The reason [the actor]
is not making assertions is not that he is doing less than that—merely expressing
thoughts, say—but that he is doing more than that—he is acting the making of
assertions’ (1973, 311). Cancellation results from the fact that the speaker is
staging or simulating the performance of an assertion, rather than genuinely
going through the performance.
On the Fregean picture defended by Soames, forceful cases such as assertion or
judgement result from an ‘endorsement’ operation that comes in addition to the
basic act of entertaining the proposition. On the alternative picture defended by
Hanks, it is forceless occurrences which involve a supplementary operation:
cancelling. Table 2.1, adapted from Recanati 2019, summarizes the difference
between the two views.⁶
⁶ A similar contrast is drawn by Gilbert (1991) between two models of belief-forming mechanisms
such as perception—the ‘Cartesian’ model and the ‘Spinozan’ model. According to both models, belief
fixation proceeds in two stages, a ‘representation’ stage and an ‘assessment’ stage (Gilbert 1991, 109).
On the Cartesian model, acceptance only occurs at the assessment stage. On the Spinozan model, the
representation stage involves automatic acceptance of the represented content, which acceptance may
be cancelled or certified at the assessment stage. Gilbert defends the Spinozan model, and so do Egan
(2008) and Mandelbaum (2014) after him.
    25
Table 2.1 The two pictures
Basic case Supplementary operation
The Fregean picture Entertaining (forceless) Endorsement (distinguishes

serious talk from fiction)
The alternative picture Assertion/judgement Force cancellation (distinguishes
(forceful) fiction from serious talk)
If we give up the Fregean picture and adopt the alternative picture instead, as
I recommend, do we have to reject Matravers’ two-stage model? No—but we need
to reinterpret it in a non-Fregean way. We can maintain that there are two levels:
the ‘basic’ level, which is the common core shared by serious talk and fictional talk,
and the second level where serious talk and fictional talk come to be differentiated.
According to the Fregean picture, the basic level corresponds to the (forceless) act
of entertaining a proposition, while the second level features the additional act of
endorsement which distinguishes serious talk from fictional talk. According to the
alternative view, the basic level corresponds to the (forceful) act of asserting or
judging, or the forceful state of believing, a proposition, while the second level
features an operation of ‘decoupling’, or ‘running off-line’, which cancels the force
of the state or act (Recanati forthcoming a, b). (Because of that operation, the state
or act in question is merely simulated.) So the two-stage model does not have to be
repudiated. The alternative view only forces us to acknowledge the presence of an
assertive component at the basic level.
I accept, indeed welcome, Matravers’ idea that understanding a discourse,
whether fictional or not, involves imagining the situation it represents, i.e. con-
structing a mental model for it.⁷ But, like Gilbert (1991), I take that process to be a
form of temporary acceptance: it consists in (provisionally) accepting the exist-
ence of certain elements corresponding to the referring expressions in the dis-
course, and ascribing to them certain properties and certain relations to other
elements, corresponding to what is predicated of them in the discourse. This
amounts to viewing things as being thus and so. Thus, if I hear or read ‘it is
raining’ (or ‘winter is coming’), I put myself in the sort of mental state I am in
when I judge or believe that it is raining (or that winter is coming). That state is
inherently assertive/judicative, even though the assertive/judicative force inherent
in it may be cancelled if it turns out that the speaker is not serious. The
intrinsically assertoric character of what goes on at the first level (common to
fictional talk and serious talk) explains why entertaining fictional scenarios carries
a certain emotional load and generates certain action tendencies. What mitigates
the fictional emotions (and distinguishes them from real emotions) is the
⁷ See the quotation from Rumelhart in Recanati 2004, 139–140; and the discussion of Bühler in
Recanati 2000, 88ff.
26  
cancellation operation at the second level, which also has the effect of inhibiting
the action tendencies generated at the first level.⁸
Corresponding to the two levels, we have to distinguish two notions of imagin-
ation or simulation: a first-level notion (imagination₁) and a second-level notion
(imagination₂). This makes the debate initiated by Matravers terminological to a
certain extent. At the first level, when we build a mental model for a discourse,
we can say (as I did several times) that we imagine the situation it represents
or (as Matravers puts it) that we ‘simulate scenarios’. Matravers is right that
imagination/simulation in that sense is not characteristic of fiction as opposed
to non-fictional discourse. Fiction comes to be differentiated from serious talk
only at the second level, when the assertive force present at the first level is
cancelled. In that case we can characterize the subject’s attitude as that of ‘mere’
imagination—that is, imagination rather than belief—and the acts of assertion or
reference as being simulated rather than genuinely performed. It is this contrast-
ive, second-level notion of imagination/simulation (imagination₂) that propon-
ents of ADT such as Currie have in mind when they say that fiction is essentially
simulative and involves imagination rather than belief. Properly understood,
therefore, ADT is compatible with the two-stage model (on its non-Fregean
interpretation). We can simultaneously accept that
(ADT)
• it is characteristic of fiction to mandate a special kind of attitude on the part
of the consumer, viz. imagination₂ rather than belief
• that attitude itself is the simulation of belief ⁹
and that
(Two-stage model)
• Understanding a representation is a matter of imagining₁ the situation it
describes; that is, building a mental model of it
• that is something we do whether the representation is fictional or not.
2.3 Parafictional statements: a problem

for the simulation approach
Merel Semeijn has elaborated a Stalnakerian analysis based on Matravers’ idea

(Semeijn 2017; Maier and Semeijn, this volume). In processing both fictional and
⁸ The simulation view thus provides a neat solution to the so-called ‘paradox of fiction’.
⁹ Recall Currie’s passage cited in section 2.2: ‘Simulation transmutes belief into imaginings . . .
Believing that it will rain has certain connections to perception and behaviour which, if they are
severed, transmutes the belief into a case of imagining that it will rain’ (Currie 1995, 149).
    27
non-fictional texts, she says, we start by constructing a mental model, which for
Semeijn means that incoming information from the text is used to update a
‘temporary workspace’ that corresponds to the text and exists alongside the
(permanent) Stalnakerian common ground. The latter, which she refers to as
the ‘official common ground’,¹⁰ corresponds to Matravers’ ‘pre-existing system of
beliefs’:
What differentiates assertions from fictional statements is how, at the end of the
discourse, they update the official common ground: whether the content of the
updated workspace is adopted as belief (for nonfiction) or as metafictive belief
(for fiction).¹¹ Hence, in the second step of the algorithm, I define assertions
and fictional statements as different ‘closure operations’ that take an ordered pair
<C,W> containing [an official common ground and] an updated, active work-
space, and return an ordered pair with a new official common ground and an
inactive workspace. (Semeijn 2017, 420)
In the second stage of processing serious discourse, the content of the temporary
workspace is adopted as belief, Semeijn says—it is used to update the common
ground. If the discourse is fictional, the content of the temporary workspace is not
adopted as belief; rather, it is embedded under a Lewisian ‘in the fiction’ operator
(e.g. ‘in the Conan Doyle stories’), and the content resulting from embedding is
adopted as parafictional belief and fed into the common ground. In a nutshell:
entertaining the proposition that it is raining in a non-fictional context gives rise to
the belief that it is raining (assertive closure), while entertaining the same propos-
ition in a fictional context gives rise to the belief that in the fiction, it is raining
(fictive closure). Either way, the content entertained in the first stage is used, in the
second stage, to update the common ground, directly (the non-fictional case) or
indirectly (the fictional case, where the content is fed into the common ground after
embedding).
If we opt for the simulation approach, however, fictive closure raises a problem.
Parafictional statements like (3), repeated here, seem to express genuine beliefs,
endowed with a truth-value:
¹⁰ Some authors, e.g. Eckardt (2015) and Stokke (2013, 2018), distinguish the common ground
corresponding to what the conversational protagonists mutually presuppose concerning the actual
world, from various additional (‘unofficial’) common grounds corresponding to what is mutually
presupposed concerning particular fictions (in the case of the Conan Doyle stories: that Holmes is a
detective, that Watson is his friend, etc.). Semeijn’s temporary workspace is meant to substitute for the
unofficial common grounds. While, at least for Eckardt, the unofficial common grounds corresponding
to various fictions are supposed to exist permanently alongside the official common ground, Semeijn’s
temporary workspace is only active during the processing of a fictional text. Semeijn’s workspace is
similar to the ‘possible world box’ posited by Nichols and Stich in their theory of pretence (Nichols and
Stich 2000, 2003).
¹¹ ‘Metafictive belief ’ is (or, rather, was) Semeijn’s name for what I call parafictional beliefs such as
‘In the Conan Doyle novels, Holmes is a detective’.
28  
(3) In Conan Doyle’s novels, Sherlock Holmes is a detective who solves mys-
teries and whose adventures are narrated by his friend Dr Watson.
The problem is that the sentence embedded under the Lewisian operator contains
fictional names, ‘Sherlock Holmes’ and ‘Dr Watson’. According to the simulation
view, these names do not actually refer to anything. Reference is merely simulated:
the author does as if he (or the narrator whose role he is playing) was referring to
particular individuals using these names. The names themselves are not genuine
proper names—they are pretend proper names, invented by the fictional writer.¹²
But if that is so, then the embedded sentence fails to be either true or false. It fails
to express the singular proposition which it purports, or pretends, to express. How
then is it possible that embedding that ‘proposition’ under a modal operator will
yield a truth-evaluable content?
Frege famously held that even if a name lacks a reference (as fictional proper
names do), they still have a sense, so the sentence in which they occur expresses a
thought even if the thought in question is neither true nor false (because the step
from sense to reference is not taken). Regarding the sense of the fictional name
‘Nausicaa’, Frege writes:
The name ‘Nausicaa’ [ . . . ] does not mean or name anything [ . . . ], but it behaves
as if it names a girl, and it is thus assured a sense. And for fiction the sense is
enough. (‘Comments on Sinn und Bedeutung’, Frege 1979, 122; emphasis mine)
This passage fits rather well with the two-stage approach discussed in the previous
sections. The mental model constructed in the first stage of processing a sentence
containing the name ‘Nausicaa’ features a certain girl to whom the name refers.
Since the sentence occurs in fiction, reference to that girl is merely simulated: it is
pretend reference—the girl does not actually exist. That is why the sentence does
not have a truth-value. But there being an individual in the mental model is
sufficient to endow the name with a sense, even though it lacks (genuine)
reference.¹³
Again, this raises the issue: how can the parafictional sentence be true if the
embedded sentence is neither true nor false because reference is merely simulated?
Frege’s likely response is that in certain contexts a name refers not to its ordinary
reference (which, in the present case, does not exist) but to its ordinary sense. The
Fregean idea of a shift in semantic value in certain contexts seems almost
inevitable if one holds the simulation view while accepting that a parafictional
¹² As Russell once wrote, ‘The fundamental falsehood in the play [Hamlet] is the proposition: the
noise “Hamlet” is a name’ (Russell 1940, 277).
¹³ When I say that there is an individual in the mental model, I simply mean that the individual is
represented; I do not mean that it exists. So the issue of the ‘metaphysical status’ of that individual
(raised by several readers of this chapter) does not arise.
    29
sentence like (3) can be true. But that means that we give up the principle of
semantic innocence: the name ‘Sherlock Holmes’ in (3) does not do the same thing
as it does in the fictional sentence which (3) embeds.¹⁴
In a fictional sentence, the name ‘Nausicaa’ is a mock proper name which
pretends to refer; since this is merely pretence, the fictional sentence itself is
neither true nor false. As Frege says,
If the sense of an assertoric sentence is not true, it is either false or fictitious, and
it will generally be the latter if it contains a mock proper name.
(‘Logic’, Frege 1979, 130)
In a footnote, Frege adds: ‘We have an exception where a mock proper name
occurs within a clause in indirect speech’, because in such contexts a name refers
to its ordinary sense, and as we have seen, even a mock proper name has a sense. If
we apply the same strategy to the case of parafictional statements, we can explain
how such statements can be true; but as I said, that means that we renounce the
idea that the name ‘Sherlock Holmes’ in (3) merely simulates reference, as it does
in the fictional statement which (3) embeds.
Many authors want to preserve semantic innocence, however. They are radical
simulationists: they maintain that the use of the fictional proper name in (3) is ‘a
continuation of the pretence’ at work in the fictional statement which (3) embeds
(Evans 1982, 365). And they extend this simulation-based analysis to the occur-
rences of blatantly empty proper names in indirect speech and attitude reports.¹⁵
How, then, can this position be made compatible with the fact that parafictional
statements (and belief sentences alike) can be true or false? It is not obvious that it
can, and radical simulationists are sometimes tempted by the view that parafic-
tional statements are not really, not literally, true or false: taken at face value, they
themselves are only pretend assertions, but such pretend assertions pragmatically
convey something which is true or false. As Crimmins says, ‘Somehow, the context
¹⁴ Semantic innocence is so called (after Barwise and Perry 1981) in reference to a passage in which
Davidson criticizes Frege’s idea that indirect discourse and attitude ascriptions involve a shift in
semantic value: ‘If we could recover our pre-Fregean innocence, I think it would seem to us plainly
incredible that the words “The Earth moves”, uttered after the words “Galileo said that”, mean anything
different, or refer to anything else, than is their wont when they come in different environments’
(Davidson 1968, 108).
¹⁵ ‘Suppose an interpreter finds an expression—say, “Mumbo-Jumbo”—which functions, syntactic-
ally, like other expressions which he can construe as names, but for which he can find no bearer, and
reasonably believes there is no bearer . . . In practice, an interpreter might say things like “This man is
saying that Mumbo-Jumbo brings thunder”, and might explain an utterance which he described that
way as expressing the belief that Mumbo-Jumbo brings thunder. [ . . . ] Such an interpreter is simply
playing along with his deluded subject—putting things his way’ (McDowell 1977, 124–127; emphasis
mine). In other words, the interpreter simulates the deluded subject’s failed act of reference to the
nonexistent entity.
30  
of pretending allows me to generate with a pretend assertion of one sort of claim a

genuine, serious assertion of a different sort of claim’ (1998, 4).
Maier and Semeijn do not have this problem because, like David Lewis, they
embrace a form of descriptivism. For Maier (2017a), the content of the imagining
triggered by a fictional statement featuring a fictional name such as ‘Holmes’ or
‘Frodo’ is an existentially quantified proposition rather than a (mock) singular
proposition. Maier writes:
On my account, the name Frodo triggers the existential presupposition that there
is someone by that name, and this presupposition gets bound or accommodated
within the imagination. As far as purely fictional statements are concerned, the
resulting interpretation I predict is roughly equivalent to classical descriptivist
analyses [ . . . ]: in all imagination worlds there is some hobbit named Frodo,
who . . . There is no rigid designation here. (Maier 2017b, 115)
It follows that the proposition expressed by the sentence embedded under ‘In the
Conan Doyle novels’ in (3) is itself an existentially quantified proposition rather
than a (mock) singular proposition. What (3) really means, according to Maier
and Semeijn, is
(3’) In Conan Doyle’s novels, there is an individual x named ‘Sherlock Holmes’

and an individual y named ‘Dr Watson’, such that x is a detective who solves
mysteries and y is x’s friend and y narrates x’s adventures.
On this view, no empty reference takes place in the scope of the Lewisian operator,
so the problem parafictional statements raise for the simulation view does not
arise.
Maier’s descriptivist approach would not be very attractive as a general theory
about the semantic contribution of proper names, since proper names are usually
considered as referential and (therefore) rigid. But Maier restricts it to fictional
proper names:
Regular proper names of course do refer rigidly to their bearers, by virtue of the
mechanisms of internal and external anchoring. [ . . . ] My uniform analysis of
fictional and other names consists in the uniform treatment of all names as
presupposition triggers, not in associating truly referential readings to all occur-
rences of names. (Maier 2017b, 115n)
Against Maier I argued for a more fully uniform analysis, according to which
fictional proper names themselves should be construed as ‘referential’ and involv-
ing a mechanism of fictitious anchoring (Recanati 2017; see also Kamp, this
volume). That is the simulation view. Still, to solve the problem parafictional
    31
statements raise for that view, one might be tempted to appeal to Maier’s descrip-
tivist analysis construed, in a more limited setting, as an implementation of the
Fregean strategy specifically concerning parafictional uses.¹⁶ According to the
resulting proposal, while the proper name in the fictional statement is a mock
proper name that pretends to refer, its status changes when it occurs in the scope
of a parafictional operator—in such contexts the name’s role is to contribute an
individual concept rather than an individual. This, of course, is incompatible with
the principle of semantic innocence.
To sum up, there is a prima facie inconsistent triad:
(1) (Simulation) In a fictional statement, a fictional name such as ‘Holmes’ or

‘Frodo’ only pretends to refer. As a result, a fictional statement involving
such a name is neither true nor false.
(2) (Semantic Innocence) A fictional name such as ‘Holmes’ or ‘Frodo’ does the
same thing in a fictional statement (‘Frodo is F’) and in the corresponding
parafictional statement (‘In The Lord of the Rings Frodo is F’).
(3) (Truth) A parafictional statement such as ‘In The Lord of the Rings Frodo is
F’ can be true or false.
Radical simulationists want to preserve both Simulation and Semantic Innocence,

at the cost perhaps of giving up Truth. Simulation and Truth can be held together
at the cost of giving up Semantic Innocence. Giving up Simulation makes it
possible to preserve Semantic Innocence and Truth. But can we do better? Can
we preserve Simulation, Truth, and Semantic Innocence? My answer, briefly
sketched in the next section, is that we can.
2.4 Exemplifying pretence for demonstrative purposes
I think it is possible to maintain both radical simulationism (the view that fictional
names pretend to refer, both in fictional and parafictional statements) and the view
that parafictional statements can be true or false. The general idea is that the
parafictional speaker engages in pretence (e.g. pretends to refer to Sherlock
Holmes and to predicate properties of ‘him’) but does so in order to show what the
fictional world of the story is like. In other words: the parafictional statement embeds a
piece of pretence (corresponding to the fictional statement) for demonstrative pur-
poses and says, truly or falsely, that this is what the world of the fiction is like.
¹⁶ My debate with Maier in Theoretical Linguistics concerned fictional uses of names such as ‘Frodo’
or ‘Holmes’ (Maier 2017a; Recanati 2017).
32  
This view relies on a framework which various philosophers have gestured at

for handling attitude reports.¹⁷ Mark Sainsbury, who calls it ‘display theory’,
presents it as follows:
In attributions of intentional states concepts are displayed, and the attribution is

correct if the concepts displayed match those in the mind of the subject. When
we say that Ursula is thinking about unicorns, we do not use the concept
UNICORNS in the normal way, the way we would use it if we said there were
unicorns in the park. Instead, we put the concept UNICORNS on display, and
our attribution is correct if Ursula is exercising that concept in her thinking. [ . . . ]
Display theory predicts that a concept that refers to nothing may be used in a
correct attribution of what someone is thinking, explaining how Ursula can think
about unicorns. (Sainsbury 2018, 1–2)
I would rather put things this way: in reporting a thought about unicorns (or
about Mumbo-Jumbo, as in McDowell’s example¹⁸), one uses and mentions the
concept UNICORN at the same time. One uses it simulatively, by ‘playing along
with the deluded subject’ whose thoughts one is reporting (McDowell 1977, 127).
But the aim of the simulation is to convey to the audience what the thought is like;
so the embedding operator itself (here, the belief operator ‘Ursula thinks that’) has
to be construed as involving a demonstrative component referring to the concepts
deployed in the current simulation, as in Davidson’s paratactic analysis of oratio
obliqua (Davidson 1968). This type of semi-quotational analysis is what we need
to reconcile radical simulationism and the truism that parafictional sentences can
be true or false. On this view the parafictional operator ‘in fiction f ’ has to be
construed, not as an ordinary intensional operator (as in Lewis 1978), but as
involving a hidden demonstrative reflexively referring to the pretence exemplified
by the fictional sentence it embeds.¹⁹ Thus understood, a parafictional statement
‘In fiction f, p’ is true just in case the piece of pretence exemplified by the use of the
embedded fictional sentence matches that which practitioners of fiction f engage
in, or, in slightly different terms, just in case what one imagines when processing
the embedded fictional statement matches what fiction f mandates its practitioners
to imagine.
On this analysis parafictional discourse turns out to be intermediate between
fictional discourse and metafictional discourse. In parafictional discourse, we do
talk about the fiction, as much as we do in metafictional discourse. When we say
‘In Conan Doyle’s novels, Sherlock Holmes is a detective’, we explicitly refer to the
fiction, by means of the phrase ‘Conan Doyle’s novels’; to that extent, we do take a
¹⁷ See e.g. Quine 1960, 92; Stich 1983, 83–84; Gordon 1986, 166–167. ¹⁸ See footnote 15.
¹⁹ As Walton says, the parafictional speaker ‘indicate[s] the relevant kind of pretense by exempli-
fying it’ (1990, 400).
    33
‘metafictional’ perspective. This is very different from being immersed in a fiction

and not reflectively thinking about it. At the same time, parafictional discourse
discloses the content of the fiction, what it represents; and this it does, in
continuity with fictional discourse, by giving to imagine what the fiction pre-
scribes its practitioners to imagine. That is what the audience of a parafictional
utterance does: she imagines a fictional state of affairs while simultaneously
‘tagging’ the imagined state of affairs as one that is depicted in the fiction.
On this view the proper name ‘Sherlock Holmes’ in a parafictional statement
like (3) simulatively ‘refers’ to the flesh-and-blood individual Sherlock Holmes,
just as it does in the fictional statement which (3) embeds. In such parafictional
discourse we ‘retell a small portion of the story in order to characterize its content’
(Everett 2013, 50–51). In so doing we take a metafictional stance towards the
Sherlock Holmes story, but the task of the audience is merely to imagine the flesh-
and-blood individual Sherlock Holmes, in the course of imagining the state of
affairs of which it is a constituent, while understanding that this is what the
Sherlock Holmes story mandates its readers to imagine. No reference is made to
‘fictional characters’ construed as cultural artefacts, in contrast to what happens in
metafictional discourse such as (1). As I wrote in my paper on that topic,
The irreducible metafictional component involved in parafictional discourse is

located in the reference to the fiction conveyed by the [parafictional operator, ‘in
fiction f ’]; all the rest is a continuation of the pretence that is constitutive of
fictional thought and talk. (Recanati 2018, 50)
2.5 Conclusion
In this chapter I have defended the view that fictional discourse is asymmetrically
dependent upon ‘serious’ (non-fictional) discourse: fictional reference and fic-
tional assertion alike are parasitic on genuine reference and genuine assertion,
which they mimic. I have shown this view to be compatible with the two-level
picture according to which processing both fictional and non-fictional discourse
involves constructing a mental model of the situations talked about: the assertoric/
judgemental force inherent in such mental representations is cancelled when the
discourse is tagged as fictional. On this view, the parafictional operators (‘in The
Lord of the Rings’, ‘in the Sherlock Holmes stories’ . . . ) through which such
tagging is effected in discourse about the fiction should not be construed as
intensional operators in the manner of Lewis (1978), but as hyperintensional
operators involving a demonstrative component referring to the pretence exem-
plified by the fictional sentence they embed. The fictional names occurring in the
sentences embedded under such operators work exactly like the fictional names
occurring in a text of fiction; they do not refer to ‘fictional characters’ construed as
34  
cultural artefacts, in contrast to what happens in metafictional discourse, but

pretend to refer to ordinary flesh-and-blood individuals, thus exemplifying and
continuing the pretence at work in the fiction they attempt to characterize.
Acknowledgements
I am indebted to Gregory Currie, Emar Maier, Merel Semeijn, Andreas Stokke, and an
Oxford University Press reviewer for discussion and comments on an earlier draft of this
chapter.
References
Barwise, Jon and Perry, John (1981). Semantic innocence and uncompromising
situations. Midwest Studies in Philosophy, 6, 387–403.
Crimmins, Mark (1998). Hesperus and Phosphorus: Sense, pretence, and reference.
Philosophical Review, 107, 1–48.
Currie, Gregory (1995). Image and Mind: Film, Philosophy, and Cognitive Science.
Cambridge: Cambridge University Press.
Davidson, Donald (1968). On saying that. In his Inquiries into Truth and
Interpretation (pp. 93–108). Oxford: Oxford University Press.
Dummett, Michael (1973). Frege: Philosophy of Language. London: Duckworth.
Eckardt, Regine (2015). The Semantics of Free Indirect Discourse: How Texts Allow Us
to Mind-Read and Eavesdrop. Leiden: Brill.
Egan, Andy (2008). Seeing and believing: Perception, belief formation and the divided
mind. Philosophical Studies, 140, 47–63.
Evans, Gareth (1982). The Varieties of Reference. Oxford: Clarendon Press.
Everett, Anthony (2013). The Nonexistent. Oxford: Oxford University Press.
Frege, Gottlob (1979). Posthumous Writings. Oxford: Blackwell.
Geach, Peter (1960). Ascriptivism. In his Logic Matters (pp. 250–254). Oxford:
Blackwell.
Geach, Peter (1965). Assertion. In his Logic Matters (pp. 254–269). Oxford: Blackwell.
Gilbert, Daniel (1991). How mental systems believe. American Psychologist, 46,
101–119.
Gordon, Robert (1986). Folk psychology as simulation. Mind and Language, 1,
158–171.
Hanks, Peter (2007). The content-force distinction. Philosophical Studies, 134,
141–164.
Hanks, Peter (2011). Structured propositions as types. Mind, 120, 11–52.
Hanks, Peter (2015). Propositional Content. Oxford: Oxford University Press.
Another random document with
no related content on Scribd:
other plants on which they have alighted. They also seize small
aquatic insects; but, although, I suspect that they disgorge in pellets
the harder parts of these, I have no proof, obtained from actual
observation, that they do so.
The holes perforated by this species for the purpose of breeding
require considerable exertion and labour. They are usually bored at
the distance of two or three feet from the summit of the bank or
surface of the ground, to the depth of about three feet, but
sometimes to that of four or even five. They are near each other or
remote, according to the number of pairs of swallows that resort to
the place, and the extent of the bank. In one situation you may find
not more than a dozen pairs at work, while in another several
hundreds of holes may be seen scattered over some hundreds of
yards. On the bluffs of the Ohio and the Mississippi there are many
very extensive breeding-places. While engaged in digging a sand-
bank on the shore of the Ohio, at Henderson, for the purpose of
erecting a steam-mill, I was both amused and vexed by the
pertinacity with which the little winged labourers continued to bore
holes day after day, whilst the pickaxes and shovels demolished
them in succession. The birds seemed to have formed a strong
attachment to the place, perhaps on account of the fine texture of the
soil, as I observed many who had begun holes a few hundred yards
off abandon them, and join those engaged in the newly opened
excavation. Whether the holes are frequently bored horizontally or
not I cannot say, but many which I examined differed in this respect
from those described by authors, for on introducing a gun-rod or
other straight stick, I found them to have an inclination of about ten
degrees upwards. The end of the hole is enlarged in the form of an
oven, for the reception of the nest, and the accommodation of the
parents and their brood.
When the birds have for a while examined the nature of the bank,
they begin their work by alighting against it, securing themselves by
the claws, and spreading their tails considerably so as, by being
pressed against the surface, to support the body. The bill is now
employed in picking the soil, until a space large enough to admit the
body of the bird is formed, when the feet and claws are also used in
scratching out the sand. I have thought that the slight ascent of the
burrow contributed considerably to enable the bird to perform the
severe task of disposing of the loose materials, which are seen
dropping out at irregular intervals. Both sexes work alternately, in the
same manner as Woodpeckers; and few ornithological occupations
have proved more pleasing to me than that of watching several
hundred pairs of these winged artificers all busily and equally
engaged, some in digging the burrows, others in obtaining food,
which they would now and then bring in their bills for the use of their
mates, or in procuring bits of dry grass or large feathers of the duck
or goose, for the construction of their nests.
So industrious are the little creatures that I have known a hole dug to
the depth of three feet four inches, and the nest finished in four days,
the first egg being deposited on the morning of the fifth. It sometimes
happens that soon after the excavation has been commenced, some
obstruction presents itself, defying the utmost exertions of the birds;
in which case they abandon the spot, and begin elsewhere in the
neighbourhood. If these obstructions occur and are pretty general,
the colony leaves the place; and it is very seldom that, after such an
occurrence, any swallows of this species are seen near it. I have
sometimes been surprised to see them bore in extremely loose
sand. On the sea coast, where soft banks are frequent, you might
suppose that, as the burrows are only a few inches apart, the sand
might fall in so as to obstruct the holes and suffocate their inmates;
but I have not met with an instance of such a calamitous occurrence.
Along the banks of small rivulets I have found these birds having
nests within a foot or two of the water, having been bored among the
roots of some large trees, where I thought they were exposed to
mice, rats, or other small predaceous animals. The nest is generally
formed of some short bits of dry grass, and lined with a considerable
number of large feathers. They lay from five to seven eggs for the
first brood, fewer for the next. They are of an ovate, somewhat
pointed form, pure white, eight-twelfths of an inch long, and six-
twelfths in breadth
The young, as soon as they are able to move with ease, often crawl
to the entrance of the hole, to wait the return of their parents with
food. On such occasions they are often closely watched by the
smaller Hawks, as well as the common Crows, which seize and
devour them, in spite of the clamour of the old birds. These
depredations upon the young are in fact continued after they have
left the nest, and while they are perched on the dry twigs of the low
trees in the neighbourhood, until they are perfectly able to maintain
themselves on wing without the assistance of their parents.
In Louisiana, or in any district where this species raises more than
one brood in the season, the males, I believe, take the principal
charge of the young that have left the nest, though both sexes
alternately incubate, all their moments being thus rendered full of
care and anxiety respecting both their offspring and the sitting bird.
The young acquire the full brown plumage of the adult by the first
spring, when there is no observable difference between them; but I
am induced to think that they keep apart from the old birds during the
first winter, when I have thought I could yet perceive an inferiority in
their flight, as well as in the loudness of their notes.
This species has no song, properly so called, but merely a twitter of
short lisping notes. In autumn it at times alights on trees preparatory
to its departure. On such occasions the individuals, often collected in
great numbers, take up the time chiefly in pluming themselves, in
which occupation they continue for hours.
I must conclude with assuring you that in my opinion, no difference
whatever exists between the Bank Swallow of America and that of
Europe. The birds from which I made the drawing for my plate were
procured on the banks of the Schuylkil River in 1824.
Hirundo riparia, Linn. Syst. Nat. vol. i. p. 344.—Lath. Ind. Ornith. vol. i. p.
575.—Ch. Bonap. Synopsis, p. 65.
Bank Swallow or Sand Martin, Hirundo riparia, Wils. Amer. Ornith. vol. v.
p. 46, pl. 38, fig. 4.
Hirundo riparia, the Sand Martin, Richards. and Swains. Fauna Bor.-
Amer. vol. ii p. 333.
Bank Swallow, or Sand Martin, Nuttall, Manual, vol. i. p. 607.
Adult Male. Plate CCCLXXXV. Fig. 1.

Bill very short, much depressed and very broad at the base,
compressed toward the point, of a triangular form with the lateral
outlines concave, when viewed from above or beneath; upper
mandible with the dorsal line considerably convex, the sides convex,
the edges sharp and overlapping, with a slight but distinct notch
close to the deflected tip; lower mandible with the angle very broad,
the dorsal line ascending and convex, the ridge broad and flat at the
base, narrowed toward the tip, which is acute, the edges inflected.
Nostrils basal, lateral, oblong.
Head of ordinary size, roundish, depressed; neck short; body
slender. Feet very small; tarsus very short, anteriorly scutellate,
moderately compressed, with a tuft of feathers behind at the lower
part; toes free, small, the lateral equal, the first much stronger; claws
long, slightly arched, much compressed, very acute.
Plumage soft and blended, without lustre. Wings very long,
extending a little beyond the tail, very narrow, slightly falciform; the
primaries tapering to an obtuse-point, the first quill longest, the
second half a twelfth shorter, the third four and a quarter twelfths
shorter than the second, the rest rapidly graduated; six of the
secondaries distinctly emarginate. Tail rather long, deeply
emarginate, the feathers tapering to an obtuse point.
Bill brownish-black. Iris hazel. Feet flesh-coloured, claws dusky. The
upper parts are greyish-brown, or mouse colour, the head and wing-
coverts darker, as are the primary coverts, primary quills, and outer
secondaries, of which the shafts are dusky above, white beneath.
The lower parts are white; the cheeks, a broad band across the
lower part of the neck and fore part of the breast, and the sides
under the wings, greyish-brown. The tail-feathers are very narrowly
edged with a lighter tint, the outer with whitish.
Length to end of tail 5 inches, to end of wings 5 1/4, to end of claws
4 1/8; extent of wings 11; bill along the ridge 3/12, along the edge of
lower mandible 6 1/2/12; wing from flexure 4 2/12; tail to the fork 1 11/12;
to the end 2 4 1/2/12; tarsus 5/12; hind toe 2 1/4/12; its claw 3 1/2/12;
3/ 1/
middle toe 4 /12, its claw 2
4 /12.
2
Adult Female. Plate CCCLXXXV. Fig. 2.

The Female cannot be distinguished from the male by any difference
in her external appearance.
Length to end of tail 4 7/8 inches, to end of wings 5 1/4, to end of

claws 4.
Young. Plate CCCLXXXV. Fig. 3.
The young when fully fledged, have the bill dusky, with the edges
yellow, the feet flesh-coloured, the claws yellowish. The colour of the
upper parts is darker, but the feathers are margined with light
greyish-brown; the quills brownish-black, the outer very faintly, the
inner broadly margined; the tail-feathers greyish-black, edged with
greyish-white. The lower parts are white, the throat faintly streaked
with dusky; the band across the breast, and the sides, coloured as in
the adult, but darker.
On very carefully comparing skins of this Swallow, with a series of
those of the Bank Swallow of Europe, procured for me by my
esteemed friend, Thomas Durham Weir of Boghead, Esq. an
enthusiastic and successful observer of the habits of birds, I can
perceive no difference whatever. Old birds compared with old, and
young with young, prove perfectly similar. There is, however, another
species closely allied to the present, and which might very readily be
confounded with it. This species, to which I give the name of Rough-
winged Swallow, Hirundo serripennis, I consider it expedient to
describe, although it has not as yet been figured by me.
In a male of the present species, from Boston, the palate is flat, the
mouth very wide, measuring 5 twelfths across. The tongue is short,
triangular, 2 1/2 twelfths long, deeply emarginate and papillate at the
base, two of the lateral papillæ much larger than the rest, the tip
bluntish and slightly slit. The œsophagus, a b c, is 1 inch 9 twelfths
long, narrow, 2 twelfths in diameter, without crop or dilatation. The
proventriculus, b, is little enlarged. The stomach, cdef, a gizzard of
moderate length, with distinct lateral muscles, and of an elliptical
form, is half an inch long, and 5 twelfths broad; its epithelium
longitudinally rugous, tough, and light red. It is filled with remains of
insects. The intestine, f g h, is 5 1/2 inches long, its greatest diameter
1 1/2 twelfth; the cœca very small, being 1 1/2 twelfth long, and 1/2
twelfth in diameter, their distance from the anus 9 twelfths. There is
no essential difference between the digestive organs of this and
other swallows, and the Flycatchers, Warblers, and other slender-
billed birds.
The trachea is 1 inch 4 twelfths long; slender, flattened, of about 55
unossified rings. The contractor and sterno-tracheal muscles are
slender; and there are four pairs of inferior laryngeal muscles
ROUGH-WINGED SWALLOW.
Hirundo serripennis.
On the afternoon of the 20th of October 1819, I was walking along
the shores of a forest-margined lake, a few miles from Bayou Sara,
in pursuit of some Ibises, when I observed a flock of small Swallows
bearing so great a resemblance to our common Sand Martin, that I
at first paid little attention to them. The Ibises proving too wild to be
approached, I relinquished the pursuit, and being fatigued by a long
day’s exertion, I leaned against a tree, and gazed on the Swallows,
wishing that I could travel with as much ease and rapidity as they,
and thus return to my family as readily as they could to their winter
quarters. How it happened I cannot now recollect, but I thought of
shooting some of them, perhaps to see how expert I might prove on
other occasions. Off went a shot, and down came one of the birds,
which my dog brought to me between his lips. Another, a third, a
fourth, and at last a fifth were procured. The ever-continuing desire
of comparing one bird with another led me to take them up. I thought
them rather large, and therefore placed them in my bag, and
proceeded slowly toward the plantation of William Perry, Esq., with
whom I had for a time taken up my residence.
The bill and feet of the Swallows were pure black, and both, I
thought, were larger than in the Sand Martin; but differences like
these I seldom hold in much estimation, well knowing from long
experience, that individuals of any species may vary in these
respects. I was more startled when I saw not a vestige of the short
feathers usually found near the junction of the hind toe with the
tarsus in the common species, and equally so when I observed that
the bird in my hand had a nearly even tail, with broad rounded
feathers, the outer destitute of the narrow margin of white. At this
time my observations went no farther.
Doubts have been expressed by learned ornithologists respecting
the identity of the Common Sand Martin of America and that of
Europe. Some of them in their treatises write Hirundo riparia? or
Cotyle riparia? which in my opinion is foolish, especially if no reason
be given for appending so crooked a character. About two years ago,
my friend the Rev. John Bachman, sent me four Swallow’s eggs
accompanied with a letter, in which was the following notice:—“Two
pairs of Swallows resembling the Sand Martin, have built their nests
for two years in succession in the walls of an unfinished brick house
at Charleston, in the holes where the scaffoldings had been placed.
It is believed here that there are two species of these birds.” The
eggs which my friend sent me differ greatly from those of our
Common Sand Martin, being so much longer, larger, and more
pointed, that I might have felt inclined to suppose them to belong to
the European Swift, Cypselus murarius. But of the birds which had
laid them no particular account was given. Time has passed; and
during the while I have been anxious to meet again with such
Swallows as I had shot near Bayou Sara, as well as to determine
whether our Common Sand Martin be the same as that of Europe.
And now, Reader, I am at last able to say, that the Sand Martin or
Bank Swallow, Hirundo riparia, is common to Europe and America;
and further, that a species, confounded with it in the latter country, is
perfectly distinct.
I perhaps should never have discovered the differences existing
between these species had I not been spurred by the remarks of
Vieillot, who, in expressing his doubts as to their identity, and
perhaps holding in his hand the bird here spoken of, says that the
tarsus is much larger than in the European Sand Martin. I have been
surprised that these doubts did not awaken in others a desire to
inquire into the subject. Had this been done, however, I should
probably have lost an opportunity of adding another new species to
those to whose nomination I can lay claim, not to speak of such as,
although well known to me previous to their having been published
by others, I have lost the right of naming because I had imparted my
knowledge of them to those who were more anxious of obtaining this
sort of celebrity. I have now in my possession one pair of these
Swallows procured by myself in South Carolina during my last visit to
that State. Of their peculiar habits I can say nothing; but, owing to
their being less frequent than the Sand Martin, I am inclined to
believe that their most habitual residence may prove to be far to the
westward, perhaps in the valleys of the Columbia River.
I regret that I have not figured this species, though it would have
proved exceedingly difficult to exhibit in an engraving the peculiar
character presented by the outer quill, unless it were much
magnified.
The specific characters of these two Swallows, so nearly allied, are
as follows:—
Fig. 1.
Hirundo riparia. Tail, Fig. 1, slightly forked, margin of first quill

smooth, tarsus with a tuft of feathers behind; upper parts greyish-
brown, lower whitish, with a dusky band across the fore part of the
neck.
Fig. 2.
Hirundo serripennis. Tail, Fig. 2, slightly emarginate, margin of first

quill, Fig. 3, rough with the strong decurved tips of the barbs, tarsus
bare; upper parts greyish-brown, lower pale greyish-brown, white
behind.
Fig. 3.
In its general appearance, including proportions as well as colour,

the Rough-winged Swallow is extremely similar to the Bank Swallow.
It differs however in having the bill considerably longer, more
attenuated toward the end, with the point of the upper mandible
more decurved. The tail, Fig. 2, is shorter and but slightly
emarginate, the lateral exceeding the middle feathers by only two-
twelfths of an inch, whereas in the other species they exceed them
by five-twelfths or even six; feathers are also broader and more
rounded at the end. The wings are longer, and extend half an inch
beyond the tail. The tarsi and toes are somewhat longer and more
slender, and there are no feathers on the hind part of the tarsus as in
the common species; the claws are much more slender.
The bill is black, the tarsi, toes, and claws dusky. The upper parts
are of the same greyish-brown, or mouse-colour, as those of the
Bank Swallow. The lower are of a very light greyish-brown, gradually
paler on the hind parts, the abdomen and lower tail-coverts being
white.
Length to end of tail 5 3/4 inches, to end of wings 6 1/4; extent of
wings 12 1/2; bill along the ridge 3 1/2/12, along the edge of lower
mandible 6 1/2/12; wing from flexure 4 4/12; tail to end of middle feather
1 10/12; to end of longest feather 2; tarsus 5/12; hind toe 2 1/4/12, its
claw 2 1/4/12; middle toe 5/12, its claw 3 1/4/12.
In a specimen, from Charleston, South Carolina, preserved in spirits,

the roof of the mouth is flat, the width of the gape 5 1/2 twelfths; the
tongue triangular, 3 1/2 twelfths long, emarginate and papillate at the
base, with two of the papillæ much larger, flat above, tapering to a
slit point, more narrow and elongated than that of the Sand Swallow.
The œsophagus is 1 inch 11 twelfths long, without dilatation, its
breadth 1 1/2 twelfth. The stomach is elliptical, muscular, 5 twelfths
long, being 4 1/2 twelfths, and placed a little obliquely; the epithelium
brownish-red, tough, longitudinally rugous, filled with remains of
insects. The intestine is 4 1/2 inches long; the cœca 1 1/2 twelfth in
length, and 8 twelfths distant from the anus.
The trachea is 1 inch 4 twelfths long, its diameter 1 twelfth; its rings
about 50; the muscles as in the other species; the bronchi very
slender, of about 12 half rings.
VIOLET-GREEN SWALLOW.
Hirundo thalassina, Swainson.

PLATE CCCLXXXV. Male and Female.
Of this, the most beautiful Swallow hitherto discovered within the

limits of the United States, the following account has been
transmitted to me by my friend Mr Nuttall. “We first met with this
elegant species within the table-land of the Rocky Mountains, and
they were particularly abundant around our encampment on Harris
Fork, a branch of the Colorado of the west. They are nearly always
associated with the Cliff Swallow, here likewise particularly
numerous. Their flight and habits are also similar, but their twitter is
different, and not much unlike the note of our Barn Swallow. In the
Rocky Mountains, near our camp, we observed them to go in and
out of deserted nests of the Cliff Swallow, which they appeared to
occupy in place of building nests of their own. We saw this species
afterwards flying familiarly about in the vicinity of a farm-house (M.
Le Boute’s) on an elevated small isolated prairie on the banks of the
Wahlamet, and as there are no cliffs in the vicinity, they probably
here breed in trees, as I observed the White-bellied Martin do. This
beautiful species in all probability extends its limits from hence to the
table-land of Mexico, where Mr Bulloch, it seems, found it.”
Dr Townsend, who afterwards had better opportunities of observing
the habits of this bird, thus speaks of it:—“Aguila chin chin of the
Chinook Indians, inhabits the neighbourhood of the Colorado of the
west, and breeds along its margins on bluffs of clay, where it
attaches a nest formed of mud and grasses resembling in some
measure that of the Cliff Swallow, but wanting the pendulous neck in
that of the latter species. The eggs are four, of a dark clay colour,
with a few spots of reddish-brown at the larger end. This species is
also found abundant on the lower waters of the Columbia. River,
where it breeds in hollow trees.”
Dr Townsend also informs me that in the neighbourhood of the
Columbia River, the Cliff Swallow attaches its nest to the trunks of
trees, making it of the same form and materials as elsewhere. From
the above facts, and many equally curious, which I have mentioned,
respecting the variations exhibited by birds in the manner of forming
their nests, as well as in their size, materials, and situation, it will be
seen that differences of this kind are not of so much importance as
has hitherto been supposed, in establishing distinctions between
species supposed by some to be different, and by others identical.
To give you some definite idea of what I would here impress upon
your mind, I need only say that I have seen nests of the Barn or
Chimney Swallow placed within buildings, under cattle-sheds,
against the sides of wells, and in chimneys; that while some were not
more than three inches deep, others measured nearly nine; while in
some there was scarcely any grass, in others it formed nearly half of
their bulk. I have also observed some nests of the Cliff Swallow in
which the eggs had been deposited before the pendent neck was
added, and which remained so until the birds had reared their brood,
amidst other nests furnished with a neck, which was much longer in
some than in others. From this I have inferred that nests are formed
more or less completely, in many instances, in accordance with the
necessity under which the bird may be of depositing its eggs.
Hirundo thalassinus, Swains. Synopsis of Mexican Birds, Philos. Mag. for

1827, p. 365.
Adult Male. Plate CCCLXXXV. Fig. 4.
Bill very short, much depressed and very broad at the base,
compressed toward the point, of a triangular form, with the lateral
outlines nearly straight; upper mandible with the dorsal line
considerably convex, the sides convex toward the end, the edges
sharp and overlapping, with a slight but distinct notch close to the
deflected acute tip; lower mandible with the angle very broad, the
dorsal line ascending and slightly convex, the ridge broad and a little
convex at the base, narrowed toward the tip, which is acute. Nostrils
basal, lateral oblong.
Head rather large, roundish; neck very short; body slender. Feet very
small; tarsus very short, anteriorly scutellate, compressed; toes free,
small, the lateral equal, the first stronger; claws rather long, arched,
much compressed, very acute.
Plumage soft and blended, on the upper parts somewhat velvety.
Wings very long, extending far beyond the tail, very narrow, slightly
falciform; the primaries tapering to an obtuse point; the first quill
longest, the second almost equal, the rest rapidly diminishing; six of
the secondaries emarginate. Tail of moderate length, emarginate, the
middle feathers four-twelfths shorter than the lateral; all rounded.
Bill black; iris brown; feet dusky. The upper part of the head deep
green gradually shaded into the dark purple of the hind neck; the
back rich grass-green, the rump and upper tail-coverts carmine
purple. The smaller wing-coverts are dusky, broadly tipped with
green, glossed with purple; the quills and larger coverts dusky,
glossed with blue; the tail also dusky, glossed with blue. A line over
the eye, the cheeks, and all the lower parts, are pure white excepting
the lower wing-coverts, which are light grey.
Length to end of tail 4 10/12, to end of wings 5 7/12; bill along the ridge
2 3/4/ , along the edge of lower mandible 5/12; wing from flexure
12
1/2
4 6/12; tail to end of middle feathers 1 1/2, to end of longest 1 10 /12;
3/ 1/
tarsus 4 /12; hind toe 2/12, its claw 2/12; middle toe 5
4 /12, its claw
2
2/12.
Adult Female. Plate CCCLXXXV. Fig. 5.

The Female is somewhat smaller, and differs considerably in colour.
The upper part of the head and the hind neck are light greyish-brown
glossed with bronzed green; the back bright green as in the male,
the rump greyish-brown; the wings and tail are as in the male, but
less glossy; as are the lower parts, which are, however, anteriorly
tinged with grey.
Length to end of tail 4 7/12, to end of wings 5; wing from flexure 4 1/4,
tail 1 8 1/2/12.
GREAT AMERICAN EGRET.
Ardea Egretta, Gmel.

PLATE CCCLXXXVI. Male.
In the third volume of this work, I have already intimated that the truly
elegant Heron which now comes to be described, is a constant
resident in the Floridas, that it migrates eastward sometimes as far
as the State of Massachusetts, and up the Mississippi to the city of
Natchez, and, lastly, that it is never seen far inland, by which I mean
that its rambles into the interior seldom extend to more than fifty
miles from the sea-shore, unless along the course of our great rivers.
I have now to add that on my way to the Texas, in the spring of 1837,
I found these birds in several places along the coast of the Gulf of
Mexico, and on several of the islands scattered around that named
Galveston, where, as well as in the Floridas, I was told that they
spend the winter.
The Great American Egret breeds along the shores of the Gulf of
Mexico, and our Atlantic States, from Galveston Island in the Texas
to the borders of the State of New York, beyond which, although
stragglers have been seen, none, in so far as I can ascertain, have
been known to breed. In all low districts that are marshy and covered
with large trees, on the margins of ponds or lakes, the sides of
bayous, or gloomy swamps covered with water, are the places to
which it generally resorts during the period of reproduction; although
I have in a few instances met with their nests on low trees, and on
sandy islands at a short distance from the main land. As early as
December I have observed vast numbers congregated, as if for the
purpose of making choice of partners, when the addresses of the
males were paid in a very curious and to me interesting manner.
Near the plantation of John Bulow, Esq. in East Florida, I had the
pleasure of witnessing this sort of tournament or dress-ball from a
place of concealment not more than a hundred yards distant. The
males, in strutting round the females, swelled their throats, as
Cormorants do at times, emitted gurgling sounds, and raising their
long plumes almost erect, paced majestically before the fair ones of
their choice. Although these snowy beaux were a good deal irritated
by jealousy, and conflicts now and then took place, the whole time I
remained, much less fighting was exhibited than I had expected from
what I had already seen in the case of the Great Blue Heron, Ardea
Herodias. These meetings took place about ten o’clock in the
morning, or after they had all enjoyed a good breakfast, and
continued until nearly three in the afternoon, when, separating into
flocks of eight or ten individuals, they flew off to search for food.
These manœuvres were continued nearly a week, and I could with
ease, from a considerable distance mark the spot, which was a clear
sand-bar, by the descent of the separate small flocks previous to
their alighting there.
The flight of this species is in strength intermediate between that of
Ardea Herodias and A. rufescens, and is well sustained. On foot its
movements are as graceful as those of the Louisiana Heron, its
steps measured, its long neck gracefully retracted and curved, and
its silky train reminded one of the flowing robes of the noble ladies of
Europe. The train of this Egret, like that of other species, makes its
appearance a few weeks previous to the love season, continues to
grow and increase in beauty, until incubation has commenced, after
which period it deteriorates, and at length disappears about the time
when the young birds leave the nest, when, were it not for the
difference in size, it would be difficult to distinguish them from their
parents. Should you however closely examine the upper plumage of
an old bird of either sex, for both possess the train, you will discover
that its feathers still exist, although shortened and deprived of most
of their filaments. Similar feathers are seen in all other Herons that
have a largely developed train in the breeding season. Even the few
plumes hanging from the hind part of the Ardea Herodias, A.
Nycticorax, and A. violacea, are subject to the same rule; and it is
curious to see these ornaments becoming more or less apparent,
according to the latitude in which these birds breed, their growth
being completed in the southern part of Florida two months sooner
than in our Middle Districts.
The American Egrets leave the Floridas almost simultaneously about
the 1st of March, and soon afterwards reach Georgia and South
Carolina, but rarely the State of New Jersey, before the middle of
May. In these parts the young are able to fly by the 1st of August. On
the Mule Keys off the coast of Florida, I have found the young well
grown by the 8th of May; but in South Carolina they are rarely
hatched until toward the end of that month or the beginning of June.
In these more southern parts two broods are often raised in a
season, but in the Jerseys there is, I believe, never more than one.
While travelling, early in spring, between Savannah in Georgia and
Charleston in South Carolina, I saw many of these Egrets on the
large rice plantations, and felt some surprise at finding them much
wilder at that period of their migrations than after they have settled in
some locality for the purpose of breeding. I have supposed this to be
caused by the change of their thoughts on such occasions, and am
of opinion that birds of all kinds become more careless of
themselves. As the strength of their attachment toward their mates
or progeny increases through the process of time, as is the case with
the better part of our own species, lovers and parents performing
acts of heroism, which individuals having no such attachment to
each other would never dare to contemplate. In these birds the
impulse of affection is so great, that when they have young they
allow themselves to be approached, so as often to fall victims to the
rapacity of man, who, boasting of reason and benevolence, ought at
such a time to respect their devotion.
The American Egrets are much attached to their roosting places, to
which they remove from their feeding grounds regularly about an
hour before the last glimpse of day; and I cannot help expressing my
disbelief in the vulgar notion of birds of this family usually feeding by
night, as I have never observed them so doing even in countries
where they were most abundant. Before sunset the Egrets and other
Herons (excepting perhaps the Bitterns and Night Herons) leave
their feeding grounds in small flocks, often composed of only a single
family, and proceed on wing in the most direct course, at a moderate
height, to some secure retreat more or less distant, according to the
danger they may have to guard against. Flock after flock may be
seen repairing from all quarters to these places of repose, which one
may readily discover by observing their course.
Approach and watch them. Some hundreds have reached the well-
known rendezvous. After a few gratulations you see them lower their
bodies on the stems of the trees or bushes on which they have
alighted, fold their necks, place their heads beneath the scapular
feathers, and adjust themselves for repose. Daylight returns, and
they are all in motion. The arrangement of their attire is not more
neglected by them than by the most fashionable fops, but they spend
less time at the toilet. Their rough notes are uttered more loudly than
in the evening, and after a very short lapse of time they spread their
snowy pinions, and move in different directions, to search for
fiddlers, fish, insects of all sorts, small quadrupeds or birds, snails,
and reptiles, all of which form the food of this species.
The nest of the Great White Egret, whether placed in a cypress one
hundred and thirty feet high, or on a mangrove not six feet above the

The Language of Fiction Emar Maier Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Language of Fiction Emar Maier Full Chapter

Uploaded by

Copyright:

Available Formats

The Language of Fiction Emar Maier

Visit to download the full and correct content document:

The Language of Fiction

The Language of Fiction

List of contributors vii

I. TRUTH, REFERENCE, AND IMAGINATION

III. PERSPECTIVE SHIFT

Márta Abrusán is a tenured research fellow at the Centre national de la recherche

viii   

Nellie Wieland is a Professor in the Department of Philosophy at California State

Language is used to represent states of affairs in the world around us.

2     

These twentieth-century trends were to a large extent inﬂuenced by linguistic

Part I: Truth, reference, and imagination

Summaries and descriptions of the contributions are given in the introductions to

1.1 Truth, reference, and imagination

It is a well-established observation that the central semantic notions of truth and

(1) a. Frodo is a hobbit.

(2) a. There are hobbits.

However, this metaphysically motivated rejection of Meinongianism leaves the

4     

aimed at incrementing alternative kinds of contextual (Bonomi and Zucchi 2003;

(4) a. Hamlet is Danish.

A particularly challenging instance of this kind of problem arises for counter-

(5) a. Hamlet might not have returned from England.

(6) a. Hamlet is younger than Sherlock Holmes.

6     

research in theoretical linguistics and philosophy of language has been concerned

1.3 Perspective shift

In parallel with efforts to better understand the linguistic characteristics of

8     

thoughts or speech. According to the standard gloss, in free indirect discourse

10     

Bal, M. (1997). Narratology: Introduction to the Theory of Narrative. Toronto:

Holton, R. (1997). Some telling examples: A reply to Tsohatzidis. Journal of

12     

van Inwagen, P. (1977). Creatures of ﬁction. American Philosophical Quarterly, 14,

(1) a. Hermione is kind of a wizard genius.

This is unsatisfactory: in the context of a discussion about the Harry Potter

14 , ,  

aesthetics and philosophy of language, it has recently been challenged by

(2) Hermione doesn’t really exist—she’s just a ﬁctional character made up by

, ,   15

oecumenical approach. Kamp partly motivates his account by demonstrating how

In “Fictional truth: In defense of the Reality Principle,” Nils Franzén revisits

16 , ,  

2.1 Fictional names and their uses

instances of ﬁctitious (non-genuine) reference to an ordinary ﬂesh-and-blood

(1) Sherlock Holmes is a ﬁctional character created by Conan Doyle in 1887.

• Statement (1) is true.

    19

2.1.1 Matravers against the simulation view

Following a venerable tradition, I have claimed that ﬁctional uses of referring

    21

The distinction between confrontations and representations is more funda-

Matravers argues for ‘a “two-stage” model of engaging with representations’,

My objection to Currie is that his functional description [ . . . ]—non standard

The two-stage model advocated by Matravers is reminiscent of the traditional

(i) it is characteristic of ﬁction to mandate a special kind of attitude on the part of

    23

asserting is more complex: it involves grasping/entertaining plus ‘endorsing’. This

Some [cognitive acts]—judging and asserting that o is red—involve further

2.2 Force cancellation