Pluralisms in Truth and Logic, Jeremy Wyatt, Nikolaj J. L. L. Pedersen, Nathan Kellen (Ed.)

Pluralisms in Truth
and Logic
Edited by
Jeremy Wyatt
Nikolaj J. L. L. Pedersen
Nathan Kellen
Palgrave Innovations in Philosophy
Series Editors
Vincent Hendricks
University of Copenhagen
Copenhagen, Denmark
Duncan Pritchard
University of Edinburgh
Edinburgh, UK
Palgrave Innovations in Philosophy is a new series of monographs. Each
book in the series will constitute the ‘new wave’ of philosophy, both in
terms of its topic and the research profile of the author. The books will be
concerned with exciting new research topics of particular contemporary
interest, and will include topics at the intersection of Philosophy and
other research areas. They will be written by up-and-coming young phi-
losophers who have already established a strong research profile and who
are clearly going to be leading researchers of the future. Each monograph
in this series will provide an overview of the research area in question
while at the same time significantly advancing the debate on this topic
and giving the reader a sense of where this debate might be heading next.
The books in the series would be of interest to researchers and advanced
students within philosophy and its neighboring scientific environments.
More information about this series at

http://www.palgrave.com/gp/series/14689
Jeremy Wyatt
Nikolaj J. L. L. Pedersen • Nathan Kellen
Editors
Pluralisms in Truth
and Logic
Editors
Jeremy Wyatt Nikolaj J. L. L. Pedersen
Underwood International College Underwood International College
Yonsei University Yonsei University
Incheon, South Korea Incheon, South Korea
Nathan Kellen
University of Connecticut,
Storrs, CT, USA
Palgrave Innovations in Philosophy

ISBN 978-3-319-98345-5 ISBN 978-3-319-98346-2 (eBook)
https://doi.org/10.1007/978-3-319-98346-2
Library of Congress Control Number: 2018963837
© The Editor(s) (if applicable) and The Author(s) 2018

This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and trans-
mission or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, express or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
Cover illustration: Bashutskyy shutterstock.com
This Palgrave Macmillan imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
Pluralism has made inroads into a number of areas in philosophy. This

edited collection brings together 18 state-of-the-art articles on pluralism
about truth and logic.
We first discussed the possibility of an edited collection shortly after
the Truth Pluralism and Logical Pluralism Conference, held at the
University of Connecticut in April of 2015. Following this conference,
several additional pluralism-related events took place, including three
workshops hosted by the Cogito Research Centre at the University of
Bologna in 2015 and 2016 and Pluralisms Week, hosted by the Pluralisms
Global Research Network and the Veritas Research Center at Yonsei
University in June of 2016. A fair number of contributors were given the
chance to present and discuss their work at one or several of these events.
The University of Connecticut, University of Bologna, and Yonsei
University all provided financial support. We gratefully acknowledge
their support. Two of the editors, Pedersen and Wyatt, would also like to
thank the National Research Foundation of Korea for support (grants no.
2013S1A2A2035514 and 2016S1A2A2911800).
We are grateful to many colleagues who share our interest in pluralism.
Their collegial, constructive ways of conducting research and discussions
are much appreciated. We are grateful to many people who, in some way
or another, have helped along the way. These include Jc Beall, Elke
Brendel, Colin Caret, Roy Cook, Douglas Edwards, Will Gamester,
v
vi Preface
Patrick Greenough, Sungil Han, Jinho Kang, Jiwon Kim, Junyeol Kim,
Seahwa Kim, Teresa Kouri Kissel, Kris McDaniel, Graham Priest, Agustín
Rayo, Greg Restall, Jisoo Seo, Stewart Shapiro, Gila Sher, Paul Simard
Smith, Erik Stei, Elena Tassoni, Pilar Terrés, Cory D. Wright, Crispin
Wright, Andy D. Yu, Luca Zanetti, and Elia Zardini. Special thanks go to
Filippo Ferrari, Michael P. Lynch, Sebastiano Moruzzi, and Joe Ulatowski.
Incheon, South Korea Jeremy Wyatt

Nikolaj J. L. L. Pedersen
Storrs, CT, USA Nathan Kellen
Contents
Part I Truth 1
Introduction 3
Nikolaj J. L. L. Pedersen, Jeremy Wyatt, and Nathan Kellen
Truth: One or Many or Both? 35

Dorit Bar-On and Keith Simmons
Truth Pluralism, Quasi-Realism, and the Problem of Double-

Counting 63
Michael P. Lynch
The Metaphysics of Domains 85

Douglas Edwards
Strong Truth Pluralism107

Seahwa Kim and Nikolaj J. L. L. Pedersen
vii
viii Contents
Methodological Pluralism About Truth131

Nathan Kellen
Normative Alethic Pluralism145

Filippo Ferrari
Truth in English and Elsewhere: An Empirically-Informed

Functionalism169
Jeremy Wyatt
Part II Logic 197
Core Logic: A Conspectus199

Neil Tennant
Connective Meaning in Beall and Restall’s Logical Pluralism217

Teresa Kouri Kissel
Generalised Tarski’s Thesis Hits Substructure237

Elia Zardini
Logical Particularism277
Gillman Payette and Nicole Wyatt
Logical Nihilism301
Aaron J. Cotnoir
Varieties of Logical Consequence by Their Resistance to

Logical Nihilism331
Gillian Russell
Contents ix
Part III Connections 363
Pluralism About Pluralisms365

Roy T. Cook
A Plea for Immodesty: Alethic Pluralism, Logical Pluralism,

and Mixed Inferences387
Chase B. Wrenn
Logic for Alethic, Logical, and Ontological Pluralists407

Andy D. Yu
Pluralisms: Logic, Truth and Domain-Specificity429

Rosanna Keefe
Aletheic and Logical Pluralism453

Kevin Scharp
Index473
Notes on Contributors
Dorit Bar-On is Professor of Philosophy, Department of Philosophy, and

Director of the Expression, Communication, and the Origins of Meaning
Research Group, University of Connecticut. Bar-On is well known for her work
in philosophy of language, mind, and metaethics. She is author of Speaking My
Mind: Expression and SelfKnowledge (Clarendon Press, 2004) and has published
in journals such as The Journal of Philosophy, Mind & Language, Philosophy and
Phenomenological Research, Noûs, Synthese, and Philosophical Studies.
Roy T. Cook is Professor of Philosophy, Department of Philosophy, University
of Minnesota. Cook specializes in the philosophy of logic, philosophical logic,
philosophy of mathematics, and aesthetics. He is the author of Paradoxes (Polity,
2013) and The Yablo Paradox: An Essay on Circularity (OUP, 2014), as well as
many papers in journals such as Mind, Analysis, Journal of Symbolic Logic,
Philosophia Mathematica, and Australasian Journal of Philosophy.
Aaron J. Cotnoir is Senior Lecturer, Department of Philosophy, University of
St Andrews. Cotnoir’s work is centered around metaphysics and philosophical
logic. He is the editor of Composition as Identity (OUP, 2014, with Donald
Baxter) and co-author of Mereology (OUP, forthcoming). His articles have
appeared in Journal of Philosophy, Mind, Noûs, Australasian Journal of Philosophy,
and more.
Douglas Edwards is Assistant Professor of Philosophy, Department of
Philosophy, Utica College. Edwards’ research centers around metaphysics, the
philosophy of language, and metaethics. He is the author of Properties (Polity,
xi
xii Notes on Contributors
2014) and The Metaphysics of Truth (OUP, 2018) and the editor of Truth: A
Contemporary Reader (Bloomsbury, under contract). His articles have appeared
in a number of journals, including the Journal of Philosophy, Australasian Journal
of Philosophy, and Pacific Philosophical Quarterly, amongst others.
Filippo Ferrari is Postdoctoral Research Fellow at the Institute of Philosophy,
University of Bonn. His research focuses primarily on two clusters of topics: the
normative aspects of enquiry and the debate about the nature of truth. He has
published his work in journals such as Mind, Synthese, Analysis, and Philosophical
Quarterly.
Rosanna Keefe is Professor of Philosophy and Head, Department of
Philosophy, University of Sheffield. Keefe specializes in philosophy of logic, phi-
losophy of language, and metaphysics. She is the author of Theories of Vagueness
(Cambridge, 2000) and numerous articles in journals such as Mind, Analysis,
Philosophical Studies, Australasian Journal of Philosophy, and Synthese.
Nathan Kellen works at the Department of Philosophy, University of
Connecticut. Kellen’s work is on truth, the philosophy of logic, philosophy of
mathematics, and ethics. Currently his main research project is an investigation
of truth pluralism and logical pluralism. He explores both of these views indi-
vidually but likewise examines how they might be connected.
Seahwa Kim is Professor of Philosophy and Dean, Scranton College, Ewha
Womans University. Kim specializes in metaphysics and the philosophy of
mathematics. Her articles have appeared in journals such as Australasian Journal
of Philosophy, Philosophical Studies, and Erkenntnis.
Teresa Kouri Kissel is Assistant Professor, Department of Philosophy, Old
Dominion University. Kouri Kissel specializes in philosophy of logic, philoso-
phy of mathematics, and mathematical and philosophical logic. Her dissertation
develops a new, neoCarnapian form of logical pluralism. Her articles have
appeared in Philosophia Mathematica, Erkenntnis, and Topoi.
Michael P. Lynch is Professor of Philosophy and Director, Humanities
Institute, University of Connecticut. Lynch’s work focuses on questions in meta-
physics, the philosophy of language, epistemology, and metaethics. He is author
of Truth in Context (MIT, 1998), True to Life (MIT, 2004), and Truth as One
and Many (OUP, 2009), as well as two books for popular audiences and a num-
ber of different articles in journals such as Philosophical Quarterly, Australasian
Journal of Philosophy, and Philosophical Studies.
Notes on Contributors xiii
Gillman Payette is Postdoctoral Fellow at the University of British Columbia.

His main research interests are philosophical logic and the philosophy of logic.
He has published in journals such as Journal of Philosophical Logic, Australasian
Journal of Philosophy, Synthese, Logique et Analyse, and Notre Dame Journal of
Formal Logic.
Nikolaj J. L. L. Pedersen is Associate Professor of Philosophy and Director of
the UIC Research Institute and Veritas Research Center, Underwood
International College, Yonsei University. He has been the principal investigator
of several collaborative research projects. Pedersen’s main research areas are truth,
epistemology, and metaphysics. He has published in journals such as Noûs,
Analysis, Philosophical Quarterly, Synthese, Erkenntnis, and The Monist. He is a
coeditor of New Waves in Truth (Palgrave Macmillan, 2010), Truth and Pluralism:
Current Debates (Oxford University Press, 2013), Epistemic Pluralism (Palgrave
Macmillan, 2017), Epistemic Entitlement (Oxford University Press, 2019), and
The Routledge Handbook of Social Epistemology (2019).
Gillian Russell is Professor of Philosophy, Department of Philosophy,
University of North Carolina at Chapel Hill. Russell specializes on the philoso-
phy of language, philosophy of logic, and epistemology. She is the author of
Truth in Virtue of Meaning: A Defence of the Analytic/Synthetic Distinction (OUP,
2008) and her articles have appeared in journals such as Journal of Philosophical
Logic, Philosophical Studies, and Australasian Journal of Philosophy.
Kevin Scharp is Reader in Philosophy, Department of Philosophy and Director,
Arché Philosophical Research Centre, University of St Andrews. Scharp special-
izes in the philosophy of language, logic, metaphysics, philosophy of science,
and the history of analytic philosophy. He is the author of Replacing Truth (OUP,
2013) and numerous articles in journals such as The Philosophical Review,
Philosophy and Phenomenological Research, Erkenntnis, and Philosophical Studies.
Keith Simmons is Professor of Philosophy, Department of Philosophy,
University of Connecticut. Simmons specializes in logic, philosophy of lan-
guage, and metaphysics. He is the author of Universality and the Liar (OUP,
1993) and has had articles appear in Philosophical Studies, Philosophy and
Phenomenological Research, and Journal of Philosophical Logic, amongst other
journals.
Neil Tennant is Arts and Humanities Distinguished Professor of Philosophy,
Department of Philosophy, Ohio State University. Tennant specializes in logic,
philosophy of mathematics, and the philosophy of language. He is the author of
xiv Notes on Contributors
a number of books, including The Taming of the True (OUP, 2002) and Changes
of Mind: An Essay on Rational Belief Revision (OUP, 2012). His articles have
appeared in many journals, including Mind, Philosophia Mathematica, Review of
Symbolic Logic, and Noûs.
Chase B. Wrenn is Professor of Philosophy, Department of Philosophy,
University of Alabama. Wrenn’s research focuses on truth, epistemology, and the
philosophy of mind and cognitive science. He is the author of Truth (Polity,
2014) and has had articles appear in journals including Australasian Journal of
Philosophy, Erkenntnis, Synthese, and The Philosophical Quarterly.
Jeremy Wyatt is Assistant Professor of Philosophy, Underwood International
College, Yonsei University. Wyatt’s main research interests are the philosophy of
language, metaphysics, and truth. His articles have appeared in Philosophical
Studies, Philosophical Quarterly, American Philosophical Quarterly, and Inquiry.
Andy D. Yu is JD student, University of Toronto. Yu completed a D.Phil. the-
sis (Fragmented Truth) at the University of Oxford. He works on philosophical
logic, the philosophy of language, metaphysics, and epistemology. He has pub-
lished in the Journal of Philosophy, Philosophical Quarterly, and Thought.
Elia Zardini is MidCareer FCT Fellow, LanCog Research Group, University of
Lisbon. Zardini specializes in logic and epistemology and has had articles appear
in many journals, including The Review of Symbolic Logic, Philosophical Studies,
Analysis, and Journal of Philosophical Logic. He is also the editor or coeditor of
Scepticism and Perceptual Justification (OUP, 2014), Substructural Approaches to
Paradox (special issue of Synthese, forthcoming), The Sorites Paradox (CUP, forth-
coming), and The A Priori: Its Significance, Grounds, and Extent (OUP,
forthcoming).
List of Figures
Truth in English and Elsewhere: An Empirically-Informed

Functionalism
Fig. 1 Alethic functionalism 173
Fig. 2 Consistency of interlinguistic and intralinguistic pluralism 177
Fig. 3 Updated functionalism 184
Core Logic: A Conspectus
Fig. 1 From classical logic to core logic 203
Fig. 2 Important system containments 204
Varieties of Logical Consequence by Their Resistance to Logical
Nihilism
Fig. 1 A truth-table proof of Modus tollens 343
Fig. 2 A paraconsistent truth-table of Modus tollens 344
xv
Part I
Truth
Introduction
Nikolaj J. L. L. Pedersen, Jeremy Wyatt,
and Nathan Kellen
1 Pluralisms
The history of philosophy displays little consensus or convergence when
it comes to the nature of truth. Radically different views have been pro-
posed and developed. Some have taken truth to be correspondence with
reality, while others have taken it to be coherence with a maximally coher-
ent set of beliefs. Yet others have taken truth to be what it is useful to
believe, or what would be believed at the end of enquiry.1 While these
views differ very significantly in terms of their specific philosophical com-
mitments, they all share two fundamental assumptions: monism and sub-
stantivism. The views all assume that truth is to be accounted for in the
N. J. L. L. Pedersen (*)
Underwood International College, Yonsei University, Incheon, South Korea
e-mail: nikolaj@yonsei.ac.kr
J. Wyatt
N. Kellen
University of Connecticut, Storrs, CT, USA
© The Author(s) 2018 3

J. Wyatt et al. (eds.), Pluralisms in Truth and Logic, Palgrave Innovations in Philosophy,
https://doi.org/10.1007/978-3-319-98346-2_1
4 N. J. L. L. Pedersen et al.
same way across the full range of truth-apt discourse (monism) and that
truth is a substantive property or relation (substantivism).
The deflationist reaction to the traditional debate is to reject substantiv-
ism and, in some cases, to endorse monism. Truth, if it has a nature at all,
has a uniform nature across all truth-apt discourse, but there is not much
to say about it. The traditional debate went off-track exactly because
truth theorists thought that there was a whole lot to say about truth—
that somehow it had a deep or underlying nature that could be uncovered
through philosophical theorizing. Instead, according to many deflation-
ists, the (non-paradoxical) instances of the disquotational schema (“p” is
true if and only p) or the equivalence schema (it is true that p if and only
if p) exhaust what there is to say about truth.2
The pluralist reaction to the traditional debate is to reject monism and
endorse substantivism. Truth pluralists, encouraged by the seminal work
of Crispin Wright and Michael Lynch, appeal to more than one property
in their account of truth. Propositions from different domains of dis-
course are true in different ways. The truth of propositions concerning
the empirical world (e.g., 〈There are mountains〉) might be accounted
for in terms of correspondence, while the truth of legal propositions (e.g.,
〈Speeding is illegal〉) might be accounted for in terms of coherence with
the body of law.3 This amounts to a rejection of monism. By contrast,
truth pluralists have traditionally endorsed substantivism. They have
appealed to properties or relations that are substantive in nature (where
this means, at least, that they directly explain certain facts entirely in
virtue of characteristics pertaining to their natures).4
The history of logic, like the history of truth, displays little consensus.
Advocates of classical logic, intuitionistic logic, and relevant logic, for
instance, have argued back and forth about the merits and demerits of their
preferred systems. Again, as in the case of truth, this seems to suggest a shared
underlying assumption of monism: there is a uniquely correct logic, and
advocates of different systems are disagreeing about which one it is. The plu-
ralist reaction—notably advocated by Jc Beall and Greg Restall (2006)—is
to reject monism and maintain that there are several equally correct logics.
Speaking more generally, pluralist views are becoming increasingly
prominent in different areas of philosophy. Pluralism about truth has been
extensively developed, defended, and critically discussed. The same goes
Introduction 5
for pluralism about logic.5 Pluralism has also made inroads into ontology
where the idea that there are several ways of being has been defended, sup-
ported, and articulated in various ways. The work of Kris McDaniel is a
particularly rich source.6 In epistemology, a variety of pluralist theses can
likewise be found in the literature. The idea that there are several epistemi-
cally good-making features of belief can be found in different guises, as
pluralism about epistemic justification, warrant, desiderata, and value.
Prominent epistemologists such as Alvin Goldman, Tyler Burge, William
Alston, and Crispin Wright all endorse one of these forms of pluralism.7
These pluralist trends are philosophically significant. They go against a
one-size-fits-all conception of their relevant areas and invite a reconsid-
eration of the nature and character of some of the most fundamental
notions in core areas of philosophy—including truth, validity, being, and
justification. This volume takes as its focus two of kinds of pluralism:
pluralism about truth and pluralism about logic. It brings together 18
original, state-of-the-art essays. The essays are divided into three parts.
Part I is dedicated to truth pluralism, Part II to logical pluralism, and Part
III to the question as to what connections might exist between these two
kinds of pluralism.
2 Truth Pluralism (Part I)

In this section, we will briefly introduce a range of ideas and issues that
have served to shape and define the debate concerning truth pluralism.
We will then introduce the contributions to Part I of the volume.
Background
Truth pluralists are engaged in critical debates on two fronts—one exter-

nal and the other internal. On the external front, we find pluralists debat-
ing monists as to which of these views of truth is superior. The argument
most commonly deployed by pluralists is the scope problem. Pluralists argue
that monist theories do not have a scope that is sufficiently wide to plau-
sibly accommodate all truth-apt discourse. Perhaps the correspondence
theory can be plausibly applied to empirical discourse, accounting for the

truth of propositions such as 〈Mt. Everest is extended in space〉. However,
it cannot plausibly cover the truth of legal propositions such as 〈Breaking
and entering is illegal〉. Now, maybe coherence with the body of law can
plausibly be applied to legal discourse. However, coherence does not seem
easily extendable to the empirical domain. Pluralists take this point to
generalize and conclude that monism is unsatisfactory. Instead, they let
several theories of truth work in tandem, restricting their range of applica-
bility to certain domains.
In connection with the internal debate, we see advocates of so-called
strong pluralism and moderate pluralism pitched against one another. They
agree that truth pluralism is the right view, but disagree over the details.
According to strong pluralists, there is no single truth property that
applies to all true propositions. Rather, there is a range of properties that
reduce or constitute truth for different propositions belonging to differ-
ent domains. Thus, the truth of 〈There are mountains〉 may reduce to
this proposition’s corresponding to reality while the truth of 〈Speeding is
illegal〉 may reduce to this proposition’s cohering with the body of law.
Crucially, there is no single property that reduces or constitutes the truth
of every true proposition. Moderate pluralists, on the other hand, endorse
a single, generic truth property. This property is possessed by every true
proposition, so that for the moderate pluralist, truth itself is one. However,
moderate pluralists also think that instances of generic truth are grounded
by instances of different truth-relevant properties. In this sense, for the
moderate pluralist, truth is also many. Thus, according to moderate plu-
ralism, truth is both one and many.8
Despite their different views on how to best articulate truth pluralism,
strong and moderate pluralists share significant commitments. One such
commitment is the commitment to domains. Domains are a crucial com-
ponent of the theoretical framework of pluralism, as reflected by the fact
that the core pluralist thesis is that the nature of truth varies across
domains.
Domains also feature prominently in some of the main challenges
faced by truth pluralists. Problems concerning mixed discourse are cases in
point. Mixed discourse is discourse that cuts across domains. Such
discourse occurs at three different levels: the levels of atomics, com-
pounds, and inferences. Consider the propositions that are expressed in:
Introduction 7
(mix-atom) π is beautiful.
(mix-comp) Mt. Everest is extended in space and Bob’s drunk driving is

illegal.
(mix-inf ) If Mt. Everest is extended in space, then Bob’s drunk driv-

ing is illegal.
Mr. Everest is extended in space.
Bob’s drunk driving is illegal.
〈π is beautiful〉 is a mixed atomic proposition. It features a mathemati-

cal concept (π) and an aesthetic concept (beauty).9 However, in light of
this, should 〈π is beautiful〉 be classified as belonging to the mathemati-
cal domain, the aesthetic domain, both of these domains, or perhaps
some other domain? This question by no means seems to be a straight-
forward one to answer. However, there seems to be considerable pressure
on the pluralist to provide an answer. After all, for atomic propositions
such as 〈π is beautiful〉, the domain membership of the proposition is
meant to determine the property that is relevant to its truth. Hence,
absent a principled story about the domain membership of mixed atom-
ics, there would be a whole cluster of propositions whose truth would
remain unaccounted for. This is the problem of mixed atomics.
〈Mt. Everest is extended in space and Bob’s drunk driving is illegal〉 is
a mixed conjunction. Its first conjunct belongs to the empirical domain,
and its second conjunct to the legal domain. The conjunction is true, as
both of its conjuncts are. However, it is not clear what story the pluralist
is going to tell about this. We can suppose that correspondence to reality
and coherence with the body of law are, respectively, the truth-relevant
properties for the two conjuncts. However, neither correspondence nor
coherence seems like a plausible candidate when we try to account for the
truth of the conjunction itself. Now, if the truth-relevant property of
neither the first nor the second conjunct is the right property, the con-
junction must have some third property. However, what property would
that be? The problem of mixed compounds challenges pluralists to tell a
story about the truth of mixed compounds.
The inference from 〈If Mt. Everest is extended in space, then Bob’s
drunk driving is illegal〉 and 〈Mr. Everest is extended in space〉 to 〈Bob’s
drunk driving is illegal〉 is a mixed inference. It is also a valid inference—
that is, necessarily, if the premises are true, then so is the conclusion. In
order to account for the validity of the inference, it would seem that there
must be some truth-relevant property that the premises and conclusion
all share which ensures that truth is preserved from premises to conclu-
sion. However, the pluralist seems to be unable to point to a property
that satisfies this constraint. For, as before, we can suppose that corre-
spondence to reality is the truth-relevant property for 〈Mr. Everest is
extended in space〉 and coherence with the body of law for 〈Bob’s drunk
driving is illegal〉. This means that one of the premises and the conclu-
sion have different truth-relevant properties. The problem of mixed infer-
ences challenges the pluralist to tell a story about the validity of mixed
inferences.10
Another fundamental problem confronting pluralists is what is some-
times called the “double-counting objection.” In essence, the objection is
that pluralists count two differences where only one is needed. They
endorse significant metaphysical differences regarding the nature of vari-
ous subject matters and, in addition, they endorse differences in the
nature of truth. However, in order to accommodate wide-ranging truth-
aptitude, differences need only be countenanced at one level—at the level
of the things themselves (numbers, trees, moral properties, laws, etc.) or
at the level of the content associated with different domains (expressivist
content vs. representational content). Drawing distinctions at the level of
truth, the objection goes, is superfluous.11
The Contributions
Dorit Bar-On and Keith Simmons’ contribution “Truth: One or

Many?” runs a version of the double-counting objection. Prominent
pluralists such as Wright and Lynch claim that one motivation for
adopting truth pluralism is that it puts one in a position to make
sense of disputes between realists and anti-realists. Adopting the the-
sis that the nature of truth varies across domains, it is possible for the
Introduction 9
pluralist to account for the “differential appeal of realist and anti-

realist intuitions” about them.12 Bar-On and Simmons counter by
arguing that truth pluralism offers no distinctive explanatory power
vis-à-vis the realism/anti-realism debate. Rather, all that is needed is a
plurality of kinds of worldly conditions that track metaphysical differ-
ences. Bar-On and Simmons’ version of the double-counting objec-
tion targets not only truth pluralists but also certain kinds of truth
monists—namely, those who think that a distinction between differ-
ent kinds of content or meaning offers explanatory power vis-à-vis the
realism/anti-realism debate. Bar-On and Simmons spell out in con-
siderable detail how their proposal differs from those of truth monists
who seek to accommodate realism/anti-realism disputes by appealing
to semantic differences.
Michael P. Lynch’s contribution “Truth Pluralism, Quasi-Realism and
the Problem of Double-Counting” offers a pluralist response to the dou-
ble-counting objection—in particular, as it might be pressed by quasi-
realists such as Simon Blackburn and global expressivists such as Huw
Price. Lynch argues that semantic diversity and cognitive unity are theses
that quasi-realists, global expressivists, and truth pluralists all seek to
accommodate and explain:
(SD) There are real differences in kind between the contents of our
beliefs and indicative statements.
(CU) All beliefs and indicative statements are subject to a single

type of cognitive normative assessment of correctness.
Blackburn proposes to accommodate semantic diversity by appealing to

two sorts of propositions or two sorts of truth-aptitude. Price proposes to
accommodate semantic diversity by appealing to two kinds of representa-
tion: i-representation and e-representation. A proposition i-represents in
virtue of its inferential or functional role, while e-representation is cashed
out in terms of co-variance with the environment. Lynch argues that, in
effect, both Blackburn’s proposal and Price’s proposal result in a form of
truth pluralism. Hence, when it comes to double-counting, truth plural-
ists turn out to be no worse off than quasi-realists or global expressivists.
Douglas Edwards’ contribution “The Metaphysics of Domains” pro-

vides a systematic account of domains and thus addresses a major lacuna
in the theoretical framework of truth pluralism. Edwards distinguishes
between a semantic and a metaphysical aspect of domains. He accounts
for these via a discussion of, respectively, singular terms and predicates
and their metaphysical counterparts, objects and properties. Edwards
argues that his proposed notion of domain is not a commitment of
pluralists only—it is implicit in a number of philosophical views.
Edwards demonstrates the significance of his account of domains by
arguing that it delivers a solution to two major challenges to truth plu-
ralism: the problem of mixed atomics and the problem of mixed
compounds.
Seahwa Kim and Nikolaj Jang Lee Linding Pedersen’s contribution
“Strong Truth Pluralism” provides a pluralist solution to the problem of
mixed compounds. Kim and Pedersen offer their solution against the
background of a certain form of strong pluralism. They thus deny that
there is any generic truth property that applies across all domains.
Instead they think that an atomic proposition’s being true reduces to its
corresponding to reality, cohering, or having some other “base-level
truth property.” The truth of a compound reduces to its having a com-
pound-specific truth-reducing property. For example, the truth of the
mixed conjunction 〈Mt. Everest is extended in space and Bob’s drunk
driving is illegal〉 reduces to its being a conjunction with conjuncts that
have their respective truth-reducing properties. Note that in this regard
mixed conjunctions are no different from pure conjunctions. Hence,
Kim and Pedersen’s response to the problem of mixed compounds is
that there is nothing special or problematic about them. They are true
in the same way that their pure counterparts are.
The contributions by Nathan Kellen, Filippo Ferrari, and Jeremy
Wyatt explore different ways to reconfigure or add a new dimension to
the debate concerning truth pluralism.
Kellen’s contribution “Methodological Pluralism About Truth” introduces
a meta-perspective on the truth pluralism debate. According to Kellen,
Wright takes anti-realism to be methodologically fundamental. He does so
in the sense that anti-realist truth is taken to be the default for all domains.
Introduction 11
This is not to say that Wright is opposed to giving truth a “non-default”

treatment within some domains. However, in order for truth to receive
this kind of treatment, there must be reasons why, within that particular
domain, the default should be abandoned. Kellen likewise attributes a
thesis of methodological fundamentality to Edwards although, in con-
trast to Wright, the default is realist truth. Kellen proposes methodological
pluralism as an alternative to anti-realist and realist default forms of plu-
ralism. According to this methodological doctrine, the pluralist should
have no default for any domain but should remain neutral until reasons
have been given one way or the other.
Filippo Ferrari’s contribution “Normative Alethic Pluralism” articu-
lates a novel view concerning the normativity of truth—what he calls
“normative alethic pluralism.” He articulates this view against the back-
ground of a rejection of normative alethic monism, the view that truth’s
normative profile is uniform across all domains, and can be captured by
a single principle that connects truth and judgment. Ferrari targets nor-
mative alethic monism by employing a normative analogue of the scope
problem. By appealing to disagreements pertaining to different domains,
Ferrari argues that the normative profile of truth varies across domains. If
people disagree over whether, for example, oysters are tasty, there is no
strong sense of fault in play. On the other hand, if two parties disagree as
to whether abortion is a morally acceptable practice, there is a strong
sense of fault in play. This is reflected by each party’s tendency toward
condemnation of the other party. This normative variability cannot satis-
factorily be accommodated within the framework of normative alethic
monism. Hence, according to Ferrari, there is reason to adopt normative
alethic pluralism. There are several ways for truth to normatively regulate
judgment. Having motivated and articulated normative alethic plural-
ism, Ferrari discusses the issue of whether it might bear significant con-
nections to pluralism about truth. He concludes that the two kinds of
pluralism are compatible and may nicely complement one another. Yet,
one does not imply the other.
Jeremy Wyatt’s contribution “Truth in English and Elsewhere: An
Empirically-Informed Functionalism” presents a refined framework that
should help to shape future work on both pluralism and functionalism
about truth. Wyatt argues that a proper defense of alethic functional-

ism—a view that was pioneered by Lynch—must be informed by empiri-
cal data. The alethic functionalist puts forward an account of the folk
theory of truth, which is meant to consist of the propositions about truth
that those who possess the folk concept of truth are ipso facto disposed to
believe upon reflection. The functionalist’s view on the folk theory of
truth is thus empirical in nature and should be evaluated using empirical
data. Wyatt discusses two kinds of existing data: data pertaining to alethic
vocabulary in English and preliminary data pertaining to the Ghanaian
language Akan. The English-related data suggest variation in the use of
alethic vocabulary among male and female English speakers. The prelimi-
nary data concerning Akan suggest that there are significant differences in
the alethic vocabulary that is used, respectively, by native Akan and native
English speakers. Wyatt argues that these lines of data lend support to
two kinds of pluralism regarding ordinary thought about truth—what he
calls intralinguistic and interlinguistic conceptual pluralism. In addition to
motivating these two sorts of pluralism, Wyatt shows how the functional-
ist, by adopting a more nuanced version of Lynch’s framework, can nicely
accommodate these pluralistic hypotheses.
3 Logical Pluralism (Part II)

Over the past decade, there has been a surge of interest in logical plural-
ism, one major factor being the publication of JC Beall and Greg Restall’s
Logical Pluralism.13 In this section, we will briefly introduce Beall and
Restall’s logical pluralism and the contributions to Part II of the volume.
In one way or another, most of these contributions engage with promi-
nent themes from Beall and Restall’s work.
Background
According to Beall and Restall, logic is plural in the sense that there are
several equally legitimate instances of what they call Generalized Tarski’s
Thesis (GTT):
Introduction 13
(GTT) An argument is validx if and only if in every casex in which

the premises are true, so is the conclusion.
Beall and Restall argue that there are at least three equally legitimate ways
to construe casex in GTT: cases as (consistent and complete) possible
worlds, cases as (possibly incomplete) constructions, and cases as (possibly
inconsistent) situations. These three notions of case deliver different log-
ics—respectively classical logic, intuitionistic logic, and relevant logic.
To shed further light on the nature of Beall and Restall’s logical plural-
ism, let us highlight a number of key features of their view: legitimacy,
logical functionalism, logical generalism, logical relativism, meaning con-
stancy, and structural rules and properties.
Legitimacy The standard of legitimacy for cases is given by three features

that Beall and Restall call “necessity”, “formality”, and “normativity.”
Roughly stated, necessity is the idea that valid arguments must be neces-
sarily truth-preserving; formality is the idea that validity is neutral with
respect to content; and normativity is the idea that invalid arguments
must involve a kind of mistake or fault.
Logical Functionalism The network approach to conceptual analysis can

be regarded as a form of conceptual functionalism. A target concept C is
characterized by a set of principles that connect C to other concepts. In
this way, C is characterized through its role or function within a larger
conceptual network. Beall and Restall’s view can be regarded as an
instance of the network approach. Generalized Tarski’s Thesis and the
three constraints of necessity, formality, and normativity characterize the
concept of validity via its connection to other concepts such as case,
truth, necessity, formality, and normativity. In this way, validity
is characterized through its function or role within a larger conceptual
network, and for this reason Beall and Restall’s logical pluralism involves
a form of functionalism.
Logical Generalism To borrow a phrase from Hartry Field, Beall and

Restall’s logical pluralism is a pluralism about all-purpose logics.14 Their
view is not that classical logic, intuitionistic logic, and relevant logic are
legitimate only when restricted to certain domains or when used for certain
purposes. Rather, their view is that classical logic, intuitionistic logic, and
relevant logic are equally legitimate across the board. This idea can be
regarded as a form of logical generalism. The legitimacy of the logics admit-
ted by (GTT) and the three constraints is meant to be completely general.
Logical Relativism Beall and Restall’s logical pluralism involves a form of

logical relativism: statements concerning the validity of arguments only
have a truth-value relative to a particular system.15 Thus, for example, as
far as logic is concerned, there is no absolute fact of the matter as to
whether double negation elimination is valid. Instead there are logic-
relative facts: in classical logic, double negation elimination is valid while
in intuitionistic logic, it is not.16
Meaning Constancy Beall and Restall are pluralists about validity and
logical consequence. However, they maintain that the meaning of the
logical connectives is constant across the various logics that qualify as
legitimate. This view contrasts with a view often attributed to Carnap—
namely, that different logics have different connectives.17 According to
the latter view, the meaning of “and”, “not”, “or”, etc. changes from one
logic to another. By contrast, on Beall and Restall’s view, logical expres-
sions share the same—but incomplete—meaning across logics. However,
clauses that govern the connectives in different logics capture different
aspects of that shared meaning.
Structural Rules and Properties It is common to distinguish between the

operational rules and the structural rules of a system. Operational rules
and properties concern specific logical operations (e.g., negation).
Structural rules and properties of a logic capture general features that
hold purely in virtue of premises and conclusions being structures of
unstructured objects that can be manipulated in certain ways. The fol-
lowing list specifies five well-known structural rules:
(Reflexivity) ϕ ⊨ ϕ
(Monotonicity) If Γ ⊨ ϕ, then Γ, Δ ⊨ ϕ

Introduction 15
(Transitivity) If Γ ⊨ ϕ and Δ, ϕ ⊨ ψ, then Δ, Γ ⊨ ψ
(Contraction) If Γ, ϕ, ϕ ⊨ ψ, then Γ, ϕ ⊨ ψ
(Commutativity) If Γ, ϕ, ψ, Δ ⊨ χ, then Γ, ψ, ϕ, Δ ⊨ χ
Reflexivity says that anything is a logical consequence of itself.

Monotonicity says that, if ϕ is a logical consequence of Γ, adding more
premises Δ does not change anything (i.e., ϕ is also logical of a conse-
quence of Γ, Δ). Transitivity says that, if ϕ is a logical consequence of Γ
and ψ is a logical consequence of Δ, ϕ, then ψ is a logical consequence of
just Δ, Γ. Contraction says that multiple occurrences of the same premise
can be cut to a single occurrence without impacting consequence.
Commutativity says that the order of premises does not matter: it can be
switched without impacting consequence. Any system that fails to accom-
modate any of the five structural rules listed above is a substructural logic.18
The Contributions
Neil Tennant’s contribution “Core Logic: A Conspectus” presents an

“absolutist pluralist” view on logic. It is absolutist in the sense that there is
a core to logic or deductive reasoning—core logic in Tennant’s terminology.
However, Tennant’s view is pluralist in that there are several legitimate
extensions of the core. Tennant is particularly interested in providing an
account of logic for mathematics. He wants to accommodate both con-
structive and non-constructive (or classical) mathematics, but argues that
all proofs should be “relevantized.” For this reason, core logic is relevant.
Thus, Tennant’s view is revisionist, as he thinks that the logics of construc-
tive and classical mathematics should be revised along relevantist lines.
The logic suitable for c onstructive mathematics is relevantized intuitionis-
tic logic (the core logic C) while the logic suitable for non-constructive (or
classical) mathematics is relevantized classical logic (the core logic C+).
Teresa Kouri Kissel’s contribution “Connective Meanings in Beall and
Restall’s Logical Pluralism” contests Beall and Restall’s meaning con-
stancy thesis, that is, their claim that logical expressions have the same
meaning across logics. She argues that this cannot be the case for intu-
itionistic negation and relevant negation. Thus, contrary to their own
claims, Beall and Restall’s logical pluralism ends up having a Carnapian
flavor.
Elia Zardini’s contribution “Generalised Tarski’s Thesis Hits
Substructure” offers an extended argument to the effect that Beall and
Restall’s logical pluralism is overly restrictive. He shows that (GTT)
implies that logical consequence has the structural features mentioned
above, that is, reflexivity, monotonicity, transitivity, contraction, and
commutativity. For instance, in every casex in which ϕ is true, ϕ is true—
thus (GTT) implies reflexivity. By the lights of Beall-Restall logical plu-
ralism, then, any system that fails to satisfy reflexivity cannot qualify as
logic properly so-called. Similarly, for any system that fails to satisfy
monotonicity, transitivity, contraction, or commutativity. In general,
Beall-Restall logical pluralism disqualifies any substructural logic from
qualifying as logic proper. Zardini argues that there are philosophical rea-
sons to doubt each of reflexivity, monotonicity, transitivity, contraction,
and commutativity. For instance, one might wish to abandon transitivity
in order to give an account of vagueness, and one might wish to abandon
contraction in order to address the semantic paradoxes. In light of these
considerations, Zardini deems Beall-Restall pluralism unacceptably
restrictive.
Gillman Payette and Nicole Wyatt’s contribution “Logical
Particularism” introduces and spells out a particularist view on logic.
According to logical particularism, there are no completely general
principles concerning validity. Rather, validity is a property of particu-
lar arguments or inferences. Since Payette and Wyatt think that differ-
ent, specific arguments call for different treatments vis-à-vis validity,
their logical particularism goes hand in hand with a particularist ver-
sion of logical pluralism. This particularist pluralism is at odds with the
kind of logical pluralism endorsed by Beall and Restall—which, as seen
above, involves a commitment to logical generalism. Logics on the
Beall-Restall view are all-purpose logics, which means that their rules
are completely general principles concerning validity. Payette and
Wyatt’s particularist pluralism is also at odds with domain-based forms
Introduction 17
of logical pluralism. Unlike logical generalists, domain-based logical

pluralists do not think that there are completely general logical prin-
ciples. However, they do think that there are somewhat general prin-
ciples, as the rules of a domain-specific logic hold for all arguments or
inferences within the specific domain to which the logic applies. This
commitment, although weaker than generalism, is likewise at odds
with logical particularism.
Aaron Cotnoir’s contribution “Logical Nihilism” makes a case for a
species of logical nihilism. According to this view, there is no logical
consequence relation that correctly represents natural language infer-
ence; formal logics are inadequate to capture informal inference.
Cotnoir gives two clusters of arguments that lend support to logical
nihilism. The first cluster consists of arguments from diversity, the sec-
ond of arguments from expressive limitations. Arguments from diver-
sity contain two steps. The first step is an argument to the effect that no
single logic can provide an adequate account of natural language infer-
ence. The second step is an argument to the effect that at most one logic
can provide an adequate account of natural language inference.
However, nothing satisfies both of these requirements—so, no formal
logic can provide an adequate account of natural language inference.
Arguments from expressive limitations also consist of two steps. The
first is an observation to the effect that natural language contains infer-
ences involving a certain phenomenon. The second step is an argument
to the effect that no formal logic has the expressive resources to capture
inferences of that kind. Hence, due to the expressive limitations of for-
mal logics, no formal logic can provide an adequate account of natural
language inference.
Gillian Russell’s contribution “Varieties of Logical Consequence by
Their Resistance to Logical Nihilism” also speaks to the theme of logi-
cal nihilism, although a different variety than the one d iscussed by
Cotnoir. Logical nihilism, as Russell approaches the view, has it that
there are no valid arguments. This view has emerged as a discussion
point in the logical pluralism debate. A criticism of Beall and Restall
is that validity should be conceived as truth-preservation across all
cases, in contrast to Beall and Restall’s strategy of pairing each type of
case with a corresponding, case-specific type of validity. If validity is

thus conceived, logical nihilism (or “minimalism”) threatens—maybe
no (or only very few) arguments are truth-preserving across all cases.
Russell investigates what resources different accounts of logical conse-
quence offer when it comes to resisting the threat of nihilism. She con-
siders Etchemendy’s interpretational and representational accounts,
Quine’s substitutional account, and Williamson’s universalist account.
Williamson’s account is found to be the one that offers the strongest
path of resistance to the threat of logical nihilism. However, Russell sug-
gests that the universalist path of resistance may come at a high cost. It
may only work because the universalist account operates at a level that
does not fully engage with the problematic phenomena that fuel the
move toward nihilism (such as empty names, vagueness and incomplete
predicates, and self-reference and overdetermination).
4 Connections (Part III)

Parts I and II of the volume respectively concern truth pluralism and logi-
cal pluralism, considered and discussed (mostly) in isolation from one
another. The contributions in Part III explore potential connections
between these two kinds of pluralism.
Background
The most widely explored form of truth pluralism is domain-based. The

core idea, as we have seen, is that the nature of truth varies across domains.
Several authors have connected domain-based truth pluralism and logic,
arguing that pluralism in the case of truth supports pluralism about logic.
The target kind of logical pluralism inherits a crucial feature from its
alethic counterpart: it is domain-based. Different domains have different
logics.
The kind of argument that truth pluralists have given to support
domain-based logical pluralism can be presented as follows19:
Introduction 19
(1) If truth is epistemically constrained in one domain and epistemically

unconstrained in another domain, then different logics govern these
domains.
(2) Truth is epistemically constrained in one domain and epistemically
unconstrained in another domain.
(3) So, there are domains that are governed by different logics.
(2) is accepted by most (domain-based) truth pluralists. Many of them take

the realism/anti-realism debate as background for their truth pluralism.
They want to accommodate the appeal of realism and anti-realism with
respect to different domains. It is an integral part of the realism/anti-realism
debate that anti-realist truth is epistemically constrained while realist truth
is epistemically unconstrained. It is with this in mind that many domain-
based truth pluralists go in for (2). The idea behind (1) is that the logic of
epistemically constrained truth is intuitionistic while the logic of epistemi-
cally unconstrained truth is classical. Both (1) and (2) can be regarded as
reflections of Dummett’s influence on the realism/anti-realism debate, as it
figures in the theoretical foundations of the pluralist program.20
Domain-based logical pluralism raises some tough questions. The view
pairs individual domains with logics. However, what is to be said about
the logic of mixed inferences (sec. 2.1)? What logic governs inferences
that cut across domains with different logics? Lynch has proposed the
following principle of modesty:
(mod) The logic of a mixed compound or inference that involves

propositions from domains with distinct logics L1, …, Ln is
the intersection of L1, …, Ln.
The logic of a given mixed compound or inference is thus minimal—it is

the core shared by all the logics of the domains involved in the com-
pound or inference. Lynch himself officially endorses two logics within
his domain-based pluralist framework, viz. intuitionistic and classical
logic. This means that there is only one way for a compound or inference
to be “logically mixed”—namely, by involving those two logics. The
modesty principle tells us, then, that the logic of mixed compounds and
inferences is always intuitionistic logic.21
We have just reflected on the potential connections between domain-

based pluralisms about truth and logic. However, what other paths into
pluralist territory might there be? There are several, it would seem. For
Beall and Restall, it is cases, rather than domains, that play an integral
role in delivering pluralism about logic. One might also think that logical
pluralism can be supported by appealing to considerations regarding con-
text. Different contexts might be governed by different logics, even if the
domain at issue is the same. Consider two contexts that both involve
mathematics. It may be that in one context, proofs are subject to the
standards of intuitionistic logic because informativeness matters and con-
structive proofs are more informative. In that context, for instance, indi-
rect proofs would not be counted as valid. However, in a different context,
proofs may be subject to the standards of classical logic, so that in that
context, indirect proofs are counted as valid.22
How about truth? Here, too, there are several paths into pluralist ter-
ritory. Let us mention three. On the basis of empirical data, Robert
Barnard and Joseph Ulatowski have suggested that there are significant
differences in ordinary people’s thought and talk about truth along vari-
ous dimensions. One such dimension is gender. In connection with cer-
tain statements, male subjects’ thought and talk about truth look to align
with a correspondence conception of truth. By contrast, female subjects’
thought and talk about truth seem to not align with a correspondence
conception in these cases. At the very least, then, these data suggest that
two different notions of truth are in play.23
John MacFarlane is well-known for advocating an assessment-sensitive
brand of relativism about truth, according to which the truth of some
statements is relative to a context of assessment. However, MacFarlane
also wants to leave room for truth simpliciter. The combination of these
two commitments suggests a form of truth pluralism, which has it that
there are two basic ways of being true: being true relative to a context of
assessment and being true simpliciter.24
Several authors have argued that the semantic paradoxes—including,
famously, the liar paradox—motivate a pluralistic account of truth. One
such author is Kevin Scharp, who has recently proposed a novel approach
to the semantic paradoxes.25 Classical logic and the unrestricted
T-schema (T(ϕ) ⟷ ϕ) generate the liar paradox.26 According to Scharp,
Introduction 21
the ordinary concept of truth is characterized by the unrestricted

T-schema. This concept is rendered inconsistent by the liar paradox and
other paradoxes. Scharp’s solution is to adopt a replacement theory. The
ordinary concept of truth should be replaced by two new concepts,
descending truth and ascending truth. The axiomatic theory that goes
with Scharp’s proposal includes the following two conditionals (but
excludes the converse conditionals):
(TD) TD(ϕ) → ϕ
(TA) ϕ → TA(ϕ)
The combination of (TD) and (TA) looks superficially like the T-schema,
but it is not. The subscripts make all the difference in the world. They
indicate that the two conditionals involve different concepts—the con-
cepts that, on Scharp’s view, should replace the inconsistent, ordinary
concept of truth. Hence, the two conditionals cannot be combined into
a biconditional to yield the T-schema. Rather, according to Scharp, truth
theorists should rest content with (TD) and (TA) and the concepts of
descending and ascending truth that they respectively concern. It is in
this sense that Scharp puts forward a pluralistic account of truth—
whereas we might have thought that we could get along with a single
truth concept, Scharp’s contention is that we actually need a pair of
replacement concepts to adequately resolve the semantic paradoxes.
Insofar as Scharp’s views are motivated by the semantic paradoxes, they
represent yet another path into pluralist territory.27
The Contributions
Roy T. Cook’s contribution “Pluralism About Pluralisms” investigates

the issue of what connections, if any, there might be between truth
pluralism and logical pluralism. His investigation focuses on domain-
based versions of both views, according to which different domains
involve respectively different ways of being true and different logics. As
seen earlier, some pluralists (e.g., Lynch and Pedersen) have argued that
domain-based truth pluralism supports domain-based logical plural-
ism. Cook contests this claim by constructing a formal model in which

domain-based truth pluralism holds, but domain-based logical plural-
ism fails. He also constructs a model in which domain-based logical
pluralism holds, but domain-based truth pluralism fails. On the basis
of these two models, he concludes that domain-based truth pluralism
and domain-based logical pluralism are independent. Neither entails
the other.
Lynch has two key commitments regarding logic. First, he is a
domain-based logical pluralist. Different domains have different log-
ics. Second, he endorses (mod) for mixed inferences: the logic of a
mixed inference is the intersection of the logics of the domains
involved in the inference. Chase Wrenn’s contribution “A Plea for
Immodesty: Alethic Pluralism, Logical Pluralism, and Mixed
Inferences” critically discusses Lynch’s commitments vis-à-vis logic.
Wrenn argues that (mod) renders intuitively valid inferences invalid
and also renders intuitively invalid inferences valid. Furthermore, he
argues that it seems natural to combine (mod) with virtual logic plu-
ralism. Virtual logic pluralism is the thesis that there is just one proper
logic, but that domains may have certain features that make them
appear to be governed by a stronger logic. Thus, for example, while
logic proper may be intuitionistic, reasoning within some domains
may conform to classical logic (for non-logical reasons). As Wrenn
notes, virtual logical pluralism goes against domain-based logical plu-
ralism. These considerations leave someone like Lynch in a tough
spot. He wants to be a domain-based logical pluralist, but domain-
based logical pluralism will not do on its own. It remains silent on
mixed inferences, so in order to deal with such inferences, (mod) is
adopted. Unfortunately, however, (mod) seems to get both certain
validity verdicts and certain invalidity verdicts wrong. Crucially, once
(mod) is adopted, virtual logical pluralism strongly suggests itself to
the self-professed domain-based logical pluralist—undermining, it
would seem, their domain-based logical pluralism.
Andy Yu’s contribution “Logic for Alethic, Logical, and Ontological
Pluralists” is primarily concerned with the issue of how truth pluralists
can provide an account of logic. Yu sets himself the task of responding to
Introduction 23
the challenges of providing a uniform, standard account of the truth of

atomics, compounds, and quantified statements, as well as a standard
account of logical consequence as necessary truth-preservation. One might
expect this to be a tall order, especially bearing in mind that mixed dis-
course has traditionally been a major obstacle for truth pluralists. Yu pres-
ents a formal framework that he takes to capture the commitments of
truth pluralists. The proposed framework behaves like classical first-order
logic and has a logical consequence relation that behaves like classical
consequence. In light of this, Yu takes his account to provide an adequate
response to the challenge of providing a pluralist account of logic that is
both uniform and standard. At the end of his contribution, Yu extends
the framework so that it can accommodate logical pluralism and sketches
how to accommodate ontological pluralism as well. He does the former
by defining a domain-specific notion of validity that allows different
domains to traffic in different kinds of validity. This amounts to a formal
model of what we have called “domain-based logical pluralism” above.
Ontological pluralism is often taken to be the view that there are several
equally fundamental existential quantifiers. Yu proposes to accommodate
this view as well by introducing domain-specific quantifiers. Following
our earlier labeling conventions, the resulting view would be a form of
domain-based ontological pluralism.
Rosanna Keefe’s contribution “Pluralisms: Logic, Truth and
DomainSpecificity” also explores domain-based logical pluralism. She criti-
cally discusses two attempts by domain-based truth pluralists to give an
account of logic: Cotnoir’s algebraic account and domain-based logical plu-
ralism. Cotnoir’s account treats semantic values as n-tuples where n is the
number of domains. Each position in an n-tuple is 1 or 0. If a proposition
belongs to the i-th domain and is true, it has 1 in the i-th position of the
n-tuple and 0 in all other places. The framework incorporates classical clauses
for the connectives and sustains a classical consequence relation, although
Cotnoir also presents a modified framework meant to allow for other log-
ics. Keefe argues that Cotnoir’s account fails. The adoption of n-tuples as
semantic values forces each sentence to have a (“sub”)semantic value in each
place of its n-tuple, and that brings about odd results. Keefe likewise
presses several objections against domain-based logical pluralism. First,
she observes that the view clashes with the widely held view that logic is
topic neutral. Second, like others, she observes that (mod)—a natural candi-
date for addressing the issue of mixed inferences—threatens to undermine
domain-based logical pluralism. Third, building on the two first points,
Keefe offers the following collapse argument: mixed inferences can cut across
all domains. By (mod), the logic of this inference would be the intersection
of all logics that hold for some domain. However, since this logic is topic
neutral, this is what logic proper is—and hence, logical monism is correct.
Fourth, Keefe questions an implicit assumption behind domain-based logi-
cal pluralism, viz. that domain membership suffices to fix the logic of infer-
ences. She points to vagueness as a phenomenon that tells against this
assumption, as this phenomenon that cuts across domains and has tradition-
ally been thought to be of relevance to logic. Lastly, Keefe presents a positive
proposal: validity is assessed relative to context. This allows for logical plural-
ism because the rules or assumptions justified in different contexts may vali-
date different arguments. Keefe observes that (1) her context-based
framework can accommodate domain-based logical pluralism in the sense
that different context-domain pairs may validate different inferences, and (2)
the context-based approach allows for logical pluralism with respect to a
single domain since different contexts may validate different inferences per-
taining to the same domain. Furthermore, Keefe suggests that the context-
based approach is also better suited to deal with vagueness.
Kevin Scharp’s contribution “Aletheic and Logical Pluralism” explores
what he calls coordinated pluralism and compares it to his own replace-
ment theory. Coordinated pluralism is the combination of context-based
truth pluralism and context-based logical pluralism. Coordinated plural-
ism is motivated by considerations related to the semantic paradoxes. As
observed earlier, if truth is characterized by the unrestricted T-schema
and classical logic or certain other logics hold, you end up with paradox.
The double-barreled pluralism that coordinated pluralism offers is meant
to block paradox by letting the nature of truth and the nature of logic
vary across contexts in such a way that there is no context in which truth
and logic have natures that generate paradox. This is meant to happen in
virtue of their natures being coordinated in any given context. If, in a
given context, truth is strong, logic is weak and vice versa. This coordi-
nated pluralism seems like a promising approach to paradox. Might it be
Introduction 25
better than Scharp’s replacement theory? Scharp shows that coordinated

pluralism falls prey to a revenge paradox, that is, a paradox generated
using the very resources that were supposed to block paradox. In light of
this, Scharp concludes that his replacement theory is in fact superior to
coordinated pluralism.
Acknowledgements Nikolaj J. L. L. Pedersen and Jeremy Wyatt have both ben-

efited from support from the National Research Foundation of Korea (grants
no. 2013S1A2A2035514 and 2016S1A2A2911800). This support is gratefully
acknowledged.
Notes
1. David 1994; Devitt 1984; Newman 2007; Rasmussen 2014; Russell
1912; Vision 2004; and Wittgenstein 1921 all endorse versions of the
correspondence theory. Coherence theorists include Blanshard 1939;
Joachim 1906; and Young 2001. (Neo-)pragmatists include James 1907,
1909; Peirce 1878; and Putnam 1981.
2. Deflationists of various stripes include Field 1986, 1994a, b; Grover
1992; Horwich 1998; Quine 1970; Ramsey 1927; Strawson 1950. For a
systematic treatment of deflationary metaphysics of truth, see Wyatt
2016.
3. We use angle brackets to represent propositions.
4. See Wright 1992, 2001; Lynch 2001, 2004b, 2009.
5. Works on pluralism about truth include Asay 2018; Beall 2000, 2013;
Cook 2011; Cotnoir 2009, 2013a, b; Edwards 2008, 2009, 2011,
2012a, b, 2013a, b, 2018; Engel 2013; David 2013; Dodd 2013; Kölbel
2008, 2013; Lynch 2000, 2001, 2004a, b, 2005a, b, 2006, 2008, 2009,
2013; Newhard 2013, 2017; Pedersen 2006, 2010, 2012a, b, 2014;
Pedersen and Edwards 2011; Pedersen and Wright 2010, 2012, 2013a,
b; Shapiro 2011; Sher 1998, 2005, 2013, 2016; Stewart-Wallace 2016;
Tappolet 1997, 2000, 2010; Williamson 1994; (C.D.) Wright 2005,
2010, 2012; Wright 1992, 1996a, b, 1998, 2001, 2013; Wyatt 2013;
Wyatt and Lynch 2016; and Yu 2017a, b. Works on logical pluralism
include Beall 2014; Beall and Restall 2000, 2001, 2006; Bueno and
Shalkowski 2009; Carnielli and Coniglio 2016; Ciprotti and Moretti
2009; Cook 2010, 2014; Eklund 2012; Field 2009; Goddu 2002;
Hjortland 2013; Humberstone 2009; Keefe 2014; Kouri 2016; Kouri
and Shapiro forthcoming; Payette and Wyatt 2018; Priest 2001, 2008,
2014; Read 2006a, b; Restall 2001, 2002, 2012, 2014; Russell 2008,
2014; Shapiro 2014; Sher forthcoming; Terrés forthcoming; van
Benthem 2008; Varzi 2002; and Wyatt 2004.
6. MacDaniel’s work on ontological pluralism spans about a decade. His
book The Fragmentation of Being (2017) brings together much of his
earlier work on the topic. See also. Turner 2010, 2012; Eklund 2009.
7. Alston 2005; Burge 2003; Goldman 1988; Wright 2004. Pedersen 2017
formulates and defends pluralism about fundamental or non-derivative
epistemic goods.
8. This slogan is the title of Lynch 2009. Having advocated moderate plu-
ralism for nearly two decades in a wide range of works, Lynch is the most
prominent advocate of the view. The distinction between these two
forms of pluralism is due to Pedersen 2006.
9. We will use small caps to denote concepts.
10. Sher 2005 raises the problem of mixed atomics. For the problems of
mixed compounds and inferences, see respectively Tappolet 2000 and
Tappolet 1997.
11. Versions of the double-counting objection have been presented by Asay
2018; Blackburn 1998, 2013; Dodd 2013, Horwich 1996; Pettit 1996;
Quine 1960; and Sainsbury 1996.
12. Wright 1998. See also Wright 1992 (Chap. 1, II) and Lynch 2004b: 386.
13. Beall and Restall 2006.
14. Field 2009.
15. In attributing a logical relativist view to Beall and Restall, we follow
Shapiro 2014.
16. It may seem odd to attribute both logical generalism and logical relativ-
ism to Beall and Restall. For, isn’t the idea that logic is general in tension
with the idea that it is relative? No. Logical generalism, as we have char-
acterized it, concerns whether a given logic legitimately issues (in)valid-
ity verdicts across the board (in the technical sense of legitimacy tied to
Generalized Tarski’s Thesis and the constraints of necessity, formality,
and normativity). For logics that are general in this sense, there is a fur-
ther issue as to the semantic status of their (in)validity verdicts—in par-
ticular, whether those verdicts have absolute or relativized truth-values.
17. Carnap 1937, 1950.
18. For a comprehensive introduction to substructural logics, see Restall
2000.
19. Lynch 2009, Pedersen 2014. The version of logical pluralism presented
in Shapiro 2014 can also be regarded as a kind of domain-based logical
Introduction 27
pluralism, although the various domains that Shapiro considers are all
mathematical.
20. The realism/anti-realism debate is, of course, one of the major themes in
Dummett’s corpus of work. See e.g., the essays in Dummett 1978.
21. See Lynch 2009 for details. The set of intuitionistic validities is a proper
subset of the set of classical validities, and hence, (mod) delivers the ver-
dict that the logic of any mixed compound or inference is intuitionistic
logic. In cases where there is overlap between two logics but no subset
relation, the intersection is identical to neither of the domain-specific
logics. This applies in the case of intuitionistic logic and relevant logic.
22. Smith forthcoming supports logical pluralism via logical contextualism.
While Beall and Restall never themselves present their logical pluralism
as a form of logical contextualism, Caret 2017 offers a contextualist read-
ing of their view and argues that it can be used to block a certain funda-
mental objection.
23. Barnard and Ulatowski 2013, Ulatowski 2017. See Kölbel 2008, 2013
for another empirically-based form of truth pluralism.
24. For details concerning MacFarlane’s assessment-sensitive relativism, see
MacFarlane 2014. Wyatt and Lynch 2016 suggest that MacFarlane is
committed to the kind of truth pluralism described.
25. Scharp 2013.
26. We’re blurring the line between use and mention here, but we trust that
the idea is clear.
27. See Wyatt and Lynch 2016 for further discussion of the relations between
truth pluralism and Scharp’s replacement theory. See also Beall 2013 and
Cotnoir 2013b for paradox-based motivations for truth pluralism.
References
Alston, W. 2005. Beyond ‘Justification’: Dimensions of Epistemic Evaluation.
Ithaca: Cornell University Press.
Asay, J. 2018. Putting Pluralism in Its Place. Philosophy and Phenomenological
Research 96 (1): 175–191.
Barnard, R., and J. Ulatowski. 2013. Truth, Correspondence, and Gender.
Review of Philosophy and Psychology 4: 621–638.
Beall, J.C. 2000. On Mixed Inferences and Pluralism About Truth Predicates.
The Philosophical Quarterly 50: 380–382.
———. 2013. Deflated Truth Pluralism. In Pedersen & Wright 2013a,
323–338. New York: Oxford University Press.
———. 2014. Strict-Choice Validities: A Note on a Familiar Pluralism.

Erkenntnis 79: 301–307.
Beall, J.C., and G. Restall. 2000. Logical Pluralism. Australasian Journal of
Philosophy 78: 475–493.
———. 2001. Defending Logical Pluralism. In Logical Consequence: Rival
Approaches Proceedings of the 1999 Conference of the Society of Exact Philosophy,
ed. J. Woods and B. Brown, 1–22. Stanmore: Hermes.
———. 2006. Logical Pluralism. Oxford: Oxford University Press.
Blackburn, S. 1998. Wittgenstein, Wright, Rorty and Minimalism. Mind 107:
157–181.
———. 2013. Deflationism, Pluralism, Expressivism, Pragmatism. In Pedersen
& Wright 2013a, 263–277. New York: Oxford University Press.
Blanshard, B. 1939. The Nature of Thought. London: George Allen and Unwin.
Bueno, O., and S.A. Shalkowski. 2009. Modalism and Logical Pluralism. Mind
118: 295–321.
Burge, T. 2003. Perceptual Entitlement. Philosophy and Phenomenological
Research 67: 503–548.
Caret, C. 2017. The Collapse of Logical Pluralism Has Been Greatly Exaggerated.
Carnap, R. 1937. The Logical Syntax of Language. New York: Harcourt, Brace
and Company.
———. 1950. Empiricism, Semantics, and Ontology. Revue Internationale de
Philosophie 4: 20–40.
Carnielli, W., and M.E. Coniglio. 2016. Combining Logics. In The Stanford
Encyclopedia of Philosophy, ed. E.N. Zalta. https://plato.stanford.edu/
archives/win2016/entries/logic-combining/.
Ciprotti, N., and L. Moretti. 2009. Logical Pluralism Is Compatible with
Monism About Metaphysical Modality. Australasian Journal of Philosophy 87:
275–284.
Cook, R.T. 2010. Let a Thousand Flowers Bloom: A Tour of Logical Pluralism.
Philosophy Compass 5: 492–504.
———. 2011. Alethic Pluralism, Generic Truth and Mixed Conjunctions. The
Philosophical Quarterly 61: 624–629.
———. 2014. Should Anti-Realists Be Anti-Realists About Anti-Realism?
Cotnoir, A. 2009. Generic Truth and Mixed Conjunctions: Some Alternatives.
Analysis 69: 473–479.
———. 2013a. Validity for Strong Pluralists. Philosophy and Phenomenological
Research 86: 563–579.
Introduction 29
———. 2013b. Pluralism and Paradox. In Pedersen & Wright 2013a, 339–350.
New York: Oxford University Press.
David, M. 1994. Correspondence and Disquotation: An Essay on the Nature of
Truth. Oxford: Oxford University Press.
———. 2005. On Truth Is Good. Philosophical Books 46: 292–301.
———. 2013. Lynch’s Functionalist Theory of Truth. In Pedersen & Wright
2013a, 42–68. New York: Oxford University Press.
Devitt, M. 1984. Realism and Truth. Oxford: Blackwell.
Dodd, J. 2013. Deflationism Trumps Pluralism! In Pedersen & Wright 2013a,
Dummett, M. 1978. Truth and Other Enigmas. Cambridge, MA: Harvard
University Press.
Edwards, D. 2008. How to Solve the Problem of Mixed Conjunctions. Analysis
68: 143–149.
———. 2009. Truth Conditions and the Nature of Truth: Resolving Mixed
Conjunctions. Analysis 69: 684–688.
———. 2011. Simplifying Alethic Pluralism. The Southern Journal of Philosophy
49: 28–48.
———. 2012a. On Alethic Disjunctivism. Dialectica 66: 200–214.
———. 2012b. Alethic vs. Deflationary Functionalism. International Journal of
Philosophical Studies 20: 115–124.
———. 2013a. Truth, Winning, and Simple Determination Pluralism. In
Pedersen & Wright 2013a, 113–122. New York: Oxford University Press.
———. 2013b. Truth as a Substantive Property. Australasian Journal of
———. 2018. The Metaphysics of Truth. Oxford: Oxford University Press.
Eklund, M. 2009. Carnap and Ontological Pluralism. In Metametaphysics, ed.
D. Chalmers, D. Manley, and R. Wasserman, 130–156. Oxford: Oxford
University Press.
———. 2012. Multitude, Tolerance and Language-Transcendence. Synthese
187: 833–847.
Engel, P. 2013. Alethic Functionalism and the Norm of Belief. In Pedersen &
Wright 2013a, 69–86. New York: Oxford University Press.
———. 1986. The Deflationary Conception of Truth. In Fact, Science and
Morality, ed. G. MacDonald and C. Wright, 55–117. Oxford: Basil Blackwell.
———. 1994a. Deflationist Views of Meaning and Content. Mind 103:
249–285.
———. 1994b. Disquotational Truth and Factually Defective Discourse.
Philosophical Review 103: 405–452.
———. 2009. Pluralism in Logic. The Review of Symbolic Logic 2: 342–359.

Goddu, G.C. 2002. What Exactly Is Logical Pluralism? Australasian Journal of
Goldman, A. 1988. Strong and Weak Justification. Philosophical Perspectives 2:
51–69.
Grover, D. 1992. A Prosentential Theory of Truth. Princeton: Princeton University
Press.
Hjortland, O.T. 2013. Logical Pluralism, MeaningVariance, and Verbal
Disputes. Australasian Journal of Philosophy 91: 355–373.
Horwich, P. 1996. Realism Minus Truth. Philosophy and Phenomenological
Research 56: 877–881.
———. 1998. Truth. 2nd ed. Oxford: Oxford University Press.
Humberstone, L. 2009. Logical Pluralism. Australasian Journal of Philosophy 87:
162–168.
James, W. 1907. Pragmatism: A New Name for some Old Ways of Thinking.
Cambridge, MA: Harvard University Press.
———. 1909. The Meaning of Truth. Cambridge, MA: Harvard University
Press.
Joachim, H. 1906. The Nature of Truth. Oxford: Clarendon Press.
Keefe, R. 2014. What Logical Pluralism Cannot Be. Synthese 191: 1375–1390.
Kölbel, M. 2008. ‘True’ as Ambiguous. Philosophy and Phenomenological Research
77: 359–384.
———. 2013. Should We Be Pluralists About Truth? In Pedersen & Wright
Kouri, T. 2016. Restall’s Proof-Theoretic Pluralism and Relevance Logic.
Kouri, T., and S. Shapiro. Forthcoming. Logical Pluralism and Normativity. To
appear in Inquiry.
Lynch, M.P. 2000. Alethic Pluralism and the Functionalist Theory of Truth.
Acta Analytica 15: 195–204.
———. 2001. A Functionalist Theory of Truth. In The Nature of Truth, ed.
M. Lynch, 723–749. Cambridge, MA: MIT Press.
———. 2004a. Minimalism and the Value of Truth. The Philosophical Quarterly
54: 497–517.
———. 2004b. Truth and Multiple Realizability. Australasian Journal of
———. 2005a. True to Life. Cambridge, MA: MIT Press.
———. 2005b. Reply to Critics. Philosophical Books 46: 331–342.
———. 2006. ReWrighting Pluralism. The Monist 89: 63–84.
Introduction 31
———. 2008. Alethic Pluralism, Logical Consequence, and the Universality of

Reason. Midwest Studies in Philosophy 32: 122–140.
———. 2009. Truth as One and Many. Oxford: Oxford University Press.
———. 2013. Three Questions for Truth Pluralism. In Pedersen & Wright
MacFarlane, J. 2014. Assessment Sensitivity: Relative Truth and its Applications.
Oxford: Oxford University Press.
McDaniel, K. 2017. The Fragmentation of Being. Oxford: Oxford University
Press.
Newhard, J. 2013. Four Objections to Alethic Functionalism. Journal of
Philosophical Research 38: 69–87.
———. 2017. Plain Truth and the Incoherence of Alethic Functionalism.
Synthese 194: 1591–1611.
Newman, A. 2007. The Correspondence Theory of Truth: An Essay on the
Metaphysics of Predication. Cambridge: Cambridge University Press.
Payette, G., and N. Wyatt. 2018. Logical Pluralism and Logical Form. Logique
et Analyse 61: 25–42.
Pedersen, Nikolaj J.L.L. 2006. What Can the Problem of Mixed Inferences
Teach Us About Alethic Pluralism? The Monist 89: 103–117.
———. 2010. Stabilizing Alethic Pluralism. The Philosophical Quarterly 60:
92–108.
———. 2012a. True Alethic Functionalism? International Journal of Philosophical
Studies 20: 125–133.
———. 2012b. Recent Work on Alethic Pluralism. Analysis 72: 588–607.
———. 2014. Pluralism × 3: Truth, Logic, Metaphysics. Erkenntnis 79:
259–277.
———. 2017. Pure Epistemic Pluralism. In Epistemic Pluralism, ed. A. Coliva
and Nikolaj J.L.L. Pedersen, 47–92. London: Palgrave Macmillan.
Pedersen, Nikolaj J.L.L., and D. Edwards. 2011. Truth as One(s) and Many: On
Lynch’s Alethic Functionalism. Analytic Philosophy 52: 213–230.
Pedersen, Nikolaj J.L.L., and C.D. Wright. 2010. Truth, Pluralism, Monism,
Correspondence. In New Waves in Truth, ed. Nikolaj J.L.L. Pedersen and
C.D. Wright, 205–217. New York: Palgrave Macmillan.
——— 2012. Pluralist Theories of Truth. In The Stanford Encyclopedia of
Philosophy, ed. E.N. Zalta. http://plato.stanford.edu/archives/spr2013/
entriesruthpluralist/.
———, eds. 2013a. Truth and Pluralism: Current Debates. New York: Oxford
University Press.
——— 2013b. Pluralism About Truth as Alethic Disjunctivism. In Pedersen &

Peirce, C.S. 1878. How to Make Our Ideas Clear. Reprinted in The Essential
Peirce, ed. N. Houser and C. Kloesel, 124–141, vol. 1, 1992. Bloomington:
Indiana University Press.
Pettit, P. 1996. Realism and Truth: A Comment on Crispin Wright’s Truth and
Objectivity. Philosophy and Phenomenological Research 56: 883–890.
Priest, G. 2001. Logic: One or Many? In Logical Consequence: Rival Approaches
Proceedings of the 1999 Conference of the Society of Exact Philosophy, ed.
J. Woods and B. Brown, 23–38. Stanmore: Hermes.
———. 2008. Logical Pluralism Hollandaise. The Australasian Journal of Logic
6: 210–214.
———. 2014. Logical Pluralism: Another Application for Chunk and Permeate.
Putnam, H. 1981. Reason, Truth and History. Cambridge: Cambridge University
Press.
Quine, W.V.O. 1960. Word and Object. Cambridge, MA: MIT Press.
Quine, W.V.O. 1970. Philosophy of Logic. Englewood Cliffs: Prentice Hall.
Ramsey, F. 1927. Facts and Propositions. Proceedings of the Aristotelian Society
Supplementary 7: 153–170.
Rasmussen, J. 2014. Defending the Correspondence Theory of Truth. Cambridge:
Cambridge University Press.
Read, S. 2006a. Monism: The One True Logic. In A Logical Approach to
Philosophy: Essays in Memory of Graham Solomon, 193–209. Berlin: Springer.
———. 2006b. Review of J.C. Beall and G. Restall, Logical Pluralism. Notre
Dame Philosophical Reviews.
Restall, G. 2000. An Introduction to Substructural Logics. London: Routledge.
———. 2001. Constructive Logic, Truth and Warranted Assertability. The
———. 2002. Carnap’s Tolerance, Meaning, and Logical Pluralism. The Journal
of Philosophy 99: 426–443.
———. 2012. Anti-Realist Classical Logic and Realist Mathematics. In The
Realism Antirealism Debate in the Age of Alternative Logics, Logic, Epistemology,
and the Unity of Science 23, ed. S. Rahman, G. Primiero, and M. Marion,
269–284. Dordrecht: Springer Netherlands.
———. 2014. Pluralism and Proofs. Erkenntnis 79: 279–291.
Russell, B. 1912. The Problems of Philosophy. Reprinted by Oxford: Oxford
University Press, 1971.
Russell, G. 2008. One True Logic? Journal of Philosophical Logic 37: 593–611.
Introduction 33
———. 2014. Logical Pluralism. In The Stanford Encyclopedia of Philosophy, ed.

E.N. Zalta. http://plato.stanford.edu/archives/spr2014/entries/logicalplural-
ism/.
Sainsbury, M. 1996. Review: Crispin Wright: Truth and Objectivity. Philosophy
and Phenomenological Research 56: 899–904.
Scharp, K. 2013. Replacing Truth. Oxford: Oxford University Press.
Shapiro, S. 2011. Truth, Function, and Paradox. Analysis 71: 38–44.
———. 2014. Varieties of Logic. New York: Oxford University Press.
Sher, G. 1998. On the Possibility of a Substantive Theory of Truth. Synthese 117:
133–172.
———. 2005. Functional Pluralism. Philosophical Books 46: 311–330.
———. 2013. Pluralism, Correspondence, and Description. In Pedersen &
———. 2016. Epistemic Friction: An Essay on Knowledge, Truth, and Logic.
———. Forthcoming. Pluralism and Normativity in Truth and Logic. To
appear in American Philosophical Quarterly.
Smith, P. Simard. Forthcoming. Assessment Context-Sensitive Logical Claims.
To appear in Inquiry. doi: https://doi.org/10.1080/0020174X.2017.1402698.
Stewart-Wallace, A. 2016. An Old Solution to the Problem of Mixed Atomics.
Strawson, P.F. 1950. Truth. Proceedings of the Aristotelian Society Supplementary
24: 129–156.
Tappolet, C. 1997. Mixed Inferences: A Problem for Pluralism About Truth
Predicates. Analysis 57: 209–211.
———. 2000. Truth, Pluralism, and Many-Valued Logic: A Reply to Beall.
———. 2010. Truth as One and Many, by Michael P. Lynch. Mind, 119:
1193–1198.
Terrés, P. Forthcoming. From Natural to Formal Language: A Case for Logical
Pluralism. To appear in Topoi.
Turner, J. 2010. Ontological Pluralism. The Journal of Philosophy 107: 5–34.
———. 2012. Logic and Ontological Pluralism. The Journal of Philosophical
Logic 41: 419–448.
Ulatowski, J. 2017. Commonsense Pluralism About Truth: An Empirical Defence.
Cham: Palgrave Macmillan.
van Benthem, J. 2008. Logical Dynamics Meets Logical Pluralism? The
Australasian Journal of Logic 6: 182–209.
Varzi, A. 2002. On Logical Relativity. Philosophical Issues 12: 197–219.
Vision, G. 2004. Veritas: The Correspondence Theory and Its Critics. Cambridge,
MA: MIT Press.
Williamson, T. 1994. Critical Study of Truth and Objectivity. International
Journal of Philo-sophical Studies 30: 130–144.
Wittgenstein, L. 1921. Tractatus Logico-Philosophicus. In Annalen der
Naturphilosophie. English trans. D.F. Pears and B.F. McGuinness, 1961.
London: Routledge & Kegan Paul.
Wright, C. 1992. Truth and Objectivity. Cambridge, MA: Harvard University Press.
———. 1996a. Précis of Truth and Objectivity. Philosophy and Phenomenological
Research 56: 863–868.
———. 1996b. Response to Commentators. Philosophy and Phenomenological
Research 56: 911–941.
———. 1998. Truth: A Traditional Debate Reviewed. Canadian Journal of
Philosophy 24 (suppl. vol.): 31–74.
———. 2001. Minimalism, Deflationism, Pragmatism, Pluralism. In The
Nature of Truth, ed. M.P. Lynch, 751–787. Cambridge, MA: MIT Press.
———. 2004. Warrant for Nothing (and Foundations for Free)? Proceedings of
the Aristotelian Society 78: 167–212.
Wright, C.D. 2005. On the Functionalization of Pluralist Approaches to Truth.
Synthese 145: 1–28.
———. 2010. Truth, Ramification, and the Pluralist’s Revenge. Australasian
Journal of Philosophy 88: 265–283.
———. 2012. Is Pluralism About Truth Inherently Unstable? Philosophical
Studies 159: 89–105.
Wright, C. 2013. A Plurality of Pluralisms? In Pedersen & Wright 2013a,
Wyatt, N. 2004. What Are Beall and Restall Pluralists About? Australasian
Wyatt, J. 2013. Domains, Plural Truth, and Mixed Atomic Propositions.
———. 2016. The Many (Yet Few) Faces of Deflationism. The Philosophical
Quarterly 66: 362–382.
Wyatt, J., and M.P. Lynch. 2016. From One to Many: Recent Work on Truth.
American Philosophical Quarterly 53: 323–340.
Young, J.O. 2001. A Defence of the Coherence Theory of Truth. The Journal of
Yu, A. 2017a. Logic for Pluralists. Journal of Philosophy 114: 277–302.
———. 2017b. Can Alethic Pluralists Maintain Compositionality? The
Truth: One or Many or Both?
Dorit Bar-On and Keith Simmons
1 Introduction: Truth—the One

and the Many
Truth pluralism is a metaphysical theory of the nature of truth. The plu-
ralist rejects the deflationist claim that truth is at best a ‘shallow’, insub-
stantial property. Indeed, the pluralist embraces a plurality of substantive
truth properties (such as superwarrant, supercoherence, or correspondence),
appropriate to different domains of discourse.
What is the intuitive motivation for pluralism? In a recent exposition,
Crispin Wright offers the following thought: Perhaps protagonists in the
traditional debates over the nature of truth (e.g., proponents of coher-
ence, pragmatist, or correspondence theories of truth) have failed to reach
agreement because their views were each appropriate to some domains
but not all. Accepting that “all the protagonists were saying locally plau-
sible things…”, one could, by embracing truth pluralism, not only
D. Bar-On • K. Simmons (*)

e-mail: dorit.bar-on@uconn.edu; keith.simmons@uconn.edu

https://doi.org/10.1007/978-3-319-98346-2_2
36 D. Bar-On and K. Simmons
accommodate the failures of specific theories of truth to extend beyond

certain paradigms, but also “resist deflationary accounts of truth”.1
As a second motivation for truth pluralism, Wright mentions
Dummett’s failure to provide a viable anti-realist construal of the mean-
ings of sentences that could replace truth-conditional semantics (which,
Dummett thinks, is unsustainable). Recognizing a plurality in kinds of
truth appropriate for different domains, Wright thinks, may allow us to
capture differences between realist and anti-realist conceptions, while
adhering to a truth-conditional semantics across the board.
The idea, then, is that pluralizing truth promises to have the following
advantages:
1. It can allow us to preserve the intuitive plausibility of the different

substantive theories of truth; and
2. It can allow us to reconstruct the realist/anti-realist debate in terms of
kinds of truth, rather than in terms of kinds of meaning.
As regards (1), one might worry that the pluralist will inherit all the
main problems of the various traditional substantivist theories of truth,
not just their advantages. In addition, however, we think that pluralism
faces distinctive difficulties, having to do, specifically, with supposing
truth to be many. In what follows, we’ll be focusing on moderate plural-
ism—a view that acknowledges, in addition to diverse truth properties, a
single property of truth (so that truth emerges as both one and many).
After presenting a recent, influential version of moderate pluralism (Sect.
2), we will articulate the main difficulties we see with this view (Sect. 3).
We believe that the difficulties we raise extend to other pluralist views,
but we do not address other pluralisms, except in passing.
Now, given the difficulties, it is natural to wonder whether there is a
metaphysically more conservative way to accommodate the core intu-
itions that originally motivated pluralism. In Sect. 4, we explore one such
way. To anticipate, we will suggest, first, that the difficulties raised in Sect.
3 speak strongly in favor of thinking of truth as one, but not also many.
At the same time, we agree with Wright and others that a certain plurality
must be recognized, if we are to allow for substantive metaphysical
debates between realists and anti-realists in various domains. However,
we think such debates should not be construed either in terms of kinds of

meaning or in terms of kinds of truth. For, properly construed, they con-
cern our conceptions of, for example, mathematical or ethical or comic
reality, or different kinds of ‘matters of fact’.
On the conservative alternative we consider, there is only one way for
sentences, propositions, beliefs, and so on, to be true; though when they
are, there may be multiple ways things can be to make them so. A joke’s
being funny is a very different sort of thing from someone’s act being
wrong, a number’s being divisible by 2, or a chair’s being red—and then
an act’s being morally wrong is a different sort of thing from it being
socially wrong; and then again a chair’s being red is a different sort of
thing from someone’s hair being red! The plurality here is at the level of
the world. The appeal to a metaphysical plurality of truth properties con-
tributes no explanatory power beyond what can be got by focusing on the
plurality in kinds of worldly conditions that are apt to render the claims
made in different domains true. Thus, if we are right, there is nothing to
prevent a truth monist—or, for that matter, even a deflationist—from
recognizing a plurality that will capture realist and anti-realist concep-
tions as appropriate to different domains.
2 Moderate Pluralism
Here we focus on moderate pluralism, the kind of pluralism that takes
truth to be both one and many. This is in contrast to strong pluralism,
where truth is many, but not one as well. It’s fair to say that most plural-
ists take moderate pluralism to be preferable, because it has ready answers
to certain objections that have been raised against pluralism. Here are
two:
(1) There appears to be some unity to the ways in which propositions can
be true—for example, it seems that any truth property would supply
a criterion for the correctness of beliefs, and it seems that any truth
property would be preserved by valid inference. So, there should be a
generic truth property to capture this unity.
(2) Consider the conjunction <Speeding is illegal and there are moun-
tains>.2 This is an example of mixed discourse: the two conjuncts are
drawn from two different domains of discourse. The first conjunct is
from the legal domain, and we may suppose that it does not have the
correspondence property. Rather, it is superwarranted: it is warranted
at the present stage of inquiry, and would remain warranted without
defeat at every successive stage of inquiry. The second conjunct is an
atomic sentence from the geologic domain, and we may suppose that
it does have the correspondence property—it is true because it cor-
responds to geologic reality. But what about the conjunction as a
whole? It is true. However, we cannot say that it has the correspon-
dence property, because its first conjunct doesn’t. Perhaps we can say
that the conjunction as a whole is superwarranted—we might allow
that the second conjunct has the correspondence property and is
superwarranted. But attributing superwarrant to the conjunction
doesn’t adequately explain why it is true, because it doesn’t adequately
explain the truth of the second conjunct. Here, the moderate plural-
ist has an answer: the truth of the conjunction is explained in terms
of generic truth.
Once we have truth as one and many, we need some account of how
the many relate to the one. Moderate pluralists offer a variety of accounts
of this relation—for example, one account takes the relation to be real-
ization, and another account takes it to be manifestation.3 Either way, we
start with certain core principles which characterize the concept of truth.
According to Michael Lynch, for example, there are at least three core
principles, or “core truisms”:
Objectivity: The belief that p is true if, and only if, with respect to the
belief that p, things are as they are believed to be. (Lynch
2009, p. 8).
Norm of Belief: It is prima facie correct to believe that p if, and only if,
the proposition that p is true. (op. cit., p. 10).
End of Inquiry: Other things being equal, true beliefs are a worthy goal
of inquiry. (op. cit., p. 12).
Most pluralists are monists about the concept of truth, and these core tru-
isms are taken to pin down this unique concept. Then, at the level of
properties, we have, according to the moderate pluralist, the generic
property of truth, call it truth-as-such, and the various properties which
relate to it—perhaps by realizing this generic property, perhaps by mani-
festing it.4
To focus the discussion, we consider Lynch’s manifestation pluralism—
though we think that what we have to say carries over to other versions of
pluralism. According to manifestation pluralism, truth-as-such is the
property that satisfies the core principles as a matter of conceptual necessity.
Now consider the manifestation relation: What is it for, say, superwarrant
to manifest truth-as-such? It is for superwarrant to have among its fea-
tures those that truth-as-such has essentially. So, suppose that a truth
about traffic laws has the superwarrant property. For the superwarrant
property to manifest truth-as-such is for superwarrant to satisfy the core
truisms, since truth-as-such satisfies these principles essentially. In legal
discourse, say, superwarrant satisfies the core truisms, and so here super-
warrant manifests truth-as-such.
Clearly, truth-as-such manifests itself—the features that truth-as-such
has essentially are among the features that truth-as-such has. So, for
example, the proposition that speeding is illegal has two truth-manifesting
properties: truth-as-such and superwarrant. Following Lynch, call a
proposition that has two truth-manifesting properties an unplain truth,
and a proposition that is only true-as-such a plain truth.
To see manifestation pluralism in action, consider the mixed conjunc-
tion <Speeding is illegal and there are mountains>. This conjunction as a
whole does not possess superwarrant or correspondence. Rather it is
plainly true: it is true-as-such, and has no other truth property. The plain
truth of the whole conjunction supervenes on the unplain truth of its
conjuncts. Each separate conjunct is what Lynch calls strongly grounded—
the truth-as-such of each conjunct is manifested by a further truth prop-
erty. <Speeding is illegal> is true-as-such because the proposition is
superwarranted; <there are mountains> is true-as-such because the prop-
osition corresponds to the facts. In contrast, the truth-as-such of the
whole conjunction is not manifested by any further truth property of the
conjunction itself. The conjunction itself is a plain truth that is weakly
grounded in the unplain truth of its components.
3 oderate Pluralism and Truth

M
Attributions
Our question now is: What should the truth pluralist say about truth
attributions? Lynch is one pluralist who has explicitly addressed this ques-
tion, so we will be focusing on what Lynch has to say about it.
Observe first the familiar equivalence between the proposition that p
and the truth attribution . There is some disagreement about
how exactly to characterize this equivalence, but we think it can be agreed
that the equivalence
 is true iff p
is at least a necessary equivalence: the two sides of the biconditional

have the same truth values in every possible world. Now consider the
proposition that speeding is illegal. According to manifestation plural-
ism—and the second-order functionalist—this proposition is true-as-
such, and it has the superwarrant property, where the superwarrant
property manifests truth-as-such. <Speeding is illegal> is an unplain
truth.
According to manifestation pluralism, things are different with the
truth attribution <<speeding is illegal> is true>. This is a plain truth: it is
true-as-such, but has no further truth property. In particular, it does not
have the superwarrant property. The attribution <<speeding is illegal> is
true> is weakly grounded in the unplain truth of <speeding is illegal>.
One might be tempted to suppose that the truth attribution <<speed-
ing is illegal> is true> should inherit the truth properties of <speeding is
illegal>. But Lynch takes the inheritance view to be ‘hopeless’.5 It is hope-
less, Lynch argues, in view of the fact that we use truth to generalize over
propositions. Suppose I say:
(G) Everything George says is true.
This is a universal generalization over propositions:
For every proposition p, if George says that p then p is true.

Now suppose that George expresses propositions about all kinds of

subject matter—about speeding, about prime numbers, about moun-
tains. According to the inheritance view, truth attributions inherit their
truth properties from their attributees. But the attributees of my truth
attribution (G) have a variety of truth properties—superwarrant, super-
coherence, correspondence. So (G) will have all those truth properties
together—and that is not possible. That is not to say that there cannot be
truths that possess all three properties; we are not ruling out the possibil-
ity that there are truths with the correspondence property that happen
also to be superwarranted and supercoherent. But (G) ranges over propo-
sitions that aren’t correspondence-apt. So (G) itself is not correspondence-
apt. So (G) cannot inherit all the alethic properties possessed by the
propositions over which it ranges. According to the manifestation plural-
ist, then, the inheritance view must be rejected.
But we think the inheritance view, or something very like it, is right.
And it is hard to see how the manifestation pluralist—or the second-
order functionalist—can resist it. Consider the proposition <speeding is
illegal>. Recall that a belief is superwarranted just when believing that p
is warranted at some stage of inquiry and would remain warranted with-
out defeat at every successive stage of inquiry. A proposition that is true
in this way is one that is warranted to believe, and remains warranted to
believe however our information is expanded or improved. Assuming
with the manifestation pluralist that the proposition <speeding is illegal>
is superwarranted, it follows that the truth attribution <<speeding is ille-
gal> is true> is also superwarranted. No further information can under-
mine either <speeding is illegal> or <<speeding is illegal> is true> without
undermining the other; if one remains warranted through every succes-
sive stage of inquiry, so does the other. This is because anyone who grasps
the concept of truth will accept the conceptual equivalence of <speeding
is illegal> and <<speeding is illegal> is true>.6 In general, is
superwarranted if and only if < is true> is superwarranted. Similarly,
if we work with the truth operator on propositions: is superwar-
ranted if and only if <it is true that p> is superwarranted. It follows, then,
that <speeding is illegal> is true and <it is true that speeding is illegal> are
both true-as-such and superwarranted. They are not plainly true.
This conclusion remains even if we accept that there is a sense in which

the truth of < is true> is dependent on the truth of , and not vice
versa. Both and < is true> are superwarranted, given the neces-
sary equivalence between and < is true>. The inheritance view is
sensitive to this dependence: < is true> inherits its truth properties
from , and not vice versa. But, to state the obvious, if <<speeding is
illegal> is true> inherits the property of superwarrant, it has the prop-
erty—it is not plainly true.
The case of supercoherence runs parallel. Following Lynch, we can char-
acterize coherence and supercoherence along the following lines. The
coherence of a framework is, of course, a complicated matter, involving
many ingredients (consistency, simplicity, completeness, mutual explana-
tory support, and so on). Assuming that we have a characterization of a
coherent framework, suppose we go on to say that a proposition 
coheres with a framework F if including in F would make F more
coherent.7 Now introduce the notion of supercoherence as follows: 
supercoheres with F if and only if coheres with F at some stage of
inquiry and would continue to do so without defeat through all succes-
sive and additional improvements to F. That is, supercoheres with F if
adding to F makes F more coherent, and will continue to do so
through all subsequent improvements to F. Now consider the proposition
 and the truth attribution < is true>. Suppose that superco-
heres with F. So, when we add to F, it makes F more coherent, and
will continue to do so through all improvements to F. But suppose instead
that we add < is true>. Then < is true> will supercohere with
F. Adding < is true> instead of will also make F more coherent,
and will continue to do so through all improvements. If either one super-
coheres with F, then so will the other. But then < is true> is not a
plain truth. And, similarly, <it is true that p> is not a plain truth.
What about correspondence? Here, the discussion cannot be clear-cut,
just because the relation of correspondence is (notoriously) unclear.
Suppose that corresponds to fact F. Given the correspondence
between and F, there will be a necessary connection between F and
the truth of . Any world that contains F is a world in which is
true; and any world that does not contain F is a world in which is
false. But given the necessary equivalence between and < is true>,
there will also be a necessary connection between F and the truth of <
is true>. That is, any world which contains F is a world in which < is
true> is true, and any world which does not contain F is a world in which
< is true> is false. It seems natural to take this as saying that there’s a
correspondence between < is true> and F. But then it follows that
< is true> is not plainly true—like , it is both true-as-such and has
the correspondence property.
At this point, the manifestation pluralist might appeal to the depen-
dence asymmetry between and < is true>: < is true> depends
for its truth on the truth of , but not vice versa. The truth attribution
< is true> is weakly grounded; but itself is strongly grounded.
Now this might support the thought that the correspondence between
< is true> and F is dependent on the correspondence between 
and F. < is true> corresponds to F only because (i) and < is
true> are necessarily equivalent, and (ii) corresponds to F. But even
if this is so, the correspondence between < is true> and F remains.
Even if this correspondence is in some way derivative, still < is true>
has the correspondence property—it is not plainly true. Again, we can
take the inheritance view as reflecting the derivative way in which <
is true> has the correspondence property: it inherits the property from its
attributee . And again we can state the obvious: If < is true>
inherits the correspondence property, it has the property.
The manifestation pluralist might claim that there’s a relevant differ-
ence between < is true> and . <<grass is green> is true> has a
proposition as its subject; <grass is green> has grass as its subject. Does this
make a difference to the correspondence relation? This depends on one’s
view of the equivalence between < is true> and . We might hold,
with Frege and many others, that the necessary equivalence between
< is true> and is an indication of the special, transparent way in
which the truth predicate works when applied to individual propositions.
The predicate ‘true’ is not an ordinary property-ascribing predicate. If we
accept that the truth predicate is transparent in this way, it is very hard to
see how and < is true> could differ in whatever truth properties
they do have—in particular, if one corresponds to F, so does the other.
If, on the other hand, one thought of ‘true’ as an ordinary property-
ascribing predicate, then, as a correspondence theorist, one should think
that < is true> corresponds to a fact involving the proposition that

 and the property truth (and not, say, a fact concerning grass and
greenness). Now < is true> corresponds to a different fact from 
−- one fact involves a proposition and the property of truth, the other
involves grass and greenness. But still < is true> has the correspon-
dence property. Again, according to the inheritance view, < is true>
will have the correspondence property in a derivative way. The proposi-
tion <grass is green> corresponds to the non-semantic facts, and that’s
why <<grass is green> is true> corresponds to the semantic fact that <grass
is green> is true. Still, <<grass is green> is true> has the correspondence
property. It is not plainly true.
We should observe again that truth attributions can take another form,
employing a sentential operator, as with <it is true that grass is green>.
Here there is no reference to propositions, and no use of a truth predicate
(property-ascribing or otherwise). And again we will have a necessary
connection, a correspondence, between the proposition <it is true that
grass is green> and the fact that grass is green.
To sum up, we’ve argued as follows:
(1) A truth attribution shares its truth properties, whatever they may be,
with its attributee.
In addition, we’ve taken the inheritance view to handle the natural

thought that the truth of < is true> depends on the truth of , but
not vice versa. And we’ve observed what is obvious, that the inheritance
view implies (1). Now, we also agree with Lynch that
(2) (1) is incompatible with a pluralist view of truth properties,
in virtue of truth’s role in generalizing over propositions, as exempli-

fied by (G) above. (Of course, given (2), and given that the inheritance
view implies (1), it follows that the inheritance view is also incompatible
with pluralism about truth properties.) If we are right about (1) and (2),
then it follows that we should reject pluralism about truth properties.
One further remark. The manifestation pluralist might take a leaf out
of the deflationist’s book, and regard a generalization such as (G) as
equivalent to an infinite conjunction of conditionals: If George says that
there are mountains then there are mountains, and if George says that
grass is green then grass is green, and … . Similarly, a generalization such
as ‘Something George said yesterday is true’ would be regarded as an
infinite disjunction. Then the problem of dealing with generalized truth
attributions reduces to the problem of dealing with truth-functional
compounds. The manifestation pluralist will say, as we saw above, that
truth-functional compounds are plainly true, while their atomic compo-
nents are unplainly true. This is so whether the compounds are mixed (as
with <Speeding is illegal and there are mountains>) or unmixed. Lynch
writes: “Compound propositions, mixed or not, are true because they are
plainly true”.8 The plain truth of a conjunction is a matter quite indepen-
dent of the discourses from which the conjuncts are drawn. But now
problems emerge for truth-functional compounds, in parallel with the
problems for truth attributions.
Suppose that at some stage of inquiry, both and <q> are separately
warranted, and would remain warranted without defeat at every succes-
sive stage of inquiry. Then it seems hard to deny that <p&q> is warranted
at the given stage and at every successive stage: that is, if and <q> are
separately superwarranted, then <p&q> is superwarranted. And if we
accept a sense in which the truth of <p&q> depends on the truth of 
and <q> separately, then this asymmetry is accommodated naturally by
the inheritance view, which preserves the superwarrant of <p&q>.
Similarly with supercoherence. If adding each of and <q> separately
to F makes F more coherent, and will continue to do so through all sub-
sequent improvements, then so will adding <p&q> to F. Suppose 
corresponds to fact F, and <q> to fact F*. Then the worlds which contain
both facts F and F* will be exactly those worlds in which <p&q> is true—
we have a correspondence between <p&q> and the facts. In each case, we
have an unmixed conjunction which is superwarranted, or supercoherent
or corresponds—they are not plainly true.
The manifestation pluralist’s claim is that any true conjunction is
plainly true. The plain truth of a conjunction is a matter supposedly inde-
pendent of the mixed or unmixed character of the conjunction. But now
we have seen that there are conjunctions that are not plainly true. So, the
manifestation pluralist’s argument for the plain truth of conjunctions has
broken down—we’ve now lost whatever motivation there was for suppos-
ing that, say, <Speeding is illegal and there are mountains> is plainly true.
And, as we saw earlier, the conjunction cannot possess the correspon-
dence property, given the first conjunct, and even if we allow that the
conjunction is superwarranted, this fails to explain the truth of the sec-
ond conjunct. The conjunction is true, but the manifestation pluralist
has no adequate account of its truth.
The point here connects to the case of truth attributions. Suppose
George only ever says two things: “Speeding is illegal” and “There are
mountains”. Now consider again:
(G) Everything George says is true.
Under the circumstances, (G) is equivalent to <<Speeding is illegal> is

true and <There are mountains> is true>, which is in turn equivalent to
<Speeding is illegal and there are mountains>. So, if manifestation plural-
ists don’t have an adequate account of the truth of this conjunction, they
will not have an adequate account of the truth of (G).
4 Wherein Plurality?
Difficulties with moderate pluralism, some independently motivated
commitments, and an attempt to accommodate the main motivations for
pluralism, lead us to propose the following set of desiderata for a view of
truth:
1. No wholesale deflationism (as per arguments we have provided

elsewhere).9
2. No commitment to quietism about realist/anti-realist debates.
3. Acknowledgement of some plurality, so as (a) to make room for sub-
stantive realist/anti-realist debates, and (b) to accommodate the idea
that, in some sense, no ‘one true size fits all’ when it comes to different
discourses.
4 . No plurality of truth properties (as per the above objections).

5. No commitment to a plurality of kinds of meaning.
We’ll address these desiderata in turn.

Regarding desideratum (1), we have argued in earlier work that the
concept of truth is richer than deflationists can allow—it has links to other
important ‘nodes’ in our conceptual scheme (such as meaning, belief, and
assertion) that cannot be ‘disquoted away’.10 We agree with Frege,
Davidson, Wright, and others, that the concept of truth is fundamental
in our conceptual scheme, and though it may not be given an analytic
definition, it can be illuminated by articulating its connections with other
concepts (e.g., in the style proposed by Strawson11). It is not clear to us,
however, whether accepting the conceptual robustness of truth commits
one to there being a substantive (or ‘nonabundant’) metaphysical prop-
erty denoted by the predicate ‘is true’ that has an underlying nature shared
by all and only true things. So, despite rejecting wholesale conceptual (as
well as linguistic deflationism), we have been prepared to leave the door
open (at least provisionally) for metaphysical deflationism: the claim that
nothing explanatory may be gained by invoking a property denoted by
the predicate ‘is true’ (and its equivalents). Here, however, we are only
concerned to deny that the conceptual robustness of truth requires recog-
nizing different kinds of truth, each appropriate in different domains, and
capturing diverse ways for statements to be true.
Turning to desideratum (2), some deflationists—notably Horwich—
maintain that “[t]he use of these labels [‘realism’ and ‘anti-realism’] within
philosophy is an unholy mess – to the point that they surely lack deter-
minate application”; indeed, Horwich says he would “like to avoid ‘isms’
altogether”.12 He thinks that in areas where anti-realism may seem tempt-
ing, “we can devise a coherent and attractive perspective combining the
most plausible contentions of the self-styled ‘realists’ with the most plau-
sible contentions of the self-styled ‘anti-realists’”, bypassing all concerns
about metaphysically ‘spooky’ facts.13 According to this ‘quietist’ view,
once we admit that it is legitimate to speak of truth, propositions, and
even beliefs and facts, in a given domain, there remains no philosophi-
cally significant issue of metaphysical relevance to be settled regarding the
discourse in question. We reject this quietism.
This leads directly to our next desideratum (3). As a default position,

we agree with Wright and others that there may well be a point to draw-
ing metaphysical distinctions among various areas of discourse, and that
we can be cautiously optimistic about the prospects of making philo-
sophical sense of debates between realists and anti-realists in various
domains. Moreover, we think that this requires acknowledging a certain
plurality.
However, we deny that satisfying desideratum (3) requires accepting a
plurality of truth properties. In addition to the objections we, and others,
have raised against moderate pluralism, there is a general worry about the
invocation of a plurality of truth properties. Suppose we ask: Why is cor-
respondence truth the right kind of truth for statements about middle-
size physical objects? Why is it not appropriate for ethics? Why is
superassertibility appropriate for comic discourse? And so on. Presumably,
the pluralist would agree that the reason has something to do with the
ontology of the relevant domain—with the relevant facts. Alethic plural-
ism is thus motivated by ontological pluralism, according to which
domains of discourse may be objective to different degrees, or exhibit
different degrees of mind-dependence. But if this is so, then it’s very
unclear what the appeal to diverse truth properties adds, explanatorily
speaking, to the already acknowledged ontological plurality. Hence desid-
eratum (4).
For related reasons, we propose desideratum (5). In agreement with
Wright, we think one should deny (contra Dummett) that the plurality
required for capturing realist/anti-realist debates is a plurality in kinds of
meaning, or one that holds “at the level of the propositions” expressed by
sentences in different discourses.14 Nothing in the semantic behavior of
sentences in different discourses suggests that they possess different kinds
of meaning. And, as with truth pluralism, if we acknowledge plurality at
the level of ontology, it is not clear what an appeal to diverse meanings
would add, explanatorily speaking.
As a default position, then, we think that the debates between realism
and anti-realism should be reconstructed neither in terms of kinds of
truth properties nor in terms of kinds of meaning. Instead, we suggest,
more conservatively, that the plurality be assigned to the relevant realms
of facts—to the worldly conditions that could render statements in given
domains true.
In a critique of Wright’s Truth & Objectivity, Pettit proposes something

along these lines:
Under the envisaged scenario, there remains only one sort of truth: that
which is defined by the platitudes-satisfying role. It is just that what truth
involves in one area – what realizes the appropriate role – may be different
from what it involves in another. The difference … will be explained by
reference to the different subject-matters: the different truth-conditions, and
the different truth-makers, in each discourse.15
Now, to gain traction against the truth pluralist, who insists on locat-
ing the relevant plurality in kinds of truth as opposed to kinds of meaning,
it is important to note that the notion of truth-conditions (as well as that
of subject-matter, and even content) is invoked in discussions of truth in
two ways that can—and, we submit, should—be separated.
A Davidsonian truth-conditional theory of meaning aims to yield as
theorems meaning-specifying biconditionals, such as
(W) “Wasser ist nass” is true if, and only if, water is wet.
The right-hand side of the biconditional is a sentence that can be used

by the theorist to specify the meaning of the sentence mentioned on the
left-hand side. Here, the right-hand side picks out the worldly condi-
tion—water’s being wet—under which the mentioned sentence is true.
However, as recognized by Davidson,16 that is not sufficient for the
biconditional to be meaning giving. Crucially, the right-hand side must
pick out that worldly condition in a way that is fit to capture the semantic
place occupied by the mentioned sentence. Contrast (W) with
(W′) “Wasser is nass” is true if, and only if, H2O is wet.
Although (W′) is a true biconditional, and its right-hand side picks

out the same worldly condition as the right-hand side of (W), (W′) is,
intuitively, not meaning-giving. We must here set aside the difficult ques-
tion whether—and how—a truth theory for a language L can, as Davidson
hoped, do all that we may expect of a theory of meaning for L. Still, we

must be careful with the familiar slogan: “The meaning of a sentence is
given by its truth-conditions”. In a Davidsonian theory of meaning,
when the meaning of a sentence is given by specifying its truth-conditions
(as does the theorem (W)), it matters how those conditions are picked
out. So, the notion of truth-conditions relevant to the familiar slogan is a
semantic one, to wit:
(i) truth-conditions as they figure in meaning-giving biconditionals—

worldly conditions picked out in a way fit for specifying the meaning of
a given sentence.17
However, when Pettit invokes truth-conditions as obviating the need

for a plurality in kinds of truth, he is presumably thinking of truth-
conditions in a different way (as suggested by his reference to ‘truth-
makers’18). The notion of truth-conditions relevant here is a metaphysical
one, to wit:
(ii) worldly conditions (objects, properties, states of affairs—if any) iden-

tified by a metaphysician as revealing the underlying nature, onto-
logical constitution, and so on, of elements in the relevant domain.
Our aim in what follows will be to take a stab at clarifying the distinc-
tion we have in mind, and to explain how this distinction bears on what
we take to be the best way to accommodate the motivations behind
pluralism.19
Begin with truth-conditions as they figure in meaning-giving bicondi-
tionals. Truth-conditions thus understood feature in Davidson’s seminal
“Truth and Meaning”, where he proposes that a theory of meaning for a
natural language should take the form of a theory of truth for that lan-
guage.20 Meaning-giving biconditionals are designed to capture the
logical place occupied by individual sentences in the whole (potentially
infinite) network of sentences of a language. Meaning-giving bicondi-
tionals are relatively neutral, metaphysically speaking. This relative neu-
trality is well-captured by Davidson himself, when he says:
If we suppose questions of logical grammar settled, sentences like ‘Bardot

is good’ raise no special problems for a truth definition. The deep differ-
ences between descriptive and evaluative (emotive, expressive, etc.) terms
do not show here. … we ought not to boggle at ‘“Bardot is good” is true if
and only if Bardot is good’; in a theory of truth, this consequence should
follow with the rest, keeping track, as must be done, of the semantic loca-
tion of such sentences in the language as a whole—of their relation to
generalizations, their role in such compound sentences as ‘Bardot is good
and Bardot is foolish’, and so on. What is special to evaluative words is
simply not touched: the mystery is transferred from the word ‘good’ in the
object-language to its translation in the meta-language.21
On the Davidsonian picture presented in “Truth and Meaning”, the

truth-conditions cited in meaning-giving biconditionals have the follow-
ing important features:
(1s22) They are arrived at through systematic logico-semantic analysis of
the relevant object language. To do its job, such an analysis will exhibit
the truth-conditions of sentences as a function of the semantic values of
their parts, in a way that reveals how they systematically interact with
other sentences and sentence parts, how they embed in various construc-
tions (such as conditionals, modal and propositional attitude contexts),
and so on.
(2s) Calling the worldly conditions that feature in the biconditionals
‘truth-conditions’ seems apt, given the involvement of truth in recovering
logical structure, entailment relations among sentences, and so on. But to
play its role here ‘truth’ need not be understood as denoting any robust
(or specific) metaphysical property. At the same time, the use of the
Tarskian truth schema to specify sentences’ meanings in no way commits
one to deflationism about truth, and on some views is incompatible with
it. (As is well known, Davidson himself has argued against deflationism
about truth, for reasons we cannot rehearse here.23)
(3s) Davidsonian semantic analysis can yield surprising results. It can
reveal covert ambiguities and context-sensitivity; it can assign sentences
logical forms that diverge radically from surface forms (think of Russell’s
analysis of sentences containing definite descriptions, or Davidson’s own
proposed analysis of action sentences);24 it can exhibit unexpected rela-
tions among sentences; and so on. The assignment of truth-conditional
meaning to mentioned sentences is not a trivial matter; it is subject to

substantive constraints, and the semantic contents it yields are not ‘thin
propositions’ that are merely the shadows of all syntactically well-formed
sentences. (e.g., the theory may rule out as meaningless sentences such as
“Twas brillig and the slithy toves did gyre and gimble”, or even “Colorless
green ideas sleep furiously”.) Thus, a Davidsonian theory is not a disquo-
tational theory of meaning.
(4s) However, there is no presumption that a Davidsonian semantic
theory will yield in every case an analytic paraphrase. Take, for example,
the sentence “Both Jack and Jill are Americans”. It’s reasonable to expect
our semantic theory to unpack the meaning of this sentence in terms of
the conditions that Jack is American and Jill is American—in contrast
with the different set of conditions for “Jack and Jill lifted the piano”. But
there is no presumption that the theory could do any better for “Jack
went up the hill” than telling us that it is true iff Jack went up the hill.
Similarly, for a sentence such as “There are ten mountains in the Taconic
range”, the theory may do no better than offer a disquotational truth-
condition. In general, there is no presumption that a truth-conditional
analysis will reveal anything interesting about the meaning of semantic
‘atoms’ such as the English terms ‘dog’, ‘mountain’, ‘water’, ‘walk’, ‘love’,
‘blue’, ‘tall’; let alone ‘happy’, ‘beautiful’, ‘funny’, ‘good’, ‘wrong’, and so
on. So, there is a sense in which the meaning-giving biconditionals are
somewhat modest, semantically speaking.
(5s) More importantly, as Davidson himself remarks, meaning-giving
biconditionals are also ontologically or metaphysically modest. For they
may not in general reveal facts about the existence, composition, or
underlying nature of the worldly conditions that the semantic theorist
invokes in her specifications of meaning. In propounding, for example,
Harrison Ford was good in Blade Runner as the truth-condition of the
English sentence “Harrison Ford was good in Blade Runner” (as well as
its translations into other languages), we, as theorists of the object lan-
guage, can remain relatively neutral on what makes for good acting. This
means that settling on the biconditionals that specify the meanings of
sentences in a given domain can leave room for substantive metaphysical
questions and debates. Having settled on the meanings of mathematical
sentences, for example, it is open to the semantic theorist (who may or
may not herself be a metaphysician) to ponder the nature of mathemati-

cal facts—whether there are numbers, what kinds of things they are, and
so on. Using the vocabulary of an area of discourse, and putting on a
metaphysician’s hat, so to speak, one can ask questions such as ‘What in
the world (if anything) makes something beautiful, morally right, funny?’.
These questions are not in general questions about language, but are
instead questions raised in the ‘material’ mode, using language.
(6s) In the case of areas of discourse thought to involve commitment
to ontologically problematic facts, there seems to be a great advantage to
recognizing the availability of the semantic notion of truth-conditions.
Acknowledging that the meanings of sentences across a wide array of
discourses can be specified by giving their truth-conditions (using the
appropriate biconditionals) allows us to accommodate undeniable logico-
semantic continuities between the allegedly problematic areas and more
straightforwardly ‘descriptive’ ones. (Unless a sentence such as “Hunting
for fun is morally wrong” can be assigned truth-conditional meaning, it
is entirely unclear how it can embed in conditionals or participate in logi-
cal inferences involving purely descriptive elements.)
(7s) At the same time, the association of truth-conditional meanings
with sentences of, for example, ethics does not automatically remove all
worries about the problematic character of putative facts in the relevant
domain. One can still be an anti-realist about ethics, even if ethical sen-
tences can be assigned truth-conditional meanings. We would argue that
securing truth-conditional meanings for sentences in an area of discourse
does not mute all significant disputes between realists and anti-realists
regarding its status.
This brings us to the second notion of ‘truth-conditions’—the one that
seems to be at work in the ‘scenario’ Pettit puts forward. On our way of
carving things up, these are worldly conditions invoked when attempting
to answer metaphysical questions about ontology, nature, constitution,
and so on—questions such as, for example, What is pain? What is color?
What makes a person happy? How are mountains to be individuated?
When does S know that p? These conditions have the following impor-
tant features:
(1m25) They are semantically innocent. They are provided—and are
offered in response to questions that are posed—in the ‘material mode’,
as opposed to responding to questions about the meanings of sentences.

Metaphysical questions can follow on the heels of assigning semantic
truth-conditions to sentences, as when we learn from the semanticist—
perhaps disappointingly—that “Torturing animals is morally wrong” is
true iff, well, torturing animals is morally wrong, and we press: but what
makes a practice morally wrong? However, this question (about the nature
of moral wrongness) can arise prior to, and independently of, recovering
the truth-conditional meaning of any sentence involving the phrase ‘mor-
ally wrong’.
(2m) Except when one’s metaphysical inquiry concerns language, the
metaphysical search for worldly conditions has nothing to do with the
analysis of meaning—though, of course, one will typically have to make
competent use of language to raise the relevant questions. Familiarly,
when the metaphysician of mind tells us that pain is a certain configura-
tion of brain states, or essentially a functional state, this is not offered as
a meaning analysis. Similarly for the utilitarian reduction of the goodness
of actions to their maximization of utility, and various other reductive
accounts.
(It may be thought that semantic externalism—a view that connects
meaning with conditions in speakers’ external environment—gives the
lie to the metaphysical neutrality of semantic analysis just suggested. But
this is a misunderstanding. Semantic externalism maintains that, in the
case of at least some terms, notably natural kind terms, their meaning is
(partly) individuated in terms of the worldly substances to which users of
the terms are causally related. So, crudely, a speaker could not mean water
by their word ‘water’, unless they were causally related to (the substance)
water. First, notice that this only provides a necessary, but not a sufficient
condition on meaning water by ‘water’. It thus falls short of giving a
semantic analysis of ‘water’ (or the conditions on competent use of the
term). But, second, as far as semantic theorizing is concerned, the
necessary condition is not to be specified metaphysically. All the external-
ist semantic theory is in a position to claim is that the meaning of ‘water’
is dependent on the nature of water, whatever that is. If water is in fact
identical to the chemical substance H2O, then being H2O is constitutive
of its metaphysical nature. But there is no expectation that ‘H2O’ should
figure in an externalist semantic account of the term ‘water’.)
(3m) The worldly conditions that figure in the truths of a semantic

theory can be seen as truth-conditions only in the sense that we can think
of them as making true some sentences/propositions/beliefs/and so on,
and not others. But, so understood, they are conditions that are individu-
ated metaphysically, not semantically. Consider: the worldly condition of
H2O being wet is, metaphysically speaking, one and the same condition
as that of water’s being wet. So this worldly condition—described either
way—makes true the sentence “Water is wet”. But, for all that, “‘Water is
wet’ is true iff H2O is wet” is not a meaning-giving biconditional for
“Water is wet”. It will not be (or at any rate should not be) a theorem
derivable from a truth theory for English.26
Where does all this leave us with respect to truth pluralism? Consider
a metaphysical inquiry into what makes something illegal. Such an
inquiry may conclude that the legality of this or that act depends in some
systematic way on our legal practices, on certain aspects of history, and so
on. Perhaps it will conclude that nothing is legal that would not be judged
legal by an ideally placed judge, so that it makes no sense to suppose that
the legality of an act could forever elude human judgment. We can sum-
marize the results of this inquiry by saying: In legal matters (in contrast
with other sorts of matters), truth is judgment-dependent. But the ques-
tion is whether putting things this way really commits us to a distinct
substantive truth property possessed by all true legal sentences (as con-
trasted with, say, sentences of everyday discourse about mid-sized dry
goods). The explanatory gain in invoking a distinct kind of truth seems to
us illusory; better, we think, to appeal to differences in what makes true
statements in the legal realm true—what objects, properties, or states of
affairs (including practices, history, and so on) make for the legality or
otherwise of this or that act or practice. Our complaint, in short, is that
nothing is added by invoking a metaphysical plurality of truth properties
over and above whatever plurality is recognized in the worldly conditions
that our metaphysicians have identified or proposed, as they investigate
different domains of discourse.
Pedersen and Lynch address what may seem like this complaint under
the heading “the double-counting objection”. Distinguishing differences
“at the level of subject matter” (which they understand as metaphysical
differences) from differences “at the level of truth” (which they consider
to be semantic differences), they address the objection that “to accommo-

date …the appeal of realism and antirealism with respect to different
domains one only needs to buy into differences in subject matter”.27
Following Wright, they respond that the metaphysical difference will
inevitably bring a semantic difference in its train.
In accounting for the circumstance that <There are mountains> is true it

seems right to say that there is a fit between the proposition and reality, and
that this fit is in no way due to us shaping, or somehow contributing to,
what the relevant tract of reality is like. Matters change when we turn to
<Speeding is illegal>.28
Our question is: What are we here adding to the claim that we have
something to do with things being illegal, but nothing to contribute with
respect to whether something is a mountain? What difference is there in
what ‘accounts for the circumstance’ that the two different propositions
are true that is not simply a matter of the difference in the constitution of
the relevant facts? Of course, we can ascend to the ‘formal mode’, and
instead of talking about what makes for the existence of mountains ask
what accounts for the truth of <There are mountains>. If we want to gen-
eralize over the whole domain, we may need to use the truth predicate,
viz. “For all p, if p is a mountain-statement, then p is true iff …”. And,
depending on how the condition is filled in, we may be able to say, for
example, that mountain statements are true in a mind-independent way.
But our point is that the possibility of characterizing the differences in
the formal mode in no way betrays commitment to a new, additional dif-
ference—one that requires postulating differences in ways of being true,
or the possession of divergent truth properties by statements in different
areas of discourse. Indeed, when Pedersen and Lynch expound the
‘semantic difference’ between the truth of <Speeding is illegal> and
<There are mountains>, they themselves immediately resort to talk in the
material mode: “Mountains are mind-independent entities while laws are
social – and so, mind-dependent – constructs”.29
Unlike others who have worried about double counting, our objection
is not motivated by a pluralist view of propositional content or a defla-
tionist view of truth.30 If we are right, there is a way of making sense of
disputes between realists and anti-realists that neither goes via a dis-
tinction at the level of propositions nor depends on deflating all truth.31
But it does not depend on invoking different kinds of truth, either. We
thus endorse alethic monism: there is only one way for true sentences,
propositions, beliefs, and so on, to be true. However, there may be
multiple kinds of worldly conditions that make them true. The rele-
vant plurality can be captured in the material mode; it doesn’t require
any semantic or alethic ascent. Of course, given the equivalence of 
and < is true>, one can advert to a formal mode and speak of the
truth of ‘x is red’ being a different sort of thing from the truth of ‘x is
divisible by 2’—indeed, sometimes putting things in terms of truth
may be unavoidable. It is the additional move, to a plurality of truth
properties, each appropriate to a different domain of discourse, that we
here oppose. This move, we maintain, is not forced on us by taking
seriously debates between realists and anti-realists. Alethic plurality
contributes no explanatory power; all we need is a plurality of kinds of
worldly conditions.32
* * *
A final remark. Wright has proposed the following analogy by way of

making pluralism about truth plausible.33 Winning a game is a unitary
concept. But in different games different things count as winning. So,
there is a plurality of ways of winning. Similarly, Wright has suggested,
there is a plurality of ways of being true, and thus a plurality of truth prop-
erties that are ‘satisfiers’ of a uniform notion of truth. A difficulty Wright
considers for the analogy is that, given a game, it is typically obvious what
constitutes winning in the game. By contrast, knowing what constitutes
being true in any given domain typically requires extensive philosophical
investigation and can be highly contentious. We suggest that our way of
portraying things can readily absorb the disanalogy. As regards games, the
key idea is what constitutes winning will vary from game to game.
Regarding different domains of discourse, the idea should not be that
what constitutes being true varies from area to area. Instead, recast in
metaphysical terms—‘at the levels of facts’ rather than at the level of prop-
ositions or truth—the idea should be that the kinds of worldly conditions
that make a statement true will vary. But then it stands to reason that it
will not in general be obvious what the ‘realizers’ of truth in various areas
are, in contrast with what constitutes winning a game. Access to what in
the world makes true statements true requires a metaphysical investiga-
tion. When it comes to games, since they are invented, and in that sense
‘of our own making’, knowing what game is being played guarantees
knowing what constitutes winning it. (That is part of what is instituted
when the game is designed.) Not so for what renders statements true in a
given domain of discourse. In general, engaging in an area of discourse
does not bring into existence the relevant worldly conditions. And, what-
ever disagreements we may have cannot be settled—as they can in the
winning case—simply through reflection on the rules of the discourse.
Where disputes arise, they are metaphysical ones.34
Notes
1. Wright (2013: 124)
2. Here and throughout, ‘’ is a name of the proposition that p.
3. See, for example, Lynch (2013).
4. See, for example, op. cit.
5. Lynch 2013.
6. The equivalence here is stronger than necessary equivalence. As
Jeremy Wyatt has pointed out to us, it’s arguable that A and B can be
necessarily equivalent without both being superwarranted for a subject
S. Consider the necessarily equivalent propositions <There’s water in the
glass> and <There’s H2O in the glass>. Suppose S grasps all the relevant
concepts. It is possible that <There’s water in the glass> is superwarranted
for S, but <There’s H2O in the glass> is not, since S may have warrant to
believe that there’s water in the glass, but not that there’s H2O in the glass.
But if S grasps the concept of truth, and < is true> will be
conceptually equivalent for S, and if one is superwarranted for S, so is
the other.
7. Lynch offers this definition of propositional coherence (with respect to
moral propositions) in Lynch 2009, p.171.
8. Lynch 2013.
9. Bar-On and Simmons (2007), Bar-On et al. (2004).
10. Bar-On and Simmons (2007).
11. Strawson (1992).

12. Horwich (2006: 195).
13. Ibid.
14. See Wright (2003: 136). Wright registers disagreement with ‘meaning
pluralists’ by saying: “the realist/anti-realist debate is not a semantic
debate in the end” (2013: 126). This can be confusing, since meaning
pluralists like Blackburn sometimes describe their disagreement with
truth pluralism by complaining that it introduces an unnecessary detour
via the semantic property of truth.
15. Pettit (1996: 886) (our emphasis), cited with approval in Wright (1996:
101f.).
16. Davidson (1984: 171–180).
17. On some views (though not Davidson’s), truth-conditions so under-
stood are what competent speakers have mastered (or internalized) and
know, at least implicitly. For relevant discussion and references, see
Bar-On (1996).
18. And witness his subsequent reference to “what it is for something to hold
in physics – what the truth-condition is …” (ibid., our emphases).
19. It will be important to bear in mind that we are not suggesting, along the
lines of e.g. two-dimensional semantics, that sentences have associated
with them two kinds of meaning, or two sets of truth-conditions.
20. Davidson (1984: 17–36)
21. Davidson (1984: 31).
22. ‘s’ for semantic.
23. Davidson (1990, 1996).
24. Davidson (1980: 105–122).
25. ‘m’ for metaphysical.
26. This is perhaps why deflationists about truth are perfectly happy to allow
that we do – and can, consistently with deflationism – speak of worldly
conditions that we loosely refer to as truth-conditions. For discussion
and references, see Bar-On et al. (2004).
27. See Pedersen and Lynch (2018).
28. op.cit.
29. op.cit.
30. Dodd (2013) advances a version of the double counting objection that
draws on deflationism.
31. For we have here only sought to question the utility of invoking a plural-
ity of truth properties over and above the property of truth that, by the
moderate pluralist’s lights, all true items possess (regardless of domain).
32. Asay (2016) also argues that all the plurality we need is to be found in
the world, not in a plurality of truth properties. But Asay’s plurality is a
plurality of truthmakers rather than truth conditions. And Asay is a
primitivist about the concept of truth and a deflationist about the prop-
erty; we don’t make these commitments.
33. See Wright (2013: VII), who follows Edwards (2011, 2013).
34. Our thanks to Jeremy Wyatt for many helpful comments, and to the
participants in the Conference on Pluralism about Logic and Truth,
University of Connecticut at Storrs, April 2015.
References
Research 96 (1): 175–191.
Bar-On, D. 1996. Anti-Realism and Speaker Knowledge. Synthese 106:
139–166.
Bar-On, D., and K. Simmons. 2007. The Use of Force Against Deflationism:
Assertion and Truth. In Truth and Speech Acts, ed. D. Greimann and
G. Siegwart, 61–90. New York/London: Routledge.
Bar-On, D., W. Lycan, and C. Horisk 2004. Deflationism, Meaning and Truth-
Conditions, reprinted with Postscript in Deflationary Truth, ed.
B.P. Armour-Garb and J.C Beall, 321–352. Chicago: Open Court Readings
in Philosophy.
Davidson, D. 1980. Essays on Actions and Events. Oxford: Oxford University
Press.
———. 1984. Inquiries into Truth & Interpretation. Oxford: Oxford University
Press.
———. 1990. The Structure and Content of Truth. Journal of Philosophy 87:
279–328.
———. 1996. The Folly of Trying to Define Truth. Journal of Philosophy 93:
263–278.
Dodd, J. 2013. Deflationism Trumps Pluralism. In Truth Pluralism: Current
Debates, ed. Nikolaj J.L.L. Pedersen and C. Wright, 298–322. Oxford:
Oxford University Press.
Edwards, D. 2011. Simplifying Alethic Pluralism. Southern Journal of Philosophy
49: 28–48.
———. 2013. Truth, Winning, and Simple Determination Pluralism. In Truth
Pluralism: Current Debates, ed. Nikolaj J.L.L. Pedersen and C. Wright,
113–122. Oxford: Oxford University Press.
Horwich, P. 2006. A World Without Isms. In Truth and Realism, ed. P. Greenough
and M.P. Lynch, 188–202. Oxford: Oxford University Press.
Lynch, M. 2013. Three Questions for Truth Pluralism. In Truth Pluralism:
Current Debates, ed. Nikolaj J.L.L. Pedersen and C. Wright, 21–41. Oxford:
Pedersen, Nikolaj J.L.L., and M. Lynch. 2018. Truth Pluralism. In The Oxford
Handbook of Truth, ed. M. Glanzberg. Oxford: Oxford University Press.
Strawson, P. 1992. Analysis and Metaphysics. Oxford: Oxford University Press.
Wright, C. 1996. Précis to Truth and Objectivity and Response to Commentators.
Philosophy and Phenomenological Research 56: 863–868, 911–941.
———. 2003. Saving the Differences. Cambridge, MA: Harvard University
Press.
———. 2013. A Plurality of Pluralisms? In Truth Pluralism: Current Debates,
ed. Nikolaj J.L.L. Pedersen and C. Wright, 123–153. Oxford: Oxford
University Press.
Truth Pluralism, Quasi-Realism,
and the Problem of Double-Counting
Michael P. Lynch
1 Unity and Diversity
Philosophical tradition has long decreed that truth comes in only one
kind. To talk about different kinds of truth is like saying that the Sphinx
has a different kind of eye—it is just another way of saying that it has
none at all.1 But philosophical tradition also decrees that traditions are
there to be opposed. Truth pluralism—or the idea, roughly, that there are
different kinds of truth—stands with the opposition in this case.
Since its introduction onto the contemporary scene by Crispin Wright
over two decades ago, truth pluralism has been connected to, and partly
motivated by, two major explanatory projects. One project involves
accommodating the intuitions that drive both realism and anti-realism
(Wright 1992). By adopting truth pluralism, the suggestion goes, we may
say that in some domains, statements or beliefs are true in a realist way,
while in other domains, truth is understood in ways traditionally cham-
pioned by “anti-realists.” Thus, for example, some statements are true
M. P. Lynch (*)
Department of Philosophy, University of Connecticut, Storrs, CT, USA
e-mail: mplynch@uconn.edu

https://doi.org/10.1007/978-3-319-98346-2_3
64 M. P. Lynch
because they represent—in some suitably thick sense of “represent”—

mind-independent natural objects (Lynch 2009), while others—such as
those concerning normative matters—are true in virtue of being superas-
sertible, or supercoherent in some fashion or other (Wright 1992, 1998b).
A second major explanatory project connected to truth pluralism
involves accommodating two distinct intuitions concerning the relation-
ship between our thought, talk, and the world. On the one hand is the
idea that our thought and talk itself come in radically diverse kinds.
Thought is not a “seamless web” (Blackburn 1998a, p. 157). Rather, dif-
ferent kinds of attitudes and speech-acts have different functions in our
cognitive lives, and that fact offers a crucial constraint on our account of
how we understand, or explain their content. Call this:
semantic diversity: there are real differences in kind between the con-
tents of our beliefs and indicative statements.
Privileging semantic diversity is an obvious starting point for certain

forms of classical expressivism, according to which the moral judgments
differ from non-moral ones in part because the latter have representa-
tional content while the former do not. However, a typical complaint
with such views is that, in their haste to privilege diversity, they tend to
deny:
cognitive unity: all such statements and beliefs are subject to a single
type of cognitive normative assessment or correctness.
Roughly put, a typical complaint against classical expressivism is that state-

ments about values are subject not only to practical reason, but also theo-
retical reason. They figure in the premises of (valid) arguments and as the
antecedents of conditionals; they are truth-functional; when we believe the
contents of these statements, we can provide epistemic justifications for our
holding of them; and so on. As we might put it, value judgments can be
correct or incorrect just as judgments about the physical world can be.
Expressivists, of course, have offered a range of responses to this chal-
lenge—a challenge that is sometimes bundled under the heading of the
“Frege/Geach” problem.2 The truth-pluralist, can, in effect, be seen as
doing the same. That is, while Crispin Wright (1992) initially introduced
Truth Pluralism, Quasi-Realism, and the Problem… 65
pluralism as an alternative to expressivism, the position can be under-

stood as offering a response to the complaint that the root insight of
expressivism is inconsistent with cognitive unity.3 Thought and talk are
unified (one) because truth is correctness, but thought and talk are
diverse (many) because correctness comes in different kinds.
There are, naturally, different ways to pursue these points. Early state-
ments of truth pluralism were typically understood as implying that there
was no single concept of truth, and that the word “true” is ambiguous
across domains (Pettit 1996; Sainsbury 1996; Tappolet 1997).4 Most cur-
rent pluralists reject this view, however, emphasizing that there is a single
concept of truth. Said concept is a functional concept, one which names
a property which can itself be realized by very different underlying prop-
erties—a correspondence property for some kinds of beliefs and state-
ments, and a coherence property for others.5 Suitably understood, this is
still consistent with the idea that truth comes in different kinds. Just as
the functionalist in the philosophy of mind can say that there are differ-
ent kinds of pain—pain as realized by neural property Φ and pain as
realized by neural property Ψ—the truth pluralist can say there are differ-
ent kinds of truth: truth as realized by property α, and truth as realized
by property β. The differences enter in at the level of the properties that
determine or realize truth. As noted, these differences are often cashed
out by appealing to metaphysically heavyweight properties of representa-
tion or idealized coherence. Strictly speaking, that is not necessary. Some
pluralists (e.g., Beall 2013) are attracted to a view according to which
there are different kinds of truth, but every kind of truth is “deflated” or
minimal.
Even so broadly sketched, pluralism has at least two prima facie bene-
fits to anyone attracted to the expressivist insight of semantic diversity.
First, there is the significant theoretical benefit of allowing us to say that
some beliefs and statements have representational content and some do
not—without complicating the semantics (see Lynch 2009). That is, all
indicative statements can be seen as alike in having their content deter-
mined (at least in part) by their truth conditions. It is just that those con-
ditions themselves come in different kinds. Second, the differences in the
metaphysics—the differences in the realizing properties—can be appealed
to in order to explain the differences in content between statements and
66 M. P. Lynch
beliefs of different kinds. Thus, rather than trying to distinguish, for

example, all normative from non-normative beliefs in advance, so to
speak—and thus possibly running into the problem cases of legal dis-
course, or discourse involving so-called thick concepts—the pluralist can
sort beliefs by appealing to the properties that realize truth for the par-
ticular content believed.6
There is a tremendous amount of work—much of it technical—to be
done in order to fully understand the details of such a view. Among other
things, pluralists must say something about so-called mixed inferences
and compounds, as well as the relationship between truth and logical
consequence. They must also offer substantive theories about the proper-
ties that determine or realize truth. Such work is still ongoing.7 In this
essay, however, I’m concerned less with the details of truth pluralism than
I am with an objection to the view that comes up before we even get to
those details. It is a criticism most recently and forcefully articulated by
Simon Blackburn, whose own quasi-realist view shares the truth plural-
ist’s desire to reconcile semantic diversity with cognitive unity.
It goes like this: There aren’t different kinds of truth; there are simply
different kinds of truths. As Blackburn recently put it, while the differ-
ences between, say, ethical judgments and judgments about the physical
world are significant and important,
…they strike at the level of the proposition: they mark distinctions of sub-
ject matter, and perhaps eventually distinctions of objectivity or the possi-
bility of cognitively fault-free disagreement. But why add to a distinction
of content, another, mirroring distinction, one only applying to kinds of
truth or conceptions of truth?” (Blackburn 2013, p. 265)
In Blackburn’s words, the truth pluralist is engaged in a kind of

double-counting.
The objection is intuitive, and a similar point has been pressed by a
number of authors (Dodd 2013; Sainsbury 1996).8 Hence it seems
incumbent on the truth pluralist to have something to say about it. On
closer inspection, however, there appears to be at least two subtly different
claims that can be made under its auspices. In what follows, I will attempt
to disentangle them. As we’ll see, only one version of the point is compat-
ible with Blackburn’s own quasi-realism; but it faces its own problems—
problems that encourage the reintroduction of the very distinctions it
was aimed against. As such, I conclude that truth pluralism need not fear
being audited for double-counting—or at least no more than the quasi-
realist. Indeed, the quasi-realist, I will argue, has excellent motivations for
adopting—or readopting, as we shall see—truth pluralism.
2 Subject-Matter First
The first way of construing the double-counting objection can be found
in the Blackburn quote given just above. It is the idea that semantic diver-
sity can be reduced, or solely explained as, diversity in subject-matter.
Quine first made this point:
There are philosophers who stoutly maintain that “true” said of logical or
mathematical laws and “true” said of weather predictions or suspect confes-
sions are two usages of an ambiguous term “true”…Why not [instead] view
“true” as unambiguous but very general, and recognize the difference
between true logical laws and true confessions as a difference merely
between logical laws and confessions? (Quine 1960, p. 131)9
In Quine’s formulation, the truth pluralist’s error lies in part in positing

“true” as ambiguous: it needlessly gives different meanings to the predi-
cate. As we’ve already noted, the contemporary pluralist can agree—“true”
should be understood as expressing a single concept. Thus, the second
aspect of Quine’s point is more salient for our purposes. Namely, we don’t
need to appeal to truth at all in order to explain the differences between
confessions and logical laws (or morals and mathematics). In Quine’s
hands, this point is clearly meant to fall out of his disquotationalist view
of truth. As Quine would put it, to say that the statement “snow is white”
is true is just another way to talk about snow. Truth is a device for disquo-
tation. If so, then why semantically ascend to the lofty heights of truth-
talk to explain the differences when we can look to the world itself?
68 M. P. Lynch
Why indeed. The answer is that sometimes there is no world to look

to. That is, in some cases, we want to explain the differences between dif-
ferent kinds of thoughts, even though—and perhaps precisely because—
we think that some of them fail to represent any independent reality. But
how do we capture this possibility if we are not allowed to appeal to the
semantic differences between the statements or beliefs in question?
Consider, by way of illustration, a different sort of philosopher from
Quine: someone who thought that truth is always and everywhere cor-
responding with propertied objects. This philosopher might say that
what marks the differences between our language games is that in physics
we have one type of object—electrons, protons, and the rest—and in
morality we have another—hedons, or, as Dworkin (1996) once memo-
rably put it, morons. Or, perhaps they might say: the difference consists
in the fact that moral discourse is just never true; there are no objects to
which it refers, poignant as it may be.10
Absent further garnishing, these are just the options the Quinean ver-
sion of the objection leaves us as well. If all we have to say about the dif-
ferences in our thought and talk is “look to the objects,” then we either
end up with blanket realism or an error theory. That may be perfectly
correct—I have nothing to say against such options here. The point,
rather, is that such a strategy leaves some options out, options that
Blackburn for one would be keen to insist upon, according to which the
semantic diversity of our thought is greater than we are so far
acknowledging.11
Of course, both Quine and our imagined correspondence theorist
might try to recapture those options by saying this: some sentences are in
the game of truth and others are not. The former has truth conditions, or
express propositions. The standard usage of the latter amounts to, for
example, planning to do something, approving of so and so, or com-
manding someone to do such and such. This, of course, is just classical
expressivism. But to endorse that view is to explain semantic diversity,
not in terms of different kinds of propositions, but by appealing to a dif-
ference between discourses that do express propositions and those that
don’t—between those that are truth-conditional and those that aren’t. It
just gives up on cognitive unity.
As a result, Quine’s “subject-matter first” point cannot be the double-

counting objection that Blackburn initially had in mind. For such a view,
while certainly coherent and no doubt attractive to many, is not aiming
to at the same explanatory goal of explaining both diversity and unity. Let
us therefore look elsewhere for the real point behind Blackburn’s
objection.
3 Content, Not Truth

We needn’t, I think, look far. Here is Blackburn invoking some heroes
from Philosophical Valhalla:
The point is that Ramsey and Wittgenstein do not need to work with a
sorted notion of truth—robust, upright, hard truth versus some soft and
effeminate imitation. They need to work with a sorted notion of a proposi-
tion, or if we prefer it a sorted notion of truth-aptitude. (Blackburn 1998b,
pp. 166–167)
Like Quine before him, Blackburn takes issue with the truth pluralist’s
view that appealing to different kinds of truth will help explain the differ-
ences. And, like Quine, he is motivated to take this view at least in part
because of his allegiance to a version of deflationism about truth. In par-
ticular, Blackburn is attracted to Horwich’s minimalism.
According to Horwich (1998), the concept of truth functions as a logi-
cal device for generalization: it allows us to overcome our merely medical
limitations and make certain generalizations, such as “every proposition
is either true or false.” We grasp that concept by being a priori disposed
to grasp instances of the equivalence schema:
ES: The proposition that p is true if and only if p.
When we understand these facts—how we grasp the concept, and its

function—we know all the essential facts about truth. Any other fact
about the truth-role can be deduced from (instances of ) ES together with
some non-truth-theoretic fact. Consequently, even if we grant that truth
70 M. P. Lynch
is a property—if a metaphysically transparent one, like the concept of

being a logical conjunction—it is a property that does no significant
explanatory work.
Blackburn, as is well known, enlists minimalism about truth in the
service of quasi-realism. He takes it that ethical, aesthetic judgments,
statements, and the like do, in fact, express propositions. They just express
unique kinds of propositions, each kind of which is still apt for the same
kind of truth. The question remains how to explain that. We’ve just seen
one way: explain differences in propositional content by appealing to dif-
ferences in the objects and properties to which those propositions (or
their structural elements) refer. But Blackburn wants a view that allows
that moral judgments, beliefs, and so on might be true without it being
the case that there are, in any serious sense, moral properties. He wants to
say that some kinds of propositions are true in very different circum-
stances, as it were, than others. Which is just to say—isn’t it?—that they
have different kinds of truth conditions. But that seems precisely to be the
view that the truth pluralist is urging, minus the simple explanation for
its obtaining that the pluralist can supply—namely, there are different
kinds of truth conditions because there are different kinds of truth.
Although it may be obvious, it is worth emphasizing that the quasi-
realist is not simply repeating her minimalism in taking on this view. She
is not simply saying: each proposition specifies its own truth condition.
One can endorse that point without being a quasi-realist. Rather, she is
insisting, in addition, that such truth conditions are themselves sorted
into discourses, domains, and subjects—in short, kinds. That is the point
of saying that moral judgments, beliefs, and statements are semantically
distinct from the non-moral versions of such things. But what would
explain this fact? Why should the conditions under which some contents
are true differ in kind from others? At this point, given that the quasi-
realist has excluded the answers above, it again seems hard to avoid the
simplest answer: because they are subject to different kinds of truth.12
Indeed, this is precisely the view that Blackburn had himself defended
earlier in his career:
The problem is not with a subjective source of value in itself, but with peo-
ple’s inability to come to terms with it, and their consequent need for a pic-
ture in which values imprint themselves on a pure passive, receptive witness…
To show that these fears have no intellectual justification means developing

a concept of moral truth out of the materials to hand: seeing how, given
attitudes, given constraints upon them, given a notion of improvement…we
can construct a notion of truth. (Blackburn 1984)
This passage (and the detailed discussion of the idea by Blackburn that
follows) is striking. The idea is precisely that of the truth pluralist and
motivated in a similar way: namely, by pointing out that those who
embrace a “subjective source” for value can respond to the obvious objec-
tions by demonstrating that moral judgments are apt for a distinctive,
but nonetheless real, kind of truth.
To sum up so far: The double-counting objection was initially framed
as the simpler alternative to truth pluralism. The thought seemed to be:
well, we can accomplish all of our theoretical goals—explaining both
semantic diversity and cognitive unity—just by appealing to something
we already have for “free”: different kinds of propositions. We don’t need
(“in addition”) kinds of truth. And that all goes swimmingly, save for the
fact that the most obvious way to press the objection—Quine’s way—
self-consciously fails (or ignores) that very explanatory task. The Quinean
route ends up either getting unity at the price of diversity (by enforcing a
blanket realism), or diversity at the price of unity (by enforcing a distinc-
tion between the truth-conditional and the non-truth-conditional).
Consequently, we arrive at the quasi-realist’s actual suggestion: truth
explains unity, kinds of propositions explain diversity. But this slogan is
already seeming less simple, and indeed, less distinct, than it did at first
blush. It is starting to sound like truth pluralism in disguise.
4 Quasi-Realism at Any Price

In part for the reasons given above, some expressivist sympathizers have
pursued other options in recent years. One option is to go hybrid (Ridge
2006, 2009) and argue that ethical commitments are comprised of a
conjunction or complex of a belief state and a desire or sentiment. Another
is to take there to be one basic, non-doxastic kind of commitment of which
ethical and non-ethical commitments are themselves species (Schroeder
72 M. P. Lynch
2008, 2010). Both of these suggestions have considerable technical merits

(and complications) all their own, and of course warrant separate discus-
sion. In any event, they are not the options taken by the quasi-realist.
Blackburn aims, he suggests, to hew closer to the expressivist’s original
core. Thus, we find him saying:
There are propositions properly theorized about in one way, and ones prop-
erly theorized about in another. The focus of theory is the nature of the
commitment voiced by one adhering to the proposition, and the different
functional roles in peoples’ lives (or forms of life, or language games) that
these different commitments occupy. (Blackburn 1998b, p. 167)
What makes A-propositions different from B-propositions is the differ-

ences in the commitments we take toward them in our lives, commitments
that have different motivational and emotional profiles, for example. Call
this the commitment-first approach. If it is to succeed, it must distinguish
kinds of commitment without relying on either a simple, unexplained
difference of kind between propositions, or on kinds of truth.
In the halcyon days of the expressivism of yore, prior to eating the fruit
of the Frege-Geach tree, we would just have said: well, some attitudes or
commitments—like belief—just are the kinds of commitments that are
truth-apt, whereas other kinds of attitudes or commitments are not. But
again, that obviously isn’t the picture the quasi-realist is offering. That’s
because the quasi-realist, remember, takes it that all kinds of commit-
ments to propositions are truth-apt. They are, after all, commitments to
propositions. Put another way, the quasi-realist is saying that we can take
what we used to innocently call the attitudes of asserting and belief
toward every kind of proposition. We deflate belief and assertion as much
as we deflate truth, as Blackburn (1998a, b) himself would be the first to
admit.
Over the last decade, some philosophers attracted to quasi-realism,
and at points Blackburn himself, have pursued another framework, one
made famous by Brandom (Brandom 1998) and recently championed by
Price (Price 2011, 2013). According to this view, what makes one atti-
tude/speech-act/commitment different from another is its special norma-
tive and social role. The common basic idea is that if you, for example,
believe or assert that p (as opposed to merely wondering whether p), then
you become liable to certain pushbacks from your interlocutors if it turns
out that not-p. Moreover, you incur certain obligations to provide them
with evidence that p should they demand it, to stand by certain easily
recognized implications of p, and so on. In short, what distinguishes one
sort of commitment/attitude/speech-act from another is the type of sta-
tus and responsibility accorded to the person who engages in that com-
mitment/attitude/speech-act. They are not distinguished in terms of
truth, reference, or direction of fit (whether word-to-world or
world-to-word).
Price, in particular, has urged that a view of this sort is the most intui-
tive extension of the quasi-realist view (see, e.g., 2011, 2013). And he has
been explicit that it is designed to capture what I’ve called semantic diver-
sity and cognitive unity. Moreover, he has urged that the view be adopted
globally—that we become global expressivists, as he has put it. Of course,
in order to take the view to help us with our present problem of distin-
guishing kinds of propositions without appeal to different kinds of truth,
we’d need to be able to show at least two things.
First, as Brandom (1998) in particular has tried to do, we would have
to show that we could build a real theory of content out of this theory of
the types of attitudes/commitments/speech-acts. Thus, for example, such
a theory must have the resources to distinguish between contents that
concern our own informational states and attitudes, and those that are
wholly independent of them. Whether it can do so remains an open
question, so let’s put this aside.
Second, the Pricean pluralist must be able to distinguish ethical and
non-ethical propositions while still maintaining the idea that, for exam-
ple, ethical propositions can still be believed and asserted, albeit in a differ-
ent way than propositions about the physical world are believed and
asserted. It is, after all, the appeal to this “different way” of believing and
asserting that makes the commitment-first approach different from the
truth-conditional approach. The point here, to clarify, is not that the
global, Pricean quasi-realist must simply distinguish between different
attitudes (like belief and desire) that are operative in different domains of
discourse; this is something that all players can grant. Rather, they must
distinguish between different kinds of one type of attitude—for example,
belief (Dreier 2004; Ridge 2006).
74 M. P. Lynch
Let’s pause here for a moment. The double-counting objection, in the

hands of the quasi-realist, maintains that the truth pluralist unjustifiably
distinguishes between kinds of truth when a distinction between kinds of
truths can do. But now we see that the quasi-realist distinguishes between
ways of asserting and believing (different kinds of assertion and belief )
when a simple distinction between assertions and beliefs would do.
Double-counting! If, contra quasi-realism, we can appeal to truth, we can
say, for example, that the differences between moral beliefs and other
sorts of beliefs doesn’t lie in the fact that there is a special “moral way” of
believing, but in the fact that moral contents are true in different kinds of
conditions— precisely the view we saw Blackburn maintain in an earlier
version of quasi-realism.
Tit for tat, however, is a game of only limited interest. The quasi-realist
double-counts, but the commitment-first approach can, after all, appeal
to real differences in our practices. But that fact, while real, can some-
times obscure a simple point. These commitments, recall, are now agreed
to also constitute moral beliefs in a deflated sense. But with what does this
deflated sense contrast? Surely there must be a contrast if we wish to count
not only different beliefs but also different kinds of belief. And it is not
enough to say that, for example, non-moral beliefs aren’t motivational.
For that would paint mathematical beliefs, modal beliefs, aesthetic beliefs,
and beliefs about the physical world with the same wide brush. Thus, in
the physical case in particular, it is tempting to add that such beliefs repre-
sent the world while moral beliefs don’t. But, as a number of authors
(Chrisman 2008; Dreier 2004) have asked: how do we say that without
giving up on our global deflationary stance toward truth, belief, assertion,
proposition, and the like?
Huw Price has replied, on behalf of the quasi-realist, that they can
make truth-free distinctions between different kinds of discourses by
appealing to a difference in kinds of representation: what he calls
i-representation and e-representation. According to Price, i-representations
represent by having “a position in an inferential or functional network”
(Price 2013) while e-representations represent by covarying with, or
tracking, the environment (Price 2013, p. 43). Price maintains that both
notions can have their uses. But, most importantly for our purposes,
appealing to the distinction allows the quasi-realist to say that “scientific
claims nevertheless have a world-tracking, e-representational character

that moral claims lack” (2013, p. 39), and that “while all assertoric vocab-
ularies are i-representational, some may be much more e-representational
than others” (2013, p. 152).
Price puts this distinction to use in various subtle and interesting ways.
But it is unclear whether it can help the quasi-realist press her contrast
with the truth pluralist. One reason for thinking this is that it seems to be
a truism about truth and representation alike that:
Represent: A belief with the content that x is F is true if and only if the
object represented by the concept x has the property repre-
sented by the concept F.
Thus, one might think, if there are different types of representation, then
we get not only different kinds of belief but also different kinds of truth,
namely:
Represent-I: An i-belief with the content that x is F is true if and only
if the object i-represented by the concept x has the prop-
erty i-represented by the concept F.
Represent-E: An e-belief with the content that x is F is true if and only
if the object e-represented by the concept x has the prop-
erty e-represented by the concept F.
How might the Pricean quasi-realist resist this charge? One way would be
to just accept that Represent-I and Represent-E result in two distinct
concepts of truth—concepts which “true” is simply ambiguous between.
Alternatively, they can insist on their minimalism about truth. Truth is
one thing, they can say, and representation is another. And all there is to
say about truth as such is still simple, and requires no appeal to the dif-
ferent kinds of representation.
But this point also won’t distinguish the Pricean quasi-realist in any sig-
nificant way from the truth pluralist. As noted earlier, truth pluralism, in its
principal formulations, isn’t an ambiguity view. It can indeed be described
as the view that there are different kinds of truth—in the same sense in
which the functionalist in the philosophy of mind can say that there are
76 M. P. Lynch
different kinds of pain. In Wright’s view and in others’, there is more than
one property that can determine, realize, or manifest the property of truth
as such. And it is difficult to see how Represent-I and Represent-E aren’t
just saying exactly that—namely, what determines the truth of i-beliefs is
i-representation and what determines the truth of e-beliefs is e-representa-
tion. But even if this is somehow resisted, the fact remains: while we started
with the thought that there was a strong contrast between truth pluralists
and quasi-realists, we have ended with the thought we might as well call
representational pluralism. The bump in the carpet has been moved.
Unsurprisingly, perhaps, Price himself suggests that the real difference
between truth pluralism and representation pluralism lies in how propo-
nents of each view tend to conceive of their project. Truth pluralists like
Wright typically see themselves as revealing the relational properties that
determine whether a belief or statement’s content is true. They are doing
a bit of metaphysics in service of their semantics. It is this that Price
wishes most to resist. For Price, representation-talk of either kind is just
that: talk. It serves a purpose, but it drops out at what he calls the “upper
level” (2013, 155ff). In saying that, for example, some assertions or atti-
tudes are more e-representational than others, we are ultimately not mak-
ing a distinction between different kinds of relational properties. Instead,
we are making claims about the “talk not the ontology” (2013, 158); we
are merely describing the roles such assertions play in our greater theoriz-
ing. This is the point of his insisting that his view represents a global
expressivism or quasi-realism.
Ultimately, then, the Pricean quasi-realist explains the differences
between kinds of thought and talk only internally to our thought and talk
themselves. Interestingly, Blackburn remains skeptical; and his skepti-
cism is worth quoting in full:
Locating thinkers or speakers in a landscape in which we only answer to

statuses accorded to us or denied to us by fellow thinkers and speakers risks
distorting our positions. For we do not just answer to each other. We
answer to each other because of what we get right or wrong about the
things we are involved with—the things we are talking about…If I am
deciding whether an object is red or square or weighs five pounds, I am not
primarily concerned with what other people will say is the case nor predict-
ing the penalties if I am out of step with them…It is not like tuning up an
orchestra, where my prime concern is to listen to whether my note is the

same as the notes of other players. We are not after democratic harmony
but getting the judgment right. (2013, p. 268)
Indeed. There are two key points here, both directed—rightly in my

view—at ways in which the Pricean view misrepresents our actual prac-
tices. The first is the worry that in globalizing expressivism we run afoul
of the claim that pragmatic views like this get things backward. In
“answering to each other,” we take ourselves to be correct in our judg-
ments because they are true—not vice versa. And second, when in par-
ticular we are concerned with matters involving the middle-sized dry
goods of everyday life, or science, we take our judgments’ correctness,
and—given the first point—their truth, to be a matter of whether we are
responding, in making those judgments, to those dry goods themselves.
Blackburn’s point here could be usefully put in terms of Price’s two
kinds of representation. When I describe an object as red or square, I
ordinarily take my description to be correct precisely because it is a response
to how things are independently of me. And “responsiveness” here surely
means, at least in part, that I am tracking the environment—I am
e-representing the middle-sized dry goods around me in just the way one
suspects that evolution has programmed me to do. And that seems, more-
over, to capture part of the difference between talk of square things and
talk of what is morally right and wrong. The latter is not plausibly gov-
erned by a norm that is cashed out in terms of responsiveness to a world
independent of us, and the former is.
5 Quasi-Realism and Truth Pluralism

As we saw above, the quasi-realist wasn’t always opposed to truth plural-
ism. In fact, the earlier Blackburn embraced it. What I’ve been suggest-
ing, in effect, is that he was right to do so. For truth pluralism gives you
both what we called earlier Cognitive Unity (and the cognitive surface
that goes with it) together with Semantic Diversity—the view that the
quasi-realist is so keen to insist on. And we don’t need to go in for the
view that there is more than one concept of truth to make the point. We
78 M. P. Lynch
can, as suggested above, simply stick with the idea that truth is a func-
tional property. To define truth functionally in the sense I am interested
in is to define it by way of its connections to other related concepts.
These connections are embodied in certain common truisms that have
played a central role in the historical discussions over truth: the equiva-
lence schema; the slogan we called “Represent”; the idea that what makes
a proposition correct to believe is that it is true; that what valid infer-
ences preserve is truth; that truth is a worthy goal of inquiry; and the
like. The idea is that these truisms, or ones very much like them, jointly
pick out the truth-role. Beliefs are true just when they are correct, when
they are the sort of beliefs we aim to have during inquiry, when the
world is as they portray it as being. That is the job description of truth,
as it were.
It seems to me that this is just the sort of view that both Blackburn’s
quasi-realist and Price’s expressivist need. It allows us to say, against tradi-
tional theorists, that truth as such may be a very thin, functional prop-
erty. But seeing truth as a functional property in this sense is entirely
consistent with there being other properties, distinct from truth, which
satisfy this description and so realize that thin property—which, in short,
play the truth-role. And that is a helpful thought, because it opens to
door to appealing to these other properties to explain what grounds the
norm in different domains or discourses. It allows us new tools for
addressing what plays the truth-role for different kinds of content and,
therefore, giving us an easy explanation for what makes those kinds dif-
ferent kinds in the first place. They are different kinds because they are
subject to being correct in different ways.
Of course, the ultimate suitability of this suggestion rides in part on
whether we can make sense of a property—distinct from correspon-
dence—that could play the truth-role for domains like the ethical. One
possibility, as already noted, is to say that truth is sometimes realized by
an extremely deflated property. Another is to follow Blackburn’s earlier
lead: to look at the “materials at hand,” and to see whether we can con-
struct such a property from properties already in play in the moral
realm—properties, for example, like coherence. This, in fact, was pre-
cisely the property that Blackburn put to such effective use in his earlier
(1984) work.
The basic recipe might be sketched as follows. Step one: look to what
warrants moral beliefs. According to a leading theory of moral epistemol-
ogy inherited from Rawls, a moral belief is warranted to the degree to
which the moral framework to which it belongs is in a state of wide
reflective equilibrium. We can, if we like, see this as a form of coherence
theory: S’s moral judgment that p is warranted to the degree that it
coheres with the rest of S’s moral and non-moral judgments. The term
“coherence” can be taken here to pick out a family of epistemic desider-
ata. It is generally thought that a framework is coherent insofar as, and to
the degree to which, its members display relations of mutually explana-
tory support, it is complete, and it is consistent. Call these coherence-
making features. Such features themselves come in degrees: members of a
framework can be more or less consistent, more or less mutually explana-
tory, and so on. A framework of judgments increases in coherence to the
degree to which it exemplifies these features, on balance, to a greater
degree. “On balance” because the features are not themselves isolated in
their coherence increasing power. A framework would not be more coher-
ent on balance, for example, simply by increasing its size (completeness)
by including consistent but explanatorily unconnected judgments.
Intuitively, by increasing its explanatorily isolated judgments, the coher-
ence of the framework would on balance remain static or decrease.
Step two: use these definitions to make sense of what it would be for a
moral framework to improve in coherence. Framework F is more coher-
ent at t2 than at t1 when, on balance, it has at t2 either more of the
coherence-making features or some of those features to a greater degree.
So, if completeness, consistency, and explanatory connectedness are
coherence-making features, adding a consistent and explanatorily con-
nected judgment to the system will increase that system’s coherence along
those dimensions. Consequently, we can say that P coheres with moral
framework F if, and only if, including P in F would, on balance, make F
more coherent.
Step three is to use this notion of improvement to build a property that
could be said to realize truth (as opposed to warrant). This is not a trivial
task. But neither is it hopeless. One suggestion, building on work by
Wright and others, would be to say that if warranted moral judgments are
those that are coherent, then true moral judgments are those that are, as
we might say, supercoherent:
80 M. P. Lynch
SUP: The moral judgment P supercoheres with F if and only if P

coheres with F at some stage of inquiry and would continue to
do so through all successive improvements of information to F,
both moral and non-moral.
Call an individual moral judgment supercoherent just when it coheres

with some moral framework. The suggestion, then, comes to this: when
a moral judgment is true, it is true in virtue of being supercoherent.
Obviously, there is much more to say about each step of this recipe,
and about the strategy overall. I’ve tried to say some of it elsewhere
(Lynch 2009). But whatever the details, I think that despite Blackburn’s
own changes of heart over the years, the case remains strong for think-
ing that the quasi-realist should embrace truth pluralism. On the one
hand, it secures the “quasi” in quasi-realism: the view is consistent with
there being no distinctively moral properties in the world. Second, it is
consistent with the thought that our value judgments have a different
function than our judgments about the natural world (Blackburn
1984). But, unlike a more deflationist quasi-realism, the theory still
allows for a sharp contrast between the moral and the non-moral: the
door is open to holding that value judgments are true in virtue of super-
coherence, and that judgments about the physical world are true in
virtue of representing entities in the natural world. Yet the view would
also secure the realist side of the equation, as supercoherence theories of
what realizes truth in the ethical domain seem to allow for the possibil-
ity of moral error. A judgment’s being supercoherent is a significant
cognitive achievement. Many of our current moral judgments may not
be supercoherent. Indeed, it is possible that none of them are. And, as
already noted, the view allows for a notion of improvement of moral
judgment.
The problem of double-counting turns out to be no more or less of a
problem for truth pluralism than it is for any other view that wishes to
capture both semantic diversity and cognitive unity. When the philo-
sophical tax-man marches through the door, pluralists can be proud of
their books. Or, at least, they can be as proud as any other philosophical
accountant who toils to record the credits and debits incurred in our mad
little business—the business of reconciling the manifest and scientific
images of ourselves.13
Notes
1. The analogy is Nietzsche (1968, p. 540).
2. Recent examples include (Bar-On and Chrisman 2009; Blackburn
1998a; Ridge 2006, 2009; Schroeder 2008, 2010).
3. See (Lynch 2013a).
4. See for example, (Wright 1998b) and (Lynch 2009, 2013b); in contrast,
see (Kölbel 2008) (Cotnoir 2009).
5. The functionalist way of putting the point was first expressed by Pettit
(1996) and Lynch (1998); Wright (2013) now interprets his view in a
similar way.
6. Expressions of the pluralist position (see Wright 1992; Lynch 2001) have
sometimes obscured this point by presenting the position as relativizing
or indexing truth realizing properties to domains. As David (2013) has
correctly noted, these flourishes are unessential to the truth pluralist’s
point. See also Lynch 2013b.
7. A recent sampling includes: Cotnoir 2009, 2013a, b; Edwards 2008,
2018; Horton and Poston 2012; Jarvis 2012; Pedersen 2006, 2012a, b;
C. Wright 2013; and C. D. Wright 2005, 2010, 2012.
8. One might also see Jamin Asay as making a similar point (Asay 2018).
9. Quine, interestingly, does not say who these philosophers were. Clearly,
though, truth pluralism was, as it were, in the air at this time.
10. Here I refer only to traditional correspondence theorists (see, e.g., Russell
1966 and Fumerton 2002). Such a view should be distinguished from a
position—which is intended to be a form of truth pluralism—that says that
correspondence itself comes in different kinds (see, e.g., Sher 2004, 2005).
11. Of course Quine himself didn’t think there are moral or modal proper-
ties, and Blackburn would agree. These are not the sorts of things we
expect to quantify over in our final, most rigorous theory of the world.
12. Wright makes a similar point in his 1998 response to Blackburn (Wright
1998a).
13. Many people have contributed to my thoughts on these matters over the
years, including notably, Nikolaj J. L. L. Pedersen, Jeremy Wyatt, and
Nathan Kellen—the editors of this volume—and Crispin Wright, Simon
Blackburn, Huw Price, Mark Timmons, Mike Ridge, Mark Chrisman,
Dorit-Bar On, and Cory Wright.
While working on this paper, I benefitted from participation in the
Pluralisms Global Research Network (National Research Foundation of
Korea grant no. 2013S1A2A2035514). This support is also gratefully
acknowledged.
82 M. P. Lynch
References
Research 96 (1): 175–191. https://doi.org/10.1111/phpr.12303.
Bar-On, D., and M. Chrisman. 2009. Ethical Neo-expressivism. In Oxford
Studies in Metaethics, ed. R. Shafer-Landau, vol. 4, 132–165. Oxford: Oxford
University Press.
Beall, Jc. 2013. Deflated Truth Pluralism. In Truth and Pluralism: Current
Debates, ed. Nikolaj J.L.L. Pedersen and C.D. Wright, 323–338. New York:
Blackburn, S. 1984. Spreading the Word. Oxford: Oxford University Press.
———. 1998a. Ruling Passions. Oxford: Oxford University Press.
———. 1998b. Wittgenstein, Wright, Rorty and Minimalism. Mind 107
(425): 157–181.
———. 2013. Deflationism, Pluralism, Expressivism, Pragmatism. In Truth
and Pluralism: Current Debates, ed. Nikolaj J.L.L. Pedersen and C.D. Wright,
Brandom, R. 1998. Making It Explicit: Reasoning, Representing, and Discursive
Commitment. Cambridge, MA: Harvard University Press.
Chrisman, M. 2008. Expressivism, Inferentialism, and Saving the Debate.
Philosophy and Phenomenological Research 77 (2): 334–358.
Analysis 69 (3): 473–479.
Cotnoir, A.J. 2013a. Validity for Strong Pluralists. Philosophy and
Phenomenological Research 86 (3): 563–579.
———. 2013b. Pluralism & Paradox. In Truth and Pluralism: Current Debates,
ed. Nikolaj J.L.L. Pedersen and C.D. Wright, 339–350. New York: Oxford
University Press.
David, M. 2013. Lynch’s Functionalist Theory of Truth. In Truth and Pluralism:
Current Debates, ed. Nikolaj J.L.L. Pedersen and C.D. Wright, 42–68.
Dodd, J. 2013. Deflationism Trumps Pluralism. In Truth and Pluralism: Current
Dreier, J. 2004. Meta-Ethics and The Problem of Creeping Minimalism.
Philosophical Perspectives 18 (1): 23–44.
Dworkin, R. 1996. Objectivity and Truth: You’d Better Believe It. Philosophy &
Public Affairs 25 (2): 87–139.

68 (2): 143–149.
Edwards, D. 2018. The Metaphysics of Truth. Oxford: Oxford University Press.
Fumerton, R.A. 2002. Realism and the Correspondence Theory of Truth. Lanham:
Rowman & Littlefield Publishers.
Horton, M., and T. Poston. 2012. Functionalism about Truth and the
Metaphysics of Reduction. Acta Analytica 27 (1): 13–27.
Horwich, P. 1998. Truth. New York: Oxford University Press.
Jarvis, B.W. 2012. The Dual Aspects Theory of Truth. Canadian Journal of
Philosophy 42 (3–4): 209–233.
Kölbel, M. 2008. “True” as Ambiguous. Philosophy and Phenomenological
Research 77 (2): 359–384.
Lynch, M.P. 1998. Truth in Context. Cambridge, MA: MIT Press.
———. 2001. Functionalist Theory of Truth. In The Nature of Truth, ed. M.P.
Lynch, 723–750. Cambridge: MIT Press.
———. 2013a. Expressivism and Plural Truth. Philosophical Studies: An
International Journal for Philosophy in the Analytic Tradition 163 (2): 385–401.
———. 2013b. Three Questions for Truth Pluralism. In Truth and Pluralism:
Nietzsche, F. W. 1968. The Will to Power. Edited by W. Kaufmann. Translated by
R. J. Hollingdale. Vintage books ed. New York: Vintage Books.
Teach Us About Alethic Pluralism? Monist 89 (1): 102–117.
———. 2012a. Recent Work on Alethic Pluralism. Analysis 72: 588–607.
———. 2012b. True Alethic Functionalism? International Journal of
Philosophical Studies 20: 125–133. (Symposium on M. P. Lynch’s Truth as
One and Many).
Price, H. 2011. Naturalism Without Mirrors. New York: Oxford University
Press.
———. 2013. Expressivism, Pragmatism and Representationalism. New York:
Quine, W. 1960. Word and Object. Cambridge: MIT Press.
Ridge, M. 2006. Ecumenical Expressivism: Finessing Frege. Ethics 116 (2):
302–336.
84 M. P. Lynch
———. 2009. The Truth in Ecumenical Expressivism. In Reasons for Action, ed.
D. Sobel and S. Wall, 219–242. New York: Cambridge University Press.
Russell, B. 1966. On the Nature of Truth and Falsehood. In Philosophical Essays,
ed. B. Russell, 147–159. London: George Allen & Unwin.
Sainsbury, R.M. 1996. Review: Crispin Wright: Truth and Objectivity.
Philosophy and Phenomenological Research 56 (4): 899–904.
Schroeder, M. 2008. Being For. Oxford: Oxford University Press.
———. 2010. Noncognitivism in Ethics. London: Routledge.
Sher, G. 2004. In Search of a Substantive Theory of Truth. Journal of Philosophy
101 (1): 5–36.
———. 2005. Functional Pluralism. Philosophical Books 46 (4): 311–330.
Predicates. Analysis 57 (3): 209–210. https://doi.org/10.1111/1467-8284.
00077.
Wright, C. 1992. Truth and Objectivity. Cambridge, MA: Harvard University
Press.
———. 1998a. Comrades Against Quietism: Reply to Simon Blackburn on
Truth and Objectivity. Mind 107 (425): 183–203.
———. 1998b. Truth: A Traditional Debate Reviewed. Canadian Journal of
Philosophy 28 (Supplement): 31–74.
———. 2013. Plurality of Pluralisms. In Truth and Pluralism: Current Debates,
ed. Nikolaj J.L.L. Pedersen and C.D. Wright, 123–156. New York: Oxford
University Press.
Synthese 145: 1–28.
———. 2009. Truth, Ramsification, and the Pluralist’s Revenge. Australasian
Journal of Philosophy 88 (2): 265–283.
———. 2012. Is Pluralism about Truth Inherently Unstable? Philosophical
Studies 159 (1): 89–105.
The Metaphysics of Domains
Douglas Edwards
1 Introduction
Pluralist theories of various sorts commit themselves to the notion of a
domain. Most prominently, truth pluralists suggest that there are differ-
ent stories to tell about how sentences1 get to be true depending on the
domain to which the sentence belongs.2 Indeed, some commitment to
the idea is ingrained in the very motivations for the view. The truth plu-
ralist’s dissatisfaction with truth monism is that one theory of truth does
not seem to fit all the different kinds of thought and talk that we ordinar-
ily take to be truth-apt, so we need to open ourselves up to the idea that,
rather than having one global theory of truth, we have a number of dif-
ferent theories of truth which apply locally, for different kinds of thought
and talk.
There is also a practical reason why truth pluralists commit themselves
to domains. Suppose for a moment that we hold that there are many
distinct truth properties, and we have no division of sentences into differ-
ent domains. On this picture, all sentences are capable of possessing all
D. Edwards (*)
Department of Philosophy, Utica College, Utica, NY, USA

https://doi.org/10.1007/978-3-319-98346-2_4
86 D. Edwards
the different truth properties. Now, the different truth properties are dis-
tinct properties, so it is possible for a sentence to possess one of the truth
properties but lack another. If we take falsity to be lack of truth, then we
have a sentence which both has a truth property and lacks a truth prop-
erty, which gives us a sentence that is both true and false. This compli-
cates matters and makes life difficult for pluralist proposals that do not
try to sort sentences into separate domains. Given that, as noted above,
truth pluralists have independent reason to hold that sentences are sorted
into domains, it seems natural to adopt them in response to this sort of
problem.
However, some pluralists are sceptical of the notion of a domain, and
Lynch (2013a) ends up rejecting domains, choosing instead to start with
specific sentences (or propositions, for Lynch), as opposed to sentences of
a particular kind. This, to my mind, is a mistake. As just noted, a key
component of pluralist views is that different kinds of subject matters are
apt for different sorts of treatment. If this idea is given up, then a central
motivation for the view is lost. As a result, I think that pluralists should
take the notion of a domain seriously as a central aspect of the view.
In this chapter, I give an account of what domains are and respond to
some problems of domain individuation. I begin by discussing the
semantic and metaphysical aspects of a domain, and then show how the
two aspects combine to yield an account of domains in general. Following
this, I will discuss two problems of domain individuation: the problem of
mixed atomics, and the problem of mixed compounds.
2 The Semantic Aspect of a Domain

I characterize domains as having two main components: a semantic com-
ponent and a metaphysical component.3 The semantic component con-
cerns the components of atomic sentences—singular terms and
predicates—and the metaphysical component concerns the things
referred to by singular terms and predicates, namely objects and proper-
ties. The discussion of the semantic aspect of a domain will thus concern
the ways in which singular terms and predicates come in different kinds,
and how this contributes to the notion of a domain.
We will begin with predicates—linguistic items, typically of the form

‘is F’, which, when combined with a singular term, form an atomic sen-
tence. My aim in this section is to substantiate the claim that predicates
come in different kinds, and that predicate kinds are distinguished by the
kinds of functional roles that predicates have. This will give us the grounds
for understanding the semantic aspect of a domain.
The general idea can be illustrated by thinking about how to distin-
guish moral predicates from physical predicates. Generally speaking,
predicates typically refer to properties. One idea is that there is something
distinctive about the properties that moral predicates pick out that ren-
ders them distinct from the properties that physical predicates pick out.
One feature might be that moral predicates pick out normative proper-
ties, where one’s recognition that an action or object has a normative
property entails that one has, in virtue of that recognition, a reason to act
or to refrain from acting. This would be distinct from a physical predi-
cate, which ascribes a property that has causal, but not normative, pow-
ers. Moral predicates will ascribe properties whose recognition gives one
a distinctly moral reason to act or not to act.
We can elaborate on what is meant by ‘moral’ here by thinking about
some general ‘platitudes’ concerning the word ‘moral’. For example,
Smith (1994) offers a set of platitudes which he takes to serve as analyses
of moral concepts:
To say that we can analyse moral concepts, like the concept of being right,
is to say that we can specify which property the property of being right is
by reference to platitudes about rightness: that is, by reference to descrip-
tions of the inferential and judgemental dispositions of those who have
mastery of the term ‘rightness’. (Smith 1994: 39)
Smith goes on to describe platitudes that concern the ‘substance’ of

morality:
For example, there are platitudes concerning the substance of morality:

‘Right acts are often concerned to promote or sustain or contribute in some
way to human flourishing’ (Foot. 1958); ‘Right acts are in some way
expressive of equal concern and respect’ (Dworkin 1977: 179–83; Kymlicka
88 D. Edwards
1989: 13, 21–9; 1990: 4–5), and the like. What these platitudes about
substance force us to admit, at the very least, is that there are limits on the
kind of content a set of requirements can have if they are to be moral
requirements at all, as opposed to requirements of some other kind (Dreier
1990). (Smith 1994: 40)
Smith offers a detailed account of the kinds of features moral proper-

ties have, which demarcates them from other kinds of properties. This
process is explicitly carried out by reference to the relevant concepts of
these properties, but, for our purposes, we can interpret Smith’s thoughts
in relation to moral predicates. The suggestion is that we can individuate
kinds of predicates in accordance with the general functional roles that
those predicates are taken to have.
Whilst we have just talked about moral and physical predicates so far,
the idea is intended to be general, in that we could construct a list of
general features for all kinds of predicates. These are intended to mark
fairly intuitive distinctions between kinds of subject matter that are
already implicit, in that we generally have a good sense of when someone
is making a moral claim, or an aesthetic claim, or a physical claim, or a
mathematical claim, to take just a few examples. Indeed, philosophers
should have a good sense of how to individuate these things, otherwise it
might be difficult to distinguish aesthetics from philosophy of mathe-
matics! The thought here is that this is, at least in part, done by the func-
tions of the words involved, particularly the predicates, in relation to the
types of features they are describing.
However, the fact that we can often individuate different kinds fairly
easily does not mean that there won’t occasionally be difficulties demar-
cating them. For example, compare religious predicates with moral predi-
cates. Both are concerned with describing how to live well, and how to
behave towards others and ourselves. In this way, they have similar func-
tional roles, which may lead one to think that they are essentially the
same. It is not much of a stretch to imagine that someone coming to their
ethics class with the impression that moral predicates just are religious
predicates. Perhaps, for instance, they think that the predicates ‘is right’
and ‘is wrong’ are equivalent in meaning to the predicates ‘is pious’ and
‘is sinful’, as they have only considered moral issues in religious contexts.
However, with some philosophical work we can separate the religious

predicates from the moral predicates. The predicates ‘is pious’ and ‘is sin-
ful’, for example, concern a person’s relationship with God, and to hold
that ‘is right’ and ‘is wrong’ are equivalent to those terms is just to hold
that actions are right or wrong insofar as they affect an individual’s rela-
tionship with God in a particular way. However, we can show that ‘is
right’ and ‘is wrong’ have a broader role to play in that we can consider
the rightness or wrongness of actions under different background views,
which show that the meanings of the predicates ‘is right’ and ‘is wrong’
are not tied to a person’s relationship with God, unlike ‘is pious’ and ‘is
sinful’.4
The idea is, then, that the first part of the semantic aspect of a domain
is understood in terms of kinds of predicates, with kinds of predicates
understood in terms of functional roles. The functional roles are under-
stood in terms of the features of the properties that the predicates are
purported to pick out. We will say more about the semantic aspect again
in a moment when we look at singular terms, but let’s say a bit first about
the metaphysical aspect.
3 The Metaphysical Aspect of a Domain

The account given above of the individuation of predicate kinds is neutral
on the substantial metaphysical matter of whether there are any proper-
ties that correspond to these predicates, and, if so, what those properties
are like. For every predicate kind F that is individuated, there are three
available options:
(i) There are objective F properties.

(ii) There are projected F properties.
(iii) There are no F properties.
Option (i) holds that there are properties ‘out there’, so to speak, and
this can explain why we have predicates of a certain kind: they are respon-
sive, in some way, to existent properties. Option (ii) holds that there are
90 D. Edwards
properties that the relevant predicates pick out, but these properties exist
because they are projections of the predicates, as opposed to being already
existing things that our predicates respond to. Option (iii) is the error-
theoretical option, where, perhaps due to the specification of the predi-
cate kind, or alternatively because the world is uncooperative, there are
no properties that correspond to the predicates of a certain kind. An error
theorist about morality, for example, holds that there are no properties
that correspond to our moral predicates.5
Our focus here will be on (i) and (ii). We can illustrate the differences
between them using the following principle about properties:
(P) The object referred to by ‘a’ falls under the predicate ‘is F’ iff the
object referred to by ‘a’ has the property referred to by ‘is F’.
Option (i) holds that there is a right-to-left order of determination on

(P): it is because A has the property of being F that A falls under the predi-
cate ‘is F’. For example, it is because this knife has the property of being
metallic that this knife falls under the predicate ‘is metallic’. We have a
property-to-predicate direction of explanation.
Option (ii), on the other hand, takes the reverse view: there is a left-to-
right order of determination on (P) as it is because A falls under the predi-
cate ‘is F’ that A has the property of being F. One example here is the
property of being cool6: Zack Morris has the property of being cool
because Zack Morris falls under the predicate ‘is cool’, rather than vice
versa. We have a predicate-to-property direction of explanation.
In order to make sense of this idea, we need to make distinctions
between kinds of properties. It is commonplace to distinguish between
abundant and sparse properties in theories of properties.7 First, let us take
the view of properties standardly called ‘predicate nominalism’.8 On this
view, there are properties insofar as there are extensions of predicates: to
have the property of being yellow is to be in the extension of ‘is yellow’;
to have the property of being mammalian is to be to be in the extension
of ‘is mammalian’, and that’s all there is to it. On this view, properties are
taken to be abundant: there are as many properties as there are extensions
of predicates.9 This allows us to be very permissive when it comes to
assigning properties to predicates, for any predicate with consistent rules

for its application will have an extension: a class of (actual and/or possi-
ble) objects that satisfy it. Thus, the predicate ‘is yellow or mammalian’
will have an extension—the class of objects that are either yellow or
mammalian—and thus will express a property. On this view of proper-
ties, it is not the case that an object is in the extension of a predicate ‘is F’
because it has the property of being F; rather the object has the property
of being F because it falls under the predicate ‘is F’.10
An opposing view is the idea that that properties form a distinct onto-
logical kind, such as universals.11 On this view, things that share a prop-
erty instantiate the same universal, and in order for two or more objects
to instantiate a universal, it must be shown that they share some signifi-
cant similarities, and that the universal plays some important causal-
explanatory role.12 If we take the only properties to exist to be those that
are universals, then this limits the number of properties to the number of
universals. As universals require some significant similarities at the meta-
physical level, the only properties that exist will have to meet this require-
ment, which renders properties sparse.
There is also room for a mixed approach, originally due to David Lewis
(1983), that allows for both abundant and sparse properties. Lewis holds
that all properties are classes, but, in some cases—abundant properties—
there won’t be anything more to say about why particular objects are
members of a class, but in others—sparse properties—there will.13 Or, to
put things in terms of predicates as we have been doing, in some cases
there will not be an account of why a particular class of objects falls under
a particular predicate, whereas, in other cases, there will.
For example, consider the predicate ‘is yellow or mammalian.’ Some
things are in the extension of this predicate, and are thus members of the
class of things that are yellow or mammalian. However, this doesn’t imply
that all of the members of this class share anything significant in com-
mon over and above their membership in the class. Contrast this now
with, for example, falling under the predicate ‘is metallic’, and being a
member of the class of metallic objects: membership of this class is not
arbitrary, and each member of the class must meet some significant con-
straints in order to be a member of the class. Indeed, it is plausible to
92 D. Edwards
think that all metallic objects share some significant feature in common.
Despite the fact that both properties are classes, there are significant
metaphysical differences between them, leading to the former’s status as
an ‘abundant’ property, and the latter’s status as a ‘sparse’ property. In
particular, sparse properties ground genuine similarities between their
bearers, and have a causal-explanatory role, whereas abundant properties
do not ground genuine similarities between their bearers, and do not
have a causal-explanatory role.
The two ways of looking at properties broadly correspond to the two
ways we considered the predicate-property relation above, with sparse
properties being those mentioned in option (i) and abundant properties
being those mentioned in option (ii). To relate this to our discussion of
predicates, I am going to use the terminology of ‘responsive’ predicates
and ‘generative’ predicates. Broadly speaking, responsive predicates
respond to sparse properties, and generative predicates generate abun-
dant properties.14
4 Relating Objects and Singular Terms

We have spoken so far about predicates and properties, but, to complete
the picture, we need to also discuss singular terms and objects. The gen-
eral function of singular terms is to refer to objects,15 and we can divide
singular terms into kinds by the sorts of objects they are used to refer to.
For example, we can distinguish between ‘the largest prime number’, ‘the
New York-Massachusetts border’, and ‘the guilty party’ as different types
of singular terms in a sense, namely mathematical, institutional, and
moral. This is partly because of how they are used: ‘the largest prime
number’ is used to refer to a mathematical object; ‘the New York-
Massachusetts border’ is used to refer to an institutional object; and ‘the
guilty party’ is used to refer to a moral object.
Moreover, the case can be made that there is a distinction between
responsive and generative singular terms, which lines up with the corre-
sponding distinction between predicates. For example, if we take causal
accounts of reference seriously, at least for physical objects, then the use
of a singular term is responsive to the existence of a physical object which
causally constrains its use. Such objects would be considered sparse

objects.
Now, not all objects will meet this condition: abstract objects, for
example, do not have causal powers, and do not look suited for a causal
theory of reference. However, we can give a different understanding of
abstract objects by using the ‘Neo-Fregean’ approach of Bob Hale and
Crispin Wright. Their idea is that to be an object is to be the referent of a
singular term appearing in a true sentence:
The lynch-pin of Frege’s platonism, according to our interpretation, is the

syntactic priority thesis: the category of objects … is to be explained as
comprising everything which might be referred to by a singular term,
where it is understood that possession of reference is imposed on a singular
term by its occurrence in true statements of an appropriate type. (Wright
1983: 53)
[O]bjects, as distinct from entities of other types (properties, relations, or,

more generally, functions of different types and levels), just are what (actual
and possible) singular terms refer to. (Hale and Wright 2005: 171)
The basic idea for our purposes here is that we can secure reference
to a mathematical object—the number 5, say—by noticing that its
associated singular term appears in at least one true sentence. Thus,
the truth of the sentence ‘the number 5 is prime’ is sufficient to secure
the existence of a referent for the singular term ‘the number 5’: the
number 5.
We can use this idea to construct the notion of an abundant object,
which pairs with the notion of an abundant property. If we hold that an
object a is abundant, then it cannot be that the singular term ‘a’ refers to
a because a exists, as we are explicitly denying that a has existence prior to
‘a’ referring to a. As a consequence, we need some grounds to establish in
virtue of what it is that ‘a’ refers to a. Hale and Wright offer an account
here in terms of truth: ‘a’ refers to a if ‘a’ appears in a true sentence.
Singular terms can thus be responsive or generative: responsive singular
terms respond to sparse objects, and generative singular terms generate
abundant objects.
94 D. Edwards
5 Summarizing Domains
I have suggested that a domain has both a semantic and a metaphysical
component, which are closely related. The semantic component of a
domain is understood in terms of singular terms and predicates. The
domain to which a singular term or predicate belongs is determined by
the kind of singular term or predicate that the relevant singular term or
predicate is identified to be. This is done by identifying the functional
role of a singular term or predicate kind to demarcate different kinds of
singular terms or predicates. The semantic aspect of a domain will thus be
understood as a singular term and predicate kind. For example, the
semantic aspect of the moral domain will be the singular terms and predi-
cates identified as being moral singular terms and predicates in virtue of
their playing the functional role associated with moral singular terms and
predicates.
The metaphysical aspect of a domain will be composed of the objects
and properties that the singular terms and predicates in the semantic
aspect of the domain refer to. Thus, the metaphysical aspect of the
moral domain, for example, will be the objects and properties referred
to by moral singular terms and predicates. This is not to say that all
there is to objects and properties of different kinds is that they are
referred to by different kinds of singular terms and predicates. This is
because, as we have seen, the relationships between singular terms and
objects, and predicates and properties, varies: in some cases, the nature
of the objects and properties is dependent on the singular terms and
predicates (the abundant model), whereas in other cases, the singular
terms and predicates are dependent on the objects and properties (the
sparse model).
This account of domains is intended to be available to theorists of vari-
ous sorts, not just truth pluralists. In the remainder of the chapter, I will
focus on a couple of problems of domain individuation that have been
posed for truth pluralism, though I think the problems remain general
problems for those who wish to individuate different kinds of subject
matter.
6 The Problem of Mixed Atomics

The problem of mixed atomics (Sher 2005; David 2013; Wyatt 2013)
puts pressure on the idea that we can assign sentences to domains. The
first class of examples of mixed atomics takes atomic sentences that are
composed of singular terms and predicates of different kinds. Two exam-
ples are ‘this crystal is beautiful’ (David 2013) and ‘Charlie is delicious’
(Wyatt 2013), where ‘Charlie’ is the name of a beet. Let us take Wyatt’s
example, ‘Charlie is delicious’. What is this sentence about? It is about
Charlie, a beet, a material object. The problematic element is what is
being said about Charlie, namely that he is delicious, which is not a mate-
rial property (we suppose). Here, we have singular term and predicate
belonging to different kinds, so what domain does the sentence belong to?
Wyatt (2013) claims that these examples should push the truth pluralist
to admit that sentences like this belong to more than one domain, even if
they are apt for only one kind of truth. I will argue though that the exam-
ples do not push the truth pluralist in that direction, and that truth plu-
ralists can still say that atomic sentences only belong to one domain.
By wearing its commitment to domains on its sleeve, truth pluralism
undoubtedly leaves itself open to these kinds of concerns about assigning
sentences to domains. However, many other theories in the area will face
the same sorts of problems. Truth pluralists take the demands for truth-
aptness to be very minimal, and focus their attention on what kind of
truth a sentence is apt for. For other views, such as classical correspon-
dence theories of truth, there is only one kind of truth, but the key ques-
tion becomes whether sentences in some domains are capable of being
true at all. Crucial to this sort of project is the ability to separate the
truth-apt sentences from the non-truth-apt ones, and this involves sepa-
rating sentences into different domains, as the examples of non-truth-apt
sentences on such views often include sentences employing moral or aes-
thetic terms.16
With this in mind, suppose we take it that physical sentences are truth-
apt, and aesthetic sentences are not. Now, take Wyatt’s example of ‘Charlie
is delicious’: is this a physical or an aesthetic sentence, or both? It
96 D. Edwards
s eemingly cannot be both, as a sentence cannot both be and not be truth-

apt, so it seems as though Wyatt’s preferred option of a sentence belong-
ing to two domains is not even in play here. The answer though seems
fairly obvious: the aesthetic predicate ‘is delicious’ immediately renders
the sentence non-truth-apt, regardless of the object involved. This sug-
gests that the key factor in determining the domain to which a sentence
belongs is the predicate, and this is a conclusion that can be reached
independently of any commitment to truth pluralism.
Pedersen and Wright suggest a solution in this spirit for truth plural-
ism itself:
Consider the following sentences:
(41) The Mona Lisa is beautiful.

(42) Speeding is illegal.
Prima facie, what determines the domain-membership of (41) and (42)

is the aesthetic and legal predicates ‘is beautiful’ and ‘is illegal’, respectively.
It is an aesthetic matter whether the Mona Lisa is beautiful; this is because
(41) is true in some way just in case the Mona Lisa falls in the extension of
the aesthetic predicate ‘is beautiful’ (and mutatis mutandis for (42)).
(Pedersen and Wright 2013)
With this in mind, remember that an atomic sentence is composed of

two elements: a singular term and a predicate. We can distinguish between
two things: what a sentence is about, and how the thing the sentence is
about is represented to be. We can grant that a sentence is about the
object referred to by the singular term, for example, ‘Charlie is delicious’
is about Charlie, but what makes a sentence a sentence in that it is a
bearer of content is that there is a way the object is represented to be:
Charlie is represented to be delicious, and this representation occurs due
to the attribution of a property to the object.
So, it is not what a sentence is about that we should be considering for
domain membership, it is rather how the thing the sentence is about is
represented, by the use of a predicate to attribute a property.17 We can
thus say of our examples that they belong to a single domain: because
deliciousness is an aesthetic predicate, ‘Charlie is delicious’ is just an aes-

thetic sentence, and because wrongness is a moral predicate, ‘torture is
wrong’ is just a moral sentence. They belong to these domains because of
the functions of the predicates involved. Consequently, simple mixes of
singular terms and predicates in a sentence should not motivate any move
from the view that each atomic sentence belongs to a single domain.
Atomic sentences are thus assigned to domains by the predicate they
contain. The singular term is not relevant to domain individuation. This
avoids the problem of mixed atomics, but there are independent reasons
to favour this account. Many different kinds of thing can be said about
the same object: a single chair can be blue, solid, beautiful, sad, danger-
ous, presidential, or singular, to name just a few examples. According to
the truth pluralist all of these things said about it will be of different
domains. Given that the nature of the singular term is held fixed, and the
only thing that varies is the predicate, this is independent reason to think
that domain-membership is determined by the predicate.
It is also worth pointing out that singular terms will nevertheless be
parts of domains. Each singular term will have its ‘home’ domain,
depending on what kind of object to which it refers. For instance, physi-
cal singular terms will have a certain character, just as physical predicates
will have a certain character. The central difference between singular
terms and predicates is that singular terms can be part of sentences of
domains other than their own, whereas predicates cannot. Whilst each
singular term has its ‘home’ domain, it does not just appear in sentences
of that domain, as detailed above.
However, one might worry that this solution to the problem of mixed
atomics is too simplistic. If we avoid the problem by holding that it is the
predicate that is relevant for determining domain-membership, then
what about predicates that appear to be members of more than one kind?
I will now attempt to allay that concern by considering an example of a
supposedly mixed predicate.
The mixed atomics with supposedly mixed predicates that I’ll examine
are ‘thick’ predicates, such as ‘courageous’, ‘kind’, or ‘lewd’. The sentence
‘Serena Williams is courageous’ might be considered mixed because ‘cou-
rageous’ expresses a property (or properties) with both moral and non-
moral aspects. We can note again that truth pluralism is not the only view
98 D. Edwards
to face this sort of problem. There is an abundance of work on whether

expressivism can deal with thick predicates, with the worry being—in
that case—that the expressive and descriptive elements are inseparable,
thus causing problems for the view that moral language is purely
expressive.18
One strategy in response is to take sentences involving thick predicates
to be compound sentences, as opposed to atomic sentences. Williams
(1985) puts this approach nicely as follows:
A statement using [thick] terms can be analyzed into something like “this
act has such-and-such a character, and acts of that character one ought not
to do.” It is essential to this account that the specific or “thick” character of
these terms is given in the descriptive element. The value part is expressed,
under analysis, by the all-purpose prescriptive term ought. (Williams 1985:
144)
A truth pluralist could adopt this solution by holding that we are no

longer dealing with a problem of mixed atomics, but rather a problem of
mixed conjunctions. In this case there will be no problem accounting for
the domains of the atomics, as—in line with what was said above—the
truth pluralist can use the properties attributed to identify the domains
the atomic sentences belong to. The only problem might be how to
account for the truth of the conjunction, but then the truth pluralist is in
familiar territory and can apply her favourite solution to the problem of
mixed conjunctions, which we will discuss below.
However, this option has some downsides. For one thing, whilst it
promises to specify the way a sentence involving a thick predicate is true,
it does not answer the question of what domain it belongs to: is it still a
mixture of a moral and a physical sentence, for example? Accordingly, it
is worth exploring an alternative account, which is inspired by Michael
Smith’s (2013) account of thick predicates.
Smith suggests that the difference between thick and thin predicates is
not as a difference in kind, but in degree. Thick predicates seem to be
thick because they impose a certain amount of descriptive content on the
entity that bears the property. For example, saying ‘Serena Williams is
courageous’ attributes not only some positive moral feature to Serena
Williams, but also some descriptive features, namely that she displays
some sort of resistance in the face of danger, for example. This may sug-
gest to some that there is some difference in kind between thick predi-
cates, like courageousness, and thin predicates, such as goodness, as one
imposes descriptive content whereas the other does not. However, this is
a mistake. Even thin properties like being morally good impose some
descriptive content on their bearers. For example, if an action is deemed
to be morally good, then this at the very least implies that the action was
carried out by an intentional agent, as opposed to simply being the result
of happenstance.19 The movement of the branch of a tree in the breeze,
for example, would not be something that could be morally good, on the
grounds that it does not meet the requirements for being a bearer of the
property of being morally good. This suggests that thin predicates like
moral goodness do impose some descriptive content on their bearers,
even if it is more minimal than the descriptive content imposed by thicker
predicates.
If we take this line, then we can say that all moral predicates will have
a degree of thickness, with the thinner predicates being predicates like
moral goodness, and the thicker predicates being predicates like coura-
geousness. There will also be predicates thicker than courageousness
which impose very specific descriptive features on their bearers, such as
the property of being a good father.20 However, we would say that, despite
the variations in degrees of thickness, these are all moral predicates. The
fact that some moral predicates imply some descriptive content does not
mean that they are not moral predicates, as predicates of pretty much any
property imply some constraints on what can bear it. The status of these
predicates as moral predicates will be determined by the functional role
that these predicates play in the cultivation and evaluation of character,
and the way we decide what we ought to do.
Moreover, this explanation generalizes to thick predicates of any kind,
and is not just intended for moral predicates. The key idea is that we look
to the function of a predicate to determine the domain to which the sen-
tence in which it occurs belongs. For example, we might take the predi-
cates ‘is sublime’ and ‘is beautiful’ to be thick predicates, as, even though
they seem to be aesthetic predicates, they have different descriptive con-
tent. However, the fact that they have different descriptive content does
not affect their status as aesthetic predicates due to the function they have
as ways to evaluate of pieces of art, for example. Once again, the fact that
100 D. Edwards
some predicates also have some descriptive content does not affect their
key function.
I have suggested that the examples of mixed atomics divide into two
main classes: (i) those where object and property are from different
domains; (ii) those where the property seems mixed. I argued that neither
of these classes poses a problem for the account of domains given above.
7 The Problem of Mixed Compounds

A separate problem is presented by the problem of mixed compounds
(Tappolet 2000). The thought here is that we can easily use logical con-
nectives to form compound sentences using atomic sentences from dif-
ferent domains, such as ‘this cat is wet and this cat is funny’, to take
Tappolet’s (2000) example. The problem was originally posed as a chal-
lenge to explain how such compounds are true on a pluralist account, but
we can also view is as a challenge to the notion of a domain: to which
domain do such mixed compounds belong?
Some have seen this problem as requiring additions to the standard
domains that truth pluralists offer, including perhaps different domains
for each kind of compound (Kim and Pedersen 2018; Pedersen and
Lynch 2018; Gamester Forthcoming). We do not need to go in this direc-
tion, though. In previous work (Edwards 2008), I gave a response to this
problem as a problem for understanding how such sentences are true by
saying that compound sentences (whether mixed or not) are true in vir-
tue of their components being true in the way(s) specified by the logical
connective in question. The truth of a compound is derivative, in the
sense that it is dependent on the truth of its components. This suggests a
difference between the ways that compounds are true (in virtue of their
components) and the way that atomic sentences are true (in virtue of
having their relevant truth property). We can extend this idea to show
that there is a difference between considering the question of which
domain an atomic sentence belongs to and considering the question of
which domain a compound belongs to.
The key idea is that domain membership is only relevant when we
consider atomic sentences. This is because domains are composed of
objects and properties, and atomic sentences are composed of singular

terms and predicates that refer to those objects and properties. Accordingly
there will be a direct connection between the content of a domain and an
atomic sentence. Compound sentences, on the other hand, form an addi-
tional layer above this framework. They are composed of atomic sen-
tences, and the characteristic of a particular compound that makes it a
particular compound is not the kinds of terms that are contained in its
atomic components, but the particular connective involved, determined
by the conditions under which the compound is true. It is the truth con-
ditions of a conjunction, for example, which mark it out from a disjunc-
tion. This is what makes it a conjunction, as opposed to another form of
compound, and this has nothing to do with the kinds of atomic sentences
that compose it. With this idea in mind, we do not face the same sorts of
questions about domain-membership with compounds as we do with
atomics. The essence of a compound sentence which makes it a com-
pound sentence is not the kinds of atomic sentences that compose it, but
the truth conditions it has.
We do face a different sort of question though, namely if compound
sentences do not belong to the conventional domains demarcated, where
do they belong? In previous work (Edwards 2009, in response to Cotnoir
2009), I suggested that they belong to the logical domain, in that the
conditions under which each compound is true are determined by one’s
preferred logical system. On this account, each compound, whether
mixed or not, is a member of the logical domain.
I think that there is something to this general idea, though it needs some
development. One thing to consider is whether we see the logical domain
in this context as a regular domain amongst the other domains, or as some-
thing distinct. I mentioned above that one way to think about compound
sentences is as being at a level of an additional layer above atomic sentences.
This is because compound sentences are essentially manipulations of
atomic sentences into new structures: they take atomic sentences as ingre-
dients, and with the additional of some specified rules, they turn them
into new, compound sentences. As this process involves operations on
atomic sentences, as opposed to a different kind of atomic sentences, it is
more natural on this model to say that compound sentences do not reside
in a domain of similar status to other domains. As a consequence, if, as
102 D. Edwards
suggested earlier, we understand the metaphysical aspect of a domain as

being the entities from which atomic sentences are built, compound sen-
tences do not have a domain in this sense. This makes them an interesting
class of sentences for a pluralist, as they are sentences that are capable of
being true, but they do not have a domain. This is where we can draw upon
the idea suggested above that compound sentences are, in a sense, ‘logical’
in nature.
The generation of compound sentences involves the taking of atomic
sentences, and connecting them in some way in accordance with some
specified rules to make new sentences. It is the ‘connecting them in some
way in accordance with some specified rules’ that contains the logical com-
ponent here, as the rules by which compounds are formed are specified by
the basic axioms of one’s background logical theory. Now, it is an open
question for truth pluralists whether all logical principles apply so univer-
sally as to be applicable to sentences from all domains, with principles like
the principle of excluded middle being a particularly controversial case,
along with the notion of logical consequence.21 However, less controversial
is the construction of compounds, and their subsequent truth condi-
tions.22 What we have here are cases of uncontroversial principles, which
apply universally, such that they allow one to take atomic propositions of
any kind and connect them in specified ways to make new sentences.23
We can extract from this the way in which compound sentences are
both domain-less, yet also logical in some sense. They are domain-less
because they are constructions out of atomic sentences, and not them-
selves atomic sentences. However, they are logical in character because
they are the results of the implementation of logical principles on atomic
sentences, and they are true insofar as they meet the standards laid down
by these logical principles.24
8 Summary
As noted earlier, the notion of a domain has been both a key and contro-
versial aspect of pluralist theories. Problems of domain individuation,
such as the problem of mixed atomics and the problem of mixed com-
pounds, along with the general difficulties of ‘dividing up’ language into
different domains have presented a significant challenge to the role of the
notion of a domain in pluralist theories. My suggestion here has been

that there is an account of domains available to pluralists, and that these
problems of domain individuation have solutions, so domains can con-
tinue to play a key role in pluralist theories.
Notes
1. I will be using sentences, as opposed to propositions, as the main examples
of truth-bearers. See Edwards (2018: Chap. 1) for more on this choice.
2. For examples of domain-based ontological pluralism, see Cotnoir and
Edwards (2015) and Edwards (2018). See Lynch (2009) for an example
of domain-based logical pluralism.
3. Note that Wyatt (2013) recommends talking of ‘topics’ and ‘domains’ as
separate things, with Wyatt’s ‘topics’ loosely corresponding to my ‘meta-
physical aspects’, and Wyatt’s ‘domains’ loosely corresponding to my
‘semantic aspects’. I choose to use the word ‘domain’ for both, because I
do not think that these aspects can be separated enough to warrant them
being called different things, as opposed to parts of the same thing. I
hope it will become clear why below.
4. See Edwards (2018) for further development of this idea in relation to
social and institutional predicates.
5. Examples of error theory in morality are Mackie (1977) and Joyce
(2001).
6. See Haslanger (2012: 89–98) for an extended discussion of coolness.
7. For more on this distinction in relation to truth, see Edwards (2013).
8. The name is due to Armstrong (1978).
9. Note that this is not the most abundant view of properties available. As
Lewis (1983) notes, if we take the view that properties are classes (class
nominalism), then properties will be more abundant than on predicate
nominalism, as there will be classes to which there is no predicate attached.
10. See Edwards (2014: Chap. 5) for more on this view.
11. See, for example, Armstrong (1978). See also Edwards (2014: Chap. 2).
12. This perhaps requires that universals are taken to be the immanent uni-
versals favoured by Armstrong (1978), as opposed to abstract universals,
and I will assume that here.
13. See Edwards (2014: Chap. 6) for more on this idea.
14. Note that this terminology still applies if we are thinking about proper-
ties as classes, and classes as mind-independent. This is because, even if
104 D. Edwards
there is a vast number of classes, we still need to make sense of a predi-

cate selecting a particular class, and thus having the particular extension
it does, which will be dependent on our practices.
15. See, for example, Hale (1994).
16. By using ‘non-truth-apt’ here I am working through a case where the
correspondence theory is paired with some form of expressivism, as
opposed to a form of error theory. Regardless of whether the correspon-
dence theory is paired with expressivism, error theory, or indeed a form
of fictionalism, a distinction will need to be made between the sentences
that are able to be true and those that are not.
17. Lynch (2009, 2013a) just holds that what matters is what a sentence is
about, with aboutness encompassing both object and property, which
leads straight into the problem of mixed atomics.
18. Eklund (2011) provides an overview. Also see Lynch (2013b) for an argu-
ment suggesting that expressivists should embrace alethic pluralism.
19. Compare this thought to the idea that it is a fact about truth that—no
matter what theory of truth you have—truth is not a property that can
be borne by shirt buttons.
20. As discussed by Stewart-Wallace (2016).
21. See, for example, Lynch (2009: Chap. 5) for discussion of these issues in
relation to truth pluralism.
22. For instance, in the discussion above, the rules for the formation of a
conjunction are those specified by the axioms of classical logic, where a
conjunction is specified to be a compound proposition that is true if and
only if each of its conjuncts are true. However, this does not commit one
to classical logic in a substantial way, as what it takes for an atomic prop-
osition to be true here can be determined in a number of different ways,
in accordance with the general approach of truth pluralism.
23. The account given here is for truth-functional compounds only.
24. Thanks to Nikolaj J. L. L. Pedersen, Jeremy Wyatt, and Nathan Kellen
for very helpful feedback on this paper. I’d also like to thank audiences at
the Pluralisms Week conference at Yonsei University and the University
of Albany philosophy colloquium. Portions of this article draw on mate-
rial originally published in Chap. 4 of my book The Metaphysics of Truth
(Oxford University Press 2018), used with permission of Oxford
University Press.
acknowledged.
References
Armstrong, D.M. 1978. Universals and Scientific Realism. Cambridge:
Cotnoir, A.J. 2009. Generic Truth and Mixed Conjunctions: Some Alternatives.
Analysis 69 (2): 473–479.
Cotnoir, A.J., and D. Edwards. 2015. From Truth Pluralism to Ontological
Pluralism and Back. Journal of Philosophy 112 (3): 113–140.
68 (2): 143–149.
———. 2009. Truth-Conditions and the Nature of Truth: Re-Solving Mixed
Conjunctions. Analysis 69 (4): 684–688.
———. 2013. Truth as a Substantive Property. Australasian Journal of Philosophy
91 (2): 279–294.
———. 2014. Properties. Cambridge: Polity Press.
Eklund, M. 2011. What Are Thick Concepts? Canadian Journal of Philosophy 41
(1): 25–49.
Gamester, W. Forthcoming. Logic, Logical Form, and the Disunity of Truth.
Analysis.
Hale, B. 1994. Singular Terms. In The Philosophy of Michael Dummett, ed.
B. McGuinness and G. Oliveri, 17–44. Boston: Kluwer Academic Publishers.
Hale, B., and C.J.G. Wright. 2005. Logicism in the 21st Century. In The Oxford
Handbook of Philosophy of Mathematics and Logic, ed. Stewart Shapiro,
Haslanger, S. 2012. Resisting Reality: Social Construction and Social Critique.
Joyce, R. 2001. The Myth of Morality. Cambridge: Cambridge University Press.
Kim, S., and Nikolaj J.L.L. Pedersen. 2018. Strong Truth Pluralism. In Pluralisms
in Truth and Logic, ed. Nikolaj J.L.L. Pedersen, Jeremy Wyatt, and Nathan
Kellen. Cham: Palgrave Macmillan.
Lewis, D. 1983. New Work for a Theory of Universals. Australasian Journal of
Philosophy 61 (4): 343–377.
Lynch, M.P. 2009. Truth as One and Many. Oxford: Oxford University Press.
106 D. Edwards
———. 2013a. Three Questions About Truth. In Truth and Pluralism: Current
———. 2013b. Expressivism and Plural Truth. Philosophical Studies 163 (2):
385–401.
Mackie, J.L. 1977. Ethics: Inventing Right and Wrong. Harmondsworth: Penguin.
Pedersen, Nikolaj J.L.L., and M.P. Lynch. 2018. Truth Pluralism. In The Oxford
Pedersen, Nikolaj J.L.L., and C.D. Wright 2013. Pluralist Theories of Truth. In
The Stanford Encyclopedia of Philosophy, ed. E.N. Zalta, http://plato.stanford.
edu/archives/spr2013/entries/truth-pluralist/.
Sher, G. 2005. Functional Pluralism. Philosophical Books 46: 311–330.
Smith, M. 1994. The Moral Problem. Malden: Wiley-Blackwell.
———. 2013. On the Nature and Significance of the Distinction Between
Thick and Thin Ethical Concepts. In Thick Concepts, ed. S. Kirchin, 97–120.
Stewart-Wallace, A. 2016. An Old Solution to the Problem of Mixed Atomics.
Acta Analytica 31 (4): 363–372.
Tappolet, C. 2000. Truth, Pluralism and Many-Valued Logics. The Philosophical
Quarterly 50 (200): 382–383.
Williams, B. 1985. Ethics and the Limits of Philosophy. London: Routledge.
Wright, C.J.G. 1983. Frege’s Conception of Numbers as Objects. Vol. 2. Aberdeen:
Aberdeen University Press.
Philosophical Studies 166 (1): 225–236.
Strong Truth Pluralism
Seahwa Kim and Nikolaj J. L. L. Pedersen
Earlier versions—or parts of this chapter—have been presented by Nikolaj J. L. L. Pedersen at the
University of St. Andrews (January 2009); University College Dublin (October 2010); University
of Tokyo (October 2012); Sungkyunkwan University (December 2012); Northern Institute of
Philosophy (July 2013); Truth & Pluralism (Pacific APA 2014); University of Barcelona
(LOGOS, July 2014); Yonsei University (December 2014); University of Toronto (April 2015);
University of Connecticut (April 2015); Nanyang Technical University (August 2015); and
Lingnan University (February 2016). An earlier version of the chapter was presented by both
authors at the 1st Pluralisms Global Research Network Workshop at Yonsei University (January
2014). Thanks to the following people for helpful discussion: Dorit Bar-On, Jc Beall, Mandel
Cabrera, Ben Caplan, Colin Caret, Roy Cook, Aaron Cotnoir, Doug Edwards, Filippo Ferrari,
Tim Fuller, Sungil Han, Joe Hwang, Lina Jansson, Jinho Kang, Junyeol Kim, Sungsu Kim, Max
Kölbel, Michael Lynch, Adam Murray, Franklin Perkins, Graham Priest, Gurpreet Rattan, Sven
Rosenkranz, Stewart Shapiro, Gila Sher, Keith Simmons, Cory Wright, Crispin Wright, Jeremy
Wyatt, Byeong-Uk Yi, Andy Yu, and Elia Zardini. Research for this chapter was supported by
grant no. 2013S1A2A2035514 (Pedersen and Kim) and grant no. 2016S1A2A2911800
(Pedersen) from the National Research Foundation of Korea. We gratefully acknowledge this
support.
S. Kim
Scranton College, Ewha Womans University, Seoul, South Korea
N. J. L. L. Pedersen (*)
e-mail: nikolaj@yonsei.ac.kr

https://doi.org/10.1007/978-3-319-98346-2_5
108 S. Kim and N. J. L. L. Pedersen
1 Truth Pluralism
A survey of traditional theories of truth reveals significant differences in
the details of the various theories. Some account for truth in terms of
correspondence to reality, others in terms of coherence. Yet others favor a
pragmatist story put in terms of what is believed at the end of enquiry or
what it is useful to believe. Interestingly, despite their differences tradi-
tional theories share a very fundamental assumption: monism about
truth. Advocates of correspondence, coherence, and pragmatist theories
have traditionally approached the truth debate as a one-player game. One
theory—and one theory only—can tell us what the nature of truth is.1
Recently, the fundamental assumption of monism has been chal-
lenged. Truth pluralists hold that there are different ways of being true.
Propositions about riverbanks might be true in virtue of corresponding
to reality while propositions about the law might be true in virtue of
cohering with the body of law. According to pluralists, monists are thus
wrong in maintaining that exactly one of the traditional theories correctly
identifies a single way of being true, applicable across all truth-apt
domains of discourse. Instead two or more accounts are correct, assum-
ing that their scope is restricted to specific domains.
Truth pluralism has attracted a great deal of attention in the last few
decades, with Crispin Wright and Michael P. Lynch being the view’s two
most prominent advocates. Wright and Lynch do not stand alone, how-
ever. A steadily increasing number of truth theorists endorse, develop, or
defend some form of pluralism.2
Different versions of pluralism have emerged. Strong pluralists give up
on the idea of truth-as-such by denying that there is a single truth prop-
erty applicable across all truth-apt domains of discourse. Truth is many,
not one. Moderate pluralists, on the other hand, hold on to the idea of
truth-as-such. There is a single generic truth property that applies across
all truth-apt discourse. Truth is one. However, at the same time truth is
many because propositions belonging to different domains may possess
the generic truth property in virtue of having distinct properties such as
correspondence or coherence.
This chapter presents and develops a form of strong pluralism. We
spell this idea out in a strongly reductionist fashion (to be given a slightly
more refined formulation in section more refined formulation in Sect. 3):
Strong Truth Pluralism 109
(TR) There is a plurality of properties T1, …, Tn such that (i) for every
proposition Φ, Φ’s being true is identical to Φ’s being Ti
(1 ≤ i ≤ n), and (ii) there is no property among T1, …, Tn that
satisfies (i) for every Φ.
(TR) is the token-token identity theory in an alethic incarnation. For

example, the truth of the proposition that there are mountains simply is
this proposition’s corresponding to reality, the truth of the proposition
that speeding is illegal simply is this proposition’s cohering with the body
of law, and neither correspondence nor coherence reduces truth for every
proposition.3
Given our endorsement of a reductionist, identity-theoretic form of
pluralism we refer to T1, …, Tn as “truth-reducing properties” or “truth
properties”.
This chapter has two aims. The first aim is to present our favored ver-
sion of strong pluralism in greater detail. In pursuing this aim we go
against the current that has so far set the course of the pluralism debate.
Strong pluralism has been somewhat neglected in the literature. It has
been widely regarded as a non-starter due to a battery of seemingly dev-
astating objections leveled against it.4 Among these objections the prob-
lem of mixed compounds is regarded as being particularly pressing—and
difficult—for the strong pluralist to deal with. The second aim of the
chapter is to give a strongly pluralist response to the problem of mixed
compounds.
We proceed as follows: in Sect. 2 we present the problem of mixed
compounds. In Sects. 3 and 4 we present two complementary metaphysi-
cal parts of our view—a specification of the range of truth-reducing
properties we endorse and an account of truth grounding. We rely on this
work in Sect. 5 where we respond to the problem of mixed compounds.
2 The Problem of Mixed Compounds

Some discourse is pure in that it concerns only one domain. 〈Mt. Everest
is extended in space and there are trees〉 is an empirical proposition, and
〈2 + 2 = 4 and 3 + 6 = 9〉 is a mathematical proposition. (Angle brackets
are used to represent propositions.) However, not all discourse is pure.
Some discourse concerns more than one domain. Consider, for example,
the following compound proposition:
( mix - Ù ) Mt.Everest is extended in space and Bob’s drunk driving is illegal.
This is a mixed conjunction. It pertains to more than one domain, the
empirical domain and the legal domain.
The problem of mixed compounds emerges when we consider the
issue of what to say about the truth of mixed compounds.5 Clearly (mix-
∧) is true—its conjuncts, after all, are both true. However, it is less clear
how exactly the strong pluralist is going to account for this. According to
our brand of strong pluralism, each conjunct belongs to a specific domain
and within that domain truth is (token-)identical to a certain property.6
Let us suppose that the truth property for the first conjunct is correspon-
dence, and that it is coherence with the body of law for the second.
However, what truth property does the conjunction itself have? Given
our favored form of pluralism this amounts to the question of which
property is identical to truth of the conjunction. Let us consider two
arguably implausible options.
One option is to say that the conjunction is correspondence-true. (We
sometimes use the label “F-true” or “F-truth” to signify that F is the
truth-reducing property for a given proposition.) This is implausible
because it completely neglects the contribution made by the second con-
junct—a conjunct that is coherence-true. Another option is to say that the
conjunction is coherence-true. However, this is implausible for the same
kind of reason: it would completely neglect the contribution made by
one of the conjuncts.
A third option is to say that the conjunction possesses some property
distinct from both correspondence and coherence. We favor a response of
this kind. However, we recognize that anyone who does so faces two
tasks:
1. Something must be said about the nature of the third property. What
property is it? Also, while it is distinct from the truth properties of
both of its conjuncts, does its nature or instances somehow depend on
them, and if so, how?
2. Since the strong pluralist rejects the idea that there is a truth property
that applies across all truth-apt domains, an assurance must be
provided that the property possessed by the compound is not such a
property.
In Sect. 5 we offer a solution that addresses both of these tasks.
3 omains, Truth Properties, and Truth

D
Grounding
The core idea behind both moderate and strong pluralism is that proposi-
tions belonging to different domains may be true in different ways.
Domains are central to both forms of pluralism, as the domain member-
ship of a proposition is a key factor in determining the way in which it is
true. On our view logical form is another. In this section we introduce an
account of domains and then offer an account of how the domain mem-
bership and logical form of a proposition determine which property
reduces truth for that proposition. Once this account is in place, a
response to the problem of mixed compounds naturally offers itself.
Propositions and Domains
Despite the central role that domains play within the standard pluralist
framework not much systematic work has been done on their nature. We
hope to do at least a little better by outlining a systematic treatment of
domains.7
We take the range of propositions that pluralists are interested in (and
typically deal with) to be the following:
Prop:
(a) atomic propositions, and

(b) complex propositions whose only constituents are atomics and truth-
functional logical operators (i.e. those represented by sentential
connectives).
Let us account for the domain membership of propositions in Prop.

Domains are sets of propositions individuated by their subject matter.
This idea seems to line up with how pluralists think about domains.
〈2 + 3 = 5〉, 〈Mt. Everest is extended in space〉, and 〈Bob’s drunk driving
is illegal〉 belong to different domains. Why? Because they concern differ-
ent subject matters or are about different kinds of states of affairs.
Domains constitute a hierarchy of sorts. There is a range of base-level
domains (D1, …, Dn). These domains are the domains that pluralists talk
about when they introduce their view—that is, the mathematical domain,
empirical domain, ethical domain, legal domain, and so on. Atomic
propositions belong to exactly one base-level domain.8 Other domains
are characterized in terms of base-level domains and set-theoretic union.
The domain of a disjunction or a conjunction is the union of the domains
of its constituents. The domain of a negated proposition is the same as
the proposition negated. That is:
(DA) For atomic Φ: D(Φ) = D(~Φ) = Di (where Di is one of the

base - level domains D1, …, Dn)
(D∨) D(Φj ∨ … ∨ Φj + m) = D(~(Φj ∨ … ∨ Φj + m)) = D(Φj) ∪ …
∪ D(Φj + m)
(D∧) D(Φj ∧ … ∧ Φj + m) = D(~(Φj ∧ … ∧ Φj + m)) = D(Φj) ∪ …
∪ D(Φj + m)
A mixed compound belongs to a domain that is distinct from the domain

of its constituents. However, pure compounds and their negations belong
to the same domain as their constituents. That is:
( D Ú Ù - pure ) If D (F j ) = ¼ = D (F j + m ) = Di , then

D (F j ∨…∨ F j+m ) = D (~ (F j ∨…∨ F j+m ))
= D (F j ∧…∧ F j+m ) = D (~ (F j ∧…∧ F j+m ))
= D (F j ) ∪…∪ D (F j+m ) = Di ∪…∪ Di

= Di
As a result, some atomic propositions and compounds belong to the

same domain. For example, 〈3 < 8〉 belongs to the mathematical domain
as does 〈22 = 4 and √4 = 2〉, 〈22 = 4 or √4 = 2〉, and any other com-
pound with only mathematical constituents.
Domains proliferate, and they do so along two dimensions. First, let-
ting n be the number of base-level domains, domains proliferate into
ones involving k base-level domains for every k ≤ n. For example, if there
are three base-level domains, domains proliferate into 1-domains,
2-domains, and 3-domains—where “k-domain” signifies a domain
involving exactly k base-level domains. Second, k-domains proliferate.
Any two k-domains that differ with respect to at least one base-level
domain are distinct. For example, D1 ∪ D2 is distinct from D1 ∪ D3 as
well as from D2 ∪ D3 (and D1 ∪ D3 from D2 ∪ D3).
How many domains are there? Answer: 2n – 1 (where, again, n is the
number of base-level domains). Thus, if there are three base-level domains,
there is a total of 7 domains: three 1-domains (D1, D2, D3), three
2-domains (D1 ∪ D2, D1 ∪ D3, D2 ∪ D3), and one 3-domain (D1 ∪ D2 ∪
D3).
Propositions and Truth Properties
We provided a general statement of our reductionist, identity-theoretic

form of pluralism in the opening section—in the form of (TR). Earlier
we specified Prop as the range of propositions that pluralists are con-
cerned with. Let us restate our general thesis, factoring in the focus on
Prop:
(T) There is a plurality of truth or truth-reducing properties T1, …, Tn

such that (i) for every proposition Φ, Φ’s being true is identical to
Φ’s being Ti (1 ≤ i ≤ n), and (ii) there is no property among T1,
…, Tn that satisfies (i) for every proposition Φ.
What properties are in the plurality of truth properties T1, …, Tn? There
is a number of base-level truth properties Tj, …, Tj + m. These include the
“standard” pluralist properties such as correspondence, coherence,
superassertibility, and so on. We refer to these properties as “base-level

truth properties” because they reduce truth for atomics:
(TA) For atomic Φ, Φ’s being true is identical to Φ’s being Tk, where Tk
is the base-level truth property that reduces truth for atomic prop-
ositions in Φ’s domain.
To illustrate consider the atomic propositions 〈Mt. Everest is extended in

space〉 and 〈Bob’s drunk driving is illegal〉. They belong respectively to
the empirical domain and the legal domain. We can suppose that their
domain membership determines respectively correspondence and coher-
ence with the body of law as the relevant base-level truth properties.
In addition to base-level truth properties there are compound-specific
truth properties. For disjunctions we endorse the following truth
property:
Φi ∨ … ∨ Φi + n’s being true is identical to Φi ∨ … ∨ Φi + n’s being

(T∨)
a disjunction with at least one disjunct that has its truth property
(as determined by its domain and logical form).
Let “T∨” label the property characterized by (T∨).

(T∨) gives us the following individually necessary and jointly sufficient
conditions for a proposition to be T∨:
(T∨*) Proposition Γ is T∨ if and only if (i) Γ has the form Φi ∨ … ∨ Φi

+ n (i.e. a disjunction), and (ii) at least one of Φi, …, Φi + n has its
truth property (as determined by its domain and logical form).
For conjunctions we endorse the following truth property:
Φi ∧ … ∧ Φi + n’s being true is identical to the property of being

(T∧)
a conjunction with conjuncts that all have their truth property
(as determined by their respective domains and logical forms).
Let “T∧” label the property characterized by (T∧).

(T∧) gives us the following individually necessary and jointly sufficient

conditions for a proposition to be T∧:
(T∧*) Proposition Γ is T∧ if and only if (i) Γ has the form Φi ∧ … ∧ Φi

+ n (i.e. a conjunction), and (ii) Φi, …, Φi + n all have their truth
property (as determined by their respective domains and logical
forms).
Why endorse the compound-specific properties T∨ and T∧? Our motiva-

tion for doing so is to take seriously the idea that disjunctions and con-
junctions are truth-functional compounds. The truth of a disjunction is a
function of its being a disjunction and the fact that at least one of its
disjuncts has its truth property. The truth of a conjunction is a function
of its being a conjunction and the fact that all of its conjuncts have their
truth property. What matters in each case are logical form and the fact
that some constituent or all constituents have the relevant truth property.
It does not matter what specific property it is or what domain the con-
stituents belong to.
Hyperintensionality
Before we move on to talk about truth grounding it is worth highlighting

an interesting feature of our view: a commitment to a hyperintensional
treatment of propositions.
Some propositions belong to the same domain but have different
truth-reducing properties. This is the case for any atomic proposition
belonging to domain Di and any compound whose constituents all
belong to that same domain. For example, consider 〈Mt. Everest is
extended in space〉, 〈Mt. Everest is extended in space or there are trees〉,
and 〈Mt. Everest is extended in space and there are trees〉. These proposi-
tions all belong to the empirical domain. However, while correspondence
to reality is the truth-reducing property for 〈Mt. Everest is extended in
space〉, T∨ reduces truth for 〈Mt. Everest is extended in space or there are
trees〉 and T∧ does so for 〈Mt. Everest is extended in space and there are
trees〉. This difference is a reflection of the fact that, on our view, which
property reduces truth for a given proposition is not determined by the

proposition’s domain membership alone. Logical form plays a role, too.
We are committed to a hyperintensional treatment of propositions. By
this we mean that we are committed to taking some intensionally
equivalent propositions to be distinct propositions. Here we take two
propositions p and q to be intensionally equivalent if and only if: for any
world w, p is true if and only if q is true. Within our strongly pluralist
framework this amounts to the following: propositions p and q are inten-
sionally equivalent if and only if: for any world w, p has its truth-reducing
property if and only if q has its truth-reducing property.
Consider an atomic proposition p and the compound proposition p ∧
p. Let p belong to domain Di and let Ti be the base-level truth property
for atomic Di-propositions. Recall that T∧ is the truth property for any
conjunction, including p ∧ p. The two propositions in question, that is,
p and p ∧ p, are intensionally equivalent. In any world w, p has its truth-
reducing property Ti if and only if p ∧ p has its truth-reducing property,
T∧. For, whenever p is Ti, it is also the case that p ∧ p is a conjunction
whose conjuncts have their truth-reducing property (namely, Ti).
Conversely, whenever p ∧ p is T∧, it is also the case that p is Ti.
Now consider the following list of statements concerning propositions
p and p ∧ p and truth-reducing properties:
(p1) Ti is the truth-reducing property for p.

(p2) T∧ is the truth-reducing property for p.
(p ∧ p1) Ti is the truth-reducing property for p ∧ p.
(p ∧ p2) T∧ is the truth-reducing property for p ∧ p.
(p1) is true while (p2) is false, and (p ∧ p1) is false while (p ∧ p2) is true.
This shows that, although p and p ∧ p are intensionally equivalent, they
cannot be the same proposition. For, what would happen if they were?
This would clash with our observations concerning (p1), (p2), (p ∧ p1),
and (p ∧ p2)—observations that are non-negotiable commitments of our
view.
In light of these considerations we cannot adopt any account that
treats p and p ∧ p as having the same content or as being the same propo-
sition. Thus, for example, we cannot adopt a standard possible-worlds
account of propositions. According to this account a proposition is iden-

tified with the set of possible worlds at which it holds.9 This would make
p and p ∧ p the same proposition.
We are committed to a hyperintensional treatment of propositions
given features that are specific to our favored version of strong truth plu-
ralism. We believe that a more general argument can be given—one that
establishes the same commitment for truth pluralism in general, whether
strong or moderate. However, we also think that this more general argu-
ment requires more work to cash out properly. For this reason, we leave
it for another day.
The issue of hyperintensionality has gone unnoticed in the literature.
However, it strikes us as interesting and significant. Our argument shows
that distinctively pluralist commitments impose a significant constraint
on the pluralist’s theory of propositions and any account they might offer
concerning their nature. This is worth noting.10
4 Truth Grounding
On the picture presented above, correspondence, coherence, superwar-
rant, and so on are truth-reducing properties only for atomic proposi-
tions. T∨ and T∧ are truth-reducing properties only for compounds of a
particular logical form, that is, disjunctions and conjunctions,
respectively.
We now turn to the issue of how compound truth is grounded. Our
account of truth grounding is complementary to our account of truth
reduction. Both accounts are needed for a comprehensive metaphysics of
truth. The issue of truth reduction for a given compound is the issue of
specifying a property to which the truth of that compound is identical.
The issue of truth grounding for a given compound is the issue of map-
ping how, or in what ways, the truth of the compound depends on the
semantic status of its constituents. Given the characterization of T∨, if p
∨ q is T∨, this is so because at least one disjunct has its truth-reducing
property. But this is to say that the instantiation of the truth-reducing
property of a disjunct suffices to ground the instantiation of T∨, that is,
disjunction-truth. Given the characterization of T∧, if p ∧ q is T∧, this is
so because every conjunct has its truth-reducing property. But this is to

say that the conjuncts’ having their respective truth-reducing properties
jointly grounds conjunction-truth. We also want to say that these because
claims cannot be reversed, that is, that truth grounding is asymmetric.
Grounding is standardly taken to be (strongly) asymmetric, (strongly)
irreflexive, and (strongly) transitive, that is,
( SA ) If éë p ùû ¬ éëq ùû, G then not : éëq ùû ¬ éë p ùû, D ( Strong asymmetry )

( SI ) Not : éë p ùû ¬ éë p ùû, G ( Strong irreflexivity )

( ST ) If éë p ùû ¬ éëq ùû, G and éëq ùû ¬ D, then éë p ùû ¬ G , D ( Strong transitivity )
“[p]” is read as the fact that p while “[p] ← [q]” is read as [p] is grounded
by [q] (or [q] grounds [p]). Upper-case Greek letters denote (possibly
empty) sets of facts. Strong asymmetry says that it is not the case that [p]
is among the grounds of [q], provided that [q] grounds [p]. Strong irre-
flexitivity says that no fact is among its own grounds, and transitivity says
that if [q], Γ ground [p] and ∆ grounds [q], then Γ, ∆ ground [p].11
These general observations about truth grounding and grounding’s
features can be used to shed light on the grounds of truth in relation to
specific compounds.
Consider the following five compounds (parentheses used to indicate
scope):
(∨1) Mt. Everest is extended in space or Bob’s drunk driving is illegal.

(∨2) Mt. Everest is extended in space or Bob’s drunk driving is legal.
(∨3) (Mt. Everest is extended in space and Bob’s drunk driving is ille-
gal) or 2 + 2 = 5.
(∧1) Mt. Everest is extended in space and Bob’s drunk driving is
illegal.
(∧2) (Mt. Everest is extended in space or 2 + 2 = 5) and (Bob’s drunk
driving is illegal or 3 + 8 = 12).
Given (T∨*) we can say something about the metaphysical ground of the
truth of (∨1) and (∨2). As earlier, suppose that correspondence and
coherence with the body of law are the truth-reducing properties for
atomic propositions belonging respectively to the empirical domain and

the legal domain. Both of (∨1)’s disjuncts have their truth-reducing
property, correspondence and coherence, respectively. The truth of (∨1)
is thus doubly grounded. It is grounded in the first disjunct’s correspond-
ing to reality and also in the second disjunct’s cohering with the body of
law. (∨2), on the other hand, has just one disjunct that possesses its truth-
reducing property. The truth of (∨2) is thus grounded solely in its first
disjunct’s corresponding to reality.
Given (T∧*) we can say something about the metaphysical ground of
the conjunction-truth of (∧1). Both of (∧1)’s conjuncts have their
truth-reducing property, correspondence and coherence, respectively.
Jointly these facts ground the truth of (∧1). (∧1) is a conjunction whose
conjuncts all have their truth-reducing property because the first con-
junct corresponds to reality and the second conjunct coheres with the
body of law.
(∨1), (∨2), and (∧1) are less complex than (∨3) and (∧2). (∨3) and
(∧2) are compounds with constituents that are themselves com-
pounds. Transitivity helps us shed light on truth grounding for such
compounds.
The truth of (∨3), like the truth of any other disjunction, is identical
to its being a disjunction with at least one disjunct that has its truth-
reducing property. There is precisely one such disjunct, namely 〈Mt.
Everest is extended in space and Bob’s drunk driving is illegal〉. This dis-
junct is itself a compound—a conjunction. This conjunction’s being true
is identical to its being a conjunction with conjuncts that all possess their
truth-reducing property. The truth of the conjuncts is identical, respec-
tively, to correspondence to reality and coherence with the body of law.
How about grounding? We have the following instances of ground-
ing: (∨3)’s being a disjunction with at least one disjunct that has its
truth-reducing property is grounded in its first disjunct, 〈Mt. Everest is
extended in space and Bob’s drunk driving is illegal〉’s being a conjunc-
tion with conjuncts that all possess their truth-reducing property. In
turn this conjunction’s being a conjunction with conjuncts that all pos-
sess their truth-reducing property is grounded jointly in the first con-
junct, 〈Mt. Everest is extended in space〉’s corresponding to reality and
the second conjunct, 〈Bob’s drunk driving is illegal〉’s cohering with the
body of law.
Applying transitivity to (∨3) we get that the truth of 〈(Mt. Everest is
extended in space and Bob’s drunk driving is illegal) or 2 + 2 = 5〉 is
grounded in 〈Mt. Everest is extended in space〉’s corresponding to reality
and 〈Bob’s drunk driving is illegal〉’s cohering with the body of law.
The truth of (∧2), like the truth of any other conjunction, is identical
to its being a conjunction with conjuncts that all possess their truth-
reducing property. (∧2) has two conjuncts, each of which is a disjunc-
tion. Their truth reduces to being a disjunction with at least one disjunct
that has its truth-reducing property. For 〈Mt. Everest is extended in space
or 2 + 2 = 5〉 there is exactly one such disjunct—the first one. This dis-
junct has its truth-reducing property, that is, 〈Mt. Everest is extended in
space〉 corresponds to reality. For 〈Bob’s drunk driving is illegal or
3 + 8 = 12〉 there is also precisely one disjunct that has its truth-reducing
property—again, the first one. This disjunct has its truth-reducing prop-
erty, that is, 〈Bob’s drunk driving is illegal〉 coheres with the body of law.
How about grounding? We have the following instances of grounding:
(∧2)’s being a conjunction with conjuncts that all have their truth-
reducing property is grounded jointly in 〈Mt. Everest is extended in
space or 2 + 2 = 5〉’s being a disjunction with at least one disjunct that has
its truth-reducing property and 〈Bob’s drunk driving is illegal or
3 + 8 = 12〉’s being a disjunction with at least one disjunct that has its
truth-reducing property. 〈Mt. Everest is extended in space or 2 + 2 = 5〉’s
being a disjunction with at least one disjunct that has its truth-reducing
property is grounded in 〈Mt. Everest is extended in space〉’s correspond-
ing to reality. 〈Bob’s drunk driving is illegal or 3 + 8 = 12〉’s being a dis-
junction with at least one disjunct that has its truth-reducing property is
grounded in 〈Bob’s drunk driving is illegal〉’s cohering with the body of
law. Applying transitivity we get that (∧2)’s being a conjunction with
conjuncts that all have their truth-reducing property is grounded in 〈Mt.
Everest is extended in space〉’s corresponding to reality and 〈Bob’s drunk
driving is illegal〉’s cohering with the body of law.
The resulting picture of truth grounding is this: the truth of any given
atomic is identical to its truth-reducing property (correspondence,
coherence, etc.). The truth of any compound is identical to its
compound-specific, truth-reducing property. However, a compound’s

having this property is ultimately grounded by facts about the instantia-
tion of base-level truth properties.12
5 Strongly Pluralist Response

A
the Problem of Mixed Compounds
The problem of mixed compounds raised the following issues for the
strong pluralist:
1. Something must be said about the nature of the truth property pos-
sessed by mixed compounds. In particular, while it is distinct from the
truth properties possessed by its constituents, does its nature or
instances somehow dependent on them—and if so, how?
2. Since the strong pluralist rejects the idea that there is a truth property
that applies across all truth-apt domains, we need some assurance that
the truth property possessed by neither disjunctions nor conjunctions
is not a property of this kind.
By using the strongly pluralist framework introduced above we can

address both of these issues. Turning to the first issue, what we have is the
following: a disjunction’s being true is simply identical to its being a dis-
junction with at least one disjunct that has its truth-reducing property. A
conjunction’s being true is simply identical to its being a conjunction
with conjuncts that all have their truth-reducing property.
This account of the nature of disjunction-truth and conjunction-truth
packs logical structure into both properties. This has the advantage of
making clear how the truth of a compound depends on the semantic
status of its constituents.
Now, there is nothing extraordinary or mysterious about the prop-
erty that reduces truth for mixed disjunctions. It is the same as the
property that reduces truth for pure disjunctions—namely, disjunc-
tion-truth. Likewise, there is nothing extraordinary or mysterious about
the property that reduces truth for mixed conjunctions. It is the same
as the property that reduces truth for pure conjunctions—namely, con-
junction-truth. In sum, from the point of view of truth reduction,

mixed and pure compounds are on a par. Problem of mixed compounds
solved.
As for grounding, mixed disjunctions may have separate, sufficient
grounds for their (disjunction-)truth pertaining to distinct domains. On
the other hand, if pure disjunctions have separate sufficient grounds for
their (disjunction-)truth, these grounds always belong to the same
domain. Consider
(PURE-∨) Mt. Everest is extended in space or London Bridge is

extended in space.
(MIX-∨) Mt. Everest is extended in space or Bob’s drunk driving is
illegal.
The truth of (PURE-∨) has two distinct, sufficient grounds pertaining to

the empirical domain. 〈Mt. Everest is extended in space〉’s corresponding
to reality is one, 〈London Bridge is extended in space〉’s corresponding to
reality another. The truth of (MIX-∨), on the other hand, has two dis-
tinct, sufficient grounds that belong to different domains—the empirical
domain and the legal domain, respectively.13
Mixed conjunctions have a sufficient ground that involves two distinct
domains whereas pure conjunctions have a sufficient ground that involves
just one domain. Consider:
(PURE-∧) Mt. Everest is extended in space and London Bridge is

extended in space.
(MIX-∧) Mt. Everest is extended in space and Bob’s drunk driving is
illegal.
The (conjunction-)truth of (PURE-∧) has a sufficient ground involving

just the empirical domain, namely 〈Mt. Everest is extended in space〉’s cor-
responding to reality and 〈London Bridge is extended in space〉’s doing so
as well. The (conjunction-)truth of (MIX-∧), on the other hand, has a suf-
ficient ground involving the empirical domain and the legal domain,
namely 〈Mt. Everest is extended in space〉’s corresponding to reality and
〈Bob’s drunk driving is being illegal〉’s cohering with the body of law.
Turn to the second issue, the issue of giving an assurance that the
truth-reducing properties possessed by mixed disjunctions and mixed
conjunctions do not apply across the board. This is straightforward. We
just need to make two observations. The first observation is that the
truth-reducing property for disjunctions, whether pure or mixed, is the
property of being a disjunction with at least one disjunct that has its
truth-reducing property. This is a property that only disjunctions can
have, and so, the truth-reducing property for mixed disjunctions is not a
generic truth property. The second observation is that the truth-reducing
property for conjunctions, whether pure or mixed, is the property of
being a conjunction with conjuncts that all have their truth-reducing
property. This is a property that only conjunctions can have, and so, the
truth-reducing property for mixed conjunctions is not a generic truth
property.
6 Conclusion
We set ourselves two aims in this chapter. The first was to present and
develop a version of strong truth pluralism. The second was to give a
strongly pluralist solution to the problem of mixed compounds. We
have achieved both aims. In Sects. 3 and 4 we introduced and devel-
oped a version of strong pluralism, an alethic incarnation of the (token-
token) identity theory. The metaphysics of the view was presented in
some detail, with a specification of its range of truth-reducing proper-
ties and an account of truth grounding. We introduced compound-
specific truth-reducing properties—identifying the truth of disjunctions
with being a disjunction with at least one disjunct that has its truth-
reducing property and the truth of conjunctions with the being a con-
junction with conjuncts that all have their truth-reducing property.
The nature of these properties put us in a position to offer a straightfor-
ward solution to the problem of mixed compounds. We thus conclude
that whatever insurmountable problems or challenges might seem to
confront strong pluralism, the problem of mixed compounds is not
one of them.
Notes
1. Correspondence theorists include David 1994, Devitt 1984, Newman
2007, Rasmussen 2014, Russell 1912, Vision 2004, and Wittgenstein
1921. Coherence theorists include Bradley 1914, Rescher 1973, Walker
1989, and Young 2001. Pragmatists or neo-pragmatists include James
1907, 1909, Peirce 1878, and Putnam 1981.
2. For example, Beall 2000, 2013; Cook 2011; Cotnoir 2009, 2013a, b;
Cotnoir and Edwards 2015; Edwards 2008, 2009, 2011, 2012, 2013,
2018; Gamester forthcoming; Kölbel 2008, 2013; Pedersen 2006, 2010,
2012a, b, 2014, ms-a, ms-b, ms-c; Pedersen and Edwards 2011; Pedersen
and Wright 2013a, b, 2016; Wyatt 2013; Yu 2017. Gila Sher and Terence
Horgan and various collaborators have developed a pluralist version of
the correspondence theory that incorporates different ways of corre-
sponding. See Sher 2005, 2013, 2016 and Horgan 2001; Barnard and
Horgan 2006, 2013. Works by Lynch and Wright include Lynch 2001,
2004, 2006, 2009, and 2013 and Wright 1992, 1994, 1995, 1996,
1998, 2001, and 2013.
3. Versions of strong pluralism are presented in Cotnoir 2009, 2013a;
Gamester forthcoming; Pedersen 2006, ms-a, ms-b, ms-c, Pedersen and
Lynch 2018 (Sect. 20.6, due to Pedersen).
4. See, for example, Lynch 2009. For surveys that provide a systematic pre-
sentation of objections to pluralism, see Pedersen 2012b, Pedersen and
Lynch 2018, Pedersen and Wright 2016.
5. The problem of mixed compounds is raised by Tappolet 2000, Sainsbury
1996, and Williamson 1994. Various pluralist options are discussed by
Cook 2011, Cotnoir 2009, Edwards 2008, 2009, 2018, Gamester forth-
coming, Lynch 2004, 2006, 2009, 2013, Pedersen 2012 b, Pedersen and
Lynch 2018, Sher 2005, 2013, and Wright 2013.
6. It is quite tedious always to use formulations that make our endorse-
ment of a token-token version of the identity theory explicit and dis-
tinguish it from its type-type counterparts. Sometimes we use
formulations that are compatible with the type-type identity theory.
However, in those cases it should be borne in mind that our reduction-
ism kicks in at the level of tokens. One reason to opt for token identity
is that it seems to integrate quite naturally with our account of truth
grounding, that is the other component of our proposed metaphysics
of truth. Grounding is usually regarded as a relation that obtains
between particular facts or states of affairs rather than types of facts or

states of affairs.
7. Wyatt (2013) and Edwards (2018), (this volume) are among the few
sources that offer a systematic discussion of domains. Lynch (2009) talks
about domains but only in passing. The formal aspects of what we say
about domains here align with Yu (2017). For a more comprehensive
treatment of domains and their relation to subject matter, see Pedersen
(ms-e).
8. The idea that atomics belong to a unique domain has been debated in
the literature. Discussion of this issue is often tied to the so-called prob-
lem of mixed atomics. Following Sher (2005), it is not clear, for example,
which unique domain 〈Causing pain is bad〉 would belong to. It involves
a mental concept (pain), a physical concept (causation), and a moral con-
cept (badness). Our view is that 〈Causing pain is bad〉 belongs to the
ethical domain. This is because the property figuring in an atomic prop-
osition determines its domain membership. Since the property of being
bad is an ethical property, 〈Causing pain is bad〉 belongs to the ethical
domain. This kind of approach is suggested in Pedersen and Wright
(2016, Sect. 4.5.1) and spelled out in more detail in Pedersen (ms-d).
Edwards (2018) favors the same approach. Wyatt (2013) develops an
account of domains according to which atomics can belong to more
than one domain. He supplements this proposal with a story about how
to determine a single truth-relevant property for propositions belonging
to more than one domain. This is not the place to pursue an in-depth
discussion of different approaches to mixed atomics and domains. For
present purposes, we rest content with simply having noted some key
differences.
9. Lewis 1986.
10. Thanks to Tim Fuller for raising the issue of hyperintensionality. Thanks
to Jeremy Wyatt for extensive comments also. Substitutability salva veri-
tate is often employed as a test for hyperintensionality. A notion or opera-
tor N on sentences is said to be hyperintensional if it does not allow
intensionally equivalent sentences to be substituted salva veritate. The
notion “propositional content of ” is hyperintensional in this sense. To see
this consider sentences “p” and “p ∧ p” whose propositional content is
respectively the proposition 〈p〉 and the proposition 〈p ∧ p〉. Although
“p” and “p ∧ p” are intensionally equivalent, “p ∧ p” cannot be substi-
tuted salva veritate for “p” in “The propositional content of ‘p’ is 〈p〉”.
This would clash with (p1), (p2), (p ∧ p1), and (p ∧ p2). Similarly, “p”
cannot be substituted salva veritate for “p ∧ p” in “The propositional

content of ‘p ∧ p’ is 〈p ∧ p〉”.
11. Rosen 2010.
12. The kind of grounding-theoretic account just provided is available to
both strong and moderate pluralists. It should be noted, however, that
there are significant differences with respect to the specifics of truth
grounding within the frameworks of respectively strong and moderate
pluralism. Crucially, within the framework of strong pluralism there is
no grounding relationship between an atomic proposition’s correspond-
ing (cohering, etc.) and its being true. For, remember an atomic proposi-
tion’s being true simply is its corresponding (cohering, etc.). Given this
identity taking a proposition’s corresponding (etc.) to ground its own
truth would violate irreflexitivity. However, grounding relations do
obtain between atomic propositions’ having their base-level truth prop-
erties and instances of disjunction-truth and conjunction-truth. For
details of a grounding-theoretic metaphysics for moderate pluralism, see
Pedersen ms-a, ms-b, ms-c and Kim and Pedersen (ms).
13. We said, “if pure disjunctions have separate sufficient grounds for their
(disjunction-)truth, these grounds always pertain to the same domain”.
Strictly speaking, this claim needs to be qualified. It is only correct if we
restrict attention to pure disjunctions with atomic constituents. To see
this, note that 〈(Bob’s drunk driving is legal or Mt. Everest is extended
in space) or (Bob’s drunk driving is illegal or there are no trees)〉 is a pure
disjunction. Its disjuncts concern the same domain—namely, the legal-
empirical domain. However, (by transitivity) the pure disjunction in
question has separate grounds for its truth in 〈Mt. Everest is extended in
space〉’s corresponding to reality and in 〈Bob’s drunk driving is illegal〉’s
cohering with the body of law. These grounds pertain to different
domains.
References
Barnard, R., and T. Horgan. 2006. Truth as Mediated Correspondence. The
Monist 89: 31–50.
———. 2013. The Synthetic Unity of Truth. In Truth and Pluralism: Current
Debates, ed. Nikolaj J.L.L. Pedersen and C.D. Wright, 180–196. Oxford:
Bradley, F. 1914. Essays on Truth and Reality. Oxford: Clarendon Press.

Cook, R.T. 2011. Alethic Pluralism, Generic Truth and Mixed Conjunctions.
Analysis 69: 473–479.
———. 2013a. Validity for Strong Pluralists. Philosophy and Phenomenological
Research 83: 563–579.
———. 2013b. Pluralism and paradox. In Pedersen & Wright 2013a, 339–350.
Cotnoir, A., and D. Edwards. 2015. From Truth Pluralism to Ontological
Pluralism and Back. The Journal of Philosophy 112: 113–140.
David, M. 1994. Correspondence and Disquotation: An Essay on the Nature of
Truth. Oxford: Oxford University Press.
Devitt, M. 1984. Realism and Truth. Oxford: Blackwell.
68: 142–149.
———. 2009. Truth-conditions and the Nature of Truth: Re-Solving Mixed
Conjunctions. Analysis 69: 684–688.
———. 2011. Simplifying Alethic Pluralism. Southern Journal of Philosophy 49:
28–48.
———. 2012. On Alethic Disjunctivism. Dialectica 66: 200–214.
———. 2013. Truth, Winning, and Simple Determination Pluralism. In Truth
and Pluralism: Current Debates, ed. Nikolaj J.L.L. Pedersen and C.D. Wright,
Gamester, W. forthcoming. Logic, Logical Form and the Disunity of Truth. To
appear in Analysis. doi: https://doi.org/10.1093/analys/anx165.
James, W. 1907. Pragmatism: A New Name for some Old Ways of Thinking.
———. 1909. The Meaning of Truth. Cambridge, MA: Harvard University
Press.
Kim, S. & Nikolaj J.L.L. Pedersen. (ms). The Return of the Many: A Critical
Appraisal of Moderate Truth Pluralism Through Metaphysics.
Research 77: 359–384.
———. 2013. Should We Be Pluralists About Truth? In Pedersen & Wright
2013a, 278–297.
Lewis, D. 1986. On the Plurality of Worlds. Oxford: Blackwell.
Lynch, M. 2001. A Functionalist Theory of Truth. In The Nature of Truth: Classic

and Contemporary Perspectives, ed. M. Lynch, 723–749. Cambridge, MA:
MIT Press.
———. 2004. Truth and Multiple Realizability. Australasian Journal of
———. 2006. ReWrighting Pluralism. The Monist 89: 63–84.
———. 2013. Three Questions for Alethic Pluralism. In Pedersen & Wright
2013a, 21–41.
Newman, A. 2007. The Correspondence Theory of Truth: An Essay on the
Metaphysics of Predication. Cambridge: Cambridge University Press.
Teach Us About Alethic Pluralism? The Monist 89: 103–117.
———. 2010. Stabilizing Alethic Pluralism. The Philosophical Quarterly 60:
92–108.
———. 2012a. True Alethic Functionalism? International Journal of Philosophical
Studies 20: 125–133.
———. 2012b. Recent Work on Alethic Pluralism. Analysis 72: 588–607.
259–277.
———. (ms-a). Grounding Manifestation Pluralism.
———. (ms-b). Grounding Determination Pluralism.
———. (ms-c). Moderate Truth Pluralism and the Structure of Doxastic
Normativity.
———. (ms-d). Moderate Pluralism About Truth and Logic: Truth and Logic
as One, Quasi-truth and Quasi-logic as Many.
———. (ms-e). Subject Matter and Domains.
Pedersen, Nikolaj J.L.L., and D. Edwards. 2011. Truth as One(s) and Many: On
Lynch’s Alethic Functionalism. Analytic Philosophy 52: 213–230.
Pedersen, Nikolaj J.L.L., and C.D. Wright, eds. 2013a. Truth and Pluralism:
Current Debates. New York: Oxford University Press.
———. 2013b. Pluralism About Truth as Alethic Disjunctivism. In Pedersen &
Wright 2013a, 87–112.
———. 2016. Pluralist Theories of Truth. In The Stanford Encyclopedia of
Philosophy, ed. E.N. Zalta, (Spring 2016 Edition). https://plato.stanford.edu/
archives/spr2016/entries/truth-pluralist/
Peirce, C.S. 1878/1992. How to Make Our Ideas Clear. Reprinted in e Essential
Peirce, ed. N. Houser and C. Kloesel, 124–141, Vol. 1. Bloomington: Indiana
University Press.
Putnam, H. 1981. Reason, Truth and History. Cambridge: Cambridge University
Press.
Rasmussen, J. 2014. Defending the Correspondence Theory of Truth. Cambridge:
Rescher, N. 1973. The Coherence Theory of Truth. Oxford: Oxford University
Press.
Russell, B. 1912. Problems of Philosophy. Reprinted by Oxford University Press,
1971.
Rosen, G. 2010. Metaphysical Dependence: Grounding and Reduction. In
Modality: Metaphysics, Logic, and Epistemology, ed. B. Hale and A. Hoffmann,
Sainsbury, M. 1996. Crispin Wright: Truth and Objectivity. Philosophy and
Phenomenological Research 56: 899–904.
Sher, G. 2005. Functional Pluralism. Philosophical Books 46 (4): 311–330.
———. 2013. Forms of Correspondence: The Intricate Route from Thought to
Reality. In Pedersen & Wright 2013a, 157–179.
———. 2016. Epistemic Friction: An Essay on Knowledge, Truth, and Logic.
Tappolet, C. 2000. Truth Pluralism and Many-Valued Logic: A Reply to Beall.
Vision, G. 2004. Veritas: The Correspondence Theory and Its Critics. Cambridge,
MA: MIT Press.
Walker, R.C.S. 1989. The Coherence Theory of Truth: Realism, Anti-realism,
Idealism. London: Routledge.
Williamson, T. 1994. A Critical Study of Truth and Objectivity. International
Journal of Philosophical Studies 30: 130–144.
Wittgenstein, L. 1921. Tractatus Logico-Philosophicus. In Annalen der
Naturphilosophie. English translation by D. F. Pears & B. F. McGuinness
(1961). London: Routledge & Kegan Paul.
Press.
———. 1994. Realism, Pure and Simple: A Reply to Timothy Williamson.
International Journal of Philosophical Studies 2: 327–341.
———. 1995. Truth in Ethics. Ratio 8: 209–226.
———. 1996. Response to Commentators. Philosophy and Phenomenological

Research 56: 911–941.
Philosophy, 24 (suppl. vol.): 31–74.
———. 2001. Minimalism, Deflationism, Pragmatism, Pluralism. In The
Nature of Truth, ed. M. Lynch, 751–789. Cambridge, MA: MIT Press.
———. 2013. A Plurality of Pluralisms. In Pedersen & Wright 2013, 123–153.
Young, J.O. 2001. A Defence of the Coherence Theory of Truth. The Journal of
Yu, A. 2017. Logic for Pluralists. Journal of Philosophy 114: 277–302.
Methodological Pluralism About Truth
Nathan Kellen
1 Introduction
Truth pluralism is the view that there are many ways of being true.
Further, these ways of being true vary with the domain of discourse.
While propositions about the physical world, for example
SE There is an even number of the stars in the universe.
may be true in virtue of some correspondence with the world, this does
not seem plausible for other domains of discourse. Consider a proposi-
tion about morality, for example
SC Sacrificing a younger child so an older child may live is morally

permissible.
N. Kellen (*)

https://doi.org/10.1007/978-3-319-98346-2_6
132 N. Kellen
If SC is true, then it seems true not in virtue of corresponding with some

fact in the physical universe, but in virtue of something else, perhaps
coherence with a set of other moral propositions.1 Let us call this feature
domain-variability:
DOMAIN-VARIABILITY: Ways of being true vary with domains.2
Nearly every truth pluralist account in the literature accepts domain-

variability and treats it as essential to their pluralism.3 But why believe
that ways of being true vary with domains? While we cannot rehearse
here every truth pluralist’s motivation for accepting DOMAIN-
VARIABILITY, I want to suggest that nearly all pluralists share a com-
mon motivation for accepting it. Consider what Crispin Wright notes as
one of his key motivations for truth pluralism:
A pluralistic conception of truth is also philosophically attractive insofar as

an account which allows us to think of truth as constituted differently in
different areas of thought might contribute to a sharp explanation of the
differential appeal of realist and anti-realist intuitions. (Wright 1998,
p. 58)
Or consider more recently (Pedersen and Lynch 2018):
A principal reason for adopting truth pluralism is that the view provides a
framework for understanding the intuitive appeal of respectively realism
and antirealism with respect to different domains. The intuitive appeal
stems in part from the observation that both traditional realist accounts of
truth, such as the correspondence theory, and traditional antirealist
accounts, such as the coherence theory, face a similar pattern of failure.
Theories that seem plausible in some domains fail to seem as plausible in
others. (Pedersen and Lynch 2018, p. 1)
Nearly every truth pluralist account in the literature explicitly

includes at least one realist and one anti-realist domain.4 My claim is
that DOMAIN-VARIABILITY is itself entailed by a more basic
principle:
Methodological Pluralism About Truth 133
(ANTI-)REALISM-VARIABILITY: The metaphysical account of

domains varies; at least one
domain is realist while at least
one other is anti-realist.
These two features are at the core of nearly every extant account of truth
pluralism. Truth pluralism, properly understood, is a combination of
realism and anti-realism into a single, coherent framework. It is this
motivation which underwrites the more obvious aspect of truth plural-
ism—that there are many ways of being true which vary by domain. It is
in this way that truth pluralism is the spiritual successor to the realism/
anti-realism debates which dominated analytic philosophy for decades
after Michael Dummett first introduced his semantic anti-realism in
(Dummett 1959). Truth pluralists are those who recognise that both
camps got something right—that there were intuitions from both realism
and anti-realism that should be captured by our theory of truth.
In what follows, I will take this understanding of truth pluralism as a
combination of realism and anti-realism and expand upon it by consider-
ing the extent to which various truth pluralisms are realist or anti-realist,
in virtue of their methodological commitments. After drawing two dis-
tinctions which will allow us to categorise various pluralist views as realist
or anti-realist I will introduce another form of truth pluralism, method-
ological truth pluralism, which does not privilege either its realist or anti-
realist aspects.
2 Truth Pluralism as (Anti-)realist?

I have argued that truth pluralism is best understood as an attempt to
combine realism and anti-realism into a single framework. But this leads
naturally to a further question: should the truth pluralist theory itself be
categorised as a form of realism, anti-realism, or neither? That is: how
should we characterise truth pluralism?
Some surveys of the truth literature have lumped pluralist views in
with anti-realist ones.5 There are a number of reasons one might do this.
Realism has been the dominant position throughout history and thus
134 N. Kellen
there tends to be a strong presumption in its favour such that any devia-
tion from it is characterised as a type of anti-realism (and of course plural-
ism is a deviation from global realism).
On the other hand, there is some reason to consider truth pluralism as
a form of realism. Anti-realism is generally characterised in terms of its
epistemic constraint: if a domain is anti-realist then all of its truths are
knowable in principle. But if a truth pluralism contains at least one realist
domain, then there will be some unknowable truths, violating the anti-
realist constraint, making it a form of realism.
While each of these positions have merit, I want to suggest a more
disciplined way of determining how a given truth pluralist theory should
be characterised. To do so, we must examine the methodological accounts
of the theories in question. As I’ve already noted, truth pluralists are plu-
ralists precisely because they wish to account for competing realist and
anti-realist intuitions. But some truth pluralists privilege one set of intu-
itions over the other, thus giving realist or anti-realist a privileged posi-
tion in their theory. Consider, for example, Crispin Wright’s
minimalism:
A basic anti-realism – minimalism – about a discourse contends that noth-

ing further is true of the local truth predicate which can serve somehow to
fill out and substantiate an intuitively realist view of its subject matter…
Because of its unassuming character, this minimalism, I suggested should
always be viewed as the “default” stance, from which we have to be shown
that we ought to move. (Wright 1992, p. 174)
Wright accepts a core Dummettian view that realism is a substantive

doctrine, one oft-assumed but little argued for, which requires further
justification over and above anti-realism. Nowhere is this clearer than in
Wright’s discussion of his anti-realist account of truth, which he calls
superassertibility, and its relation to other accounts of truth:
If nothing bars the interpretation of a discourse’s truth predicate as superas-

sertibility, then it is open to us to think of the truth of its statements as
consisting merely in their durably meeting its standards of warranted asser-
tion – a property for which all minimally assertoric sentences are eligible.
But if truth and superassertibility can be prised apart – if it can be shown

that superassertibility is a bad interpretation of the truth predicate in ques-
tion – then the thought is at least strongly suggested that what confers
truth on a statement is not a matter of its meeting standards internal to the
language game, as it were, but its fit with an external reality. (Wright 1992,
p. 142)
Here Wright repeatedly emphasises that superassertibility—his plural-

ism’s anti-realist aspect—has methodological priority over alternative
accounts. For example, each sentence is couched in conditionals (“But if
truth and superassertibility can be prised apart…”). For Wright, if a
domain of discourse is eligible for truth, then it is first granted superas-
sertibility as its truth predicate. If later it is shown that this will not do—
that superassertibility fails as an interpretation of the truth predicate for
the discourse—then we move to another, more realist account of truth.
As Wright claims, “it is realism which must try to make good its case” by
showing that anti-realism is unacceptable.6
While Wright’s pluralism is amenable to having realist aspects, its core
is firmly anti-realist.7 Wright gives clear methodological priority to his
anti-realism; anti-realism is the default from which we must move. This
of course does not rule out his adopting realism in a particular domain,
but it does make it more difficult to do so. It is in this way that Wright’s
pluralism is best understood as a form of anti-realism.
Not all truth pluralists follow Wright in privileging anti-realism. It is
perfectly coherent to formulate a truth pluralism which not only does not
privilege anti-realism but instead privileges realism; consider for example
Edwards (2011):
…although the notion of truth as correspondence to the facts might fit our
domain of discourse about the material world, a different notion of truth –
perhaps one with less metaphysical baggage, constructed out of coherence,
or justification or warrant – may fit the domains in which the correspon-
dence notion looks problematic. (Edwards 2011, pp. 31–2)
The principle here is the same as in the case of Wright and anti-realism.
The realist account of truth (correspondence) is the default, and one from
136 N. Kellen
which we move if it cannot be made to work in some discourse, for exam-

ple, in the moral domain. Edwards’ truth pluralism is, at its core, a realist
account.
When a pluralist account of truth sets either realism or anti-realism as
the default for theory it is treating that aspect of the theory as being more
methodologically fundamental. Wright holds anti-realism to be method-
ologically fundamental and realism to be secondary and vice versa for
Edwards. This is a methodological (or epistemic) attitude that the plural-
ist has towards one aspect of their view: they believe that we more entitled
to it, or that it’s prima facie more plausible, or simply that it is the default
view. We can thus say that a view holds that (anti-)realism is methodologi-
cally fundamental if (anti-)realism is the default account of truth for a
given domain.
It may be helpful to introduce an analogy here. Imagine the truth plu-
ralist constructing her theory as an assembly line worker. Down the line
come various boxes, our domains, in a random order. The truth pluralist
knows that each domain box will need to be stamped with a label: this
one here is a realist box, that one is an anti-realist box, and so on. Like
most pluralists, she has two stamps: one for her realist boxes and the one
for her anti-realist boxes. The productive assembly line worker will put
the stamp most likely to be used in her dominant hand so as to speed up
the process. So which stamp goes in her dominant hand? Well it depends
on whether she thinks the default status of the boxes is realist or anti-
realist. The stamp she privileges will be the one which is what I am calling
methodologically fundamental in her practice.
We set out with the purpose of being able to determine whether truth
pluralism was itself a realist or anti-realist view. With the notion of meth-
odological fundamentality in hand we are able to distinguish pluralist
views into two camps: those which hold realism to be methodologically
fundamental and those which hold anti-realism to be methodologically
fundamental. We can call the former views “realist” and the latter “anti-
realist” for shorthand. This allows us to formulate an answer to the ques-
tion of whether a given truth pluralist account is realist or anti-realist,
depending on its methodological commitments.
Now I wish to introduce a further distinction, built upon method-
ological fundamentality, which will allow us to sort truth pluralist views.
Consider (Lynch 2009)’s manifestation functionalist pluralism. Lynch

introduces a representationalist correspondence theory as his first account
of being true. As he notes, this way of being true seems incompatible with
the possibility of moral propositions being true, leading him to consider
a coherence theory of truth for moral propositions. He proposes the
property of supercoherence:
SUPERCOHERENCE: A moral proposition P supercoheres with a

moral framework F iff P coheres with F at
some stage of inquiry in and continues to do
so at any epistemically improved stage in + 1.
(Lynch 2009, pp. 171–2)
While it is easy to make sense of how supercoherence may be a way of

being true, Lynch finds the account lacking, as it is “too permissive”, and
leaves the possibility that even “craziest moral views” built on false empir-
ical judgments may turn out to be supercoherent, and thus true.8 This
leads him to conclude that
…it is not enough for the fabric of our moral thought to be woven tightly –
to be durably coherent – it must also be nailed down, or grounded on a
firmer floor. (Lynch 2009, p. 173)
Lynch thus introduces his final, anti-realist account of truth,

concordance:
CONCORDANCE A moral proposition P is concordant iff P supercoheres

with a moral framework F and F′s morally relevant non-moral judgments
are true. (Lynch 2009, p. 176)
Note that “true” appears in the definition of concordance, which

means that it can only be (non-circularly) understood when combined
with another theory of truth, like his representational correspondence
theory. Lynch defines concordance out of his representational correspon-
dence theory, and thus the correspondence theory plays a crucial and
ineliminable role in his second account of truth. Lynch’s concordance is
138 N. Kellen
thus essentially made up of two parts: an anti-realist theory (supercoher-

ence) and a realist theory (representational correspondence).
Obviously, this requires that Lynch treat his realist aspect as method-
ologically fundamental, but interestingly it also goes much further. For
Lynch, realism is what I will call theoretically fundamental, insofar as
another part of his theory is defined (partially) in virtue of the realist
aspect of his pluralism. This is to be differentiated from those hold that
realism is methodologically fundamental, for example (Edwards 2011).
Edwards does not suggest adopting an anti-realist account built out of
some sort of correspondence, but rather one built out of entirely separate,
epistemic resources like coherence or warrant.9 Likewise, Wright’s plural-
ism, while methodologically anti-realist, does not attempt to build up a
notion of correspondence from superassertibility.
Let us take stock now. I have introduced two distinctions to help us
determine whether a given pluralist account is realist or anti-realist.
Methodological (anti-)realists treat (anti-)realism as the default, method-
ologically privileged account of truth. Theoretical (anti-)realists not only
treat (anti-)realism as methodologically privileged, but they build their
secondary account of truth out of the default one. Separating these into
exclusive categories, we can sum the discussion up with the following
table:
Methodologically fundamental Theoretically fundamental
Realist Edwards (2011) Lynch (2009)
Anti-realist Wright (1992) ???
As of now, there is no truth pluralism in the literature which treats

anti-realism as theoretically fundamental. It’s not clear how such an
account would go or what would motivate it over a run-of-the-mill realist
theory like a form of correspondence. Nonetheless, it seems to be a con-
ceptually possible account of truth.
It is worth stopping here to consider how this divvying up of the con-
ceptual landscape compares with the other, major way of categorising
truth pluralist views. All truth pluralists agree that there are multiple
ways of being true, that is, there are multiple truth properties. Strong
truth pluralists (e.g. Pedersen and Kim 2018) claim that there is no truth
property had by all the true propositions, while moderate truth pluralists
(e.g. Lynch 2009) believe that there is a generic truth property which all
the true propositions have, in virtue of having one of the other, varying
truth properties. While much of the literature up until this point has
been dominated by moderate truth pluralisms, strong truth pluralism
has recent defenders, including Pedersen and Kim (2018), Cotnoir
(2013) and Ferrari, Morruzzi, Pedersen (ms.). We can also create a table
for this distinction10:
Strong truth pluralism Moderate truth pluralism
Wright (1992), Cotnoir (2013) Lynch (2009), Wright (2013)
Pedersen and Kim (2018) (Edwards 2011)
(Ferrari, Morruzzi, Pedersen, ms.)
Notice that this way of dividing truth pluralist views cuts across the
fundamentality distinctions I introduced. The strong/moderate distinc-
tion does not track the realism/anti-realism distinction whatsoever. While
Lynch (2009) and Wright (1992) differ according to the strong/moderate
distinction, it is not because they fall on opposite sides of the realist/anti-
realist debate, but rather because one advocates for a generic truth prop-
erty and another does not.11 If we are concerned with distinguish types of
truth pluralism based on their core commitments, I suggest that we do so
in virtue of my fundamentality criteria.
3 Methodological Pluralism About Truth

In the previous section I introduced two ways in which a truth pluralist
can be (anti-)realist: by holding (anti-)realism as methodologically or
theoretically fundamental. While these distinctions allow us to answer
the question of whether a given truth pluralist theory is realist or anti-
realist, and they allow us to distinguish various truth pluralist views in a
new way according to their core, methodological commitments, the pic-
ture I have provided is not yet sufficient to capture the full breadth of
truth pluralist views. So far when discussing realism and anti-realism I
have only talked about pluralist views which privilege either one or the
other theory. But must every truth pluralist pick a side?
140 N. Kellen
It may very well be that our language is so diverse that there is no rea-
son to hold a privileged attitude towards realism or anti-realism with
respect to arbitrary domains. Perhaps realism and anti-realism, contra
(Wright 1992; Edwards 2011; Lynch 2009) are on a par. Neither realism
nor anti-realism are the default, and neither require more evidence to
establish themselves as the metaphysics of a domain than the other;
instead, what they require is simply different evidence.
Consider again the assembly line metaphor. Absent some knowledge
of the general features of the boxes such that they are more likely to be
say, realist than anti-realist, it makes little sense for the worker to assume
that the next box coming down the line is realist. This of course does not
mean that she ought to assume that it is anti-realist instead. What she
ought to do is examine each box as it comes down the line and determine
how to categorise it, based on its own particular features.
So too for the truth pluralist, one might argue. Both realists and anti-
realists have claimed methodological superiority for their views. Realists
often note that realism better comports with common sense intuition,
while anti-realists claim that anti-realism is more metaphysically and
epistemically respectable. Perhaps what the pluralist ought to do is to step
back from each of these claims and instead take a methodologically neu-
tral stance, and examine each theory’s case with respect to each individual
domain of discourse. Call such an approach to truth pluralism method-
ological pluralism about truth.
Methodological pluralism about truth may have many virtues. It does
not claim—nor need to provide support for such a claim—that our lan-
guage and conceptual frameworks are structured in such a way that either
realism or anti-realism are more likely to hold in a particular domain. The
methodological pluralist about truth remains neutral (or silent) on that
issue. Given the difficulty of establishing such conclusions, and the long-
standing debates over their success, this should count as a point in favour
of the methodological pluralist.
Further, the methodological pluralist about truth is, in a sense, even
more pluralist than its rival views. Pluralism was developed because
monists failed to see that their theories of truth are implausible globally
but plausible locally. The truth pluralist who treats realism or anti-realism
as fundamental may commit the same sin as the monist, although on a
smaller scale, in failing to recognise the true diversity of our language.

They impose a type of monism—defeasible monism, but monism none-
theless—on our language; the methodological pluralist demurs from
such a claim. This does not require that the methodological pluralist
about truth cleanly divide up the language. It may simply turn out to be
a fact of our language that most domains ought to be construed as realist.
This does not undermine methodological pluralism.
My methodological pluralism about truth is similar to Field’s (1994)
methodological deflationism. It is a working hypothesis: do not assume
that either realism or anti-realism will win out for a given domain absent
any evidence one way or another. In this way, methodological pluralism
about truth is actually compatible not only with pluralist accounts which,
accept realism in all domains but one or vice versa, but in fact with
monistic accounts of truth. While monistic accounts legislate that the
same theory of truth holds globally, over all domains; the methodological
pluralist does not, but nor does she assume that truth will in fact vary
with domain.12 Methodological pluralism about truth, unlike method-
ological deflationism, does not come with it a commitment to a particu-
lar theory being the default for our working hypothesis, however.
4 Conclusion
I have suggested that truth pluralism is best understood as the spiritual
successor to the realism and anti-realism debate, insofar as it attempts to
combine the best aspects of each theory. This raises a natural question:
how ought we understand this new, mixed view? To answer this, I sug-
gested that we again look at the core of truth pluralism: its realist and
anti-realist motivations, and how it treats these views in theory construc-
tion. I suggested that some truth pluralists hold one of the two to be
methodologically fundamental, that is, as the default stance from which we
move. Others go further, in not only giving a privileged methodological
status to one theory but to build the secondary account out of the first,
thereby making the primary aspect theoretically fundamental. I then
showed how these two distinctions can be used to categorise various types
of truth pluralism in a way that cuts across the standard way of under-
standing the literature.
142 N. Kellen
I closed by considering one final view: methodological pluralism about

truth, which does not privilege either its realist or anti-realist forebears.
This methodological pluralism about truth may not be a completely new
form of truth pluralism; many truth pluralists do not take a stand on
whether realism or anti-realism comes first. They simply note that no
account of truth works everywhere, and thus pluralism must be adopted.
Perhaps if each pluralist spent more time reflecting on their methodology
they would determine that they privilege one account over the other, or
perhaps they would opt to remain methodologically neutral. What I have
done here is made a plea to pay attention to the methodological commit-
ments and motivations of various truth pluralisms, and to give a posi-
tion—currently occupied or not—a name, and some arguments in its
favour.13
Notes
1. Barring the truth of some fully reductive moral naturalism, that is.
2. This is similar to what (Beall 2013, p. 324) calls “language-relative truth
pluralism”.
3. The only exception I am aware of is Beall (2013), if one accepts that his
deflated truth pluralism is a pluralism about truth as opposed to a plural-
ism about truth-predicates which may be unrelated to the actual philo-
sophically robust concept of truth. Note also that Beall’s pluralism does
not accept DOMAIN-VARIABILITY, which again sets it far apart from
other views in the literature.
4. This includes the most developed accounts of truth pluralism in Wright
(1992) and Lynch (2009), as well as, for example Cotnoir and Edwards
(2015), Edwards (2011), Pedersen (2006). Cotnoir (2013), who does
not explicitly call for dividing domains by realism/anti-realism but pro-
vides a semantic framework for truth pluralisms which have certain
domains which maintain classical logic and others which have paracom-
plete logics, a conclusion commonly held to follow from adopting real-
ism and anti-realism respectively. Another potential outlier is Gamester
(2017), who does divide up his truth pluralism into realist and anti-
realist parts, but the anti-realist parts are motivated by expressivism
rather than the traditional metaphysical/semantic debates which the

other pluralists are concerned with.
5. See, for example, Burgess and Burgess (2014) and Künne (2005).
6. Wright (1992, p. 174).
7. It is not even clear that Wright (1992) is truly a pluralist account, insofar
as he never explicitly calls for more than one truth predicate. He notes
ways in which we may work our way up from a superassertibility as the
truth predicate of a given domain to a more realist truth predicate, for
example, some form of correspondence, but never explicitly adopts a
second truth predicate. Wright is a pluralist because he believes the con-
cept of truth does not rule out the possibility of plurality of predicates;
as he notes: “Minimalism is thus at least in principle open to the possi-
bility of a pluralist view of truth” (Wright 1992, p. 25, emph. original).
8. Lynch (2009, p. 173).
9. Edwards does not suggest any particular theory of truth to be contrasted
with correspondence, although he mentions some candidates which
have previously been advocated as monistic views of truth (Edwards
2011, p. 32).
10. Cotnoir (2013) reads Wright (1992) as a strong truth pluralist. It is not
clear that this is the only reading of the text, and Wright’s later work is
explicitly a form of moderate pluralism. However, the reading is sup-
ported enough to worth including here.
11. As I note in the previous footnote, the placement of Wright (1992) is
contentious. I prefer to read Wright as a moderate pluralist, who would
thus end up in the same camp as Lynch (2009). This is despite their
views being quite different when it comes down to methodology and
motivations; nearly complete opposites in fact. On either way of inter-
preting Wright I do not believe that the strong/moderate distinction best
captures the disagreement between these views.
12. This is analogous to the way in which Field’s (1994) methodological
deflationism is compatible with inflationism about meaning or content.
13. Thanks to audiences at the University of Connecticut and Yonsei
University for helpful feedback on this chapter. Thanks especially to
Dorit Bar-On, Douglas Edwards, Filippo Ferrari, Will Gamester,
Sebastiano Moruzzi, Nikolaj J. L. L. Pedersen, Joe Ulatowski, Cory
Wright, Crispin Wright, Jeremy Wyatt and Andy Yu. Thanks most of all
to Michael Lynch for countless discussion and feedback.
144 N. Kellen
References
Beall, J. 2013. Deflated Truth Pluralism. In Truth and Pluralism: Current
Burgess, A., and J. Burgess. 2014. Truth. Princeton: Princeton University Press.
Cotnoir, A.J. 2013. Validity for Strong Pluralists. Philosophy and Phenomenological
Research 86 (3): 563–579.
Dummett, M. 1959. Truth. Proceedings of the Aristotelian Society 59 (1):
141–162.
Edwards, D. 2011. Simplifying Alethic Pluralism. The Southern Journal of
Philosophy 49 (1): 28–48.
Field, H. 1994. Deflationist Views of Meaning and Content. Mind 103 (411):
249–285.
Gamester, W. 2017. The Diversity of Truth: A Case Study in Pluralistic
Metasemantics. PhD Dissertation, University of Leeds.
Künne, W. 2005. Conceptions of Truth. Oxford: Clarendon Press.
Teach Us About Alethic Pluralism? Monist 89 (1): 102–117.
Pedersen, Nikolaj J.L.L., and S. Kim. 2018. Strong Truth Pluralism. In Pluralisms
in Truth and Logic, ed. J. Wyatt, Nikolaj J.L.L. Pedersen, and N. Kellen,
107–130. London: Palgrave Macmillan.
Press.
Philosophy 28 (sup1): 31–74.
Wright, Crispin. 2013. A Plurality of Pluralisms. In Truth and Pluralism: Current
Normative Alethic Pluralism
Filippo Ferrari
1 Introduction
Some philosophers have argued that truth is a norm of judgement.1 This
thesis has been given in a variety of formulations—that true judgements
are the correct ones; that it is better to judge truly than to judge falsely;
and that the truth is what judges ought to pursue in enquiry. I will assume
that truth somehow functions as a norm of judgement, and I will be
focusing on two core questions concerning the judgement-truth
norm—namely:
(i) what are the normative relationships between truth and judgement?
(ii) do these relationships vary or are they constant?
I argue for a pluralist picture—what I call normative alethic pluralism2

(henceforth NAP)—according to which (i) there is more than one cor-
rect judgement-truth norm and (ii) the normative relationships between
F. Ferrari (*)
Universität Bonn, Bonn, Germany

https://doi.org/10.1007/978-3-319-98346-2_7
146 F. Ferrari
truth and judgement vary in relation to the subject matter of the

judgement. By means of a comparative analysis of disagreement in three
core areas of the evaluative domain—refined aesthetics, basic taste, and
morality—I show that there is an important variability in the normative
significance that disagreement has in these areas—I call this the variabil-
ity conjecture. By presenting a variation of Lynch’s scope problem for
alethic monism, I argue that a monistic approach to truth’s normative
function is unable to vindicate the normative variability conjecture. I
then argue that NAP provides us with a very promising model to account
for such a conjecture.
I leave the discussion of both the metaphysical issue of what grounds
these normative relationships and the epistemic issue of what justifies our
beliefs about them aside. In particular, I take no stand on the question of
whether truth’s normative function reflects a characteristic intrinsic to the
nature of truth or whether it is grounded in some features external to
truth’s nature. I intend the framework developed here to be compatible
with rejecting the thesis that truth is an intrinsically normative notion.3
2 The Truth-Norm
Advocates of the thesis that truth is a norm disagree about how the norm
should be conceived, and many formulations of it have been given. The
following list provides a sample of formulations that can be found in the
recent debate (my emphases):
Horwich We ought to want our beliefs to be true (and therefore not-
want to have any false ones).4
James The true is the name of whatever proves itself to be good in
the way of belief.5
Loewer True belief is valuable and, other things being equal, it is a
rational doxastic policy to seek true beliefs.6
Lynch End of Inquiry: True propositions are those we should aim
to believe when engaging in inquiry.
Norm of Belief: True propositions are those that are correct
to believe.7
Normative Alethic Pluralism 147
McHugh [T]he attitude of belief sets truth as the standard that a

proposition must meet in order for it to be a fit object of
that attitude.8
Wedgwood For any S, p: If S considers whether p, then S ought to
believe that p if and only if p is true.9
Williams Beliefs aim at truth.10
While philosophers in the list take the truth norm to apply to belief,
I take judgements to be the things that truth is primarily normative
over. Following Shah and Velleman, I consider a judgement to be a
cognitive mental act of affirming a proposition. It is a mental act
because “it involves occurrently presenting a proposition, or putting it
forward in the mind. It is a cognitive act because it involves presenting
the proposition as true”.11 In this respect, judgements differ from
beliefs, which are cognitive mental attitudes. The precise relationship
between judgement and belief is a complex issue which won’t concern
us in this chapter.12 Somewhat simplistically, I take it that the end
product of a judgement is typically a belief whose content is a proposi-
tion. Since the examples I will consider involve an evaluative process, I
take them to be fairly typical cases of beliefs formed through an act of
judging. Thus, I use the expression ‘norms governing judgement’ as a
catchall expression to indicate norms governing the formation, main-
tenance, and relinquishing of beliefs formed by means of an act of
judging.
Two things about this list and the current debate on the judgement-
truth norm deserve some discussion. The first concerns the variability in
the use of the normative vocabulary involved in the formulations above.
Some formulations employ terms like ‘should’ or ‘ought’—thus using
deontic terms; others involve notions such as ‘valuable’ and ‘good’—thus
using an axiological vocabulary; others still talk in terms of truth’s being
the aim of beliefs—using a teleological vocabulary; and, finally, others
formulate the truth-norm in terms of correctness or fittingness—taking
these notions to be normatively independent of any axiological or deon-
tic element. I introduce the term criterial to indicate this latter category
of normative notions. Mixed formulations—that is, formulations mixing
different normative vocabularies—are also possible, as Lynch’s way of
148 F. Ferrari
cashing out the normative role of truth both in terms of aim and in terms
of correctness seems to suggest.
Abstracting from the many issues concerning the precise formulation
and interpretation of the truth norm, we can capture the highlighted
variability in the ways in which the norm is conceived by distinguishing
the following normative dimensions13:
teleological Judgement aims at truth.

criterial It is correct (fitting) to judge that p (if and) only if p is
true.
axiological It is valuable (good) to judge that p (if and) only if p is
true.
deontic One ought to judge that p (if and) only if p is true.
These are four distinct dimensions of the normative constraint that

truth can exert on judgement. In section “Minimal Normative Alethic
Pluralism”, I show that the criterial, the axiological, and the deontic
dimensions of the normative role of truth are normatively independent
of each other. What about the teleological dimension? I take it to be
(partly) constitutive of the act of judging. In judging, a subject per-
forms an act that aims, constitutively, at the truth of the subject matter
at issue. This means that all judgements, as I understand them here,
deal with truth-apt discourse. In this respect, the teleological dimen-
sion is always present whenever a proper judgement is performed and
for this reason I leave this dimension aside in the discussion that
follows.
With this in hand, I call a normative principle (NP) any principle
expressing the normative constraint that truth exerts on judgement in
terms of one or more of the three aforementioned dimensions—that is,
criterial, axiological, and deontic. The notion of a NP so defined allows
for some flexibility. In this sense, all the various norms in the list above
count as NPs as I have defined the term.
The second thing worth mentioning is that the philosophers in the list
above assume—at least implicitly, in the context of their work—that
their preferred NP applies uniformly to all judgements regardless of their
subject matter. In other words, they all seem to share a commitment to a

monistic conception of the normative function of truth—what I call
Normative Alethic Monism (NAM). Such a conception encompasses the
following two theses:
Singularity: There is only one NP expressing truth’s normative func-

tion on judgement.14
Uniformity: NP applies uniformly to all judgements in all areas of
truth-apt discourse.
I take singularity and uniformity to provide a characterisation of NAM in

terms of jointly sufficient and individually necessary conditions.
Moreover, singularity and uniformity are independent theses: one might
deny uniformity but endorse singularity by maintaining that with respect
to a specific set of judgements the truth-norm is normatively silent15; or
one might deny singularity but endorse uniformity by maintaining that
truth uniformly exerts a plurality of normative functions for all judge-
ments. However, neither of these ways of rejecting NAM amounts to
what I take to be a genuine form of normative pluralism (more on this
later).
To make NAM more perspicuous, let us consider two specific applica-
tions of it. First, say that an alethic deontologist is a philosopher who
thinks that the unique normative alethic principle governing all judge-
ments is deontic—that is, it provides thinkers with some prescriptions
about what to judge. An alethic deontologist, then, takes truth and falsity
to always line up, respectively, with the obligatoriness and the impermis-
sibility of the judgement in question. Second, those—alethic axiologists—
who think of truth’s normative function in purely axiological terms—that
is, in terms of what it is good or bad to believe—will take truth and falsity
to always line up, respectively, with the good and the bad in the way of
judgements.
NAM imposes a structural rigidity in the normative function of truth
which, I argue, makes it unable to account for differences in the norma-
tive significance of disagreement.
150 F. Ferrari
3 isagreement and Its Normative

D
Significance
By the expression ‘normative significance of disagreement’ I mean the
extent to which engaging in a disagreement licenses participants to nega-
tively evaluate each other’s view. The kind of disagreement that matters
for the purposes of this chapter is something analogous to what
MacFarlane calls the simple view of disagreement (i.e. doxastic
noncotenability)16:
DIS To disagree with someone’s judgement that p is to issue judge-

ments whose contents are jointly incompatible with p.17
Disagreement, in the DIS model, is thus minimally normatively signifi-

cant in the following way: a commitment to judging that p is a commit-
ment not only to the truth of p but also to the falsity of every content q
which is incompatible with p, and consequently a commitment to assess-
ing anybody endorsing q as judging falsely. However, what other kinds of
negative normative assessment or reactive attitude18 to a contrary view are
associated with an attribution of falsity and thus licensed by the presence
of a disagreement varies in relation to the subject matter at issue. Call this
the variability conjecture.
With this in hand, let’s now turn to a discussion of some examples. The
examples I will consider involve evaluative judgements—that is, judge-
ments about evaluative matters such as basic taste, refined aesthetics, and,
more controversially, fundamental morality.19 These are judgements for
which a substantive kind of objectivity may be hard to sustain and, for
this reason, they are taken to lack robustly representational content—in
the sense of Wright (1992, Chap. 4). In this respect, these judgements are
not subject to the possibility of a more fundamental kind of failure—a
failure deriving from a misrepresentation of how things are objectively,
that is, in mind-independent reality.
Moreover, I should point out that these are just a few examples that I
have chosen because I think they are particularly fit to show the kind of
variability in the normative significance of disagreement I am interested
in. Nothing hinges on which particular example we choose to highlight

this variability. Also, I am not claiming that all examples of moral, aes-
thetic, and taste disagreement should be understood in the way suggested.
Perhaps there is also substantive intra-domain variability. This, though,
wouldn’t weaken my proposal—the more variability the better. However,
in order to keep the discussion reasonably simple and smooth, I shall talk
as if the three examples I discuss are paradigmatic of the three domains in
question.
4 Disagreement (1): Fundamental Moral

Consider the following disagreement about some fundamental moral
value between Julie and Jill:
Julie thinks that torturing people is always a morally deplorable practice.

Jill disagrees, thinking that sometimes torturing people is after all morally
acceptable.
I take this to be an example of some deep incompatibility between moral

values concerning fundamental human rights. Thus, we may ask: what
kind of normative assessment of Jill’s contrary view would we typically
expect from Julie in such a situation of disagreement? We would expect
Julie to issue a strongly negative assessment of Jill’s view. Not only would
Julie think that Jill is completely wrong, but she would also feel com-
pelled to think that Jill ought to change her mind about that issue. In
fact, she would find it quite deplorable that Jill holds such a morally bad
view.
What this suggests is that the kind of reaction we would expect from
Julie amounts to a substantial criticism of Jill’s contrary view, which, in
some cases, leads to a disposition to urge Jill to revise her view on torture.
Even assuming that Julie and Jill consider each other equally knowledge-
able on most of the relevant non-moral facts concerning torture, once the
disagreement comes to light they will typically cease to consider each
other as equally respectable moral judges, at least with respect to a range
of topics closely connected to the subject matter of their disagreement.
152 F. Ferrari
An appealing explanation of the appropriateness of this kind of sub-

stantive criticism of our opponent’s moral standing might be linked to a
distinctive feature of moral value—namely, what we might call the higher-
order heritability of moral value.20 The thought is that the moral condem-
nation of a certain action or practice in these fundamental cases engenders
a commitment to judging the holding of any view favourable to that
action or practice as alethically impermissible. In this respect, strong
moral criticism of a certain practice carries with it a strong alethic criti-
cism of the holding of any judgement supporting that practice. The criti-
cism is typically conveyed by means of an attribution of falsity to such a
judgement, which, in this case, goes hand in hand with assessing the
judgement as impermissible.
That said, it is important to keep in mind that this is a reconstruction
of what we might typically expect in a situation of disagreement about
fundamental moral issues. In other words, this is merely a conjecture con-
cerning the typicality of a certain pattern of response to a situation of
disagreement in some radical cases of incompatibility of fundamental
moral values. There certainly may be a variety of contextually salient fac-
tors that would impact on the kind of critical assessment that we would
deem appropriate—but for the sake of simplicity I will leave these atypi-
cal cases aside.
5 Disagreement (2): Basic Taste

Consider now a situation in which Julie and Jill have a dispute about
basic taste. I here draw an intuitive distinction between judgements
about matters of basic taste and judgements about matters of refined
aesthetics. Without endorsing any controversial thesis about where to
draw the boundary between basic taste and refined aesthetics—there
might be no sharp boundary—I will discuss two examples that might
be taken as paradigmatic of each category. Let’s begin with the basic
taste case:
Julie thinks that oysters are delicious.

Jill disagrees, thinking that they are tasteless.21
In sharp contrast to the fundamental moral case, the kind of negative

assessment of Jill’s contrary view we would typically expect from Julie
would be paradigmatically weak—in fact nothing more than the realisa-
tion that they have divergent views on the taste of some particular oys-
ters, and this alongside with an appreciation, in typical circumstances,22
of the legitimacy of Jill’s contrary view. If this approximates what people
would typically expect in a situation of disagreement about basic taste, it
shows that no substantive criticism of a contrary view should be expected.
Absent pragmatic issues of coordination,23 a live and let live attitude is
deemed appropriate. We may take as evidence of this that a rather natural
follow-up that we might expect from Julie in coming to know about Jill’s
contrary view would be a first-person qualification of the judgement—
for example, “Oh, well, I just wanted to say that I like these oysters”—24
as a way of avoiding, in the dialectical context of the disagreement, the
emergence of what might be considered an unjustified and futile quarrel.
In fact, the mere presence of disagreement does not typically make either
Jill or Julie less confident about her own opinion concerning oysters.
What might happen sometimes is that, whenever it is plausible to assume
a certain commonality of taste within the context of the disagreement, we
experience a sense of surprise or curiosity in coming to know that some-
one we thought had a similar taste in fact has a radically different opinion
about the food in question.
6 Disagreement (3): Refined Aesthetics

Last, consider the following dialogue about a matter of refined
aesthetics:
Julie thinks that Gould’s 1955 execution of Goldberg Variations is unequalled.

Jill disagrees, thinking that the 1955 execution is too virtuosic. The 1981
version is preferable as a more mature interpretation.
What kind of negative assessment of someone’s holding a contrary view

should we expect in a disagreement about matters of refined aesthetics?
Here it seems plausible to conjecture that the kind of critical reaction to
154 F. Ferrari
Jill’s contrary view that people would deem appropriate for Julie to have
is one that falls somewhere in the middle of the two extremes of the fun-
damental moral case and the basic taste case. Thus, we may typically
expect Julie to be somehow surprised by Jill’s contrary judgement regard-
ing Gould’s 1955 execution of Goldberg Variations. This is because she
would think that Jill’s judgement is somehow off-colour. Thus, a certain
degree of criticism would be regarded as appropriate. In particular, it
would be natural for Julie to consider Jill’s aesthetic sensibility on this
occasion to be not as good as her own. In this respect, contrary to the
basic taste case, disagreement about refined aesthetics seems to give rise to
an attribution of fault. Although it seems appropriate for Julie to con-
tinue to regard Jill as well-informed as she is about Glenn Gould’s musi-
cal production, by being committed to her own judgement about Gould
and thus to her own scale of aesthetic value, she is committed to assessing
Jill’s opinion as inferior—indeed, one that it would be better not to have.
As a result, it might happen that in some cases the conversation continues
by each subject trying to persuade the opposite party to change her mind.
However, in contrast with the fundamental moral case described above,
the degree of intensity of the reactive attitude that would be appropriate
for Julie to have in response to Jill’s contrary view is significantly lower
than in the fundamental moral case. We could certainly understand, if
not justify, a high degree of heat in disputes about fundamental moral
issues, but not in disputes about refined aesthetics. Regardless of how
much aesthetically off-colour I consider your contrary view, granted that
we are equally knowledgeable about the subject matter in question I
would still feel some pressure to regard you as permitted to hold to the
view you endorse.25
7 ormative Alethic Monism

N
and the Normative Scope Problem
I will now argue that NAM falls prey to a serious objection that originates
in a particular application of what is known in the truth pluralism debate
as the scope problem for monistic (substantivist) conceptions of truth. The
thought at the core of the objection is that no single substantive notion
of truth—for example, correspondence, superassertibility, or coher-

ence—seems to be fully adequate to play the truth role across all different
domains.26 For instance, whereas correspondence might be the adequate
notion of truth for our judgements about the physical constitution of the
world, it might score less well as the notion of truth for the domain of
basic taste and other evaluative domains. Similarly, a version of the coher-
entist conception of truth might be the adequate notion of truth in
mathematics or in the moral domain, but it seems utterly inadequate for
contingent judgements.27 Lastly, superassertibility might provide the
right model for truth in some evaluative domains or in mathematics,28
but it is inadequate for judgements concerning matters that are epistemi-
cally inaccessible to us—for example, claims about the past for which no
evidence has survived. Thus, although all these various notions of truth
might provide a locally adequate model of truth, none of them gives us
an adequate model of truth across all different domains.
My contention is that the monistic conception of the normative role
of truth faces an analogous problem. If something along the lines of the
variability conjecture described in sections “Disagreement and Its
Normative Significance”, “Disagreement (1): Fundamental Moral”,
“Disagreement (2): Basic Taste”, and “Disagreement (3): Refined
Aesthetics” is accurate, then it is easy to see why NAM is in trouble. It
would in fact require thinkers to adopt a Procrustean attitude towards the
negative assessment of a contrary view in the presence of a disagreement,
regardless of the specific subject matter at issue. This would mean that,
depending on which NP we take to characterise truth’s normative func-
tion, some of the normative assessments described in the examples would
be deemed inappropriate.
To illustrate: recall that alethic deontologism is the view according to
which truth exerts only a deontic constraint over all judgements, cashed
out in terms of ‘ought’, ‘permissible’, and ‘impermissible’. Then this view
would have problems accounting for the kind of normative assessment
that seems appropriate in the domain of basic taste—and, arguably, in
the domain of refined aesthetic matters as well. This is because such a
view would say that false judgements are always impermissible and thus
that a reactive attitude of rational condemnation to a contrary judgement
would be always legitimate.29 But this is not what we would typically
156 F. Ferrari
expect in situations of disagreement about basic taste or, more controver-

sially, refined aesthetics.
Secondly, recall that alethic axiologism is the view according to which
truth exerts only an axiological constraint over all judgements, cashed out
in terms of ‘valuable’, ‘good’, and ‘bad’. This view would have problems
in accounting for the normative assessment typical of disagreement about
basic taste—and, arguably, about fundamental moral matters. This is
because such a view would say that false judgements are always bad, but
not impermissible. However, this prediction seems inadequate for the
basic taste case—because no substantive criticism is forthcoming—and
also for the fundamental moral case—because a more substantive nega-
tive reaction would be expected.
Last, let us call alethic criterialism the view according to which truth
exerts only a criterial constraint on judgement, cashed out in terms of
‘correctness’, ‘fittingness’, and ‘incorrectness’. Since the kind of norma-
tive assessment of a contrary view predicted by this view would be para-
digmatically weak, it would be inadequate to account for the kind of
normative assessment that occurs in disagreements about fundamental
moral issues and refined aesthetic matters—where some (more or less)
substantive criticism to a contrary view would typically be deemed appro-
priate. The core of the normative scope problem is thus that none of these
monistic views is adequate as a general view about truth’s normative func-
tion. This puts considerable pressure on NAM.
8 Normative Alethic Pluralism

In this and the following section “Introduction” introduce a pluralist
framework for modelling the normative function(s) that truth exerts over
judgements—Normative Alethic Pluralism—and show how this frame-
work can be put to use to solve the scope problem. At the core of NAP
are the following two theses:
Plurality There is more than one NP expressing truth’s normative

function on judgement.
Variability Truth exerts a variable normative function on judgements
from different areas of truth-apt discourse.
Plurality and Variability are meant to provide individually necessary and

jointly sufficient conditions for NAP. They amount to the negation of
Singularity and Uniformity respectively. Thus, according to NAP there is
more than one correct NP expressing truth’s normative function and,
moreover, different areas of discourse are governed by different NPs.
NAP predicts that while the normative function of truth in one area of
discourse might be expressed by a NP that encompasses the deontic
dimension, in some other areas it might be expressed by a NP which
encompasses the axiological dimension, and in other areas still it might
be expressed by a NP which is exhausted by the criterial dimension—
thus, without any deontic or axiological constraint. Interestingly, NAP
leaves open the possibility of an area of discourse where truth is norma-
tively inert, as it were, while still functioning as the aim of judgements in
that area—and thus still exerting its teleological function. This means
that it allows for a purely deflationary notion of truth (and falsity) that
operates, locally, in some domains of discourse.30
In order to defend the cogency of NAP, it is important to show that the
criterial, the axiological, and the deontic dimensions enjoy a certain
degree of normative independence from each other. There are two ways
in which these three dimensions can be normatively independent:
(IND 1) Criterial⇏Axiological; Criterial⇏Deontic; Axiological⇏Deontic31

(IND 2) Deontic⇏Axiological; Deontic⇏Criterial; Axiological⇏Criterial
Arguing for both (IND 1) and (IND 2) would get us full normative inde-
pendence, which means the strongest available form of NAP—call it
strong NAP. Otherwise we could argue for only one direction of norma-
tive independence, and we would get two possible moderate forms of
NAP—the first endorses (IND 1) without endorsing (IND 2), while the
second endorses (IND 2) without endorsing (IND 1).32 Since the rele-
vant direction of normative independence that we need in order to deal
with the variability illustrated in the examples above is that expressed by
(IND 1), in what follows I will argue for (IND 1) only, and I will remain
silent on (IND 2). Moreover, to provide a full defence of the relevant
moderate version of NAP we need an argument to show that (IND 2)
fails—that is, to show that we have at least one of the following directions
of normative dependence:
158 F. Ferrari
(DEP 1) deontic⇒axiological
(DEP 2) deontic⇒criterial
(DEP 3) axiological⇒criterial
Even though I believe that these normative dependences are plausible, I

won’t argue for them. What I will do in the next section is to defend a
minimal version of NAP that is sufficient to solve the impasse generated
by the normative version of Lynch’s scope problem. Such a minimal ver-
sion can then be elaborated either in the direction of strong NAP or in
that of moderate NAP.
9 Minimal Normative Alethic Pluralism

To argue for minimal NAP we need to establish the following two claims:
(i) that the presence of the criterial dimension by itself does not enforce
either the axiological dimension or the deontic dimension and (ii) that
the axiological dimension by itself does not enforce the deontic one.
Let’s discuss the first claim first. The criterial dimension provides us
with a standard for categorising judgements as correct or incorrect in
accordance with the truth-value of their propositional content. There is
an ongoing debate about how to interpret the correctness norm and in
what sense it is normative33—whether in deontic terms, such as ‘ought’,
or in what I call axiological terms, such as ‘good’, ‘bad’, and ‘valuable’. I
reject both interpretations. More precisely, I do not think that we are
forced to adopt either interpretation. I follow McHugh34 here in under-
standing correctness as a distinct normative property in its own right, as
neither deontic nor axiological. This thesis presupposes that the category
of the normative should be conceived—as Thomson (2008) suggests—as
more variegated than philosophers have traditionally sustained. Taking
this on board, it may be argued that for a judgement to be correct is for
it to have a normative property. But it is not correct because one ought to
hold it, or because it would be good if one held it. Correctness is distinct
from these other normative properties. For one thing, quite intuitively,
while tipping might be the correct thing to do in a certain restaurant,
provided that you have received adequate service, it’s not something
obligatory—thus refraining from tipping would be somehow incorrect,

given the circumstances, but not impermissible. In this respect, the fact
that it is correct to judge that p does not imply that the subject ought to
judge that p. Moreover, something might be a correct thing to do,
because, for example, it is within your rights to do so, without this imply-
ing that it is also a valuable or good thing. The correct way of setting the
table is to put the knife to the right of the plate and the fork on the oppo-
site side, but intuitively there’s nothing valuable in setting the table that
way.35 In this respect, the criterial dimension does not entail the axiologi-
cal one. In general, there should not be any deep difficulty in the thought
that something could be correct without being valuable or obligatory, or
incorrect without being disvaluable or impermissible. However, further
inquiry is needed into what specific kind of sui generis normative notion
correctness may be.36
Let us now turn to a discussion of the second normative independence
claim—that is, that the presence of the axiological dimension by itself
does not enforce the presence of the deontic one. The axiological dimen-
sion tells us that it is, at least pro tanto, good, or valuable, to judge accord-
ing to what is true, and that it is, at least pro tanto, bad, or disvaluable,
to judge falsely. Again, how to understand the axiological dimension is a
much-discussed topic.37 However, for the purpose of this chapter, we can
leave the many, undoubtedly important, issues raised by that debate
aside, and focus on the question of whether the axiological dimension per
se entails the deontic dimension. There are reasons for thinking that it
does not. As McHugh points out
Something may be bad without its badness being a matter of anyone’s hav-
ing done anything they ought not have done, and without its being the
case that there is anyone who ought to change it; some prospective state of
affairs or object may be good without its being the case that there is anyone
who ought to produce it or bring it about.38
Thus, a false judgement may be bad or disvaluable without its badness or

disvalue being a matter of someone having done anything they ought not
to have done in judging so. Analogously, a true judgement may be good
or valuable without its being the case that anyone ought to judge that way.
160 F. Ferrari
This thought can be substantiated by drawing a parallel between the

moral case and the alethic case by means of the notions of supererogation
and suberogation. In ethics, a supererogatory act is generally conceived as
an act that goes beyond the call of duty and brings about a considerable
moral good. As Chisholm puts it, supererogatory acts are “non-obligatory
well-doing”.39 By contrast, following Chisholm’s characterisation, we can
say that suberogatory acts are “non-forbidden ill-doing”, or, in Driver’s
words “suberogatory acts are bad to do, but not forbidden”.40 Heroic acts
are typical examples of the supererogatory: if I sacrifice my life in order to
save a group of children who would otherwise die, this would be consid-
ered extremely good and valuable, but clearly non-obligatory—even if
sacrificing my life is the only way to save those children. As an example
of a suberogatory act, consider the following scenario. Julie has two per-
fectly healthy kidneys, but she refuses to donate one kidney to her sister
Jill, who desperately needs one. Julie’s refusal would be clearly suberoga-
tory. Since Julie is under no obligation to donate her kidney, her refusal
is not an impermissible thing to do. However, we would certainly con-
sider such a refusal morally disvaluable. This reflects the widespread
thought that it would be better if Julia donated her kidney to Jill.
What the categories of the supererogatory and suberogatory show is
that there are acts that are intuitively morally optional but still open to an
axiological assessment as to whether they are good or bad. In the alethic
case, we may say that an alethically supererogatory judgement—under-
stood, remember, as a cognitive mental act—is a good judgement to
make, because true, but not obligatory, while an alethically suberogatory
judgement would be a judgement that is bad in that it violates the axio-
logical dimension of the normative function of truth without thereby
being a judgement that ought not to be performed.41
Unless we can find an argument to show that the concept of alethic
suberogation is incoherent we should remain open to the possibility that in
some areas of discourse there are cases of suberogatory judgements. This
means that we should be wary of assuming that compliance with or viola-
tion of the axiological dimension automatically engenders compliance with
or violation of the deontic dimension. As a consequence, we have the idea
that truth’s normative function in some domain may encompass the axio-
logical dimension while at the same time being deontically silent.
No doubt, more need to be said to have an exhaustive picture of how

these dimensions of truth’s normative role are related to each other.
However, I take the remarks that I have just offered as sufficient to estab-
lish the reasonableness of the partial normative independence claim
(IND 1). This granted, we have at least the following four NPs governing
truth’s normative function:
(NP-1) Truth exerts a deontic constraint on judgements—perhaps

together with both an axiological and a criterial constraint if we
endorse (DEP 1 and DEP 2).
(NP-2) Truth exerts an axiological but not a deontic constraint on
judgements—perhaps together with a criterial constraint if we
endorse (DEP 3).
(NP-3) Truth exerts only a criterial constraint on judgements.
(NP-4) Truth exerts no normative constraint—besides its teleological
role.
10 ormative Alethic Pluralism

N
and the Normative Scope Problem
How does NAP help in dealing with the normative version of Lynch’s
scope problem as illustrated in section “Normative Alethic Monism and
the Normative Scope Problem”? The specific kind of negative normative
assessment that I have in mind here can be understood by means of the
notion of normative fault. The conjecture concerning the variable
normative significance of disagreement discussed above can be reformu-
lated as follows: the kind of fault that is attributable to someone holding
a contrary view in the context of a disagreement might vary in relation to
the subject matter of the disagreement. What does this variation in nor-
mative fault amount to? How should we model it? Once we take norma-
tive alethic pluralism on board, we are open to the idea that there is more
than one dimension in which truth governs judgement. Correspondingly,
we have a plurality of ways in which someone holding a view that is
judged false might be said to be at fault. Thus, we have the following four
categories of fault-attribution:
162 F. Ferrari
deontic fault In judging not-p the subject is judging in a way she

ought not to.
axiological fault In judging not-p the subject is doing something
disvaluable.
criterial fault In judging not-p the subject is judging incorrectly.
no fault It is perfectly fine for the subject to judge the way
she does.
As we saw in the previous section, the various dimensions of truth’s nor-

mative profile are to a certain extent independent of each other. The vari-
ous kinds of attributions of normative fault I have just introduced enjoy
an analogous kind of independence. In particular, in line with the mini-
mal version of NAP I am endorsing here, the three main categories of
fault-attribution listed above are independent of each other in the follow-
ing way:
(IND 3) c
riterial fault⇏axiological fault; criterial
fault⇏deontic fault; axiological fault⇏deontic fault.
Whether the other direction of independence also stands depends on

whether we argue for a moderate or a strong version of NAP. I intend to
remain neutral on this issue.
Thus, we might say that one’s recognition that one disagrees with
another person about certain subject matters licenses an attribution of
criterial fault without licensing an attribution of axiological or deontic
fault. Reflecting on the kind of negative assessment that seems appropri-
ate in the disagreement between Julie and Jill about oysters, this could be
the right prediction. Thus, the (NP-3) model of the normativity of truth
seems to be the adequate one for the domain of basic taste.
By contrast, disagreements about other matters might license an
attribution of both criterial and axiological fault without licensing an
attribution of deontic fault. And this might be what happens in typical
cases of disagreement about refined aesthetic matters. Although it would
seem appropriate for Julie to assess Jill’s contrary but equally well-
informed view as somehow inferior—and thus to attribute axiological
fault—it would seem inappropriate for Julie to assess Jill’s contrary view
as impermissible. Thus, no deontic fault seems to be forthcoming. In

this respect, the (NP-2) model of the normativity of truth would be the
adequate one for the domain of refined aesthetics.
Lastly, some disagreements might license an attribution of all these
kinds of fault. This seems to be what happens in disagreements about
fundamental moral issues where both parties would deem it appropriate
to assess the opposite party’s contrary view not only as incorrect and mor-
ally less valuable but also as one that ought not to be had. If this is right,
then (NP-1) would be the right model for the normativity of truth in the
fundamental moral domain.
An important question remains as to whether sense can be made of the
(NP-4) model of the normativity of truth—that is, whether we can have
sensible disagreements involving judgements teleologically constrained
by truth where no fault is involved. This ultimately depends on whether
we can have a normatively inert truth property operating locally, in some
areas of discourse.42
11 Conclusions
I have discussed what I take to be an interesting variability in the norma-
tive significance of disagreement and I have suggested that such variabil-
ity requires some flexibility in the normative function that truth exerts in
different areas of discourse. Because a monistic view on truth’s normative
function—what I have called normative alethic monism—forces us to
adopt a Procrustean attitude towards the normative significance of dis-
agreement, it is inadequate as a general model of the normativity of truth.
I have argued for this by presenting a variation of what is known in the
truth pluralism literature as the scope problem.
I have outlined a pluralist framework for understanding the normativ-
ity of truth—normative alethic pluralism—that promises to score better
than NAM in addressing the variability in the normative significance of
disagreement discussed above. The key point of my proposal is to under-
stand variation in normative fault by looking at which dimensions of
truth’s normative profile operate in a given domain of discourse. Once we
adopt this pluralistic stance towards the normativity of truth, we obtain
164 F. Ferrari
a framework which adequately models the variability in the normative

significance of disagreement across different areas of discourse. In other
words, NAP provides us with all the necessary tools for addressing the
normative version of the scope problem.43
Notes
1. For example: Dummett 1959; Gibbard 2005; Horwich 2013; Shah and
Velleman 2005; Wedgwood 2007; Wright 1992.
2. Since the label can be misleading in one important respect, I should
clarify that with it I mean a pluralist account of the normative function
that truth plays in relation to judgements. It is not part of the proposal
to claim that this pluralist account requires or entails a pluralistic account
of the nature of truth—although, the two views, taken together, gives a
highly coherent and neat package.
3. See Ferrari (2016b) for an account of the normativity of truth—espe-
cially of what I call the axiological dimension of the normativity of truth
(see below, section “The Truth-Norm”)—which is compatible with the
minimalist conception of truth advocated by Horwich (in, e.g., Horwich
1998). For further discussion of value and Horwich’s minimalism, see
Ferrari 2018.
4. Horwich 2013: 17.
5. James 1975: 42.
6. Loewer 1993: 266.
7. Lynch 2013: 24.
8. McHugh 2014: 177.
9. Wedgwood 2007.
10. Williams 1973: 136.
11. Shah and Velleman 2005: 503.
12. On this, see Chrisman 2016 and Sosa 2015, part III.
13. I focus on judgement rather than belief (as all the authors mentioned
above do) to avoid intricate issues concerning the doxastic voluntarism
versus non-voluntarism debate.
14. Singularity resembles Williamson’s simple account, but without the con-
stitutivist element. See Williamson 2000: 240.
15. See Williams 2012 for an account of normative silence.
16. MacFarlane 2014: Chap. 6.
17. Where ‘incompatibility’ here is understood in semantic terms, either as

contrariness or contradictoriness of the propositional contents involved.
18. In the sense discussed by Strawson 1962.
19. NAP can be integrated into a realist conception of morality and aesthet-
ics and help in getting a handle on the difference in normative signifi-
cance between these kinds of disagreement and disagreement about
some other factual matters.
20. This label has been suggested to me by Crispin Wright (personal
conversation).
21. I assume, contra contextualists and expressivists, that even in the case of
basic taste we can make sense of disagreement in terms of doxastic non-
cotenability. Defending this claim would take me too far away—but see
MacFarlane 2014 and Ferrari and Wright 2017 for some arguments
against a contextualist and expressivist treatment of basic taste
judgements.
22. Here by typical circumstances I mean circumstances in which no appre-
ciable defeater—for example, one of the parties being under anaesthetics
or her gustatory sensibility being temporarily impaired because of the
effect of a strong cough syrup that alters her taste—is in place.
23. E.g. a situation in which Julie and Jill have to decide whether to take
their best friend to a French bistro.
24. A discussion of this point can be found in Ferrari and Wright 2017.
Wyatt 2018, pp. 263–5 makes a similar prediction about first-person
qualifications in cases of disagreement about basic taste.
25. See Ferrari 2016a for a more detailed discussion of the comparison
between disagreement in basic taste and disagreement in refined
aesthetics.
26. See Lynch 2009: 34–36.
27. A classical objection to coherentist accounts of truth can be found in
Wright 1998.
28. If, for example, you endorse a constructivist account of mathematics.
29. It is important to highlight that the legitimacy of having a certain reac-
tive attitude does not entail the legitimacy of expressing that attitude.
There might be reasons (e.g. prudential, moral, or other kinds of con-
textual factors), that are independent of the norms governing judge-
ments, that would make the expression of my reactive attitude
inappropriate even though it would be legitimate for me to have such
an attitude.
166 F. Ferrari
30. The cogency of this option depends on whether we can make sense of a
purely non-normative notion of truth. This is a debated issue among
philosophers working on truth: see, for instance, Dummett 1959;
Wright 1992; Lynch 2009; Wrenn 2015. For some replies, see, e.g.,
Horwich 1998; Ferrari 2016b; Ferrari and Moruzzi (2018).
31. ‘⇏‘should be read as: ‘does not enforce…’.
32. It is helpful to point out that the views that I call ‘strong NAP’ and ‘mod-
erate NAP’ are quite different from the views that are often called ‘strong
alethic pluralism’ and ‘moderate alethic pluralism’ in the alethic plural-
ism debate. Two remarks are especially relevant on this: first, that strong
NAP doesn’t entail strong alethic pluralism and, second, that moderate
NAP doesn’t entail moderate alethic pluralism.
33. See Thomson 2008 for a discussion of the various kinds of normativity
in relation to judgements.
34. See McHugh 2014 and McHugh and Way 2015.
35. There is a similar, and familiar, contrast in the normative ethics debate:
one might think that right actions are those that maximise utility but ask
what is good about doing what is right.
36. McHugh (2014: 177) suggests that we should understand correctness in
terms of fittingness: ‘the attitude of belief sets truth as the standard that
a proposition must meet in order for it to be a fit object of that attitude
[…] For an attitude to be fitting is for it to have a normative property.
But it is not fitting because you ought to hold it, or because you may
hold it, or because it would be good if you held it. Fittingness, I main-
tain, is distinct from these other normative properties’.
37. See, for instance, David 2005; Hazlett 2013; Kvanvig 2003; Lynch
2005; McHugh 2012.
38. McHugh 2012: 10.
39. Chisholm 1963: 3.
40. Driver 1992: 286.
41. Turri first applied the category of the suberogatory to the case of the
normativity of assertions, but differently, and with different aims; see
Turri 2013.
42. Ferrari and Moruzzi (2018).
43. This paper has enormously benefitted from discussions with Elke
Brendel, Matthew Chrisman, Massimo Dell’Utri, Douglas Edwards,
Matti Eklund, Andreas Fjellstad, Patrick Greenough, Thomas
Grundmann, Paul Horwich, Nathan Kellen, Michael Lynch, Giacomo
Melis, Anne Meylan, Moritz Müller, Carol Rovane, Andrea Sereni, Erik
Stei, Elena Tassoni, Joe Ulatowski, Giorgio Volpe, Jack Woods, Chase
Wrenn, Cory Wright, Jeremy Wyatt, Luca Zanetti, Dan Zeman. Special
thanks are due to Carrie Ichikawa Jenkins, Sebastiano Moruzzi, Nikolaj
J. L. L. Pedersen, Eva Picardi and Crispin Wright. Moreover, I would
like to acknowledge the generous support of the Deutsche
Forschungsgemeinschaft (DFG—BR 1978/3–1) for sponsoring my
postdoctoral fellowship at the University of Bonn.
acknowledged.
References
Chisholm, R. 1963. Supererogation and Offense: A Conceptual Scheme for.
Ethics. Ratio 5: 1–14.
Chrisman, M. 2016. Epistemic Normativity and Cognitive Agency. Noûs.
Doihttps://doi.org/10.1111/nous.12184.
David, M. 2005. On Truth Is Good. Philosophical Books 46 (4): 292–301.
Driver, J. 1992. The Suberogatory. Australasian Journal of Philosophy 70 (3):
286–295.
Dummett, M. 1959. Truth. Proceeding of the Aristotelian Society 59: 141–162.
Ferrari, F. 2016a. Disagreement About Taste and Alethic Suberogation. The
Philosophical Quarterly 66 (264): 516–535.
———. 2016b. The Value of Minimalist Truth. Synthese. Doi https://doi.
org/10.1007/s11229-016-1207-9.
———. 2018. The Value of Minimalist Truth. Synthese 195 (3): 1103–1125.
https://doi.org/10.1007/s11229-016-1207-9.
Ferrari, F., and S. Moruzzi. 2018. Ecumenical Alethic Pluralism. Canadian
Journal of Philosophy. https://doi.org/10.1080/00455091.2018.1493880.
Ferrari, F., and C. Wright. 2017. Talking with Vultures. Mind 126 (503):
911–936.
Gibbard, A. 2005. Truth and Correct Belief. Philosophical Issues 15: 338–350.
Hazlett, A. 2013. A Luxury of the Understanding: On the Value of True Belief.
Horwich, P. 1998. Truth. 2nd ed. Oxford: Oxford University Press.
———. 2013. Belief-Truth Norms. In The Aim of Belief, ed. T. Chan, 17–31.
168 F. Ferrari
James, W. 1975. Pragmatism and the Meaning of Truth. Cambridge MA: Harvard
University Press.
Kvanvig, J. 2003. The Value of Knowledge and the Pursuit of Understanding.
Cambridge: Cambridge University Press.
Loewer, B. 1993. The Value of Truth. Philosophical Issues 4: 265–280.
Lynch, M. 2005. True to Life. Cambridge MA: The MIT Press.
———. 2013. Three Questions for Truth Pluralism. In Truth and Pluralism, ed.
Nikolaj J.L.L. Pedersen and C.D. Wright, 21–41. Oxford: Oxford University
Press.
MacFarlane, J. 2014. Assessment Sensitivity: Relative Truth and Its Applications.
McHugh, C. 2012. The Truth Norm of Belief. Pacific Philosophical Quarterly 93:
8–30.
———. 2014. Fitting Belief. Proceedings of the Aristotelian Society 114 (2):
167–187.
McHugh, C., and J. Way. 2015. Fittingness First. Ethics 126 (3): 575–606.
Shah, N., and D. Velleman. 2005. Doxastic Deliberation. Philosophical Review
114 (4): 497–534.
Sosa, E. 2015. Judgement and Agency. Oxford: Oxford University Press.
Strawson, P. 1962. Freedom and Resentment. Proceedings of the British Academy
48: 1–25.
Thomson, J.J. 2008. Normativity. La Salle: Open Court.
Turri, J. 2013. Knowledge and Suberogatory Assertion. Philosophical Studies 3:
1–11.
Wedgwood, R. 2007. The Nature of Normativity. Oxford: Oxford University
Press.
Williams, B. 1973. Problems of the Self. Cambridge: Cambridge University Press.
Williams, R. 2012. Indeterminacy and Normative Silence. Analysis 72 (2):
217–225.
Williamson, T. 2000. Knowledge and Its Limits. Oxford: Oxford University Press.
Wrenn, C. 2015. Truth. Cambridge: Polity Press.
Press.
Philosophy Supplementary 28: 31–74.
Wyatt, J. 2018. Absolutely Tasty: An Examination of Predicates of Personal
Taste and Faultless Disagreement. Inquiry 61 (3): 252–280.
Truth in English and Elsewhere:
An Empirically-Informed Functionalism
Jeremy Wyatt
We should reconcile ourselves with the fact that we are confronted, not with
one concept, but with several different concepts[.] [W]e should try to make
these concepts as clear as possible (by means of definition, or of an axiomatic
procedure, or in some other way); to avoid further confusions, we should
agree to use different terms for different concepts; and then we may proceed to
a quiet and systematic study of all concepts involved, which will exhibit their
main properties and mutual relations. (Tarski 1944, p. 355)
This chapter explores the future of functionalism about truth. I’ll aim to
defend functionalism while also explaining why and how functionalists
should rely on empirical evidence regarding ordinary thought about
truth. I’ll begin by sketching the functionalist framework and outlining
some of its signature virtues. Next, I’ll raise what I take to be the most
serious problem with functionalism—that although it entails empirical
claims regarding ordinary thought about truth, its main proponent
Michael Lynch has offered no empirical evidence in support of the view.
J. Wyatt (*)

https://doi.org/10.1007/978-3-319-98346-2_8
170 J. Wyatt
I’ll then venture to patch this lacuna by discussing some extant empiri-
cal data on ordinary thought about truth. These data, while inconclusive
and in need of supplementation, support two underinvestigated pluralist
views of ordinary thought about truth. Once we integrate these pluralist
approaches with the functionalist framework, we get a much more
nuanced view that is appropriately sensitive to relevant empirical
findings.
To close my defense, I’ll take a look at a second, pressing objection to
functionalism that has been advanced by Cory Wright (2010) and explain
why it doesn’t point to any serious flaws in the view. My overall conclu-
sion, then, will be that functionalism is a distinctive and promising
approach to the study of truth that can withstand philosophical scrutiny
and cleanly integrate empirical inquiry into our ongoing search for the
nature of truth.
1 Truth: Concept and Property

The ur-question of truth theory is, of course, ‘What is truth?’ In contem-
porary work on truth, a distinction is often drawn between the ordinary,
or folk concept truth and the property (or relation) truth that is alleg-
edly represented by this concept.1 The folk concept truth is meant to be
a concept that we employ prior to receiving any philosophical training
and that professional philosophers tend to use when not thinking about
philosophical issues pertaining to truth.2
William Alston (2002, p. 11) provides a paradigmatic motivation for
drawing the concept-property distinction in truth theory, noting that3:
[A] property might have various features not reflected in our concept of
that property. To choose a well worn example, heat (the property of a phys-
ical object’s being more or less hot) is revealed by physics to be an average
kinetic energy of constituent molecules, even though our ordinary concept
of heat involves no such component (that’s not the way we ordinarily iden-
tify heat).
Alston’s insight is that since truth is a folk concept, it may fail in

some respects to adequately represent truth. Analogies are especially help-
ful here, as Alston’s own analogy with heat shows. A similar and familiar
Truth in English and Elsewhere: An Empirically-Informed… 171
case is that of jade. We ordinarily conceive of jade as being a single, typi-

cally green kind of mineral that is, for example, often used in Asian art.
Yet while it is true that many familiar varieties of jade are green and that
jade is often used in Asian art, chemists have discovered that there are
actually two kinds of mineral—jadeite and nephrite—that we ordinarily
call ‘jade’. This means that although our folk concept jade does provide
a partially accurate representation of the nature of jade, there is more to
the nature of jade than what it represents.
As a result, when truth theorists set about their business, they actually
confront not the single, traditional question ‘What is truth?’ but rather
the following series of comparatively fine-grained questions:
(i) What is the nature of the folk concept truth?

(ii) What is the nature of the property truth that it allegedly represents?
(iii) How might inquiry about truth inform inquiry about truth?
2 Alethic Functionalism
The Framework
The most developed response to these questions comes from Michael

Lynch (2000, 2001, 2004a, b, 2005, 2009, 2013), who advocates func-
tionalism about truth, or alethic functionalism. Alethic functionalism is
inspired by the functionalist theories of mind and morality, respectively,
championed by David Lewis (1966, 1970, 1972, 1994) and Frank
Jackson (1998). Like these theories, alethic functionalism consists of two
stages that involve the method of Ramsification.4
The first stage of alethic functionalism is the Conceptual Stage. The aim
here is to offer an account of the folk concept truth. We begin by amass-
ing a set S of sentences that contain ‘true’ or ‘truth’ (or both). The sen-
tences in S must express propositions p1,…,pn that all and only the
possessors of truth are disposed to believe upon reflection. Following
Lynch, we can call p1,…,pn the folk theory FT of truth. It is worth empha-
sizing that according to the functionalist, possession of truth doesn’t
require that one explicitly believe p1,…,pn. What is required, rather, is a
172 J. Wyatt
disposition to believe p1,…,pn, provided that one has the concepts that
one needs to understand these propositions and has also reflected on the
propositions long enough to understand them.
To illustrate, Lynch takes FT to include the propositions expressed by
the following sentences5:
(Objectivity) A belief is true iff with respect to that belief, things are as
they are believed to be.
(Norm of Belief ) It is prima facie correct to hold a belief iff that belief is
true.
(End of Inquiry) Other things being equal, holding true beliefs is a wor-
thy goal of inquiry.
The second stage of alethic functionalism is the Metaphysical Stage. The

ambition here is to determine whether any actual properties have the
features that are ascribed to truth in FT. To clarify what these features are,
we construct the Ramsey sentence of FT.
The first step is to regiment the grammar of the sentences that pertain
to FT. In particular, we replace every instance of ‘true’ with ‘truth’ and
adjust the sentences’ grammar appropriately. We then conjoin them to
deliver the postulate PT of FT:
(PT) A belief exemplifies truth iff with respect to that belief, things are as
they are believed to be and it is prima facie correct to hold a belief iff
that belief exemplifies truth and other things being equal, holding
beliefs that exemplify truth is a worthy goal of inquiry.
Next, we replace every instance of ‘truth’ in PT with an objectual vari-

able y and we bind these variables with an existential quantifier. This
delivers the Ramsey sentence RT of FT, which has the following form:
(RT) There is a y (a belief exemplifies y iff with respect to that belief, things
are as they are believed to be and it is prima facie correct to hold a
belief iff that belief exemplifies y and other things being equal, holding
beliefs that exemplify y is a worthy goal of inquiry).
The final step is to determine whether any actual properties witness RT.
If we discover that exactly one property does so, then we say that the folk
theory FT is uniquely realized. In this case, we can simply call the realizing
property truth. Accordingly, call this possibility monism. However, the
functionalist also leaves room for two further possibilities. It may be that
many properties witness RT; in this case, we say that FT is multiply real-
ized, or that we’ve confirmed a certain sort of pluralism about truth. The
third possibility is that upon investigation, we find that no property wit-
nesses RT. In this case, we say that FT is unrealized, that is, that no actual
property is represented by the folk concept truth. Call this result defla-
tionism.6 Figure 1 summarizes the Conceptual and Metaphysical Stages
of functionalism.7
One attractive feature of functionalism is that it incorporates
Ramsification, which is a precise and well-understood method.8 It also
comports with Alston’s insight by allowing for the possibility that the
representation provided by the folk concept truth is partially accurate,
even as there is more to say about the metaphysics of truth than what is
articulated in this concept. If FT is uniquely or multiply realized, then the
properties that realize it may have additional features that aren’t men-
tioned in FT. On the other hand, if FT is unrealized, then that, too, will
probably come as a surprise to many possessors of the folk concept
truth, insofar as they presumably take some property in the world to
answer to their thought and talk about truth. This, then, would also be a
sense in which the truth about the nature of truth diverges from the rep-
resentation afforded by truth.
Conceptual Stage Metaphysical Stage
Identify the folk theory FT of truth Construct the Ramsey sentence RT of FT
Determine whether FT is uniquely realized, multiply

realized, or unrealized
Monism Pluralism Deflationism
Fig. 1 Alethic functionalism

174 J. Wyatt
In connection with this second virtue, I would emphasize a point that

is sometimes overlooked in discussions of alethic functionalism. This is
that functionalism itself is completely neutral regarding the metaphysics
of truth, apart from the constraints imposed by FT. In principle, inquiry
along functionalist lines could lead us to accept a monist, pluralist, or
deflationary account of truth’s nature, and it’s clear that these metaphys-
ics of truth are strikingly different. Rather than being regarded as a meta-
physics of truth, then, functionalism is best described as a method for
inquiring about truth that cleanly articulates a way in which conceptual
inquiry about truth can inform metaphysical inquiry about truth.9
Functionalism’s Empirical Debt
We’ve seen, then, that functionalism enjoys some significant virtues, and
this is all to the good. However, I want now to describe what I take to be
functionalism’s most serious shortcoming and to then explain how we
should remedy it.
The functionalist’s main ambition at the Conceptual Stage is to offer
an account of the folk concept truth. Their driving idea is that this con-
cept is underwritten by the folk theory FT of truth, and Lynch articulates
a view as to which propositions comprise FT. We should now ask: what
evidence do we have that FT in fact consists of the propositions expressed
by (Objectivity), (Norm of Belief ), and (End of Inquiry)?
Lynch offers two lines of evidence in support of his view on FT. One is
that he suspects that the propositions at issue are among our most funda-
mental (implicit) beliefs about truth.10 The other is that certain influen-
tial philosophers such as Aristotle, James, and Peirce have taken these
propositions to be central truths about truth.11
The basic problem is that these lines of evidence are inadequate in
principle. Lynch’s view about FT is empirical—it is a view about how
ordinary thinkers are, in fact, disposed to think about truth. However, we
have little reason to believe that philosophers enjoy special insight into
this issue. When we receive our professional training, we no doubt acquire
a range of new abilities, but we don’t tend to spend much time in gradu-
ate school systematically examining how untrained, native speakers think
about philosophically relevant issues. Accordingly, Lynch’s suspicion that

we tend to believe the propositions expressed by (Objectivity), (Norm of
Belief ), and (End of Inquiry) provides only very weak evidence that we
actually tend to do so.
Similarly, we have little reason to believe a priori that influential phi-
losophers’ views about truth will closely align with those of ordinary
thinkers. For this reason, the fact that Aristotle, James, and Peirce believed
the propositions at issue provides only very weak evidence that ordinary
thinkers are disposed to do the same. Functionalism, then, incurs an
empirical debt that Lynch’s lines of evidence simply fail to settle.
The good news, however, is that the functionalist can settle their
empirical debt, provided that they are willing to do the proper sort of
work. This work involves gathering empirical data on ordinary thought
about truth and then using this data at the Conceptual Stage. In light of
the results at the Conceptual Stage, the functionalist will then need to
re-examine the options that are available at the Metaphysical Stage.
In the remainder of the chapter, I want to outline some of the data that
we currently have on ordinary thought about truth and to indicate what
lessons they may hold. I’ll then explain how all of this affects the func-
tionalist framework. To anticipate, I take the main upshot for functional-
ism to be that once we appreciate the trajectory along which our current
data are leading us, we are able to glimpse some previously obscured
options at the Metaphysical Stage.
3 Data and Conceptual Pluralism

Interlinguistic and Intralinguistic Conceptual Pluralism
As we’ll see, our current data point toward a notable degree of complexity
in ordinary thought about truth. Accordingly, it will be useful to clearly
articulate a pluralist view of ordinary thought about truth that has two
main features. First, it should at least be supported by many of our data.
Second, such a view should be able to serve as a guiding hypothesis in
future studies of ordinary thought about truth.
176 J. Wyatt
The basic thesis behind what we’ll call conceptual pluralism about truth
is that there is more than one actual, folk truth concept. Notice, then,
that conceptual pluralism isn’t the claim that there could be thinkers who
use a truth concept that differs from the one used by, say, contemporary
English speakers. Rather, it is a claim about the actual world to the effect
that it contains thinkers some of whom use a different truth concept than
do others.
When we evaluate conceptual pluralism empirically, the following
basic procedure should serve us well.12 We begin by examining ordinary
speakers’ usage of ‘true’ and ‘truth’, as well as words in other languages
that can be directly translated using ‘true’ or ‘truth’. Call any such word a
piece of alethic vocabulary. If we find significant differences in speakers’
usage of alethic vocabulary, then we should go on to determine whether
some speakers associate different concepts with these words than others
do. Any concepts that we discover we’ll provisionally call truth
concepts.13
We can distinguish between two kinds of conceptual pluralism. The
first is what we’ll call interlinguistic conceptual pluralism:
(Inter) There are at least two actual linguistic communities L1 and L2 such
that some L1-members use a truth concept T1, whereas some L2-
members use a distinct truth concept T2.
Accordingly, if we discover two linguistic communities L1 and L2 such

that all L1-members use a truth concept T1 and all L2-members use a
distinct truth concept T2, this will verify interlinguistic conceptual plu-
ralism. However, the thing to notice is that we needn’t arrive at precisely
this discovery in order to show that interlinguistic conceptual pluralism
is true. Rather, all that is needed to verify this view is that there be a lin-
guistic community L1 some (but not necessarily all) of whose members
use a truth concept T1 in addition to another linguistic community L2
some (but not necessarily all) of whose members use a distinct truth con-
cept T2.
A second variety of conceptual pluralism is what we can call intralin-
guistic conceptual pluralism:
T1 T2 T2 T3
L1 L2
Fig. 2 Consistency of interlinguistic and intralinguistic pluralism
(Intra) There is at least one actual linguistic community L such that some
L-members use a truth concept T1, whereas other L-members use a
distinct truth concept T2.
We can make one preliminary observation about the logical relation-

ship between interlinguistic and intralinguistic pluralism—namely, that
they are mutually consistent. Figure 2 brings this fact out by detailing
two linguistic communities L1 and L2 that we might observe upon inves-
tigation. L1 and L2 individually confirm intralinguistic pluralism and
they jointly confirm interlinguistic pluralism.
In the rest of this section, I’ll be arguing that both intralinguistic and
interlinguistic pluralism enjoy notable support from our existing data. As
matters stand, the proper assessment is that these data are suggestive,
though quite far from conclusive. This means that further study is needed
before we can be very confident in any view of ordinary thought about
truth. However, what I hope to bring out is that both intralinguistic and
interlinguistic pluralism are very live theoretical options.
Intralinguistic Pluralism: Correspondence and Gender
In discussing intralinguistic pluralism, I’ll be drawing on a recent study

by Robert Barnard and Joseph Ulatowski (2013).14 Barnard and Ulatowski
sought to investigate the extent to which their subjects agreed with what
they call the correspondence root15:
178 J. Wyatt
(CR) Truth involves agreement, copying, or correspondence between

what (in the mind or in language) is true and what (in the mind or in
the world) makes it true.
They elicited responses from 200 English-speaking subjects to one of

nine vignettes. To assess their subjects’ agreement with (CR), Barnard
and Ulatowski presented them with a probe statement that is meant to be
similar to (CR) but framed in non-technical language:
(1) If a claim reports how the world is, then it is true.
It is worth pausing at this stage to note that there is a significant bug

in Barnard and Ulatowski’s experimental design. Insofar as it states that
truth involves agreement, copying, or correspondence, (CR) suggests that
correspondence is a necessary condition for truth. By contrast, the probe
statement (1) states that ‘reporting how the world is’ is a sufficient condi-
tion for truth. As a result, the reactions of Barnard and Ulatowski’s sub-
jects to (1) wouldn’t seem to provide evidence as to the attitudes that they
would take toward (CR). Rather, Barnard and Ulatowski’s data look to
provide insight into how their subjects would think about the converse of
(CR), which we might call the modified correspondence root:
(MCR) If a thought or claim copies, agrees with, or corresponds to what

(in the mind or in the world) would make it true, then that thought or
claim is true.
When evaluating Barnard and Ulatowski’s findings in what follows,

then, we will take their findings to concern (MCR), rather than (CR).
Future studies could profitably investigate subjects’ attitudes toward (CR)
by using the converse of (1), rather than (1), as a probe statement.
Two of Barnard and Ulatowski’s findings involve the Bruno case and
the Donna case16:
Bruno case: Bruno has just finished painting his house. Bruno painted his
house the same color as the sky on a clear summer day. Bruno claims
his house is blue.
Donna case: Donna is traveling in Germany, but does not speak German.
She watches as a sailor asks for der Stadtplan and is handed what looks
like a map of the city. Donna asks for der Stadtplan in a shop and is
sold a city map. Donna still speaks no German, but believes that ask-
ing for der Stadtplan is a good way to obtain a city map from a German
shopkeeper.
In the Bruno case, Bruno makes an ordinary, observational claim.

Similarly, in the Donna case, Donna forms a belief about an ordinary,
practical issue. Accordingly, Barnard and Ulatowski hypothesized that
these cases would elicit high degrees of agreement with (1) among their
subjects.17
However, what they found was that when presented with these cases,
male subjects expressed stronger agreement with (1) than female subjects.
In particular, when presented with the Donna case, male subjects tended
to agree with (1), whereas female subjects tended to disagree with (1).
Using a 5-point Likert scale with a midpoint of 3, Barnard and Ulatowski
found that the mean male and female responses to the Bruno case were
3.96 and 3.07, respectively. Likewise, the mean male and female responses
to the Donna case were 3.56 and 2.60, respectively.18 These differences
are statistically significant (Bruno: N = 41, t(39) = 2.305, p < 0.027;
Donna: N = 40 (25 male, 15 female), t(37.701) = 2.521, p < 0.016).
Pending further study of English speakers’ thought about truth, no
single interpretation of Barnard and Ulatowski’s findings is clearly supe-
rior to its competitors. However, it is clear that their findings lend some
support to intralinguistic pluralism. To explain their data along intralin-
guistic pluralist lines, we first hypothesize that there are at least two folk
truth concepts TCL and TCA. To possess TCL, one must be disposed to
accept (MCR) upon reflection (at least when thinking about certain top-
ics). To possess TCA, by contrast, one need not be disposed to do so.19
Thus, we might call TCL a correspondence-leaning concept and TCA a
correspondence-averse concept. The proffered explanation of Barnard and
Ulatowski’s findings would then be that among English speakers, males
are more likely than females to possess TCL, whereas females are more
likely than males to possess TCA.
180 J. Wyatt
Future studies should reevaluate this hypothesis by attempting to rep-

licate Barnard and Ulatowski’s results among English speakers. They
should also test similar hypotheses about other linguistic communities.
Additionally, I would note that in evaluating the intralinguistic pluralist
explanation of Barnard and Ulatowski’s findings, we should determine
whether there are plausible explanations as to why male English speakers
would be more likely to possess TCL, whereas female English speakers are
more likely to possess TCA.
Interlinguistic Pluralism: English and Akan
English Speakers on Correspondence
Regarding interlinguistic pluralism, the first thing to say is that to my

knowledge, there are no published, empirical studies that seek to evaluate
this hypothesis.20 However, in a series of pioneering papers, Kwasi Wiredu
(1985, 1987, 2004) brings to light two plausible conjectures that support
interlinguistic pluralism. Wiredu’s conjectures concern English and the
Ghanaian language Akan. Like Barnard and Ulatowski, Wiredu draws
our attention to English and Akan speakers’ respective views on truth and
correspondence. In doing so, he considers a claim that differs slightly
from both (CR) and (MCR), which we can call the correspondence thesis:
(CT) True propositions are those that correspond to facts.
Wiredu’s first conjecture concerns English speakers’ attitudes toward

(CT). Suppose that an ordinary English speaker were presented with the
following attempt at a definition of truth:
(2) True propositions are those that are true.
It’s reasonable to hypothesize that most such speakers would recognize

upon reflection that (2) is viciously circular, which would cause them to
refrain from agreeing that (2) is an accurate definition of truth. By con-
trast, Wiredu would conjecture that adult English speakers will tend to
evaluate (CT) differently, if asked to consider it as a definition of truth.

His first conjecture is21:
(WC1) Few (say, < 50% of ) native English speakers are disposed to believe
upon reflection that (CT) is viciously circular.
Notice, then, that (WC1) is not a conjecture about English speakers’

agreement or disagreement with (CT). Rather, it is a conjecture about the
percentage of English speakers who are disposed to believe upon reflec-
tion that (CT) is viciously circular. Presumably, if one did believe that
(CT) is viciously circular, then one would refrain from agreeing that it is
an accurate definition of truth. However, what is worth observing is that
an English speaker might refrain from believing that (CT) is viciously
circular, even though they don’t accept it as an accurate definition of
truth. This would be the position, for instance, of an English speaker who
favors an alternative (e.g. relativist) definition of truth but grants that
(CT) is a way that one might reasonably attempt to define truth. This
shows that (WC1) doesn’t entail that most English speakers are disposed
to believe upon reflection that (CT) is an accurate definition of truth
(this, of course, is a virtue of (WC1), given the data from Barnard and
Ulatowski (2013).
(WC1) awaits direct empirical confirmation. However, it is weakly
supported by two lines of empirical data, due respectively to David
Bourget and David Chalmers and Barnard and Ulatowski. Bourget and
Chalmers (2014) surveyed professional philosophers in 99 leading phi-
losophy departments, most of which (89) are located in English-speaking
countries.22 One of their survey questions was ‘Truth: correspondence,
deflationary, or epistemic?’ 50.8% of their respondents (± 1.5%) indi-
cated that they either accepted or leaned toward some sort of correspon-
dence theory of truth (ibid. pp. 477, 498). This means that either a sizable
minority or a slim majority of their respondents favored something like
(CT) and would thus presumably take (CT) to be free from vicious
circularity.
In the present context, Bourget and Chalmers’ results are admittedly
limited. One reason is that they didn’t explicitly use (CT) in their study.
Another is that their target group consisted of professional philosophers,
182 J. Wyatt
whereas (WC1) is a conjecture about native English speakers more gener-

ally. Despite these limitations, though, Bourget and Chalmers’ findings
provide weak support for (WC1).
Additionally, Barnard and Ulatowski sought in a recent study to exam-
ine the degree to which both professional philosophers and non-
philosophers take truth to be objective. In doing so, they examined their
subjects’ agreement and disagreement with various probe statements, one
of which is:
(3) When a claim is true, it expresses a fact.
Barnard and Ulatowski (2017, Tables 3 and 5) found that 86% of their
philosophers and 85.8% of their non-philosophers either agreed or
strongly agreed with (3). It’s thus plausible that most of these subjects
took (3) to be free from vicious circularity.
These findings are also admittedly limited in the present context, given
that (3)—in contrast to (CT)—doesn’t explicitly mention correspon-
dence and wasn’t explicitly presented as a (partial) definition of truth.
However, like Bourget and Chalmers’ findings, they also provide weak
support for (WC1). In light of the limitations of the existing evidence
that bears on (WC1), it would be especially helpful for future studies to
directly examine the degree to which English speakers take (CT) to be
viciously circular.
Akan Speakers on Correspondence
When exploring English and Akan speakers’ views on truth and corre-
spondence, Wiredu (1987, p. 29) also considers a direct translation of
(CT) into Akan:
(CTA) Asem no te saa kyerese ene nea ete saa di nsianim.
In (CTA), ‘true’ and ‘fact’ are translated using the same Akan expression
‘ete saa’, which means it is so. Thus, as Wiredu points out, (CTA) can also
be faithfully translated into English as ‘That a proposition is so amounts
to its coinciding with what is so’. This observation leads Wiredu to offer
the following insightful conjecture:
This has the beauty of a tautology, but it teaches little wisdom. It seems
to me unlikely that thinking in this language, one could be easily
tempted into correspondence formulations of this sort. Indeed, in this
language, it is pretty clear that the problem of truth must be the problem
of clarifying the idea of something being so. (Ibid. Cp. 1985, p. 47;
2004, pp. 48–9)
We can thus render Wiredu’s second conjecture along the following lines:
(WC2) Most (> 50% of ) native Akan speakers are disposed to believe
upon reflection that (CTA) is viciously circular.
Like (WC1), (WC2) awaits direct empirical confirmation. However, in

their own investigations of Akan alethic vocabulary, J.T. Bedu-Addo
(1985, pp. 70, 76–7) and Safro Kwame (2010) express sympathy with
Wiredu’s contention that ‘true’ and ‘fact’ should be translated using a
single Akan expression.23 We thus have strong preliminary evidence that
(CTA)’s vicious circularity will be apparent to native Akan speakers. As a
result, (WC2), like (WC1), is a plausible conjecture and should be empiri-
cally evaluated.
(WC1) and (WC2) lend support to interlinguistic pluralism, which can
explain them along the following lines. We first hypothesize that most
English speakers possess some truth concept T1 as well as a distinct
concept fact that is independent of T1 in the sense that it is definable
without reference to T1. They deploy these concepts when interpreting
(CT) and will accordingly take (CT) to be free from vicious circularity.
By contrast, it seems that when they reflect on (CTA), most Akan speakers
don’t deploy a truth concept as well as a distinct fact concept. Rather,
they deploy a single concept being so that doubles as a truth concept and
a fact concept. Since they deploy being so when interpreting (CTA), they
will accordingly take (CTA) to be viciously circular. It then follows that T1
≠ being so, which confirms interlinguistic pluralism. As a result, although
additional empirical work is needed to fully assess (WC1) and (WC2), it
is clear that interlinguistic pluralism promises a neat explanation of
Wiredu’s conjectures.
184 J. Wyatt
4 An Updated Functionalism

We’ve now seen that both intralinguistic and interlinguistic pluralism
enjoy a good deal of plausibility. For this reason, it will be interesting to
see how we can integrate them into the broader functionalist framework.24
In § 2, we saw that the standard functionalist approach consists of two
stages—the Conceptual and Metaphysical Stages. In essence, the
Conceptual and Metaphysical Stages now decompose into various sub-
stages, which increases the number of options that are available at the
Metaphysical Stage. Figure 3 offers a graphical depiction of this updated
functionalist framework, which I’ll now briefly describe.
Suppose that upon investigation, we discover that there are n actual,
folk truth concepts T1,…,Tn, where n > 1. At conceptual sub-stage c1, we
identify the folk theory FT1 of T1. We then move to the corresponding
metaphysical sub-stage m1 and construct the Ramsey sentence RT1 for T1.
Next, we determine whether FT1 is uniquely realized, multiply realized,
or unrealized. We apply this procedure at all conceptual and metaphysical
sub-stages c1,…,cn and m1,…,mn.
Conceptual Stages Metaphysical Stages

c1: Identify the folk theory FT1 m1: Construct the Ramsey sentence RT1 of FT1
of T1
. Determine whether FT1 is uniquely realized, multiply
. realized, or unrealized
. .
cn: Identify the folk theory FTn .
of Tn .
mn: Construct the Ramsey sentence RTn of FTn
Determine whether FTn is uniquely realized, multiply

realized, or unrealized
mf : Evaluate results at m1-mn
Monism Pluralism1 Pluralism2 Pluralism3 Deflationism
Fig. 3 Updated functionalism

Lastly, we want to evaluate the metaphysical results at m1,…,mn. Call

the metaphysical sub-stage that we now occupy the final (pending further
investigation) metaphysical sub-stage mf. Interestingly, at mf, we now
have five basic possibilities:
(a) Monism: All of the folk theories FT1,…,FTn are uniquely realized by
the same property P;
(b) Pluralism1: All of the folk theories are uniquely realized, but some are
realized by a different property than others;
(c) Pluralism2: Some, but not all, of the folk theories are multiply
realized;
(d) Pluralism3: All of the folk theories are multiply realized; and
(e) Deflationism: All of the folk theories are unrealized.
Options (a) and (e) are the most familiar, insofar as they respectively
represent monism and a pure sort of deflationism. We can helpfully think
of options (b)–(d) as representing increasingly complex grades of plural-
ism about the properties that realize the various folk theories. These
grades of pluralism mark the basic degrees to which multiple realization
can creep into our metaphysics of truth when we try to navigate the
divide between monism and deflationism.25
What I would emphasize at this stage is that before we endorse any of
these metaphysics of truth, we must do a great deal of further
metaphysical—and empirical—work. My hope is that the updated func-
tionalist framework will serve as a useful instrument for this work.
5 Functionalism and Epistemic Circularity

Wright’s Criticism
Thus far, my defense of alethic functionalism has mainly consisted of an

effort to outline how the functionalist should pay off their considerable
empirical debt. I want to wrap up this defense by looking at an objection
to functionalism that comes from Cory Wright. This objection deserves
extended reflection because it strikes at the conceptual bedrock of the
186 J. Wyatt
functionalist program—namely, the functionalist’s use of Ramsification.

Accordingly, if it succeeds, then all that I’ve said in defense of functional-
ism is essentially for naught.
While Wright would presumably allow that theorists such as Lewis
and Jackson can fruitfully use Ramsification to investigate the natures of
mental states or moral rightness, he takes there to be a distinctive prob-
lem with using Ramsification to inquire about the nature of truth. Wright
succinctly articulates his misgiving as follows26:
Functionalists about truth implicitly define their (nominalized) τ-term

true by using Ramsification to produce a sentence that indicates its denota-
tion. But any implicit definition proceeds on the basis of explicit decisions
that the principles constitutive of [the given theory] T are themselves true.
Hence the circularity. In turn, making any explicit decisions that they are
true requires already knowing in advance what truth is. Hence the epis-
temic circularity. If we suppose further that knowing in advance what truth
is entails knowing what the τ-name truth denotes, then it becomes unclear
why functionalists about truth ever required an implicit definition via
Ramsification in the first place. Hence the problem.
To make our discussion of Wright’s challenge more fluid, we can sim-

plify matters by supposing that the functionalist posits only one folk
truth concept truth. Given the data from § 3, it is not obvious that this
should be the functionalist’s actual stance. However, if the functionalist
endorses conceptual pluralism, then a version of Wright’s objection will
loom at each metaphysical sub-stage m1,…,mn, so the objection is press-
ing in either case.
Wright takes functionalism to suffer from what he calls ‘epistemic cir-
cularity’. He points out that when the functionalist details the postulate
PT at the Metaphysical Stage, it’s fair to think of them as attempting not
only to clarify the folk concept truth, but also to implicitly define the
word ‘truth’.27 The problem is supposed to arise from the alleged fact that
in offering this implicit definition, the functionalist is claiming that they
know that the conjuncts of PT are true. To know that the conjuncts of PT
are true, Wright insists, the functionalist must have prior knowledge of
the nature of the property truth.
The question then becomes: how is the functionalist supposed to know

what truth’s nature is? The functionalist aims to acquire knowledge of
truth’s nature by constructing PT, using PT to construct the Ramsey sen-
tence RT, and then determining which property witnesses RT. Yet if
Wright is correct, then when they follow this procedure, the functionalist
claims (implicitly) that they know that the conjuncts of PT are true.
The core of the problem, then, is this: knowledge that PT’s conjuncts
are true must, says Wright, be based on prior knowledge of truth’s
nature—which, if the functionalist is correct, must be based on prior
knowledge that PT’s conjuncts are true. It thus seems that in relying on
Ramsification, the functionalist is claiming to have knowledge—that PT’s
conjuncts are true—that they simply can’t acquire, given the strictures of
their own view. To close, I’d like to mount a direct defense of functional-
ism against Wright’s worry. That is, I’ll aim to explain why functionalism
isn’t viciously circular in the way that he suggests.28
Why Functionalism Isn’t Epistemically Circular
For the sake of argument, I will suppose with Wright that the functional-
ist takes all of PT’s conjuncts to be true.29 We then pose Wright’s problem:
to know that PT is true, mustn’t the functionalist already know the nature
of the property that ‘truth’ denotes, that is, truth? But isn’t their view that
our knowledge of truth’s nature is based on our prior knowledge of PT? If
so, then it seems that to come to know that PT is true, they must already
know that PT is true, which is impossible.
The functionalist should dispel this impression of circularity by point-
ing out that they can know that PT is true without already knowing the
nature of truth. This follows from a more general principle. Assuming
that A is a competent user of the word ‘true:’30
(4) If A knows that p and they know that the content of sentence ‘S’ is
that p, then A knows that ‘S’ is true.
To see the plausibility of (4), suppose that a high school student Van
looks at the whiteboard in his classroom and sees the sentence ‘The chem-
istry exam will begin at 1:30 p.m. today’. Van knows that the content of
188 J. Wyatt
this sentence is that the chemistry exam will begin at 1:30 p.m. on that
day. If he also knows that the exam will in fact begin at 1:30 p.m. on that
day, then he knows that the sentence on the whiteboard is true. It’s sim-
ply not the case that to know that the sentence is true, Van must know
what the nature of truth is. It’s enough that he’s a competent user of the
word ‘true’ and thus understands how this word applies to English
sentences.
Let’s now think in the same vein about PT. Suppose that PT is the con-
junction ‘S1 and S2 and S3’ and that the functionalist maintains that PT is
true. We ask them: how do you know that PT is true? To explain this, the
functionalist needs only point out (i) that the contents of its conjuncts
are respectively that S1, that S2, and that S3 and (ii) that they know that
that S1, that S2, and that S3. From (i), (ii), and (4), it follows that they
know that PT is true. Consequently, it’s simply not the case that to acquire
knowledge of PT, the functionalist must have antecedent metaphysical
knowledge of truth’s nature. Like Van, what they need is an understand-
ing of how ‘true’ applies to English sentences. They acquire this under-
standing simply in virtue of being a competent, mature English
speaker—and thus prior to embarking on their analyses of truth and
truth.31
6 Conclusions
In this chapter, I hope to have conveyed why alethic functionalism prom-
ises to be a powerful tool for theorists studying the nature of truth.
Functionalism delivers an insightful treatment of the concept-property
distinction in truth theory and in doing so, sets our sights on the con-
spicuous value of empirically studying ordinary thought about truth.
Despite considerable advances by experimental truth theorists, we’ve
managed only to break ground on this avenue of inquiry. This means that
we’ll need to expend a great deal of effort to see it to fruition—and in
doing so, we should direct a sizable portion of our effort to the evaluation
of conceptual pluralism. What awaits us as we pursue this line of inquiry
is a sharper understanding of ordinary thought about truth, the nature of
truth, and the interconnections between these topics. Whether the final
metaphysics of truth is monist, pluralist, or deflationary, we will be able

to confidently say that it is the product of clear-eyed metaphysical inves-
tigation disciplined by painstaking empirical research.32
Notes
1. In what follows, I’ll use italics to denote properties and relations and
small caps to denote concepts.
2. Cf. Lynch (2009, p. 7).
3. For further illuminating reflections on this distinction, see Asay (2013,
Chap. 1; 2018); Bar-On and Simmons (2007); Eklund (2017, § 2); and
Lynch (2005, 2009, Chap. 1).
4. For a canonical account of Ramsification, see Lewis (1970).
5. (2009, Chap. 1). I’ve altered the wording of the sentences, though this
won’t influence the concern that I’ll develop for functionalism in § 2.
In making these alterations, I’ve been guided by the assiduous study of
Lynch’s functionalism by Marian David (2013).
6. More specifically, this metaphysics of truth amounts to what Künne
(2003, pp. 3–4, Chap. 2) calls “nihilism,” or what I prefer to call pure
deflationism. Not all deflationists (e.g. Horwich 1998) are pure defla-
tionists in this sense, but some (e.g. Quine (1948, 1970) and Strawson
(1949, 1950)) have endorsed or would endorse the view. According to
the taxonomy here, deflationists such as Horwich who commit to the
existence of a single property truth should be classified as monists. Thus
to fully flesh things out, we’d need to distinguish between deflationary
monists and substantivist monists (and between deflationary and sub-
stantivist pluralists). Doing so requires a good bit of effort; for details, see
Wyatt (2016).
7. As I’m using the expressions ‘pluralism about truth’ and ‘truth pluralism’,
one may endorse what is often called ‘alethic pluralism’ without thereby
endorsing truth pluralism. I take truth pluralism to be the view that
more than one actual property realizes FT. In other words, truth plural-
ism is the view that more than one actual property fits the job descrip-
tion for truth—full stop, and hence not merely relative to some domain
or other (on this distinction, see David (2013) and Edwards (2011,
pp. 34–7)). Extant alethic pluralists—strong and moderate alike—thus
stop short of endorsing truth pluralism. Strong alethic pluralists (e.g.
190 J. Wyatt
Cotnoir (2013) and Kim and Pedersen, this volume) take no actual
property to realize FT simpliciter and moderate alethic pluralists (e.g.
Edwards (2011) and Lynch (2009)) take exactly one such property to do
so. This, I would stress, isn’t a criticism of these views, only an observa-
tion about their structure.
It’s also worth noting that truth pluralism, in the present sense, is
structurally similar to Beall and Restall’s logical pluralism, insofar as the
properties that would realize FT would do so with respect to all domains.
8. In his most recent work (2009, 2013), Lynch has elected to not rely
explicitly on Ramsification, thereby departing from the strategy in his
earlier, pathbreaking work on functionalism (2000, 2001, 2004b, 2005).
However, I would point out that in his recent work, Lynch still speaks of
properties playing the ‘truth role’. The truth role is supplied by PT and
playing it looks to be nothing over and above witnessing RT, so it seems
that Lynch is still loyal at heart to Ramsification, even though he chooses
not to wear the badge.
9. Lynch (2009, p. 84; 2013, p. 27) and Wright (2005, n. 14) also allude
to this point (see also Devlin (2003)). However, other authors are a bit
too quick in assimilating functionalism to pluralism. I have in mind here
the otherwise illuminating critical discussions by Caputo (2012); Horton
and Poston (2012); Newhard (2013, 2014, 2017) and Wright (2005, §§
4.1, 4.2). What I would point out is that the criticisms advanced by
these authors affect the functionalist only if they commit to some sort of
alethic pluralism—a move that Lynch does make, but which is neverthe-
less entirely optional.
10. Lynch (2004a, Chap. 1; 2009, p. 8).
11. Lynch (2009, pp. 10–14; 2013, p. 24).
12. Cp. Barnard and Ulatowski (2013, 2017); Fisher, et al. (2017); Kölbel
(2008); and Mizumoto (ms). See especially Ulatowski (2017, pp. viii, ix,
2, 8–9).
13. I say ‘provisionally’ because it is at this point that we should look for
underlying similarities among the concepts that speakers deploy when
they use alethic vocabulary. If we detect such similarities, we should
then—and only then—advance a general account of what makes a con-
cept a truth concept. To do otherwise would amount to gratuitous the-
ory building in the absence of sufficient data.
14. Moltmann’s investigations (2015, 2018) into truth predicates, and
related kinds of predicate, in various natural languages also look to bear
significantly on intralinguistic pluralism. The same goes for the work of
Dzobo (1992, pp. 79–83) and Wiredu (1985, 1987, 2004) on the
Ghanaian languages Ewe and Akan.
15. 2013, p. 621.
16. Ibid. p. 631.
17. Ibid. p. 633.
18. Ibid. Figs. 2 and 3.
19. I include the parenthetical qualification to flag a further sort of variance
that Barnard and Ulatowski (ibid. Fig. 1) found among their subjects.
They found that their subjects (male and female alike) were more likely
overall to agree with (1) when presented with the Bruno case than when
presented with a case involving a simple arithmetical calculation. We
might call this kind of variance topic-sensitivity. What I would point out
is that this topic-sensitivity is independent of intralinguistic pluralism,
insofar as the former may be present across the community of English
speakers. For further investigation of topic-sensitivity, see Ulatowski
(2017).
20. That said, Mizumoto (ms) has gathered interesting data pertaining to
Japanese and English that do look to support interlinguistic pluralism. I
should also note the fascinating collection of papers surveyed by Maffie
(2001). By contrast, Matthewson and Glougie (forthcoming) investigate
interesting cross-linguistic uniformities in the use of alethic vocabulary.
21. I take (WC1) to be a conjecture that Wiredu would be willing to make,
although he doesn’t explicitly advance it. He comes extremely close to
doing so at 2004, pp. 47–9. Cp. 1985, pp. 47–8, 49–50; 1987, p. 28.
22. Bourget and Chalmers (2014, § 2) note that in principle, anyone was
allowed to take their survey, though the target group about which they
mainly report are the professional philosophers from the mentioned
departments.
23. Bedu-Addo argues that the best Akan expression to use is ‘nokware’, a
view with which Kwame looks to be sympathetic.
24. My thinking here has been influenced by some suggestive remarks due to
Wright (2005, pp. 18–21).
25. The grades of pluralism in (b)–(d) are notably different from the main-
stream pluralist truth theories that have been developed thus far. One
major difference is that the pluralisms in (b)–(d) make no reference to
the notion of a ‘domain’, which figures prominently in mainstream plu-
ralist theories. Rather, these pluralist views integrate Ramsification with
conceptual pluralism, an approach that hasn’t been attempted by main-
stream alethic pluralists.
192 J. Wyatt
26. (2010, p. 8). Cp. (2005, pp. 21–2).

27. Cp. Lewis (1970, p. 429).
28. In this way, my reply to Wright differs from that of Lynch (2013, n. 12),
who grants that functionalism is viciously circular, but insists (less than
convincingly, to my mind) that every other theory of truth is viciously
circular in the same respect.
I should also note that my reply on behalf of functionalism differs
from the reply that Wright (2010, § 6) finds most convincing. I lack the
space to examine this reply, but my basic concern is that it rests on a
mischaracterization of the functionalist’s ambitions.
29. Admittedly, it seems to me that the functionalist need not do so. By way
of analogy, think of a non-racist psychologist studying the race-related
concepts used by ordinary, racist subjects. It would be consistent for
them to analyze these concepts along functionalist lines while refraining
from commitment to the truth of their postulates. It would seem that
the alethic functionalist enjoys a similar sort of freedom when analyzing
ordinary thought about truth.
30. (4) is grounded in (one direction of ) what is sometimes called the trans-
parency of the word ‘true’, as that word applies to sentences. It should
also be helpful to remember the importance of the concept-property
distinction in truth theory, as discussed in §1.
31. There is admittedly a further issue that must be explored—namely,
how the functionalist could know that S1, that S2, and that S3. Though
I can’t pursue this issue at length here, it’s plausible that the functional-
ist can acquire this knowledge in a rather familiar way. They can do so
by reflecting on cases that pertain e.g. to the objectivity of truth or to
the connections between A’s belief being true and its being correct for
A to hold that belief. Put a bit more generally, the functionalist could
come to know that PT is true by competently deploying the truth con-
cept for which PT is the postulate in reflecting on cases that pertain to
PT’s conjuncts.
32. I am grateful to audiences at the University of Bologna, Hong Kong
University, Texas Christian University, and Yonsei University for their
feedback on this chapter. Those who have helped me in developing the
chapter include Wes Cray, Filippo Ferrari, Max Deutsch, Richard Galvin,
Will Gamester, Patrick Greenough, John Harris, Blake Hestir, Ole
Hjortland, Michael Lynch, Kelly McCormick, Sebastiano Moruzzi,
Shyam Nair, Francesco Orilia, Giorgio Volpe, and Crispin Wright. I owe
particular thanks to Jamin Asay, Teresa Kouri, Nikolaj J. L. L. Pedersen,

Joe Ulatowski, and Cory Wright. Also, while working on this paper, I
received support from the National Research Foundation of Korea (NRF
2013S1A2A2035514 and 2016S1A2A2911800). This support is grate-
fully acknowledged.
References
Alston, W. 2002. Truth: Concept and Property. In What Is Truth? ed. Schantz,
11–26. New York/Berlin: de Gruyter.
Asay, J. 2013. The Primitivist Theory of Truth. Cambridge: Cambridge University
Press.
———. 2018. truth: A Concept Unlike Any Other. Synthese Special Issue
Truth: Concept Meets Property, Wyatt, ed. https://doi.org/10.1007/
s11229-017-1661-z.
Barnard, R., and J. Ulatowski. 2013. Truth, Correspondence, and Gender.
Review of Philosophy and Psychology 4 (4): 621–638.
———. 2017. The Objectivity of Truth, a Core Truism? Synthese Special Issue
Truth: Concept Meets Property. https://doi.org/10.1007/s11229-017-1605-7.
Bar-On, D., and K. Simmons. 2007. The Use of Force Against Deflationism:
Assertion and Truth. In Truth and Speech Acts: Studies in the Philosophy of
Language, ed. Greimann and Siegwart, 61–89. London: Routledge.
Bedu-Addo, J.T. 1985. Wiredu on Truth as Opinion and the Akan Language. In
Philosophy in Africa: Trends and Perspectives, ed. Bodunrin, 68–90. Ile-Ife:
University of Ife Press.
Bourget, D., and D. Chalmers. 2014. What Do Philosophers Believe?
Caputo, S. 2012. Three Dilemmas for Alethic Functionalism. The Philosophical
Quarterly 62 (249): 853–861.
Cotnoir, A. 2013. Validity for Strong Pluralists. Philosophy and Phenomenological
Research 83: 563–579.
Devlin, J. 2003. An Argument for an Error Theory of Truth. Philosophical
Perspectives 17 (1): 51–82.
194 J. Wyatt
Dzobo, N.K. 1992. Knowledge and Truth: Ewe and Akan Conceptions. In
Person and Community: Ghanaian Philosophical Studies I, ed. Wiredu and
Gyeke, 73–84. Washington, DC: The Council for Research in Values and
Philosophy.
Edwards, D. 2011. Simplifying Alethic Pluralism. Southern Journal of Philosophy
49 (1): 28–48.
Eklund, M. 2017. What Is Deflationism About Truth? Synthese Special Issue
Truth: Concept Meets Property, Wyatt, ed. https://doi.org/10.1007/
s11229-017-1557-y.
Fisher, M., J. Knobe, B. Strickland, and F. Keil. 2017. The Influence of Social
Interaction on Intuitions of Objectivity and Subjectivity. Cognitive Science
41: 1119–1134.
Horton, M., and T. Poston. 2012. Functionalism About Truth and the
Metaphysics of Reduction. Acta Analytica 27 (1): 13–27.
Horwich, P. 1998. Truth. 2nd ed. Oxford: Oxford University Press.
Jackson, F. 1998. From Metaphysics to Ethics: A Defence of Conceptual Analysis.
Research 77 (2): 359–384.
Künne, W. 2003. Conceptions of Truth. Oxford: Oxford University Press.
Kwame, S. 2010. Nokware. In The Oxford Encyclopedia of African Thought, ed.
F. Abiola Irele and B. Jeyifo. New York: Oxford University Press. Accessed at
http://www.oxfordaasc.com/article/opr/t301/e275.
Lewis, D. 1966. An Argument for the Identity Theory. The Journal of Philosophy
63 (1): 17–25.
———. 1970. How to Define Theoretical Terms. The Journal of Philosophy 67:
427–446.
———. 1972. Psychophysical and Theoretical Identifications. Australasian
———. 1994. Reduction of Mind. In Companion to the Philosophy of Mind, ed.
Guttenplan, 412–431. Malden: Blackwell.
Lynch, M. 2000. Alethic Pluralism and the Functionalist Theory of Truth. Acta
Analytica 24: 195–214.
———. 2001. A Functionalist Theory of Truth. In Truth and Pluralism: Current
———. 2004a. True to Life. Cambridge, MA: MIT Press.
———. 2004b. Truth and Multiple Realizability. Australasian Journal of
———. 2005. Alethic Functionalism and Our Folk Theory of Truth. Synthese
145: 29–43.
———. 2009. Truth As One and Many. Oxford: Oxford University Press.
———. 2013. Three Questions for Truth Pluralism. In Truth and Pluralism:
Maffie, J. 2001. Truth from the Perspective of Comparative World Philosophy.
Social Epistemology 15 (4): 263–273.
Matthewson, L., and J. Glougie. Forthcoming. Justification and Truth: Evidence
from Languages of the World. In Epistemology for the Rest of the World, ed.
Stich, Mizumoto, and McCready. Oxford: Oxford University Press.
Mizumoto, M. (ms) A Prolegomenon to the Cross-Linguistic Study of Truth.
Moltmann, F. 2015. ‘Truth Predicates’ in Natural Language. In Unifying the
Philosophy of Truth, ed. Achourioti et al., 57–83. Dordrecht: Springer.
———. 2018. Truth Predicates, Truth Bearers, and Their Variants. Synthese
Special Issue Truth: Concept Meets Property. https://doi.org/10.1007/
s11229-018-1814-8.
Newhard, J. 2013. Four Objections to Alethic Functionalism. Journal of
———. 2014. Alethic Functionalism, Manifestation, and the Nature of Truth.
Acta Analytica 29 (3): 349–361.
———. 2017. Plain Truth and the Incoherence of Alethic Functionalism.
Synthese 194 (5): 1591–1611.
Quine, W.V.O. 1948. On What There Is. Review of Metaphysics 2 (5): 21–36.
Reprinted in From a Logical Point of View, 2nd Revised ed. Quine, 1–19.
New York: Harper and Row.
———. 1970. Philosophy of Logic. 2nd ed. Englewood Cliffs: Prentice Hall.
Strawson, P.F. 1949. Truth. Analysis 9 (6): 83–97.
———. 1950. Truth. Proceedings of the Aristotelian Society Supplementary 24:
129–156.
Tarski, A. 1944. The Semantic Conception of Truth: And the Foundations of
Semantics. Philosophy and Phenomenological Research 4 (3): 341–376.
Ulatowski, J. 2017. Commonsense Pluralism About Truth: An Empirical Defence.
Cham: Palgrave Macmillan.
Wiredu, K. 1985. The Concept of Truth in the Akan Language. In Philosophy in
Africa: Trends and Perspectives, ed. Bodunrin, 43–54. Ile-Ife: University of Ife
Press.
———. 1987. Truth: The Correspondence Theory of Judgment. African
Philosophical Inquiry 1 (1): 19–30.
196 J. Wyatt
———. 2004. Truth and an African Language. In African Philosophy: New and
Traditional Perspectives, ed. Brown, 35–50. Oxford: Oxford University Press.
Synthese 145 (1): 1–28.
Wright, C. 2010. Truth, Ramsification, and the Pluralist’s Revenge. Australasian
Wyatt, J. 2016. The Many (Yet Few) Faces of Deflationism. The Philosophical
Quarterly 66 (263): 362–382.
Part II
Logic
Core Logic: A Conspectus
Neil Tennant
1 Introduction
As the title of this study implies, there is a much longer and more detailed
treatment of Core Logic, to be found in Tennant (2017). Here we simply
try to explain and summarize the main points and their relation to debates
about monism and pluralism.
The title of the Storrs conference in 2015 of which this volume is the
proceedings was ‘Pluralism in Truth and Pluralism in Logic’. But the
prevailing intellectual climate is so pluralist that the singulars in the con-
ference title have metastasized so as to occasion, in the title of this vol-
ume, the plural noun ‘Pluralisms’. In this heady atmosphere, the present
author presents a more conservative, but still accommodating, ‘absolutist
pluralism’ about logic—albeit a reformist one. He believes there is such a
thing as ‘core logical’ deductive reasoning, and that it is relevant, in a
sense heretofore not satisfactorily explicated. It comes, however, in two
N. Tennant (*)
Department of Philosophy, Ohio State University, Columbus, OH, USA
e-mail: tennant.9@osu.edu

https://doi.org/10.1007/978-3-319-98346-2_9
200 N. Tennant
varieties: constructive and non-constructive, as typified most importantly

in mathematics, where mathematical theorems are deduced from math-
ematical axioms.
Locating the Core Systems ℂ and ℂ+
In the present author’s view, the decision tree for choice of logic, for an
unmitigated absolutist, looks like this1:
Is there one correct logic?

Absolutist: Yes Pluralist: No
Is it Classical Logic?
Quietist: Yes Reformist: No
Is it relevant?
Yes No
Is it constructive?
Yes No Yes No
C Core Classical I Other

Core
The formal-logical reforms that need to be carried out on Classical Logic

C and Intuitionistic Logic I have to do with relevantizing formal proofs.
The ‘absolutist pluralism’ adverted to above differs from this unmitigated
absolutism by making the following concessive recommendations:
So you’re a constructive mathematician? Then the right logic for you to use
is Core Logic ℂ; or …
So you’re a non-constructive, i.e. classical mathematician? Then the right
logic for you to use is Classical Core Logic ℂ+.
The relevantizing effected within these two core systems is non-negotiable.

Classical Core Logic ℂ+ is simply the ‘classicized extension’ of Core Logic ℂ.
Core Logic: A Conspectus 201
From now on, when we write ‘Core logician’, we intend this to be

understood disjunctively, as ‘user of Core Logic for constructive reason-
ing, or user of Classical Core Logic for non-constructive reasoning’. This
disjunctiveness is our concession to pluralism; while the commitment to
relevance remains absolute.
On Being Conservative in One’s Aim
This author is concerned (as Gentzen was) to complete satisfactorily

Frege’s original task of regimenting the expert deductive reasoning of
mathematicians in as natural a way as possible. And all such reasoning is
relevant. Gentzen’s aim with his systems of natural deduction was clear:
Wir wollen einen Formalismus aufstellen, der möglichst genau das wirk
liche logische Schliessen bei mathematischen Beweisen wiedergibt.
(Gentzen 1934, p. 183)
This was translated by M. E. Szabo as
We wish to set up a formalism that reflects as accurately as possible the

actual logical reasoning involved in mathematical proofs.
The difference between the Core logician’s approach and that of Gentzen
is that the Core logician takes seriously the fact that such reasoning
(whether or not it is constructive) is always fully relevant, in the way it
proceeds from premises (say, mathematical axioms) to conclusions (say,
mathematical theorems).2 One’s logical system should furnish formal
proofs that fill in all the gaps that occur in the typically informal proofs
that mathematicians provide, even when they are being (by their own
disciplinary standards) very rigorous. Formal logicians have more
demanding disciplinary standards. Yet their formal proofs should just be
appropriately more detailed homologues of mathematicians’ informal
proofs. Note that this observation applies both to the constructive math-
ematical reasoner and to the non-constructive, or classical one.
So much for the conservativeness of the Core Logician’s formalizing
project.
202 N. Tennant
On Being Accommodating in One’s Aim
The Core Logician is accommodating in the following important regard.

Since the time of Frege, intuitionistic or constructive mathematical rea-
soning has achieved a special systemic status, through the work of Heyting
following Brouwer, resulting in the system I of Intuitionistic Logic. So
there should be a formal-logical account of constructive mathematical rea-
soning conducted relevantly, and a more expansive formal-logical account
of non-constructive (classical) mathematical reasoning, likewise conducted
relevantly. These accounts are distilled, respectively, as the systems of
Core Logic ℂ and Classical Core Logic ℂ+. The core systems ℂ and ℂ+
respectively relevantize the two more familiar and traditional systems I
and C. And they do so in the same way, ‘at the level of the turnstile’.
What Is Wrong with R and Its Ilk
By contrast, the relevance logics (such as R) in the Anderson-Belnap tra-

dition (see Anderson and Nuel D. Belnap, Jr. (1975)) seek to relevantize
at the level of the object-language conditional, thereby changing its Fregean
sense. In relevant logics such as R, the following inferences fail, despite
the fact that they directly render the Fregean stipulation that a condi-
tional is true when its antecedent is false, and is also true when its conse-
quent is true:
Øj y
j ®y j ®y
In these relevant logics, one is also deprived of Disjunctive Syllogism:
j Ú y Øj
y
despite its ubiquity in mathematical reasoning (e.g. about ordering rela-

tions). There is no known metatheorem to the effect that ordinary math-
ematical reasoning can be fully formalized in any system like R. Actually,
there have been negative results in this regard—see Friedman and Meyer
(1992). This is in stark contrast to the situation with ℂ (for constructive
mathematical reasoning) and ℂ+ (for non-constructive mathematical
reasoning).
The Uniformity of Relevantizing
The right logic for constructivist reasoning must bear to the right logic
for non-constructivist reasoning the same sort of relation (represented by
the solid horizontal arrows below) that Intuitionistic Logic I bears to
Classical Logic C.
Likewise, the right logic for constructivist reasoning should bear to
Intuitionistic Logic I the same sort of relation (represented by the dashed
vertical arrows below) that the right logic for non-constructivist reason-
ing bears to Classical Logic C.
Our claim is that ℂ is the right logic for constructivist reasoning, and
ℂ+ is the right logic for non-constructivist reasoning.
Note that Core Logic contains the aforementioned Disjunctive
Syllogism (φ ∨ ψ, ¬φ : ψ) and Fregean truth-table respecting infer-
ences (¬φ : φ → ψ and ψ : φ → ψ); but it does not contain either one
of the two closely related Lewis paradoxes φ, ¬φ : ψ and φ, ¬φ : ¬ψ (see
Figs. 1 and 2).
I C
Fig. 1 From classical logic to core logic
204 N. Tennant
Fig. 2 Important system containments
What Is Wrong with M
In addition to the four systems already mentioned (C, I, ℂ, and ℂ+), the
reader might wonder how Johansson’s system M of Minimal Logic fits
into the picture. Our contention is that M was intended by Johansson to
be a relevantized system of constructive reasoning. And in this project, it
failed, because, although it avoids the positive form of the first Lewis
paradox (φ, ¬φ : ψ), it nevertheless contains the negative form (φ, ¬φ : ¬ψ).
Moreover, M also fails to validate Disjunctive Syllogism. But the rel-
evant validity of Disjunctive Syllogism is unimpeachable; and it is
indispensable to mathematical reasoning (consider: b<a ∨ a≤b, b ≮ a:
a ≤ b).
Figure 2 summarizes these and other well-known differentiae among
the five systems under discussion.
In formulating the core systems ℂ and ℂ+, we take as our point of
departure, but significantly modify and thereby improve upon, Gentzen’s
systems of natural deduction (and the corresponding sequent calculi) for
I and C. The system M, as we have just stressed, is an aberration, consist-
ing as it does of sub-optimally formulated introduction and elimination
rules for the logical operators. Those rules have to be fundamentally re-
thought, and particular ones importantly modified, in order to ensure
relevance even after eschewing the infamous Absurdity Rule
^
j
It is the presence, or absence, of the Absurdity Rule—given Gentzen’s

original formulation of the introduction and elimination rules for the
logical operators—that respectively yields I, or M. Johansson’s system M
results from simply taking Gentzen’s rules for I and removing the
Absurdity Rule.
n an Isomorphism Between Sequent Proofs and Natural

O
Deductions
A remarkable feature of the natural deduction and sequent systems for

the core systems ℂ and ℂ+ is that their natural deductions are ‘isomor-
phic’ to their corresponding sequent proofs (once one has snipped off the
‘twigs’ to those leaf-nodes in natural deductions that are occupied by any
major premises of eliminations). In a rigorously definable sense, their
trees of rule-applications correspond 1-1 (eliminations to left rules, intro-
ductions to right rules). The only real difference is that nodes of a proof-
tree are labelled by single sentences in natural deductions but by the
whole sequents thus far proved in the corresponding sequent proofs. In
this conspectus, for ease of presentation, we shall focus on just natural
deductions and the rules of inference by means of which they are
constructed.
2 F amiliar Rules That Are Not in Core

Logic ℂ
Strictly Classical Negation Rules
Classical Reductio ad Absurdum (CR):

206 N. Tennant
(i )
Øj

^
(i )
j
Law of Excluded Middle (LEM):
j Ú Øj
Dilemma (Dil):
(i ) (i )
j Øj
 
y y
(i )
y
Double Negation Elimination (DNE):
ØØj
j
Ex Falso Quodlibet (EFQ), a.k.a. the Absurdity Rule

^
j
3 Graphic Rules for Core Logic ℂ

Diamonds [resp., boxes] on discharge strokes indicate that vacuous dis-
charge is permitted [resp., prohibited].
208 N. Tennant
Rules for Classical Core Logic ℂ+
In order to obtain Classical Core Logic from Core Logic, it suffices to

add either Classical Reductio or Dilemma. These two strictly classical
rules are interderivable in Core Logic. In each of these rules, it is the
sentence φ that is its ‘classical focus’. This is because the reasoner who
applies the rule is presuming that φ is determinately true, or false, even
though it might not be known (or indeed, even knowable) which is the
case. (See Tennant (1997) for extended philosophical discussion of this
point.)
(i )
Øj
( CR )
^
(i )
j
(i ) (i )
j Øj

y y
( Dil ) (i )
y
(i ) (i )
j Øj

y ^
(i )
y
Here are some important points to note about the core systems:
1. Major premises for eliminations (MPEs) stand proud. That is, they do
not stand as conclusions of non-trivial proof-work above them.
2. So, all core proofs are in normal form. That is, no core proof contains
any sentence-occurrence standing both as an MPE and as the conclu-
sion of an application of any other rule, be it an elimination, or an
introduction, or a classical negation rule.3
3. Vacuous discharge of assumptions is prohibited where need be.
4. The rules of →-I and ∨-E are liberalized. That is, these rules are formu-
lated in such a way as to obviate the need (in the Gentzen-Prawitz
formulations) to resort to strategic applications of EFQ before they
can be applied.
5. The rule Ex Falso Quodlibet is banned.
6. So, all proofs are relevant—in a very exigent sense about to be

explained.
210 N. Tennant
4 he Special Kind of Relevance Ensured

T
by Core Proofs
In order to expand on point (6), we need some tailor-made syntactic
notions that will allow us to explicate the idea that the set Δ of premises
of a (classical) core proof is relevantly connected both within itself and to its
conclusion φ.
In the propositional case, we can limit ourselves to speaking of atoms,
for these are the only non-logical expressions occurring in a sentence. But
the treatment of ‘vocabulary sharing’ to be expounded briefly here extends
to first-order logic, with predicates taking over from atoms (or being con-
sidered in addition to atoms, if atoms remain in the picture). We assume
the reader is familiar with the notions of positive and negative occur-
rences of subformulae within formulae. The status of positive vs. negative
is called the polarity of the subformula-occurrence in question.
We shall say φ ≈ ψ just in case some atom enjoys an occurrence in φ
and an occurrence in ψ of the same polarity. And we shall say φ ⋈ ψ just
in case some atom enjoys an occurrence in φ and an occurrence in ψ of
opposite polarities. A set of sentences is ⋈-connected just in case any two
distinct members of it are connected by a ⋈-path within it. Clearly any
set of sentences admits of a unique partition into inclusion-maximal
⋈-connected sets, which we shall call its ⋈-components. (Note that for
any θ ∈ Δ, if θ does not bear ⋈ to any other member of Δ, then the
singleton {θ} is a ⋈-component of Δ.)
The expression ±φ means that some atom occurs both positively and
negatively in φ. The expression ♮Δ means that if Δ is a singleton {φ},
then ±φ; otherwise, Δ is ⋈-connected. The expression φ◂Δ means that
every ⋈-component of Δ contains some sentence ψ such that φ ≈ ψ.
Point (6) involves a very exigent relevance condition  ( D, j ) on the
premise set Δ and the conclusion φ of any (classical) core proof. The
condition holds (by definition) just in case exactly one of the following
three conditions is satisfied:
1. Δ is non-empty, φ is ⊥, and ♮Δ.

2. Δ is non-empty, φ is not ⊥, and φ◂Δ.
3. Δ is empty, φ is not ⊥, and ±φ.
Theorem 1 For every classical core proof of φ from Δ, we have  ( D, j ) .
There is no relevantist ‘variable-sharing’ result in the literature, espe-

cially concerning any of the relevance logics within the Anderson-Belnap
tradition, that establishes a connection of relevance as demanding as this
one between the premises and conclusions of proofs. The condition
 ( D, j ) is of course violated by A, ¬A: B (provable in Intuitionistic,
hence also in Classical, Logic) and by A, ¬A: ¬B (provable in Minimal
Logic).
The reader will find more details, including examples of the foregoing
syntactic notions and rigorous proof of Theorem 1, in Tennant (2015b).
5 Further Metatheorems About Core Logic

Definition 1 Unless otherwise indicated, we shall assume of the deducibility
relation ⊢ that it satisfies no more than the rules of the system ℂ of Core Logic.
(To be precise, Δ ⊢ φ means that φ is core-deducible from premises that are
drawn from Δ. The set of premises in question may be a proper subset of Δ.)
Definition 2 We write Δ1, Δ2 for Δ1 ∪ Δ2, and we write Δ, φ for Δ ∪ {φ}.
Theorem 2. (Cut Elimination for Classical Core Proof) There is an effec-

tive method [ , ] that transforms any two core proofs
Δ φ, Γ
Π Σ (where φ ∉ Γ and Γ may be empty)
φ θ
into a core proof [Π, Σ] of θ or of ⊥ from (some subset of ) Δ ∪ Γ.

212 N. Tennant
Proof See Tennant (2012b) for Core Logic, and see Tennant (2015a) for
the generalization to Classical Core Logic. The results can also be found
in Tennant (2017).
Corollary 1. (Cut with Potential Epistemic Gain) If Δ ⊢ φ and

Γ, φ ⊢ ψ, then either Δ, Γ ⊢ ⊥ or Δ, Γ ⊢ ψ.
Corollary 2. (Cut for Absurdity) If Δ ⊢ φ and Γ, φ ⊢ ⊥, then Δ, Γ ⊢ ⊥.
Corollary 3. (Cut on Consistent Premises) If Δ ⊬ ⊥ and Δ ⊢ Γ and

Γ ⊢ φ, then Δ ⊢ ψ.
Theorem 3 If Δ ⊢I φ, then for some Γ ⊆ Δ, either Γ ⊢ φ or Γ ⊢ ⊥.
Theorem 4 If Δ ⊢C φ, then for some Γ ⊆ Δ, either G + j or G + ^ .
6 ummary of Notable Features of the Core

S
Systems
We shall end with a list of notable features of the core systems, including
certain points that we have not had space to expound upon above. The
interested reader will find more detail, in support of all these claims, in
Tennant (2017).
1. The so-called loss of unrestricted transitivity of deduction in Core

Logic brings with it epistemic gain, in the form of proofs of subse-
quents of what one might have expected to be able to prove (by
appeal to unrestricted transitivity).
2. Core Logic suffices for Intuitionistic Mathematics.

3. Classical Core Logic suffices for Classical Mathematics.
4. Classical core proofs display the most exigent connection of rele-
vance among their premises and conclusions that has ever been for-
mally explicated.
5. Core Logic suffices for the hypothetico-deductive testing of scientific
theories against empirical evidence.
6. Core Logic is the logic of ‘conceptual constitution’: it suffices for
neo-logicist derivation of number-theoretic axioms from deeper logi-
cal principles governing the operator #xΦ(x) (the number of Φs).
7. Core Logic has the nice property that its natural deductions and
sequent proofs are structurally isomorphic.
8. Core Logic suffices for the reasoning involved in the logical and
semantic paradoxes.
9. Core Logic is obtained by smoothly generalizing an inferential read-
ing of the truth-tables.
10. Core Logic is the minimal inviolable core of logic that is needed, and
suffices, for rational belief revision. (See Tennant (2012a), Ch. 9.)
11. There is an automated proof-finder for (propositional) Core Logic,
whose decision problem is PSPACE-complete (like that of
Intuitionistic Logic).
12. Any classical core proof provides a means of ‘quasi-’effectively trans-
forming any verifications of its premises, relative to a model M, into
a verification of its conclusion, relative to M.
The verifications mentioned in point (12) are those of Tennant (2010)

and Tennant (2018). The adverb ‘quasi-effectively’ is appropriate when
the domain of the model M is infinite. If, however, M is finite, the prefix
‘quasi-’ may be dropped.
In conclusion: we submit for consideration, and as a challenge to both
absolutists and pluralists recommending other systems of deductive logic,
the following view:
The main aim of any formal system of deductive logic is to be able to regi-
ment in ultimately fine-grained detail, and in a structure-respecting and
214 N. Tennant
perfecting way, the informal but rigorous expert reasoning conducted by

mathematicians. But mathematicians fall into two communities, between
which the question of allegiance need not, for present purposes, be
addressed: the constructive and the non-constructive, or classical.4 The
main questions are: How best does one regiment the expert deductive rea-
soning in which they respectively engage? And what is the relationship
between the respective systems that result?
The Core logician (of the disjunctive kind explained above) offers definite
answers to these questions. The challenge, to both absolutists and plural-
ists advancing different choices of system(s), is to show that their rival
choices come anywhere near matching the core systems on the collective
criteria implicit in the notable features just listed above.5
Notes
1. This diagram should be interpreted as involving successive choice points
in response to the questions posed in vertically descending order. Each
question is posed within the overall context created by the successively
narrowing columns containing it.
2. In general, of course, what mathematicians call lemmas and corollaries
can feature both as premises of proofs and as conclusions of proofs.
3. Only in non-core systems like I and C is there the possibility that abnor-
mality could arise by having an MPE stand also as the conclusion of an
application of EFQ.
4. Even a committed anti-realist can recognize the legitimacy of the demand
that one should be able to formalize classical deductive reasoning (on the
part of the realist) in the way described here. The philosophical disagree-
ment between the realist and the anti-realist can be set aside while such a
normative-cum-descriptive project is undertaken. Once the project is
completed, however, one can embark on that deferred debate. The follow-
ing question will then loom large: what is the underlying methodological
or metaphysical or meaning-theoretic commitment involved in acquiesc-
ing in applications of any of the strictly classical negation rules? Our own
answer is given in Tennant (1996): applications of such rules manifest
one’s realist metaphysical outlook, to the effect that the ‘litmus sentences’
involved enjoy determinate truth values, regardless of our ability to deter-

mine what these values are, even in principle.
5. In section “What Is Wrong with R and Its Ilk” we made the telling objection
that there is no known metatheorem to the effect that ordinary mathemati-
cal reasoning can be fully formalized in any system like the Anderson-Belnap
system R. A further detracting contrast for R, when compared with Classical
Core Logic ℂ+, can be pointed out even at the propositional level. The deci-
sion problem for theoremhood in the propositional system R (let alone its
first-order counterpart) is undecidable. See Urquhart (1984). But the same
decision problem for ℂ+, as for C, is co-NP-complete.
References
Anderson, A. R., and N. D. Belnap, Jr. 1975. Entailment: The Logic of Relevance
and Necessity I. Princeton: Princeton University Press.
Friedman, H., and R.K. Meyer. 1992. Whither Relevant Arithmetic? Journal of
Symbolic Logic 7: 824–831.
Gentzen, G. 1934. Untersuchungen über das logische schliessen. Mathematische
Zeitschrift I:176–210. In The American Philosophical Quarterly 1, Translated
by M. E. Szabo as ‘Investigations into Logical Deduction’, 228–306, 1964.
Tennant, N. 1996. The Law of Excluded Middle Is Synthetic a Priori, If Valid.
Philosophical Topics 24: 205–229.
———. 1997. The Taming of the True. Oxford: Oxford University Press.
———. 2010. Inferential Semantics. In The Force of Argument: Essays in Honor
of Timothy Smiley, ed. J. Lear and A. Oliver, 223–257. London: Routledge.
———. 2012a. Changes in Mind: An Essay on Rational Belief Revision. Oxford:
———. 2012b. Cut for Core Logic. Review of Symbolic Logic 5 (3): 450–479.
———. 2015a. Cut for Classical Core Logic. Review of Symbolic Logic 8 (4):
236–256.
———. 2015b. The Relevance of Premises to Conclusions of Core Proofs.
Review of Symbolic Logic 8 (4): 743–784.
———. 2017. Core Logic. Oxford: Oxford University Press.
———. 2018. A Logical Theory of Truthmakers and Falsitymakers. In The
Oxford Handbook of Truth, ed. M. Glanzberg, 355–393. Oxford: Oxford
University Press.
Urquhart, A. 1984. The Undecidability of Entailment and Relevant Implication.
Journal of Symbolic Logic 49: 1059–1073.
Connective Meaning in Beall
and Restall’s Logical Pluralism
Teresa Kouri Kissel
1 Introduction
Jc Beall and Greg Restall (2006) spell out and defend a version of logical
pluralism. A feature of their particular pluralism is that corresponding
connectives in admissible logics mean the same thing. This means, for
example, that negations in classical, intuitionistic and relevant logic are
the same. This is in contrast to a “Carnapian” view, on which change in
logic is a change in the meanings of the logical terms.1
Unfortunately, Beall and Restall have not provided a conception of the
meaning of logical connectives on which this is true. I will show that, as
they describe it, negation in intuitionistic logic and negation in relevant
logic cannot mean the same thing on any reasonable interpretation of
their account of meaning. The fact of the matter is that they have not
settled just what “same meaning” amounts to, and any reasonable way of
spelling out what they say results in connectives which mean different
things.
T. Kouri Kissel (*)
Department of Philosophy and Religious Studies, Old Dominion University,
Norfolk, VA, USA

https://doi.org/10.1007/978-3-319-98346-2_10
218 T. Kouri Kissel
I will start by describing their pluralism, and what they take the logical
connectives to mean. I will then show that, as defined, the negation con-
nective in intuitionistic logic does not mean the same thing as the nega-
tion connective in relevant logic. Finally, I suggest that, as described, they
have no way around this problem. I close with some suggestions about
what the truth of the matter might be.
2 Beall and Restall’s Pluralism

Beall and Restall characterize their logical pluralism as “the view that
there is more than one genuine deductive consequence relation, and that
this plurality arises not merely because there are different languages, but
rather arises even within the kinds of claims expressed in the one lan-
guage” (Beall and Restall 2006, p. 3). Since they hold that the logical
connectives are part of the language of a logic, they claim that logics
which share connectives are “in one language” (or at least this is not ruled
out by connective meanings).
This view contrasts with what is typically taken to be Rudolf Carnap’s
view about logical pluralism. As Restall puts it
To put it graphically, as a pluralist, I wish to say that A, ¬A ⊢C B but A, ¬A

⊬RB. A and ¬A together, classically entail B, but A and ¬A together do not
relevantly entail B. On the other hand, Carnap wishes to say that A, ¬C A ⊢
B but A, ¬RA ⊬B. A together with its classical negation entail B, but A
together with its relevant negation need not entail B. (Restall 2002, p. 432)
That is, he takes it that Carnap holds that variation in logic is due to
variation in connective meanings, which he represents by subscripting
the connectives. Restall here is attributing to Carnap what Shapiro (2014)
calls the “Dummett-Quine-Carnap” perspective (see, e.g., Shapiro 2014,
p. 108). This perspective is one which assumes that, if the logic has
changed, then the meanings of the connectives have changed as well.
From this perspective, then, no two distinct logics can have connectives
with exactly the same meanings.
Connective Meaning in Beall and Restall’s Logical Pluralism 219
Restall, though, takes his position to be one where the variation in

logic is not due to such connective meaning change, which he represents
by subscripting the turnstile.2 What Beall and Restall are after is some-
thing stronger: that the logic can vary even while holding the connective
meanings fixed.
Beall and Restall need to identify something which changes when the
logic changes. This “something” cannot be the connective meanings. The
crux of their view comes from the Generalized Tarski’s Thesis
GTT: An argument is validx if and only if in any casex in which the prem-
ises are true, so is the conclusion (Beall and Restall 2006, p. 29)
What changes when the logic changes, for Beall and Restall, is what class
of cases we are considering. Cases are, at a minimum, “‘things’ in which
claims may be true” (Beall and Restall 2006, p. 89). An admissible class
of cases is one which generates a logical consequence relation which is
necessary, normative and formal. There are at least three classes of cases,
on this view, one which generates a classical logical consequence relation,
one intuitionistic relation and one relevant relation.
Cases need to be the type of thing such that changing the cases in ques-
tion does not necessarily change the meanings of the connectives in ques-
tion. For, if changing cases resulted in changing connective meanings, we
could not accommodate the “in one language” thesis. Cases, whatever
they are, must be the things such that changing them does not change
connective meaning.
Classical logic is generated by taking cases to be Tarski models.
Intuitionistic logic is generated by taking cases to be what Beall and
Restall call “stages in constructions” (though the clauses they give for
stages are just the clauses for nodes in Kripke structures, see Beall and
Restall 2006, pp. 62–64). Finally, relevant logic is generated by taking
cases to be situations similar to those of Barwise and Perry (1983), but
allowing for inconsistent situations.
Beall and Restall’s description of Tarski models follows the usual
semantics.3 Some examples of clauses, for some model M and some valu-
ation s, are
220 T. Kouri Kissel
M, s ⊨ ¬A if and only if it is not the case that M, s ⊨ A

M, s ⊨ A ∧ B if and only if M, s ⊨ A and M, s ⊨ B.
The intuitionistic cases, which they call stages in a construction, are nodes
in Kripke structures. A Kripke structure is a pair < W, ≤ > such that W is
a set of nodes, and ≤ is an ordering on those nodes. ≤ is reflexive and
transitive. They give as clauses (for w, u ∈ W)
w ⊨ A ∧ B if and only if w ⊨ A and w ⊨ B

w ⊨ ¬A if and only if, for all u such that w ≤ u, u ⊭ A.
Lastly, situations are “simply parts of the world” (Beall and Restall
2006, p. 49), except that situations can be inconsistent, and parts of pos-
sible worlds cannot be. There is an important relationship between situa-
tions, called compatibility. Roughly, two situations are compatible if they
do not “contradict” one another.4 Compatibility, according to Restall
(1999), is probably non-reflexive, symmetric and such that every situa-
tion is compatible with something. Some examples of clauses on this
semantics, for s and t situations and C the compatibility relation, are
s ⊨ ¬A if and only if, ∀t such that sCt, t ⊭ A

s ⊨ A ∧ B if and only if s ⊨ A and s ⊨ B.
What we need now, to answer the question of whether the connectives in

these logics do indeed share meanings, is to know what they take the
identity conditions of the meanings of the connectives to be. Unfortunately,
they do not say much on this front. What they do say is
The clauses can… [all] be true of one and the same [connective] simply in
virtue of being incomplete claims about… [the connective]. What is
required is that such incomplete claims do not conflict, but the clauses gov-
erning negation do not conflict. The classical clause gives us an account of
when a negation is true in a [model]…, and the constructive clause gives us
an account of when a negation is true in a construction. Each clause picks
out a different feature of negation. (Beall and Restall 2006, p. 98)
This means that, somehow, there is one thing which is (for example)
negation, and the various negation clauses given above all express differ-
ent aspects of that one thing.
To make sense of this claim, we need to fill in more details. I take it
that Beall and Restall (or, at least Restall) want to take the meanings of
the connectives on this picture to be maximal or minimal truth condi-
tions, as described in Restall (2002). The maximal truth condition for a
connective is something like the disjunction of all the clauses for the con-
nectives in the admissible cases. That is, the maximal truth condition for
negation (assuming the only admissible logics are classical, intuitionistic
and relevant) would be something like
(M is a Tarski model and M, s ⊨ ¬A if and only if it is not the case that M,

s ⊨ A) OR (w is a node in a Kripke structure and w ⊨ ¬A if and only if, for
all u such that w ≤ u, u ⊭ A) OR (s is a situation and s ⊨ ¬A if and only if,
∀t such that sCt, t ⊭ A)
The meaning of negation is then the disjunction of all three negation

clauses given above.5
A minimal truth condition, on the other hand, is a bare biconditional
between a sentence and an attribution of truth. So, for example, the min-
imal truth condition for negation would be something like
¬A is true in x if and only if A is not true in y for y compatible with x
The minimal truth condition straightforwardly captures the relevant

negation clause. The classical negation clause is also straightforwardly
captured: since classical situations are complete consistent situations, any
situation which is compatible with a classical situation will not make
anything true that the classical situation does not, so “compatible with”
in this restriction will be reducible to “equal to”. Finally, we can capture
an intuitionistic clause by again restricting ourselves to explicitly consis-
tent situations. In this case, we have to assume that two explicitly consis-
tent situations are compatible only in the cases where one can be “built”
out of the other. As I will claim later, in order to be “added to” the maxi-
mal truth condition, a clause for a connective must have certain features,
222 T. Kouri Kissel
and I will argue that at least one of those features is being able to be made
to fit a certain type of minimal truth condition.
If this is acceptable, then we need to ensure that there is no conflict
among clauses where they overlap. This is because Restall (2002) holds
that “the two accounts [classical and relevant] agree where they overlap”
(p. 438), which seems to accord with the claim above that incomplete
claims do not conflict (from Beall and Restall 2006, p. 98). There are two
questions we need to ask now: what, exactly, does it mean for two clauses
to overlap? And do the intuitionistic and relevant clauses agree where
they overlap? I think the answer to the second question is no, but it
depends in large part on our answer to the first. It is to this which I now
turn.
3 Overlapping Clauses
The method that Beall and Restall use to explain overlap involves describ-
ing all three consequence relations in a situation semantics (Restall (1999)
and Beall and Restall (2001)). They state
We… provide a model of how these things could be… to show how our
three different claims about the behavior of negation can be true together…
A world [classical model] is a complete, consistent situation… [We] take
all constructions to be situations… [that are] not only consistent, but
explicitly so. (Beall and Restall 2001, pp. 10–11)
Though this is ultimately the downfall of the view, we need the details to
explain the problem.
This method of explaining overlap involves restricting the clauses for
the connectives in the situation semantics in certain ways to generate
the clauses for the connectives in the other logics. At the very least, the
road is simple enough for finding a classical consequence relation in
such a semantics. We simply restrict ourselves to considering complete
consistent situations, and restrict compatibility to be the identity
relation.
The case for finding an intuitionistic consequence relation in such a

semantics is much harder. Beall and Restall (2001) claim we need to
restrict ourselves to considering explicitly consistent situations, where
each explicitly consistent situation makes ¬(A ∧ ¬A) true, for all A (Beall
and Restall 2001, p. 11). Complete consistent situations (those which
generate classical logic) are also explicitly consistent. Presumably, Beall
and Restall hold that this generates the intuitionistic clauses for the con-
nectives. In Sect. 4, I will show that this is not correct.
To figure out whether two clauses for a connective overlap, we first, we
need to assess what all the negation clauses have in common; what is it
that makes them disjuncts in the maximal truth condition?6 As noticed in
Hjortland (2013), all three negation clauses can be made to “fit” the fol-
lowing clause
(*) for x ∈ D, x ⊨ ¬A if and only if, ∀y such that xRy, y ⊭ A
This is something like the minimal truth condition alluded to above. Let’s
look at precisely how this will work. First, we take the default to be that
D is the entire situation space, and that R is compatibility (we will be
forced to loosen these requirements later). This means that the default of
(*) generates a relevant negation clause. Classical negation can then be
recovered by letting D range over complete, consistent situations, and
letting R be the identity relation. So far so good.
Recall that Beall and Restall (2001) claim that the intuitionistic clause
can be recovered by restricting our attending to explicitly consistent situ-
ations. In the language of (*), this means letting D range over explicitly
consistent situations, and letting R be such that, if c is a stage, and cRs,
then sRs, so that stages only “see” consistent situations.
We are now in a position to articulate a definition for when two clauses
overlap. Two clauses overlap when the intersection of their domains is
non-empty. I think Beall and Restall would agree up to this point. We
can also see now that all three clauses overlap, since the intersection of
situations with complete consistent situations and/or with explicitly con-
sistent situations will not be empty. This means what is left to determine
whether all three negations have the same meaning is to assess whether or
not they conflict where such overlap occurs.7
224 T. Kouri Kissel
4 The Problem
I will focus on showing that the intuitionistic clause conflicts with the
relevant clause for negation. I will demonstrate this by showing the addi-
tion of the intuitionistic clause to a maximal truth condition which
already has the relevant clause as one of its disjuncts makes negation arbi-
trary. If we add the intuitionistic clause to the maximal truth condition
which already has the relevant clause, then we will be able to add clauses
to the maximal truth condition which we ought not to consider part of
the meaning of negation. I take it, showing that the relevant negation
becomes arbitrary upon the addition of the intuitionistic clause demon-
strates that the clauses conflict. So, both clauses cannot be part of the
meaning of negation.
In order to assess whether the intuitionistic clause for negation con-
flicts with the relevant clause, we need to settle just what stages-as-
situations are. Beall and Restall claim that they are merely explicitly
consistent situations. If stages are merely explicitly consistent situations,
then changes in the relationship R in (*) need to be made. I call this horn
“Option 1”. If they are not explicitly consistent situations, then there will
be no overlap between the intuitionistic negation clause and the relevant
negation clause. I call this horn “Option 2”. I will consider each option
in turn, and show that both options make negation arbitrary, and thus no
matter how Beall and Restall go, they will not be able to succeed in their
connective meaning project.
Suppose first that stages are merely explicitly consistent situations. To
do this, we need to make sure that the intuitionistic clause for negation
can appropriately fit (*) when D ranges over explicitly consistent situa-
tions. This means that, in addition to R being restricted in such a way
that stages are only compatible with consistent situations, we also need to
insist that R not be symmetric. Were the parallel relation to be symmet-
ric, we would lose any notion of constructability. The ≤ relation is sup-
posed to be something like a “constructed-from” or “built from” relation,
so that we have w ≤ v just in case v is “built from” w. Suppose that the
relationship between explicitly consistent situations (the “built-from”
relationship) was indeed symmetric. Then anytime v was “built from” w,
w would also be “built from” v. We could never construct anything new,

so to speak, because each explicitly consistent situation would be built
from those before and after it. Everything would be true at the first situ-
ation in the ordering. So, for (*) to properly capture the intuitionistic
clause, R cannot be symmetric. If we want the negation clause for intu-
itionistic cases to properly fit (*), we need to restrict D to explicitly con-
sistent situations, and R to some subrelation of compatibility which is not
symmetric.
If we allow this restriction, though, we have a problem. If all it takes to
be part of the meaning of negation is to be “moldable” into (*) (where R
can be any subrelation of C), then lots of things will fit this mold.8 Not
all of them will be things we want to include as part of the meaning of
negation.
For example, let D be anything we like, and let R be empty.9 The empty
relation is certainly a subrelation of compatibility. The clause here is ∀d ∈
D, d ⊨ ¬A if and only if, ∀t such that dRt, t ⊭ A. Since R is empty, this
means the right-hand side of this conditional is always vacuously satis-
fied. The “empty-R” negation clause becomes d ⊨ ¬A for any A and any
d ∈ D. This means that negated sentences are always true. Once again, if
Beall and Restall are right, then this operator, which when appended to
anything produces a true sentence, must be part of the meaning of nega-
tion. This makes negation arbitrary. An operator which, when appended
to any sentence, makes that sentence true ought not be part of the mean-
ing of negation. If intuitionistic stages are explicitly consistent situations
ordered by a subrelation of the compatibility relation, and we can restrict
R in this way, then we can see that intuitionistic negation and relevant
negation cannot share a meaning.
This suggests quite strongly that intuitionistic stages simply cannot be
situations in the way Beall and Restall describe, at least not if we want the
relevant and intuitionistic negations to mean the same thing. If they
were, they would need to be ordered by a proper subrelation of compat-
ibility, and that would allow us to add arbitrary clauses to negation. This
rules out Option 1, and so we must now turn to Option 2.
Even though intuitionistic stages are not situations, we can still make
the intuitionistic clause for negation “fit” the (*) mold. We simply let D
be nodes in Kripke structures, and let R be the ≤ relation. Here, we avoid
226 T. Kouri Kissel
the previous problem: nodes will only be related to the nodes in the
future of their own Kripke structure, and the relationship will not be
symmetric.
There will be no overlap between the clause for negation in situations
and the clause for negation in nodes of Kripke structures. The intersec-
tion of the domains will be empty, since one contains only situations, and
the other contains no situations.
Again, though, we have a similar issue. If all it takes to be part of the
meaning of negation is to be “moldable” into (*) (with no requirements
on the domain) then there is a problem. Again, there are lots of clauses
which would fit this mold, and it is not the case that we would like to
allow all of them to be part of the meaning of negation.
Consider the following. Let us assume that the situation semantics is
rich enough such that every atomic sentence is not true at some situation.
Then, for any list of negated atomic sentences, we can construct a
situation-in-a-model at which they will all be true.10 We do so simply by
restricting our domain in certain ways. Suppose, for example, that the list
we were given contained ¬A1,…,¬An. Then, we simply restrict our domain
to the set/class of situations which make none of A1,…,An true. This will
be non-empty, since because of our assumption, there is at least one situ-
ation which does not make true the conjunction of A1,…,An.11 Further, at
any situation in this domain, the negations of A1,…,An are true, since the
compatibility relationship will be restricted to our domain, and no situa-
tion in the domain makes A1,…,An true. In this case, negation is arbi-
trary. We can make the negation of any sentences we would like true, just
by restricting the domain in certain ways. Thus, allowing restrictions to
the domain in order to get the intuitionistic clause for negation requires
allowing restrictions that gets arbitrary clauses for negation. I claim that
Beall and Restall cannot allow this and still claim that they have suc-
ceeded in pinning down the meaning of negation.
Summing up, we see that if stages are either explicitly consistent situa-
tions with the accessibility relation being a subrelation of compatibility,
or if they cannot be situations at all, negation as described by Beall and
Restall will be arbitrary if it includes both the intuitionistic and relevant
clause. Given the way Beall and Restall discuss stages, and intuitionistic
cases, these seem to be the only options. So, either negation is arbitrary
or intuitionistic negation cannot be part of the “disjunction of all nega-

tion clauses” meaning. This means that either negation is arbitrary or
Beall and Restall have not produced a pluralism without language change
after all.12
5 A Different Situation Semantics?

There is another option here, and that is to go the way of Dunn (1993).13
There, he presents a method by which we might “fit” both relevant and
intuitionistic negations into one model.
This system has two relations, C and ⊑, so there is no need to make
one “fit” the other, and we do not have the problem we saw above with
trying to make the Kripke structure “in-the-future-of ” relation fit with
our compatibility relation. Each relation corresponds to one “type” of
negation: ¬∗A (defined by x ⊨ ¬∗A if and only if x∗ ⊭ A) and ¬⊥A (defined
by x ⊨ ¬⊥A if and only if ∀y(y ⊨ A implies y⊥x)) (p. 331). Roughly, ∗ can
be thought of as the operator from the Routley-Meyer relevant logic
semantics, and so we use it to generate ⊑, which can be thought of as A
⊑ B ↔ everything A makes true, B makes true. ⊥ can be thought of as an
orthogonality relation (see, e.g., Goldblatt’s semantics for orthologic),
which we can use to interpret this C as a compatibility relation. By vary-
ing restrictions on the relations C and ⊑, and we can get classical, intu-
itionistic and relevant negation in this model. Moreover, it works even
when we add ∧ and ∨. One might think now that this is all well and
good, but that we still have two negations in play: ¬∗ and ¬⊥. Here is
where Dunn’s system shines: one can show, under suitable circumstances,
that these two negations really come to the same thing. Thomas (2015)
shows how one could extend such a system to include a conditional con-
nective as well. So Beall and Restall may be saved after all!
Thomas’s (2015) expansion is subtle and elegant, but it unfortunately
will not do for Beall and Restall’s position. In effect, it “lets” the quanti-
fiers influence the meaning of negation. We can construct, for example, a
model with two domains (Thomas’s term for something like our sets of
situations above), such that φ ⊨ ¬¬φ holds when φ is quantifier free, but
does not hold when it is not.14 We take two domains, call them ω and α,
228 T. Kouri Kissel
such that Thomas’s N relationship (the equivalent of our “R”, above) con-
tains exactly < ω, α > and < α, α >, and then let the other relationships
in the Thomas system (the range of each quantifier, and the tripartite
relationship which is used to make sense of the arrow) be identity rela-
tions. Then, supposing all objects are P at ω, but that we introduce some
new object, a, at α such that ¬Pa, then we will have both ω ⊨ ∀xPx and
ω ⊨ ¬∀xPx. This in effect means that negation will not be compositional,
since φ ⊨ ¬¬φ holds when φ is quantifier free, but does not hold when it
is not. This seems to be allowing the quantifiers to influence the meaning
of negation. Though this is not an issue for Thomas, it poses a problem for
Beall and Restall. A non-compositional negation would not match natural
language negation, and since Beall and Restall motivate their position in
part by the applicability of their various logical systems, the connectives in
their system must at least partly match the natural language connectives.
Since natural language negation is arguably compositional, a non-compo-
sitional negation will not do the trick, and should not be admitted as part
of the meaning of negation. Moreover, Thomas proves that all logics
obtained by this system are subclassical, so this logic will be necessary, nor-
mative and formal as desired.15 Again, we see that if we try to use (*) to
capture the meaning of negation, even in this new system, we will wind up
letting things in which ought not to count as negation.
6 Merely Technical Meanings

It is open to Beall and Restall at this point to simply state that their defini-
tion of the connectives is merely technical, and that it does not matter that
the meaning of negation really is arbitrary. This would allow them to say
that it does not matter that negation is arbitrary, and maintain that their
pluralism really is one in which the connectives across logics are pairwise
synonymous, where synonymy amounts to sameness of meaning.
However, on top of being unsatisfying for more general reasons, this
move is not open to Beall and Restall because they motivate their plural-
ism in part by the ability of each logic to capture deductive reasoning in
various applications. In particular, this means that the connective mean-
ings have to have something in common with what the connective
s peakers actually use. If they want their theory to be applicable, then the
connectives cannot be merely stipulative. So they cannot just define
negation technically, it has to match up with something in the world.
7 A Possible Solution
There is an interesting moral to draw here. There does seem to be some-
thing fundamentally different about intuitionistic negation and relevant
negation. I strongly suspect this has something to do with the fact that
intuitionistic and relevant consequences are both restrictions of classical
consequence in different ways. Intuitionistic logic allows explosion but
not double negation elimination, while relevant logic allows double nega-
tion elimination but not explosion. However, Beall and Restall have pro-
vided a meaning for the connectives on which classical and relevant
negations mean the same thing. We can also do this for classical and
intuitionistic negations. Here, we would use nodes in Kripke models as
intuitionistic cases, and recover classical consequence, and the classical
connective clauses, by restricting ourselves to one-node Kripke structures
(see Beall and Restall 2006, p. 98). If both of these systems are legitimate,
then we have a pluralism which is pluralistic in both logic and language.
We have two model theories, that of the situations semantics and that of
the Kripke structures, and two logics in each, classical and relevant in the
first, and classical and intuitionistic in the second. On the one hand, we
have what Beall and Restall want: two consequence relations with con-
nectives which share a meaning, if sameness of meaning of the connec-
tives in two logics amounts to being able to use the same model theory
for both logics. On the other hand, we have what our earlier characteriza-
tion of Carnap wants: sometimes, logic change does require connective
meaning change. In a sense, this type of system might be thought of as a
synthesis between Carnap and Beall and Restall.
As a sketch of how it might go, consider the following.16 We might
think that the meanings of the connectives are context sensitive in a cer-
tain way; for our purposes, the meanings are sensitive to the goals of
certain uses of logic. That is, the meanings are sensitive to what the logic
in question is being used to do (e.g., formalize a mathematical system,
230 T. Kouri Kissel
analyze a pattern of deductive inferences in natural language, etc.), or

what we are trying to figure out when we are talking about logic. Consider
two particular uses of logic: using classical logic to do classical analysis,
and using intuitionistic logic to do smooth infinitesimal analysis.
Interestingly, with smooth infinitesimal analysis, we can prove some nega-
tions of classical theorems, so that, for a certain group of elements, x,
¬∀x(x = 0 ∨ x ≠ 0) is true. Now, we might ask, what is the meaning of the
connectives for each of these logics on the occasion of each use of logic?
Well, that will be dependent on the contexts in which these uses take
place. Depending on the goals of the pursuers of those goals, sometimes
it might make sense to treat the logical connectives as meaning the same
thing. On the other hand, sometimes it might make sense to treat the
connectives as not sharing a meaning. Kouri Kissel (2018) develops a
theory of how this might work based on a specific pragmatic linguistic
framework, and Shapiro (2014) develops this on the basis of what type of
mathematical goal each context involves. In this way, the conversational
goals might be said to affect how many uses of each connective get to
count as the same negation. Thus, sometimes the connectives will mean
the same thing, and sometimes they will not.
8 Conclusion
Beall and Restall (2006) propose a logical pluralism where the connec-
tives for each logic are pairwise synonymous. Due to the manner in which
connectives are given their meaning, relevant negation and intuitionistic
negation cannot mean the same thing. Thus, their pluralism is a plural-
ism of languages and logics, not just logics as desired.17
Notes
1. I argue elsewhere (see Kouri (2018)) that this is not the best way to
understand the Carnapian picture. For the purposes of this chapter,
though, it suffices.
2. It is not clear that Restall is right in such an assessment of Carnap. At the
very least, Restall ought to grant that Carnap’s pluralism is a pluralism of
logics and languages. Restall’s position, on the other hand, is just a plu-
ralism of logics. Additionally, it is the case that Carnap could claim “A
and ¬A together, classically entail B, but A and ¬A together do not rele-
vantly entail B”. He would claim this is the case precisely because the
connectives mean something different in each logic. In some sense, what
this quote does for Restall is to show that Restall disagrees with the sec-
ond half (“A together with its classical negation entail B, but A together
with its relevant negation need not entail B”.) and not that Carnap dis-
agrees with the first. Thanks to Hannes Leitgeb for pushing me on this
issue.
3. There is a reason to think that the metalanguage Beall and Restall are
using is also classical. Though they never explicitly make a claim about
which metalanguage they are working with, it seems their description of
stages in constructions and situations as subclassical suggests that classi-
cal logic is the strongest logic admissible on their picture, and their rejec-
tion of contra-classical logics suggests the same. If this is the case, then it
seems that it would make a good choice for a metalanguage. For a criti-
cism of their (probable) choice in metalanguage, see Read (2006). It
should be pointed out here that Beall and Restall never explicitly make
the claim that their metalanguage is classical. A referee has suggested to
me that in personal correspondence, Restall has gone so far as to claim
that there is no metalanguage associated with logical pluralism. I am not
entirely sure what to make of this, as presumably, we need some language
in which to discuss the logics which are admissible. Whether the meta-
language is classical, though, or even if there is no metalanguage, most of
what follows can be re-stated accordingly, and so I will not pursue this
complication further here.
4. Unfortunately, very little about the compatibility relation is committed
to by Restall. This poses some problems for Beall and Restall’s project. In
particular, it will turn out that it is important whether compatibility is
primitive, or somehow determined by negation.
5. There is another way we can view this (if Beall and Restall’s metalan-
guage is classical, these will be equivalent). It might be rather than a
disjunction, we have a conjunction of conditionals. The maximal truth
condition for negation would then be
(if M is a Tarski model then M, s ⊨ ¬A if and only if it is not the case
that M, s ⊨ A) AND (if w is a node in a Kripke structure then w ⊨
¬A if and only if, for all u such that w ≤ u, u ¬ A) AND (if s is a situ-
ation then s ⊨ ¬A if and only if, ∀t such that sCt, t ¬ A)
232 T. Kouri Kissel
Since I take it these are equivalent in their classical metalanguage (see

endnote 3), I will focus on the disjunctive form.
6. This is in part necessary to rule out “junky” connectives, formed by dis-
joining the “wrong” clauses to the maximal truth condition. For exam-
ple, it is required to prevent having to claim that the disjunction of the
clause for conjunction in Tarski models with the clause for the arrow in
situations is a legitimate connective meaning. Without this, the maximal
truth condition is not a good candidate condition for the meaning of
any connective, and we can dismiss Beall and Restall on those grounds.
7. There is an immediate response available on behalf of Beall and Restall
here (thanks to Alexandru Radulescu for the suggestion). Beall and
Restall claim that what they are doing is providing a precisification of the
vague notion of logical consequence. So why not also assume that they
are giving a precisification of a vague account of negation? Then, we
would not need such a precise definition of negation, as I have suggested
with maximal truth conditions and (*), and the examples I give in Sect.
4 can all be dismissed as borderline cases or outliers. However, I think
there will still be a problem here, though it is different from any I will
address here. There will be a tension between their firm claim that the
connectives must mean the same thing, and the vague meaning of those
connectives. In order to make the first claim, we must assume that there
is something precise about the nature of the connectives, while to make
the claim about vagueness, we must dismiss this preciseness. In effect, I
think that if we want to pursue the claim that the connective meanings
are vague, we are better to adopt something like the tentative conclusion
I give in Sect. 7.
8. Additionally, of course, the logical consequence relationship will have to
satisfy GTT and be necessary, normative and formal. For Beall and
Restall, necessity is truth preservation in all cases, normativity is the abil-
ity to go wrong if you accept the premises and reject the conclusion of a
valid argument, and formality is either schematicity or one of: providing
norms for thought as such, indifference to identities of objects, or con-
tentlessness (Beall and Restall 2006, Chap. 2).
9. This logical consequence relationship is also necessary, normative and
formal, and satisfies the GTT. It satisfies the GTT since the cases in
question are just whichever situations are “blind”. The empty-R relation
is just a sublogic of what Beall and Restall refer to as relevant logic, which
we had already assumed was necessary and formal, and so these charac-
teristics are preserved. Additionally, it is normative, since we “go wrong”
by assuming that the situations are compatible with something.
10. Thanks to Graham Leach-Krouse for suggested an example of this type.

11. These restricted domains of situations still abide by the necessary-
normative-formal requirements on being an admissible logic (see defini-
tion in endnote 9), and satisfy the GTT since the cases are simply
situations of a specific type. These situations produce a consequence rela-
tion which is formal (since it is in a model) and normative, since we “go
wrong” by assuming that the situations we are considering are not
restricted to not making true a certain set of sentences. Finally, it is nec-
essary. Beall and Restall define necessity as “the truth of the premises
necessitates the truth of the conclusion” (p. 14), in other words, as long
as it is not possible for the premises to be true and the conclusion to be
false (p. 40). But in this case, this is not possible, since the possibilities in
play are a subset of the original domain, so there are no possible cases
where things are different.
12. One might think at this point that we might be able to accomplish the
Beall and Restall project if we framed everything in terms of nodes in
Kripke structures rather than situations. After all, terminal nodes in
Kripke structures are classical models (see Beall and Restall 2006, p. 98).
However, we find we have the same problem once again. The default
instantiation of (*) would then be D=nodes in Kripke structures, R =≤.
We would have to expand ≤ so that it could be symmetric, since com-
patibility is symmetric. However, Kripke structures give us no clues as to
how to expand ≤ to appropriately capture compatibility, and so we can
proceed as we wish, and expand it to any relationship. Then we simply
take R to be the universal relation, and re-run the previous counterex-
ample. Another option would be to consider cases as classes of pointed
frames (thanks to Shay Logan for this suggestion). Then, we could cap-
ture the various logics by restricting what gets admitted to the class of
pointed frames in question. Moreover, the two counter examples pre-
sented above would be less odd, since one ought to expect that odd
frames come equipped with odd negations. However, we will still have a
problem here (thanks to Beau Madison Mount for this suggestion). It is
still the case that we will have “overlap” (e.g., pointed Kripke frames for
propositional logic with the null signature will overlap with classical
frames with a unary relation), and thus we will still be able to develop
sentences which are true when the frame is considered as one kind of
model, and not true when considered as the other.
13. Thanks to Aaron Cotnoir for suggesting this option.
14. Thanks to Ethan Brauer, Steven Dalglish and Giorgio Sbardolini for help
in constructing this example.
234 T. Kouri Kissel
15. They will be necessary, normative and formal on the basis of Beall and
Restall’s definitions. Formality here seems relatively simple. Necessity
and normativity, on the other hand, are a bit odd, and this is the result
of Beall and Restall’s definitions of them. For Beall and Restall, any
sublogic of classical logic will be necessary, since they hold that relevant
logic is necessary because worlds are a special type of situation. But here,
since, on Thomas’s system, classical logic can be recovered as a special
type of model, and since no model we are concerned with proves any-
thing contra-classical, all models ought to be truth preserving, and so
necessary on Beall and Restall’s picture. This reasoning mimics their
discussion of relevant logic, where they say “The question is, in other
words, whether truth-preservation over all situations guarantees truth-
preservation over all possible worlds. The answer is ‘yes’ if possible
worlds count as (perhaps special) situations” (Beall and Restall 2006,
p. 54). Here, we have swapped “models” for situations, and “classical
models” for possible worlds, but the reasoning is similar. Normativity,
for Beall and Restall, comes down to there being some sort of mistake
associated with reasoning invalidly on this type of logic. But they are
not terribly specific about what types of things count as mistakes. The
mistake associated with relevant logic is irrelevant reasoning, and the
mistakes associated with constructive logic are “mistakes of constructiv-
ity” (Beall and Restall 2006, 70). In this case, then, I claim that not
reasoning properly according to this new logic is reasoning composi-
tionally when one shouldn’t. As one referee has suggested, there is a
good reason for denying that all subclassical logics are normative and
necessary. Though this seems plausible (and many commentators agree
that Beall and Restall’s definitions of necessary and normative are prob-
lematic, see, e.g., Bueno and Shalkowski (2009) and Keefe (2013)),
strictly speaking here all we only need is the Thomas system to match
Beall and Restall’s definitions, and it seems to. On the other hand, if the
Thomas system does not produce necessary, normative and formal log-
ics, then it is not an option for Beall and Restall to begin with.
16. For more details, see Kouri Kissel (2018) and Shapiro (2014).
17. Acknowledgments: I would like to thank Stewart Shapiro, Shay Logan,
Chris Pincock, Kevin Scharp, Neil Tennant and a referee for this volume
for feedback on earlier drafts. Additionally, audiences at the 2016
Central APA, the UConn Conference on Logical Pluralism and Truth
Pluralism, the Munich Center for Mathematical Philosophy and the
2017 Canadian Philosophical Association meeting provided incredibly
helpful comments.
References
Barwise, J., and J. Perry. 1983. Situations and Attitudes. Vol. 78. Cambridge,
MA: MIT Press.
Beall, J.C., and Greg Restall. 2001. Defending Logical Pluralism. In Logical
Consequence: Rival Approaches. Proceedings of the 1999 Conference of the Society
of Exact Philosophy, ed. John Woods and Bryson Brown, 1–22. Stanmore:
Hermes.
———. 2006. Logical Pluralism. Oxford: Clarendon Press.
118 (470): 295–321.
Dunn, J.M. 1993. Star and perp: Two treatments of negation. Philosophical
Perspectives 7 (May): 331–357.
Hjortland, O.T. 2013. Logical Pluralism, Meaning-Variance and Verbal
Keefe, R. 2013. What Logical Pluralism Cannot Be. Synthese 191 (7):
1375–1390.
Kouri, T. 2018. A New Interpretation of Carnap’s Logical Pluralism. Topoi:
1–10. Online first.
Kouri Kissel, T. 2018. Logical Pluralism from a Pragmatic Perspective.
Australasian Journal of Philosophy: 96 (3): 578–591.
Read, S. 2006. Review of Logical Pluralism Notre Dame Philosophical Reviews.
https://ndpr.nd.edu/news/logical-pluralism/.
Restall, G. 1999. Negation in Relevant Logics (How I Stopped Worrying and
Learned to Love the Routley Star). In What Is Negation?, ed. H. Wansing and
D. Gabbay, Volume 13 of Applied Logic Series, 53–76. Dordrecht: Kluwer
Academic Publishers.
———. 2002. Carnap’s Tolerance, Language Change and Logical Pluralism.
Shapiro, S. 2014. Varieties of Logic. Oxford: OUP.
Thomas, M. 2015. A Generalization of the Routley-Meyer Semantic Framework.
Journal of Philosophical Logic 44 (4): 411–427.
Generalised Tarski’s Thesis Hits
Substructure
Elia Zardini
Earlier versions of the material in this chapter have been presented in 2015 at the Veritas
Pluralism, Language and Logic Workshop (Yonsei University); in 2018, at the LanCog Seminar
(University of Lisbon), at the LOGOS Workshop Pluralism and Substructural Logics (University
of Barcelona), at the fifth SBFA Conference (Federal University of Bahia) and at the Workshop
Disagreement within Philosophy (Rhine Friedrich-Wilhelm University of Bonn). I’d like to thank
all these audiences for the very stimulating comments and discussions. Special thanks go
to Agustín Rayo, Colin Caret, Bogdan Dicher, Catarina Dutilh Novaes, Luís Estevinha, Filippo
Ferrari, Ole Hjortland, Luca Incurvati, José Martínez, Ricardo Miguel, Sergi Oms, Nikolaj J. L.
L. Pedersen, Hili Razinsky, Lucas Rosenblatt, Sven Rosenkranz, Diogo Santos, Ricardo Santos,
Erik Stei, Célia Teixeira, Pilar Terrés, Zach Weber, Jack Woods and Jeremy Wyatt. I’m also
grateful to the editors Nathan Kellen, Nikolaj J. L. L. Pedersen and Jeremy Wyatt for inviting me
to contribute to this volume and for their support and patience throughout the process. The
study has been funded by the FCT Research Fellowship IF/01202/2013 Tolerance and Instability:
The Substructure of Cognitions, Transitions and Collections. Additionally, the study has been funded
by the Russian Academic Excellence Project 5-100. I’ve also benefited from support from the
Project FFI2012-35026 of the Spanish Ministry of Economy and Competition The Makings of
Truth: Nature, Extent, and Applications of Truthmaking, from the Project FFI2015-70707-P of the
Spanish Ministry of Economy, Industry and Competitiveness Localism and Globalism in Logic and
Semantics and from the FCT Project PTDC/FER-FIL/28442/2017 Companion to Analytic
Philosophy 2.
E. Zardini (*)
LanCog Research Group, Philosophy Centre, University of Lisbon,
Lisbon, Portugal
International Laboratory for Logic, Linguistics and Formal Philosophy, School
of Philosophy, National Research University Higher School of Economics,
Moscow, Russian Federation

https://doi.org/10.1007/978-3-319-98346-2_11
238 E. Zardini
1 Starting Off
Generalised Tarski’s Thesis
Beall and Restall (2006) (henceforth, ‘BR’) have influentially defended a

specific version of logical pluralism, a general view according to which,
roughly stated, there is more than one legitimate relation of logical conse-
quence. More in detail, BR endorse (p. 35)1 the rather convoluted thesis
(henceforth, ‘BRT’) that legitimate relations of logical consequence are all
and only those that:
BRT1 Drop out of Generalised Tarski’s Thesis:

GTT An argument is validx iff, in every casex where every
premise is true, so is the conclusion2
(given a suitable assignment of a range of cases to ‘x’);
BRT2 Additionally satisfy some other conditions traditionally associ-
ated with logical consequence, i.e. necessity, normativity and
formality.
By way of some initial skirmishes, there are actually considerations I

regard as decisive against all the three conditions of BRT2. As for the
necessity condition,3 the argument from ‘Snow is white’ to ‘Actually,
snow is white’ is valid, but, for no context C, ‘If ‘‘Snow is white’’ is true-
as-uttered-in-C, ‘‘Actually, snow is white’’ is true-as-uttered-in-C’ is nec-
essary4 [the point is due essentially to Kaplan (1989), pp. 507–510; see
Zardini (2012) for elaboration and variations].5
As for the normativity condition, the argument from ‘P0’, ‘P1’, ‘P2’ …,
‘Pi’ and ‘All my scientific beliefs are that P0, that P1, that P2 …, that Pi’ to
‘All my scientific beliefs are true’ is valid, but it is not the case that one
ought to believe that all one’s scientific beliefs are true [the point is due
essentially to Makinson (1965); see Zardini (2018a) for elaboration and
variations]. BR do mention the preface paradox (pp. 16–18) and appar-
ently reject, on the basis of the considerations marshalled by the paradox,
that all their scientific beliefs are true, but, in spite of their insisting that
one also ought to believe that all one’s scientific beliefs are true, those
Generalised Tarski’s Thesis Hits Substructure 239
considerations would seem to show not only that one ought to reject that
all one’s scientific beliefs are true, but also that one is under no pressure at
all to believe the extremely strong, unlikely, unreliable, arrogant—and
indeed rather silly—proposition that all one’s scientific beliefs are true—
there is no dilemma about believing that not all one’s scientific beliefs are
true, that is just the unambiguously right thing to believe.6
As for the formality condition, there would seem to be a logic of, say,
metaphysical necessity as a particular necessity distinct from, say, epistemic
necessity and governed by principles different from those governing other
necessities (there is no interesting logic of a thin, formal ‘necessity’; it is only
thick, material necessities that have interesting logics). However, I know of
no reasonable sense in which the notion of metaphysical necessity is for-
mal [see Zardini (2018c) for elaboration and variations]. It’s as simple as
that. Having noted all this about BRT2, in the following (really, from the
last paragraph of this section onwards), I’ll rather focus on BRT1.
I’m very sympathetic to logical pluralism as the general view roughly
stated in the fourth last paragraph. To take a natural example, I think that
classical analysis is legitimate (and true)—and so, a fortiori, that its accom-
panying classical logic is legitimate—and that intuitionist analysis [in the
style of e.g. Brouwer (1927)] is also legitimate (and true), and so, a for-
tiori, that its accompanying intuitionist logic is legitimate. Their apparent
incompatibility can quite satisfactorily be explained away by the very
plausible hypothesis that they describe different mathematical structures.7 It
is true that some members of either party make extremely bold claims to
the effect that, as a matter of principle, the mathematical structures pos-
tulated by the other party cannot exist [e.g. Dummett (2000), 250–269].
But I find such claims very implausible and only backed up by very weak
arguments, so that I think that it can safely be concluded that such mem-
bers are plain wrong (there’s room for everyone in Plato’s heaven!). Here
as elsewhere, first-order tolerance requires a modicum of second-order
intolerance. In effect, I thus accept the ‘regionalist’ thesis [already sug-
gested by Putnam (1968)] that different logics apply to different domains. I
submit that the opposite idea that the same logic applies to every logical
domain gets whatever spurious plausibility it has from comparison with
the much more plausible idea that the same physics applies to every physical
domain. But, while physical domains would seem connected and homoge-
240 E. Zardini
neous enough to make the latter thesis rather likely, the whole range of
logical domains—going from the most exotic physical particles through
baldness and truth to the most esoteric mathematical structures—are so
completely disconnected and highly heterogeneous as to make the former
thesis rather unlikely. Given the bewildering variety of logical domains, it
is just plausible to expect that a single logical operation can obey a certain
principle on a certain domain and not obey it on another domain.
I’m not quite sympathetic to logical pluralism as typically understood in
the contemporary debate—where BR have been a major player—for, on
that understanding, even when taking a particular argument and even
when choosing a particular ‘level of formality’, as when, for example, it is
debated in philosophy of logic [e.g. Beall (2007)] whether the particular
argument from ‘The Liar sentence is true iff it is not’ to ‘The Earth is flat’
is valid at the level of the logic of truth, it is supposed to be legitimate to
hold that the argument is valid and also legitimate to hold that it is not
valid, and so, given that in this case presumably it is legitimate to accept
the premise of the argument if (and only if ) it is legitimate to hold that
the argument is not valid,8 it is in effect supposed to be legit to accept the
problematic instances of the T-schema and also legit to reject them! In
addition to undermining the significance of what is in fact one of the most
significant debates in philosophy of logic, pending first arguments such
specific claim is in itself so helplessly relativist that it would seem it ought
to lead to the rejection of any general view entailing9 it.
I’m even less sympathetic to BRT (in particular, to its ‘only’-component).
For, quite generally, BRT is yet another example of trying to impose a for-
mat to which every instance of a kind has to conform (in this case, every
instance of the kind Legitimate Relation of Logical Consequence). Besides
the telling considerations of Wittgenstein (1953), §66 (which would not
seem to have adequately been heeded in much subsequent analytic philoso-
phy), I think that the track record of such attempts in the history of thought
in general and in philosophy in particular (and even more in particular
about such philosophically central kinds as Truth, Knowledge, Cause, etc.)
is extraordinarily negative. And I don’t see any reason for thinking that the
specific case of the kind Legitimate Relation of Logical Consequence will
deviate from this general tendency. Indeed, in the following I’ll corroborate
this point by discussing a particular area that systematically violates GTT.
The discussion is ultimately meant to achieve something more interesting
than simply shooting a sundry series of cheap counterexamples at GTT,

and even more interesting than pointing to a relatively unified area that
systematically violates it; to wit, it is meant to uncover a fundamental diver-
gence between logical consequence and truth preservation, and, relatedly, to
invert the nowadays dominant order of priority between logical consequence
and logical truth.10
Substructure
The area that, in section “Generalised Tarski’s Thesis”, has been referred to
as systematically violating GTT is the one of substructural logics [see Restall
(2000); Paoli (2002) for a couple of general, philosophically sensitive sur-
veys]. To recall, operational properties are those properties of a logic which
concern particular logical operations (for example, reductio ad absurdum is
an operational property in that it concerns negation). By contrast, struc-
tural properties are those properties of a logic which concern general fea-
tures that only depend on considering premises and conclusions as a
manipulable structure of unstructured objects (for example, reflexivity is a
structural property in that it only depends on the premise and conclusion
being the same object). Classical logic and many non-classical logics (for
example, intuitionist logic) have a series of noteworthy structural proper-
ties: reflexivity, monotonicity, transitivity, contraction, commutativity and
others. A logic is substructural iff it lacks one of the properties in that series.
I’ll argue then that GTT systematically rules out all kinds of philosophically
interesting substructural logics as legitimate relations of logical consequence.
Given some of my work, I actually have a particular emotional attach-
ment to this exclusion from the realm of ‘logical consequence’ :-( But, as
in many other cases, it’s really not that interesting to determine exactly to
which logics the label ‘legitimate relation of logical consequence’ applies
(either in virtue of the vagaries of the use of that expression or in virtue of
a sheer stipulation).11 What’s interesting is what kind is singled out by
GTT, and how central it is for philosophical theorising about logic. In that
respect, my cumulative argument will show that the kind singled out by
GTT is not central for philosophical theorising about logic. And, as has
already been foreshadowed, there will also be other consequences for
broadly Tarski-inspired theses.
242 E. Zardini
2 The Crash
Reflexivity
A first structural property is reflexivity:
(I) φ ⊢ φ holds.
There are philosophical reasons to doubt (I) [Moruzzi and Zardini (2007),
pp. 180–182].12 Firstly, (I) refers to circular arguments—so circular that
they might be thought to violate epistemic and metaphysical principles
governing logical consequence. Epistemically, one might try to forge some
connection between validity and transmission [Wright (2000)] and think
that a valid argument must be such that one can use it in at least some
context to acquire a new justification for believing its conclusion13 [see
Martin and Meyer (1982) for very broadly similar considerations].14 But
in no context can one acquire a new justification for believing φ by infer-
ring it by (I) from φ.15 Metaphysically, one might try to forge some con-
nection between Ableitbarkeit and Abfolge [Bolzano (1837)] and think
that a valid argument must be such that, if its premises are true in virtue
of non-logical facts, its conclusion must be true (also) simply in virtue of
the truth of the premises.16 But x-is-true-in-virtue-of-the-truth-of-y is non-
reflexive (i.e. such that, for some x, it is not the case that x is true in virtue
of the truth of x), even restricting to sentences that are true in virtue of
non-logical facts.17 Secondly, notice that (I) is an atypical structural prop-
erty, in that it licences the validity of rules (things of the form ‘Γ ⊢ φ
holds’) rather than metarules (things of the form ‘If Γ0 ⊢ φ0, Γ1 ⊢ φ1, Γ2
⊢ φ2 … hold, ∆ ⊢ ψ holds’). That would seem already in itself problem-
atic: if at least almost all allegedly valid structural properties licence meta-
rules rather than rules, that is good evidence that all allegedly valid
structural properties licence metarules rather than rules. And it becomes
even more problematic once it is realised that, for this reason, (I) violates
the plausible principle that an argument is valid in virtue of some of the
semantic properties of some expressions occurring in its premises or conclu-
sion—principle that is entailed by some understandings of the plausible
principle that argument validity is a matter of analyticity, for example by
the understanding that an argument’s validity consists in the fact that its
conclusion preserves the truth of its premises in virtue of the meaning of

some expressions occurring in its premises or conclusion.18
That is all exciting stuff, but enter now GTT. In every casex where φ is
true, so is φ. That holds no matter what x is, and so non-reflexive logics
cannot drop out of GTT, and hence, by BRT1, non-reflexive logics are not
legitimate relations of logical consequence. That’s an unhelpful ruling!
The objection is sometimes raised that logical consequence is only con-
cerned with truth preservation—i.e. with the fact that the conclusion is true
except the premises are not—and variations thereof, whereas the reasons
given in the second last paragraph in favour of non-reflexive logics clearly
advert to quite different styles of considerations. Whatever its other mer-
its [Zardini (2012)], the objection relies on an untenable separation of
truth preservation from other features (such as those adverted to in the
second last paragraph). For at least some of the same styles of consider-
ations given in favour of non-reflexive logics also speak in favour, for at
least some kind of implication, of the claim that, for some P, ‘If P, then
P’ is not true. But that implication will also presumably satisfy the deduc-
tion theorem of its home logic, and so its home logic had better be non-
reflexive on pain of validating ‘If P, then P’, and so on pain of failing to be
truth preserving. Failure of desiderata for implication different from truth
preservation thus amplifies, via the deduction theorem, into failure of the
desideratum for logical consequence of truth preservation. Say that an
argument immediately fails to be truth preserving iff its premises are true and
its conclusion is not (which is presumably tantamount to the argument not
being [truth preserving as per the idea presented at the beginning of this
paragraph]), and say that an argument ultimately fails to be truth preserv-
ing iff its validity entails the validity of an argument that immediately fails
to be truth preserving. Then, the point of this paragraph can be summed
up by saying that, while (I) does not immediately fail to be truth preserv-
ing, it does ultimately fail to be truth preserving.
Monotonicity
A second structural property is monotonicity:
(K) If Γ ⊢ φ holds, Γ, ∆ ⊢ φ holds.

244 E. Zardini
(K) is uncontroversially not valid for many relations of non-deductive

consequence (i.e. relations where the truth of the premises supports but
does not guarantee the truth of the conclusion). That is not in itself par-
ticularly relevant, since BRT is only about logical consequence, which is
a relation of deductive consequence (i.e. a relation where the truth of the
premises does guarantee the truth of the conclusion). However, even in
the case of logical consequence, there are philosophical reasons to doubt
(K), as it licences arguments with irrelevant premises [a comment analo-
gous to the one in section “Reflexivity” applies concerning the connec-
tion with truth preservation, cf Moruzzi and Zardini (2007), p. 184;
Zardini (2015a), pp. 228–229].19
That is all exciting stuff, but enter now GTT. In every casex where
every element of Γ, ∆ is true, so is every element of Γ. Therefore, if, in
every casex where every element of Γ is true, so is φ, then, in every casex
where every element of Γ, ∆ is true, so is φ. That holds no matter what x
is, and so non-monotonic logics cannot drop out of GTT, and hence, by
BRT1, non-monotonic logics are not legitimate relations of logical con-
sequence. That’s an unhelpful ruling!
It is an especially unhelpful ruling for anyone with an interest in rele-
vant logics [e.g. Anderson and Belnap (1975)]: for, while these are tradi-
tionally more concerned with implication rather than with logical
consequence and consequently presented axiomatically (pace the [logical-
consequence]-first doctrine I’ll criticise in section “From Logical
Consequence Back to Logical Truth”!), they can also naturally be pre-
sented in such a way that (K) is not valid in them (a presentation which
philosophically makes a lot of sense, since as I’ve in effect noted in the
second last paragraph, irrelevance is not more benign for logical conse-
quence than it is for implication). Given that relevant logics are—together
with intuitionist logic—BR’s main witness for logical pluralism, it is
actually somewhat strange that they do not directly thematise this issue.
The issue does obliquely creep up in their discussion of the status of LEM
in their own version of relevant logics (pp. 53, 100–101): in that version,
because they want to deny that φ ⊢ ψ ∨¬ψ holds, but implicitly assume
(K), they must deny that ∅ ⊢ ψ ∨¬ψ holds, and so, I take it, deny that
LEM is a logical truth20 (both of which they actually do). They sweeten
the pill by saying that LEM is a necessary truth. But that’s not good
enough. One problem with BR’s position is that many relevant logics do
count LEM as a logical truth: for example, the relevant logic BX is defined
in terms of adding LEM as a logical truth21 to the basic relevant logic B
(analogously, a few other relevant logics …X are defined in terms of add-
ing LEM as a logical truth to the relevant logic …), and the even more
influential relevant logic R can be—and often is—partially characterised
in terms of having LEM as a logical truth. Be that as it may with LEM in
particular, the same issue arises with almost any alleged logical truth (save for
absolutely trivial things like ‘For some P, P’). Since the whole hierarchy
of relevant logics as standardly conceived relies on such alleged logical
truths being logical truths in the pertinent logics, and since anyways I’d
like to think that there are relevant logical truths in addition to absolutely
trivial things like ‘For some P, P ’, I deem BR’s position untenable (even
without going into the fact that it does not even speak to the case where
(K) is applied to an argument with premise and conclusion). They should
either, pace philosophical common sense (and their own choice of wit-
ness), deny that relevant logics are legitimate relations of logical conse-
quence or, pace GTT, deny (K).
Transitivity
A third structural property is transitivity:
(S) If Γ ⊢ φ and ∆, φ ⊢ ψ hold, ∆, Γ ⊢ ψ holds.
(S) is uncontroversially not valid for some relations of non-deductive

consequence [Zardini (2015a), pp. 231–233]. Again, for the reason seen
in section “Monotonicity”, that is not in itself particularly relevant.
However, even in the case of logical consequence, there are philosophical
reasons to doubt (S), as, in the presence of extremely compelling princi-
ples, it gives rise to the paradoxes of vagueness [e.g. Zardini (2008)]. For
example, (S) gives rise to the Sorites paradox in the presence of the
extremely compelling principle of tolerance (if something has a vague
property, so does something relevantly similar to it).
246 E. Zardini
That is all very exciting stuff, but enter now GTT. If, in every casex
where every element of Γ is true, so is φ, then, in every casex where every
element of ∆, Γ is true, so is every element of ∆, φ. Therefore, if, in every
casex where every element of ∆, φ is true, so is ψ, then, in every casex
where every element of ∆, Γ is true, so is ψ. That holds no matter what x
is, and so non-transitive logics cannot drop out of GTT, and hence, by
BRT1, non-transitive logics are not legitimate relations of logical conse-
quence. That’s a very unhelpful ruling!
Contraction
Up to now, we’ve been implicitly assuming that premises are combined in

a very coarse-grained way, suitably modelled by sets of sentences. But one
of the hallmarks of a few substructural logics is precisely the rejection of
that assumption. To open up the possibility of a more fine-grained way of
combining premises, let’s henceforth replace in our formal modelling sets
of sentences with series of sentences. A fourth structural property is then
contraction:
(W) If Γ, φ, φ ⊢ ψ holds, Γ, φ ⊢ ψ holds.
There are philosophical reasons to doubt (W). Firstly, in the presence of

extremely compelling principles, (W) gives rise to the semantic paradoxes
[e.g. Zardini (2011)]. For example, (W) gives rise to the Liar paradox in
the presence of the extremely compelling principle of correspondence (‘P’
is true iff P). Secondly, (W) forecloses an attractive account of what it is
for an object to be changing from one state to its opposite [Zardini
(2018d)].
That is all very exciting stuff, but enter now GTT. In every casex where
every element of Γ, φ is true, so is every element of Γ, φ, φ.22 Therefore,
if, in every casex where every element of Γ, φ, φ is true, so is ψ, then, in
every casex where every element of Γ, φ is true, so is ψ. That holds no
matter what x is, and so non-contractive logics cannot drop out of GTT,
and hence, by BRT1, non-contractive logics are not legitimate relations
of logical consequence. That’s a very unhelpful ruling!
Commutativity
A fifth structural property is commutativity:
(C) If Γ, φ, ψ, ∆ ⊢ χ holds, Γ, ψ, φ, ∆ ⊢ χ holds.
There are philosophical reasons to doubt (C). Firstly, there might be con-
text change between an occurrence of a premise and another occurrence of
a premise [e.g. Zardini (2014a)]. For example, if each next occurrence
occurs on the day after the day on which the previous occurrence occurs,
〈‘Tomorrow it’ll rain’, ‘It’s not raining’〉 is naturally regarded as inconsis-
tent (and so, let’s assume, as entailing everything), but 〈‘It’s not raining’,
‘Tomorrow it’ll rain’〉 isn’t.23 Secondly, (C) forecloses an attractive account
of what it is for an object to be changing from one state to its opposite
[Zardini (2018d)].
That is all very exciting stuff, but enter now GTT. In every casex where
every element of Γ, ψ, φ, ∆ is true, so is every element of Γ, φ, ψ, ∆.
Therefore, if, in every casex where every element of Γ, φ, ψ, ∆ is true, so
is χ, then, in every casex where every element of Γ, ψ, φ, ∆ is true, so is χ.
That holds no matter what x is, and so non-commutative logics cannot
drop out of GTT, and hence, by BRT1, non-commutative logics are not
legitimate relations of logical consequence. That’s a very unhelpful
ruling!
3 Aftermath
Unireliabilism
In all the cases we’ve considered in Sect. 2, GTT fails in that its satisfac-
tion by a logic in effect requires that the logic capture the fact that, for
every model in a certain class, it is not the case that the premises are true in
the model and the conclusion is not. Ascending to a more abstract perspec-
tive, GTT fails in those cases arguably at bottom because it is thus a spe-
cies of a more general kind of occurrence-insensitive, single-value,
single-model reliabilism: ‘reliabilism’ because it only considers whether the
248 E. Zardini
conclusion has the relevant value when the series of premises do (rather
than considering stronger connections between the conclusion having the
relevant value and the series of premises doing); ‘occurrence-insensitive’
because it only considers which sentences the premises are (rather than
considering, for example, how many times and in what order they occur);
‘single-value’ because it only considers one value relevant for all the series
of premises and the conclusion (rather than considering different values
relevant for different occurrences of the premises or the conclusion); ‘sin-
gle-model’ because, on each test,24 it only considers one model for all
the series of premises and the conclusion (rather than considering different
models for different occurrences of the premises or the conclusion). Since
such reliabilism is hostile to plurality in occurrences, values and models in test-
ing for logical consequence, it may well be labelled for short ‘unireliabilism’.
The issue surfaces slightly when BR note (p. 91) that, because of its
‘preservationism’, GTT rules out non-reflexive and non-transitive logics.
But that’s just the tip of the substructural iceberg—as we’ve seen in Sect.
2, GTT systematically rules out all kinds of philosophically interesting
substructural logics! Quite generally, GTT clashes with substructurality.
And the culprit is not some vague, benign and acceptable ‘preservation-
ism’—as we’ve seen in the last paragraph, GTT clashes with substructur-
ality because it is a species of the much more precise, malign and
questionable unireliabilism, which, obliterating as it does plurality in
occurrences, values and models, as well as intensional connections between
premises and conclusion, makes any version of ‘pluralism’ based on it barely
worth its name. Indeed, once GTT’s unireliabilism has been identified as
the culprit for its clash with substructurality, it becomes clear that, just as
GTT generalises in a very natural way the classical idea that an argument
is valid iff, for every classical model, it is not the case that the premises are
true in the model and the conclusion is not (by relaxing its ‘classical-
model’-component), one can generalise further in equally natural ways uni-
reliabilism itself (by relaxing some of its components). And at least the
specific substructural logics alluded to in Sect. 2—rather than represent-
ing conceptions of logical consequence totally alien to GTT—can actu-
ally be thought of as instances of such further generalisation. To wit,
non-reflexive logics of the kinds alluded to in Sect. 2 can be thought of as
requiring that the value of the conclusion be better than the values of the
premises (thus clashing with the ‘single-value’-component of unireliabi-

lism). Non-monotonic logics require that the value of the conclusion be
due to the values of the premises (thus clashing with the ‘reliabilism’-
component of unireliabilism). Non-transitive logics of the kind alluded
to in Sect. 2 can be thought of as allowing that the value of the conclu-
sion be worse than the value of the premises (thus clashing with the
‘single-value’-component of unireliabilism). Non-contractive and non-
commutative logics are sensitive to the number and order of occurrences of
a premise (thus clashing with the ‘occurrence-insensitive’-component of
unireliabilism). Non-commutative logics of the first kind alluded to in
Sect. 2 can be thought of as testing different occurrences of the premises
or the conclusion at different models (thus clashing with the ‘single-
model’-component of unireliabilism).25 I see no reason for stopping at
unireliabilism and not embracing these further dimensions of plurality
which give rise to substructurality.26
Logic-Relative Version of GTT; an Absolute-Truth

A
Formulation of a Logic-Relative Version of GTT
Although it concerns a variety of logics (most of which non-classical),

BRT itself is a theory in classical logic [in this, it follows a well-trodden
path: for a variety of reasons—some better than others according to e.g.
Zardini (2018b)—the usual metatheory of many non-classical theories is
indeed classical]. Classical (more accurately, non-substructural) reason-
ing was in fact crucial at several places in my arguments in Sect. 2. For
example, when I argued that ‘in every casex where φ is true, so is φ’, I was
simply helping myself to the classical logical truth ‘Every F is F’. Thus,
taking our cue from a theme already emerged in fn 19, can’t we simply fix
the problem by adopting instead a logic-relative version of GTT, requiring
only that GTT hold for a logic once GTT is interpreted27 and evaluated
using that very same logic (so that, for example, in discussing GTT for
non-reflexive logics, ‘Every F is F’ is no longer available as a logical truth)?
No, we can’t. To begin with, we’d need to supplement the logic with
theories of syntax, of cases and of truth-in-a-case, and not every interesting
logic will be able adequately to sustain such supplementations, since the
required expressive resources might be lacking from the logic or they
250 E. Zardini
might be governed by too weak principles of the logic. Not every interesting
logic has been developed with the purposes of doing syntax or
semantics.
Moreover, picking up from the second disjunct of the second last sen-
tence of the last paragraph, it’s completely up for grabs which logics would
satisfy the logic-relative version of GTT. Indeed, it is likely that quite a
few interesting logics will not satisfy its logic-relative version, and we in
fact know that even some of those that did not create problems for GTT
will no longer satisfy its logic-relative version. For example, the Bočvar
logic B3 [Bočvar (1938)] and the Kleene logic K3 [Kleene (1938)] will not
satisfy the logic-relative version of GTT, since they lack a suitably weak
implication vindicating claims—so crucial to GTT—along the lines of
‘Every F is G’ (for example, ‘In every casex where φ is true, so is φ’). In
fact, we also know that some interesting logics that did create problems
for GTT will still not satisfy its logic-relative version. For example, non-
contractive logics will not satisfy the logic-relative version of GTT, since,
letting ⊗ be a multiplicative conjunction, in typical non-contractive log-
ics φ ⊢ φ ⊗ φ does not hold, but, in the lattice-theoretic semantics asso-
ciated with typical non-contractive logics [e.g. Ono (2003)], ⊗ is
idempotent on designated value (i.e., if φ gets designated value, φ ⊗ φ gets
the same value), so that ‘If φ gets designated value in a certain casex, so
does φ ⊗ φ’ and hence—since, presumably, φ is true in a case iff φ gets
designated value in that case—‘If φ is true in a certain casex, so is φ ⊗ φ’
hold (and presumably do so even when using a non-contractive logic in
the metatheory). GTT’s rulings were definitely unhelpful, but those of its
logic-relative version are likely to be helpless.
Furthermore, while the logic-relative version of GTT is designed to
preserve its letter, it arguably abandons its spirit. For GTT’s spirit argu-
ably incorporated unireliabilism, but the new substructural interpreta-
tions of its universal quantification and implication deprive GTT of one
component or other of unireliabilism. Indeed, there is barely anything sub-
stantial in common at all, unireliabilist or otherwise, between, say, the sense
in which a non-reflexive logic thinks it satisfies GTT and the sense in
which a non-transitive logic thinks it does. The fact that the logic-relative
version of GTT holds for the logics for which it does would seem simply
to show that such logics are in the relevant respects expressively adequate
rather than show that they all conform to the same format. GTT is paid lip
service, but wildly different doctrines are actually pursued behind the
chants of truth preservation.
Finally, it is hard to see what roles cases are playing anyways in the logic-
relative version of GTT. Taking a conditional of the form ‘If every prem-
ise is true, so is the conclusion’, say that its interpretation in terms of
classical [universal quantification and implication] is an ‘[individual-
truth]-preservation conditional’ (in the sense that the individual truth of
all the premises is preserved by the conclusion), whereas its interpretation
in terms of non-classical [universal quantification and implication] (as
these are available in the logic in question) is a ‘[collective-truth]-genera-
tion conditional’ (in the sense that the collective truth of the series of
premises generates the truth of the conclusion).28 (Make the same dis-
tinction if universal quantification is replaced by conjunction.)29 Then,
for the non-classical logics that satisfy the logic-relative version of GTT,
the non-classicality of plain [collective-truth]-generation conditionals
(which are typically much stronger than the corresponding [individual-
truth]-preservation conditionals) will presumably suffice to rule out the
undesired classical arguments, without quantification over cases and with
absolute truth instead of truth (or designated value) relative to cases.30 True,
that might still rule in a lot of other undesired supra-classical arguments
(i.e. arguments whose form is not valid even in classical logic), as, for
example, the argument from ‘Italy will play against Sweden’ to ‘Italy will
lose’ (since the [collective-truth]-generation conditional ‘If “Italy will
play against Sweden” is true, so is “Italy will lose”’ sadly holds). But, for
better or worse, such arguments will be ruled out by BRT2 (especially, by
the necessity and formality conditions). Cases are supposed to be nothing
less than the cornerstone of GTT (pp. 23–24), but they are nothing more
than a frill of its logic-relative version.
From Logical Consequence Back to Logical Truth
The logic-relative version of GTT cannot thus rescue BRT-style plural-

ism. Still, I think that, especially in its final, absolute-truth formulation
(without cases or designated values, as per the last paragraph of section “A
252 E. Zardini
Logic-Relative Version of GTT; an Absolute-Truth Formulation of a

Logic-Relative Version of GTT”), the logic-relative version of GTT
affords an important insight into some of those logics that satisfy it31
(which include the specific substructural logics alluded to in Sect. 2).32
But, before I can adequately explain this in the ninth next paragraph, I
first need to lay out some background.
A mantra in contemporary philosophy of logic [e.g. Etchemendy
(1988), pp. 74–78] is represented by the [logical-consequence]-first doc-
trine already mentioned in section “Monotonicity”, according to which
the central property in logic is logical consequence, and logical truth is
grounded33 in it as being merely a degenerate limit case of logical conse-
quence: those valid arguments that have no premises and one conclusion
(where a logical truth is taken to be the conclusion of any such valid
argument)34 [from a historiographical point of view, it is usual to attri-
bute to Tarski the shift in focus from logical truth to logical conse-
quence against the evil Frege-Russell tradition, see e.g. Tarski (1930)].
The great insight is supposed to be that logic is not a body of truths but
a body of rules, and so, in this respect, more akin to etiquette than to
mathematics.
As I’ve already mentioned and referenced in fn 20, BR themselves
seem actually wary of the [logical-consequence]-first doctrine, apparently
for the reason that one could equally well ground logical consequence in
logical truth by saying that, given a logic L, the fact that φ0, φ1, φ2 …, φi
⊢L ψ holds is grounded in the fact that φ0 & φ1 & φ2 … & φi → ψ is a
logical truth in L. While my main aim in this section is to argue against
the [logical-consequence]-first doctrine, I should emphasise that by no
means do I intend to defend an opposite universalist view along the lines
of a [logical-truth]-first doctrine, and, in particular, that I regard the
grounding just mentioned as untenable.
Firstly, and less importantly, the grounding in question does not cover
cases of logical consequence with infinitely many premises. My two cents
is that this is probably best dealt with by extending the logic with infini-
tary conjunction—after all, logics without conjunction would already
have had to be extended with (normal) conjunction to deal with cases of
logical consequence with at least two premises. In for a penny, in for a
pound. Of course, there might be issues as to whether every logic with
valid arguments with at least two premises supports some such extension,
and, if it does, whether it supports a privileged one that can be relied on
in vindicating the grounding in question.35
Secondly, and more importantly, the grounding in question runs afoul
of logics with non-standard behaviour of conjunction or implication. As for
conjunction, a good example is given by Sb (fn 6), which is a philosophi-
cally interesting and technically well-behaved logic, but in which φ & ψ
→ χ can be a logical truth without φ, ψ ⊢ χ holding (essentially because
φ, ψ ⊢Sb φ & ψ does not hold). As for implication, good examples are
given by B3 and K3, which are philosophically interesting and techni-
cally well-behaved logics, but in which φ ⊢ ψ can hold without φ → ψ
being a logical truth (essentially because φ → ψ is indeterminate as soon
as either (B3) or both (K3) of φ and ψ are). (A dual, equally good exam-
ple is given by the logic of paradox LP [Asenjo (1966)], which is a philo-
sophically interesting and technically well-behaved logic, but in which φ
→ ψ can be a logical truth without φ ⊢ ψ holding (essentially because, in
LP, φ → ψ is true as soon as φ is false, even if it is also true).) Notice that
all these examples depend on taking as operative conjunction and impli-
cation those immediately available in the logics. It is certainly technically
possible to extend the logics with further operations with a more stan-
dard behaviour; the problem is that there would typically seem to be no
privileged way of doing so [witness the plethora of non-material implica-
tions proposed for extending K3 or LP; see e.g. Field (2008); Priest
(2006) respectively, and Zardini (2016) for some critical discussion],
which undermines the grounding of logical consequence in logical truth,
since it leaves one with no determinate grounding basis. To elaborate on a
particularly resilient version of the problem, suppose that, for example, a
particular implication ⇒0 does at least the extensional part of the job for
some such logic L, so that, even without restricting to the ⇒0-free lan-
guage, φ0, φ1, φ2 …, φi ⊢L ψ holds iff φ0 & φ1 & φ2 … & φi ⇒0 ψ is a
logical truth in L. A particularly resilient version of the problem is that
there typically is at least one other, equally natural implication ⇒1 whose
logic and so whose meaning differ, respectively, from the logic and so from
the meaning of ⇒0 but which also does at least the extensional part of the
job, so that, even without restricting to the ⇒1-free (or ⇒0-free) lan-
guage, φ0, φ1, φ2 …, φi ⊢L ψ holds iff φ0 & φ1 & φ2 … & φi ⇒1 ψ is a
254 E. Zardini
logical truth in L. One would then be faced with the riveting question of
whether logical consequence in L is grounded in the ⇒0-implicational
logical truths of L or in the ⇒1-implicational logical truths of L (or in
both, or in either, or partly in one partly in the other, etc.).36
Thirdly, and even more importantly, the grounding in question is shat-
tered by logics with a standard behaviour of the determinacy operator [see
Zardini (2014c) for some contrast with a non-standard behaviour of that
operator]. A good example is given by the supervaluationist logic Sp [Fine
(1975)], which is a philosophically interesting and technically well-
behaved logic, but in which φ ⊢ ψ can hold without φ → ψ being a logi-
cal truth (essentially because φ ⊢Sp 𝒟 φ holds but φ → 𝒟 φ is not a
logical truth in Sp). Points analogous to those made in the last paragraph
hold concerning the possibility of extending the logic with a further
implication satisfying the deduction theorem.
Does the [logical-consequence]-first doctrine fare any better than the
[logical-truth]-first one? Well, the Tarskian shift from logical truth to logical
consequence is certainly beneficial insofar as one should also look at logical
consequence, which, for the reasons mentioned in the last three paragraphs,
in a few logics includes features that genuinely go over and beyond those of logi-
cal truth.37 This observation, however, falls dramatically short of vindicating
the [logical-consequence]-first doctrine in its full strength, since, for example,
the doctrine insists that, even for those logics where there is no such differ-
ence between logical consequence and logical truth, logical truth is grounded
in logical consequence. I’m going to attack the [logical-consequence]-first
doctrine precisely by arguing that, for some of those logics (in particular, for
the specific substructural logics alluded to in Sect. 2), it is logical conse-
quence that is grounded in logical truth rather than vice versa.
My argument starts with the plausible if somewhat controversial
assumption that logical consequence is itself grounded in facts about truth,
rather than being primitive [e.g. Field (2015)] or grounded in other kinds
of facts like, for example, facts concerning proof [e.g. Prawitz (2005)]. In
turn, this standard, general semantic conception of logical consequence is
typically accepted in a specific version, in classical logic, according to
which logical consequence is grounded in guaranteed preservation of the
truth of the premises by the conclusion [following in this Tarski again, espe-
cially Tarski (1936)].
What for our purposes is interesting to observe is that such truth-

preservation account of logical consequence is very close to the one got by
reading (as it is in the spirit of BRT, see pp. 23–24) a dependence claim
in the left-to-right direction of GTT. Let’s make this observation more
accurate. The truth-preservation account cannot plausibly relax GTT and
unireliabilism’s ‘reliabilism’-component, since the account (‘guaranteed
preservation of the truth of the premises by the conclusion’) is all about
the conclusion being true except the premises are not, a condition which,
as I’ve already in effect observed in section “Reflexivity” (and reinforced
in fn 19), is presumably tantamount to the condition that it is not the
case that the premises are true and the conclusion is not.38,39 The truth-
preservation account cannot plausibly relax GTT and unireliabilism’s
‘occurrence-insensitive’-component, since the account (‘guaranteed pres-
ervation of the truth of the premises by the conclusion’) is all about the
truth of the premises (and the conclusion), and that is a question on
which, for example, φ and φ, φ necessarily agree.40 They might disagree
on some novel understanding of premises as tokens (fn 22), but such
understanding is, at the present stage of inquiry, sheer speculation. The
truth-preservation account cannot plausibly relax GTT and unireliabi-
lism’s ‘single-value’-component, since the account (‘guaranteed truth pres-
ervation from the premises to the conclusion’) is all about identity in
value between the premises and the conclusion.41 The truth-preservation
account can plausibly relax GTT and unireliabilism’s ‘single-model’-com-
ponent, since nothing in the account (‘guaranteed preservation of the
truth of the premises by the conclusion’) prevents that the guarantee of
truth preservation consists in its holding over a certain class of tests that
might look at different models for different occurrences of the premises
or the conclusion. Summing up, the truth-preservation account can plau-
sibly [diverge from GTT and unireliabilism while remaining faithful to
itself ] only with respect to their ‘single-model’-component: it is still
essentially a species of occurrence-insensitive, single-value reliabilism.
Therefore, the truth-preservation account clashes with substructurality
for the same reasons for which GTT does (minus the ‘single-model’-prob-
lem with the first kind of non-commutative logic alluded to in section
“Commutativity”). Substructural logical consequence is not grounded in
truth preservation.42
256 E. Zardini
In what, pray, is it then grounded? It is at this juncture that the

[collective-truth]-generation conditionals can play a role, at least for those
substructural logics (extendable) with conjunction and implication
behaving in the way mentioned in fn 31 (which include the specific sub-
structural logics alluded to in Sect. 2; the proposal I’m about to develop
is supposed to apply mainly to those logics, and possibly to some more
substructural logics (extendable) with conjunction and implication
behaving in the same way). For, although, for those logics, the truth-
preservation account of logical consequence must be given up, we can
still uphold the more general semantic conception of logical consequence
by appealing to the idea that logical consequence is grounded in guaran-
teed generation of the truth of the conclusion by the collective truth of the series
of premises (i.e. by replacing in the truth-preservation account an [individ-
ual-truth]-preservation conditional with a [collective-truth]-generation
conditional). But how are we to spell out this idea more precisely?
As a first stab, we could say that, in every specific substructural logic L
alluded to in Sect. 2, the fact that φ0, φ1, φ2 …, φi ⊢L ψ holds is grounded
in the fact that the [collective-truth]-generation conditional T ⌜φ0⌝ & T
⌜φ1⌝ & T ⌜φ2⌝ … & T ⌜φi⌝ → T ⌜ψ⌝, as interpreted by L, holds over the
class of tests relevant for L. To say this would probably be the most direct
modification of the most standard way of spelling out the truth-
preservation account of logical consequence, but it is only a first stab as it
would have a couple of serious problems. Firstly, even when L does have
a ‘class of relevant tests’, these often involve as models rather exotic math-
ematical structures with no direct informal interpretation, and which can
therefore hardly serve as a credible ground for logical consequence (con-
trary to the class of tests relevant for, say, classical logical consequence,
which involve as models structures that can directly be informally inter-
preted as mundane possible situations). Secondly, as I’ve stressed in fn 30,
the only notion of truth straightforwardly available in L is the one of
absolute truth, whose interaction with the models involved in the tests
relevant for L is however ill-defined (what is well-defined is the notion of
truth-in-a-model, but it is well-defined in the classical metatheory of L, not
in L itself, and, even if it could be defined in L itself, it would give the
wrong results, for example, as per the argument in section “A Logic-
Relative Version of GTT; an Absolute-Truth Formulation of a Logic-
Relative Version of GTT”, for typical non-contractive logics).
What is causing both problems is reliance on a class of tests, and the

solution to both problems is therefore to get rid of it. But just how?
Notice first that the only point of quantification over such class in the
most standard way of spelling out the truth-preservation account of logi-
cal consequence is to ensure that logical consequence has the desired
modal force. Observe then that, for the purposes of grounding, there is
arguably no way of defining the class in purely non-logical terms (for one
thing, as we’ve seen in fn 5, it won’t do to define it in terms of metaphysi-
cally possible worlds).43 If so, the most standard way of spelling out the
truth-preservation account in effect amounts to grounding logical conse-
quence in (among other things) logical possibility (of the models involved
by the relevant tests). It does not offer a grounding of logical facts in
non-logical facts; it only offers a grounding of logical facts in general in a
privileged class of facts concerning logical possibility and individual-[truth-
in-a-model] preservation. But then it would seem that nothing of philo-
sophical value is lost by taking as more fundamental not logical possibility
as applied to models but logical necessity (aka logical truth) as applied to
sentences,44 and say instead that the fact that φ0, φ1, φ2 …, φi ⊢L ψ holds
is grounded in the fact that the [collective-truth]-generation conditional
T ⌜φ0⌝ & T ⌜φ1⌝ & T ⌜φ2⌝ … & T ⌜φi⌝ → T ⌜ψ⌝ is a logical truth in L
[cf Zardini (2014d), pp. 371–382 for the development of a congenial
proposal in the specific case of the non-contractive approach to the
semantic paradoxes mentioned in section “Contraction”],45 thereby
grounding logical facts in general in a privileged class of facts concerning
logical necessity and collective-[absolute-truth] generation (and, along the
way, getting rid of classes of tests).46
If you’ve read fn 46, you might now be tempted by the thought that
semantic ascent is dispensable, and so that we should rather say that the
fact that φ0, φ1, φ2 …, φi ⊢L ψ holds is grounded in the fact that the
object-language conditional φ0 & φ1 & φ2 … & φi → ψ is a logical truth
in L, thereby grounding logical facts in general in a privileged class of facts
concerning logical necessity and conjunction-cum-implication (and so jetti-
soning the semantic conception of logical consequence altogether). The
temptation should be resisted, as the proposed alternative grounding
arguably associates some facts about logical consequence with the wrong
facts about logical truth. Let’s take as example a typical non-contractive
logic LW (with ‘and⊗’ and ‘if→ …, then→ …’ as multiplicative conjunction
258 E. Zardini
and implication, respectively) and the fact that, in LW, by adjunction,
‘Snow is white’, ‘Grass is green’ entail ‘Snow is white and⊗ grass is green’.
According to my proposal, that is grounded in (the logical necessity of )
the fact that, if→ ‘Snow is white’ is true and⊗ ‘Grass is green’ is true, then→
‘Snow is white and⊗ grass is green’ is true, whereas, according to the alter-
native proposal, that is grounded in (the logical necessity of ) the fact that,
if→ snow is white and⊗ grass is green, then→ snow is white and⊗ grass is
green. The former account of the matter is clearly superior, since it
respects the difference in the argument between the series of premises and the
conclusion, and in particular it respects the point of the argument which is
to the effect that the simple ‘Snow is white’ and ‘Grass is green’ together
suffice for the complex ‘Snow is white and⊗ grass is green’. As in the classi-
cal case, it just is illuminating to use in the metalanguage the notions of
conjunction and truth for combining two premises and then, mentioning
a certain single sentence of the object language which involves an opera-
tor expressing the very same notion of conjunction, explaining why that
conclusion follows from the two premises—it just is illuminating to use
the idea of two sentences being both true to make a point about a single
sentence involving ‘and’.47 All this is completely missed by the latter
account, which flattens such wealth of structure concerning the interaction
between premise combination and conjunction into a tautology that has
nothing to do with either.48 Semantic ascent allows us to discern the fine
structure of logical consequence: its combination of premises and its
entailment from premises to conclusion.49
Having resisted the temptation, it only remains to state with due
solemnity a consequence of the approach we’ve been pursuing which has
been looming large for a while (and which, for that matter, is also a con-
sequence of the alternative proposal discussed in the last paragraph).
Namely, and quite literally, on this revolutionary approach, the [logical-
consequence]-first doctrine is turned upside down: in the specific sub-
structural logics alluded to in Sect. 2, not only logical truth is not
grounded in logical consequence, but, vice versa, it is logical consequence
that is grounded in logical truth.50 In those logics, the central property is
logical truth, and logical consequence is grounded in it as being merely a
degenerate limit case of logical truth: those logical truths that are [collective-
truth]-generation conditionals (where the series of premises and the con-
clusion of a valid argument are taken to be, respectively (and modulo

semantic descent), the series of conjuncts in the antecedent and the con-
sequent of any such logical truth).51 In those logics just as elsewhere, rules
are only legitimate insofar as they are based on truths.
Notes
1. Throughout, page references are to BR unless otherwise stated.
2. Throughout, I follow BR in assuming, merely for simplicity, a multiple-
premise, single-conclusion framework. The extension of my points to the
(superior but) more complex multiple-premise, multiple-conclusion
framework is straightforward.
3. BR claim that necessity is important for logical consequence at least
partly because it guarantees that logic applies unrestrictedly in hypothetical
reasoning (pp. 15–16). But it arguably doesn’t in the first place, since, say,
the law of excluded middle (henceforth, ‘LEM’) is a logical truth but it is
arguably wrong to reason that, if Brouwer were right, it would be the
case that either Goldbach’s conjecture holds or it doesn’t. Notice that the
hypothesis is perfectly possible (Brouwer might have held correct views),
so it is not even the case that logic applies unrestrictedly to possible hypotheses
(which would anyways raise the issue of why it should be more impor-
tant unrestrictedly to apply to possible hypotheses rather than to, say,
interesting hypotheses).
4. Throughout, I work with a standard framework for context-dependent
languages where the truth of a sentence is relative both to a context of
utterance and to a circumstance of evaluation, and I make such relativities
explicit with the relevant parts of the construction ‘φ is
true-as-uttered-in-C-as-evaluated-at-E’.
5. I’d like to forestall two likely reactions to this point. Don’t say that what
is true is ‘For every context C, if “Snow is white” is true-as-uttered-in-
C-as-evaluated-at-C, “Actually, snow is white” is true-as-uttered-in-C-as-
evaluated-at-C’. The proposed Ersatz has the embarrassing problem
(emerging in the opening ‘what is true [my emphasis, EZ]’) that neither
it nor its embedded conditional is non-trivially necessary (so that the
spirit if not the letter of the necessity condition would seem violated).
What the proposed Ersatz indeed offers is universality over contexts, but
that is a far cry from any interesting necessity: for every context C, ‘I exist’
260 E. Zardini
is true-as-uttered-in-C-as-evaluated-at-C, but there is no interesting

sense in which the existence of a first person is necessary [Zardini (2012),
pp. 266–268 will be excused if he perhaps got too carried away by his
‘semantic modality’]. Along these lines, the proposed Ersatz is not even
of the right logical form: for the necessity condition is presumably to the
effect that there is a certain unique truth-theoretic property (truth-as-
uttered-in-such-and-such-context) which is preserved at every circumstance
from premises to conclusion (cf fn 3), whereas the proposed Ersatz is to the
effect that every circumstance is such that at it some truth-theoretic property
or other (truth-as-uttered-in-a-context-containing-that-circumstance) is pre-
served from premises to conclusion, with different such properties being pre-
served at different circumstances. That’s like proposing to understand the
idea of universal access to healthcare not to the effect that that there is a
certain unique health system that is accessed by everyone, but to the
effect that everyone accesses some health system or other, with different
such systems being accessed by different people. Also, don’t say that,
while, for no context C, ‘If “Snow is white” is true-as-uttered-in-C,
“Actually, snow is white” is true-as-uttered-in-C’ is metaphysically neces-
sary, for every context C it is logically necessary. For, without further ado
(which, as far as I know, has never been made), the claim that logical
consequence is logically necessary is totally vacuous, as it is tantamount to
the claim that logical consequence holds with the very special force of …
logical consequence (in exactly the same way, also defeasible arguments
could be claimed to be ‘defeasibly necessary’!).
6. Such failure of normativity spells disaster for normative constraints on
logical consequence of the kind ‘If φ0, φ1, φ2 … ⊢ ψ holds, one should
not accept φ0, φ1, φ2 … and reject ψ’, which are endorsed, for example,
by Restall (2005); Beall (2015) [see Zardini (2016), pp. 313–314, fn 9
for some further discussion]. Notice that, if one proposed (what BR
don’t) that, actually, ‘one should not accept φ0 & φ1 & φ2 … and reject
ψ’, even setting aside various issues related to the possible infinity of
φ0, φ1, φ2 … it would be hard to see how such proposal does not in effect
amount to a wholesale rejection of the normativity of multiple-premise
arguments (since the source of the normative fact that one should not
accept φ0 & φ1 & φ2 … and reject ψ is presumably the fact that the
single-premise argument φ0 & φ1 & φ2 … ⊢ ψ holds), which would be
quite sad since the preface paradox can do nothing to undermine the
incontrovertible fact that quite a few times multiple-premise arguments
do have normative force (in particular, are such that one should not
accept each of their premises and reject their conclusion, independently

of whether one de facto accepts the conjunction of their premises).
Indeed, while, as far as the issues raised by the preface paradox are con-
cerned, in classical logic multiple-premise arguments can be taken to have
normative force (when and) only when one should [accept the conjunction
of their premises if one accepts each of them] (so that, if a multiple-premise
argument has normative force and one accepts each of its premises, since
one should then accept the conjunction of its premises, the proposal in
question—at least when crucially supplemented by some sort of account
of when one should [accept the conjunction of the premises of an argu-
ment if one accepts each of them]—by in effect appealing to the norma-
tivity of the corresponding single-premise argument can vindicate the
claim that one should not reject its conclusion), that is no longer so in
some logics with non-standard behaviour of conjunction such as the sub-
valuationist logic Sb [Jaśkowski (1948)]. In Sb, it is possible, for example,
that φ, ψ entail χ with normative force, that one accepts each of φ and ψ
but that φ & ψ is inconsistent (let alone entailed by φ, ψ), so that, for
non-epistemic reasons (and so for reasons not concerned by the issues
raised by the preface paradox), it is not the case that one should accept
it. In such a situation, since the argument from φ, ψ to χ has normative
force and one accepts each of φ and ψ, one should not reject χ, but, since
it is not the case that one should accept φ & ψ, the proposal in question
can no longer in effect appeal to the normativity of any φ & ψ-[single-
premised] argument to vindicate the claim that one should not reject χ.
7. To be clear, the idea is that the objects are different, but the logical opera-
tions on them are the same: for example, there is a single operation of dis-
junction, which obeys LEM on the real numbers of classical analysis,
whereas it does not obey it on the real numbers of intuitionist analysis.
An alternative idea would be that the objects are the same, but the logical
operations on them are different: for example, classical analysis talks about
classical disjunction which obeys LEM on the real numbers, whereas
intuitionist analysis talks about intuitionist disjunction which does not
obey LEM on the same numbers. Such alternative is not only in itself
unnatural, it is also incoherent in view of the well-known fact [Popper
(1948)] that, on a shared domain of objects and properties, intuitionist
logic would seem to collapse on classical logic: letting ‘orC’, ‘orI’, ‘identi-
calC’, ‘identicalI’, ‘notC’ and ‘notI’ express classical and intuitionist dis-
junction, identity and negation respectively, by using the deductive rules
appropriate for each logical expression we can derive ‘r is identicalI with
262 E. Zardini
0 orI r is notI identicalI with 0’ from ‘r is identicalC with 0 orC r is notC

identicalC with 0’, thereby making a hash of the intuitionist contin-
uum. While the classical-analysis/intuitionist-analysis case is only one
example of logical pluralism, I take the overall thrust of this discus-
sion to provide some evidence to the effect that, in those cases where
a logic is motivated by the aim of accounting for the specific behaviour
of objects and properties in a certain domain (as the specific logics
alluded to in sections “Transitivity” and “Contraction” are), plurality
of legitimate relations of logical consequence is accommodated by dis-
tinguishing domains rather than by multiplying logical operations. In
other cases, a logic is motivated instead by the aim of accounting for
certain general, domain-unspecific features of logical consequence (as the
specific logics alluded to in sections “Reflexivity”, “Monotonicity”
and “Commutativity” are)—in those cases, plurality of legitimate
relations of logical consequence is indeed most plausibly accommo-
dated by multiplying logical operations. In both kinds of cases, con-
trary to logical pluralism as typically understood in the contemporary
debate (as per the next paragraph in the text), plurality of legitimate
relations of logical consequence is ultimately accommodated by dis-
tinguishing which fully interpreted sentences they apply to.
8. While there is usually a slide in going from its being legitimate to hold
that φ does not entail an absurdity to its being legitimate to accept φ (for
example, if φ is ‘The number of stars in the universe is odd’), there is no
slide here in going from its being legitimate to hold that ‘The Liar sen-
tence is true iff it is not’ does not entail ‘The Earth is flat’ to its being
legitimate to accept ‘The Liar sentence is true iff it is not’, since the pos-
sibility that doing the former is not legitimate is essentially the only rea-
son for doubting that doing the latter is legitimate.
9. Throughout, by ‘entail’ and its relatives I mean the converse of the rela-
tion of logical consequence. By ‘imply’ and its relatives I mean instead an
operation expressed by a conditional operator.
10. Given this ultimate aim (and the fact that the abundant literature on
GTT has already done a good job in this respect), in this chapter I won’t
delve into other very plausible counterexamples, nor into the very obvi-
ous fact that, just as GTT is a very natural generalisation of an unduly
restrictive notion of logical consequence, there are very natural generali-
sations of GTT itself in several directions (a fact to be handled with
some caution, see fn 26).
11. Although, should a substructural logic prove to be part of the best solu-
tion to, for example, the semantic paradoxes, I myself would find it ter-
minologically misguided not to label it ‘legitimate relation of logical
consequence’ (surely, whatever logic governs truth deserves to be labelled
‘legitimate relation of logical consequence’!).
12. In the text, I’ll briefly mention what I regard as the best such reasons not
to commit to denial of (I)—an issue which obviously lies well beyond
the scope of this chapter—but simply to make more vivid how interest-
ing philosophical positions denying (I) would unhelpfully be outlawed
by GTT. Analogous comments apply for the other structural properties
to be considered below in the text.
13. That is, by relying in a non-deviant way on one’s belief in the premises of the
argument and on one’s inference from them to the conclusion: one can per-
haps acquire a new justification for believing, say, ‘There are circular
arguments whose conclusion I’ve inferred from their premises’ by ‘going
through’ the relevant instance of (I), but one would thereby be relying in
a clearly deviant way on one’s belief in the premise of the argument and
on one’s inference from it to the conclusion. As elsewhere [see e.g.
Chisholm (1966), p. 30 for the much discussed case of deviant causa-
tion, in whose debate it would ironically seem presupposed that deviance
only affects causation and not also rational connection], it is a tricky issue,
lying beyond the scope of this chapter, to spell out exactly what such
deviance is.
14. Even the crooked simplification argument, say, φ & ψ ⊢ φ passes muster,
since one can be told that Al met with Bob and Cate and acquire a new
justification for believing that Al met with Bob by inferring it by simpli-
fication from what one has been told.
15. This uncontroversial fact would easily be accounted for if, when trans-
mission occurs, the justification for believing the conclusion were simply
identical with the justification for believing the premises, for then it would
be obvious that the justification for believing the conclusion of (I) can-
not be new with respect to the justification for believing its premise. Pace
Moruzzi and Zardini (2007), p. 181, that would seem however a sim-
plistic view of what happens when transmission occurs (a mistake for
which I assume the sole responsibility!), since, presumably, also the justi-
fication one has for inferring the conclusion from the premises is part of the
justification one acquires for believing the conclusion. While the ‘mereol-
ogy of justification’ is still in its infant days, a natural speculation is that
264 E. Zardini
nothing having as a part a justification j for believing φ can be a new justi-

fication with respect to j for believing φ, given which it still follows that the
justification for believing the conclusion of (I) cannot be new with
respect to the justification for believing its premise (assuming through-
out, extremely plausibly, that one can only acquire a justification for
believing the conclusion if one has a justification for believing the prem-
ises, and that the latter is then part of the former). Notice that, if the
necessary condition for validity in question is the stronger one to the
effect that one can use a valid argument in at least some context to
acquire a first justification for believing its conclusion, the relevant
uncontroversial fact is instead that in no context can one acquire a first
justification for believing φ by inferring it by (I) from φ, which is easily
accounted for, since it is obvious that the justification for believing the
conclusion of (I) cannot be first with respect to the justification for
believing its premise [see e.g. Zardini (2014b) for a recent discussion of
transmission].
16. Again, even simplification (fn 14) passes muster (if vacuously so), since,
arguably, φ & ψ is always true in virtue of the (sometimes non-logical)
fact described by φ, the (sometimes non-logical) fact described by ψ and
the (always logical) fact that φ and ψ entail φ & ψ.
17. I’m not saying that x-is-true-in-virtue-of-the-truth-of-y is irreflexive (i.e.
such that, for every x, it is not the case that x is true in virtue of the truth
of x), since it isn’t [Zardini (2018b)].
18. To belabour the point, φ does not preserve the truth of φ in virtue of the
meaning of some expressions occurring in it, even if it is crucial for that
preservation that, given the specific meanings involved in φ, the very
same meanings are involved in φ. Compare: I don’t have the same age as
my mother’s only son in virtue of being 38-year old, even if it is crucial
for that sameness that, given my specific age, that very same age is had
by my mother’s only son.
19. Read (1981,2003) offers a very different defence of the connection
between non-monotonic logics and truth preservation by reinterpreting
truth preservation in terms of a relevant implication (‘If every premise is
true, so is the conclusion’). Such reinterpretation would seem however
miles away from the original idea of truth preservation as presented in
section “Reflexivity” (whose intelligibility Read does not contest), which
is simply about the conclusion being true except the premises are not:
that requires a ‘relevance connection’ between the premises being true
and the conclusion being true just as little as the fact that an object
weighs 1,171,979 grams except it is not now in my right pocket (which

is presumably tantamount to every object that is now in my right pocket
weighing 1,171,979 grams) requires a ‘relevance connection’ between an
object being now in my right pocket and its weighing 1,171,979 grams.
Having noted this, Read’s idea of reinterpreting truth preservation in
terms of non-classical notions is in my view insightful in the respects I
myself will exploit in sections “A Logic-Relative Version of GTT; an
Absolute-Truth Formulation of a Logic-Relative Version of GTT” and
“From Logical Consequence Back to Logical Truth”.
20. That follows, for example, under the [logical-consequence]-first doc-
trine, according to which logical truth is grounded in logical consequence
as being merely a limit case of logical consequence with no premises and
one conclusion. BR seem actually wary of the doctrine (pp. 12–13,
despite p. 3), but they do accept at least that logical truths are co-exten-
sional with the single conclusions that follow from no premises, which
amply suffices for the implication at issue in the text.
21. And not simply as a necessary truth, for it is crucial for what is the logic of
BX (i.e. what follows from what in BX, what is a logical truth in BX,
what is a logical falsity in BX, etc.) that LEM have the same status as the
other axioms of BX, which it is again crucial for what is the logic of BX
that it be the one of logical truth. Analogous comments apply to the next
two occurrences in the text of ‘as a logical truth’.
22. Throughout, I understand the elements of a series 〈x0, x1, x2 …〉 to be the
members of the set {x0, x1, x2 …}, and, crucially, I understand the prem-
ises 〈φ0, φ1, φ2 …〉 to be the members of the set {φ0, φ1, φ2 …}, so that,
for example, the only premise in 〈φ, φ〉 is φ (as a consequence of both
understandings, I identify the premises 〈φ0, φ1, φ2 …〉 with the elements
of that series). The understanding of premises just flagged as crucial
would seem mandated by the traditional idea that premises are sentences
[which, in general, I myself have defended in Zardini (2018e); BR,
pp. 8–12, also consider judgement types and propositions, which however
mandate the same understanding of premises]. In connection with non-
contractive logics, sometimes [e.g. Girard (1995), p. 2] a different idea is
aired to the effect that premises are ‘tokens’ (of some kind or other). I
find extant proposals in this direction deeply problematic [see Zardini
(2018b) for details], and I am much more attracted to seeing non-
contractive logics as being sensitive to the number of occurrences of the
same premise (i.e. sentence), so that, for example, the series 〈φ, φ〉 does
not represent two tokens of φ, say φ19 and φ79 (whatever these may
266 E. Zardini
be!), but the double occurrence of φ. I’ll henceforth presuppose such

broad understanding of non-contractive logics. The understanding of
course invites the question how, when testing for logical consequence,
there can be sensitivity to the number of occurrences if one only consid-
ers whether the premises (i.e. certain sentences) are true. My reply begins
with the observation that, when, in the metatheory, one considers whether
‘the premises’ φ0, φ1, φ2 … are true, one is not simply considering
whether φ0 is true, whether φ1 is true, whether φ2 is true …; one is com-
bining φ0, φ1, φ2 … (to be sure, in that order) and considering whether
they are all true—i.e. whether φ0 is true together with φ1 together with φ2
… It is then the mode of such combination that, if one’s metatheory is in
a substructural logic, can be sensitive to the number of occurrences (and,
as for the logics in section “Commutativity”, to their order): for example,
if φ occurs twice as a premise as in 〈φ, φ〉, when testing for logical con-
sequence one should not consider whether simply φ is true, but whether
φ is true together with φ, and one might thereby be using a mode of
combination under which the latter question may receive an answer dif-
ferent from the one received by the former question. I’ll make important
use of this reply in the broader context of sections “A Logic-Relative
Version of GTT; an Absolute-Truth Formulation of a Logic-Relative
Version of GTT” and “From Logical Consequence Back to Logical
Truth”.
23. Clearly, on this kind of logic of context change, also all the other struc-
tural properties considered in this chapter are not valid, but I think that
it is more insightful to place it in the non-commutative camp rather than
in any other substructural camp.
24. For our purposes, it is important to distinguish between models and tests.
Roughly, models are single evaluations of the sentences of the whole lan-
guage, whereas tests are partial checks for logical consequence looking
whether, if the series of premises have the relevant value in the relevant
model(s), the conclusion has the relevant value in the relevant model.
Importantly, one test might look at different models for different occur-
rences of the premises or the conclusion. As a baby example, a language
might have only two different models, and logical consequence for that
language be determined by the tests checking for the fact that it is not
the case that the series of premises are true in one model and the conclu-
sion is not true in the other model [see Zardini (2014a) for a more adult
example].
25. I should note that the generalisations required by non-monotonic and

non-contractive logics would seem in an importantly different ballpark,
since their implementation is much less straightforward, and typically
requires the adoption, at some level or other, of the operations of, respec-
tively, intensional implication and multiplicative conjunction character-
istic of those logics (I’ll myself follow this strategy in sections “A
Logic-Relative Version of GTT; an Absolute-Truth Formulation of a
Logic-Relative Version of GTT” and “From Logical Consequence Back
to Logical Truth”).
26. All this of course does not mean that we should simply replace GTT
with some sort of GGTT that allows for all possible generalisations in
these directions (or others), since a few such generalisations will yield
relations that are obviously not instances of the kind Legitimate Relation
of Logical Consequence (but are not ruled out by BRT2 either). In a
familiar oscillation, in the attempt at weakening too strong a condition
that undergenerates with respect to a target kind, we end up with too
weak a condition that overgenerates with respect to the kind. There is no
(non-trivial) essence of Legitimate Relation of Logical Consequence.
27. Throughout, bear in mind that, as per regionalism (section “Generalised
Tarski’s Thesis”), in some cases such ‘interpretation’ does not involve dif-
ferent logical operations and is more akin to the way in which, say, play-
ing football and playing chess represent two different ‘interpretations’ of
playing even though these do not involve different properties playing0
and playing1 exemplified (on the one hand by people and on the other
hand) by football and chess, respectively.
28. That is only a very rough gloss and you shouldn’t read too much into it,
not the least because the sense in which ‘the collective truth of the series
of premises generates the truth of the conclusion’ varies dramatically
from logic to logic (recall the last paragraph in the text). Still, the gloss is
evocative at least for the specific substructural logics alluded to in Sect.
2, constituting a programmatic slogan congenial to the foundational role
I’ll argue in section “From Logical Consequence Back to Logical Truth”
these conditionals have. Thanks to Luís Estevinha for feedback on this
matter.
29. If the logic in question has more than one operation of universal quanti-
fication, implication or conjunction, I assume that there is a most appro-
priate one for the role these conditionals are supposed to play, and I
focus on that one (the assumption is arguably satisfied for all the specific
substructural logics alluded to in Sect. 2).
268 E. Zardini
30. To set aside distracting issues concerning the opacity of absolute truth
[e.g. Zardini (2015b)], in this discussion I assume that it makes sense to
extend the target logics with a quoting singular term ⌜φ⌝ for each sen-
tence φ of their original language and with a truth predicate T such that,
for every sentence φ of their original language, T⌜φ⌝ is intersubstitut-
able with φ. Such extensions are straightforward (indeed, on my view,
just as logical as, say, the extension of the conjunction-free fragment of
intuitionist logic with conjunction), contrary to those that would be
needed to develop a theory of truth-in-a-case (as per the third last para-
graph in the text). (Such extensions are also harmless, since they only
licence the intersubstitutability of T ⌜φ⌝ with φ if φ is T-free.) Moreover,
focusing on languages for which such extensions make sense is justified
since, in this discussion, my aim is to defend anti-universalist claims
rather than universalist ones. Thanks to José Martínez and Ricardo
Santos for feedback on some of these issues.
31. I should really be a bit more precise about what it is for a logic to satisfy
the absolute-truth formulation of the logic-relative version of GTT. For our
purposes, anticipating a bit, a natural way of making that notion precise
is to say that a logic L satisfies the absolute-truth formulation of the
logic-relative version of GTT iff [φ0, φ1, φ2 …, φi ⊢L ψ holds iff T ⌜φ0⌝
& T ⌜φ1⌝ & T ⌜φ2⌝ … & T ⌜φi⌝ → T ⌜ψ⌝, as interpreted by L, is a logi-
cal truth in L] (where, by fn 30, the last claim is equivalent with the
claim that φ0 & φ1 & φ2 … & φi → ψ, as interpreted by L, is a logical
truth in L, which makes satisfaction of the absolute-truth formulation of
the logic-relative version of GTT by L a matter of L’s conjunction and
implication correlating in the familiar ways to premise combination and
entailment in L, respectively). Thanks to Ricardo Santos for urging this
clarification.
32. Yes, including non-contractive logics, even given what I’ve said in sec-
tion “A Logic-Relative Version of GTT; an Absolute-Truth Formulation
of a Logic-Relative Version of GTT”, since we’re considering a formula-
tion of the logic-relative version of GTT in terms of plain [collective-
truth]-generation conditionals, without cases or designated values: while
it is the case that, if φ gets a designated value, so does φ ⊗ φ, we do not
have that, if φ is true, so is φ ⊗ φ [Zardini (2011)].
33. Throughout, I understand ‘ground’ and its relatives in a suitably broad
fashion, so as to include also the case of reduction (as is particularly plau-
sible in the case of the [logical-consequence]-first doctrine).
34. The second main conjunct in its full strength is not guaranteed by the
first one, and so, if at all desired, it must be added separately. Still, it
represents the arguably most plausible grounding of logical truth in logi-
cal consequence, and that’s why, throughout, we’re focussing on it.
Notice that ‘most plausible’ doesn’t imply ‘plausible’: following, but from
no premises, would seem to make just as little sense as arriving, but at no
places! Sometimes, the problem is fudged by invoking the empty set and
saying that logical truths are those sentences that follow from the empty
set; however, while it is useful formally to model logical consequence as
a relation between sets (or, in view of substructurality, series), only an
empty-set mystic would think that logical truths are characterised by a
distinctive logical relation to the empty set (as opposed to many other
more relevant objects). Some other times, the problem is fudged instead
by invoking the 0ary operator t and saying that logical truths are those
sentences that follow from t; however, while it is useful to introduce t in
the formal study of logical consequence (especially in view of substruc-
turality), its informal understanding is as of the ‘conjunction’ of all logical
truths, which arguably prevents grounding a sentence being a logical
truth in its following from t. Notice that an analogous problem affects
the related idea that logical falsity is grounded in logical consequence as
being merely a limit case of logical consequence with one premise and no
conclusions (whereas, FWIW, the more general notion of inconsistency of
a series of premises can still be grounded in logical consequence and logi-
cal falsity in terms of the series entailing the ‘disjunction’ of all logical
falsities). Thanks to Ricardo Miguel for pressing me on the formulation
of the [logical-consequence]-first doctrine.
35. Having noted all this, I’ll (pick up from fn 31 and) henceforth focus on
the finite case myself.
36. Even setting aside the problem raised in the text, the extension of these
logics with the desired operations typically requires complications going so
far beyond the basic, natural framework of the logics as to make the [logical-
truth]-first doctrine hardly credible for them. Moreover, the resulting
operations are typically so tailor-made to fit the target relation of logical
consequence that, given on the one hand the valid arguments of one of
these logics and on the other hand its putatively grounding logical truths,
the by far most plausible account is that the putatively grounding logical
truths are what they are because the valid arguments are what they are
rather than vice versa.
270 E. Zardini
37. Since, in arguing against the [logical-truth]-first doctrine, we’ve also

touched on some dramatic cases (B3 and K3) which, at least on their
standard sentential fragment, have no logical truths whatsoever but still
have a wealth of valid rules, it’s just fair to mention that similar dramatic
cases of logics [like TS of Cobreros et al. (2012)] which, at least on their
standard sentential fragment, have no valid rules whatsoever but still have
a wealth of valid metarules argue against the [logical-consequence]-first
doctrine [as does the fact that the valid rules of ST of Cobreros et al.
(2012) coincide with those of classical logic in spite of ST quite clearly
being different from classical logic!]. The point iterates at higher orders. A
thoroughly diverse picture of logics thus emerges, on which, for some
logics, what is most fundamental are their logical truths (as I’ll argue start-
ing from the next paragraph in the text); for some other logics, their valid
rules; for yet some other logics, their valid metarules … Thanks to Bogdan
Dicher, Sergi Oms and Lucas Rosenblatt for insisting on these points.
38. A further reinforcement: the condition that the conclusion is true except
the premises are not would seem to be all about getting things right—no
matter how inelegantly—in going from the premises to the conclusion; that
is, just not getting them wrong; that is, not going from premises that are
true to a conclusion that is not.
39. Notice that both the truth-preservation account of logical consequence
and GTT constitute a very specific version of reliabilism, where the rele-
vant property a valid argument is reliable about is truth (rather than
knowledge, unfalsity, truth together with unfalsity, etc.).
40. That is, φ and φ, φ agree that their only premise, φ, is true. The addi-
tional strength of φ, φ over φ consists in representing φ as being true also
as occurring twice. While that entails φ’s truth, it is not entailed by it.
That’s not enough to make the additional strength of φ, φ relevant for
the truth-preservation account of logical consequence: for example, the
property of being known by me is also such that, while my knowledge of
φ entails φ’s truth, it is not entailed by it, but that property is clearly
irrelevant for the account. Having made this point, there is indeed an
important question as to how to understand the idea that a premise is
true ‘as occurring twice’, which I’ve addressed in fn 22.
41. It’s worth noting that, even if preservation is dropped in favour of a rela-
tion relating possibly distinct values (say, connection), the immediately
resulting account (‘guaranteed connection between the truth-theoretic
values of the premises and of the conclusion’) is still all about ‘truth-
theoretic’ values such as truth, falsity, unfalsity, etc., whereas the distinction
among designated values in the model theory of some non-reflexive and
non-transitive logics does not sustain any such interpretation [for exam-
ple, think of the non-[truth-theoretic] interpretation offered by Zardini
(2008), pp. 347–349 of the distinction between two different kinds of
designated values relevant for a wide family of non-transitive logics].
42. We can further argue not only that substructural logical consequence is
not grounded in truth preservation, and not only that substructural logi-
cal consequence is not co-extensional with truth preservation, but also
that substructural logical consequence does not even require truth pres-
ervation. The best example for this is arguably offered by the non-con-
tractive approach to the semantic paradoxes mentioned in section
“Contraction” [especially as further developed in Zardini (2018f )], on
which one can warrantedly accept, say, that the Liar sentence λ is true
while holding that the argument from λ, λ to ‘Snow is black’ is valid, and
so while holding that the premise (which occurs twice, see fn 22) of the
argument is true and its conclusion is not. Substructural logical conse-
quence does not require truth preservation. Thanks to Sven Rosenkranz
for prompting this expansion of the argument.
43. To be clear, the claim is not that the intended class is not co-extensional
with any class defined in purely non-logical terms—not the least because,
under very plausible mathematical assumptions, such co-extensionality
does obtain in the case of many logics! Especially for such logics, the
claim is not that logical consequence is not correlated with individual-
truth preservation over a class of tests defined in purely non-logical
terms; it is rather that logical consequence is not grounded in such truth
preservation, as it is only correlated with it because the class of tests
defined in purely non-logical terms just so happens to be co-extensional
with the intended class, which is in turn only characterisable as such partly
in logical terms [for example, as the class of all and only those tests involv-
ing logically possible models, cf Etchemendy (1990), pp. 107–124].
44. This is not to deny that something of heuristic value can be gained by
working with the notion of a sentence being true in every logically possible
model rather than with the notion of a sentence being logically necessary,
just like something of heuristic value can be gained by working with the
notion of a sentence being true at every metaphysically possible world
rather than with the notion of a sentence being metaphysically necessary:
in both cases, quantificational reasoning, based on the well-understood
272 E. Zardini
notion of truth-in-a-model or truth-at-a-world, can be easier than modal

reasoning. But in neither case is the heuristics plausibly taken as evidence
of the grounding of modal facts in quantificational ones concerning
models or worlds.
45. Notice that this proposal does not require us to use L (in particular, its
tricky conjunction and implication) in our own theory: it’s sentences that
are logical truths, and so, on this proposal, we only need to mention—
rather than use—the relevant sentences and hence mention, rather than
use, the relevant operators and the logic governing them. Instead of
employing ourselves the straightjacket of [individual-truth] preserva-
tion, we refer to L’s own interpretation and evaluation of [collective-
truth] generation. (Don’t complain that this might give us less of a clue
as to which arguments are valid in L: the proposal is about what grounds
the fact that an argument is valid in L, not how we get to know it.) Having
noted this, as will emerge in particular in the next paragraph in the text,
the proposal does rely on a canonical specification of the relevant sen-
tences (grounding is sensitive to how the sentences are presented!), so
that its understanding does require appreciation of some general semantic
features of such sentences.
46. That this philosophical approach to grounding logical consequence for the
specific substructural logics alluded to in Sect. 2 is viable is confirmed by
technical facts concerning the formal definitions of some of these logics.
To give one glaring example (whose obvious import for the foundational
issues we’re investigating seems to me to have hitherto been grossly over-
looked), as I’ve already in effect remarked in section “Monotonicity”,
relevant logics are traditionally presented basically in terms of their logical
truths, and, when they are taken to be non-monotonic, their relation of
logical consequence is typically defined in terms of φ0, φ1, φ2 …, φi ⊢L
ψ holding iff φ0 → (φ1 → (φ2 … → (φi → ψ))) …) is a logical truth in L.
To give another glaring example (whose obvious import …), typical
non-contractive logics have a semantic presentation employing a certain
family of lattices (section “A Logic-Relative Version of GTT; an Absolute-
Truth Formulation of a Logic-Relative Version of GTT”): in such pre-
sentation, their relation of logical consequence is not defined, as is usual
in broadly algebraic treatments of logics, in terms of preservation of des-
ignated value from φ0, φ1, φ2 … and φi to ψ in every model (for, as per
the argument in section “A Logic-Relative Version of GTT; an Absolute-
Truth Formulation of a Logic-Relative Version of GTT”, that definition
would give the wrong results), but in terms of φ0 ⊗ φ1 ⊗ φ2 … ⊗ φi →
ψ getting designated value in every model (that is, being a logical truth).
47. While this deft movement from use to mention is illuminating, the light
it sheds is admittedly somewhat faint. But then there is only so much
light to be had at these depths.
48. Don’t say that the fact that φ0, φ1, φ2 …, φi ⊢L ψ holds is grounded
instead in the fact that the object-language, conjunction-free conditional
φ0 → (φ1 → (φ2 … → (φi → ψ))) …) is a logical truth in L (cf fn 46), so
that, in particular, the fact that, in LW, ‘Snow is white’, ‘Grass is green’
entail ‘Snow is white and⊗ grass is green’ is grounded in (the logical
necessity of ) the fact that, if→ snow is white, then→, if→ grass is green,
then→ snow is white and⊗ grass is green. Setting aside the adhocness of
such deviation given the route we’ve followed starting from GTT, the
deviation stumbles on exactly the same problem raised in the text when
grounding the fact that, in LW, by modus ponens, ‘If→ snow is white,
then→ grass is green’, ‘Snow is white’ entail ‘Grass is green’, since it
grounds it in (the logical necessity of ) the tautologous fact that, if→, if→
snow is white, then→ grass is green, then→, if→ snow is white, then→ grass
is green. Thanks to José Martínez for comments on some of these issues.
49. Semantic ascent could probably not fulfil this function if alethic defla-
tionism held. Ergo, by modus tollens …
50. In turn, it is plausible that logical truth, even in those logics, is not primi-
tive, and we may expect that its account will appeal to properties of the
logical operations used or mentioned in a logical truth (the latter disjunct
being particularly relevant when the logical truth involves [collective-
truth]-generation conditionals). If so, whether structural properties are
valid is grounded in whether certain sentences involving [collective-
truth]-generation conditionals are logical truths (see fn 51 for a specific
proposal), which is in turn grounded in the properties of logical operations.
Substructurality is a logically interesting but philosophically shallow
phenomenon caused by logically hidden but philosophically active
underlying logical operations. Structure is grounded in operations.
Thanks to Ricardo Miguel for pushing me on this.
51. Although a full treatment of the status of metarules (and metametarules,
and metametametarules …) lies beyond the scope of this chapter, a natural
way of extending the approach we’ve been pursuing to metarules for the
specific substructural logics alluded to in Sect. 2 is to say that the fact that a
metarule is valid (over and above its being admissible) in L is grounded in
the fact that the conditional having as antecedent the conjunction of the
series of [collective-truth]-generation conditionals corresponding to the
‘premise rules’ of the metarule and as consequent the [collective-truth]-gen-
eration conditional corresponding to the ‘conclusion rule’ of the metarule is
274 E. Zardini
a logical truth in L (and then iterate this strategy for metametarules, metam-
etametarules, metametametametarules …). Thanks to Bogdan Dicher
and Lucas Rosenblatt for their questions about the status of metarules.
References
Anderson, A., and N. Belnap. 1975. Entailment. Vol. I. Princeton: Princeton
University Press.
Asenjo, F. 1966. A Calculus of Antinomies. Notre Dame Journal of Formal Logic
7: 103–105.
Beall, J., ed. 2007. Revenge of the Liar. Oxford: Oxford University Press.
———. 2015. Free of Detachment: Logic, Rationality, and Gluts. Noûs 49:
410–423.
Beall, J., and G. Restall. 2006. Logical Pluralism. Oxford: Oxford University
Press.
Bočvar, D. 1938. Ob odnom trexznačnom isčislenii i ego primenenii k analizu para-
doksov klassičeskogo rasširennogo funkcional’nogo isčislenija. Matematičeskij
sbornik 4: 287–308.
Bolzano, B. 1837. Wissenschaftslehre. Vol. II. Sulzbach: Seidel.
Brouwer, L. 1927. Über Definitionsbereiche von Funktionen. Mathematische
Annalen 97: 60–75.
Chisholm, R. 1966. Freedom and Action. In Freedom and Determinism, ed.
Keith Lehrer, 11–44. New York: Random House.
Cobreros, P., P. Égré, D. Ripley, and R. van Rooij. 2012. Tolerant, Classical,
Strict. Journal of Philosophical Logic 41: 347–385.
Dummett, M. 2000. Elements of Intuitionism. 2nd ed. Oxford: Oxford University
Press.
Etchemendy, J. 1988. Tarski on Truth and Logical Consequence. The Journal of
———. 1990. The Concept of Logical Consequence. Cambridge, MA: Harvard
University Press.
Field, H. 2008. Saving Truth from Paradox. Oxford: Oxford University Press.
———. 2015. What Is Logical Validity? In Foundations of Logical Consequence,
ed. Colin Caret and Ole Hjortland, 33–69. Oxford: Oxford University Press.
Fine, K. 1975. Vagueness, Truth and Logic. Synthese 30: 265–300.
Girard, J. 1995. Linear Logic: Its Syntax and Semantics. In Advances in Linear
Logic, ed. Jean-Yves Girard, Yves Lafont, and Laurent Regnier, 1–42.
Jaśkowski, S. 1948. Rachunek zdań dla systemów dedukcyjnych sprzecznych.

Studia Societatis Scientiarum Torunensis 1: 55–77.
Kaplan, D. 1989. Demonstratives. In Themes from Kaplan, ed. Joseph Almog,
John Perry, and Howard Wettstein, 481–563. Oxford: Oxford University
Press.
Kleene, S. 1938. On a Notation for Ordinal Numbers. The Journal of Symbolic
Logic 3: 150–155.
Makinson, D. 1965. The Paradox of the Preface. Analysis 25: 205–207.
Martin, E., and R. Meyer. 1982. S (for Syllogism). ms.
Moruzzi, S., and E. Zardini. 2007. Conseguenza logica. In Filosofia analitica,
ed. Annalisa Coliva, 157–194. Rome: Carocci.
Ono, H. 2003. Substructural Logics and Residuated Lattices—An Introduction.
In Trends in Logic. 50 Years of Studia Logica, ed. Vincent Hendricks and Jacek
Malinowski, 193–228. Dordrecht: Springer.
Paoli, F. 2002. Substructural Logics: A Primer. Dordrecht: Kluwer.
Popper, K. 1948. On the Theory of Deduction, Part II. The Definitions of
Classical and Intuitionist Negation. Indagationes Mathematicae 10: 44–54.
Prawitz, D. 2005. Logical Consequence from a Constructivist Point of View. In
The Oxford Handbook of Philosophy of Mathematics and Logic, ed. Stewart
Shapiro, 671–695. Oxford: Oxford University Press.
Priest, G. 2006. In Contradiction. 2nd ed. Oxford: Oxford University Press.
Putnam, H. 1968. Is Logic Empirical? In Proceedings of the Boston Colloquium
for the Philosophy of Science 1966/1968, Vol. V of Boston Studies in the
Philosophy of Science, ed. Robert Cohen and Marx Wartofsky, 216–241.
Dordrecht: Reidel.
Read, S. 1981. Validity and the Intensional Sense of ‘and’. Australasian Journal
———. 2003. Logical Consequence as Truth-Preservation. Logique et Analyse
183: 479–493.
Restall, G. 2000. An Introduction to Substructural Logics. London: Routledge.
———. 2005. Multiple Conclusions. In Logic, Methodology and Philosophy of
Science: Proceedings of the Twelfth International Congress, ed. Petr Hájek, Luis
Valdés Villanueva, and Dag Westerståhl, 189–205. London: College
Publications.
Tarski, A. 1930. Fundamentale Begriffe der Methodologie der deduktiven
Wissenschaften. I. Monatshefte für Mathematik und Physik 37: 361–404.
———. 1936. O pojęciu wynikania logicznego. Przegląd filozoficzny 39: 58–68.
Wittgenstein, L. 1953. Philosophische Untersuchungen. Oxford: Blackwell.
276 E. Zardini
Wright, C. 2000. Cogency and Question-Begging: Some Reflections on

McKinsey’s Paradox and Putnam’s Proof. Philosophical Issues 10: 140–163.
Zardini, E. 2008. A Model of Tolerance. Studia Logica 90: 337–368.
———. 2011. Truth without Contra(di)ction. The Review of Symbolic Logic 4:
498–535.
———. 2012. Truth Preservation in Context and in Its Place. In Insolubles and
Consequences, ed. Catarina Dutilh-Novaes and Ole Hjortland, 249–271.
London: College Publications.
———. 2014a. Context and Consequence. An Intercontextual Substructural
Logic. Synthese 191: 3473–3500.
———. 2014b. Confirming the Less Likely, Discovering the Unknown.
Dogmatisms: Surd and Doubly Surd, Natural, Flat and Doubly Flat. In
Scepticism and Perceptual Justification, ed. Dylan Dodd and Elia Zardini,
———. 2014c. Evans Tolerated. In Vague Objects and Vague Identity, ed.
Kensuke Akiba and Ali Abasnezhad, 327–352. Dordrecht: Springer.
———. 2014d. Naive Truth and Naive Logical Properties. The Review of
———. 2015a. Breaking the Chains. Following-from and Transitivity. In
Foundations of Logical Consequence, ed. Colin Caret and Ole Hjortland,
———. 2015b. The Opacity of Truth. Topoi 34: 37–54.
———. 2016. Restriction by Noncontraction. Notre Dame Journal of Formal
Logic 57: 287–327.
———. 2018a. Forthcoming. Closed without Boundaries. Synthese.
———. 2018b. Forthcoming. Instability and Contraction. Journal of
Philosophical Logic.
———. 2018c. Forthcoming. The Underdetermination of the Meaning of
Logical Words by Rules of Inference. In The A Priori: Its Significance, Grounds,
and Extent, ed. Dylan Dodd and Elia Zardini. Oxford: Oxford University
Press.
———. 2018d. Changing without Contra(di)ction. ms.
———. 2018e. The Bearers of Logical Consequence. ms.
———. 2018f. Unstable Knowledge. ms.
Logical Particularism
Gillman Payette and Nicole Wyatt
1 Introduction
Methodology in logic has, on the whole, conceived of a logic as a system
of rules or principles. A correct logic is one which includes correct prin-
ciples, and logical reasoning is reasoning in accord with such correct prin-
ciples. The contemporary versions of this approach use a mathematically
precise formal language to express these principles, which are understood
similarly to the mathematical expression of laws of nature. We call this
orthodox view ‘logical generalism’.
The assumption of generalism is so broadly made that it is difficult to
make sense of giving it up. Generalism might seem to be constitutive of
G. Payette
Department of Philosophy, University of British Columbia,
Vancouver, BC, Canada
e-mail: gpayette@dal.ca
N. Wyatt (*)
Department of Philosophy, University of Calgary, Calgary, AB, Canada
e-mail: nicole.wyatt@ucalgary.ca

https://doi.org/10.1007/978-3-319-98346-2_12
278 G. Payette and N. Wyatt
logic, and the idea of a non-generalist approach to logic incoherent. In

this chapter, we aim to undermine the idea that logic and generalism go
hand in hand by exploring in detail an approach to logic that is particu-
larist rather than generalist.
We subscribe to five central theses. First, we are anti-exceptionalists
about logic—that is, we take the study of logic to be contiguous with
other sciences. Second, we reject the idea that logical theories aim at cor-
rectness. Third, we endorse a view of explanation on which scientific
theories need not be correct in order to be explanatory. Fourth, we deny
that natural language inferences have logical forms independently of the
use of a particular formal logic to model them. Fifth, we take logical
validity to be a property of particular inferences/arguments, rather than a
property of logical forms or schema.
We have discussed the first three in Payette and Wyatt (2018) and the
fourth in Wyatt and Payette (2018). Our focus in this chapter is on five.
What is the particularist understanding of validity, and in what way does
formal logic contribute to our understanding of valid inference? We also
compare our view to that of the local logical pluralist.
2 Validity
Our concern is with the validity of natural language arguments. This is
not to deny that formal logics have many other important applications,
nor that logical systems which have no relevance to natural language
may be of considerable interest. But it is only when logics are under-
stood as theories of a consequence relation in natural language that the
debates between logical monists and logical pluralists make sense: for
any logics to be correct there has to be something that they are correct
with respect to. Following Priest (2014), we call this the canonical appli-
cation for logic.
Monists and pluralists agree that at least one logic is a correct theory
of natural language validity. In contrast, logical nihilists and logical par-
ticularists share the view that no logic provides such a theory. For nihil-
ists like Curtis Franks, this is a reason to give up the canonical application
entirely:
Logical Particularism 279
fixing our sights on [natural language validity] saddles logic with a burden
that it cannot comfortably bear, and that logic, in the vigor and profundity
that it displays nowadays, does and ought to command our interest pre-
cisely because of its disregard for norms of correctness. (Franks 2015, 148)
From Franks’ point of view, focus on the canonical application is both

futile and pernicious. With respect to the latter, he argues that the pursuit
of a correct logic obscures the fact that much of the real interest in logic
lies in the interrelations between logics. For example, consider the enor-
mous, indeed uncountable, variety of logics that appear by removing
from classical propositional logic the law of excluded middle, thus gener-
ating intuitionistic propositional logic and those logics in between the
two. Logicians can focus on particular ways of presenting these logics in
terms of proof systems, and indeed on particular kinds of proof systems,
for example, sequent calculi. They can show that the effect of removing a
logical law can be reproduced by changing the structural parameters of
one system that has all of the same proof rules for the connectives: allow-
ing multiple formulas on the right-hand side of the sequents gives classi-
cal logic, while restricting the right-hand side to just one formula gives
intuitionistic logic.1 One of the crucial observations logicians can make
from this perspective is that classical logic collapses many distinctions
that one may want to make. The most famous example: ¬¬p is logically
distinct from p in intuitionistic logic, but not in classical logic. These
results, and many like them, are part of our logical knowledge, but,
Franks argues, they cannot be understood as part of some correct logic
for natural language (Franks 2015, 154–5).
With respect to the futility of pursuing the canonical application,
Franks argues that even with respect to our most common and successful
patterns of reasoning, we have no grounds for maintaining that these pat-
terns are anything other than useful:
it is a preconception that science is made possible by ahistorical norms of

right reasoning. Once one considers the possibility that logic may be stud-
ied with patterns of thought adapted to what we learn along the way, it
becomes hard to understand what special status the rudimentary principles
we find ourselves reflexively appealing to are supposed to have. (Franks
2015, 159)
Franks’ position, then, is that no logical system is likely to be correct.

First, because they are put together partly by accident, and largely by
efficacy. The dominant systems do not have—or cannot be seen to have,
at least—some special metaphysical status. And second, because there is
no evidence that any of the classes of inference rules we gather together
have some property that makes them the valid rules: “the evidence sug-
gests that the arrangement of our [logical] toolkit is a highly contingent
matter” (Franks 2015, 159).
In contrast, as particularists we are realists about the existence of a con-
nection between premises and conclusions of valid arguments, and we
think that logic is meant to help us study that connection. That is, we
take it that there is a fact of the matter as to whether, for the utterance of
a specific argument, that argument is valid or not. However, contra the
standard view, we maintain that validity is a property of argument tokens,
not argument types. Further, we agree with Franks that there is no feature
or group of features that are sufficient or necessary for validity and also
agree that there are no correct logical systems. We do not however share
his view that the only way in which logic could contribute to our under-
standing of natural language is by offering a correct theory of true logical
principles.
On our view, logic studies the connections between premises and con-
clusions of valid arguments, which we understand in preservationist
terms. While logical orthodoxy takes validity to be a matter of truth pres-
ervation between premises and conclusions, preservationists take a
broader perspective on the sorts of features which a consequence relation
can preserve.2 For the particularist, valid arguments are those in which
there is some feature which obtains of the conclusion in virtue of obtain-
ing of the premises. However, we treat that feature as variable. That is to
say, some particular concrete arguments are valid because they preserve
truth, some because they preserve relevance, and so on. Moreover, when
one makes a syntactically identical argument on a subsequent occasion, it
may fail to be valid, because on that particular concrete occasion, preserv-
ing truth (or relevance) may not be important to validity, whereas on the
prior occasion it was. Logic can study both the wide variety of features
that valid arguments preserve and the connections between premises and
conclusions without needing to meet the burden of correctness Franks
rightly rejects.3
3 Logics
Our view is that formal logics are mathematical models of validity; admit-
tedly, that is not an uncommon view. But there are important distinc-
tions between our view and other versions of the logic-as-modelling view.
Consider, as a representative example, the view of Roy Cook (2002). On
his account logic aims to model natural language and in particular argu-
ments constructed in natural language—including technical languages
like that of mathematics.4
According to Cook, logical models—proof theory, model theory, for-
mal language—contain certain mathematical objects that accurately rep-
resent some aspect of language/reality. These accurate representations are
the representors.5 Other components of the models are merely mathe-
matical ‘artifacts’ not intended to represent any part of reality. The stan-
dard metaphor is that of a model ship: the model may have certain
components which do not exist in the real ship, and may lack certain
components of the real ship.
Which elements are the representors and which the artifacts in a model
depends on what we want a model to do. For example, a model ship may
not have engines. If we are looking for a model that represents the exte-
rior of a ship, engines are unnecessary. But another model that is meant
for hydrodynamic ‘tank tests’ of the ship may also lack engines, but
include a lead block where the engines are in the actual ship to represent
the position and mass of the engines (though not the moving parts of the
engines since those don’t matter for this purpose). The real ship does not
have a lead block where its engines are; but the lead block is a representa-
tion of a particular, relevant aspect of the real boat. But now suppose we
want to put a model of the ship on a shelf, and we want to make sure that
it doesn’t fall over, so we put in a lead block which happens to be where
the engines are. The representors in this case are again the details on the
exterior of the model. Before, the block was a representor; now it is an
artifact.
Given this understanding of modelling, in order for a logic to correctly
represent validity in natural language, the elements of the model that
stand in relations of consequence must be representors rather than arti-
facts. This means the formulas of the formal languages must represent the
sentences of natural language. On the standard picture, they do so by
representing the logical forms of these sentences. Among other things,

this means that natural language sentences must have logical forms; that
every natural language sentence represented by a given formula must have
the same logical form; and that the formulas must also be composed of
parts which themselves accurately represent the parts of the logical forms.6
There are a number of difficulties with this picture of logics as models.
First, the distinction between representors and artifacts assumes that
models represent, to the extent that they do, via their parts. It is true that
in the model ship on the shelf the lead block doesn’t represent the engines.
But it does combine with rest of the model ship structure to represent the
usual (and preferred!) orientation of the actual ship. The lead block is not
by itself a representor, but classifying it as an artifact seems a mistake: it
does play some role in how the model functions as a model of the ship.
The distinction rests on an overly simple picture of modelling; impor-
tantly, it may be the outcome of interactions between components that
represents instead of a single component. Even in the case of hydrody-
namic tank testing of model ships, it is really the interaction between the
shape of the ship and the distribution of mass in the ship that is playing
the key role. Representation can be an emergent property—A, B, and C
can be combined to represent something without being used individually
to represent.
Second, in the specific case of logics, the picture advocated by Cook
takes for granted that the formulas of formal logics are representors, not
artifacts, and that what they represent are the logical forms of natural
language sentences. However, as we have argued in detail elsewhere
(Wyatt and Payette 2018), there is no basis for ascribing logical forms to
natural language sentences independently of the use of a particular logical
system. Ascriptions of logical form are theory-laden.
From the point of view of the particularist, the idea that models, or
parts of models, explain by representing is a mistake. We think the idea
that the model and its components must represent is driven by the
assumption that only correct theories can be explanatory. Our view is
that logical theories cannot explain validity on theories of explanation in
which correctness is a prerequisite.7 Once we have diverged from think-
ing of explanatory logical theories as correct, we can more easily question
the requirement that models represent.
4 Modelling
We have denied that logics represent features of natural language. Contra
the nihilist, we maintain that logics do shed explanatory light on natural
language consequence. To understand how non-representational models
can explain, we turn to modelling in the sciences. We focus on three key
features of explanatory but inaccurate or non-representative models:
abstraction, imagination, and scaffolding.
As Godfrey-Smith observes (2009), scientists from a variety of disci-
plines spend time considering things that don’t exist: examples include
ideal gases, frictionless planes, infinitely large biological populations,
wholly rational agents, and biologically implausible neural networks.
Theoretical investigation of these apparently increases our understanding
of the behavior of actual gases, the movement of physical objects, evolu-
tion in finite populations, human behavior, and human and animal cog-
nition. All of these are part of model-based science, in which one system is
explored as a basis for understanding another system.
This seems puzzling—how can studying things that do not exist lead
to knowledge of things that do? Unsurprisingly, philosophers of science
disagree on exactly how ‘false’ models—models that rest on known false
assumptions—function in scientific practice. But there are a number of
salient themes in the literature.
Firstly, many false models rest upon abstraction in various ways. When
we posit ideal gases and wholly rational agents, we are abstracting away
from details which matter in the actual world. Abstraction depends on
making false assumptions and ignoring details, but increases the
tractability of the problem. For example, in ecology scientists face two
obstacles for studying real systems: (a) the time scale on which they oper-
ate tends to exceed the time available for study, and (b) their complexity
makes it difficult to manipulate them systematically (Odenbaugh 2005,
233). Abstract mathematical models in ecology allow both problems to
be overcome. Moreover, even when a more detailed and perhaps accurate
model is available, an abstracted model may be more cognitively and
computationally tractable, and thus more useful, as Woody (2013,
1574–5) argues with respect to the ideal gas law.8
Secondly, false models support contrastive explanations. In many

cases, scientists seek to answer not ‘why this?’ but ‘why this rather than
that?’. Consider Galileo’s theory of falling bodies, which predicts that all
bodies will fall at the same speed. Of course that is not what we observe
in the actual world. But Galileo’s theory allows us to explain why in our
world anvils fall faster than feathers precisely via attention to the unreal-
istic assumption that falling happens in a vacuum. Galileo’s false model
points us to the correct explanation—air resistance—and away from the
incorrect one—mass (Hindriks 2008, 342–3).9
Another aspect of this feature of false but productive models is that
they support counterfactual reasoning of various sorts. Bokulich (2011),
discussing Bohr’s model of the hydrogen atom, points out that while the
model is fictional, it allows us to answer a wide range of ‘what-if-things-
had-been-different’ questions. There is, she suggests, a pattern of coun-
terfactual dependence between the emission spectrum of hydrogen and
the elements of Bohr’s model (Bokulich 2011, 43–44). Odenbaugh, dis-
cussing ecology, calls this how-possibly modelling (Odenbaugh 2005,
236–7). Both contrastive explanations and counterfactual reasoning are
made possible by the support models provide for imagination.
Finally, false models may provide explanatory scaffolding for scientists.
In ecology, mathematical models of communities led to the articulation
of a number of concepts of community stability and complexity, and thus
to alternative hypotheses concerning the relationships between them, all
of which have gone on to provide an improved conceptual framework for
studying actual populations (Odenbaugh 2005, 250). False models can
also guide scientists in the construction of explanations by exhibiting the
shape that good explanations should take. In chemistry, the ideal gas law
provides, among other things, inferential scaffolding, in that it allows
chemists to conceptualize actual gas properties as deviations from the
ideal, and thus unifies the treatment of all gases under a single false model
(Woody 2013, 2015).
Scaffolding can be seen at work in graphical models as well: Woody
(2014) shows how the visual organization of the periodic table functions
like a map. It helps chemists navigate an aspect of their discipline “not by
presenting the most detailed representation possible but by highlighting
features of the landscape that are taken to be significant” (Woody 2014,
142). And the table can function in this way despite the fact that the so-
called periodic law that underlies it does not seem to reveal any causal
structure or identify any mechanisms, nor is it precise enough to support
scientific predictions (Woody 2014, 133–4, 143).10
Some arguments for logical nihilism are motivated by the failure of
formal logic(s) to offer accurate predictions for natural language conse-
quence. But this assumes that the only roles that can be played by logics
with respect to natural language validity are prediction and description.
Franks’ argument for nihilism is, in part, that no system can capture the
variety of logical properties that we might be interested in. But he still
thinks that were a logic useful for explaining natural language validity, it
would have to be correct, that is, as a theory it would be true or as a
model it would have to be accurate. In contrast, the literature on scien-
tific modelling suggests that false models and theories can play a role in
scientific explanation in a variety of ways.
5 Logicians
One common thread in our rapid survey of false models in science is the
attention paid to the way in which scientists make use of these models,
and the roles that abstraction, imagination, and scaffolding play. This
echoes Ronald Giere’s views about modelling in general, which character-
ize representation as something scientists do with models for specific pur-
poses, rather than as a timeless property of a model (Giere 2004, 743).11
When it comes to logics, the question is not whether the formal sys-
tems correctly represent some feature of natural languages. The right
question is how these models are used to explain. Our claim is that logi-
cians interested in natural language use logics to represent in order to
accomplish their explanatory goals.12 Meeting those goals involves
abstraction, imagination, and scaffolding. It turns out that this perspec-
tive makes sense of three practices of logicians which at first may not
make sense on ‘representationalist’ accounts of modelling and veridical
explanatory theories.
Models in logic can be any or all of formal languages, proof theories,
or model theories. Given the target phenomena of logical study we are
considering, formal languages are rather important. Those languages

carry the relation of consequence. But we must not forget the wide vari-
ety of models in use by logicians. There are various versions of formal
language: for example, first-order versus propositional. There are many
kinds of model theory: for example, algebraic versus possible worlds. And
of course there are various kinds of proof theories: for example, Fitch
versus axiomatic.
That variety brings us to our first practice: generating many kinds of
models used for specialized purposes. For example, modal logicians use
multiple separate propositional languages to capture fragments of natural
language: for example, temporal and deontic. These idealizations allow
logicians to focus on certain phenomena rather than needing to express
the modal connections in terms of first-order formulas, or deal with
interactions between the various modal operators. Abstraction is essential
to tractability, just as in other scientific contexts.
Of course there is considerable investigation of the relationship
between modal logics and first-order logic. This manifests in the practice
of translating modal formulas into first-order formulas that express con-
ditions on Kripke frames.13 But logicians also understand the limitations
of that connection: certain first-order conditions on frames—like all
worlds are related—cannot be represented via a modal formula. That
brings us to another of the roles of modelling: imagination.
Logicians use models to understand or imagine the variety of ways
things might be. The natural language arguments are the only things that
are observable in the science of logic: we do not see the semantics, we do
not see logical forms, we do not see generalized proof rules. All of those
are models that represent a way of thinking about validity. But we try to
draw connections between different arguments; we try to extend our
understanding in one case to other cases. Logicians also ask how it is that
things could be different between arguments. These differences could be
in the way the world is, but they could also be differences in what matters
in the connection or disconnection between premises and conclusions.
The model and proof theories are used to make sense of those questions.
It is here that the second practice of logicians becomes particularly
salient: representation theorems. Representation theorems are important
in mathematics. They allow one to draw connections between different
structures. In social choice theory, one lays out a set of axioms for how,
for example, a choice function behaves, and then shows that those axioms
correspond to a certain kind of preference relation.14 In logic, we see rep-
resentation theorems in a variety of forms.
One form we have already mentioned: modal correspondence theory.15
Some object language formulas, when true at every world in every model
on a frame, mean that the relation in that frame has a certain property:
transitivity, reflexivity, and so on. But the most common representation
theorem one sees in logic is completeness.16 Generally completeness
involves the representation of a certain semantic notion of consequence
in terms of a proof-theoretic form of consequence—or semantic truth in
terms of theoremhood.17 Another form of representation theorem is
when one looks for a set of proof rules, usually in a sequent calculus, that
can be characterized by a certain set of properties. Such characterizations
were what Michael Dummett used to argue for the supremacy of intu-
itionistic logic in (Dummett 1991). How do these connect to
imagination?
Representation theorems show the range of variation for a given use of
a model. A completeness theorem shows which logical relationships
between formulas are in and which are out. If we want to include differ-
ent forms, we know we have to change the model theory in some way.
Similarly for sets of sequent proof rules. If we want the set of proof rules
to have an extra property, we know that the consequence relation will
have to change. Correspondence theory shows a closer connection
between model theory and formal language than completeness theorems.
It shows that sometimes general facts about the model theory can be
brought into the object language: properties about the frame relation can
be expressed by the formulas. If we want to change the model theory,
then, we know the logically true formulas will change.
The final role of models, scaffolding, explains why logicians haven’t
discarded classical propositional logic (CPL) despite its limitations. They
add to it and subtract from it in various ways. Sometimes they add prop-
ositional connectives as in modal logics. Sometimes they model sub-
propositional structure as in first-order logic or branching quantifier
logics. Subclassical logicians may weaken it by rejecting, for example,
excluded middle. Logicians give systems which differ from CPL, and
philosophers of logic tend to debate about the virtues and vices of rival
logics in their propositional forms. We know the fit between proposi-
tional languages and natural language arguments isn’t perfect, but we
keep them around. More importantly, logicians don’t get rid of CPL: it is
what every logician cuts her teeth on. It also acts as a logica franca for
metalogical investigations into logics.
CPL acts as a conceptual scaffolding for understanding other logics.
CPL is really big: you can prove a lot using it. Most of the propositional
logics that are studied are sub-logics of CPL. That is largely because it is
post-complete18 and it is structurally complete.19 Often, logics are studied
by how their proof theory or semantics must be different from those for
classical logic; it is our logical jumping-off point.
In the end, we see all of those modelling practices from other sciences,
which also use inaccurate models, at work in the practices of logicians.
But all of our talk about applications and modelling may be reminiscent
of another perspective in the philosophy of logic that we want to distin-
guish from our view: pluralism.
6 Pluralism
Let us take stock. Like the nihilist, we are convinced by the existence
of counterexamples to every putative logical law that no logic describes
a consequence relation in natural language. Unlike the nihilist, we do
not conclude that logic contributes nothing to our understanding of
natural language validity. According to the particularist, logics can
and do reveal something about the genuine connection between prem-
ises and conclusions in natural language. Logics are best understood as
explanatory but inaccurate/non-representative/fictional models; they
function like other ‘false’ models in the sciences. Before we turn to
pluralism we have to highlight a couple of important features of
particularism.
The common understanding of logics as models assumes, one, natural
language utterances have logical forms, and two, what matters to conse-
quence is stable across contexts and arguments. We think both assump-
tions are false. The former is argued against elsewhere.20 The latter is a
common ground for the varied particularist projects. As Dancy puts it,
“[t]he core of particularism is its insistence on variability” (2013).21
Whether moral or logical, particularism carries this commitment to
the variable relevance of features. In the logical context, having the form
of modus ponens, under some system of logical analysis, is a relevant
feature to the assessment of some arguments, but not in the assessment of
others. Faced with counterexamples, the particularist feels no need to
offer increasingly recondite analyses of the arguments in order to escape
the conclusion that modus ponens is invalid.22 As Dancy emphasizes, the
particularist takes the view that moral—or in our case logical—reasons
function in ways “not noticeably different from the way in which other
reasons function—more ordinary reasons for action, say, or reasons for
belief rather than for action” (Dancy 2013).
We now want to say a little more about what separates our view from
that of the pluralist. Following Haack (1978), we distinguish between
global pluralism—the view that there are multiple logics which are correct
come what may—and local pluralism, the view that logics are only correct
relative to a particular domain of inquiry. Much of the more recent work
on pluralism takes the view that global pluralism is the only pluralism
worthy of the name, on the grounds that it is only the global pluralist
who claims that a specific argument could be simultaneously valid and
invalid.23 But Shapiro (2014) offers a robust defense of a particular ver-
sion of local pluralism.
Global pluralists and monists share the common assumption that there
is a general-purpose consequence relation for natural language that
logic(s) aim to describe. They differ on how many such relations there
are, but agree that any such relation applies in any context of reasoning.
In contrast, the local pluralist maintains that there are different
consequence relations that apply in different contexts of reasoning. This
might, on first glance, look very much like the particularist view. After
all, we have said that validity is a matter of preserving some property or
other in the move from premises and conclusion, and that what property
must be preserved changes from case to case.
The point of fundamental disagreement between us and the local plu-
ralist lies with the notion of correctness. While the local pluralist allows
what matters to consequence to vary from context to context, they think
the aim of logic is accurate prediction and adjudication for natural lan-
guage validity. To avoid the pitfalls of the global pluralist, the local plural-
ist needs some way of demarcating the area where each correct logic is in
force. For example, a local pluralist might maintain that there are differ-
ent logics for different areas of inquiry (which might be epistemologically
or ontologically distinguished); or different logics for different sets of
logical connectives (and perhaps that a single natural language connective
can represent different logical connectives); or different logics for differ-
ent sets of possible truth values (e.g. one might think that claims about
the natural world must either be true or false, but that vagueness in social
facts requires a plurality of truth values).
The need for demarcation makes local pluralist vulnerable to the
arguments of the nihilist. As Russell (2017) argues, we can ‘validate’
any purported logical law by restricting the substitution class into that
law. For example, if we allow only true sentences to be substituted for
φ and ψ in the form φ → ψ, ψ ⊨ φ, then this ‘law’—affirming the con-
sequent—would seem to be correct. That isn’t a convincing justification
of affirming the consequent precisely because the restriction on substi-
tution is ad hoc. But then the question becomes, what restrictions are
legitimate?
The local pluralist needs a method of demarcation independent of any
judgments about validity. For example, the local pluralist cannot restrict
substitution to sentences where the laws of classical logic apply, and use
that as an argument that classical logic is appropriate for that domain.
Such a demarcation may be possible, but we are skeptical because any
such demarcation would depend upon a logic neutral account of logical
form.24
In contrast, the particularist has no need of an independent means of
determining which logic applies to a particular argument; from our
perspective the role of logic is not to adjudicate or predict validity but to
explain it. Features and properties have variable relevance to validity, and
we cannot determine in advance of considering a specific argument which
properties are relevant. But, once we have decided what is relevant, logics
can explain how the relevant validity-making properties could be pre-
served between premises and conclusions.25
7 Particularist Logic
One natural objection to our view rests on the idea that logics are not just
explanatory but also action guiding. Certainly logicians, qua scientists,
may act as we have described, but non-logicians use logic to give argu-
ments. Moreover, they use formal logic to give and check their arguments.
That means they use a logic’s rules to ensure the validity of their argu-
ments by following the rules of that logic. Now for what might seem to
be a U-turn: this practice can be acceptable.
We take logic to be the study of connections between premises and
conclusions. Formal systems show how connections between premises
and conclusions can be made. In classical logic, for example, we use a
conception of ‘connection’ as truth preservation. Classical logic also pro-
vides a precise account of what ‘truth preservation’ means.26 This model
sets a narrow boundary for what can count when evaluating an argument
for validity. The flaw in the generalist way of thinking is the assumption
that model must be applied everywhere, always.
What about premises that can’t all be true simultaneously? Something
other than truth must matter in that case. Here we see logic not just as a
science, but as engineering. We look for some property that might suit
our purposes, and then we see what we can do to model it as a connec-
tion. A favorite example is the notion of level of consistency or just
‘level’—from Jennings and Schotch (1984). Level is the answer to the
question: what is the size of the smallest partition of the set in which all
the sets in the partition are consistent? It measures inconsistency. The
level of {p, ¬p} is two, because it takes at least two sets in a partition to get
consistent sets. The level of {q ∧ p, p ∧ ¬ q ∧ r, p ⊃ (¬q ∧ ¬ r)} is three.
One relation that preserves level is defined by saying t follows from a set
of premises Γ when t follows classically from some set in every partition
of Γ that is the same size as the one used to determine the level of the set,
and each element of the partition is consistent.27 The level of the set of
consequences on this definition will be the same as the level of the
premises.
This level-preserving approach to inconsistent consequence builds one
model atop another as scientists are wont to do. It gives an altogether
different interpretation to ‘preserving’ since the property that is preserved

is between sets, not sentences, but why not? The truth property in the
classical model stops working in the case of inconsistency. This is not the
only place where something other than truth is used, but for different
reasons. In deontic logic it has long been recognized that rules are not
really true, and so some have tried to interpret satisfaction between mod-
els and formulas, not as truth, but as rule legitimacy—whatever that may
mean.28 Others have added speech act satisfaction to model theories for
speech acts, in addition to truth.29 With yet another focus, there are log-
ics concerned with metaphysical grounding rather truth, see Fine (2012).
These logics use the scaffolding provided by classical logic, just in differ-
ent ways.
In each of these cases, a way of connecting premises to conclusions is
chosen. A model is engineered to understand the way that connection
behaves across a range of applications. But each model focuses only on
that one connection. The particularist doesn’t change the methodology of
formal logic, she changes the interpretation of the end results to recog-
nize the limitations of the model. But when is it legitimate to use a logic
to construct an argument?
There is a wonderful property that some logics have: conservative exten-
sion. Start with a language L, with a relation of consequence ▷ between
the sentences of L. Then extend L to L* with some new bits.30 An exten-
sion of ▷ is a relation over L*, represented by ⊢, that contains ▷. We
say that ⊢ is a conservative extension of ▷ when for each argument (A,
b) in L, A ⊢ b iff A ▷ b. The new relation doesn’t make any arguments
that were not already valid in ▷, valid.
The particularist methodology is to offer an explanation of conse-
quence for (a fragment of ) natural language, and is basically the standard
methodology in logic. Let’s say ▷ is consequence in natural language.
The logician focuses attention on some subset of L, L′, and gives a model
that is extensionally adequate for the particular arguments in ▷ restricted
to L′. The elements of L′ are replaced by their mathematical models con-
structed from the new bits in L*. What we then get is a language made
from the union of L without the bits in L′, and L*. The arguments from
▷ restricted to L′ are replaced by their counterparts in L* that are related
by ⊢.31
The language of L* is formal, so it will have substitution instances

from outside of L′. ⊢ is an extensionally adequate explanation of the
arguments from L′; it may extend ▷, and that is why conservative
extension is important. Building arguments using the new parts from
L* allows her to be certain that she won’t offend the un-targeted portion
of ▷. The explanation of natural language argument is fragmentary; ⊢
has limited application. Any arguments that extend ▷ are given or
derived using rules which are chosen to characterize the abstract model,
not reality per se.
Argument construction proceeds, on the particularist account, by first
deciding what connection matters between premises and conclusions.
Then one deploys a logic developed to model that connection. Arguments
constructed with this process in mind aren’t guaranteed validity—par-
ticularism doesn’t assume logics are correct. The goal of the particularist
approach to argument construction is epistemological transparency: it
wears its commitments on its sleeve.32
8 Conclusion
Reasons have variable relevance. The reason that a glass dropped on a
cement floor broke is that it is fragile and the cement is hard. In another
case, I might drop an identical glass and it might not break. In this sec-
ond case, the fact that it is fragile and I dropped it on hard cement is not
a reason that the glass broke, for the simple reason that the glass is intact.
Dancy (2013) has pointed out that reasons can have this variability in the
moral context, and we are pointing out that they have that variability in
the logical context.
The logical particularist maintains that an argument analyzed as hav-
ing the form of modus ponens when using classical logic might well be
the reason it is valid in one case. In another case, an argument with that
form in classical logic is not valid, and the classical analysis provides no
reason for taking it to be.
Nonetheless, we think that the standard methodology of logic is pro-
ductive and contributes to the historical goal of explaining validity in
natural language reasoning. Logical particularists use logics as explanatory
tools and take the explanations to be on par with those offered by other
sciences. Logics do not explain validity on their own. In this chapter,
we’ve sketched an account of the functioning of logics as models which is
continuous with the use of such models in other scientific endeavors.
Specifically, we have argued that models in logic serve our explanatory
goals without aiming at, or succeeding at, representation.
Notes
1. Although we have argued elsewhere that giving the same rules to connec-
tives does not mean that the connectives have the same meaning qua
natural language connectives (Wyatt and Payette 2018), we do not doubt
the mathematical significance of the construction and the purely math-
ematical conception of proof-theoretic meaning used in the construc-
tion. It is the mathematical significance of the constructions that interests
Franks, if we have understood him correctly. We are not at odds on this
point.
2. See Payette and Schotch (2007), Jennings et al. (2009).
3. As will become clear from the subsequent discussion, we do reject the
view on which formal logic sheds light on the validity of arguments by
directly representing the semantics of natural language sentences. See
Stokhof (2007) for a discussion of the relationship between formal lan-
guages and natural language semantics that is somewhat parallel to the
view of the relationship between formal logics and natural language
argument advocated here.
4. Cook’s account of how logics model validity is very similar to that of
Shapiro (2014): the primary difference is that Shapiro takes logics to be
only in the business of modelling mathematical argument, while for
Cook mathematical argument is a special case of a more general goal.
5. Cook refers to models as ‘mock-ups’ (Cook 2002).
6. Notice that this picture is absolutely neutral on the nature of the rela-
tionship between the phonological realization or the syntax of natural
language sentences and their logical forms.
7. See Payette and Wyatt (2018) for more discussion of this point.
8. Abstraction here covers both idealization and fictionalization in science,
where fictional models are those in which no amount of de-idealization—
that is, no amount of adding in details left out or correcting false assump-
tions—would recover a correct description of the actual world (Bokulich
2011).
9. Hindriks (2008) also argues that models in the economics context
can be better when containing a larger number of unrealistic
assumptions.
10. The periodic table does not, contra the usual historical story, indicate
directly the presence of missing elements, nor does it generate any pre-
dictions of specific quantitative properties. See Woody (2014) for fur-
ther discussion.
11. Giere proposes that the representative features of models fall under the
general schema: S uses X to represent W for purposes P. Some of this
purpose relativity is already present in Cook’s view since which compo-
nents are representors and which are artifacts may depend on the goal of
the model. That goal is, presumably, imposed on the model. It is possible
that Cook would be in closer agreement than we have suggested here,
but note that critics, for example, Smith (2011), describe Cook’s view as
one where representation is a relation between (parts of ) models and
reality, rather than the 4-ary relation above. We won’t enter any further
into debates about how to interpret Cook here since our goal is to
advance our own view.
12. This is not a descriptive claim; that is, we are not claiming that this is
how logicians understand what they are doing, whether collectively or
individually. Rather this is a claim about how we should understand
logical practice. In line with our anti-exceptionalism, we are approaching
the epistemological and metaphysical questions in logic in the same way
in which the philosopher of biology approaches those questions in biol-
ogy. In both cases one must attend to the messy business of how science
actually gets done. For further discussion of our methodology, see Payette
and Wyatt (2018).
13. For example, □p ⊃ p is true at every world in every model on a Kripke
frame (W, R) iff ∀w, wRw.
14. See Sen (1970). One can also consider Cozic and Hill (2015) to under-
stand the larger role of representation theorems in science.
15. See van Benthem (1997) for an overview.
16. Here we are using ‘completeness’ in the (common) sense which includes
both completeness and soundness theorems.
17. See Franks (2010) for more on interpretations of completeness.
18. You cannot add a formula that isn’t a theorem to CPL as a new theorem
without being able to then prove everything.
19. All of the admissible rules are derivable rules. For more on that distinc-
tion, see Iemhoff and Metcalfe (2009).
20. See Wyatt and Payette (2018).
21. See, for example, the discussions in McKeever and Ridge (2005), Cullity
and Holton (2002) for moral particularisms, and Lakatos (1976) and
Larvor (2001, 2008) for particularism in mathematics. John Norton’s
material theory of induction is, we think, a particularist approach to
inductive argument (2003, 2010).
22. See Kolodny and MacFarlane (2010), Dreier (2009), Cantwell (2008),
or McGee (1985) for some possible counterexamples to modus ponens.
23. See, for example, Cook (2010), Beall and Restall (2006).
24. We argue in Wyatt and Payette (2018) that attributions of logical form
are only possible from within a logic.
25. Logical particularism can be seen as a form of instrumentalism, but not
at the level of theory justification. That is, we don’t take it that a ‘correct’
logic is the most useful one. Rather we take it that the practices and
methods of formal logic are instrumentally justified because of their
effectiveness in explanation.
26. Of course one could give other interpretations of the connection mod-
eled by classical logic, and some people disagree with the classical inter-
pretation of truth preservation, see, for example, Read (1994).
27. The examples above are not very interesting sets. Each has only its parti-
tion into unit sets, and so something follows when it follows classically
from at least one of the formulas.
28. This is at least true for certain parts of the models in Castañeda (1981).
29. This is done in Searle and Vanderveken (1985).
30. a.k.a. mathematically defined logical connectives.
31. Note that we did not say that ⊢ explicates the arguments from L′. We
agree with Woody, if we have interpreted her correctly, that what is
often called ‘explication’ is simply a stage in explanation. Models, par-
ticularly mathematical ones, become part of the understanding of
phenomena.
32. This is similar, we think, to the project in Field (2015) where commit-
ments to relationships between degrees of belief are translated into logi-
cal rules. Particularism assumes that logical connections can be about
things other than degrees of belief.
References
Press.
Bokulich, A. 2011. How Scientific Models Can Explain. Synthese 180 (1):
33–45.
Cantwell, J. 2008. Changing the Modal Context. Theoria 74 (4): 331–351.
Castañeda, H.N. 1981. The Paradoxes of Deontic Logic: The Simplest Solution
to All of Them in One Fell Swoop. In New Studies in Deontic Logic, ed.
R. Hilpinen, 37–85. Chichester: Wiley-Blackwell.
Cook, R.T. 2002. Vagueness and Mathematical Precision. Mind 111 (442):
225–247.
———. 2010. Let a Thousand Flowers Bloom: A Tour of Logical Pluralism.
Philosophy Compass 5 (6): 492–504.
Cozic, M., and B. Hill. 2015. Representation Theorems and the Semantics of
Decision-Theoretic Concepts. Journal of Economic Methodology 22 (3):
292–311.
Cullity, G., and R. Holton. 2002. Particularism and Moral Theory. Aristotelian
Society Supplementary Volume 76: 169–209.
Dancy, Jonathan. 2013. Moral Particularism. In The Stanford Encyclopedia of
Philosophy (Fall 2013 Edn.), ed. E.N. Zalta. Metaphysics Research Lab,
Stanford University.
Dreier, J. 2009. Practical Conditionals. In Reasons for Action, ed. D. Sobel and
S. Wall, 116–133. Cambridge: Cambridge University Press.
Dummett, M. 1991. The Logical Basis of Metaphysics. Harvard: Harvard
University Press.
Field, H. 2015. What Is Logical Validity? In Foundations of Logical Consequence,
ed. C.R. Caret and O.T. Hjortland, 33–70. Oxford: Oxford University Press.
Fine, K. 2012. The Pure Logic of Ground. Review of Symbolic Logic 5 (1): 1–25.
Franks, C. 2010. Cut as Consequence. History and Philosophy of Logic 31 (4):
349–379.
———. 2015. Logical Nihilism. In Logic Without Borders: Essays on Set Theory,
Model Theory, Philosophical Logic and Philosophy of Mathematics, ed.
A. Villaveces, R. Kossak, J. Kontinen, and A. Hirvonen, 147–166. Berlin: De
Gruyter.
Giere, R.N. 2004. How Models Are Used to Represent Reality. Philosophy of
Science 71 (5): 742–752.
Godfrey-Smith, P. 2009. Models and Fictions in Science. Philosophical Studies

143 (1): 101–116.
Haack, S. 1978. Philosophy of Logics. Cambridge: Cambridge University Press.
Hindriks, F. 2008. False Models as Explanatory Engines. Philosophy of the Social
Sciences 38 (3): 334–360.
Iemhoff, R., and G. Metcalfe. 2009. Proof Theory for Admissible Rules. Annals
of Pure and Applied Logic 159 (1–2): 171–186.
Jennings, R.E., and P.K. Schotch. 1984. The Preservation of Coherence. Studia
Logica 43 (1–2): 89–106.
Jennings, R.E., B. Brown, and P. Schotch, eds. 2009. On Preserving: Essays on
Preservationism and Paraconsistency. Toronto: University of Toronto Press.
Kolodny, N., and J. MacFarlane. 2010. Ifs and Oughts. Journal of Philosophy
107 (3): 115–143.
Lakatos, I. 1976. Proofs and Refutations: The Logic of Mathematical Discovery.
Larvor, B. 2001. What Is Dialectical Philosophy of Mathematics? Philosophia
Mathematica 9 (2): 212–229.
———. 2008. Moral Particularism and Scientific Practice. Metaphilosophy 39
(4–5): 492–507.
McGee, V. 1985. A Counter Example to Modus Ponens. Journal of Philosophy
82 (9): 462–471.
McKeever, S., and M. Ridge. 2005. The Many Moral Particularisms. Canadian
Norton, J.D. 2003. A Material Theory of Induction. Philosophy of Science 70 (4):
647–670.
———. 2010. There Are No Universal Rules for Induction. Philosophy of Science
77 (5): 765–777.
Odenbaugh, J. 2005. Idealized, Inaccurate but Successful: A Pragmatic Approach
to Evaluating Models in Theoretical Ecology. Biology and Philosophy 20 (2–3):
231–255.
Payette, G., and P.K. Schotch. 2007. On Preserving. Logica Universalis 1 (2):
295–310.
Payette, G., and N. Wyatt. 2018. How Do Logics Explain? Australasian Journal
of Philosophy 96 (1): 157–167.
Priest, G. 2014. Logical Pluralism: Another Application for Chunk and
Permeate. Erkenntnis 79 (SUPPL.2): 331–338.
Read, S. 1994. Formal and Material Consequence. Journal of Philosophical Logic
23 (2): 247–265.
Russell, G. 2017. An Introduction to Logical Nihilism. In Logic, Methodology,

and the Philosophy of Science: Proceedings of the Fifteenth International Congress,
ed. H. Leitgeb, I. Niiniluoto, P. Seppäla, and E. Sober, 1–10. London:
College Publications.
Searle, J.R., and D. Vanderveken. 1985. Foundations of Illocutionary Logic.
Sen, A. 1970. Collective Choice and Social Welfare. San Francisco: Holden Day.
Shapiro, S. 2014. Varieties of Logic. Oxford: Oxford University Press.
Smith, Peter. 2011. Squeezing Arguments. Analysis 71 (1): 22–30.
Stokhof, M. 2007. Hand or Hammer? On Formal and Natural Languages in
Semantics. Journal of Indian Philosophy 35 (5–6): 597–626.
van Benthem, J. 1997. Correspondence Theory. In Handbook of Philosophical
Logic, ed. D.M. Gabay and F. Guenthner, vol. 2, 2nd ed., 325–408.
Dordrecht: Klewer.
Woody, A.I. 2013. How Is the Ideal Gas Law Explanatory? Science & Education
22 (7): 1563–1580.
———. 2014. Chemistry’s Periodic Law: Rethinking Representation and
Explanation After the Turn to Practice. In Science After the Practice Turn in the
Philosophy, History, and Social Studies of Science, ed. L. Soler, S. Zwart,
M. Lynch, and V. Israel-Jost, 123–150. London: Routledge.
———. 2015. Re-Orienting Discussions of Scientific Explanation: A Functional
Perspective. Studies in History and Philosophy of Science Part A 52: 79–87.
Wyatt, N., and G. Payette. 2018. Logical Pluralism and Logical Form. Logique
et Analyse 241: 25–42.
Logical Nihilism
Aaron J. Cotnoir
Ordinary language has no exact logic.

Strawson (1950)
1 Outlining the View
The philosophy of logic has been dominated by the view that there is One
True Logic. What is meant by ‘One True Logic’ is sometimes not made
entirely clear—what is a logic and what is it for one of them to be true?
Since the study of logic involves giving a theory of logical consequence
for formal languages, the view must be that there is one true theory of
logical consequence. What it means for a logic to be true is, roughly, for
it to correctly represent something or other.1 What do logics represent? It
is clear from the various uses of applied logic, they can represent many
different sorts of phenomena. But for the purposes of traditional logic,
A. J. Cotnoir (*)
Department of Philosophy, University of St Andrews, St Andrews, UK
e-mail: ac117@st-andrews.ac.uk

https://doi.org/10.1007/978-3-319-98346-2_13
302 A. J. Cotnoir
though, theories of consequence are frequently taken to represent natural

language inference.
As Beall and Restall (2006) note:
Logic, in the core tradition, involves the study of formal languages, of

course, but the primary aim is to consider such languages as interpreted:
languages which may be used either directly to make assertions or denials,
or to analyse natural languages. Logic, whatever it is, must be a tool useful
for the analysis of the inferential relationships between premises and con-
clusions expressed in arguments we actually employ. If a discipline does not
manage this much, it cannot be logic in the traditional sense. (8)
If this is right, it means that logic is connected to inferential practice of

natural language speakers.2 Now perhaps logical theorizing is not an
entirely descriptive enterprise and contains some element of normativity.
But we think it is clearly true that a logic would have claims to be ‘correct’
only if it is constrained in some way by actual inferential practice. For the
purposes of this chapter, we will follow Beall and Restall (and others3) in
holding this characterization of logic as traditionally conceived.4
Once we have pinned down our subject matter, we see a number of
possible outcomes.
Logical Monism There’s exactly one logical consequence relation that

correctly represents natural language inference.
Logical Pluralism There’s more than one logical consequence relation

that correctly represents natural language inference.
Logical Nihilism There’s no logical consequence relation that correctly

represents natural language inference.
Much of the discussion in the philosophy of logic over the last decade
has been devoted to the debate between logical monism and logical plu-
ralism. But logical nihilism hasn’t been given nearly as much attention,
even though the view has historical roots and is philosophically
defensible.
Logical Nihilism 303
A clarification: there is another view which can rightly be called

‘Logical Nihilism’ which is not the view under consideration here.
Mortensen (1989) has argued that anything is possible based on a kind of
thorough-going empiricism. On one reading, Mortensen believes every-
thing is logically possible.5 And if everything is possible, then for any
argument from premises X to conclusion A there will be some possible
case according to which X is satisfied, but A not. Thus, the logical conse-
quence relation is empty. This view plausibly deserves the title ‘Logical
Nihilism’6 and in fact has been defended under that name by Estrada-
González (2011) and Russell (2017). So, I think it is worth distinguish-
ing two types of logical nihilism:
Logical Nihilism 1 There’s no logical consequence relation that correctly

represents natural language inference; formal logics are inadequate to
capture informal inference.
Logical Nihilism 2 There are no logical constraints on natural language

inference; there are always counterexamples to any purportedly valid
forms.
On the second view, there is a consequence relation (namely the empty

one) which gets natural language inference ‘right’. The focus of this chap-
ter is on logical nihilism in the first sense.
In what follows, I present and defend a number of arguments in favor
of logical nihilism. The arguments are grouped into two main families:
arguments from diversity (§2) and arguments from expressive limitations
(§3). The arguments are often simple syllogisms, pointing to fundamen-
tal differences between natural languages and formal consequence rela-
tions. Many of the arguments involve familiar problems in the philosophy
of logic. The arguments, taken individually, are interesting in their own
right; they each highlight an important way in which the formal methods
of logic can be seen to be inadequate to modeling natural language infer-
ence. But the arguments taken jointly are more significant; by presenting
all the arguments together, we can build something of a cumulative case
for logical nihilism. Of course, if any of these arguments are sound then
logical nihilism is correct. But the arguments reinforce one another, such
304 A. J. Cotnoir
that logical nihilism presents us with a unified view across a broad range
of issues in philosophy of logic. I conclude (§4) by considering related
philosophical issues and sketching a general outlook on logic and formal
methods that is nihilist-friendly.
Before presenting these arguments, however, let me respond to an
immediate worry. In presenting and endorsing arguments for a view, one
ordinarily takes them as good arguments, where good arguments are (at
the very least) valid arguments. But if there is no correct theory of logical
consequence, then presumably there are no valid arguments either. Thus,
one might suspect that any attempt to argue for logical nihilism under-
mines itself.
This worry could be serious if logical nihilism entailed that there were
no valid arguments (as in Logical Nihilism 2). But notice that logical
nihilism (of both sorts) is consistent with their being standards governing
good inference in natural language. And Logical Nihilism 1 merely claims
that there is no formal theory that perfectly captures these standards. And
that’s perfectly compatible with these arguments being formally valid in
some regimented language that is adequate for more restricted purposes.
It’s also compatible with arguments conforming to inherently informal
standards on good inferential practice, or an abductive methodology in
the philosophy of logic (e.g. Williamson 2017). So the worry is
ill-founded.
2 Arguments from Diversity
The first batch of arguments for logical nihilism follows a simple recipe.
The first ingredient is an argument for logical pluralism; this establishes
that no single logic can be an adequate theory of inference. The second
ingredient is a constraint on logical consequence that rails against logical
pluralism—that is, a constraint that requires there be at most one single
adequate theory of inference. Combine these two ingredients and stir;
the result is logical nihilism. For if there cannot be one single correct
theory, and any correct theory must not be plural, then there cannot be
any correct theory at all. Fortunately for us, such ingredients are not hard
to come by.
There are a number of types of logical pluralism that have been

defended in the literature; I won’t consider all of them. I will focus on
Beall and Restall’s (2000, 2006) case-based pluralism, and Lynch’s (2009)
domain-based pluralism. I’ll also briefly consider Varzi’s (2002) logical
relativism and Russell’s (2008) truth-bearer dependent version of logical
pluralism.7
An Argument from Necessity
Beall and Restall (2006) suggest that the notion of logical consequence
may be analyzed by the Generalized Tarski Thesis (GTT).
GTT An argument is validx if and only if in every casex in which the

premises are true, the conclusion is true.
Of course, we have yet to specify what sorts of cases are under consid-
eration here. Logicians are frequently concerned with models, but the
notion of consequence itself doesn’t determine that models are the only
possible option. Once this is granted, there is a straightforward argument
to pluralism.
The idea is that settled core of logical consequence (‘the intuitive or
pre-theoretic notion’ (Beall and Restall 2000)) is given by GTT. An
instance of GTT is obtained by a specification of casesx in GTT, and a
specification of the relation of being true in a case. An instance of GTT is
admissible if it satisfies the settled role of consequence, and if its judg-
ments about consequence are necessary, normative, and formal (in some
sense or other). A logic, then, is an admissible instance of GTT.
Beall and Restall contend that there are at least three admissible
instances of GTT. They defend this claim by appeal to the diverse pur-
poses logic may be put to, each of which corresponds to a different kind
of case. There are the complete and consistent cases of classical logic (i.e.
worlds), the incomplete cases of intuitionistic logic (i.e. constructions), and
the inconsistent (and/or incomplete) cases of relevant logic (i.e. situations).
As a result, there is no single correct consequence relation, but many cor-
rect consequence relations relative to which kind of cases we intend.
306 A. J. Cotnoir
Suppose we agree with Beall and Restall that there are a number of
different ways of specifying the notion of a case, each with its own speci-
fication of truth-in-a-case, and hence a number of instances of GTT. We
might still disagree that any one of these instances is admissible because
the resulting consequence relations fail to be necessary. Clearly, it’s a plau-
sible constraint on logic that it must be necessary; an argument is valid
whenever it is necessarily truth preserving. But on the surface, the neces-
sity constraint appears to require us to look at all kinds of cases, if they
really are genuine cases.
This is, in effect, a version of an objection to logical pluralism devel-
oped by Bueno and Shalkowski (2009).
Thus, on Beall and Restall’s account, none of the major families of logics
they consider satisfy the necessity constraint. By their own standard, none
of these are logics at all. Only a very weak consequence relation survives
this scrutiny, according to their accounting of the necessity constraint as
quantification over all cases. Once the partisan spirit of logical monism is
replaced with the open-minded embrace of cases suitable to alternative log-
ics, no commonly promulgated consequence relation seems to satisfy the
necessity constraint. Hence, according to their own accounting of the con-
straints on relations of logical consequence, there are no such relations.
(p. 299f )
It is worth noting that there are actually two sorts of objections here, each
resulting in a different form of logical nihilism. The first objection is that,
on an absolutely general reading of necessity—necessity in the broadest
possible sense—as quantifying over all cases of any type, no instance of
GTT counts as admissible. And hence there are no logics; this is a form
of Logical Nihilism 1.
On the other hand, we might instead specify a general notion of case
as follows: c is a case* iff c is a casex for some x. Similarly, we can give a
general specification of truth-in-a-case as follows: A is true-in-a-case* iff c
is a casex for some x and c represents A to be true-in-a-casex. This seems to
suffice for an instance of GTT. And indeed, this will constitute an admis-
sible instance especially with regard to necessity (read again in the broad-
est sense). But, if we take seriously the vast range of logics which have
been put forward by various logicians, then given corresponding notions

of casesx and truth-in-a-casex, it is doubtful that any substantive logical
principles will survive.8 This would lead to a situation close to Logical
Nihilism 2. But even if we simply stick with the cases Beall and Restall
have already admitted, it seems clear that the logic resulting from quan-
tifying over all such cases is still going to be far too weak to adequately
account for natural language inference.
I do not mean to suggest that Beall and Restall have nothing to say in
response to this style of argument—indeed, they do (2006). Whether
their response is ultimately successful I leave to the reader to decide. But
I merely want to show that arguments for case-based pluralism might
naturally be converted to an argument from diversity that attempts to
establish logical nihilism.
An Argument from Topic-Neutrality
A similar sort of strategy equally applies to domain-based pluralists. Some

pluralists about truth have suggested that there are viable forms of logical
pluralism that sit naturally with their views. Pluralism about truth is the
view that different domains of inquiry (e.g. science, mathematics, ethics,
history, etc.) have different truth properties associated with them; so
being true in one domain might consist in corresponding to a fact, whereas
being true in a different domain might correspond to a kind of general-
ized coherence. Lynch (2008, 2009) has recently argued that, because
truth properties in domains of inquiry are so vastly different, which infer-
ences are valid (i.e. which inferences preserve these properties) can vary
from domain to domain.9 This is a domain-based logical pluralism.
However, many have contended that logical consequence must be
topic-neutral. Indeed, we might agree with Sher (1996) that ‘formality
and necessity play the role of adequacy conditions: an adequate defini-
tion of logical consequence yields only consequences that are necessary
and formal’ (p. 654). That is, what arguments are valid can depend only
on their form, and cannot depend on what the arguments are about. How
to cash out this topic-neutrality constraint is somewhat controversial (see
MacFarlane 2002 for various proposals), but it seems clear that any
308 A. J. Cotnoir
domain-based logical pluralism will not meet such a constraint. Of

course, one might accept that within a domain logical consequence does
not depend on any further facts about the content of those propositions.
But there is still the residual fact that which argument forms are valid and
invalid will depend on the domain of inquiry one is in. The truth plural-
ist argues that any adequate theory of consequence must be domain-
relative. By topic-neutrality, there can be no adequate theory of
consequence. As a result, one can convert an argument for domain-based
pluralism into an argument from diversity for logical nihilism.
Other Arguments from Diversity
Another strategy along the same lines applies to Varzi’s (2002) brand of
logical pluralism. According to logical relativism, the correct inferences
depend crucially on what one takes to be logical—as opposed to the non-
logical—vocabulary. Varzi advocates for a kind of skepticism about any
objective criterion on where to draw the logical/non-logical divide.10 Of
course, one might be perfectly happy to accept that there are natural lan-
guage consequence relations for each way of drawing the divide. Indeed,
one might rather appeal to instead to ‘meaning postulates’ or ‘semantic
constraints’ as (cf. Sagi 2014) for all vocabulary to generate semantic
entailment relations. Formal semanticists often deliver such theories.
But there are general reasons why formal semantics does not deliver
the logic of natural language. Glanzberg (2015, §2.2) argues that such
meaning postulates or semantic constraints will, at best, deliver analytic
entailments which should not intuitively count as valid (e.g. ‘John cut the
bread’ lexically entails ‘The bread was cut with an instrument’ because of
constraints on the meaning of ‘cut’). This seems to dictate against count-
ing such semantic ‘entailment’ relations to be logical consequence proper.
Glanzberg (2015, §2.1) also argues that natural language semantics,
whether in its neo-Davidsonian or type-theoretic guises, are primarily
concerned with absolute semantics. That is, they give semantic clauses for
sentences expressing the meanings that speakers actually understand
those sentences to have. They do not give relative semantics, that is,
semantic clauses that are generally specified over a range of models, which
is precisely what is required in order to specify logical consequence.11 The

end result is that, if logic must be formal, then there is no logical conse-
quence relation that is adequate to natural language.
At the risk of belaboring the point, allow me to give a final argument
from diversity. Russell (2008) argues that different views on the relata of
the consequence relation (e.g. sentences, propositions, etc.) yield differ-
ent consequence relations. If one takes sentences to be the relata of con-
sequence, then contextual aspects of language can validate inferences
which are intuitively invalid (e.g. ‘I am here now’ is a logical truth in
Kaplan’s LD).12 Alternatively, if one takes propositions to be the relata of
consequence, then propositions can validate inferences which are intui-
tively invalid as well (e.g. ‘Hesperus is Hesperus’ entails ‘Hesperus is
Phosphorous’). Russell draws logical pluralism as her conclusion, but I’d
contend that the arguments better support logical nihilism: both of these
consequence relations deliver intuitively invalid validities. Of course, a
nihilist need not reject that such semantic or metaphysical necessitation
relations exist; it is just that they do not appear adequate to the role of
logic as traditionally conceived.
The various available arguments from diversity are arguments on-the-
cheap, as it were. They piggyback on arguments for logical pluralism,
together with monist constraints about the concept of logical conse-
quence. Monists may not be sympathetic to the underlying pluralistic
motivations, and pluralists may not be sympathetic to the monistic read-
ing of the constraints on logical consequence. Though I wouldn’t expect
monists or pluralists to agree, I think it is striking that logical nihilism
can make sense of the motivations behind both lines of thought.
3 Arguments from Expressive Limitations

The second class of arguments follows a different sort of recipe. It starts
from a given phenomenon of natural language inferential practice, and
argues that no formal language can exhibit that phenomena. Many argu-
ments of this sort have been given in the past and recent literature. But
again, I think it is worth putting them all in one place to display their
cumulative weight. Every argument from expressive limitations needs to
310 A. J. Cotnoir
be refuted in order to defend against logical nihilism. And I hope to dis-

play just how difficult this task is.
The Argument from Semantic Closure
The first argument from expressive limitations for logical nihilism appeals
to considerations around semantic closure.
1 . Natural languages are semantically closed.

2. No formal language is semantically closed.
3. So, no formal language is adequate to natural language.
The argument appeals to an alleged fundamental difference between

formal languages (from which logical consequence relations are derived)
and natural language. The idea is that natural languages are capable of
representing their own semantics; we could express all the truths about
English in English. But formal languages cannot generally be supple-
mented with semantic notions that apply to themselves on pain of
triviality.
Consider premise 1. Many find it straightforwardly obvious that natu-
ral languages like English are semantically closed in the sense that we can
use natural languages like English to talk about the semantic properties
and relations of sentences of English. Scharp (2013) argues that it is a
fundamental presupposition of formal semantics that such a thing is pos-
sible. In attempting to give a complete theory of natural languages as a
whole, couched in a given natural language, we must presuppose that
that language can serve as a metalanguage for itself. Take the semantic
property ‘is true’ as an example. Eklund (2002) has argued that in order
to be competent with the use of the English truth predicate, one must be
willing to apply it to all sentences of the English language.
As for premise 2, there are a number of different ways of supporting
this claim. Tarski showed clearly that the truth predicate could not be
expressed in any formal language (of certain expressive richness) if the
consequence relation was to be classical due to semantically ‘rogue’ sen-
tences like the Liar sentence. This means that, at the very least, premise 2
is true with respect to classical formal languages.13 This result has led
many to reject that classical logic is the One True Logic.
But the well-known phenomenon of revenge plagues non-classical for-
mal languages as well. The literature is predictably populated with articles
that find inexpressible concepts for non-classical logics that purport to be
semantically closed. Even absent a general recipe for finding a revenge
paradox for every formal language, the sheer volume of failed attempts at
semantic closure would justify a plausible pessimistic meta-induction14: if
premise 2 has proven true for every formal language so far, then there’s
good reason to think the issue is systemic.
There are somewhat more general recipes for finding revenge problems
even for non-classical approaches to paradox. Here is a basic one: we begin
with three desiderata for any semantically closed formal language 𝕷.
Characterization There is a property φ s.t. all and only ‘rogue’ sentences

of 𝕷 are φ.
Semantic Closure φ is expressible in 𝕷; there is a name ⟨α⟩ for every

sentence α in 𝕷.
Revenge Immunity No sentence attributing φ to a ‘rogue’ sentence is

itself φ.
Let Γ be the class of ‘rogue’ sentences. Now, [by SEMANTIC

CLOSURE] construct the following sentence λ ∶ φ⟨λ⟩—the sentence
which says of itself that it is a rogue sentence.
Now, it will typically be the case that λ is a paradigm case of a rogue
sentence; λ ∈ Γ. So, [by characterization], φ⟨λ⟩ is true. But [by REVENGE
IMMUNITY], φ⟨λ⟩ ∉ Γ. Contradiction. One might object that this is
merely a proof that, contrary to appearances, λ is not rogue. But then
consider the sentence γ ∶ φ⟨γ⟩ ∨ ¬T⟨γ⟩, where T is our truth predicate.
If γ is not rogue, then it must fall under some other semantic category
(i.e. true or false). But in either case, we get the standard Liar reasoning
yielding a contradiction. So γ must be rogue. But since γ is an attribution
of φ to a rogue sentence, we have a violation of REVENGE
IMMUNITY. This is perhaps especially problematic since, if γ is rogue,
312 A. J. Cotnoir
that’s precisely what one disjunct of γ says, and thus it should intuitively
be true. The characterization of paradoxical sentences seems destined to
break down.
There are, of course, a number of options for replying to a revenge
charge like this. The first option is to reject that Characterization is a
desiderata at all. Of course, the characterization constraint has two direc-
tions. One might reject the ‘all’ direction, and attempt to leave some
paradoxical sentences uncharacterized. An immediate reply would be
that one has thereby failed to give a complete account of the semantic
paradoxes. On the other hand, one might reject the ‘only’ direction, and
claim that we must throw out some babies (non-rogues) when throwing
out the bathwater (rogues). Two immediate replies come to mind: first,
this will result in a situation where being a ‘rogue’ sentence is not express-
ible in 𝕷 (even if φ is); after all φ will apply to some non-rogues. Second,
being a ‘rogue’ sentence appears not to be doing any work in the resulting
theory, the characterization of paradoxes happens using φ.
One might opt for second response to revenge: reject Revenge
Immunity—that is, accept that some attributions of φ are themselves
‘rogue’. An immediate reply to this approach, however, is that since the
theory itself includes attributions of φ to the paradoxical sentences, one’s
theory relies on exactly the problematic phenomenon. That is, one is
forced to use rogue sentences in giving one’s theory.
A third response to these revenge charges would be simply to reject
Semantic Closure. This is Tarski’s solution (for formal languages) in that
he rejects any such φ is expressible. But again linguistic appearances sug-
gest the contrary for natural languages, and so this will lead us to a ver-
sion of logical nihilism.
Perhaps one might attempt to agree with Priest (1984) that this merely
shows semantic closure can be had only on pain of inconsistency; an implicit
argument for dialetheism. But Beall (2015) has outlined a general argu-
ment barring semantic closure for formal theories on pain of triviality. The
rough idea is to classify ‘rogue’ sentences of a formal semantic theory
(couched in a formal language together with its consequence relation) as
trivializer-sentences: sentences that, relative to that language’s consequence
relation, yield the theory containing all sentences of the language. Then
there can be no trivializer predicate for that theory characterizing all and
only the trivializing sentences of the language. That’s because a sentence

which says of itself that it is a trivializer for that very theory will be unchar-
acterized by that predicate. The upshot is that any formal language is either
semantically incomplete or absolutely inconsistent (i.e. trivial).
Another related issue for semantic closure is that there are very general
reasons for thinking that the validity predicate for a logical consequence
relation of a formal language will not expressible in the language itself.
One reason for thinking this has been put forward by Field (2008).
Gödel’s second incompleteness theorem says that no sufficiently strong
formal language can prove its own consistency. If we had semantic clo-
sure (including an adequate truth and validity predicates), we would be
able to do just such a thing. After all, if we could show in a formal lan-
guage that the validity relation was truth preserving, we would be able to
assert all the axioms of a logical theory to be true, and then be confident
that by closing the theory under logical consequence, we did not intro-
duce any inconsistencies.
Beall and Murzi (2013) have also argued that no validity predicate will
be expressible in any formal language (satisfying certain structural rules
like contraction and transitivity) that contains its own truth predicate.15
This is due to the fact that one can use such a validity predicate to con-
struct variants of Curry’s paradox. Now, it may be that by rejecting c ertain
structural rules might allow one to express a validity predicate.16 But even
still, the resulting (non-contractive, or non-transitive) logical conse-
quence relation will be highly non-standard, and one wonders whether
such a logic could be adequate to natural language inference. The under-
lying worry is that by attempting to accommodate semantic closure in a
formal language, we are required to weaken the logical consequence rela-
tion for that language to such a degree that it no longer resembles our
patterns of natural language inference.
It is no accident that the literature regarding semantic closure contin-
ues to grow, for these issues are directly tied up with the adequacy of
formal methods for giving an adequate treatment of natural language
inference. It is worth saying that quite a lot of brilliant technical work has
gone into these problems, and that the debate is far from settled. But
logical nihilism is not outside the bounds of reason even considering the
best work on these problems.
314 A. J. Cotnoir
The Argument from Quantification
The second argument from expressive limitations trades on another limi-

tation of formal languages to do with absolute generality.
1 . Natural languages have absolutely unrestricted quantifiers.

2. No formal language has absolutely unrestricted quantifiers.
3. No formal language is adequate to natural language.
Each premise is, of course, controversial so let’s look at the case for
each.17
What evidence is there for premise 1? Certain uses of natural language
quantification appear to require an unrestricted reading. For example,
part of the main function of philosophical uses of quantification is to rule
out the existence of certain objects: ‘The universal set does not exist’
should not be true if the universal set does exist but we just can’t quantify
over it. Even the rejection of unrestricted quantification seems to presup-
pose the availability of unrestricted quantification; saying ‘one cannot
quantify over absolutely everything’ appears to presuppose that there is
something that one cannot quantify over.18 Additionally, McGee (2003)
argues that natural language universal quantification would be unlearn-
able if it weren’t unrestricted.
There are also general reasons for accepting premise 2. Here is an argu-
ment that has been put forward repeatedly in the literature, based on two
key claims.
All-in-One Formal quantification involves quantification over sets (or at

least domains that are set-like).
No Universe There is no universal set (or set-like domain).
What reason is there for thinking that formal quantification involves

quantification over sets? Grim (1991) and Priest (2002) have both pro-
vided arguments along these lines. In fact, Priest (2002) has argued for a
stronger claim, what he calls the ‘Domain Principle’, according to which
all quantification (whether formal or not) involves sets.19 Far and away,
the best formal theory of natural language quantifiers is Generalized

Quantifier Theory; and generalized quantifiers are relations between sets.
However, whether this is required for any adequate theory of natural
language quantification is controversial.
There are a few available lines of response. One might reject All-In-
One, and suggest that plural quantifiers provide a counterexample to this
claim (Boolos 1985).20 It is, of course, relevant that plural quantification
is isomorphic to second-order quantification over non-empty monadic
predicates. And second-order quantification (at least on its standard
semantics) intrinsically involves quantification over sets. Many have
argued that plural quantifiers are no more innocent than higher-order
quantifiers, or are at least ‘set-like’ in the relevant sense.21
To be sure these issues are complicated, and debate is ongoing. But
there are many who simply do not see a way to preserve unrestricted
quantification, and opt rather for an indefinitely extensible hierarchy of
quantifiers.22 Perhaps one might suggest a proof- theoretic approach to
quantification without domains. Of course, there is the worry about the
failure of any proof-theory to be complete with respect to higher-order
models. A potential suggestion would be to give rules (schematic or proof-
theoretic) for quantifiers that are open-ended, in the sense they govern the
behavior of all possible extensions of the language (e.g. higher-order
quantifiers) all at once.23 But even on this proposal, it will still be true
that no single formal language can achieve quantification over absolutely
everything. At best, the view delivers a single concept of quantification
which will be retained no matter how one’s formal language is expanded
to ever more inclusive domains.
Instead of rejecting All-in-One, one might instead reject No Universe
and consider formal theories with set-like universal objects. In ordinary
set theory, we reject a universal set because it would have to be a member
of itself. And for Russell-paradox related reasons, no set can be a member
of itself. Class theory can have a universe of sets, but not a universe of
classes since proper classes are not members of anything including other
classes. Similarly, mereology has a universal object. Could one use mereo-
logical models for quantification? Uzquiano (2006) shows that this
method (supplemented with some natural assumptions) can only recover
quantification over sets up to certain types of limit cardinals; but if there
316 A. J. Cotnoir
are strongly inaccessible cardinals, then this quantification is not truly

unrestricted.
Another option would be to appeal to alternative set theories.
Paraconsistent set theories can non-trivially contain universal sets guar-
anteed to exist by Naïve Comprehension (Brady 2006; Weber 2010,
2012). Such sets will be inconsistent, but in a paraconsistent setting we
can have inconsistent sets (including the Russell set) non-trivially. So, in
this case it appears we can have unrestricted quantification (Priest 2007).
But even if this proposal looks promising, it comes with a similar cost—
we lose a plausible form of restricted quantification.
Any adequate theory of restricted quantifiers requires the following:
Modus Ponens All As are Bs, x is an A; hence x is a B.
Weakening Everything’s a B; hence all As are Bs.
Contraposition All As are Bs; hence all non-Bs are non-As.
That is, any theory that rejects the above principles will be inadequate
to natural language inference. But any theory that accepts them reintro-
duces a form of Ex Falso Quodlibet.
B
A→B
¬B ¬B → ¬A
¬A
Here, on the assumption that B and ¬B, we can infer ¬A. But this is
something that simply cannot be valid in a dialetheic setting, since it
would entail the truth of (the negation of ) every sentence
whatsoever.24
There are set theories based on classical logic with a universal set:
Quine’s (1937) New Foundations, and descendants (see Forster 1992).
But Linnebo (2006, p. 156f ) notes, they are technically unappealing
and lack a unified intuitive conception. More importantly, they have a
niversal set only by placing further restrictions on Comprehension, a

u
limitation which Williamson (2003, p. 425f ) argues would also make
any account of ordinary restricted quantification impossible.25
Other Arguments from Expressive Limitations
Another key argument from expressive limitations appeals to an obvious

phenomenon of natural language: vagueness. The philosophical literature
on vagueness is enormous, and the number of theoretical options are too
numerous to outline here. I want to highlight some views of vagueness
that are broadly in line with the nihilist outlook on logic. One key player
is, of course, Dummett who argued that the use of vague predicates in
natural language is intrinsically inconsistent.
What is in error is not the principles of reasoning involved, nor, as on our

earlier diagnosis, the induction step. The induction step is correct, accord-
ing to the rules of use governing vague predicates such as ‘small’: but these
rules are themselves inconsistent, and hence the paradox. Our earlier model
for the logic of vague expressions thus becomes useless: there can be no
coherent such logic. (Dummett 1975, p. 319f )
If there can be no coherent logic of vague expressions, but natural lan-

guage contains vague expressions, then we have a straightforward argu-
ment to logical nihilism. More recently, Eklund (2002) argued that
languages with vague expressions are inconsistent languages. Ludwig and
Ray (2002) suggest that no sentence involving a vague term can be true.
Aside from these strong claims about the inconsistency or incoherence
of natural language, however, one might want to contend there is still a
mismatch between formal logics and natural language with how they
treat vagueness. It’s not outlandish to suggest no formal language, with its
mathematical precision, could ever hope to perfectly represent natural
language vagueness. Or one might agree with Sainsbury (1996) and Tye
(1994) that any semantics for vague expressions would have to appeal to
a similarly vague metalanguage.26
Hiding in the background to most formal approaches to vagueness is
the ubiquitous revenge problem of higher-order vagueness. It may be that
318 A. J. Cotnoir
such problems inevitably render precise formal methods inadequate for

the vague aspects of natural language inferences.27
There are probably other arguments from expressive limitations that
can be made; I have simply pointed to three of the more difficult prob-
lems in philosophical logic: semantic closure, unrestricted quantification,
and vagueness. And as emphasized, these problems are not new, and the
debate over the correct treatment of such phenomena ongoing. Still the
logical nihilist permits a unified perspective: the expressive resources of
natural language and our best formal languages might not perfectly coin-
cide, and that might be because there are important and deep differences
between natural language inference and logical consequence.
4 Related Issues
An issue I’ve largely set aside is the fact that logic is taken to be normative.
Perhaps this unreasonably privileges psycho-semantic facts about how we
in fact infer, rather than how we ought to infer. On this conception,
appeals to linguistic data and evidence for ‘semantic intuitions’ can appear
irrelevant. The question of logic is: ‘what is best inferential practice?’
But I want to stress that even if logical consequence is conceived as a
fundamentally normative relation, there might well be arguments for
logical nihilism. For example, we might contend that an objective norma-
tive relation would be metaphysically and epistemically strange. Roughly,
we might apply Mackie’s (1977) arguments to logical consequence itself.
This case is made strongly by Field (2009):
Quite independent of logic, I think there are strong reasons for a kind of
antirealism about epistemic normativity: basically, the same reasons that
motivate antirealism about moral normativity, or about aesthetic good-
ness, extend to the epistemic case. (For instance, (i) the usual metaphysi-
cal (Humean) worry, that there seems no room for ‘straightforward
normative facts’ on a naturalistic world-view; (ii) the associated epistemo-
logical worry that access to such facts is impossible; (iii) the worry that such
normative facts are not only nonnaturalistic, but ‘queer’ in the sense that
awareness of them is supposed to somehow motivate one to reason in a
certain way all by itself.) (p. 354)
And if we have good reason to reject entities which are metaphysically or

epistemically or motivationally strange, we would have good reason to
reject logical consequence (qua objective and normative).28
If logical consequence is a subjective normative relation, then whether
that relation holds would be relative either to our aims or our conven-
tions. This view is called ‘logical pragmatism’ by Haack (1978). Here, the
aim of formal logic is not so much correct representation, but rather util-
ity. But this view is not opposed to logical nihilism: it may well be that
the best inferential practice for natural language users (i.e. the most use-
ful) has no formal analogue. After all, semantic closure, unrestricted
quantification, and vagueness are useful features of natural language. But
here again what counts as ‘best’ in the pragmatic sense will depend in
large part on what natural language is used for.29
In defending logical nihilism, I do not intend to impugn any honor-
ific status of natural language inference. This is over against so-called
Inconsistency Theories (with which logical nihilism has much in com-
mon) according to which natural language is inconsistent because of, for
example, the truth predicate or vague expressions.30 Logical nihilists are
free to disagree with the claim that natural languages are inconsistent,
perhaps because they agree with Tarski (1944) that ‘the problem of con-
sistency has no exact meaning with respect to this [natural] language’
and only arises with the sufficiently precise formulation of a formal
language. Some inconsistency theorists claim that natural language is
incoherent or meaningless because it has no precise logic.31 Here again,
logical nihilists are free to disagree. We might well agree with Wright
(1975) that natural language inference can be unprincipled without
being incoherent.
Nor am I impugning formal methods in the study of natural language,
in philosophy, or elsewhere. As Glanzberg (2015) rightly notes, the appli-
cation of formal methods has been one of the greatest successes in the
study of natural languages of the last half-decade. The application of for-
mal techniques to, for example, epistemology, metaphysics, perhaps vir-
tually every area of philosophy has likewise yielded significant progress.
But we should be aware that formal languages serve a representational
function, and—like modeling tools used in any science—their applica-
tions have limits.
320 A. J. Cotnoir
This is something the early analytic pioneers of logic, Frege and Tarski,
understood well. Consider Frege:
I believe I can make the relationship of my Begriffsschrift to ordinary lan-

guage clearest if I compare it to that of the microscope to the eye. The lat-
ter, due to the range of its applicability, due to the exibility with which it is
able to adapt to the most diverse circumstances, has a great superiority over
the microscope. Considered as an optical instrument, it admittedly reveals
many imperfections, which usually remain unnoticed only because of its
intimate connection with mental life. But as soon as scientific purposes
place great demands on sharpness of resolution, the eye turns out to be
inadequate. The microscope, on the other hand, is perfectly suited for just
such purposes, but precisely because of this is useless for all others. (Frege
2002, §V)
Formal languages require a significant degree of abstraction and idealiza-

tion from natural language inference.32 Only certain aspects of the model
are intended to represent the phenomena being modeled. This perspec-
tive might even help to handle some of revenge problems involving
semantic closure, absolute generality, vagueness, and so on. Too-easy
revenge appeals to artifacts of the model and argues they are inexpressible
in the model. But once we see the clear separation, we can see that there’s
no obligation for the theorist to take these burdens on.33 The view of
logic-as-modeling is sometimes billed as a form of logical pluralism (e.g.
Cook 2010). And while it is clear my sympathies lie with this view, a
rather more drastic opinion of the limitations of formal methods inclines
me to say that no formal language gets it right, rather than to say that
many formal languages get it right more-or-less.
The nihilist accepts that there are no correct or completely general for-
mal theories. But that needn’t mean there aren’t a lot of reasonably good,
useful, and explanatory ones. In certain areas of inquiry, some of the
more useful features of natural language are unnecessary and may even be
detrimental to the theoretical purposes at hand. Science and mathematics
(both pure and applied), and even many areas of philosophy, require the
kind of precision inherently informal languages cannot provide. Formal
logic has its best application in areas where we can regiment our language
and revise our practice. But we shouldn’t forget that not all of our inquiry
is like this: sometimes regimentation leaves us with an expressive loss, and

sometimes revising our practice is not possible or even desirable. Logical
nihilism reminds us to respect the differences between model and
reality.34
Notes
1. Of course, there are many ways one might think about truth; one needn’t
think of truth in terms of correct representation. But it is a reasonable
way of thinking about the issue, and I think it is what the intuitive slo-
gan ‘One True Logic’ is getting at.
2. This quote strongly suggests that Beall and Restall would not disagree
with the characterization of logics as being in the business of correct
representation of natural language inferential practice.
3. Compare recent authors like Bueno and Colyvan (2004), “The aim of
logic is taken to be to provide an account of logical consequence that
captures the intuitive notion of consequence found in natural language”
(p. 168). Or Resnik (2004), “As practitioners of inference we make spe-
cific inferences […] As logicians we try to formulate a systematic account
of this practice by producing various rules of inference and laws of logic
by which we presume the practice to proceed. This aspect of our work as
logicians is like the work of grammarians” (p. 179). Or consider Cook
(2010) “[A] logic is ‘correct’, or ‘acceptable’, etc., if and only if it is a cor-
rect (or acceptable, etc.) codification of logical consequence. The idea
that the philosophically primary (but obviously not only) goal of logical
theorizing is to provide a formal codification of logical consequence in
natural language traces back (at least) to the work of Alfred Tarski”
(p. 195).
4. Cook (2010, p. 495f ) gives a detailed account of what it means to say a
logical consequence relation is adequate to natural language inference. I
am assuming something like his definition is suitable for this purpose.
5. Another way of reading Mortensen is as arguing that real broad possibil-
ity outstrips pure logical possibility. In this case, then, there may be logi-
cally impossible scenarios that are not, broadly speaking, impossible.
Mortensen would then not count as a logical nihilist in the sense above.
6. Parallel disputes over the metaphysics of composition. Here universalism
states that composition always occurs, whereas nihilism claims that the
322 A. J. Cotnoir
composition relation is basically empty. ‘Emptyism’ just doesn’t have the

same ring to it. And unfortunately, ‘Noneism’—the most natural name
for the view defended in this chapter—is already taken.
7. There is also Eklund’s (2012) Carnapian language-relative approach;
Shapiro’s (2011, 2014) contextual approach, and Cook’s (2010) logic-as-
modeling approach. Later I’ll touch on some issues that directly relate to
their motivations.
8. Similar arguments have been made by Read (2006, p. 208f ) and Priest
(2006, Chap. 12).
9. In Cotnoir (2013), I outline general approach to validity motivated by
this idea. See also Pedersen (2014) for more motivations.
10. This was Tarski’s (1936) early view.
But I also consider it to be quite possible that investigations will bring
no positive results in this direction, so that we shall be compelled to
regard such concepts as logical consequence as relative concepts. The
fluctuation in the common usage of the concept of consequence
would in part at least be quite naturally reflected in such a compul-
sory situation [of a relatively-defined concept of consequence].
(p. 420)
See also Etchemendy (1990) and Dutilh-Novaes (2012).
11. Glanzberg’s (2015) rich and carefully argued paper is concerned with
rejecting the view that natural language (a structure with a syntax and a
semantics) determines a logical consequence relation. His position is
very similar to logical nihilism of the sort I’m defending here, and he is
probably one of the view’s closest allies. But strictly speaking Glanzberg’s
view is compatible with logical monism and logical pluralism, since
there could be one (or more) correct theory of natural language infer-
ence, even if it isn’t possible to simply read such a thing off from natural
language itself.
12. Compare Zardini (2010, 2014) who argues that mid-argument context
shifts invalidate standard inferences (e.g. “I am sitting” can fail to entail
itself ).
13. See also Bacon (2015) who argues that there can be no ‘linguistic’ theo-
ries of paradox based in a classical language due to revenge problems.
14. Thanks to Cory Wright for suggesting this way of framing the issue.
15. Related arguments first appeared in Whiǣle (2004), and Shapiro (2011).
See also Murzi (2014). For dissenting voices see Cook (2014), Wansing
and Priest (2015).
16. See e.g. Zardini (2013), Weber (2014).

17. For good introductory discussions of these issues see Florio (2014) and
Rayo and Uzquiano (2006, p. 1). The following discussion is indebted to
them in various ways.
18. See Lewis (1991) and Williamson (2003).
19. Priest himself rejects ‘No Universe’ for his preferred set theory; for dis-
cussion see below.
20. See Uzquiano (2009), Rayo (2002).
21. For example, Jané (1993), Resnik (1988), and more recently Linnebo
(2003).
22. For example, one might utilize hyperplural quantification (Rayo 2006),
which under certain assumptions, isomorphic to the type hierarchy
(Linnebo and Rayo 2012). There are also modal approaches (Linnebo
2010 and Studd 2013) which allow set-theoretic domains to be indefi-
nitely extensible.
23. For some defenders of this view, see Lavine (2006), McGee (2006), and
Williamson (2006).
24. This problem is discussed, and some possible lines of response explored
in Beall et al. (2006).
25. See also Weir (2006).
26. See Cook (2002) for discussion.
27. Wright (2010) contends that higher-order vagueness worries are pseudo-
problems; it is a revenge problem that only arises for views which misun-
derstand first-order vagueness. The logical nihilist can afford sympathy
to such claims.
28. Field takes this to be an argument for pluralism (because of an under-
lying antirealist pluralism about epistemic norms), one might well
think such considerations provide better reasons to be nihilist about
logic.
29. The comparison with morality is instructive: consider the relevance of
anti-theory views in ethics (e.g. Clarke 1987) to logical nihilism, or even
the similarities between particularism (e.g. Dancy 1983) and Hofweber’s
(2007) view that we should give up the ideal of deductive inference as
exceptionless and monotonic. Natural language inferences need not be
exceptionless or monotonic, and often are not. But they might still be
generically valid, in the sense that generics like “Humans are bipeds” are
true.
30. E.g. Chihara (1979), Eklund (2002), and Ludwig (2002).
324 A. J. Cotnoir
31. E.g. Scharp (2013) who thinks defective concepts like ‘truth’ need to be
replaced, or Patterson (2009) who argues that we understand natural
language using a false semantic theory, such that strictly speaking natural
language sentences have no meanings.
32. See especially Glanzberg (2015, §IV), but also Cook (2002, 2010),
Shapiro (2006), and Scharp’s (2013) metrological naturalism.
33. This point is made clearly and forcefully in Beall (2007) with respect to
the semantic paradoxes. Cook (2002) argues for a similar perspective
with respect to vagueness.
34. I’d like to thank audiences at the Truth Pluralism and Logical Pluralism
Conference at the University of Connecticut, the Swiss Society for Logic
and Philosophy of Science at the University of Neuchâtel, the Northern
Institute of Philosophy 2011 Reading Party, the University of St Andrews
Philosophy Society, and the students in my 2014 and 2017 Philosophy
of Logic seminars. Their comments and questions led to many improve-
ments and developments in the paper. Special thanks to Colin Caret,
Roy Cook, Matti Eklund, Ole Hjortland, Michael Lynch, Julien Murzi,
Stephen Read, Gillian Russell, Gil Sagi, Kevin Scharp, Stewart Shapiro,
Keith Simmons, Crispin Wright, and Elia Zardini for discussions on
these topics over a number of years. Thanks also to an anonymous referee
for helpful comments on a previous version of the paper. The biggest
debt of gratitude is owed to my PhD supervisor Jc Beall, who disagrees
with many of the ideas in this paper. A reaction against one’s academic
upbringing can be a sign of deep respect; and I hope this paper is taken
in that spirit.
References
Bacon, A. 2015. Can the Classical Logician Avoid the Revenge Paradoxes?
Philosophical Review 124 (3): 299–352.
Beall, J. 2007. Prolegomena to Future Revenge. In Revenge of the Liar, ed.
J. Beall, 1–30. Oxford: Oxford University Press.
———. 2015. Trivializing Sentences and the Promise of Semantic Completeness.
Analysis 75: 573–584.
Beall, J., and J. Murzi. 2013. Two Flavors of Curry Paradox. Journal of Philosophy
110: 143–165.
Beall, J., and G. Restall. 2000. Logical Pluralism. Australasian Journal of
Philosophy 78 (4): 475–493.

Beall, J., R.T. Brady, A.P. Hazen, G. Priest, and G. Restall. 2006. Relevant
Restricted Quantification. Journal of Philosophical Logic 35 (6): 587–598.
Boolos, G. 1985. Nominalist Platonism. The Philosophical Review 94: 327–344.
Brady, R. 2006. Universal Logic. Stanford: CSLI Publications.
Bueno, O., and M. Colyvan. 2004. Logical Non-apriorism and the Law of Non-
contradiction. In The Law of Non-contradiction: New Philosophical Essays, ed.
Graham Priest, Jc Beall, and Bradley P. Armour-Garb, 156–175. Oxford:
118 (470): 295–321.
Chihara, C. 1979. The Semantic Paradoxes: A Diagnostic Investigation.
Clarke, S.G. 1987. Anti-theory in Ethics. American Philosophical Quarterly 24
(3): 237–244.
Cook, R.T. 2002. Vagueness and Mathematical Precision. Mind 111: 227–247.
———. 2014. There Is No Paradox of Logical Validity. Logica Universalis 8
(3–4): 447–467.
Research 86 (3): 563–579.
Dancy, J. 1983. Ethical Particularism and Morally Relevant Properties. Mind 92
(368): 530–547.
Dummett, M. 1975. Wang’s Paradox. Synthese 30 (3–4): 301–324.
Dutilh-Novaes, C. 2012. Reassessing Logical Hylomorphism and the
Demarcation of Logical Constants. Synthese 185 (3): 387–410.
Eklund, M. 2002. Inconsistent Languages. Philosophy and Phenomenological
Research 64 (2): 251–275.
———. 2012. The Multitude View on Logic. In New Waves in the Philosophy of
Logic, ed. G. Restall and G. Russell, 217–240. Palgrave Macmillan.
Estrada-González, L. 2011. On the Meaning of Connectives (Apropos of a
Non-necessitarianist Challenge). Logica Universalis 5 (1): 115–126.
Etchemendy, J. 1990. The Concept of Logical Consequence. Cambridge, MA:
Harvard University Press.
Field, H. 2008. Saving Truth from Paradox. Oxford: Oxford University Press.
———. 2009. Pluralism in Logic. Review of Symbolic Logic 2 (2): 342–359.
Florio, S. 2014. Unrestricted Quantification. Philosophy Compass 9 (7):
441–454.
326 A. J. Cotnoir
Forster, T.M. 1992. Set Theory with a Universal Set. Oxford: Oxford University
Press.
Frege, G. 2002. Begriffsschrift, a Formula Language, Modeled upon that of
Arithmetic, for Pure Thought. In From Frege to Gödel: A Source Book in
Mathematical Logic, 1879–1931, ed. J. Van Heijenoort, 1–82. Cambridge,
MA: Harvard University Press.
Glanzberg, M. 2015. Logical Consequence and Natural Language. In
Foundations of Logical Consequence, ed. C.R. Caret and O.T. Hjortland,
Grim, P. 1991. The Incomplete Universe: Totality, Knowledge, and Truth.
Cambridge, MA: MIT Press.
Haack, S. 1978. Philosophy of Logics. Cambridge: Cambridge University Press.
Hofweber, T. 2007. Validity, Paradox, and the Ideal of Deductive Logic. In
Revenge of the Liar, ed. J. Beall, 145–158. Oxford: Oxford University Press.
Jané, I. 1993. A Critical Appraisal of Second-Order Logic. History and Philosophy
of Logic 14: 67–86.
Lavine, S. 2006. Something About Everything: Universal Quantification in the
Universal Sense of Universal Quantification. In Absolute Generality, ed.
A. Rayo and G. Uzquiano, 98–148. Oxford: Oxford University Press.
Lewis, D.K. 1991. Parts of Classes. Oxford: Blackwell.
Linnebo, Ø. 2003. Plural Quantification Exposed. Noûs 37 (1): 71–92.
———. 2006. Sets, Properties, Unrestricted Quantification. In Absolute
Generality, ed. A. Rayo and G. Uzquiano, 149–178. Oxford: Oxford
University Press.
———. 2010. Pluralities and Sets. Journal of Philosophy 107 (3): 144–164.
Linnebo, Ø., and A. Rayo. 2012. Hierarchies Ontological and Ideological. Mind
121 (482): 269–308.
Ludwig, K. 2002. What Is the Role of a Truth Theory in a Meaning Theory. In
Meaning and Truth: Investigations in Philosophical Semantics, ed. O. Campbell,
D. Shier, and J.K. Campbell. New York: Seven Bridges Press.
Ludwig, K., and G. Ray. 2002. Vagueness and the Sorites Paradox. Philosophical
Perspectives 16: 419–461.
Lynch, M.P. 2008. Alethic Pluralism, Logical Consequence, and the Universality
of Reason. Midwest Studies in Philosophy 32 (1): 122–140.
MacFarlane, J.G. 2002. What Does It Mean to Say that Logic Is Formal? PhD
thesis, Citeseer.
Mackie, J.L. 1977. Ethics: Inventing Right and Wrong. New York: Pelican Books.
McGee, V. 2003. Universal Universal Quantification. In Liars and Heaps, ed.

M. Glanzberg and J. Beall, 357–364. Oxford: Oxford University Press.
———. 2006. There Is a Rule for Everything. In Absolute Generality, ed. A. Rayo
and G. Uzquiano, 179–202. Oxford: Oxford University Press.
Mortensen, C. 1989. Anything is possible. Erkenntnis 30 (3): 319–337.
Murzi, J. 2014. The Inexpressibility of Validity. Analysis 74 (1): 65–81.
Patterson, D. 2009. Inconsistency Theories of Semantic Paradox. Philosophy and
Pedersen, Nikolaj J.L.L. 2014. Pluralism x 3: Truth, Logic, Metaphysics.
Erkenntnis 79 (2): 259–277.
Priest, G. 1984. Semantic Closure. Studia Logica: An International Journal for
Symbolic Logic 43 (1/2): 117–129.
———. 2002. Beyond the Limits of Thought. Oxford: Oxford University Press.
———. 2006. Doubt Truth to Be a Liar. Oxford: Oxford University Press.
———. 2007. Review of Absolute Generality. Notre Dame Philosophical Reviews.
https://ndpr.nd.edu/news/absolute-generality/.
Quine, W.V.O. 1937. New Foundations for Mathematical Logic. The American
Mathematical Monthly 44: 70–80.
Rayo, A. 2002. Word and Objects. Noûs 36 (3): 436–464.
———. 2006. Beyond Plurals. In Absolute Generality, ed. A. Rayo and
G. Uzquiano, 220–254. Oxford: Oxford University Press.
Rayo, A., and G. Uzquiano. 2006. Introduction. In Absolute Generality, ed.
A. Rayo and G. Uzquiano, 1–19. Oxford: Oxford University Press.
Read, S. 2006. Monism: The One True Logic. In A Logical Approach to Philosophy,
ed. D. DeVidi and T. Kenyon, 193–209. New York: Springer.
Resnik, M.D. 1988. Second-Order Logic Still Wild. The Journal of Philosophy 85
(2): 75–87.
———. 2004. Revising Logic. In The Law of Non-contradiction: New
Philosophical Essays, ed. Graham Priest, J.C. Beall, and Bradley P. Armour-
Garb, 178–193. Oxford: Oxford University Press.
Russell, G. 2008. One True Logic? Journal of Philosophical Logic 37 (6): 593–611.
———. 2017. An Introduction to Logical Nihilism. In Logic, Methodology and
Philosophy of Science, ed. P.S. Hannes Leitgeb, Ilkka Niiniluoto, and E. Sober.
Sagi, G. 2014. Formality in Logic: From Logical Terms to Semantic Constraints.
Logique et Analyse 56 (227): 259–276.
Sainsbury, R.M. 1996. Concepts Without Boundaries. In Vagueness: A Reader,
ed. R. Keefe and P. Smith, 186–205. Cambridge, MA: MIT Press.
328 A. J. Cotnoir
Shapiro, S. 2006. Vagueness in Context. Oxford: Oxford University Press.

———. 2011. Varieties of Pluralism and Relativism for Logic. In A Companion
to Relativism, ed. D. Hales, 526–552. Malden: Wiley-Blackwell.
———. 2014. Varieties of logic. Oxford: Oxford University Press.
Sher, G. 1996. Did Tarski Commit ‘tarski’s Fallacy’? Journal of Symbolic Logic 61
(2): 653–686.
Strawson, P.F. 1950. On Referring. Mind 59 (235): 320–344.
Studd, J. 2013. The Iterative Conception of Set. Journal of Philosophical Logic
42: 697–725.
Tarski, A. 1936. On the Concept of Logical Consequence. 2nd ed, 409–420.
Indianapolis: Hackett.
———. 1944. The Semantic Conception of Truth: And the Foundations of
Semantics. Philosophy and Phenomenological Research 4 (3): 341–376.
Tye, M. 1994. Sorites Paradoxes and the Semantics of Vagueness. Philosophical
Perspectives 8: 189–206.
Uzquiano, G. 2006. The Price of Universality. Philosophical Studies 129:
137–169.
———. 2009. Quantification Without a Domain. In New Waves in the
Philosophy of Mathematics, ed. O. Bueno and Ø. Linnebo, 300–323. London:
Palgrave Macmillan.
Varzi, A.C. 2002. On Logical Relativity. Noûs 36 (1): 197–219.
Wansing, H., and G. Priest. 2015. External Curries. Journal of Philosophical
Logic 44: 453–471.
Weber, Z. 2010. Transfinite Numbers in Paraconsistent Set Theory. Review of
Symbolic Logic 3 (1): 71–92.
———. 2012. Transfinite Cardinals in Paraconsistent Set Theory. Review of
Symbolic Logic 5 (2): 269–293.
———. 2014. Naive Validity. The Philosophical Quarterly 64 (254): 99–114.
Weir, A. 2006. Is It Too Much to Ask, to Ask for Everything? In Absolute
Generality, ed. A. Rayo and G. Uzquiano, 333–368. Oxford: Oxford
University Press.
Whittle, B. 2004. Dialetheism, Logical Consequence, and Hierarchy. Analysis
64 (4): 318–326.
Williamson, T. 2003. Everything. Philosophical Perspectives 17 (1): 415–465.
———. 2006. Absolute Identity and Absolute Generality. In Absolute Generality,
ed. A. Rayo and G. Uzquiano, 369–390. Oxford: Oxford University Press.
———. 2017. Semantic Paradoxes and Abductive Methodology. In The
Relevance of the Liar, ed. B. Armour-Garb, 325–346. Oxford: Oxford
University Press.
Wright, C. 1975. On the Coherence of Vague Predicates. Synthese 30 (3/4):

325–365.
———. 2010. The Illusion of Higher-Order Vagueness. In Cuts and Clouds:
Vagueness, Its Nature, and Its Logic, ed. R. Dietz and S. Moruzzi. Oxford:
Zardini, E. 2010. Truth Preservation in Context and in Its Place. In Insolubles
and Consequences, ed. C. Dutilh-Novaes and O. Hjortland, 249–271.
———. 2013. Naive Logical Properties and Structural Properties. The Journal
———. 2014. Context and Consequence: An Intercontextual Substructural
Logic. Synthese 191: 3473–3500.
Varieties of Logical Consequence
by Their Resistance to Logical Nihilism
Gillian Russell
Recent work on logical pluralism has suggested that the view is in

danger of collapsing into logical nihilism, the view on which there are
no valid arguments at all.1 The goal of this chapter is to argue that the
prospects for resisting such a collapse vary quite considerably with
one’s account of logical consequence. The first section lays out four
varieties of logical consequence, beginning with the approaches
Etchemendy (1990) called interpretational and representational, and
then adding a Quinean substitutional approach as well as the more
recent universalist account given in Williamson (2013, 2017). The
second section recounts how the threat of logical nihilism arises in the
debate over logical pluralism. The third and final section looks at the
ways the rival accounts of logical consequence are better or worse
placed to resist the threat.2
G. Russell (*)
Department of Philosophy, University of North Carolina at Chapel Hill,
Chapel Hill, NC, USA
e-mail: gillian_russell@unc.edu

https://doi.org/10.1007/978-3-319-98346-2_14
332 G. Russell
1 Varieties of Logical Consequence

In a volume on logical pluralism, it is not unusual to write or assume that
there are different views of logical consequence; classical and relevant
logicians disagree about the extension of the relation, and monists and
pluralists disagree in a different way about whether there can be more
than one consequence relation. Still, these varieties of logical consequence
are not the main ones that I wish to draw your attention to in this section.
Two monists who agree on a logic can have different views of logical con-
sequence in the sense I have in mind. Essentially, that is because they
might agree that an argument form—for example, modus ponens—is
valid, but disagree about the features in virtue of which it is valid. We
might say that they disagree about what logical consequence is or, to bor-
row a phrase from Etchemendy, that they disagree about the concept of
logical consequence.3
So what is logical consequence? In informal philosophy it has often
been said—in fact, I’ve said it quite emphatically myself—that an argu-
ment is valid just in case any possible world in which all of the premises
are true is one in which the conclusion is true. I’ll call this characterisa-
tion the modal slogan. Once we get down to work, the modal slogan gets
us into trouble. One problem is that there are sentences which express
necessary truths which are not logically valid, such as 2 + 2 = 4, Hesperus
is Phosphorus and Water is not CO2.4 This problem has a natural counter-
part for consequence: there are no possible worlds where Hesperus is bright
is true but Phosphorus is bright is not—and the latter is still not a logical
consequence of the former. A second problem for the modal slogan is the
converse of the first: indexical logics like Kaplan’s Logic of Demonstratives
have logical truths which need not express necessary truths, such as Ap ↔
p, Np → p, NLocated(I, Here) and dthat[α] = α.5 This problem too has a
counterpart for logical consequence: there are possible worlds in which I
am not here now, but in which snow is white. Still, since I am here now is
a logical truth (in LD), it is a logical consequence of any arbitrary set of
sentences, including {snow is white}. This gives us an example of an argu-
ment where Α ⊨ Β even though there are possible worlds where Α is true
but Β is not. Hence the modal slogan both under- and over-generates
instances of logical consequence.6
Varieties of Logical Consequence by Their Resistance to Logical… 333
In practice, these problems present few difficulties when we are doing

formal logic, and that is because the modal slogan is dropped in favour of
a characterisation in terms of models. Exact definitions vary slightly from
author to author, but they are generally acknowledged to be descendants
of Tarski’s definition:
The sentence X follows logically from the sentences of the class K if and
only if every model of the class K is also a model of the sentence X. (Tarski
1936, p. 417)
Tarski uses ‘model of a sentence X’ to mean something (in his paper it

was a sequence of objects, in contemporary formulations it is a model)
which satisfies X. Where—as is usually the case in an argument—X is a
sentence (a formula with no free variables), a model of X will be a model
which makes X true. Hence we might say:
The sentence X is a logical consequence of the set of sentences K just in case

every model that makes K true also makes X true.
The structural similarity to the modal slogan is clear, but it would be

dangerous, for the reasons given above, to assume that models represent
possible worlds.
Our first two varieties of logical consequence differ in their views of
what models do represent. Roughly, on the interpretational account,
models represent different assignments of meanings to the non-logical
expressions in our language. On the representational account, models rep-
resent different arrangements of the world.7
The difference is perhaps best illustrated with an example, so suppose
that we have a first-order language, L, with logical expressions (∧, ¬, ∀,
=) as well as non-logical individual constants (a, b, c, …), non-logical
predicates (P, Q, R, …), first-order variables (x, y, z, …) and punctuation
((, ), , ). We adopt a standard definition for wffs (well-formed formulas
and sentences).8 A model M for L consists of a pair (D, I) with D a non-
empty set (the model’s domain) and I an interpretation function which
assigns appropriate extensions from D to all the non-logical expressions
in the language.9
334 G. Russell
Now consider two models, K and J. K = 〈DK, IK〉 and is such that the

sentence Fa is true in K, because a I K (the denotation assigned to a in K)
is a member of F I K (the extension assigned to the one-place predicate F
in K). Model J differs on this: Fa is false in J, that is, IJ(Fa) = F, since IJ
assigns an object to a which is not in the set F I J .
On the representational conception of logical consequence, K and J
don’t differ on the meaning of any of the expressions in Fa, rather they
differ in representing situations in which the referent of a ‘is F’ or not. On
this way of thinking about it, models represent something like states of
the world (admittedly in a highly abstract fashion, but that is to be
expected in logic) and so when a set of premises entails a conclusion, that
is because no matter how the world is, if the premises are true, the con-
clusion is.
On the interpretational conception—by contrast—K and J differ in
assigning different meanings to the expressions a and F. With the meanings
assigned to the expressions in K, the sentence is true, but with the mean-
ings assigned to it in J, the sentence is false. On this view, if the conclusion
of an argument is true in all the models in which the premises are all true,
then that is because no matter how we interpret the non-logical expres-
sions in the argument, if the premises are true, then the conclusion is.
These two views of formal models fit naturally with two different intui-
tive conceptions of what logical consequence is. On one, ¬ (Snow is white
∨ Grass is green) entails ¬ snow is white because of the meanings of ∨ and
¬. Since the meanings of ∨ and ¬ are sufficient to guarantee that the argu-
ment is valid, it won’t matter what interpretation the sentences Snow is
white and Grass is green receive. So on this intuitive conception, there are
two natural ways to test for validity: (i) keep the logical constants the
same but vary the interpretation of the non-logical expressions or (ii)
keep the logical constants the same but substitute different expressions
for the non-logical expressions.
We might call these the meaning variation and word substitution tests,
respectively. The latter is an approach that Quine endorsed:
The relation of implication in one fairly natural sense of the term, viz. logi-
cal implication is readily described with the help of the auxiliary notion of
logical truth. A statement is logically true if it is not only true but remains
true when all but its logical skeleton is varied at will; in other words, if it is
true and contains only logical expressions essentially, any others vacuously.
Now one statement may be said to logically imply another when the truth-
functional conditional which has the one statement as an antecedent the
other as consequent is logically true. (Quine 1981, §5)10
There’s a natural problem with this substitutional approach though,

identified already by Tarski11: it makes the relation of logical consequence
depend on the richness of the available language.
This condition may in fact be satisfied only because the language with
which we are dealing does not possess a sufficient stock of extra-logical
constants. The condition […] could be regarded as sufficient for the sen-
tence X to follow from the class K only if the designations of all possible
objects occurred in the language in question. This assumption, however, is
fictitious and can never be realised. (Tarski 1936, pp. 415–416)
The meaning-variation test avoids this problem. Rather than substituting,

for example, a name that already refers to the object which is not in the
extension of ‘F,’ we reinterpret the original name, a, so that it refers to that
object, giving us our counterexample. This was Tarski’s approach. He
conceived of a valid argument as one that preserved truth no-matter how
we reinterpreted the non-logical expressions.
Though we usually regard contemporary definitions of logical conse-
quence as descendants of Tarski’s 1936 definition, there are several strik-
ing differences between it and contemporary definitions. The first is that
rather than assigning alternative interpretations to the non-logical expres-
sions directly, Tarski first forms the sentential function corresponding to
each sentence by replacing the non-logical expressions with syntactically
appropriate variables (like constants with like variables, unlike with
unlike). For example, we might start with the sentence ¬Odd(2) ∧
Prime(2) and form the sentential function:
ØXx Ù Yx (1)
Here the non-logical predicates Odd and Prime have been replaced
with the predicate variables X and Y, and the individual constant 2 has
been replaced with the individual variable, x, to form a sentential
function.
336 G. Russell
Sentential functions are then satisfied (or not) by the closest thing in
Tarski’s system to a modern-day model: a sequence of objects. A sequence
of objects which would satisfy (1) above would be one which began (the
set of planets, Ceres, the set of moons, …)—assuming the ‘objects’ in this
sequence are assigned to the non-logical expressions in the order they
appear in the sentence. A sequence of objects which would not satisfy (1)
would be (the set of even numbers, 4, the set of prime numbers). Such
sequences can easily be reconceptualised as functions from the non-
logical expressions to suitable meanings, and hence as an analogue to our
modern-day interpretations.
Another difference between Tarski’s definition and contemporary ones
is that Tarski’s models have no counterpart for the contemporary domain
of quantification. As Etchemendy notes, domains are a very natural addi-
tion to a model if you are thinking of consequence representationally—
we can represent worlds with fewer or more objects—but seem
unmotivated on the interpretational account Tarski favoured12; we
wouldn’t expect a reinterpretation of the non-logical expressions to affect
the number of objects there are. Still, we normally think that variable
domains are important for achieving the right extension for ‘logical con-
sequence’—we wouldn’t want ∃x ∃ y(x ≠ y) (or any other sentence equiv-
alent to the claim that there are at least n objects for some n ∈ ℕ) to be
true in every model (and thus a logical truth) or false in every model (and
thus a logical falsehood). If we think of models representationally, this
naturally motivates the inclusion of a domain, which then allows us to
have variable domain models, resulting in (what we normally think of as)
a sensible extension for the consequence relation.
On the other hand, the representational conception of a model looks
less natural—and the interpretational conception more natural—when it
comes to thinking about why two names, say a and b, might be assigned
the same referent in one model and a different referent in another. On the
interpretations account, this simply represents an alternative semantic
value for at least one of the names, a reinterpretation like any other. Just
as 2 could be a name for Julius Caesar, so Hesperus could be a name for
Mars. On the representationalist account—where we assume that the
semantic value of a name is an object—this looks much harder to explain.
If a refers to Venus and b refers to Venus, it is hard to see how we can
retain those interpretations (as the spirit of the account instructs us to do)
and somehow—by changing the world—end up with a model in which
I(a) ≠ (b).
Here is where we are: we have three varieties of logical consequence:
interpretational, representational, and substitutional. On the interpreta-
tional account, we vary the interpretations of the non-logical expressions
in an argument, and if every way of doing so makes the conclusion true
if the premises are all true, then the conclusion is a logical consequence
of the premises. This view is the historical ancestor of contemporary
model theory, and it makes some limited sense of that model theory, but
it is hard to motivate contemporary models containing variable domains
from this perspective. On the substitutional account, we keep the logical
expressions fixed, and substitute syntactically appropriate alternative
expressions uniformly throughout the argument. If doing so can result in
true premises but a non-true conclusion, the argument is not valid. This
account was endorsed by Quine, but makes the extension of the relation
of logical consequence hostage to the richness of the language. On the
representational account we retain the interpretation of all expressions in
the argument, but consider various different states the world might be in.
If every state that makes all the premises true makes the conclusion true,
then the conclusion is a logical consequence of the premises. This moti-
vates the variable domains of contemporary models quite naturally, but it
is hard to see why different models should be allowed to assign different
interpretations to expressions which actually share an interpretation
(which they are permitted to do in contemporary model theory). Now it
only remains to introduce the fourth and final conception of logical con-
sequence that I will consider.
Universalist Logical Consequence
On the accounts we have looked at so far, the relation of logical conse-

quence is a metalinguistic one, concerning preservation of the property
of truth between a set of premises and a conclusion. Timothy Williamson
(2013) proposes an alternative, non-metalinguistic conception of logical
consequence, one which is again a recognisable descendent of Tarski’s
338 G. Russell
approach. The view is easier to understand if we start with Williamson’s

technical definition, and this is in turn easier to grasp if we begin with
logical truth, and then generalise to logical consequence afterwards.
Call a sentence A a logical truth just in case a related sentence A′, gen-
erated by uniformly replacing all the non-logical vocabulary with appro-
priate new variables (distinct vocabulary with distinct variables) and
closing the result under universal quantification, is true. For example,
each of the sentences on the left below is a logical truth just in case its
counterpart on the right is true:
a=a "x ( x = x )
Fa Ú ØFa "X"x ( Xx Ú ØXx )
Af ® f "X ( AX ® X )

Among other things, these illustrations make it clear that we are to

quantify into all non-logical positions, including predicate and sentential
positions. Call the result of replacing all the non-logical expressions with
appropriate variables the sentence’s shell. Then we can express the new
definition succinctly:
Definition 1 (Logical Truth) A is a logical truth just in case the universal

closure of A’s shell is true.13
The universal closure of a sentence will not (in general) be a metalin-

guistic claim but merely a very general one, and hence the truth condi-
tions for claims about logical truth do not require metalinguistic facts to
hold. One might reasonably wonder what is required to hold so that
∀X ∀ x(Xx ∨ ¬ Xx) or ∀X(X ∨ ¬ X) are true. How is quantification into
predicate position or sentence position to be understood? One view
would be that we are quantifying over properties and propositions,
respectively, and that, for example, ∀X(X ∨ ¬ X) is true just in case what-
ever proposition we assign to the variable X, X ∨ ¬ X is true. Similarly,
whatever object we assign to x and property to X, Xx ∨ ¬ Xx is true.
This raises questions about what properties and propositions are, but
just as first-order logic can make a lot of progress while taking the idea of
an object for granted, so we can pursue this approach further without
endorsing particular views on these matters. All we need to say for now is
that propositions are the values of sentential variables, much as objects
are the values of first-order variables, and similarly properties are the val-
ues of predicate variables.
Now we generalise to logical consequence. This must be done with
some care since there is no obvious way to close an argument’s shell using
quantifiers (we normally quantify into sentences, not arguments). So
consider an argument with premises, such as modus tollens: P → Q,
¬Q ⊨ ¬ P. Call the result of uniformly replacing the non-logical vocabu-
lary with variables throughout the argument the argument’s shell. The
shell for modus tollens is X → Y, ¬Y ⊨ ¬ X. Then we say that the conclu-
sion of the argument is a logical consequence of the premises just in case
every assignment of appropriate values (objects to first-order variables,
properties to predicates, etc.) that makes both X → Y and ¬Y true also
makes ¬X true. Here is another illustration, this time with a first-order
case:
a=b
Fa
Fb
The shell for the argument is:
x=y
Xx
Xy
An assignment will be a function that assigns objects to x and y and a

property to X. Our new definition of logical consequence tells us that the
argument is valid just in case there is no such assignment that makes x = y
and Xx true, and Xy not true. In general:
Definition 2 (Logical Consequence) A sentence is a logical consequence of

a set of premises just in case there is no assignment that makes all the premises
of the argument’s shell true and the conclusion not true.
340 G. Russell
This is our fourth variety of logical consequence. It is much newer

than the other three varieties, and no doubt its merits will become
clearer over time. In this chapter, I want to look at how it fares—in
comparison with the other approaches—against the threat of logical
nihilism.
2 What Is Logical Nihilism?

Logical nihilism—as we will understand it here—is the view that there
are no laws of logic, or equivalently that the relation of logical conse-
quence is empty. For any arbitrary set of premises and conclusion, the
nihilist holds that the premises do not entail that conclusion.14 Putative
laws of logic—laws endorsed by others, but not by the nihilist—include
modus ponens, and the law of excluded middle, conjunction elimination
and the principle of non-contradiction, and more generally, any well-
formed atomic sentence in which ⊨ is the main predicate, flanked by a
name for a set of sentences on the left, and a name for a sentence on the
right, as in:
A ® B, A  B ( MP )
 A Ú ØA ( LEM )
AÙB  A ( ÙE )
 Ø(A Ù ØA ( PNC )
A, ØA  B ( Explosion )
A  A ( ID )

A ® B, B  A ( AC )
Call sentences of this form E-sentences (‘E’ for entailment). Many log-
ics reject one or more of these putative laws. Classical logic rejects (AC),
paraconsistent logics reject (Explosion), Strong Kleene and intuitionistic
logics deny that (LEM) is a law. Logical nihilism rejects them all, includ-
ing relatively uncontroversial principles, such as conjunction elimination
and identity. More comprehensively, for any set of sentences K and any
conclusion X, the nihilist is committed to:
K  X
What this amounts to will depend on what logical consequence is, that
is, on what ⊨ means. Different conceptions of consequence correspond
to different versions of logical nihilism. Representational nihilism is the
view that for any argument, there is always a case—roughly, a state of the
world—in which the premises are true and the conclusion is not.
Interpretationalist nihilism is the view that in any argument there is always
an interpretation of the non-logical expressions which makes the prem-
ises true and the conclusion not. Substitutional nihilism is the view that
there is always a way to substitute expressions for the non-logical expres-
sions in an argument which makes the premises true and the conclusion
not true. And universalist nihilism will be the view that there is always an
assignment of values to the non-logical expressions in an argument which
makes the premises of an argument’s shell true without making the con-
clusion true.
Nihilism, then, is relatively simple to state and understand, even if, like
logical consequence, it comes in several flavours. But one might wonder
why anyone would entertain such an apparently wild view. My goal in
this section will not be to convince you of nihilism itself, but rather to
convince you that the view is a live option, so that you might be inter-
ested in safeguards against it.15
One of the reasons it is a live option stems from the extreme generality
of logical laws. Claims that are very general are, in one sense, very strong,
and hence take much to establish and little to refute. For example, it is
not sufficient for the truth of modus ponens that the truth of
snow is white
and
snow is white ® grass isgreen

342 G. Russell
guarantees that of
grass isgreen
—though I’m confident that it does—because this argument is only one

instance of modus ponens. In order for modus ponens to be a logical law,
every single instance would need to be truth-preserving.16 Similarly, it is
not sufficient for the validity of modus ponens that every single argument
you can think of that takes that form is truth-preserving. There is always
a chance that the counterexample involves some case, or some interpreta-
tion, or some assignment that you haven’t thought of.
Do we have reason to think that every instance of modus ponens is
truth-preserving—including those instances that feature Sorites pred-
icates, truth and other metalinguistic predicates, self-reference, sen-
tences which generalise over the mathematical objects in infinite sets,
context-sensitivity, fictional names, and so on? The history of the
paradoxes—many of which have motivated the rejection of classical
laws of logic—suggests any such faith could be misplaced. While we
might feel confident that the instance above is truth-preserving (and I
do feel confident about that), I am not similarly confident that every
argument of the same form—no matter how rich the language in
which it is formulated—is also so. To the extent one is not confident
of that, one is not confident that modus ponens is a logical law.
Do We Have Proofs of the Logical Laws?
In mathematics, one way to establish even very general claims is by giving

a proof, and certainly we give proofs of logical laws. These proofs are of
two kinds. The first is formal, that is, in a particular formal proof system.
We might, for example, give the following sequent calculus proof of the
Law of Excluded Middle:
cut
Fig. 1 A truth-table proof of Modus tollens
But such a formal proof of the Law of Excluded middle serves as a

proof (in the intuitive sense) of the Law of Excluded Middle only on the
assumption that the proof rules are sound. Logicians who deny the law—
like strong Kleene logicians, intuitionist logicians, and nihilists—reject
this assumption, and so this kind of proof will not establish a law of logic.
A second standard way to establish a law of logic is via informal (though
still rigorous) model-theoretic proof. One example is to use a truth-table
to display classical interpretations (one per row; see Fig. 1 above). A second
example is to use classical first-order models, as in this model-theoretic
proof of Fa ⊨ ∃xFx:
Take an arbitrary model M = (D, I) which makes Fa true. Then
|a|M ∈ |F|M. Consider an assignment v of elements from D to variables
such that v(x) = |a|M. Then Fx is satisfied by to M on v, and hence M
makes ∃xFx true. Generalising, any model which makes Fa true makes
∃xFx true.
Still, paraconsistent logicians don’t take the truth-table above to show
that modus tollens is a logical law, because they hold that it fails to dis-
play all the relevant interpretations. When the truth-table is expanded to
include the interpretations on which sentence letters may receive the
truth-value ‘Both’ (B), we find interpretations on which the premises
may both be ‘Both’ while the conclusion is false—a counterexample to
the law (Fig. 2).
Similarly, with the model-theoretic proof, a free logician will claim
that we have failed to consider relevant models. What about models in
which a’s referent is not a member of the domain of quantification?
So though we have proofs of general logical laws, both formal and
informal, unless we know that one particular set of models, or one par-
ticular proof system, is the right one, we do not know that these are
344 G. Russell
Fig. 2 A paraconsistent truth-table of Modus tollens
c orrect. And if we don’t have proofs of the logical laws, then we don’t have
proofs that logical nihilism is false.
Logical Nihilism and Logical Pluralism
A second reason to concern ourselves with nihilism is that an argument

has been given for it in the literature on logical pluralism.17 On one
much-discussed version of logical pluralism—the view due to Beall and
Restall (2000)—validity is defined using what they call the Generalised
Tarski Thesis (GTT):
(GTT) An argument is valid x if and only if, in every case x in which
the premises are true, so is the conclusion.
This is naturally understood as assuming a version of the representation-

alist approach to consequence. The (GTT) leads to pluralism on Beall
and Restall’s view because the word case is an everyday, imprecise expres-
sion which can be made more precise in a variety of ways. We might
consider truth-preservation over possible worlds, paraconsistent or para-

complete models and more. More than one of these precisifications of
case is, they think, legitimate, and the different legitimate precisifications
lead to different extensions for the word valid.
Priest has pressed an objection to this view18:
The obvious reply to this argument is that it is only truth-preservation over

all situations that is, strictly speaking, validity. One of the points about
deductive logic is that it will work come what may: we do not have to
worry about anything except the premises. (Priest 2006, p. 202)
The objection draws on a view of logic that is both intuitive and has his-
torical precedent, namely, that logic is supposed to be completely general;
if an argument K ⊨ X preserves truth in (say) arithmetic but not in geom-
etry, or in classical cases but not in dialethic ones—then it is not really a
law of logic after all. Logic is supposed to work in absolutely all cases.
Beall and Restall respond with a reductio ad absurdum on the assump-
tion that logic is completely general: they say that this assumption will
lead to absurdity—in particular to logical nihilism, or some similar
absurdly weak logic:
…we see no place to stop the process of generalisation and broadening of

accounts of cases. For all we know, the only inference left in the intersec-
tion of (unrestricted) all logics might be the identity inference. (Beall and
Restall 2000, p. 92)
Elsewhere I’ve argued that not even identity would remain (Russell
2017), but one might think that this is not of the greatest importance.
Most philosophers and logicians regard both all-out logical nihilism and
a view on which only identity is left (or some other very minimal set of
E-sentences—call such a view logical minimalism) as absurd. They think
that if your philosophy of logic leads you to either logical nihilism or
logical minimalism, it must be wrong.
Though discussions of pluralism have tended to assume a case-based
(and so representationalist) approach to consequence, this is inessential to
the debate. A structurally similar issue—nihilism begot by generality—
346 G. Russell
arises for other varieties of consequence as well. It is easy to see this if we

reformulate the GTT to get different conceptions of pluralism for each of
the conceptions of consequence discussed above:
(GTT-s) An argument is validx if and only if on every uniform substi-

tutionx (of appropriate non-logical expressions for the non-
logical expressions in the argument) on which the premises
come out true, so does the conclusion.
(GTT-i) An argument is validx if and only if on every interpretationx
of the non-logical expressions on which the premises are
true, so is the conclusion.
(GTT-r) An argument is validx if and only if in every casex in which
the premises are all true, so is the conclusion.
(GTT-u) An argument is validx if and only if on every assignmentx of
values to the non-logical expressions on which the premises
are true, so is the conclusion.
To the extent that we think that the words substitution, interpretation,

and assignment are—like case—still imprecise and tolerant of various rea-
sonable precisifications, the way is open for those precisifications to result
in different extensions for valid.
Moreover, the monist may object, as above, that this does not take the
every in the respective definitions seriously enough. It is only a logic
which takes all of the different precisifications into account which is—
strictly—logic. And so then the pluralist can suggest again that this com-
mitment to absolute generality in logic will lead to nihilism.
3 Modes of Resistance
So much for the big picture: a commitment to generality in logic seems
to push us towards nihilism, or at least towards weaker and weaker logics.
Still, when we look at the details, the availability, plausibility, and persua-
siveness of certain kinds of counterexample vary with our conception of
consequence. We are now in a position to explore some of this variation
in the third and final section of this chapter. In the coming subsections, I
will look at the issues associated with empty names, vagueness, and self-
reference, in that order. In each case I will begin by sketching a possible

view (or views) of the topic that would lead to a rejection of a classical
entailment principle. Since my goal is to look at safeguards that logics
might have against such things, I won’t argue for these sketches of views
on empty names, vagueness, and so on. Rather, my aim is to put them on
the table so that we can look at the internal resources that each concep-
tion of logic has for resisting them. Things will get quite complicated, but
I will finish by drawing out a general pattern: if our conception of conse-
quence quantifies over elements that go ‘deeper’ than the phenomenon
that is thought to cause trouble, then the phenomenon does not threaten
our logic. Deeper logics, then—and especially Williamson’s universalist
variety—have more resistance to pressures to weaken and, ultimately,
against nihilism. But the cost of this depth is complimentary: it doesn’t
allow us to study the effects of interesting surface phenomena on
consequence.
Empty Names
Classical logic does not model empty names. In classical model theory, an
interpretation must assign an extension to every individual constant.
Natural languages have empty names—Pegasus, Santa, and so on—and
we might want to know about entailment relations on sentences which
contain them. In particular, a commitment to generality in logic could
lead us to ask what will happen to the logic if we consider all names,
including the empty ones.
What will happen depends on one’s conception of logical consequence.
For example, on the substitution-conception, we are supposed to be able
to substitute any name uniformly for another name throughout a valid
argument without losing validity. But if we substitute Pegasus for a
throughout this classical validity:
a = a  $x ( x = a ) (2)

we appear to have a classically valid argument whose premise is a clas-

sical logical truth, but whose conclusion is false. Thus if we reject the
classical restriction on which kinds of terms can be substituted—that is,
348 G. Russell
if we are serious in our commitment to generality—then we come under

pressure to reject this principle of entailment.
Things look different if we are employing (GTT-u). Then the problem
of empty names gets no traction, because (GTT-u) doesn’t quantify over
linguistic expressions, but over assignments of referents to them. Whatever
object we assign to Pegasus in ∃xx=Pegasus (Bismarck, Du Bois, the num-
ber 3, the Eiffel Tower, etc.), the sentence comes out true.
Intuitively, the substitutional conception is saying that a = a is a logical
truth if it is true whatever name we substitute for a, and hence it is mak-
ing a claim about names, including empty names. But the universalist
conception aims to make no claims about names, only about objects. It
says that a = a is a logical truth if (roughly) every object is identical to
itself. It bypasses the problem of empty names, and hence is resistant to
the counterexample.
What about the remaining two varieties, interpretational and repre-
sentational? Let’s begin with interpretational, that is, let’s suppose that
contemporary models represent different ways the non-logical parts of
languages can be. It is fairly clear that the models do this by assigning
different referents to names, that is, it is built into our standard concep-
tion of a model that an interpretation for a name is a referent. That makes
space for the interpretationalist to say that when they consider only com-
plete interpretation functions, they already are considering all (really all)
the possible interpretations for the names.
Still, this response on behalf of (GTT-i) rings a bit hollow, since the
possibility under consideration is precisely that a name has not a referent.
If the goal is to represent all the ways a language can be, having a part that
doesn’t have an interpretation (in the sense of having a referent) is prima
facie one of those ways. So it looks as if the interpretational conception
of consequence ought to consider partial interpretation functions if they
are really to consider all the ways language can be, and this would lead to
a rejection of (2).
The representational variety can seem more ‘worldly’ in spirit than the
interpretational. It regards models as representing different states of the
world—different cases. If Pegasus is an empty name, then there is no
Pegasus. Is there still a case in which Pegasus flies and another in which
Pegasus does not fly? My inclination is to say no. Making ‘Pegasus flies’ true
is a not condition that a case can meet or fail to meet if ‘Pegasus’ has no
meaning, any more than making the sky dfsndfnsf is a condition that
something can meet or fail to meet.
But perhaps you think it is too quick to say that the name Pegasus has no
meaning. It could be associated with some descriptive condition—a
sense—that, as it happens, isn’t satisfied, but could be in some cases. Perhaps
you want to say that Pegasus doesn’t exist, but that there are still models on
which Pegasus flies is true, namely, models in which something satisfies the
condition associated with Pegasus, and that thing flies. Such a conception
of names could still be associated with interpretation functions that assign
objects to names, but we might expect that names would be permitted to
lack a referent, that is, the requirement that the interpretation function be
a complete function over the set of individual constants would be dropped.
This is consistent with the representational approach to models. Just think
of the possible changes to the states of the world as including the deletion
or inclusion of the referents of names, and we keep the language the same
by retaining the name’s descriptive sense (which is ignored in the model).
This isn’t my view of names, for broadly Kripkean reasons. But if you like
this view, then you might think that the representational approach is as
threatened by the existence of empty names as the interpretational one.
To sum up the results of this section, empty names motivate a decisive
weakening of the logic in the case of the substitutional conception of
consequence, but not on the worldly conception. Meanwhile defenders
of the interpretational and representational conceptions can make moves
to avoid admitting incomplete interpretation functions on names. The
defender of the interpretational conception would deny that ‘empty’ is a
legitimate interpretation of a name (because interpretations of names are
referents), and the defender of the representational conception would
deny that empty names can be used to place conditions on cases. It is
perhaps worth noting, first, that both these ways of avoiding weakenings
have a decidedly Millian flavour, and second, that on the conceptions of
logical consequence that permit us to avoid weakening in the face of
empty names, we do this by finding a way of thinking of logic and/or
names that allows us to ignore the phenomenon. This is somewhat unsat-
isfying, and this is a point that I will return to at the end of this chapter.
For now, let’s look at a new threat.
350 G. Russell
Vagueness and Incomplete Predicates
Predicates in classical models must have complete extensions, where this

means that the extension (the set of things of which the predicate is true)
and the anti-extension of the predicate (the set of things of which the
negation of the predicate is true) exhaust the domain of the model: every-
thing in the domain is in one set or the other.
But one might think that the extensions of vague predicates in natu-
ral languages are incomplete. On one version of the idea, the conven-
tional meanings associated with our colour predicates determine a range
of shades which are red, and a range of shades which are not red, but do
not legislate on all of the shades which fall in between. We might imag-
ine a series of colour cards progressing gradually from the brightest of
reds at one end to the pinkest of pinks at the other. (Represented in
black and white below.) Anything which is pink is not red, but while
there are shades on one end which fall in the conventional extension of
‘red’, and shades at the other end which fall in the extension of ‘pink’
(and so in the anti-extension of ‘red’) there are shades in the middle
about which our linguistic conventions say neither that the shade is red,
nor that it is not red.
red not-red
This suggests that the interpretation function for a predicate P might

assign an extension |P|+ (the set of things that ‘are P’) and an anti-extension
|P|− (the set of things that ‘are not P’) though these two sets need not
exhaust the domain of the model.19 Then if |a| ∈ |P|+, Pa is true, if
|a| ∈ |P|−, then ¬Pa is true, but if |a| is in neither the extension nor the
anti-extension of the predicate, then both Pa and ¬Pa are neither true nor
false—a third truth status, which I will abbreviate to neither.
To argue for a third truth-status is not yet to have made a case for a
non-classical logic, because a philosopher who believes in three truth-
statuses could still hold that all and only the classical entailments are
valid. But there is a well-worn path from a third truth-status to a non-
classical logic. For if Pa is neither on some interpretations, then what of
super-sentences containing it, such as ¬Pa be or Pa ∨ ¬ Pa? One approach
adopts the Strong Kleene truth-tables, on which ¬Pa is neither if Pa is and
a disjunction takes neither if both disjuncts do. Assuming that a logical
truth is true on all interpretations we have a counterexample to the clas-
sical law of excluded middle when Pa gets the neither truth status.20
 A Ú ØA (3)
Different varieties of consequence have different levels of resistance to

this kind of counterexample. There are some similarities between the
responses open to them here and the responses available in the previous
empty names case. For example, the (GTT-s) is especially vulnerable; if
there are predicates with incomplete extensions and logic is supposed to
consider all uniform substitutions of predicates for predicates, it is hard
to see how we can justify rejecting counterexamples to a logical law based
on substituting incomplete predicates.
The (GTT-i) account of logical consequence is subject to similar pres-
sures. If there are natural language predicates with incomplete extensions,
and we are thinking of models as representing different ways the language
could be, it seems artificially restrictive to ban ways on which predicates
get indeterminate extensions, that is, incomplete interpretation functions
in models. However, just as in the case of empty names, there is an option
here if the interpretationalist has the will to take it: if they take a Russellian
view of propositions, they might wish to maintain that an interpretation
for a predicate simply is a property. They can hold (as in Russell 1918)
that just as the world contains objects, so it contains properties, which
352 G. Russell
serve as the meanings of predicates. To interpret a predicate just is to

assign it one of these worldly properties (much as to interpret a name is
to assign it a referent) and, on the assumption that properties have deter-
minate extensions, interpreted predicates must similarly have determi-
nate extensions. Still, perhaps even more so than in the case of names,
this rings a little hollow: we are considering modifying a logic to take
account of predicates with incomplete extensions, since it seems that this
is one of the ways language can be. Though a strict Russellian might
simply redescribe this view of vagueness as one on which one of the ways
language can be is uninterpreted (or improperly or incompletely inter-
preted), this is still, intuitively, one of the ways language can be. To insist
on restricting the interpretations we look at to incomplete ones seems
incompatible with a commitment to generality.
On the representational account, the meanings of all the expressions in
an argument we are testing remain as they are, and we consider whether
there are any ways the world could be which make the premises true and
the conclusion not. If the P in Pa ∨ ¬ Pa is a determinate predicate (say
‘is prime’), it would seem that no matter what worldly changes we make
to the referent of a (or its surrounding world)21 Pa will be true or false (but
not neither), and so Pa ∨ ¬ Pa would always be true. The problem is that
still won’t mean that the law of excluded middle is a logical truth, given
the way the representational approach is normally implemented. Because
representationalists concede that logic is supposed to be formal, so that
what we would need for the LEM to continue to hold is the stronger
claim that all sentences of the form A ∨ ¬ A are true in all models, and a
sentence Qa ∨ ¬ Qa, where Q is an indeterminate predicate, will provide
us with a counterexample.
Finally, let’s look at the universalist account, for here things are differ-
ent again. On this conception, when we ask whether Qa ∨ ¬ Qa is a
logical truth, we are asking whether there is any assignment of proper-
ties and objects to X and x, respectively, in the sentence’s shell—
Xx ∨ ¬ Xx—such that the shell does not come out true. Here there is
no direct reference to indeterminate predicates, and the question natu-
rally arises, whether there is any such thing as an indeterminate property.
No doubt there is much to be said either way, but there is clearly space
for a fan of the worldly account to take the following view: while there
are indeterminacies in the interpretations of predicates—for example, in

the meanings of predicates—there is no such indeterminacy or vague-
ness in the world itself. Nothing about a cloud, or a rainbow, or a furry
cat, is really indeterminate. The world is as it is, the various molecules
and photons and hairs are where they are, and if the truth-value of a
sentence about a mountain or a heap or the colour of a playing card is
indeterminate, that is just because the rules for when to call something
‘red’ (or ‘a mountain’ or ‘a heap’) don’t tell us enough about what it is to
be called in these worldly cases. This view has several interesting conse-
quences (and perhaps for some they are too interesting), but one of
them is that the existence of indeterminate predicates need have no
impact on logic. Perhaps ‘red(a) ∨ ¬ red(a)’ contains an indeterminate
predicate, but when we ask whether it has the form of a logical truth, we
are asking whether there are any assignments on which Xx ∨ ¬ Xx fails
to come out true, and if there are no assignments which have indetermi-
nate properties in their ranges—because there are no indeterminate
properties—the answer will be ‘no.’
Self-Reference and Overdeterminacy
We will look at one more example before drawing some conclusions. It is

well known that some metalinguistic sentences and expressions—such as
‘This sentence is not true’ and ‘heterological’—generate paradoxes. These
paradoxes have led logicians to weaken logics, in two ways. On the way I
will focus on here—dialetheism—the paradoxes are taken to show that
there are true contradictions. Dialetheists hold that the contradictory
conclusions of the paradoxes are true.22 We can model this using a similar
idea to the one employed to handle indeterminacy: suppose that predi-
cates like ‘true’ and ‘heterological’ are associated with conventional rules
which give a sufficient condition for being ‘F’ and a sufficient condition
for being ‘not F’ but—the paradox shows us—these conditions need not
be exclusive. Some things might meet the condition for falling in the
extension of ‘heterological’, making ‘x is heterological’ true of them.
Some things fall in the anti-extension, making ‘x is not heterological’ true
of them, and some things—like ‘heterological’ itself—fall in both the
354 G. Russell
extension and the anti-extension of ‘F’, making both the sentence and its
negation true.
heterological heterological
palindrome meaningful
monosyllabic heterological polysyllabic
long English
We might call such predicates inconsistent. A pressing worry about this

view is that it leads to triviality via (Explosion):
A,Ø A  B (4)
So dialetheists generally reject (Explosion), and indeed, this seems well

motivated given their view: if there are models in which A is true and ¬A
is true, then there are models in which both A and ¬A are both, and these
need not be models in which B is true. So once again, we have a view
which suggests abandoning a principle of classical logic.
The situation with respect to the various conceptions of logical conse-
quence is in some ways similar here to that in the case of incomplete
predicates. For example:
• if the substitutionalist is committed to generality, and inconsistent

predicates exist, it is hard to see how they can resist the pressure to
include them.
• the interpretationalist can either allow inconsistent interpretations
(naturally leading to a weakening of the logic) or make the Russellian
move of insisting that interpretations just are assignments of worldly
properties and combine it with the metaphysical view that there are no
inconsistent worldly properties.
• the representationalist might hold that there are no cases in which

explosion arguments employing consistent predicates have true prem-
ises and non-true conclusions, but since they allow that logic is formal,
and there will be arguments of the same form with inconsistent predi-
cates, this is not sufficient to save explosion.
• the universalist again has a natural defence: argue that there are no
inconsistent properties and then point out that since their conception
of consequence quantifies over properties, the existence of inconsistent
predicates does not provide any new counterexamples to explosion
(just as the existence of empty names doesn’t show that it’s not true that
every object is self-identical).
The major difference, I think, between the incompleteness and incon-

sistency cases, is that philosophers are generally more adverse to the idea
that there are inconsistent properties than they are to the idea that there
are incomplete properties. That is, despite the parallels between the cases,
people are more sympathetic to the idea that there might be an incom-
pleteness in the world than they are to the idea that there might be incon-
sistency ‘out there.’
4 Conclusion
Things have become quite complicated, but this is only, I think, because
there are four varieties of logical consequence on the table, and some of
these offer more than one option when confronted with a linguistic
expression which does not meet the standard assumptions of classical
logic. The beginnings of a pattern are emerging from this complexity, so
let me try to draw that out. You can take logical consequence to be a rela-
tion that quantifies over various things: expressions, meanings of various
kinds, or over bits of the world (objects and properties and the like.)23
The substitutional view takes it to quantify over expressions, the univer-
salist view takes it to quantify over bits of the world, and the interpreta-
tionalist view can either be understood as quantifying over meanings
or—if you take the line that a meaning just is a bit of the world (this is
what I’ve been calling the ‘Russellian’ take)—over bits of the world. The
356 G. Russell
representationalist view might be thought of as quantifying over cases

(which could be thought of—tolerating some metaphysical hair—as bits
of the modal universe), but in practice the assumption that logic is formal
means that it usually inherits the features of quantifying over
expressions.
Any view which takes consequence to quantify over meanings has the
option of saying that it doesn’t quantify over the kinds of meaning that
result in the problems. For example, if the issue is empty names, then the
problem is arising at the level of reference. An interpretationalist view
could hold that consequence quantifies over assignments of reference,
and hence needn’t take into account interpretations which assign no ref-
erent to a name. If, on the other hand, they quantified over what Carnap
called ‘individual concepts’ or senses of names (in either case an aspect of
meaning which determines the referent) the question would naturally
arise: what about individual concepts which fail to determine a referent
in some circumstances?
And finally, we also sometimes have the option of combing the view of
logical consequence with a view of the metaphysics of the problem. Take
incompleteness. On some views, there are incomplete predicates but no
incomplete meanings—the predicates are incomplete precisely because
they fail to link up to a unique meaning. On others, there are incomplete
predicates, incomplete meanings, but no incompleteness in the world
itself. And on others still there are incomplete predicates, incomplete
meanings, and incomplete worldly properties, like being a mountain or
being orange. We might speak of linguistic vagueness, semantic vague-
ness, and ontic vagueness. We can also talk of linguistic dialetheism,
semantic dialetheism, and ontic dialetheism.
Suppose you think that there are incomplete predicates, but no incom-
plete meanings or worldly properties. Then you can resist counterexam-
ples based on incompleteness by adopting a conception of consequence
that quantifies over meanings or worldly properties—not predicates.
Suppose you think instead that there are incomplete predicate meanings,
but not worldly properties. Then you will probably think that there are
incomplete predicates (the ones that express those meanings) but you can
avoid counterexamples by adopting a conception of consequence that
quantifies over worldly properties. Finally suppose you think that there
are incomplete worldly properties (as you might if you believe in ontic
vagueness). Then you are likely to be stuck with any counterexamples

(and hence stuck with a weaker logic) based on incompleteness, since
there will be predicates which attribute those worldly properties.
So here is a high-level trend (much complicated, as we have seen, in
the details): you can define consequence by quantifying over things on a
vertical spectrum that has linguistic expressions at the top, various layers
of meanings in the middle, and worldly things at the bottom. Various
putative counterexamples arise from phenomena like incompleteness,
inconsistency, and empty names, and for some of these phenomena it is
controversial how ‘deep’ they go. For example, it is controversial whether
incompleteness is a phenomenon restricted to expressions, or whether
there are also incomplete meanings and incomplete properties. Similarly,
there are names that lack referents, and Frege thought there were mean-
ings without referents (names with sense but no referent; though Millians
deny this), but it is hard to see how there could be objects but no refer-
ent—the reference failure phenomenon goes at most as deep as
meanings.
But if you think the phenomenon which generates a counterexample
goes at most as deep as level n, then you can avoid it by selecting a con-
ception of logical consequence that quantifies over a level that is lower
than level n. Say, level n − 1.24 Finally, let me note that there appears to
be a cost to opting to define consequence in a way that avoids the prob-
lems of, for example, empty names and incomplete and inconsistent
predicates, namely, that it merely avoids the problems. Avoidance is sen-
sible when we are engaged in some projects, but more generally, if such
things exist then they can feature in arguments, and we might want to
know the extension of valid over such arguments. A conception of logical
consequence which simply doesn’t apply doesn’t answer these questions—
it merely declines to consider them.
Notes
1. For example, Beall and Restall (2006, p. 92) and Russell (2016).
2. A note on terminology: in this chapter I will follow the common practice
of treating the expressions ‘(logical) consequence,’ ‘entails,’ and ‘valid’ as
intertranslatable—an argument is valid just in case its conclusion is a
358 G. Russell
(logical) consequence of its premises, and this is so just in case the premises
together entail the argument’s conclusion. Hence the different views of
logical consequence that I will be looking at are just as much different
views of validity and different views of entailment. There are two other
words that it might be tempting to use here: implication (perhaps being
Quineanishly careful to use this for entailment rather than for a
conditional) and inference. But following Harman (1986) I take inference
to be a different topic altogether, and I’ll be cautious about using ‘impli-
cation’ because I think Quine’s suggested regimentation (i.e. restricting
its use to talk of the entailment relation) is less entrenched than would
be ideal for successful communication (Quine 1966, pp. 165–166;
Quine 1981, §5).
3. For example, in the title of Etchemendy (1999). I recognise that
Etchemendy’s word ‘concept’ is fighting talk in some philosophical circles.
I mean to talk—as Etchemendy did—about different views of what fea-
tures an argument has to have to be valid and have no special commitment
to using the word ‘concept’ or to any particular construal of that phrase.
4. Sometimes people respond to this problem by distinguishing different
kinds of necessity. 2 + 2 = 4 and Hesperus is Phosphorus, they might
claim, are metaphysically necessary, but not logically necessary, and it is
the logical modality in terms of which logical consequence is defined.
There are two problems with this response. The first is that it replies on
a controversial claim about the kind of necessity possessed by the propo-
sitions expressed by these sentences. In, for example, Kripke’s modal
argument of the necessity of Hesperus is Phosphorus, the ‘□’ that that
sentence inherits is the very same one applied to Hesperus is Hesperus
(Kripke 1980). The second problem is that the account threatens to be
circular. What is logically necessary if it is not the things which hold in
virtue of logic alone—the logical truths. But this assumes that we already
have an account of logical consequence.
5. Kaplan (1989). Here A is the actuality operator, N the now operator,
and α a singular term, so that informal instances of these sentences might
be Actually it is snowing if, and only if, it is snowing., if it is snowing now,
then it is snowing, I am here now and dthat[the shortest spy]=the shortest spy.
6. An additional problem for the modal slogan is that: necessity and logical
truth come apart in the model theory for modal logics. □p may be true
at some points in a model and false at others. This would make no sense
if logical truth were necessity—necessity might be world relative but
logical truth is not.
7. The terminology here is from Etchemendy (1999). I find this terminol-

ogy somewhat difficult to remember, but one useful mnemonic is to
note that the one view is interpretational because it speaks of different
interpretations of the linguistic items, while the other is representational
because it talks about the things that get represented.
8. Here is one: (i) if Π is an n-place predicate and t1,…,tn are n terms, the
Πt1,…,tn is an atomic wff; (ii) if t1 and t2 are terms, then t1 = t1 is an
atomic wff; (iii) if A is a wff, then ¬A is a wff; (iv) if A and B are wffs,
then A∧B is a wff; (v) if A is a wff and x is a first-order variable, then ∀xA
is a wff. If A contains no quantifiers, then all the variables in A are free.
If ∀ is immediately followed by a variable ξ, then we say that it is a
ξ-binding quantifier. In a wff ∀ξA, the ξ-binding quantifier ∀ is said to
bind all free occurrences of ξ in A. A variable which is bound in A is not
free in A. Any wff with no free variables is a sentence.
9. That is, nI will be an element of D, where n is a name, PI will be a subset
of Dn, where P is an n-place predicate, and fI will be a complete function
from Dn to D, where f is an n-place function.
10. See also “Their characteristic [logical truths] is that they not only are true
but stay true even when we make substitutions upon their component
words and phrases as we please, provided merely that the so-called ‘logi-
cal’ words ‘=’, ‘or’, ‘and’, ‘not’, ‘if-then’, ‘everything’, ‘something’, etc.,
stay undisturbed” (Quine 1950, p. 4).
11. Tarski too considers a variation of the word-test approach:
“If, in the sentences of the class K and in the sentence X, the con-
stants—apart from purely logical constants (like signs being every-
where replaced by like signs), and if we denote the class of sentences
thus obtained from K by ‘K′’, and the sentence obtained from X by ‘X′’,
then the sentence X′ must be true provided only that all the sen-
tences of class K′ are true.” (Tarski 1936, p. 415)
12. Etchmendy (1999, pp. 68–69; 111–114).
13. It is a bit loose to speak of ‘the’ universal closure of ‘the’ shell, since there
will be many trivial variants, for example, ∀Y(Y → Y) and ∀Z(Z → Z) are
also universal closures of a shell of P → P. If desired, we might enumerate
the variants and count only the first (or the fifth) as the universalisation.
Since the issue won’t be of importance here, I won’t mention it again.
14. In his chapter in this volume, Aaron Cotnoir distinguishes two views
called ‘logical nihilism’ and focuses on one. The view I have in mind here
is the other one.
360 G. Russell
15. I’ve looked in more depth at one argument for logical nihilism here:
(Russell 2017)
16. All of our conceptions of logical consequence agree that this is necessary
for logical consequence, though some may require more than this.
17. Admittedly, that argument was intended to be a reductio on someone
else’s view. But one person’s reductio ad absurdum is another person’s
argumentum ad absurdum.
18. There are similar lines of argument in Read (2006, p. 61) and Bueno and
Shalkowski (2009, p. 11).
19. The problems of higher order vagueness make it clear that this is not suf-
ficient to accommodate and explain vagueness, but there are still many
views on which the move is regarded as necessary.
20. And in fact Strong Kleene logic provides counterexamples to all the clas-
sical logical truths.
21. We leave to the previous section discussion of what happens if a does not
exist.
22. That is, The Liar sentence is true and the Liar sentence is not true or
‘Heterological’ is heterological and ‘heterological’ is not heterological.
23. And one of the reasons things are especially complicated is that the most
familiar views can be understood as quantifying over either meanings or
bits of the world.
24. It would be especially interesting in future work to examine this pattern
against the phenomenon of context-sensitivity. Context-sensitivity is a
paradigmatically linguistic, rather than worldly, issue, but there are still
views (such as MacFarlane’s non-indexical context-sensitivity) which
allow it to go deeper than others. Moreover, context-sensitivity is known
to have interesting non-classical effects on logic.
References
Bueno, O., and S. Shalkowski. 2009. Modalism and Logical Pluralism. Mind
118: 295–321.
Etchemendy, J. 1990. The Concept of Logical Consequence. Cambridge: Harvard
University Press.
———. 1999. On the Concept of Logical Consequence. Stanford: CSLI.

Harman, G. 1986. Change in View. Cambridge, MA: MIT Press.
Kaplan, D. 1989. Demonstratives: An Essay on the Semantics, Logic,
Metaphysics, and Epistemology of Demonstratives. In Themes from Kaplan,
ed. J. Almog, J. Perry, and H. Wettstein. New York: Oxford University Press.
Kripke, S.A. 1980. Naming and Necessity. Oxford: Blackwell.
Priest, G. 2006. Doubt Truth to be a Liar. Oxford: Oxford University Press.
Quine, W.V.O. 1950. Methods of Logic. Cambridge, MA: Harvard University
Press.
———. 1966. Three Grades of Modal Involvement. In The Ways of Paradox, and
Other Essays, 158–176. Cambridge, MA: Harvard University Press.
———. 1981. Mathematical Logic. 3rd ed. Cambridge: Harvard University
Press.
Read, S. 2006. Monism: The One True Logic. In A Logical Approach to Philosophy:
Essays in Honour of Graham Solomon, ed. D. DeVidi and T. Kenyon, 193–209.
Berlin: Springer.
Russell, B. 1918. The Philosophy of Logical Atomism. Chicago/LaSalle: Open
Court Classics.
Russell, G. 2016. Logical Pluralism. In The Stanford Encyclopedia of Philosophy,
ed. E.N. Zalta, CSLI, Summer 2013 edn.
———. 2017. An Introduction to Logical Nihilism. In Logic, Methodology and
Philosophy of Science—Proceedings of the 15th International Congress. College
Publications.
Tarski, A. 1983/1936. On the Concept of Logical Consequence. In Logic,
Semantics and Metamathematics, ed. J. Corcoran, 2nd ed., 409–420.
Williamson, T. 2013. Modal Logic as Metaphysics. Oxford: Oxford University
Press.
———. 2017. Semantic Paradoxes and Abductive Methodology. In Reflections
on the Liar, ed. B. Armour-Garb, 325–346. Oxford: Oxford University Press.
Part III
Connections
Pluralism About Pluralisms
Roy T. Cook
1 Introduction
In recent years various sorts of pluralisms have come into vogue in the
philosophy of logic. For present purposes, two in particular stand out:
• Logical Pluralism: The claim that two (or more) distinct logics are cor-
rect (or legitimate, or best).
• Alethic Pluralism: The claim that two (or more) distinct accounts of
truth are correct (or legitimate, or best).
Pluralisms, of either variety, come in two flavors1:
• Domain-specific Pluralism: There are two (or more) distinct accounts of

[insert the phenomenon in question], and each applies to a distinct
domain (typically, one not overlapping the domains to which the other
accounts apply).
R. T. Cook (*)
Department of Philosophy, University of Minnesota, Minneapolis, MN, USA
e-mail: cookx432@umn.edu

https://doi.org/10.1007/978-3-319-98346-2_15
366 R. T. Cook
• Domain-independent Pluralism: There are two (or more) distinct

accounts of [insert the phenomenon in question], and each applies
everywhere.
Interestingly, the most well-developed account of alethic pluralism,

Michael Lynch’s account as developed in Lynch (2008, 2009) and else-
where, is a domain-specific alethic pluralism, while the first (and per-
haps still most well-developed) account of logical pluralism—JC Beall
and Greg Restall’s account as developed in Beall and Restall (2000,
2001, 2006)—is a domain-independent logical pluralism. Stewart
Shapiro’s logical relativism—presented in Shapiro (2015)—presents a
rich, domain-specific logical pluralism as an alternative to the Beall-
Restall domain-independent approach. Domain-independent alethic
pluralism remains the most under-examined of the four options,
although Sher (2013) is a promising recent exploration in this
direction.2
Here we shall focus on the domain-specific versions of both alethic
pluralism and logical pluralism. The reason is not that these accounts are
in some way superior to domain-independent pluralisms about the same
subject matter.3 The reason that we will restrict our attention for the most
part to domain-specific pluralisms is that we are interested in the connec-
tions, if any, between alethic pluralism and logical pluralism—and in
particular, whether the former entails the latter or vice versa—and the
only extant arguments that I am aware of for such connections focus on
domain-specific accounts. I have no doubt that arguments for or against
the existence of connections between domain-independent alethic plu-
ralism and domain-independent logical pluralism (or between combina-
tions of domain-specific and domain-independent pluralisms) can be
formulated, and hopefully some of the work carried out here will prove
helpful in such endeavors. But this chapter will focus primarily on the
domain-specific case, although I will return briefly to other combina-
tions, noting some apparent difficulties in formulating similar arguments
for independence, in the concluding section.
If domain-specific alethic pluralism entails domain-specific logical
pluralism, and vice versa, then in some sense we really only have one
Pluralism About Pluralisms 367
pluralism, not two. If, however, the two sorts of pluralism are indepen-
dent of each other, then we truly have two distinct kinds of pluralism—
that is, we have a plurality of pluralisms. The purpose of this chapter is to
argue that domain-specific alethic pluralism does not entail domain-
specific logical pluralism (contrary to Lynch (2008, 2009) and Pedersen
(2014)), nor does domain-specific logical pluralism entail domain-
specific alethic pluralism, and hence we do have such a plurality of
pluralisms.4
To accomplish this, in Sect. 2 I will present a simple model that shows
both how one might be a domain-specific pluralist about logic while
being a monist about truth and how one might be a domain-specific
pluralist about truth while being a monist about logic. I will then, in
Sect. 3, use this model to identify the fallacy in the arguments found in
Lynch (2008, 2009) and Pedersen (2014).5 Section 4 will then further
flesh out the model, distinguishing between different senses in which a
domain might be epistemically constrained. Finally, in a short conclud-
ing section, I will tie up some loose ends and gesture in the direction of
the aforementioned difficulties in extending the present account to
domain-independent pluralisms (and combinations of domain-specific
and domain-independent pluralisms).
Before moving on, a methodological note is in order: Although this
chapter is a response to arguments found in Lynch (2008, 2009) and
Pedersen (2014), the arguments made below do not only apply to ver-
sions of domain-specific alethic pluralism that take truth to be a func-
tional concept manifested by different lower-level properties in
different domains. The arguments are much more general than this,
applying to any view according to which truth obeys different prin-
ciples in different domains, regardless of whether that is a result of
their being different sui generis truth properties in different domains,
or a single truth property that behaves differently in different domains,
or anything in between (e.g. the various accounts of domain-specific
alethic pluralism inspired by Wright (1992)). In short, the argument
is a logical one, and is independent of much of the metaphysical theo-
rizing that is a requisite of spelling out this-or-that particular brand of
alethic pluralism in detail.
368 R. T. Cook
2 Three Theories
Our argument for the independence of domain-specific alethic pluralism
and domain-specific alethic pluralism (i.e. for a plurality of pluralisms)
will proceed by constructing a philosophical model—one consisting of
three distinct domains with differing logics and differing accounts of
truth—and showing that one substructure of the model satisfies domain-
specific alethic pluralism but not domain-specific logical pluralism, while
another substructure of the model satisfies domain-specific logical plural-
ism but not domain-specific alethic pluralism. In order to construct such
a model, we need to consider three ingredients:
1 . The domains in question

2. The logics that govern each domain
3. The accounts of truth that govern each domain
Our model will require three distinct domains, which we shall simply
call D1, D2, and D3. Once we have carried out a bit more of the formal
work, there will be much more to say about what sorts of domain D1, D2,
and D2 are.
Now, for each domain, we need to decide on a logic that governs that
domain, and on a theory of truth that governs that domain. First, for
concreteness, let’s assume that we are working in some standard formal
language LPA+T(x) for arithmetic that contains6:
∀, ∃, ¬, ∧, ∨, →
as primitive logical vocabulary (plus the usual countable infinity of vari-

ables, etc.), plus:
0, s ( x ) , +,×

as non-logical arithmetic vocabulary, plus a (non-logical) truth predicate

T(x).7 We will use LPA to refer to the sublanguage containing all and only
those formulas of LPA+T(x) not containing the truth predicate T(x).
The non-truth-theoretic axioms are the standard axioms of Peano

Arithmetic PA with induction restricted to formulas not containing T(x),
and we assume that all three domains satisfy the axioms of PA (so
understood).
Second, we need to settle on some logics. Here we will only need to
consider two familiar logics: First-order classical logic (C) and first-order
intuitionistic logic (H). We will assume that the logic that applies to both
D1 and D2 is H, and the logic that applies to D3 is C.
Finally, we need two theories of truth. We will base these two theories
on the familiar compositional axioms for truth8:
[A1] (∀x)(AtomPA(x)→ (T(x) ↔ T0(x)))

[A2] (∀x)(SentPA(x)→ (T(‘ ℜx’)
 ↔ ¬T(x)))
[A3] (∀x)(∀y)(SentPA(x)⋀Sent(y)→ (T(‘x ∧ y’) ↔ (T(x)⋀T(y))))
[A4] (∀x)(∀y)(SentPA(x)⋀Sent(y)→ (T(‘x ∨ y’) ↔ (T(x)∨T(y))))
[A5] (∀x)(∀y)(SentPA(x)⋀Sent(y)→(T(‘x →  y’) ↔ (T(x)→T(y))))
[A6] (∀x)(FormPA(x)→T(‘( ∀x 
)x’) ↔ (∀x)T(sub(x, v)))
[A7] (∀x)(FormPA(x)→T(‘( ∃x
)x’) ↔ (∃x)T(sub(x, v)))
Following standard usage, we will call this theory CT↾.9 This is our
first theory of truth. Our second theory of truth, which we shall call
CT↾+, is simply CT↾ supplemented with the Arithmetic Law of Bivalence:

(
Biv PA : ( ∀x ) Sent PA ( x ) → (T ( x ) ∨ T ( ¬
 x )) )
We will assume that CT↾ is the theory of truth that applied to D1, and
CT↾+ is the theory of truth that applies to D2 and D3. We can sum all of
this up in the following table:
Logic Truth theory
D1 H CT↾
D2 H CT↾+
D3 C CT↾+
Before moving on, it is worth making a technical point. The theory of

D2 and the theory of D3 are not identical. In the context of CT↾ with
370 R. T. Cook
intuitionistic logic H, BivPA entails that excluded middle holds for each
sentence in LPA, but it does not entail that excluded middle holds for all
sentences in LPA+T(x). Another way of putting this is as follows: We do not
have:
For all Φ ∈ LPA+T(x):
CT + H Φ iff CT + C Φ

although we do have:
For all Φ ∈ LPA:
CT + H Φ iff CT + C Φ

(The latter result can be proven by a simple, albeit tedious induction

which is left to the reader.) Thus, the theories of each domain are
distinct.
Now, what does all of this have to do with whether domain-specific
alethic pluralism entails domain-specific logical pluralism (or vice versa)?
The answer is simple:
• If the only legitimate domains are D1 and D2, then domain-specific

logical pluralism is false, since there is only one logic (H), yet domain-
specific alethic pluralism is true, since there are two domain-specific
theories of truth: CT↾ (for D1) and CT↾+ (for D2).
• If the only legitimate domains are D2 and D3, then domain-specific
alethic pluralism is false, since there is only one theory of truth (CT↾+),
yet domain-specific logical pluralism is true, since there are two
domain-specific logics: H (for D1) and C (for D2).
Thus, domain-specific alethic pluralism does not entail domain-

specific logical pluralism (since the first observation provide a counterex-
ample to the validity of the inference), and domain-specific logical
pluralism does not entail domain-specific alethic pluralism (since the sec-
ond observation provides a counterexample). In short, we have shown
that with respect to logic and truth, we have a plurality of pluralisms.
Before moving on, a methodological clarification is in order. In the

construction given above, we have equated a theory of truth, in the rele-
vant sense, with a formal theory codified in the language LPA+T(x). As a
result, one might worry that the argument just given has missed the
mark, since most (if not all) extant defenses of alethic pluralism in the
literature treat truth as a substantive concept and a theory of truth as
more than merely a formal theory (e.g. Wright (1992), Lynch (2009)).10
This is of course right, but it is not a problem, since it is also the case that
philosophically substantial accounts of the nature of truth “bring with
them” formal theories that codify (the formal portions of ) that account,
and sufficiently different such informal, philosophical accounts of truth
will typically be accompanied by correspondingly different formal theo-
ries (such as CT↾ and CT↾+).
In particular, although Lynch (2009)—our main target here—does
not specify the formal theories of truth that correspond to his informal,
substantial accounts in any detail, it is clear that the two types of truth
theory on which he focuses the majority of his attention—accounts
where truth is epistemically constrained and accounts where it is not—
will correspond to formal theories that differ from one another in exactly
the manner in which CT↾ and CT↾+ differ (i.e. in terms of whether they
include some version of bivalence as a truth-theoretic principle). Thus,
although the argument given above only applies to the accounts of truth
developed by Lynch (and Wright) indirectly, via the link between infor-
mal philosophical accounts of truth and the formal theories correspond-
ing to such informal accounts, it nevertheless does apply.11
3 What Went Wrong?

The observations of the previous section are enough to see that there is no
direct connection, in either direction, between the two sorts of pluralism
at issue—at least in a formal sense. But it is worth examining each of
these three domains in a bit more detail to understand the independence
just established a bit more deeply. Fleshing out the philosophical aspects
of typical formulations of domain-specific pluralism (of both sorts) will
help us in this task.
372 R. T. Cook
Discussions of domain-specific alethic pluralism—and, in particular,

the discussions of the connection between domain-specific alethic plural-
ism and domain-specific logical pluralism found in Lynch (2008, 2009)
and Pedersen (2014)—typically focus on two kinds of truth: correspon-
dence and superwarrant. I will assume that correspondence is familiar to
the reader (or, at least, familiar enough for our purposes here), but super-
warrant deserves a little bit more attention. A proposition being super-
warranted is defined by Lynch as follows:
The proposition that p is superwarranted just when it is warranted without

defeat at some stage of inquiry and would remain so at every successive
stage of inquiry. A stage of inquiry is a state of information: it is always
open to extension, and potentially incomplete: At any particular stage of
inquiry, we may have no warrant for a proposition and no warrant for its
negation. A belief is warranted without defeat at a stage of inquiry as long as
any defeater for the belief at a given stage is itself undermined by evidence
available at a later stage. (Lynch 2008, 124–125)
We need not delve into these details too deeply, however, since for our
purposes the crucial consequence of truth being superwarrant in a par-
ticular domain is that truth is epistemically constrained in that domain—
that is, the following principle of epistemic constraint for truth holds for all
Φ in the language appropriate for that domain:
ECT: If T(‘Φ’) then it is feasible to have a warrant for Φ. (see Lynch (2008,
134))
In domains where truth is correspondence, however, epistemic constraint

might fail to hold (but in another sense it might not—more on this
anon!). Both Lynch (2008) and Pedersen (2014) argue that alethic plural-
ism entails logical pluralism (at least, if our alethic pluralism contains at
least one domain where truth is correspondence and at least one domain
where truth is superwarrant, see Lynch (2008, 133–135)) along the fol-
lowing lines12:
1. There is a domain D1 where truth is superwarrant, hence epistemically

constrained.
2. There is a domain D3 where truth is correspondence, hence not epis-

temically constrained.
3. Not every instance of excluded middle is true in D1.
4. Every instance of excluded middle is true in D3.
(C) Hence, the logic of D1 is distinct from the logic of D3.
Let’s call this the epistemic constraint argument. Before moving on to

the real problem with the epistemic constraint argument, it is worth not-
ing two smaller worries.
First, epistemic constraint holding of a domain does not entail that
excluded middle necessarily fails in that domain. After all, traditional
intuitionists are logical monists—they do not think that excluded middle
is logically true on any domain, even though there are many domains
that are intuitionistically provably decidable and hence all instances of
excluded middle are true (and, for mathematical domains, necessarily
true) in that domain. Hence, there is some reason to doubt the quick
move from (1) to (3).
Along similar lines, the failure of excluded middle does not immedi-
ately entail that truth is not correspondence. Neil Tennant, for example,
has developed an understanding of intuitionistic logic that seems com-
patible with the correspondence theory of truth. In particular, he argues
that:
We are maintaining that these instances of [excluded middle] are better

thought of as being held to be (necessarily) true on a priori rather than on
‘logical’ grounds. Indeed, the holding true (as a matter of necessity) of
every such instance [of excluded middle]
… expresses an essentially metaphysical belief. This belief is to the effect
that the world is determinate in every expressible regard. (Tennant 1996,
213)
If this is right, then truth could be correspondence, but the world might
not be determinate enough (in the relevant sense) to guarantee that, for
each sentence Φ, either Φ corresponds to the way the world is or Φ fails
to correspond to the way the world is. As a result, the inference from
374 R. T. Cook
truth being correspondence to all instances of excluded middle being

true—that is, the inference from line (2) to line (4)—is also
questionable.
These worries being noted, let’s set them aside and, for the sake of
argument, grant that:
• If truth is correspondence in domain D, then all instances of excluded

middle hold for sentences about D.
• If truth is superwarrant in domain D, then some instances of excluded
middle fail to hold for sentences about D.
The key to seeing how these observations are compatible with the
independence of domain-specific alethic pluralism and domain-specific
logical pluralism lies in how, exactly, we understand “hold”.
As the subscripts I use in the above reconstruction of the Lynch/
Pedersen argument indicate, the epistemic constraint argument presup-
poses that, with respect to correspondence and superwarrant, there are
only two types of domain possible:
• Those like D1, where the logic is H and (at least some) instances of
excluded middle fail to hold.
• Those like D3, where the logic is C and (all) instances of excluded
middle hold.
In short, the epistemic constraint argument assumes that we cannot

have a domain like D2, where the logic is intuitionistic but all instances
of excluded middle nevertheless hold (for sentences about D2—i.e. those
sentences in LPA—although not necessarily for all sentences about the
semantics of the language describing D2, i.e. not necessarily all sentences
in LPA+T(x)). But why could there not be such a domain? What prevents
there being a domain where instances of excluded middle are not logical
truths, but are nevertheless guaranteed to be true in virtue of principles
in our theory of truth?
Michael Lynch attempts to rule out something very much like domains
like D2. It is worth quoting the passage in question at some length:
… why not say that [intuitionistic logic] is not only the default logic, it is
the only logic, and that domains whose logic appears classical only do so
because we are employing additional principles such as Bivalence which
aren’t part of the one true logic.
Alethic pluralism … is certainly consistent with this suggestion. But the
suggestion comes at a price. According to the suggestion, [intuitionistic
logic] holds in all domains. Bivalence is not recognized as a logical princi-
ple by [intuitionistic logic]. Therefore, in every domain, Bivalence is not
recognized as a logical principle. Therefore in domains which, according to
this suggestion, nonetheless appear classical – and therefore abide by biva-
lence – Bivalence must be true for some non-logical reason. And one might
wonder what that reason might be.
A full assessment of this suggestion, therefore, requires drawing the
boundaries of logic, an issue well beyond the scope of the current essay.
But one small point is worth making: it seems natural that if one domain
allows some inferences as valid and another does not, they have different
logics. And domains where Bivalence holds will allow some inferences as
valid that other domains (which don’t sanction Bivalence) will not. So
the natural thought is that they have different logics. (Lynch 2009,
103–104)
There is a lot that needs unpacking and clarifying here, and in particular,
we are going to do some of the boundary-drawing with regard to logic
that Lynch places outside the scope of his own project. As we shall see,
this will allow us to easily identify the reason that Bivalence (or some-
thing like it) might be true of a domain without being a logical truth
(with respect to that domain).
Lynch’s argument against the truth-but-not-logical-truth of Bivalence
can be reconstructed as follows13:
1 . Bivalence is a logical principle.

2. If a logical principle is true in domain D, then it is valid in D.
(C) Hence, the logic of any domain where Bivalence is true is C.
It turns out that both premises in this argument fail, however.

376 R. T. Cook
First, Bivalence is not a logical principle in the relevant sense. The dif-
ference between classical logic and intuitionistic logic is not the validity
of Bivalence, it is the validity of (all instances of ) excluded middle.
Bivalence is a metalogical, semantic principle—one that utilizes the truth
predicate—and hence Bivalence is not a logical truth (not even for the
classical logician). I have argued at length for a similar claim about the
Tarskian T-schema in Cook (2012)—the same argument applies here. To
put this point bluntly, a logical principle is either (i) a sentence that con-
tains only logical vocabulary, or (ii) a schema that contains only logical
vocabulary and metavariables.14 The truth predicate T(x) is not a piece of
logical vocabulary (nor, obviously, a metavariable)—hence Bivalence is
not a logical principle (again, see Cook (2012) for more discussion).
But perhaps the first premise in our reconstruction of Lynch’s argu-
ment, more carefully formulated, is that Bivalence entails a logical prin-
ciple—namely, the logical schema known as excluded middle—and that
if Bivalence holds of a domain D, then excluded middle is true in D.15
Hence, we could reformulate the argument as:
1 . Bivalence entails the truth of a logical principle: excluded middle.

2. If a logical principle is true in domain D, then it is valid in D.
(C) Hence, the logic of any domain where Bivalence is true is C.
This still won’t do, however. The reason is simple: Even if we reformu-
late (1) in this way, premise (2) is still false.
We should remember that what is at issue here is logical validity—not
merely truth (or even necessary truth). Tarski provided us with a useful
first approximation of what we mean by logical consequence (and hence
of logical validity as the special empty-premise case of logical
consequence):
Consider any class ∆ of sentences and a sentence Φ which follows from the
sentences of this class. From an intuitive standpoint it can never happen
that both the class ∆ consists only of true sentences and the sentence Φ is
false. Moreover, since we are concerned here with the concept of logical,
i.e. formal, consequence, and thus with a relation which is to be uniquely
determined by the form of the sentences between which it holds … the
consequence relation cannot be affected by replacing the designations of

the objects referred to in these sentences by the designations of any other
objects. (Tarski 1936, 414–415)
According to Tarski, logical consequence is a relation that holds between

statements (and collections of statements) of our actual language and
involves two crucial ideas:
A (natural language) statement Φ is a logical consequence of a set of (natu-

ral language) statements ∆ if and only if:
• Necessity: The simultaneous truth of every member of ∆ guarantees the
truth of Φ.
• Formality: This guarantee follows solely from the logical form of Φ and
of the members of ∆.
An immediate corollary is that a principle Φ (e.g. excluded middle) is a

logical truth if and only if (i) Φ is necessarily true, and (ii) its (necessary)
truth is somehow guaranteed by its logical form.
But there are plenty of cases where logical principles are true—even
necessarily true—in some domain without being logically true. We have
already mentioned one such case: All instances of excluded middle are
true in intuitionistic domains that are decidable, but this in no way
entails that the logically monist intuitionist accepts that excluded middle
is logically true in such domains.
Another example is provided by anyone who accepts a platonist
account of standard mathematics and also accepts that the linguistic
resources of standard second-order logic are, properly speaking, logical
resources (not everyone accepts this, of course, e.g. Quine (1970)). I
myself am such a person (as are many others—e.g. Shapiro (1991, Shapiro
1997)). Any such view will entail the truth—in fact, the necessary
truth—of the following logical principle:

( ∃f ) ( ∀x ) ( ∃y ) ( f ( x ) = f ( y ) → x = y ) ∧ ( ∃x ) ( ∀y ) ( ¬f ( y ) = x )
expressing the fact that the mathematical universe is Dedekind infinite.
But the vast majority of people holding such views will reject that this
claim is logically true.16
378 R. T. Cook
Now, why does the intuitionist deny the logical truth of excluded mid-
dle even in cases when the domain in question is decidable? Why does the
(second-order) platonist deny the logical truth of the claim that the uni-
verse is Dedekind infinite? Tarski has already given us the answer: In each
case, the principle in question is true, and even necessarily true, but the
principle is not guaranteed to be true merely in virtue of its form. On the
contrary, in the first case we need to carry out a proof of the decidability
of the theory of the domain in question in order to see that all instances
of excluded middle are true, and in the latter case we need to argue that
platonism is, in fact, true (and, in addition, that at least one true mathe-
matical theory has an infinite domain).
Hence, anyone who wants to accept something along the lines of
domain D2 needs to only argue that Bivalence is true of the domain in
question, but is not true (on that domain) in virtue of its logical form.
Perhaps it is true because we have a (non-logical) warrant for the domain
being decidable, or perhaps it is true because we have some sort of meta-
physical guarantee that the domain in question is determinate in certain
respects (along the lines of (Tennant 1996)). Or perhaps we have some
other sort of non-logical guarantee that Bivalence, and hence all instances
of excluded middle, is true of the domain. None of these would in any
way imply that excluded middle is a logical truth, and none of these are
particularly implausible. Hence, the epistemic constraint argument fails,
and we do indeed have a plurality of pluralisms.
4 Pluralism and Epistemic Constraint

In the previous two sections, we formulated an argument that domain-
specific alethic pluralism and domain-specific logical pluralism are inde-
pendent, and critiqued an argument due to Lynch and Pedersen for the
opposite conclusion. If all we cared about was whether the two kinds of
pluralism were independent, then we could stop here. But presumably
the reason we are interested in the connections between domain-specific
alethic pluralism and domain-specific logical pluralism is that we find
one or the other of these views interesting.17 And the three toy domains
constructed in Sect. 2 and used in the argument for independence (and
the debunking of the epistemic constraint argument) also illustrate some

interesting, and as far as I can tell novel, observations about the role that
epistemic constraint can and should play in accounts of domain-specific
pluralism of either sort.
As we have already noted, the alethic pluralist believes that truth comes
in at least two varieties—correspondence and superwarrant—and argues
that epistemic constraint for truth:
ECT: If T(‘Φ’) then it is feasible to have a warrant for Φ.
holds of domains where truth is superwarrant, but not necessarily for

domains where truth is correspondence (and we agreed, for the sake of
argument and Tennant (1996) notwithstanding, to accept that epistemic
constraint does not hold in domains where truth is correspondence).
Much of the point of the previous section was to carefully drive a
wedge between logical principles that are merely true, or even necessarily
true, and those that are logically true (i.e. necessarily true in virtue of
logical form). Given this, it makes sense to ask whether logical truth is
similarly epistemically constrained in particular discourses. In other
words, we can also ask whether the following principle—we shall call
epistemic constraint for logical truth—holds in a particular discourse:
ECLT: If Φ is a logical truth, then it is feasible to have a warrant for Φ.
The first thing to notice is that ECT entails ECLT: If Φ is a logical truth,
then it is a truth, and so by ECT it is feasible to have a warrant for Φ. But
there seems to be no reason to think that ECLT entails ECT.
Let’s assume that, if ECT holds, then the right logic for the domain in
question is H. The reason is simple and follows the (correct core of ) rea-
soning found in Lynch (2008, 2009) and Pedersen (2014): If logical
truth is epistemically constrained in a domain D, then it would seem that
there is no guarantee that we can know, of every sentence Φ about D, on
purely logical grounds, that either Φ is true or ¬Φ is true. But, of course,
as discussed in the previous section, there might be other non-logical
grounds on the basis of which we can conclude that Bivalence holds, and
hence (if our theory of truth is powerful enough otherwise, as CT↾+ is)
380 R. T. Cook
that all instances of excluded middle (for sentences about the domain, at
least—although not necessarily for sentences involving the truth predi-
cate T(x)) are true.
This leaves three remaining combinations, corresponding to each of
the domains constructed in Sect. 218:
1. Both ECLT and ECT hold of D. Then the logic of D is H, and the
theory of truth should not include BivLD. Hence, the domain is (in
relevant respects) like D1.
2. ECLT holds but ECT does not hold of D. Then the logic of D is H, and
the theory of truth should include BivLD. Hence, the domain is (in
relevant respects) like D2.
3. Neither ECLT nor ECT holds of D. Then the logic of D is C, and the
theory of truth should include BivLD. Hence, the domain is (in rele-
vant respects) like D3.
Thus, once we understand that logical truth being epistemically con-

strained does not imply that truth is epistemically constrained (although
the converse implication does hold), we can see that the notion of epis-
temic constraint, generally speaking, divides up domains into three sorts,
not two, and this in turn explains how domain-specific alethic pluralism
might hold while domain-specific logical pluralism might fail, or vice
versa.
5 Concluding Remarks
In the previous section, we have seen that, in principle, at least, domain-
specific alethic pluralism is independent of domain-specific logical plu-
ralism, and we also noted that the one simple way to understand this
phenomenon is by noting that logical truth being epistemically con-
strained does not entail that truth is likewise epistemically constrained.
In short, with respect to the domain-specific versions of the two types of
pluralism at issue, we do have a pluralism about pluralisms. We’ll con-
clude the chapter by examining whether similar stories might be told
about various other combinations of domain-specific pluralism, domain-

independent pluralism, and monism. Since we can choose any of these
three options for each of logic and truth, this gives us nine distinct pos-
sible combinations:
DS Alethic Plur. DI Alethic Plur. Alethic Mon.
DS Logical Plur. 1 2 3
DI Logical Plur. 4 5 6
Logical Mon. 7 8 9
We have already shown that domain-specific logical pluralism is com-

patible with alethic monism (option 3), and domain-specific alethic plu-
ralism is compatible with logical monism (option 7). In addition,
although the arguments of Lynch (2008, 2009) and Pedersen (2014) fail
to show that domain-specific alethic pluralism covaries with domain-
specific logical pluralism, the discussion found there makes a strong case
for their compatibility. More simply, we can just consider a view that
accepts all of D1, D2, and D3 (or relevantly similar domains) as legitimate
(option 1). And no one doubts that the combination of alethic monism
and logical monism (option 9) is coherent, even if many readers of this
chapter (and this volume) will no doubt think such a view is likely incor-
rect. In short, we’ve covered the corners of this chart, but have addressed
none of the options in the “+”-shaped region remaining.
It appears, however, that if we restrict our attention to the two kinds of
truth (epistemically constrained or not) and the two kinds of logic (epis-
temically constrained or not) that have concerned us thus far, then only
two of the remaining options are coherent.
If one is a domain-independent logical pluralist, then this implies that
there are at least two logics that apply everywhere. Assume that one is
epistemically constrained and the other not. Then this means that,
according to at least one of these everywhere applicable logics, logical
truth is not epistemically constrained (everywhere). But this means truth
must (everywhere) fail to be epistemically constrained (since truth being
epistemically constrained implies logical truth being epistemically
constrained). And if truth must fail to be epistemically constrained, we
must be alethic monists. This rules out options (4) and (5), but leaves (6)
intact.
382 R. T. Cook
Along similar lines, if one is a domain-independent alethic pluralist,

then there are two accounts of truth that apply everywhere. Assume
one is epistemically constrained and the other not. Then, according to
at least one of these everywhere applicable accounts of truth, truth is
epistemically constrained (everywhere). But this means that logic must
(everywhere) be epistemically constrained (since, again, truth being
epistemically constrained implies logical truth being epistemically
constrained). And if logic must be epistemically constrained, we must
be logical monists. This rules out options (2) and (5) (again), but
leaves (8) intact.
I do not, in fact, think these brief arguments actually show that
domain-independent logical pluralism entails truth monism, or that
domain-independent alethic pluralism entails logical monism.19 Rather,
what I think that these brief, somewhat un-careful arguments show is
that, if we are to develop a view that combines a domain-specific plural-
ism of one sort with a domain-independent pluralism of the other sort
(i.e. option 2 or 4), or one that combines domain-independent pluralism
of both sorts (i.e. option 5), then defending that combination as a coher-
ent position on the 3 × 3 grid of possible positions above will require
additional theoretical tools not mobilized here. For example, if we are
domain-independent alethic pluralists, so truth both is epistemically con-
strained and is not epistemically constrained (depending on which notion
we use), perhaps there is a way to keep the former notion from “bleeding”
into logical truth, entailing that logical truth must be epistemically con-
strained everywhere (for any legitimate logic).
I am optimistic that such theoretical tools can be found, and that
accounts of logic, truth, and their interrelations corresponding to options
(2), (4), and (5) can (and likely will be) carefully formulated and defended
as coherent. But for the present I will be content to have shown that,
contrary to the suggestion found in Lynch (2008, 2009) and Pedersen
(2014), options (3) and (7) are coherent, leaving examination of the
other options for another day.20
Notes
1. Further, other sorts of pluralisms—such as metaphysical pluralisms—
also fall on one or the other side of the domain-specific/domain-inde-
pendent divide.
2. See Edwards (2012) and Pedersen and Wright (2013a, b) for good sur-
veys of alethic pluralism and Cook (2010) and Russell (2016) for good
surveys of logical pluralism.
3. On the contrary, I am rather unsympathetic to domain-specific logical
pluralism, and rather sympathetic to domain-independent logical plu-
ralism, despite the good efforts of (Shapiro 2015). And I am wholly
unsympathetic to either sort of alethic pluralism. For some of that story,
see (Cook 2014).
4. It is important to note that I am not arguing that we have a plurality of
true pluralisms (see the previous footnote). The point is that the accounts
in question—domain-specific alethic pluralism and domain-specific
logical pluralism—are distinct positions, regardless of whether they are
correct or not.
5. I am, in what follows, going to concentrate on the argument as given in
Lynch (2008, 2009). Pedersen (2014) characterizes itself as presenting a
slightly more detailed and cleaned up version of Lynch’s argument (and
extending the conclusions to metaphysical pluralism, which is interest-
ing but orthogonal to our concerns here). I would like to note that
Pedersen’s essay was extremely helpful in sorting out how, exactly, the
argument in question is supposed to work.
6. We need both quantifiers and all four propositional connectives in
LPA+T(x) if we are to allow the logic in question to be full intuitionistic
logic, since the familiar equivalences that allow for various (classical)
definitions of one operator in terms of another fail intuitionistically.
7. For an argument that the truth predicate is non-logical, see (Cook 2012).
We will return to discuss the relevance of this fact in detail below.
8. A brief explanation of the notation in these axioms:
• T0(x) is the arithmetically definable truth predicate for ∅0 sentences

0
• AtomPA(x) is the (purely arithmetic) predicate expressing “x is the code

of an atomic sentence of PA”.
• SentPA(x) is the (purely arithmetic) predicate expressing “x is the code
of a sentence of PA”.
384 R. T. Cook
• FormPA(x) is the (purely arithmetic) predicate expressing “x is the code

of a formula of PA with exactly one free variable”.
• ℜ is the recursive function that maps the code of a formula to the code
of its negation (and similarly for ∧ , ∨ , and →
 ).
• ∀v is the recursive function that maps the code of a formula with one

free variable to the code of the universal quantification of that formula

(and similarly for ∃v ).

• sub(x, y) is the recursive function that maps a number v and the

code of a formula with one free variable to the code of the formula
obtained by replacing the free variable with the numeral for that
number.
Note that our presentation of CT↾ contains more axioms than in

many standard presentations, since we are not treating the conditional as
defined in terms of negation and disjunction.
9. Nothing hinges on which particular theory of truth we chose, so long as
combining that theory with Heyting Arithmetic HA (i.e. the axioms of
PA plus the logic H) and combining that theory with classical Peano
Arithmetic (i.e. the axioms of PA plus classical logic C) give extension-
ally distinct theories. The simplicity (and conservativeness) of CT↾
makes it a convenient choice, however.
10. Thanks are owed to an anonymous referee for pressing this point.
11. Further, the discussion of the next two sections serves to provide an
explanation of the formal construction just given in terms of philosophi-
cal notions mobilized in Lynch’s account—in particular, in terms of the
role played by epistemic constraint.
12. We shall see the reason for the odd numbering of the domains shortly.
13. This reconstruction involves a simplification—one harmless in the pres-
ent context—namely that whatever logics we are taking to be candidate
logics for our domains, if we supplement any of those logics with the law
of excluded middle, we obtain C. This is, of course, not true for many
logics (i.e. substructural logics), but it is true of H.
14. Of course, non-logical vocabulary can be substituted for the metavari-
ables in instances of the latter sort.
15. In fact, this won’t work simply because the version of Bivalence we added
to CT↾—that is, BivPA—does not entail excluded middle for all sen-
tences in LPA+T(x). Let’s set aside this technical observation, however, since
we have more important, more philosophical fish to fry.
16. A slightly snarkier way of making the point: If the truth, or even the
necessary truth, of a logical principle entailed its logical truth, and if
second-order logic is logic, then any platonist would presumably also be
some sort of logicist.
17. I certainly find them interesting, even if, as I noted in endnote 3, I ulti-
mately think neither of them is correct.
18. Here LD is the language about domain D—that is, all sentences about D
not including the truth predicate T(x), hence BivLD is Bivalence restricted
to the sentences in LD.
19. This, despite the fact that I think that the combination of domain-
independent logical pluralism and alethic monism (i.e. option 6) is the
right option. The point is that I don’t think the other options are incoher-
ent—I just think that they are wrong.
20. Thanks are owed to Nikolaj J. L. L. Pedersen, Nathan Kellen, and an
anonymous referee for helpful comments on early versions of this
material.
References
———. 2001. Defending Logical Pluralism. In Logical Consequence: Rival
Approaches. Proceedings of the 1999 Conference of the Society of Exact Philosophy,
ed. J. Woods and B. Brown, 1–22. Stanmore: Hermes.
Cook, R. 2010. Let a Thousand Flowers Bloom: A Tour of Logical Pluralism.
———. 2012. The T-schema Is Not a Logical Truth. Analysis 72 (2): 231–239.
———. 2014. Should Anti-realists Be Anti-Realists About Anti-realism?
Edwards, D. 2012. Pluralist Theories of Truth. In Internet Encyclopedia of
Philosophy, ed. J. Fieser and B. Dowden. http://www.iep.utm.edu/plur-tru/
Lynch, M. 2008. Alethic Pluralism, Logical Consequence, and the Universality
of Reason. Midwest Studies in Philosophy 32: 122–140.
Pedersen, Nikolaj J.L.L. 2014. Pluralism × 3: Truth, Logic, Metaphysics.
386 R. T. Cook
Pedersen, Nikolaj J.L.L., and C.D. Wright. 2013a. Pluralist Theories of Truth.

In The Stanford Encyclopedia of Philosophy, ed. E.N. Zalta. https://plato.stan-
ford.edu/entries/truth-pluralist/.
———. 2013b. Truth and Pluralism: Current Debates. Oxford: Oxford
University Press.
Quine, W. 1970. Philosophy of Logic. Cambridge, MA: Harvard University Press.
Russell, G. 2016. Logical Pluralism. In The Stanford Encyclopedia of Philosophy
(Winter 2016 edition), ed. E.N. Zalta. https://plato.stanford.edu/archives/
win2016/entries/logical-pluralism/.
Shapiro, S. 1991. Foundations Without Foundationalism: The Case for Second-
Order Logic. Oxford: Oxford University Press.
———. 1997. Philosophy of Mathematics: Structure and Ontology. Oxford:
———. 2015. Varieties of Logic. Oxford: Oxford University Press.
Sher, G. 2013. Forms of Correspondence: The Intricate Route from Thought to
Reality. In Truth and Pluralism: Current Debates, ed. Nikolaj J.L.L. Pedersen
and C.D. Wright, 157–179. Oxford: Oxford University Press.
Tarski, A. 1936. On the Concept of Logical Consequence. In Logic, Semantics,
Metamathematics, ed. A. Tarski, 2nd ed., 409–420. Indianapolis: Hackett.
———. 1983. Logic, Semantics, Metamathematics. Ed. J. Corcoran, 2nd ed.
Tennant, N. 1996. The Law of Excluded Middle Is Synthetic A Priori, If Valid.
Philosophical Topics 24 (1): 205–229.
Woods, J., and B. Brown, eds. 2001. Logical Consequence: Rival Approaches.
Proceedings of the 1999 Conference of the Society of Exact Philosophy, Hermes
Stanmore.
Press.
A Plea for Immodesty: Alethic Pluralism,
Logical Pluralism, and Mixed Inferences
Chase B. Wrenn
1 Introduction
There is a natural, if inchoate, intuition that truth might have a variable
nature, depending on what the subject matter is. Given the great differ-
ences between natural science, mathematics, ethics, and aesthetics, does
it not stand to reason that ‘Electrons have negative charge,’ ‘1 + 1 = 2,’
‘Torture is wrong,’ and ‘Duck Soup is funny’ are all true, but the truth of
each is a different property from the truth of the others?
The arch-deflationist W. V. Quine had the intuition,1 but it was Crispin
Wright (1992) who brought the prospects of pluralism about truth into
contemporary discussion. These days, a variety of different ways of work-
ing out the details of a pluralist conception of truth are on offer. By
alethic pluralism (AP), I mean the view that there are different, but sub-
stantial, truth-properties in different discourses.
C. B. Wrenn (*)
Department of Philosophy, The University of Alabama, Tuscaloosa, AL, USA
e-mail: cwrenn@ua.edu

https://doi.org/10.1007/978-3-319-98346-2_16
388 C. B. Wrenn
There is also a natural, if inchoate, intuition that logic is variable,

depending on what you are talking about. As alternative logics have
proliferated, often motivated by the concerns of specific domains of
inquiry, one might guess that the logic appropriate for some inquiries just
isn’t the same as the logic appropriate for others.
The distinctive feature of a logic is the relation of logical consequence
it describes. Equivalently, logics can be distinguished in terms of which
inferences qualify as valid in them—provided we understand ‘inferences’
not as psychological events but as sets of premises and conclusions. By
logical pluralism, I mean the view that there are multiple, legitimate rela-
tions of logical consequence. Domain-specific logical pluralism (DLP) is
the view that different domains of inquiry are governed by different rela-
tions of logical consequence.
Both AP and DLP face the problem of mixed inferences. This problem
arises when we try to explain the validity of inferences whose premises
and conclusion are not uniformly drawn from the same domain of
inquiry. Some versions of alethic pluralism, such as Michael Lynch’s man-
ifestation functionalism, are specifically designed with this problem in
mind. However, there is a natural route from AP to DLP and, as Lynch
is aware, the addition of DLP to AP threatens to reintroduce the problem
of mixed inferences, even for manifestation functionalism. Nevertheless,
Lynch contends, the threat can be avoided through modesty: when there
are multiple candidate logics on the table, we should be cautious and
employ only the weakest one of them or, if none is weakest, we should
endorse as valid only those inferences validated by all candidates (Lynch
2008, 2009, Ch. 5).
The modest approach, I will argue, is unsatisfactory on two counts. (1)
It is bound to invalidate certain intuitively valid inferences. (2) It threat-
ens to collapse into a view I call virtual logical pluralism, which isn’t really
a form of logical pluralism at all. On the other hand, there appears to be
no good alternative to the modest approach for the advocate of AP +
DLP. If that is so, then it seems alethic pluralists should steer clear of
domain-specific logical pluralism—or vice versa.
A Plea for Immodesty: Alethic Pluralism, Logical Pluralism… 389
2 Alethic Pluralism
The distinctive claim of alethic pluralism is that truth is (in some sense of
‘is’) a different property for claims in different discourses or domains of
inquiry—but truth is also more substantial than deflationists claim it to
be. Pluralists differ in their understanding of the ‘is’ in their distinctive
claim. Some take it to be a matter of identity: the truth predicate desig-
nates different properties when applied to claims from different discourses
(Wright 1992, 2013). Some take it to be a matter of realization or mani-
festation: the property playing the role of truth in one discourse need not
be the same property as what plays the truth-role in another (Lynch 2001,
2009).
Not only can alethic pluralists differ on what it means to be truth for
a discourse, but they can differ on which properties are truth in which
discourses. One pluralist might think correspondence is truth in mathe-
matics, while another thinks mathematical truth is something epistemic,
for example. A main motivation for pluralism is the idea that realism
seems appropriate for some discourses (such as physical science) but not
others (such as aesthetics). Let us call a property a truth-candidate if it is
the sort of property pluralists think might be truth for some discourse or
other. Two prominent truth-candidates are robust correspondence and
superwarrant. A claim robustly corresponds to the world when it repre-
sents the world as being a certain way, the world is that way, and the
world’s being that way is independent of our actual or possible possession
of evidence that that is how it is. (Henceforth, I will omit ‘robust’ and
refer to this as ‘correspondence.’) The notion of superwarrant comes from
Wright’s notion of superassertability (1992). A claim is superwarranted
when there is a stage of inquiry in which it is warranted without defeat
and it would remain that way in all successive stages of inquiry (Lynch
2009, p. 38; see also Pedersen 2014).
Plausibly, something like correspondence is truth in an area such as
physics, chemistry, or everyday discourse about middle-sized dry goods
and their determinate properties. Superwarrant is a plausible contender
in areas where truth seems evidentially constrained. For example, no joke
390 C. B. Wrenn
could be secretly funny—funny despite the fact that no one could know.
Truth and knowability walk together in discourse about what is funny,
and one might suppose the truth about humor must coincide with what-
ever our most considered, unimprovable judgments would say about it.
3 Mixed Inferences and Logical Pluralism

The following argument illustrates the problem of mixed inferences
(Tappolet 1997):
Mix
( 1) The one about the nun at the dude ranch is funny.

(2) Either snow is white or the one about the nun at the dude ranch isn’t
funny.
(3) Therefore, snow is white.
Mix’s conclusion and its first premise come from difference discourses.
Suppose truth is something epistemic, such as superwarrant, in discourse
about what is funny, and it is something realist, such as correspondence,
in discourse about the color of snow. Then the argument appears valid,
but it also appears not to transmit truth from its premises to its conclu-
sion, given pluralism. Truth for (1) is superwarrant, but truth for (3) is
not. So, alethic pluralists have to find a way to explain the validity of such
inferences.
An especially promising approach is functionalism about truth (Lynch
2001, 2008, 2009). In Lynch’s version, a number of ‘truisms’ serve to
define the role or job description of truth. They include, among others:
Objectivity: The belief that p is true if, and only if, with respect to
the belief that p, things are as they are believed to be.
and
End of Inquiry: Other things being equal, true beliefs are a worthy goal
of inquiry.
Whatever property best satisfies the truisms for a given discourse’s

atomic claims is said to ‘manifest’ truth for those claims.2 Correspondence
might manifest truth in some discourses, while superwarrant manifests it
in others. Truth itself is the property that essentially satisfies the truisms.
In a discourse where correspondence satisfies the truisms, propositions
have the property of truth by having the property of correspondence, not
very differently from the way that objects can have the property of redness
by having the property of being some particular shade of red.3
Every true proposition has the property of truth, and the property of
truth plays the truth-role essentially. For mixed compounds such as (2),
and in mixed inferences such as Mix, there may be no property other
than truth itself that manifests truth. Premise (2) is the sort of compound
that is true whenever it has a true component. The fact that different
properties manifest truth for each of its parts doesn’t change the fact that,
if either component has a truth-manifesting property, then the com-
pound is true. The truth of the component, however it might be grounded,
grounds and determines the truth of (2). Likewise, even if different prop-
erties manifest truth for (1), (2), and (3), the inference from (1) and (2)
to (3) has the feature crucial to deductive validity: If (1) and (2) each have
the property of truth, then (3) must also have the property of truth. That
is how the functionalist tries to solve the problem of mixed inferences.
Just as alethic pluralists think there are multiple truth-properties, logi-
cal pluralists believe there are multiple relations of logical consequence.
Nevertheless, any relation of logical consequence will have these core fea-
tures (Beall and Restall 2006):
GTT: An argument is valid iff every casex in which its premises are true
is a casex in which its conclusion is true.4
NNF: The logical consequence relation is normative, necessary, and

formal.
The ‘x’ subscript in GTT indicates a parameter for different logical

systems. Each system brings with it a different conception of which ‘cases’
are relevant. Different systems count different arguments as valid and
deliver different logical consequence relations.
392 C. B. Wrenn
As Beall and Restall’s discussion illustrates, it can be difficult to state

NNF’s notions of normativity, necessity, and formality precisely. The
general ideas, though, are clear enough. Logical consequence is ‘norma-
tive’ in the sense that it is somehow wrong to accept the premises of a
valid argument while rejecting its conclusion (ibid., p. 16). It is ‘neces-
sary’ in the sense that valid reasoning tells us not only that if a set of
premises are true, then so is a conclusion, but also that if those premises
were true, then the conclusion would be true as well (ibid., pp. 15–6).
Following MacFarlane (2000), Beall and Restall identify three different,
and non-equivalent, ways in which we might plausibly clarify the idea
that logic is ‘formal’ (ibid., pp. 18–23). What will matter for my pur-
poses, though, is the idea that logic abstracts away from the particular
semantic content of thought, though of course there remains room to
debate just where to draw the line between logical form and semantic
content.
Despite the possibility of such debate, I will make some assumptions
about which features of a claim are ‘formal’ and which involve its ‘seman-
tic content.’ According to pluralists, what discourse a claim belongs to,
and so what truth-candidate is relevant to the claim, depends on what the
claim is about. For example, here is Lynch:
What makes a proposition a member of a particular domain? The obvious

answer: the subject matter it is about … That said, it would be nice to say
at least something more about what, at least in general, distinguishes prop-
ositions of one domain from another. A natural answer is this: the kind of
concepts (moral, legal, mathematical) that compose the proposition in
question. (2009, pp. 79–80)
I will assume that the question of what domain a claim belongs to is a

question about its semantic content, not its logical form.
For a relation between premises and a conclusion to be formal, it must
be independent of the actual content of those claims. Two arguments
alike in logical form will be alike in terms of whether that relation holds
between their premises and their conclusion. Both classical and intu-
itionistic entailment are formal in this way.
Logical pluralism is just the view that there is more than one genuine
relation of logical consequence. Domain-specific logical pluralism (DLP)
holds that different domains of inquiry or discourses have different log-
ics. The logical consequence relation in ethics, for example, might not be
the same as the logical consequence relation in physics. It might be intu-
itionistic entailment in the former case, classical entailment in the latter.
Such a view could be attractive to alethic pluralists on purely intuitive
grounds: if logic tells us about the preservation of truth in inference, and
different discourses have different truth-properties, why should we expect
different discourses to have the same logic?
Lynch describes a more direct path from AP to DLP, provided we
assume that truth is sometimes non-epistemic and sometimes superwar-
rant (2008, pp. 134–5; 2009, pp. 94–6; see also Pedersen 2014). In
domains where truth is non-epistemic, we might expect the classical logi-
cal laws to hold, and so logical consequence in those domains would be
classical. For example, the classical Law of Excluded Middle (LEM) holds
for those discourses. But in domains where superwarrant is truth, logical
consequence amounts to the following (given that ‘cases’ amount to pos-
sible worlds):
SWLC: P is a logical consequence of Q iff every possible world in which

Q is superwarranted is a possible world in which P is
superwarranted.
There is, however, no general guarantee that either a proposition or its

negation must be superwarranted. So, LEM does not hold for discourses
in which superwarrant is truth, and SWLC’s relation of logical conse-
quence is not the same as classical logical consequence.
Though DLP may be tempting to alethic pluralists, it also threatens to
reintroduce the problem of mixed inferences. Functionalism aimed to
explain the validity of Mix in a way that didn’t require assigning it to a
discourse. But given DLP, there are multiple relations of logical conse-
quence that vary from discourse to discourse. So, we need to know which
relation of logical consequence is relevant to evaluating a mixed inference
such as Mix. Its premises and conclusion come from different discourses,
with different, proprietary relations of logical consequence. DLP does
not provide a discourse-neutral relation of logical consequence, and so we
394 C. B. Wrenn
seem not to have a relation of logical consequence appropriate for the

mixed inference itself. And with no such relation on the table, we have no
way of accounting for the inference’s validity. What is to be done?
4 Modesty
One might be tempted to try for DLP the same functionalist move Lynch
uses to resolve AP’s problem of mixed inferences. In that case, one would
consider ‘logical consequence’ to be a higher-order relation. One proposi-
tion is a logical consequence of another just in case they bear a relation to
one another that plays the role of (or manifests) logical consequence for
them. Such a move is unhelpful in this case. We need to know what rela-
tion plays the role of logical consequence for Mix. It is hard to see how an
answer to that question could be given while respecting the formality
requirement.
There are two, related worries here. First, the functionalizing move
does not shed much light on the question of whether (3) is, in fact, a logi-
cal consequence of (1) and (2). On the functionalizing move, it is a con-
sequence if it bears the relation to them that plays the role of logical
consequence appropriate for the content of the argument. But that is
merely to avoid the real question: What logical consequence relation is
appropriate for the content of that inference?
The second worry concerns the formality of the higher-order relation
of consequence. Classical and intuitionistic consequence are formal; what
P and Q are about is irrelevant to whether P is a classical consequence of
Q and to whether it is an intuitionistic consequence of Q. However,
whether P and Q are related by a relation that plays the role of logical
consequence for them does depend on their content. If P has a form that
would make it a classical consequence of Q, but not an intuitionistic
consequence of Q, then the question whether P bears a relation to Q that
plays the role of logical consequence for them is one whose answer
depends on what P and Q are about. Two inferences could be identical in
terms of their logical forms, while one of them has a conclusion that bears
the functionalized consequence relation to its premises and the other
does not. So, the functionalized relation fails the formality condition on
being a relation of logical consequence.
Rather than functionalizing logical consequence, Lynch’s own pro-
posal is to be as modest as possible: “[W]hen reasoning across domains,
logical caution is in order: We want to limit the number of logical truths
that we endorse, so as to respect those domains which, by virtue of the
property that plays the truth-role within them, enforce less [sic] logical
laws than others” (2008, p. 138; 2009, p. 101). He codifies that idea in
the following way:
MODEST*: Where a compound proposition or inference contains propo-

sitions from distinct domains, the default governing logic is that comprised
by the intersection of the domain-specific logics in play. (Lynch 2008,
p. 139; 2009, p. 102)5
We can identify a logic with the inferences it validates. A logic is ‘in

play’ for an inference when it governs the domain of at least one of the
inference’s components (i.e., premises, conclusion, or a component of a
premise or conclusion). The intersection of the logics in play would be a
logic that validates all and only those inferences validated by all the logics
in play. A mixed inference, then, qualifies as valid if and only if all the
logics in play would count it as valid. That, we should note, is enough to
validate Mix, assuming Disjunctive Syllogism and Double Negation
Introduction are valid in all the logics in play for it.
It also avoids the formality problem, because each logic’s proprietary
relation of logical consequence can be formal, while the question of
which one is relevant to a given argument will depend on the argument’s
content. Since DLP itself already says that the governing logic of an
unmixed inference depends on its content, MODEST* doesn’t violate the
formality condition in any way that an advocate of DLP should find
troubling.
This way of tackling the problem is certainly modest and cautious.
But, I think, it is too modest and cautious. Its zeal not to validate any
invalid inferences leads it to invalidate certain intuitively valid inferences.
It also risks collapsing DLP into what I will call virtual logical pluralism.
396 C. B. Wrenn
5 MODEST* Is Too Modest

There are two kinds of valid mixed inferences that the modest approach
invalidates. For illustrative purposes, let us suppose ‘Jack went up the hill’
is governed by classical logic and correspondence truth, and let us sup-
pose that ‘1 + 1 = 2’ is from a domain governed by intuitionistic logic, in
which Double Negation Elimination (DNE) is not a valid form of
inference.
Consider this argument:
Inessential Mix
( 4) 1 + 1 = 2 and it’s not the case that Jack did not go up the hill.
(5) Therefore, Jack went up the hill.
Getting from (4) to (5) requires an application of DNE, and so the

argument is not intuitionistically valid. Since intuitionistic logic gov-
erns one of (4)’s conjuncts, MODEST* tells us the inference is valid if
and only if it is both intuitionistically and classically valid. So, we have
to count it as invalid, which seems wrong. I call this a case of ‘inessen-
tial mixing,’ since the proposition that 1 + 1 = 2 appears to play no
essential role in the inference, apart from compelling a logic too weak
to validate it.
The recipe for inessential mixing is just this: Take an inference that is
valid in a given logic L, and then add a (non-contradictory) proposition
from a domain with a logic that does not validate that inference. The
proposition can be added as a new premise, or it can be conjoined with
one of the old premises. Either way, the addition of the new proposition
requires, according to MODEST*, a weakening of the logic so that the
inference is no longer valid.
There are also cases of ‘essential mixing.’ Here is one:
Essential Mix
( 6) Either it’s not the case that Jack did not go up the hill or 1 + 1 ≠ 2.
(7) 1 + 1 = 2.
(8) Therefore, Jack went up the hill.
In this case, we again have an inference that is not intuitionistically

valid, but now the intuitionistic proposition that 1 + 1 = 2 is essential to
the argument. By MODEST*’s lights, the argument is invalid, and that
seems wrong too.
The recipe for essential mixing is also straightforward. Start with an
inference that is valid in a given logic, L1, whose components all come
from a domain governed by L1, and that has an essential premise that is
not a component of the conclusion. Uniformly replace all occurrences of
that premise in the argument (including as a component of another
claim) with a (non-contradictory) proposition from a discourse governed
by a logic, L2, that does not validate the inference. This alteration, by
MODEST*, requires the new inference to be governed by L2 rather than
L1; it makes the new inference invalid. But as Essential Mix illustrates,
such inferences can be valid after all.
Lynch’s MODEST* proposal is too modest to validate all the mixed
inferences that, intuitively, we should want validated. Indeed, it invali-
dates even Mix, on the assumption that propositions about what is funny
are governed by a logic that does not validate Disjunctive Syllogism.6
6 Virtual Logical Pluralism

A second problem for MODEST* is one Lynch himself anticipates. If we
are retreating to the weakest logic in play anyway, why not give up on
logical pluralism altogether and hold that the One True Logic is some-
thing extremely weak, such as Neil Tennant’s ‘core logic’ (1987, 2012a,
b), while “domains whose logic appears to be classical only do so because
we are employing the classical portion” of a weaker-than-classical logic
with classical logic as a proper part (Lynch 2008, p. 139)?
Lynch’s reply is that, if we made such a move, we would be adopting a
One True Logic in which classical principles such as Bivalence do not
hold. As logical principles, they are incorrect. So, in domains where
Bivalence does hold, it must hold “for some non-logical reason. And one
might wonder what that reason might be” (Lynch 2008, p. 139).
Virtual logical pluralism (VLP) is the view that there is only one rela-
tion of logical consequence, but some discourses have features that make
398 C. B. Wrenn
them appear to be governed by stronger logics, as special cases of the One

True Logic.
Here are some examples to illustrate how VLP might work:
Semantic Spandrels: The overall logic of a language containing a transpar-

ent truth predicate is paraconsistent, but such that the segment of the lan-
guage lacking the predicate behaves perfectly classically. Here, the lack of a
transparent truth predicate makes non-semantic discourse appear classical,
as a special case of the non-classical One True Logic. (Beall 2009)
Quasi-Quinean Intuitionism: Quine was no intuitionist, but suppose the

One True Logic is intuitionistic, while Bivalence and the Law of Excluded
Middle have proven indispensable to the practice of science. Then scientific
discourse can be seen as including them as axiom schemata, or seen as pre-
supposing their instances. The classical logic of science is a special case of
the intuitionistic One True Logic, justified on the grounds of its
indispensability.
For present purposes, it isn’t necessary for us to evaluate VLP’s merits

and demerits, but only to note that endorsing VLP involves giving up
DLP. MODEST* invites a retreat to VLP from DLP, in the face of the
problem of mixed inferences. Is Lynch’s reply sufficient to block that
retreat?
It appears not to be. First, as Semantic Spandrels indicates, there can
be logical reasons for the Law of Non-Contradiction to hold in some
discourses but not others, even when the One True Logic is paraconsis-
tent. In this case, the logical reason is precisely what Beall (2009) goes to
such great pains to show: truth-value gluts emerge in semantic discourse
only because that discourse includes a logical device, ‘transparent truth,’
characterized by the intersubstitutability of Tp and p in all nonopaque
contexts. The Law of Non-Contradiction holds in non-semantic dis-
courses, not for a substantive, non-logical reason, but rather for the logi-
cal reason that their languages don’t include that logical device.7 So,
Lynch’s claim that we would need non-logical reasons for classical logical
principles to hold in some discourses rather than others appears to be
incorrect.
Second, Quasi-Quinean Intuitionism reminds us that a generally

Quinean outlook toward logic is always available. If we want to explain
why classical principles hold in a given area of inquiry, the answer could
be that we have found those principles to be so fruitful in that area that
the principles have become central theoretical commitments. From the
Quinean perspective, logical principles are distinctive only in their cen-
trality to our web of belief. If the One True Logic isn’t classical, there is
no in-principle reason our web couldn’t still have pockets of classicism.
The classical principles in those areas would have the same status as cen-
tral and very-unlikely-to-be-revised structuring theories in science. A
chemist is no more likely to give up the belief that hydrogen has atomic
number 1 than to give up DNE. If DNE isn’t part of the one true logic,
but it is important to chemistry, then let it be an important part of chem-
ical theory that, when we’re talking about chemistry, we accept every
instance of~~p ⊢ p. For the Quasi-Quinean, even the principles of the
One True Logic are justified by way of their centrality and importance to
our web of belief. On this view, the distinction between ‘logical’ and
‘non-logical’ reasons for accepting an inference is too tenuous to do much
serious theoretical work. A Quasi-Quinean would thus find Lynch’s way
of blocking the retreat to VLP unsatisfying.
There is no compulsion for an advocate of DLP to retreat to VLP, but
the move is hard to resist. It is especially hard to resist if the advocate of
DLP also endorses MODEST*. In the end, it is hard to see the difference
between accepting DLP and MODEST* or, instead, just embracing
VLP. It is thus hard to see why an alethic pluralist wouldn’t just give up
on logical pluralism.
7 Chains of Inference
A striking feature of Inessential Mix and Essential Mix is that there are
chains of inferences from the premises to the conclusion that are valid,
even by the lights of MODEST*. The premises of both arguments intu-
itionistically entail:
(9) It’s not the case that Jack did not go up the hill.
400 C. B. Wrenn
And that claim, from a domain whose governing logic is classical, clas-
sically entails the conclusion of both arguments. An advocate of AP +
DLP might hope to find a solution to the problem of mixed inferences in
that fact. The result would be a modification of MODEST* in the fol-
lowing way:
MODEST**: Where a compound proposition or inference contains prop-

ositions from distinct domains, the default governing logic is that com-
prised by the intersection of the domain-specific logics in play. An
inference Γ ⊢ p is valid if and only if either (i) it is valid in its default gov-
erning logic, or (ii) there is a proposition q such that the inference Γ ⊢ q is
valid in its default governing logic and q ⊢ p is valid.8
MODEST** gives a recursive characterization of validity. In the base

case, Γ ⊢ p is valid in the inference’s default governing logic, as in
MODEST*. The recursive clause (ii) allows for the case that there is a
chain of individually valid inferences from the premises to the conclu-
sion, which may not be valid in the default governing logic of Γ ⊢ p, as
there is for both Inessential Mix and Essential Mix by way of (9).
Clause (ii) requires only that q ⊢ p be valid, not that it be valid in its
default governing logic. That allows for the chain of inferences to include
more than one intermediate step. For example, let I, R, and C be claims
from domains governed by intuitionistic, relevantistic, and classical logic,
respectively, and let S be a classical but not relevantistic consequence of
C. We should want I & ~~(R&C) ⊢S to be valid. Although the premise
entails ~~(R&C) in all the logics in play, it does not entail C in all the
logics in play. Further, ~~(R&C) ⊢S is not valid in its (relevantistic)
default governing logic. To get from ~~(R&C) to S, we must go by way
of the relevantistically valid ~~(R&C) ⊢C and then the classically valid
C⊢S. Requiring q ⊢ p to be valid in its default governing logic would not
deliver the desired result.
MODEST** ensures that p is a logical consequence of Γ so long as
there is a chain of individual inferences, each valid in its default govern-
ing logic, from Γ to p. That likewise ensures that logical consequence for
mixed inferences is transitive.
There are at least two disadvantages to such an approach.

First, an advocate of DLP might want to allow for the possibility that
logical consequence in some discourses is not transitive. For example, one
might think that doing so is the best way to avoid sorites paradoxes in
vague discourse.9 In such a logic, the following argument would be
invalid
Pure Bald
( 10) A person with 0 hairs is bald.

(11) For all x, if a person with x hairs is bald, then so is a person with x+1
hairs.
(12) Therefore, a person with 100,000 hairs is bald.
Pure Bald is invalid by the lights of a non-transitive logic, because get-

ting from (10) and (11) to (12) requires intermediate steps, such as:
( 13) If a person with 0 hairs is bald, then a person with 1 hair is bald.
(14) A person with 1 hair is bald.
(15) If a person with 1 hair is bald, then a person with 2 hairs is bald.
(16) A person with 2 hairs is bald.
and so on to:
(17) A person with 999,999 hairs is bald.
The point of having a non-transitive logic is to stop such chains of

inferences.
Given MODEST**, though, we are guaranteed transitivity of logi-
cal consequence when we deal with propositions from multiple
domains. So, consider this inference, where ‘P(x)’ stands for ‘A person
with x hairs is bald,’ ‘M’ stands for ‘Murder is wrong’ (or some other
claim from a non-vague discourse), and ‘…’ abbreviates the obvious
conjuncts:
402 C. B. Wrenn
Mixed Bald
(18) P(0) &M

(19) (P(0) &M→P(1) &M) & … & (P(99,999) &M→P(100,000) &M)
(20) P(100,000)
In the non-transitive default governing logic, (18) and (19) can entail10:
(21) (P(1) & M) & (P(1) & M→P(2) & M) & … & (P(99,999)& M →
P(100,000) & M)
The inference from (21) to (22) is valid in its default governing logic:
(22) (P(2) & M) & (P(2) & M → P(3) & M) & … & (P(99,999)& M
→ P(100,000) & M)
By similar steps, each valid in its default governing logic, we ultimately

reach the conclusion P(100,000) & M, and then P(100,000). Since there
is such a chain of individual inferences, each valid in its default governing
logic, MODEST** counts Mixed Bald as valid. But that just means
MODEST** undoes the very work that having a non-transitive logic for
vague discourses was supposed to do.
This disadvantage might not be very worrisome to some pluralists.
Beall and Restall claim that non-transitive logics cannot possibly satisfy
GTT, and so non-transitive logics are not really logics at all (2006, p. 91).
A fan of AP + DLP could say the same.
A second, more serious disadvantage is that this approach still invali-
dates some valid mixed inferences, such as this one:
Disjunctive Mix
( 23) Either it’s not the case that Jack didn’t go up the hill or 1 + 1 = 2.
(24) Therefore, either Jack went up the hill or 1 + 1 = 2.
The inference from (23) to (24) is not intuitionistically valid, and (23)
doesn’t have any intuitionistic consequences from which (24) can be val-
idly inferred. Both MODEST* and MODEST**, then, count Disjunctive
Mix as invalid. But that seems wrong; the argument is valid.
8 Conclusion
The problems discussed here arise for DLP if we try to solve the problem
of mixed inferences by way of MODEST* or a modification of it. The
advocate of AP + DLP might then try to find a different solution to the
problem. Unfortunately, such a solution looks hard to come by.
The trouble is that mixed inferences can have unmixed conclusions. So,
we want to guarantee that no sound mixed inference has a conclusion
that is untrue in its home discourse. Short of MODEST* and its relatives,
there may be no way to guarantee that. The problem can be seen by con-
sidering an argument such as this (Lynch 2008, pp. 136–7):
NIX: If it is not the case that Sophie’s choice is morally right, then grass is
not green. But grass is green; so Sophie’s choice is morally right.
Suppose (a) that Sophie’s choice is to decide which of her children to

give over to the Nazis (as in the novel), (b) that ‘Sophie’s choice is morally
right’ is neither true nor false, and (c) that conditionals with antecedents
that are neither true nor false are true.11 Then the argument’s premises are
true, but its conclusion is not true. In evaluating NIX, we need to use a
logic that does not count it valid.
The argument for MODEST*, then, is that if we evaluate inferences
according to any logic stronger than the logic that applies to any of their
components, then we will wind up endorsing certain conclusions that are
not true in their home discourses. So, to maintain AP + DLP without
losing the ability to evaluate any mixed inferences as valid at all, we need
to take the cautious route embodied by MODEST*.
AP + DLP thus faces an unfortunate dilemma. On the one hand, it
needs MODEST* in order to validate any mixed inferences at all
without giving up on its AP component. On the other hand,
MODEST* invalidates some inferences that, intuitively, the advocate
of AP + DLP should want to count as valid, and it invites rejecting
DLP in favor of VLP.
There may be a way out of this dilemma. Maybe a creative modifica-
tion of AP + DLP can validate the various valid mixed inferences dis-
cussed here without validating NIX. MODEST* does not require, but
404 C. B. Wrenn
only invites, the retreat from DLP to VLP, and maybe that is an invitation
easily declined. Nevertheless, it does appear to be a serious problem for
AP + DLP, and there is one obvious way to avoid it: just don’t combine
AP and DLP in the first place.
Notes
1. “Science, thanks to its links with observation, retains some title to a cor-
respondence theory of truth; but a coherence theory is evidently the lot
of ethics” (Quine 1981, p. 63).
2. Lynch (2013) downplays the role of discourses in pluralist theories, but
most pluralists (including earlier time-slices of Lynch) suppose that, at
least for atomic propositions, what truth-property is relevant to a propo-
sition is a function of what discourse the proposition belongs to.
3. Different versions of functionalism characterize the relationship between
truth itself and the various other properties that “realize,” “manifest,” or
“play the role of ” truth differently. For discussion see Wright (2013).
4. GTT gives a generic characterization of logical consequence because an
argument is valid if and only if its conclusion is a logical consequence of
its premises.
5. Lynch also considers a solution on which the governing logic is the
weakest logic among those governing the various domains involved in
the compound proposition or inference. That solution, however, is just a
special case of MODEST*, where the logics in question are ordered in
such a way that one of them is weakest.
6. Some paraconsistent logics reject Disjunctive Syllogism. Maybe some
jokes both are and aren’t funny. Then such a logic might well be the right
one for discourse about comedy.
7. There is room for some debate as to what does or does not constitute a
“logical reason” for a logical law to hold. Here I assume that law holds
for a logical reason when its instances are logical truths. On Beall’s
approach, the instances of ~(p& ~p) are logical truths, so long as they
don’t involve the transparent truth device.
8. I owe thanks to Jack Lyons, Stewart Shapiro, and Sarah Wright for point-
ing out the need to consider versions of this move on behalf of AP +
DLP.
9. See Zardini (2008) and Cobreros et al. (2012).

10. Whether they do entail it might depend on the details of the non-
transitive logic in question. The entailment holds for a logic whose non-
transitive consequence relation is the ⊨ct of Cobreros et al. (2012).
11. These are the suppositions Lynch makes in presenting the case.
References
Beall, J.C. 2009. Spandrels of Truth. New York: Oxford University Press.
Beall, J.C., and G. Restall. 2006. Logical Pluralism. Oxford: Clarendon Press.
Cobreros, P., P. Egré, D. Ripley, and R. van Rooij. 2012. Tolerant, Classical,
Strict. Journal of Philosophical Logic 41 (2): 347–385.
Lynch, M.P. 2001. A Functionalist Theory of Truth. In The Nature of Truth, ed.
M.P. Lynch, 723–749. Cambridge, MA: MIT Press.
———. 2008. Alethic Pluralism, Logical Consequence and the Universality of
Reason. Midwest Studies In Philosophy 32 (1): 122–140.
———. 2009. Truth as One and Many. Oxford/New York: Oxford University
Press/Clarendon Press.
Pedersen, Nikolaj J.L.L. 2014. Pluralism x 3: Truth, Logic, Metaphysics.
Quine, W.V. 1981. On the Nature of Moral Values. In Theories and Things.
Predicates. Analysis 57 (3): 209–210.
Tennant, N. 1987. Natural Deduction and Sequent Calculus for Intuitionistic
Relevant Logic. Journal of Symbolic Logic 52: 665–680.
———. 2012a. Changes of Mind: An Essay on Rational Belief Revision. Oxford:
———. 2012b. Cut for Core Logic. The Review of Symbolic Logic 5 (3):
450–479.
Press.
406 C. B. Wrenn
———. 2013. A Plurality of Pluralisms. In Truth and Pluralism: Current

Zardini, E. 2008. A Model of Tolerance. Studia Logica 90 (3): 337–368.
Logic for Alethic, Logical,
and Ontological Pluralists
Andy D. Yu
1 Introduction
Following Wright (1992), one might maintain that the debate between
realists and antirealists about a given domain of subject matter—whether
the domain is about scientific, mathematical, or ethical subject matter—
is best formulated as a debate about the nature of the truth property
associated with the domain. This motivates the standard understanding,
due to Wright (1992), of alethic pluralism, according to which there are
distinct, domain-specific truth properties, where domains are associated
with subject matter.1 To illustrate, the correspondence truth property is
associated with the scientific domain, the coherence truth property is
associated with the mathematical domain, and the superassertible truth
property (which takes truth to be epistemically constrained) is associated
with the ethical domain.
Before continuing, let me clarify two simplifying assumptions. First,
pluralists reject truth as it is represented in the familiar account of logic
in favor of an alternative representation that captures its distinctness and
A. D. Yu (*)
Faculty of Law, University of Toronto, Toronto, ON, Canada

https://doi.org/10.1007/978-3-319-98346-2_17
408 A. D. Yu
domain specificity. I set aside the suggestion that pluralists can unquali-
fiedly grant truth as it is represented in the familiar account, as just T
(“truth”). Not least, pluralists themselves concede that they should say
something more. Crucially, the suggestion does not do justice to plural-
ists’ commitment that there are distinct domain-specific truth properties.
For the commitment suggests that something is missing from the familiar
representation of truth, so that the familiar account stands in need of
revision. Given the intertwining of truth with logic, alethic commitments
motivate corresponding logical commitments. Second, for simplicity,
truth bearers are sentences. That truth bearers are sentences is compatible
with propositions and other entities also being truth bearers.
My focus in this chapter is on how pluralists of various stripes might
go about doing logic. As I explain in what follows, alethic pluralists face
challenges to provide accounts of atomic sentences, quantified sentences,
complex sentences, and logical consequence that capture their view of
truth. But the challenges might not be limited to these. For alethic plural-
ists might also be logical pluralists, in taking distinctions between truth
properties to induce corresponding distinctions between logical conse-
quence relations. Similarly, alethic pluralists might also be ontological
pluralists, in taking distinctions between truth properties to induce cor-
responding distinctions between quantifiers. Alethic pluralists who are
also logical pluralists, or ontological pluralists, should provide a unified
account of logical consequence, or of quantified sentences, that captures
their view of truth and logical consequence, or of truth and quantifiers.
My plan for this chapter is as follows. In Sect. 2, I review the account
I propose on behalf of alethic pluralists.2 In Sect. 3, I extend the account
to accommodate logical pluralism as well as ontological pluralism.
2 Alethic Pluralism
In this section, I outline logical challenges for alethic pluralists and review
my answers to them on behalf of alethic pluralists. In the first subsection,
I outline four challenges, involving atomic sentences, quantified sen-
tences, complex sentences, and logical consequence. In the second sub-
section, I review my answers to them.
Logic for Alethic, Logical, and Ontological Pluralists 409
Logical Challenges for Alethic Pluralists
Alethic pluralists face four challenges, which arise from considering mix-
tures of sentences about distinct subject matters.
The first challenge, to maintain a unified account of atomic sentences,
arises from considering atomic sentences, which can be about distinct
subject matters.3
Pluralism is often illustrated using straightforward sentences. Consider
the following sentences:
(1) Water is H2O.

( 2) 1 + 1 = 2.
(3) Murder is wrong.
These sentences are about exactly one subject matter, and so—given that
they are true—are true in exactly one way. Given that (1) is a scientific
sentence, it is correspondence true; given that (2) is a mathematical sen-
tence, it is coherence true; and given that (3) is an ethical sentence, it is
superassertible true. But what about when matters are less straightfor-
ward? Not all sentences are about exactly one subject matter. Atomic
sentences can be mixed, in that the sentences are about distinct subject
matters; these contrast with sentences that are pure, which are not about
distinct subject matters. But if mixed sentences are true, it is unclear in
what way or ways they are true.
Consider the following mixed atomic sentence:
(4) The number π is beautiful.
Apparently, (4) is about distinct subject matters, namely, mathematical

subject matter and aesthetic subject matter. It is unclear in what way or
ways true mixed sentences, such as (4), are true. Given that (4) is true, in
what way or ways is it true? Perhaps it is both correspondence true and
superassertible true. Alternatively, perhaps it is just superassertible true.
Finally, perhaps it is true in neither of these ways, but if so, then in
what way or ways is it true?
410 A. D. Yu
The challenge is for pluralists to maintain an account of mixed atomic

sentences, and more generally, a unified account of atomic sentences,
whether pure or mixed. We can put the challenge thus:
First logical challenge. Maintain a unified account of atomic sentences.
The second challenge, to maintain a unified account of quantified sen-

tences, arises from considering quantified sentences, which can be about
distinct subject matters.4
Consider the following mixed quantified sentence:
(5) Everything is self-identical.
As with (4), (5) also seems to be about distinct subject matters.

Straightforwardly, (5) seems to be about everything. But if (5) is about
everything, and so is associated with a generic domain, then the domain-
specific property associated with the generic domain seems to be a generic
truth property. While even pluralists might grant the existence of a generic
truth property, it remains unclear to what extent they should grant such
a property theoretical significance.5
The challenge is for pluralists to maintain an account of mixed quanti-
fied sentences, and more generally, a unified account of quantified sen-
tences, whether pure or mixed. We can put the challenge thus:
Second logical challenge. Maintain a unified account of quantified

sentences.
The third challenge, to maintain a standard account of complex sen-

tences, arises from considering logical combinations of sentences, which
can be about distinct subject matters.6
Consider the following mixed complex sentences:
(6) Water is H2O and 1 + 1 = 2.
Given that the scientific sentence (1) is correspondence true and the
mathematical sentence (2) is coherence true, their conjunction (6) is also
true. But in what way or ways is (6) true? The options available here are
similar to the options available when considering mixed atomic sen-
tences. Perhaps (6) is both correspondence true and coherence true.
Alternatively, perhaps it is just correspondence true. Finally, perhaps it is
true in neither of these ways, but if so, then in what way or ways is it true?
Here, one version of this last option is salient. Given that a conjunction
is true iff both conjuncts are true, perhaps there is a truth property com-
mon to (1), (2), and (6). But such a property seems to be a generic truth
property. Again, it is unclear to what extent pluralists should grant such
a property significance.
The challenge is for pluralists to maintain an account of mixed com-
plex sentences, and more generally, a standard account of complex sen-
tences, whether pure or mixed. We can put the challenge thus:
Third logical challenge. Maintain a standard account of complex sentences,

according to which the following obtains:
(i) a negation is true iff the negand is false;
(ii) a conjunction is true iff each conjunct is true;
(iii) a disjunction is true iff either disjunct is true.
Plausibly, what counts as a standard account should be understood in

a looser rather than a stricter sense. To permit a reasonable amount of
flexibility, the standard account of complex sentences sought need not be
identical to the account from the familiar account of logic. Let me
elaborate.
Let us assume a sentential language ℒ, with atomic sentences
Atomic(ℒ) and sentences Sent(ℒ). Also, let us adopt the standard con-
vention of using the metavariables α, β, γ, … (with or without subscripts)
to range over atomic sentences, ϕ, ψ, χ, … (with or without subscripts)
to range over all sentences, and Γ to range over sets of sentences. For
simplicity, we focus on the logical operators that are negation, conjunc-
tion, and disjunction. The material conditional and the material bicondi-
tional are defined in terms of these operators in the usual way, as follows:
ϕ → ψ: = ¬ϕ ψ and ϕ ↔ ψ: = (ϕ → ψ) (ψ → ϕ).
412 A. D. Yu
On the familiar account of complex sentences, a model M: Atomic(ℒ)

→ {T, F } is a function from atomic sentences to truth values, where T
represents the generic truth property and F represents the generic falsity
property.7 Given a model M, we get the semantic value function |·|M :
Sent(ℒ) → {T, F} from sentences to truth values based on the model M.
We define | · |M by |α|M = M(α) (where α is atomic) and the following:
(LO) (i) |¬ϕ|M = T iff |ϕ|M = F;

(ii) |ϕ ∧ ψ|M = T iff |ϕ|M = T and |ψ|M = T;
(iii) |ϕ ∨ ψ|M = T iff |ϕ|M = T or |ψ|M = T.
Recalling the point about permitting a reasonable amount of flexibility, it

is enough that the account of complex sentences sought is similar enough
to the familiar account to count as standard. Granted, whether an account
is similar enough to the familiar one is vague. Still, examples show that
we have enough of a grip of what counts as standard for the notion to be
fruitfully employed. The strong Kleene account of complex sentences is a
standard account, but a noncompositional account probably is not.
The fourth challenge, to maintain a standard account of logical conse-
quence, arises from considering mixed arguments, where the premises
and conclusion can be about distinct subject matters.8
Consider the following argument, which has premises (1) and (2) and
conclusion (6):
( 1) Water is H2O.
(2) 1 + 1 = 2.
(6) Water is H2O and 1 + 1 = 2.
Clearly, the argument from (1) and (2) to (6) is valid: (6) is a logical con-
sequence of (1) and (2). On the standard account of logical consequence,
logical consequence necessarily preserves truth. So perhaps some single
truth property is preserved from (1) and (2) to (6). But what property is
that? Perhaps the truth property is common to (1), (2), and (6). If so,
then the property seems to be a generic truth property. Once again, it is
unclear to what extent pluralists should grant a generic truth property
significance. The challenge is for pluralists to maintain an account of

mixed arguments, and more generally, a standard account of logical con-
sequence. We can put the challenge thus:
Fourth challenge. Maintain a standard account of logical consequence,

according to which logical consequence necessarily preserves truth.
To permit a reasonable amount of flexibility, the standard account of

logical consequence sought need not be identical to the account from the
familiar account of logic. On the familiar account of logical consequence,
we have the following:
(LC) Γ ⊨ ϕ iff for every model M, if |γi|M = T for every γi in Γ, then
|ϕ|M = T.
Logical consequence preserves truth across all models, and is classical.

It is enough that the account sought is similar enough to the familiar
account to count as standard. Examples show that we have enough of a
grip of what counts as standard for the notion to be fruitfully employed.
A many-valued account of logical consequence in terms of designation is
standard, but not an account that validates the inference from ϕ to ¬ϕ.
In sum, the four logical challenges for pluralism are to maintain uni-
fied accounts of atomic sentences and quantified sentences, and standard
accounts of complex sentences and of logical consequence.
Logic for Alethic Pluralists
I now review the account that I propose on behalf of alethic pluralists.9
For a Sentential Language
I begin by reviewing the account for a sentential language.

On the account I propose on behalf of pluralists, there is a one-to-one
correspondence between domains, domain-specific truth properties, and
domain-specific falsity properties. Pure domains are associated with
414 A. D. Yu
exactly one subject matter, while impure domains are associated with two
or more subject matters. Where domains are either pure or impure, pure
domains generate all domains. Each atomic sentence is assigned to exactly
one domain. Negations are always assigned to the same domain as the
negand, while conjunctions and disjunctions may or may not be assigned
to the same domain as each operand, depending on whether or not the
operands are assigned to the same domain. Each atomic sentence is then
assigned a domain-specific truth value, where the relevant domain is the
one it is assigned. The domain-specific truth values of negations, con-
junctions, and disjunctions are determined by the domain-specific truth
values of the operands. Logical consequence necessarily preserves domain-
specific truth.
To incorporate talk of domains, we assume a structure 𝕯 = ⟨𝒟, ⊙⟩,
where D is a nonempty set and ⊙ is a binary operation on 𝒟. Intuitively,
𝒟 is the set of domains, and ⊙ is a mixing operation on domains. To
capture the idea that there are distinct domains (and associated domain-
specific truth properties, as well as domain-specific falsity properties), we
require that there are at least two members in 𝒟: |𝒟| ≥ 2. Let Di denote
a member of 𝒟. Intuitively, Di ⊙ Dj is the mixed domain for Di and Dj.
Since we want the mixed domain for identical domains to be the same
domain and do not want the order in which we mix domains to matter,
we require that 𝕯 is a semilattice:
(DS) (i) Di ⊙ Di = Di (idempotency);

(ii) Di ⊙ Dj = Dj ⊙ Di (commutativity);
(iii) (Di ⊙ Dj) ⊙ Dk = Di ⊙ (Dj ⊙ Dk) (associativity).
As usual, a partial order ≤ on 𝒟 is induced by setting the following:
(PO) Di ≤ Dj iff Di ⊙ Dj = Dj.
Note that ≤ in turn induces a strict partial order <, where Di < Dj iff
Di ≤ Dj and Di ≠ Dj. The following desirable feature follows from the fact
that 𝕯 is a semilattice:
(SG) If Di ≰ Dj and Dj ≰ Di, then Di < Di ⊙ Dj and Dj < Di ⊙ Dj.

This feature captures the idea that distinct domains where neither
domain is induced by the other induce new domains.
To incorporate talk of subject matter, we assume that there is a distin-
guished subset Pure(𝒟) of pure domains from which all domains are
generated. Accordingly, 𝒟 is the smallest superset of Pure(𝒟) closed
under the mixing operation ⊙. Intuitively, pure domains are associated
with exactly one subject matter, while impure domains are associated
with two or more subject matters. Pure domains are incomparable, in
that for distinct pure domains Di and Dj, we have both Di ≮ Dj and Dj ≮
Di. To determine which pure domains, or subject matters, a given domain
is associated with, we define the subject matter extraction function m: 𝒟
→ ℘ (Pure(𝒟)) from domains to sets of pure domains by m(Di) = {Dj ∈
Pure(𝒟 ) | Dj ≤ Di}.
To incorporate talk of domain-specific truth and domain-specific fal-
sity, we associate with 𝕯 two isomorphic structures, 𝕿 = ⟨𝒯, ⊕⟩ and 𝕱
= ⟨ℱ, ⊖⟩. The semantic values, and more specifically domain-specific
truth values, are drawn from 𝒯 ∪ ℱ. Intuitively, 𝒯 is the set of domain-
specific truth properties, and ℱ is the set of domain-specific falsity prop-
erties. Let TDi as well as Ti—both notations are useful—denote the
domain-specific truth value that represents the truth property associated
with domain Di. Similarly, let FDi as well as Fi denote the domain-spe-
cific truth value that represents the falsity property associated with
domain Di. Since 𝕿 and 𝕱 are isomorphic to 𝕯, the partial orders on
domain-specific truth values mirror the partial order on domains:
(OR) Di < Dj iff Ti < Tj iff Fi < Fj.
A model M is a pair M = ⟨d, v⟩, where d: Sent(ℒ) → 𝒟 is a domain

assignment function from sentences to domains and v: Atomic(ℒ) → 𝒯
∪ ℱ is a valuation function from atomic sentences to domain-specific
truth values. Intuitively, d assigns a sentence to the domain it is associated
with. We define d by the following:
(DO) (i) d(α) ∈ 𝒟

(ii) d(¬ϕ) = d(ϕ);
(iii) d(ϕ ∧ ψ) = d(ϕ) ⊙ d(ψ);
416 A. D. Yu
(iv) d(ϕ ∨ ψ) = d(ϕ) ⊙ d(ψ).
This definition permits even atomic sentences to be assigned to any

domain, whether pure or impure. Intuitively, v assigns an atomic sen-
tence to a domain-specific truth value. Accordingly, we require that v(α)
∈ {Td(α), Fd(α)}.
Given a model M, we get the semantic value function | · |M: Sent(ℒ)
→ 𝒯 ∪ ℱ from sentences to domain-specific truth values based on M.
To capture the idea that a sentence is either domain-specifically true or
domain-specifically false, we first require that |ϕ|M ∈ {Td(ϕ), Fd(ϕ)}. Then,
we define | · |M by |α|M = v(α) and the following:
(LO*) (i) |¬ϕ|M = Td(¬ϕ) iff |ϕ|M = Fd(ϕ);

(ii) |ϕ ∧ ψ|M = Td(ϕ ∧ ψ) iff |φ|M = Td(ϕ) and |ψ|M = Td(ψ);
(iii) |ϕ ∨ ψ|M = Td(ϕ ∨ ψ) iff |φ|M = Td(ϕ) or |ψ|M = Td(ψ).
Intuitively, a negation is domain-specifically true iff the negand is domain-

specifically false, a conjunction is domain-specifically true iff each con-
junct is domain-specifically true, and a disjunction is domain-specifically
true iff either disjunct is domain-specifically true.
Finally, we define logical consequence as follows:
(LC*) Γ ⊨ ϕ iff for every model M, if |γi|M = Td(γ)for every γi in Γ,

then |ϕ|M = Td(ϕ).
Intuitively, an argument is valid iff in every model where each premise

is domain-specifically true, the conclusion is also domain-specifically
true. Logical consequence preserves domain-specific truth across all
models.
In Yu (2017), I show that the overall account meets a number of con-
straints and desiderata, and that the account of logical consequence is
classical. I also submit that the account answers three of the four logical
challenges—all of the four challenges except the challenge to maintain a
unified account of quantified sentences.
For a First-Order Language
I now extend the account for a sentential language to a first-order setting.

I show that the resulting account provides a unified account of quantified
sentences as well.
The proposed first-order extension is conservative with respect to the
accounts of complex sentences and of logical consequence, but intro-
duces richer models, and variable assignment functions. Following Sher
(2005); Shapiro (2009, 2011), I suggest that pluralists about truth should
also be pluralists about satisfaction: there are distinct, domain-specific
satisfaction relations. The main task is to formally develop this
suggestion.
Let us move to a countable first-order language ℒ+, which includes
quantified formulas. In this setting, the metavariables ϕ, ψ, χ, … range
over all formulas; τ1, τ2, … range over all terms (i.e., constants and vari-
ables); υ ranges over variables; and Φn ranges over n-ary predicates. The
structures 𝕯, 𝕿, and 𝕱 are much as before.
A model is now a triple M = ⟨d, U, I⟩, where d is a domain assignment
function, U is a nonempty set of objects, and I is an interpretation func-
tion. I discuss domain assignment functions and interpretation functions
in more detail in what follows.
An interpretation function maps constants to members of U and n-ary
predicates to functions from sets of n-tuples of objects to domain-specific
truth values:
(I′) (i) I(ϵ) ∈ U if ϵ is a constant;

(ii) I(Φn) ∈ {f | f : Un → 𝒯 ∪ ℱ}
To motivate the definition of domain assignment functions, consider

that the domain that a formula is assigned is determined by both the
predicates and the relevant objects. The predicates are not enough to
determine the domain. While “Water is H2O” and “0 = {}” contain
exactly the same predicates (i.e., the identity predicate), they are assigned
distinct domains. That the former sentence is about scientific subject
matter suggests that it is associated with the scientific domain (but it is
418 A. D. Yu
not associated with the mathematical domain). And that the latter sen-
tence is about mathematical subject matter suggests that it is associated
with the mathematical domain (but it is not associated with the scientific
domain). Similarly, the relevant objects are not enough to determine the
domain. While “2 is a prime number” and “The number 2 is beautiful”
are about exactly the same objects (i.e., the number 2), they are associated
with distinct domains. That the former sentence involves a mathematical
predicate suggests that it is associated with the mathematical domain (but
it is not associated with the aesthetic domain). And that the latter sen-
tence involves an aesthetic predicate suggests that it is associated with the
aesthetic domain.
To capture the above, a domain assignment function now needs to
assign formulas together with variable assignment functions to domains,
where a variable assignment function a based on a model M assigns vari-
ables to members of U. Since only the terms’ denotations (rather than the
terms themselves) matter for determining the domain, we go through
denotation functions and auxiliary domain assignment functions. A
denotation function [·]a based on variable assignment function a is
defined by the following:
(DN) (i) [ϵ]a = I(ϵ) for all constants ϵ;

(ii) [υ]a = a(υ) for all variables υ.
An auxiliary domain assignment function x from n-ary predicates and

n-tuples of objects to domains, meanwhile, is required to satisfy the fol-
lowing, where ⟨u1, … , un⟩ ∈ Un:
(XD) x(Φn, ⟨u1, … ,un⟩) ∈ 𝒟
So, setting aside quantified formulas for now, we define d by the

following:
(DO′) (i) d(Φnτ1…τn, a) = x(Φn, 〈[τ1]a, …, [τn]a〉);

(ii) d(¬ϕ, a) = d(ϕ, a);
(iii) d(ϕ ∧ ψ, a) = d(ϕ, a) ⊙ d(ψ, a);
(iv) d(ϕ ∨ ψ, a) = d(ϕ, a) ⊙ d(ψ, a).
In the special case where n = 0, so that an atomic formula Φ0 is just a

sentence letter, d assigns Φ0 to a domain irrespective of the variable
assignment function.
The way d works with quantified formulas needs more explanation.
The idea is that the domain assigned to a given quantified formula is
determined by the domains assigned to the embedded formula. For
example, the domain assigned to “Every x is such that x is identical to x”
is determined by the domains assigned to the formula “x is self-identical”
relative to all variable assignment functions that differ at most in their
assignment to the variable x. Since there might be infinitely many such
variable assignment functions, we assume that the mixing operation ⊙
on domains is defined for infinite sets of domains. Let a ∼υ b mean that
variable assignment function a differs from variable assignment function
b at most in the assignment to the variable υ. Then, now focusing on
quantified formulas, given that the universal quantifier and the existential
quantifier are duals, and negation preserves domain membership, we
define d by the following:
(DO′) (v) d(∀υϕ, a) = ⦿ {d(ϕ, ai)| ai∼υa};

(vi) d(∃υϕ, a) = ⦿ {d(ϕ, ai)| ai∼υa}.
Given a model M, the semantic value function | · |M,a based on M and

variable assignment function a treats terms and predicates as usual:
(SV) (i) |τ|M,a = [τ]a for all terms τ;

(ii) |Φn|M,a = I(Φn) for all predicates Φn.
To treat atomic formulas, we require that | · |M,a satisfies the following:
(ATM) |Φn τ1 ⋯ τn|M,a = Td(Φnτ1⋯τn,a) iff |Φn|M,a(|τ1|M,a,⋯,|τn|M,a) ∈ 𝒯
The account of negation, conjunction, and disjunction is much as

before:
(LO*′) (i) |¬ϕ|M, a = Td(¬ϕ, a) iff |ϕ|M, a = Fd(ϕ, a);

(ii) |ϕ ∧ ψ|M, a = Td(ϕ ∧ ψ, a) iff |ϕ|M, a = Td(ϕ, a) and |ψ|M, a = Td(ψ, a);
(iii) |ϕ ∨ ψ|M, a = Td(ϕ ∨ ψ, a) iff |ϕ|M, a = Td(ϕ, a) or |ψ|M, a = Td(ψ, a).
420 A. D. Yu
The account of the universal and existential quantifiers is as follows:
(LO*′) (iv) |∀υϕ|M, a = Td(∀υϕ, a) iff |ϕ|M, b = Td(ϕ, b) for all b where b ∼υ a;

(v) |∃υϕ|M, a = Td(∃uϕ, a) iff |ϕ|M, b = Td(ϕ, b) for some b where b ∼υ a.
Finally, the account of logical consequence is much as before:
(LC*′) Γ ⊨ ϕ iff for every model M, if |γi|M, a = Td(γ, a) for every γi in
Γ, then |ϕ|M, a = Td(ϕ, a).
In Yu (2017), I submit that the account—for a first-order language—

proposed on behalf of pluralists answers all four of the above logical chal-
lenges for pluralism, including the challenge to maintain a unified
account of quantified sentences.
3 ccommodating Logical or Ontological

A
Pluralism
In this section, I propose two extensions of the account for alethic plural-
ists. Each extension accommodates one of logical pluralism and ontological
pluralism, each of which one might maintain in addition to alethic plural-
ism. I focus on domain-specific versions of logical pluralism and ontologi-
cal pluralism, which are motivated by considerations similar to the ones
that motivate alethic pluralism. I emphasize, however, as those who suggest
combining alethic pluralism with either logical pluralism or ontological
pluralism emphasize, that alethic pluralism is independent of either.10
In the first subsection, I propose an extension that accommodates logi-
cal pluralism. In the second subsection, I propose an extension that
accommodates ontological pluralism. While I do not pursue it, one might
combine all three of alethic pluralism, logical pluralism, and ontological
pluralism.11 I hope the details I provide suffice to show how an extension
that accommodates the combination of all three pluralisms might go.
Recall Wright’s (1992) motivation for alethic pluralism: one might
maintain realism about some domains but antirealism about others. But
perhaps domain-specific truth properties are linked to domain-specific

logics, and domain-specific quantifiers.
Insofar as Wright’s consideration regarding realism motivates alethic
pluralism, it might also motivate logical pluralism, a salient version of
which is the view that there are distinct logics associated with distinct
domains.12 To illustrate logical pluralism, classical logic is associated with
the scientific domain, intuitionistic logic is associated with the mathe-
matical domain, and some nonclassical logic is associated with the ethical
domain. To illustrate the combination of alethic pluralism with logical
pluralism, realist domains, which are associated with the correspondence
truth property, are associated with classical logic, while antirealist
domains, which are associated with the superassertible truth property, are
associated with a weaker, nonclassical logic.13
Similarly, insofar as Wright’s consideration motivates alethic pluralism,
it might also motivate ontological pluralism, a salient version of which is
the view that there are distinct logically or metaphysically “perspicuous”
quantifiers associated with distinct domains.14 To illustrate ontological
pluralism, quantifiers that range only over concrete objects are associated
with the scientific domain, quantifiers that range only over abstract
objects are associated with the mathematical domain, and quantifiers that
range only over mind-dependent entities of an appropriate kind are asso-
ciated with the ethical domain.15 To illustrate the combination of alethic
pluralism with ontological pluralism, each domain comes with a distin-
guished ontology of its own, inducing distinctions in the associated
domain-specific truth property.16
To clarify, I am only explaining why one who maintains alethic plural-
ism might also maintain logical pluralism or ontological pluralism. One
might combine alethic pluralism with logical pluralism or ontological
pluralism. Then again, one might not. I do not think alethic pluralists
must maintain any of the combinations here.17
Accommodating Logical Pluralism
The proposed extension of the account for a sentential language accom-

modates logical pluralism, so that some domains but not others are asso-
ciated with a nonclassical logic, by allowing certain domains to be
422 A. D. Yu
associated with a third domain-specific property—an “indeterminacy”

property that indicates indeterminacy—that is neither domain-specific
truth nor domain-specific falsity, and employing (versions of ) the strong
Kleene clauses for complex sentences.
Recall that the pair of structures 𝕿 and 𝕱 are both isomorphic to 𝕯
(and to each other), and provide domain-specific truth values that repre-
sent domain-specific truth properties and domain-specific falsity proper-
ties, respectively. We now associate with 𝕯 further isomorphic structures
to provide further domain-specific truth values. In the simple case, we
associate a single further isomorphic structure 𝕴, which provides domain-
specific truth values that represent domain-specific indeterminacy prop-
erties. The semantic values, and more specifically domain-specific truth
values, are then drawn from 𝕿 ∪ 𝕴 ∪ 𝕱. For convenience, let I Di as well
as Ii denote the domain-specific truth value that represents the indetermi-
nacy property associated with domain Di. Since 𝕴 is isomorphic to 𝕯,
the partial order on 𝕴 mirrors the partial order on 𝕯:
(OR*) Di < Dj iff Ii < Ij.
A model is a pair M = ⟨d, v⟩, much as before, except that now the valua-
tion function v: Atomic(ℒ) → 𝕿 ∪ 𝕴 ∪ 𝕱 is from atomic sentences to
domain-specific truth values in 𝕿 ∪ 𝕴 ∪ 𝕱. We require that v(α) ∈
{Td(α), Id(α), Fd(α)}.18 Given a model M, the semantic value function | · |M:
ℒ → 𝕿 ∪ 𝕴 ∪ 𝕱 is much as before, except that now it is a function from
atomic sentences to domain-specific truth values in 𝕿∪𝕴 ∪𝕱. To cap-
ture the idea that a sentence is either domain-specifically true, domain-
specifically indeterminate, or domain-specifically false, we first require
that |ϕ|M ∈ {Td(α), Id(α), Fd(α)}. Then, we define | · |M by |α|M = v(α) and the
following:
ìTd Øf iff f = Fd f ;
ïï ( ) M ( )
(LO**) (i) Øf M = í Fd ( Øf ) iff f M = Td (f ) ;

ï
ïî I d ( Øf ) otherwise.
ìTd f Ùy iff f = Td f and y = Td (y ) ;

ïï ( ) M ( ) M
(ii) f Ùy M
= í Fd (f Ùy ) iff f M = Fd (f ) or y M
= Fd (y ) ;
ï
ïî I d (f Ùy ) otherwise.
ìTd f Ú y iff f = Td f or y = Td y ;
ïï ( ) M ( ) M ( )
(iii) f Ú y M = í Fd (f Ú y ) iff f M = Fd (f ) and y M = Fd (y ) ;
ï
ïî I d (f Ú y ) otherwise.
Next, we define domain-specific logical consequence as follows:
(LC**) G Di f iff for all M such that d (g j ) = d (f )
= Di for all g j in G , if g j = TDi
M
for all g j in G , then f M = TDi .

Intuitively, an argument is domain-specifically valid iff in all models that
assign all the sentences in the argument to the relevant domain, if each
premise is domain-specifically true, the conclusion is also domain-
specifically true. Domain-specific logical consequence preserves domain-
specific truth across all models that assign all the sentences in the argument
to the relevant domain.
I submit that the resulting account provides a unified account of logi-
cal consequence that captures the commitments of one who is both an
alethic pluralist and a logical pluralist.
Accommodating Ontological Pluralism
The proposed extension of the account for a first-order language accommo-

dates ontological pluralism, so that there are distinct quantifiers associated
with distinct domains, by maintaining that quantifiers are domain-specific,
in that they consider variable assignment functions that assign the relevant
variables to objects in domain-specific partitions of a model’s universe.
We assume that the language now contains domain-specific quantifiers
" Di and $Di , where Di is a domain in 𝒟. A preliminary revision to the
424 A. D. Yu
account is to assume a partition of the universe U in each model on the

basis of domains. Given a model ⟨d, U, I⟩, each cell U Di in the partition
corresponds to a domain Di in 𝒟. Then we revise the clause for quanti-
fied formulas to the following:
(LO***) (i) |∀Di υϕ|M, a = Td(∀υϕ, a) iff |ϕ|M, b = Td(ϕ), b for all b where b ∼υ a
and b(υ) ∈ U Di ;
(ii) |∃Di υϕ|M, a = Td(∃υϕ, a) iff |ϕ|M, b = Td(ϕ), b for some b where b ∼υ a and
b(υ) ∈ U Di .
Intuitively, domain-specific quantifiers only consider variable assignment

functions where the relevant variables are assigned to objects in domain-
specific partitions of a model’s universe.
I submit that the resulting account provides a unified account of quan-
tified sentences that captures the commitments of one who is both an
alethic pluralist and an ontological pluralist.
Notes
1. Sometimes, it is added that a truth property is whatever satisfies certain
truth-characterizing principles. However, this possibly extra feature does
not concern me here, so I omit mention of it for simplicity.
2. See Yu (2017).
3. This challenge is discussed in Lynch (2005b), Sher (2005, 2013), Wyatt
(2013), Pedersen and Wright (2013b).
4. This challenge is discussed in Lynch (2009), Shapiro (2009, 2011),
Cotnoir (2013), Pedersen and Wright (2013b).
5. See Tappolet (2000), Edwards (2008, 2009, 2012), Cotnoir (2009,
2013), Pedersen (2006, 2010, 2012), Pedersen and Wright (2013a),
Wright (2012) on the extent to which pluralists should grant a generic
truth property theoretical significance.
6. This challenge is discussed in Williamson (1994), Tappolet (2000),
Lynch (2001, 2004, 2005b, 2008, 2009, 2013), Edwards (2008, 2009),
Cotnoir (2009), Pedersen and Wright (2013b), Wright (2013).
7. Where there is little risk of confusion, I am somewhat loose with the
distinction between truth properties themselves and truth values that
formally represent them. For example, I may write of T as either being a
truth property or representing a truth property. When discussing alethic

pluralism informally, it is more natural to talk about truth properties,
but when discussing the provision of a formal semantics, it is more natu-
ral to talk about truth values.
8. This challenge is discussed in Sainsbury (1996), Tappolet (1997), Beall
(2000), Lynch (2000, 2004, 2005a, 2008, 2009), Pedersen (2006),
Cotnoir (2013), Wright (2013).
9. See Yu (2017).
10. See Pedersen (2014), Cotnoir and Edwards (2015).
11. See Pedersen (2014) for the suggested combination of all three
pluralisms.
12. See Beall and Restall (2006), Shapiro (2014) on logical pluralism more
generally. The versions they discuss are not all domain-specific versions
of logical pluralism.
13. Wright (1992, 44) seems to suggest this combination of alethic plural-
ism with logical pluralism. Lynch (2008, 2009) and Pedersen (2014)
both suggest the combination.
14. See McDaniel (2009, 2010a, b, 2017), Turner (2010, 2012) on onto-
logical pluralism.
15. In the scientific case, I assume for the sake of illustration that mathemat-
ics—or at least commitment to the existence of abstract mathematical
objects—is dispensable; see Field (1980) for a defense of this view. In the
ethical case, the relevant entities might be mind-dependent overall but
have some basis in mind-independent entities as well. For example, per-
haps the fact that an action constitutes murder is mind-dependent but
has basis in the mind-independent fact that death is brought about.
16. Pedersen (2014), Cotnoir and Edwards (2015) suggest this combination
of alethic pluralism with ontological pluralism.
17. See Pedersen (2014), Cotnoir and Edwards (2015).
18. We might restrict sentences associated with realist domains to only
domain-specific truth values that represent either domain-specific truth
or domain-specific falsity.
References
Beall, J. 2000. On Mixed Inferences and Pluralism About Truth Predicates.
Beall, J., and G. Restall. 2006. Logical Pluralism. Oxford: Oxford University Press.
426 A. D. Yu
Cotnoir, A.J. 2009. Generic Truth and Mixed Conjunctions: Some Alternatives.
Analysis 69 (3): 473–479.
———. 2013. Validity for Strong Pluralists. Philosophy and Phenomenological
Research 86 (3): 563–579.
68 (2): 143–149.
———. 2009. Truth-Conditions and the Nature of Truth: Resolving Mixed
Conjunctions. Analysis 69 (4): 684–688.
———. 2012. On Alethic Disjunctivism. Dialectica 66 (1): 200–214.
Field, H.H. 1980. Science Without Numbers: A Defense of Nominalism. Princeton:
Princeton University Press.
Lynch, M.P. 2000. Alethic Pluralism and the Functionalist Theory of Truth.
———. 2001. A Functionalist Theory of Truth. In The Nature of Truth: Classical
and Contemporary Perspectives, ed. M.P. Lynch, 723–750. Cambridge, MA:
MIT Press.
———. 2004. Truth and Multiple Realizability. Australasian Journal of
Philosophy 82 (3): 384–408.
———. 2005a. Functionalism and Our Folk Theory of Truth: Reply to Cory
Wright. Synthese 145: 29–43.
———. 2005b. Précis to True to Life, and Response to Commentators.
Philosophical Books 46 (289–291): 331–342.
———. 2008. Alethic Pluralism, Logical Consequence and the Universality of
Reason. Midwest Studies in Philosophy 32: 122–140.
McDaniel, K. 2009. Ways of Being. In Metametaphysics: New Essays on the
Foundation of Ontology, ed. D. Chalmers, D. Manley, and R. Wasserman,
———. 2010a. Being and Almost Nothingness. Noûs 44 (4): 628–649.
———. 2010b. A Return to the Analogy of Being. Philosophy and
———. 2017. The Fragmentation of Being. Oxford: Oxford University Press.
Teach Us About Alethic Pluralism? The Monist 89 (1): 102–117.
———. 2010. Stabilizing Alethic Pluralism. Philosophical Quarterly 60 (238):

92–108.
———. 2012. Recent Work on Alethic Pluralism. Analysis 72 (3): 588–607.
———. 2014. Pluralism × 3: Truth, Logic, Metaphysics. Erkenntnis 79: 259–277.
Pedersen, Nikolaj J.L.L., and C.D. Wright. 2013a. Pluralism About Truth As
Alethic Disjunctivism. In Truth and Pluralism: Current Debates, ed. Nikolaj
J.L.L. Pedersen and C.D. Wright, 87–112. Oxford: Oxford University Press.
———. 2013b. Pluralist theories of truth. In Stanford Encyclopedia of Philosophy,
ed. E.N. Zalta. (Spring 2013 ed.). Stanford, CA: Stanford University.
Sainsbury, R.M. 1996. Crispin Wright: Truth and Objectivity. Philosophy and
Shapiro, S. 2009. Truth as One and Many, by Michael P. Lynch. Notre Dame
Philosophical Reviews.
———. 2011. Truth, Function and Paradox. Analysis 71 (1): 38–44.
Sher, G. 2005. Functional Pluralism. Philosophical Books 46 (4): 311–330.
———. 2013. Forms of Correspondence: The Intricate Route from Thought to
Reality. In Truth and Pluralism: Current Debates, ed. Nikolaj J.L.L. Pedersen
and C.D. Wright, 157–179. Oxford: Oxford University Press.
Predicates. Analysis 57 (3): 209–210.
———. 2000. Truth Pluralism and Many-valued Logics: A Reply to Beall.
Turner, J. 2010. Ontological Pluralism. Journal of Philosophy 107 (1): 5–34.
———. 2012. Logic and Ontological Pluralism. Journal of Philosophical Logic
41: 419–448.
Williamson, T. 1994. Critical Notice: Truth and Objectivity by Crispin Wright.
International Journal of Philosophical Studies 2 (1): 130–144.
Press.
———. 2012. Is Pluralism About Truth Inherently Unstable? Philosophical
Studies 159: 89–105.
———. 2013. A Plurality of Pluralisms. In Truth and Pluralism: Current
Yu, A.D. 2017. Logic for Alethic Pluralists. Journal of Philosophy 114 (6): 277–302.
Pluralisms: Logic, Truth
and Domain-Specificity
Rosanna Keefe
1 Introduction
Logicians and Philosophers have developed a large and rich array of logi-
cal systems—intuitionistic logic, relevance logics, free logics, a vast assort-
ment of many-valued logics, multiple conclusion logics, extensions of
classical logic of various types including temporal and modal logics and
many others besides. These are naturally referred to as “logics” and can be
considered and studied in the abstract, giving clear-cut notions of what
follows from what according to the various different logics. Talk of “logic”
(in the singular) often signals something more than a system considered
in the abstract. Logic as the study of what follows from what, looks to
concern more than what follows from what relative to some chosen system.
Similarly for talk of, for example, the study of the logic of our language
or of the world. Not all of the abstract systems are on a par and philoso-
phers have often provided detailed arguments for regarding their chosen
logic as of particular or unique significance, where many of the positions
R. Keefe (*)
Department of Philosophy, University of Sheffield, Sheffield, UK
e-mail: r.keefe@sheffield.ac.uk

https://doi.org/10.1007/978-3-319-98346-2_18
430 R. Keefe
advocated in this way are incompatible.1 How do we determine which of

the candidate logics we should select?
It is natural to suppose that we should treat many of the abstract logi-
cal systems as just that—perhaps unsuccessful candidates for capturing
the genuine logical consequence relation. The possibility of being a logical
pluralist seems to allow us to reject none, or at least fewer of such candi-
dates. The pluralist maintains that there is no unique consequence
relation.
In this chapter, I will explore one type of logical pluralism. There are
other logical pluralist positions that I will not cover here, and, in particu-
lar, I will put aside Beall and Restall’s form of logical pluralism (2006).
They advocate a form of pluralism based on capturing the notion of con-
sequence with (GTT)—an argument is valid iff in every case in which the
premises are all true, the conclusion is also true—and recognising mul-
tiple acceptable ways to spell out the notion of “case” in this definition.
They maintain that this legitimates a wide-ranging logical pluralism,
endorsing at the same time a wide range of apparently conflicting logics
(including classical, intuitionistic and relevance logics). I rejected their
position in an earlier paper (2014), arguing that they do not give us a way
to endorse such a range of different logics. On the other hand, I have
previously defended the viability of a much narrower form of pluralism,
where, within some particular logical framework, we can endorse more
than one different account of validity and logical consequence, because
several alternatives may be equally good candidates for the right generali-
sation of the traditional classical account of validity once the nonclassical
framework is adopted. For example, in a many-valued setting, we can
agree on the number of truth-values and the semantics of the connec-
tives, but leave open whether to regard logical consequence as necessary
preservation of some designated value—where there may be several
acceptable choices for the designated value(s)—or as preservation of truth
in such a way that, on one option, the conclusion must not be less true
than the least true premise.2 What options are there for a pluralism whose
scope is between these widest and narrowest alternatives, or for a rela-
tively wide scope based on a different framework from Beall and Restall’s?
In this chapter, I ask whether we should see different logical systems as
appropriate for different domains (or perhaps in different contexts) and
Pluralisms: Logic, Truth and Domain-Specificity 431
whether this would amount to a form of logical pluralism. One, though

not the only, route to this type of position, is via pluralism about truth.
Given that truth is central to validity, the commitment the typical truth
pluralist has to different notions of truth for different domains may sug-
gest differences regarding validity in those different domains. Indeed, as
we’ll see, the differences between the proposed multiple notions of truth
are often of a type that is clearly significant in relation to logical features,
such as whether or not a constructive notion of truth is at issue. I set out
and assess these issues in the next section.
2 Truth Pluralism and Logical Pluralism

Crispin Wright, Michael Lynch and others defend an important form of
pluralism about truth, which, very roughly, takes the following form. The
conditions required for a sentence to count as true vary with the type of
sentence, or the domain or region of discourse into which that sentence
falls. For a sentence about concrete physical objects to be true is different
from what it takes for a sentence about morality to be true, or an asser-
tion that joke is funny, or perhaps a mathematical statement; some of
these notions may involve correspondence, some may be anti-realist,
with scope for a range of other differences.3
Truth pluralism has taken a variety of different forms. Sometimes the
variation across domains in the properties in virtue of which sentences
are true is combined with retaining a universal concept, or even property
of truth, whereas a strong pluralist denies any common truth property
across domains.4 All versions are to be distinguished from a pluralism or
relativism about what is true, where the same proposition may count as
both true (relative to one thing) and false (relative to another).
Some pluralists about truth have maintained that logical pluralism fol-
lows or at least is a very natural companion position.5 In particular, a key
difference between notions of truth for different domains concerns a dif-
ference over whether it is a constructive notion (e.g. superassertibility or
superwarrant) or a more “objective” one. For example, you might think
that “this is funny” cannot have a completely epistemically inaccessible
truth-value, so you may be inclined to intuitionistic logic in that domain
432 R. Keefe
and, for example, deny that “p or not-p” must always be true, since
(roughly) p and not-p may both be unknowable. Within certain scientific
domains, by contrast, there may be no such constraints and there may
instead be reasons to presuppose bivalence.
Lynch coins the phrase “domain-specific logical pluralism” (DLP) for
the position at issue here. And Pedersen writes, “Logical pluralists main-
tain that the distinction between different kinds of domains is not merely
verbal, but that it bears on inference. What inferences come out valid
may vary from domain to domain”.6 This may be a misleading portrayal
of most logical pluralists, for whom this does not capture the role of
domains (see below); but it is nonetheless worth considering this domain-
specific logical pluralism and some preliminary objections it faces.
One important general question is the following. If different logics
govern different domains, what logic governs arguments that involve
statements from more than one different domain? This question can be
usefully addressed alongside consideration of the “mixed inference chal-
lenge” to pluralism about truth.7 That challenge takes an argument with
premises from different domains, thus subject to different truth predi-
cates, and considers what validity can amount to if it isn’t necessary pres-
ervation of a single property of truth. Tappolet considers, for example,
“Wet cats are funny; this cat is wet; therefore, this cat is funny”, where the
first premise and conclusion are from a different domain with different
standards of truth than the second premise.
One kind of response to the mixed inference challenge maintains that
what matters for validity is necessary preservation of falling under the
concept of truth, even if what it takes to fall under that concept is differ-
ent for different types of sentence.8 Tappolet complains, “why should we
need the many truth predicates instead of the one that does the inferen-
tial job?” (2000, p. 384): that one thing is necessarily preserved in valid
arguments may seem to provide enough unity to undermine truth plural-
ism. But advocates have shown how truth pluralism nonetheless remains
viable on most plausible forms of the view.
The logic governing a mixed inference, on this approach, would be the
intersection of the logics governing each of the domains involved in the
argument, for arguments valid in only some domains will fail to necessar-
ily preserve some kind of truth in some other domain. This will often be
the weakest of the logics for each of the two (or more) domains; for
example, for an argument with realist and anti-realist domains, the logic
would be intuitionistic.9 If the picture is to be as general as possible and
permit a wide range of different logics as applicable to different domains,
including paraconsistent logics, it will not always be the case that the
logic of one domain is contained within the other. If we had an argument
spanning a range of domain representing the full diversity of consequence
relations, the intersection may be nearly empty (compare Beall and
Restall 2006, p. 92.)10
It may be natural, with this approach, to assign particular significance
to that logical consequence relation which is the intersection of the logics
across all domains. Perhaps we should regard that as the logical conse-
quence relation. That would be to abandon domain-specific logical plu-
ralism. It would then be inaccurate to present the key consequence
relation as the intersection across the specific consequence relations, since
those latter will not after all be logical consequence relations. The stron-
ger, apparently justified principles of reasoning within particular domains
would then have to count as non-logical. Lynch says, “… in domains
which, according to this suggestion, nonetheless appear classical … biva-
lence must be true for some non-logical reason. And one might wonder
what that reason might be.” Suppose, then, we resist this option and
uphold domain-specific logical pluralism and the different logics for dif-
ferent domains; there will potentially be other different logics for differ-
ent mixed inferences, depending on what overlapping logics are at issue
and mixed inferences in which it is the very weakest (intersection) logic
at play will be rare.
In the next section, keeping domain-specific logical pluralism on the
table, we consider a different approach.
3 Cotnoir’s Algebraic Approach

I turn next to a different, interesting approach, offered in Cotnoir 2013.
He seeks to tackle the problem of mixed inferences while avoiding the
charge of being committed to a single (possibly disjunctive) notion of
truth as what is preserved in common to all the domains. He focuses on
434 R. Keefe
“Strong Pluralism”—pluralism about truth which is not committed to a

unique property of truth, a view which he endorses and attributes to
Wright.11 The basic idea is to model the language and logic with a more
complex range of truth-values than those employed in a bivalent—or
even typical multi-valent—system. His values are n-tuples, where there is
a place in the n-tuple for each domain. We can then represent a sentence
as true in one particular domain but not others, by giving it a value of
1 in the place corresponding to the domain in which it is true, but not in
the other places. Cotnoir uses this idea to go on to define an algebraic
notion of validity.
Cotnoir’s “admissible valuations” are the valuations satisfying certain
definitions of the connectives that extend the classical definitions. In brief
(for the model he largely focuses on): values in each place are either 0 or
1, and the constraints imposed by the definitions of the connectives mir-
ror at each place of the n-tuple the classical rules for the connective, so,
for example, the negation of a sentence “flips” between 0 and 1 at each
place of the n-tuple and each place of the n-tuple for (A v B) takes 1 at
the relevant place of the n-tuples if it is 1 at that place for A or for B (or
for both). He then introduces a notion of ordering, ≤, between n-tupled
values, where (using v(A) for the value of A), v(A) ≤ v(B) iff v(A V B) =
v(B). There is a top value, with 1 in every place of the n-tuple, and a bot-
tom value—0 in every place; there are lots of values between which have
different mixtures of 0s and 1s. These values aren’t well-ordered; for
example, simplifying to n = 2 with v(A) = <0,1> and v(B) = <1,0>, we
would have neither v(A) ≤ v(B) nor v(B) ≤ v(A) since v(A V B) = <1,1>
which is not equal to v(A) or v(B). The logical consequence relation is
then defined as follows: A is the consequence of some premises iff v(the
conjunction of the premises) ≤ v(conclusion).
Cotnoir goes on to show how the logic resulting from this is classical
because the framework is a Boolean algebra.12 The details of this argu-
ment do not matter for our purposes. We can grant that the framework
he describes, given the assumptions he makes, does indeed deliver classi-
cal logic. Running a simplified illustration will show how some classically
valid inferences come out valid, but not necessarily in the expected way.
If v(A) = <0,1> and v(B) = <1,0>, v(A & B) = <0,0>. A,B trivially entails
A&B because whether the entailment holds turns on whether the
c onclusion is at least as true as the conjunction of the premises, when in

this case the conclusion is identical to the conjunction of the premises. So
the entailment holds even though the premises are each true in some way
and the conclusion has the very bottom value.
In accommodating inferences involving several domains, Cotnoir has
provided a framework that has a universal logic. But this does not sit well
with the position considered above, according to which logical pluralism
is a natural partner for pluralism about truth given the different logical
behavior central to the differences in the truth properties (e.g. whether it
is a constructive notion of truth). Cotnoir seeks to accommodate logical
pluralism by generalising the framework in certain ways to be described.
I shall argue that these fail and that they illustrate a more general problem
in relation to pluralisms about logic and truth.
Cotnoir anticipates the following important objection to his semantic
values, as used to capture the domain-specific properties of truth. On his
account, a sentence, S, falling squarely into domain i and true according
to the corresponding notion of truth will take 1 in the ith place of its
semantic value and zeros in all the other places while its negation will take
0 in the ith place and 1 in all others. This fails to capture the idea that S
and its negation are squarely within that one domain and thus only truth-
apt in the corresponding sense. As Cotnoir summarises: “why think that
the negation of a true descriptive proposition must be morally true,
mathematically true etc.…?” (Cotnoir 2013, p. 574) In response, he sug-
gests introducing a third value, ½, reflecting the status of a sentence in a
domain for which it is not truth-apt. S will then take value ½ in all places
except i, as will its negation. Cotnoir goes on to describe the resulting
many-valued logic, which turns out to be both paracomplete and para-
consistent. For example, the Law of Excluded Middle fails, as he illus-
trates with an example of a sentence taking ½ in all places of its n-tuple
value, where its negation will thus take the same value, as will the disjunc-
tion of these two sentences.13
This development of the framework does not have the advantages
advertised, however. Most importantly for our purposes, it merely shifts
the universal logic from classical logic to this paraconsistent and para-
complete alternative, for this is the logic that determines whether an
argument is valid or sentence is logically true—even if the premises and
436 R. Keefe
conclusions are from within a single domain rather than being a mixed
inference. Cotnoir appears to deny this, however, and says, “One can, if
one wishes, insist that propositions in some domains are always classically
evaluated while allowing for non-classical domains of discourse. One
merely stipulates that, at some coordinates in the n-tuples, components
are always only selected from 1 and 0” (p. 14). But that doesn’t make for
bivalence within the domain in the desired sense that all sentences from
that domain are true or false; rather all sentences at all are true or false in
that domain. As he earlier asks, “why think that the negation of a true
descriptive proposition must be morally true, mathematically true etc.?”
It should be compatible with taking mathematical truth to be bivalent to
deny that either “Jack is funny” or “Jack is not funny” is mathematically
true. This seems to make a mockery of the idea of mathematical truth—
limited to the mathematical domain—within a pluralism about truth.
Cotnoir also claims to provide a framework that can accommodate
domains in which truth is constructive and intuitionistic logic holds,14
but, I will argue, the same problem threatens. The original Boolean alge-
bra delivers classical logic and so leaves no room for a constructive notion
of truth, and thus fails to capture the notion of truth that Wright—whose
position he seeks to stay close to—defends for a domain such as humor
and treats as key to his central notion of superassertibility. Cotnoir’s
response to this kind of problem is to develop a variant on the framework
described above which validates intuitionistic logic. Whereas the original
classical version was based on a Boolean algebra, an intuitionistic version
is based on a Heyting algebra. The resulting consequence relation, ǀ=I, is
intuitionistic.
The mechanics of the system and the construction of semantic values
is somewhat different in this case from the previous versions. In the clas-
sical version, the n-tuple consists of 1s and 0s, where the third possibility
of ½ is added in the many-valued variant. The talk of the Boolean algebra
is then introduced at the level of the behavior of the resulting many-place
values. This is how Cotnoir gets quickly to the conclusion that classical
logic (captured by that Boolean algebra) holds for mixed inference, that
is, allowing for the full range of n-tuples that could be assigned. In the
intuitionistic alternative, we are told that the ith place of the n-tuple are
members of the ith Heyting algebra, where for some domains, this could
be the special case of a Boolean algebra. This suggests an approach to the

logic of a single domain that is determined not by the n-tuple values of
the corresponding sentences, but only the component corresponding to
that domain. I will argue that this approach is problematic.
Note, first, a complication. The Heyting algebra framework encom-
passes Boolean algebras and thus classical logic as a special case. Standard
many-valued domains do not fit this framework, so to also make room
for a many-valued domain, a more general, weaker type of algebra is
needed. I will not go into the details here.
To assess the viability of the suggested solution to allowing a classical
domain alongside intuitionistic ones (and other domains, when further
generalised), it will be useful to ask the following question: what would it
be for an argument to be strictly within one domain alone? There are
general issues for pluralism about truth arising from the specification of
domains, and I won’t enter into those here. A natural informal reply is
that the sentences within domain i are truth-apt within that domain only,
so only receive a value on that co-ordinate. But that won’t fit within
Cotnoir’s framework, where each sentence has an n-tupled value, so must
have some value in place k even if it isn’t truth-apt within that domain
(where, on the original version it would be 0 and it would be ½ on the
many-valued version). Cotnoir says, “As in the semantics prior, 1 in the
ith place represents that the proposition is in domain i and has the prop-
erty truei. Likewise for 0 in the ith place and falsei. To see why this won’t
do, suppose m is a domain respecting classical logic, then Hm, the corre-
sponding Heyting algebra, will be a Boolean algebra with only the values
0 and 1. But in that case, every sentence will take 0 or 1 in the mth place
and, by this definition, count as in the domain.15 For example, if m is the
mathematical domain, moral statements would not be truem and so
would take 0 in the mth value and thus count as part of that domain after
all. This illustrates a general problem regarding assignment of any kind of
semantic value in relation to a particular truth predicate when the sen-
tence is not truth-apt in that respect.
Perhaps we could take a different approach to understanding what it is
to be in a particular domain, where this cannot be read off from the
semantic value assigned. If we had a classification of sentences—maybe
just atomic sentences—into domains that was independent of the
438 R. Keefe
a ssignment of the semantic values, then we could allow sentences that

were assigned 1 or 0 in the mth-placed value, while not being part of the
mth domain. Such a position muddies the waters with regard to the pre-
viously fairly appealing idea of being truthm-apt (i.e. apt for propositions
within the mth domain). For a sentence from outside domain m could be
truem (taking value 1 in the mth place). Being truem, it may naturally be
counted as truthm-apt despite the pluralist’s typical presentation of truthm-
aptness as reserved for sentences in domain m. If, on the other hand, it
counts as neither truem nor truthm-apt, despite receiving value 1 at the
corresponding place of the semantic value, then that renders those values
mystifying, standing in, as they do, sometimes for truth of the appropri-
ate kind and sometimes for something else entirely. But to focus specifi-
cally on validity within a particular domain, an argument would then
count as valid within that domain—validm—iff necessarily if the premises
and conclusion are all within the domain and the premises are true, so is
the conclusion. If we preserved Cotnoir’s framework in conjunction with
this, the very same argument could count as validm but not validSP, that is,
according to the strong pluralist’s notion of validity used for mixed infer-
ences. Whether an argument is validm would depend on its content not
just its form; we cannot straightforwardly say that not-not-p so p is validm
as it is only validm if p is from the right domain. This consequence and the
general structure outlined here is not dependent on the Cotnoir frame-
work, so I will return to it below.
To summarise, I have found Cotnoir’s framework to be unsuccessful
and suggest one moral is that you shouldn’t try to capture something not
being truth-apt with a semantic value at all. The algebraic approach
requires us to make sense of question whether a truth predicate applies
outside its domain: some value has to be assigned at that place. This does
not respect the spirit of truth pluralism, a flaw that will be shared by
other algebraic approach that models validity by considering behavior
over all domains in this way. I have examined Cotnoir’s position because
it provides an unusual attempt to model the logic associated with plural-
ism about truth in such a way as to accommodate arguments that are not
purely within one domain and also to allow for different logics for differ-
ent domains. I turn next to some more general problems for
“domain-specific logical pluralism”, whether or not that is seen as arising
from pluralism about truth.
4 Domains and Their Logics

As we have seen, DLP is the kind of logical pluralism that we might
expect to be combined with pluralism about truth given the key role for
different domains for truth pluralism. This form of logical pluralism
could also be defended without commitment to pluralism about truth,
however. For you could think that different logics govern different
domains even though it is the same, univocal truth property which is
involved: in different domains, different rules for necessary truth-
preservation may apply. This kind of position might be appealing when
we consider the range of different phenomena that have prompted phi-
losophers to seek nonclassical logics, and we reflect on the difficulties of
trying to confront them all at once. Consider, for example, future contin-
gents, vagueness, discourse involving empty names, quantum behavior
and statements about what is funny: these have all been prompts for
adopting a nonclassical logic of some type. The debates on these indi-
vidual issues are messy enough on their own and it looks desirable to
avoid having to accommodate all these phenomena at once within the
single all-encompassing logic. DLP seems to offer us a way to carve off
the different problems and treat them individually. This could also vindi-
cate the common practice of ignoring some problematic phenomena—
for example, ignoring vagueness—for if we isolate off a particular domain,
we don’t need to worry about the logic specific to that domain when
considering arguments outside it.
The idea of different logics for different domains is often ignored in
debates about logical pluralism that aren’t linked in to pluralism about
truth.16 Field, for example, argues that the most interesting versions of
logical pluralism will maintain that several “all-purpose logics” can all be
right and, as such, don’t really disagree, despite appearances. The idea of
an “all-purpose logic”, as encapsulated in its name, is a logic applicable
across all domains and all contexts and in the light of all messy phenom-
ena. As Field emphasises, a defender of a particular all-purpose logic may
have some role for other logics without them capturing logical conse-
quence, perhaps because extra premises hold. For example, he explains
how Putnam’s quantum logic could fit the bill, and yet in dealing with
macroscopic objects, classical logic still be appropriately used because it
440 R. Keefe
holds to such a high degree of approximation in such contexts.17 I agree

that this kind of usability of a logic in a given domain—an enhanced
consequence relation—can be explained in ways that are compatible with
commitment to a logical monism involving a different (all-purpose)
logic: that two different logic “apply” in this kind of way in two different
domains does not demonstrate the rightness or viability of DLP and we
need this to be reflected in the characterisation of DLP. But, even if a
logical pluralism advocating several different all-purpose logics would be
particularly striking, domain-specific logical pluralism could still be a
very interesting position. If the one and only logic for some domain is
different from that of another domain (so that in each case it is not merely
that a different logic can be used because it makes for a good approxima-
tion or because we can reasonably make certain additional assumptions),
that will show that there can be no all-purpose-logic, which would be a
very significant conclusion worthy of the title of “logical pluralism”.
A bundle of central problems for domain-specific logical pluralism
surround the individuation of domains which I will not discuss in any
detail here.18 There are general problems in marking out domains. Most
arguments concerning moral matters will involve factual (non-moral)
claims: should those claims thereby fall into the moral domain after all?
It leaves us with even less grasp of the idea of such a domain if “this cat is
wet” in Tappolet’s example above is counted as falling in the moral
domain. The alternative of regarding those arguments as cross-domain
and thus not necessarily subject to the logical laws of morality threatens
to leave very few arguments within the moral domain.
Whereas domains such as mathematics, morality and perhaps humor
might be thought to be individuated by their subject-matter, logically
challenging phenomena such as vagueness, future contingents and empty
names surely aren’t unified by their subject-matter. Vague language can be
about anything, for example, so the phenomenon and its logical reper-
cussions cuts across several domains. Moreover, staying outside the
domain of vagueness would need to be a matter of more than just a voiding
sentences that are actually borderline—for example, sticking to premises
and conclusions that happen to be definitely true or false—as assessment
of validity requires consideration of all possible valuations and a clear
nonborderline sentence is often possibly vague, thereby impacting on the
logic of the argument. So, we should not assume that the subject-matter
of an argument can always determine the right logic, as phenomena such
as vagueness can also impact on the argument while cutting across sub-
ject-matters. And the logical relevance of a phenomenon such as vague-
ness is even harder to isolate (in the absence of a specific subject-matter),
since the possible vagueness of a nonborderline sentence is still relevant
to the assessment of validity.
Even if we found a way to isolate sentences within particular domains,
we would again face the problem that typical arguments involving those
sentences also involve sentences from outside those domains. Might there
then be another logic governing cross-domain argument? We have seen
above that treatments of “mixed inferences” can offer a logic that differs
from the logics of other domains.19 Should we then call this a new
domain—the domain of mixed inferences—allowing that sentences are
in this domain as well as in their more exclusive domain? After all, the
idea of DLP was that different logics could be upheld because they were
limited to different domains, so a mixed inference logic suggests a mixed
inference domain. If so, that would be an extreme example of a domain
with no unity of subject-matter. If that domain contains all sentences—
since all sentences could be part of some mixed inference in combination
with sentences from any other domain—then it looks to yield a universal
logic. Other logics will then just look like the kind of enhanced conse-
quence relations described above.20 In short, we need a logic of mixed
inferences since arguments needn’t remain within a single pure domain.
But such a logic threatens to undermine the pluralism of the other so-
called consequence relations since it has a universality.
We can frame this as we did earlier (Sect. 3). Take an argument, A, B
so C, whose premises and conclusion fall within domain m. We then
assess it as validm since it is validated by the enhanced rules specific to that
domain. But the argument is also in the broader domain, SP, and that
same argument is not validSP, since not validated by the weaker rules.21 If
we regard both as genuine logical consequence relations, we have a logical
pluralism beyond domain-specific logical pluralism, since the very same
argument is both valid and not valid. The earlier discussion suggests the
natural move of identifying genuine validity with the weaker consequence
relation. Lynch, however, suggests that we should regard A, B so C as
442 R. Keefe
valid simpliciter because there is a legitimate notion of validity with

respect to which it is valid. If truthm is the notion of truth at issue in rela-
tion to A, B and C, why not take that argument to be valid simpliciter if
it necessarily preserves that property of truth? I respond that the notion
of truth relevant to assessing the argument may not be settled relative to
the narrowest interpretation of the argument itself. Even if all the prem-
ises and conclusions are from within a particular domain, we may want
to consider whether adding an extra premise to that argument would
preserve validity, and since the extra premise could be from outside the
domain, we should think of a broader notion of validity as key. Moreover,
if the argument is of a form that is not generally necessarily truth-
preserving, then this may be enough to prevent it qualifying as a valid
argument.
There will be principles that hold in a domain (e.g. bivalence) that look
logical but must be taken to hold for some other reason. Lynch objects
that bivalence would then hold in some realm for some non-logical reason
and it isn’t clear what kind of reason that could be. But there can be a
range of very general features of a domain that needn’t be logical. They
can be metaphysical or, for example, reflect the fact that there is no scope
of vagueness within that domain. Not all arguments that necessarily pre-
serve truth can count as genuine logical consequence relations. “Jon is a
bachelor, therefore Jon is unmarried” necessarily preserves truth but there
is no space to reflect this in the true logical consequence relation. It is
typically recognised that the logical consequence relation must be formal
and an argument such as this is surely not valid “in virtue of form”, but
rather because of the special meaning of the terms “bachelor” and “unmar-
ried”. Even if integral to the notion of truth relevant to the domain,
principles that hold in a domain in virtue of features specific to that
domain threaten to fail tests of formality.22
Finally, I turn to a related, more general, simple worry about domain-
specific logical pluralism. One of the central features of logic is often
taken to be its topic neutrality. A logic that is confined to a single domain
cannot have that neutrality. Although that objection may seem simplistic
(can’t we just reject that characterisation of logic as wrong?), it ties in with
the worries about pinning down domains. For an argument to be valid, it
must be the case that on no interpretation of the non-logical vocabulary
are the premises all true and the conclusion false. Within DLP, there must
be a restriction to interpretations within the domain; but if the domain is
not clearly demarcated, the test cannot be adequately applied. Suppose,
for example, that we hope to put aside vagueness and focus on the “pre-
cise” domain. Any sentence could be vague, at least as far as its structure
is concerned. So, even if the premises are precise, the consideration of all
possible interpretations will include some vague ones. We thus would
need to limit the interpretations quantified over for the test to the precise
ones. But then we can only reach a conclusion about what follows from
what on the assumption of necessary precision of the relevant elements of
the argument. Since this is clearly a false assumption, the suggestion that
this captures the genuine consequence relation for that domain looks
questionable. When the domains are demarcated by subject-matter (e.g.
morality or humor), then the limitation to interpretations within the
domain may be less problematic than the previous case. But this will still
only do if we can isolate the sentences in that domain in such a way as to
encompass sentences involved in typical arguments concerning that sub-
ject-matter; as indicated above, this is often not easy. Topic neutrality
allows for the kind of generality required.
I have argued that the claim that the right logic varies across different
domains is not sustainable, whether in a version stemming from plural-
ism about truth or not. It is still open to the logical pluralist to reject
topic neutrality as a requirement on logic,23 but I put aside that option
and turn, in the next section, to consider a model whereby different rules
of reasoning are permitted in different domains or contexts, where this
can be superimposed over a universal (and topic-neutral) logic.
5 Relative Validity, Contexts and Domains

This section considers a treatment of arguments from different domains
that adopts the suggestion that different rules of reasoning are called
upon and justified in different contexts. Timothy Smiley offers an account
of relative validity which tells an appealing story about formalising argu-
ments with suppressed rules, according to which arguments are modelled
as valid relative to rules that are justified in the context.24 I will suggest
that this account may be of use in tackling issues considered above.
444 R. Keefe
Just as there are contexts when a reasoner is justified in assuming some-

thing which can act as a suppressed premise of their argument, so a rea-
soner can sometimes be justified in using a specific rule which—though
not strictly valid—needs no further justification in the context. Smiley
emphasises that a formalisation of someone’s everyday argument as
involving a suppressed rule can be better than one diagnosing a sup-
pressed premise. The push to do the latter instead can have the artificial-
ity of traditional logician’s attempt to regard arguments as disguised
syllogisms. The contextually justified rule can, he maintains, be a casual
ad hoc one specific to the particular circumstances (e.g. “it’s Tuesday, so
this is Paris”—justified by a timetable) or it can be a formal system of
axioms or rules. We can consider the rule employed as part of system R,
which is reasonably assumed in the context, and endorse the argument,
in the context, because it is valid relative to system R. If R contains a rule
that is not itself necessarily truth-preserving, then an argument can be
valid relative to R, without being strictly valid.
The key notions are spelled out as follows. The premises of an argu-
ment imply the conclusion by R—that is, the argument is valid relative
to R—iff there is no way to falsify that inference without falsifying R. An
inference counts as being falsified if there is an assignment of Ts and Fs to
sentences such that the premises are assigned T and the conclusion is
assigned F. A rule is falsified if some instance of it is falsified. When con-
sidering what can be falsified, “the only restriction [on the assignment of
Ts and Fs] is that all occurrences of the same sentence should be assigned
the same truth-value” (Smiley 1995, p. 730).
The strategies for assessing an argument explicitly stated as merely “A,
so B” are then the suppressed premise strategy (SPS) and the suppressed rule
strategy (SRS). SPS requires finding a missing premise, P, which is true,
needs no further justification in the circumstances and is such that “A, P;
so B” is valid. SRS involves finding a rule, R, which is truth-preserving,
needs no further justification in the circumstances and which results in
“A; so B”, being valid by R. (It is worth noting that system R includes
rules and axioms, so SPS is a special case of SRS.) As an example, take
“Alan is taller than Benji, Benji is taller than Carmine; so Alan is taller
than Carmine”. SRS detects rule “from X is taller than Y and Y is taller
than Z, infer X is taller than Z”. SPS needs to call upon, for example, “for
all X, Y and Z, if X is taller than Y and Y is taller than Z, then X is taller

than Z” and use the rule of universal instantiation and modus ponens.
Smiley claims that “the rule strategy thus takes the argument as it comes”
(p. 731) and can result in a formalisation that may seem closer to the
original intentions of the subject.
Here are three ways that relative validity and formal validity can come
apart.
1) Argument is relatively valid but falls short of being absolutely valid.
The argument uses a rule which is justified in the circumstances, but

not absolutely valid. Smiley’s case involving “it’s Tuesday, so this is Paris”
is an example of this that is dependent on a casual rule justified in quite
specific circumstances. As we will explore below, this category of argu-
ments can include cases where rules are justified given the domain over
which we are operating.
2) Argument is relatively valid and absolutely valid (i.e. necessarily truth-

preserving), but not formally valid (by the chosen system).
The “taller than” case above is an example of this.
3) The argument involves a notion of relative validity that is stronger

than the notion of validity assumed by the classical logician.
This is a possibility given the definition of falsifying a rule—restricted

only by the requirement that all occurrences of the same sentence are
assigned the same truth-value—such that for some instance of the rule,
the premises are true and the conclusion false. This allows the falsification
of, for example, “A; so A or B”, and so there could be systems relative to
which that rule is not valid. So, for example, we can consider validity rela-
tive to a relevance logic and there may be contexts in which these stricter
conditions are justified.
Should we take this to show that the genuine notion of logical conse-
quence is weaker than classical logic and, indeed, is a minimal conse-
quence relation, weaker than most standard alternatives? The natural
approach here is to identify this with the consequence relation that
446 R. Keefe
endorses only the minimal logic common to all contexts (compare the
discussion above).
To settle the correct consequence relation, however, we would need to
determine what rules are common to all contexts. Someone convinced
that classical logic is the true logical consequence relation might argue, for
example, that reductio ad absurdum is always justified, even if it is dis-
puted. Conversational participants may together have some control over
what rules are added because justified in the context (e.g. the timetable
rule), but, it might be thought, not have the same control over excluding
other rules. Arguably, however, there are contexts in which classical logic
is too strong—it is said, for example, that reasoning about an inconsistent
database requires rules of a relevance logic that block explosion, so that it
is not the case that the recording of contradictory facts on the database
warrants inferring any conclusion at all. Or take another application of a
relevant logic: Lewis argues that a paraconsistent logic is a suitable “logic
for equivocators”.25 If we cannot rule out that A is ambiguous across the
premises not-A and A or B, then we should not infer B; rather the rules of
a paraconsistent logic are better placed to guide our reasoning if we cannot
be sure we are avoiding equivocation. The classicist can respond by
explaining how these cases do not show that there is any argument genu-
inely of the classical form in question (e.g. disjunctive syllogism or reduc-
tio) that fails to be valid. For example, the database argument really needs
its premises and conclusion appended with “according to the database”, so
that what look like P and not-P are actually the non-contradictory “accord-
ing to the database P” and “according to the database not-P”. Similarly,
corresponding to the potentially equivocating English argument, there are
several different propositional contents (corresponding to the various dis-
ambiguations) and none of those invalidate disjunctive syllogism.
The alternative position which does identify logical consequence with
the very weakest common core of all candidate logics may seem truer to
the spirit of the approach to relative validity as explained above, for exam-
ple, in the attempt to “take the argument as it comes”. I will not here try
to settle the choice between these options (and others)—it would require
tackling much-disputed questions about the essential nature of logical
consequence. I have instead illustrated how Smiley’s account of relative
validity may allow for a flexibility of theoretical position with which it is

combined, as well as flexibility for the everyday reasoner in what counts
as valid relative to the context they are in. We may thus see this as inde-
pendent of the question what the true logical consequence relation is, or,
indeed, what the true logical consequence relations (plural) are (since
logical pluralism may also remain a live option). For the approach may be
compatible with a form of logical pluralism that is not domain-specific:
if there is no uniquely correct logical consequence relation, we may con-
sider this story about relative validity superimposed over different logics,
where the latter variation is not a matter of variation between contexts or
over different domains.
Let us return to our earlier issues of truth pluralism and domain-
specific logical pluralism in the light of this framework. The truth plural-
ist emphasised different features of truth in different domains, where
many of these appeared logical in character. For example, in a domain
where a constructive notion of truth is needed, the rule of reductio ad
absurdum is not justified, whereas in other domains it is justified. In those
latter domains, we can take the justified system R to be one that includes
the additional rule of reductio, and arguments can then be justifiably
formalised as involving that rule and thus as valid relative to R. We could
then accommodate the difference in appropriate reasoning in different
contexts. The contexts here can be determined by the domains under
consideration. A context purely dealing with mathematics justifies rules
applicable to that domain, whereas a context concerned with attributions
of humor does not. A mixed inference will, in effect, combine two or
more domains, thus undermining the justification of any rules that are
only valid in only one of them.
The truth pluralist talks of different domains, whereas this approach to
relative validity focuses on different contexts. But we can see the latter as
more general by considering the domains relevant to a context. A single
domain, say mathematics, may be at issue in some context. This can
mean that the evaluation of the same argument (same premises, conclu-
sions and relevant domains) can be relatively valid in one context and not
in another—it is not dictated purely by the domain in which the prem-
ises and conclusion are situated, if statements from other domains may be
448 R. Keefe
relevant. The focus on contexts rather than domains is also more suitable
for dealing with phenomena such as vagueness and the thought that in
many contexts it is appropriate to ignore vagueness and reason with rules
that are legitimate in the absence of vagueness.
We can then allow that the same (weak) underlying logic governs all
domains, strictly speaking, while granting that additional rules are war-
ranted within a specific domain. Even if we maintain that these rules do
not count as logical, since not reflective of absolute validity, they share
many features of logical rules and play a comparable role. They appear a
long way down the continuum from the ad hoc timetable rule “it’s
Tuesday so this is Paris”. In reasoning and in evaluating arguments it is
often relative validity that is paramount and that is most appropriate for
formalising arguments.
Regarding pluralism about truth specifically: the role of the minimal
core logic here mirrors the approach to the mixed inference challenge
that declares an argument valid if it preserves falling under the concept of
truth (allowing for different properties of truth for different domains).
The truth pluralist would thus then need a response to the standard
objection that truth should then be identified with that property that is
necessarily preserved in validity, but if such a response can be given the
framework of relative validity could offer an additional way to approach
the variety of arguments involving the variety of types of truth.
The above context-based framework can thus be employed by the
domain-specific logical pluralist, but that framework also allows for logi-
cal pluralism that does not involve several domains. Different contexts
may demand rules of classical logic, where others only justify intuitionis-
tic rules, even when the same domain (e.g. mathematics) is at issue.26
According to Lynch, the claim that “there is more than one logic gov-
erning our reasoning” is the key claim of logical pluralism (2008, p. 132).
The framework of relative validity allows us to accommodate the idea
here without commitment to multiple logical consequence relations. For
principles and rules can govern our reasoning without their reflecting
logical consequence: other informal rules may be crucial to our reasoning
in certain contexts for pragmatic reasons and stories about how to reason
will involve more than just principles of logic.
Notes
1. Uses of logical systems can go beyond capturing logical consequence
relations. For example, it might be useful to employ a paraconsistent
logic to work with databases that may contain inconsistent data. But that
is not to say that there are true contradictions, even though it can be on
the record that p and on the record that not-p. See Keefe 2014, footnote
13.
2. See Keefe 2000, chapter 5, and on pluralism about supervaluationist
notions of consequence, see Keefe 2001. Hjortland 2012 offers the help-
ful term “intra-theoretic pluralism” for this kind of pluralism. A different
relatively modest logical pluralism is one that maintains that there is no
uniquely correct choice of logical constants and so we can construct dif-
ferent equally good logics by selecting different sets from the putative
logical constants.
3. See, for example, Lynch 2009, Wright 1992, 2013 and the volume of
papers Pedersen and Wright 2013.
4. For example, Wright maintains that there is one concept but many prop-
erties and Lynch adopts a functionalism about truth whereby different
concepts play the role.
5. Pedersen (2014, p. 262) maintains that “alethic pluralism argues that
any alethic pluralism that accepts both realist and anti-realist concep-
tions of truth “brings on a commitment to logical pluralism” and Lynch
writes: ‘The alethic functionalist … is not required to endorse …logical
pluralism. But it is likely that she will.’ (2009, p. 104).
6. Pedersen 2014, p. 260.
7. Tappolet 1997, Williamson 1994.
8. See, for example, Wright 2013, p. 133. Beall’s many-valued approach
(2000)—allowing different truth-values, all designated values, for differ-
ent predicates, and validity as preservation of designated value—argu-
ably also falls into this category of approach.
9. See Lynch 2008, p. 137.
10. Lynch 2008, p. 139 responds to the objection that the first approach he
considers—the logic of a mixed inference is the logic governing the
weakest element of it—requires “the assumption that the logics in ques-
tion can be ordered, in the sense that the stronger logics are extensions of
the weaker logics” by maintaining, without further explanation, that this
is not “an unreasonable constraint on those logics that apply only to
specific domains of inquiry”. But he also considers the above objection
450 R. Keefe
of taking the logic to be the intersection of the logics governing the ele-
ments of the inference.
11. Note that this denial of a unique property is compatible with accepting
a unified concept of truth. On some conceptions of properties (e.g. con-
cept nominalism), that would surely be enough for commitment to a
single general property of truth, in which case the distinction between
weak and strong pluralism would collapse.
12. Along the way to showing this, he shows that his set of values with con-
junction and disjunction forms a lattice with the top value taking 1 in
every place and the bottom value taking 0 in every place.
13. This illustration is an odd case of a sentence that doesn’t fall into any
domain. The Law of Excluded Middle also fails for more typical sen-
tences within some domain, as the first disjunct, and thus the second
disjunct too, will typically take value ½ in some of the places and thus
the disjunction will not take the top value (1 in all places). The only cases
where it will take the top value is when the disjuncts have 1 or 0 in each
place, which will be the rare cases of sentences simultaneously in all
domains, for example, perhaps itself a disjunction with a disjunct from
each domain.
14. “In this section, I extend the algebraic account of validity to nonclassical
domains, showing how the account can handle domains for which para-
complete, paraconsistent and intuitionistic logic seem most appropriate”
(p. 13).
15. Might it be represented instead by alternative Boolean algebra, so that
the logic is still classical, even though bivalence does not hold? If biva-
lence is a logical feature of a domain (as Lynch argues, 2008), then surely
only the two-valued Boolean algebra will be able to capture the domain.
16. Domain-specific logical pluralism is not the kind of position Beall and
Restall are interested in, for example. They would, I think, regard it as a
form of relativism, since what is valid is relative to the domain/context.
Beall and Restall’s logical pluralism turns on the different interpretations
of case in the defining principle of logical consequence (GTT): “An
argument is validx if and only if, in every casex in which the premises are
true, so is the conclusion.” Pedersen (2014) seeks to adopt their frame-
work by explaining how different notions of case could be appropriate to
different domains. But (GTT) commands a universality in its bicondi-
tional, and limiting the notion of case to a single domain does not pro-
vide a “precisification” of “case” in the manner required for Beall and
Restall’s framework.
17. Field 2009, p. 344.
18. Issues about domains are addressed in Lynch 2009 and Wyatt 2013,
among many other places.
19. The arguments below do not require that the logic of mixed inferences is
distinct from all of the logics of the pure domains—it could, rather, be
the weakest, for example, intuitionistic logic if the domain-specific
options are just classical logic and intuitionistic logic. The same ques-
tions arise as to the status of the stronger logics.
20. Perhaps, then, we should deny that there is any mixed domain—the
domain-specific logical pluralist could consider the logic governing
mixed inferences (the intersection of the logics of the pure domains)
without requiring a corresponding domain. Different logics may then
result from different domains either by being logics of those different
domains or resulting from the interaction of different domains in a more
complex way. Whether there is a mixed domain or not, the argument
regarding the universality of the logic of mixed domains still holds.
21. Or, if SP is not strictly a domain, then the assessment via the most gen-
eral logic still applies.
22. Beall and Restall (2006), for example, identify the settled core features of
logical consequence relations as necessity, normativity and formality: to
qualify as one of the genuine consequence relations of their logical plu-
ralism, a relation must have these features. Although details of Beall and
Restall’s “settled core” are controversial, a requirement of formality is
widely accepted and the domain-specific relations in question meet none
of the candidate more detailed specifications of the formality criterion.
23. See, for example, Shapiro 2014 for a view taking this line, focusing on a
range of fruitful mathematical theories that use different logics (e.g.
intuitionistic analysis).
24. Smiley 1995. See also Keefe 2010 for further discussion of Smiley’s
account of relative validity.
25. Lewis 1982.
26. For other recent papers connecting logical pluralism with ideas about
context, see Caret 2017 and Simard Smith 2018.
References
Beall, J. 2000. On Mixed Inferences and Pluralism About Truth Predicates. The
Beall, J., and G. Restall. 2006. Logical Pluralism. Oxford: Clarendon Press.
452 R. Keefe
Caret, C. 2017. The Collapse of Logical Pluralism Has Been Greatly Exaggerated.
Research 86: 563–579.
Field, H. 2009. Pluralism in Logic. Review of Symbolic Logic 2: 342–359.
Hjortland, O.T. 2012. Logical Pluralism, Meaning-Variance, and Verbal
Keefe, R. 2000. Theories of Vagueness. Cambridge: Cambridge University Press.
———. 2001. Supervaluationism and Validity. Philosophical Topics 28: 93–105.
———. 2010. Relative Validity and Vagueness. In The Force of Argument: Essays
in Honor of Timothy Smiley, ed. J. Lear and A. Oliver, 127–143. London:
Routledge.
———. 2014. What Logical Pluralism Cannot Be. Synthese 191: 1375–1390.
Lewis, D.K. 1982. Logic for Equivocators. Noûs 16: 431–441.
Lynch, M.P. 2008. Alethic Pluralism, Logical Consequence and the Universality
of Reason. Midwest Studies in Philosophy 32: 122–140.
Pedersen, Nikolaj J.L.L. 2014. Pluralism × 3: Truth, Logic, Metaphysics.
Pedersen, Nikolaj J.L.L., and C.D. Wright, eds. 2013. Truth and Pluralism.
Shapiro, S. 2014. Varieties of Logic. Oxford: Oxford University Press.
Smiley, T.J. 1995. A Tale of Two Tortoises. Mind 104: 725–736.
Smith, P. Simard. 2018. Assessment Context-Sensitive Logical Claims. Inquiry
https://doi.org/10.1080/0020174X.2017.1402698.
———. 2000. Truth Pluralism and Many-Valued Logics: A Reply to Beall. The
Williamson, T. 1994. A Critical Study of Truth and Objectivity. International
Journal of Philosophical Studies 30: 130–144.
Wright, C.J.G. 1992. Truth and Objectivity. Cambridge, MA: Harvard
University Press.
———. 2013. A Plurality of Pluralisms. In Truth and Pluralism, ed. Nikolaj
J.L.L. Pedersen and C.D. Wright, 123–156. Oxford: Oxford University
Press.
Wyatt, J. 2013. Domains, Plural Truth and Mixed Atomic Propositions.
Aletheic and Logical Pluralism
Kevin Scharp
There are two main objectives. One is to provide a unified framework for
thinking about pluralisms with respect to truth and pluralisms with
respect to logic—aletheic and logical pluralism. It is tempting to think
that there is a single underlying pluralism that gets applied in these two
different cases, but I think that is wrong. The aletheic pluralists are, for
the most, working with one model, and the logical pluralists are working
with a different model. There are exceptions in each case but the domi-
nant views are distinct. I want to try to bring out this difference and
think about which of these versions is more plausible.
The second objective is to use the version of pluralism that I think is
more plausible to describe a view on the truth paradoxes I have been
defending lately. And then use it again to formulate a position on the
paradoxes that would be a rival to the one that I have advocated. It seems
K. Scharp (*)
Arché Philosophical Research Centre, University of St Andrews,
St Andrews, UK
Department of Philosophy, University of St Andrews, St Andrews, UK
e-mail: ks70@st-andrews.ac.uk

https://doi.org/10.1007/978-3-319-98346-2_19
454 K. Scharp
like this view I will present would be a better than my own view. I defend
this rival view a bit and then argue that it is actually not a better approach
to the paradoxes.
1 Pluralisms
Roy Cook writes: “One is a pluralist about a particular phenomenon X iff
one thinks that there is more than one correct account of X” (Cook 2010:
493). Cook then defines logical pluralism as “[T]here is more than one
‘correct’, ‘best’, or ‘legitimate’ logic” (Cook 2010: 492), and Gillian
Russell offers something very similar: “[T]here is more than one correct
logic” (Russell 2016: 1).
On the other hand, aletheic pluralism has been defined by Pedersen
and Wright as “[T]here is more than one way of being true” (Pedersen
and Wright 2013: 1) and in Pedersen “There are several properties … in
virtue of which propositions are true” (Pedersen 2012: 1) and “What
property makes propositions true may vary across domains or from sub-
ject matter to subject matter” (Pedersen and Wright 2013: 2).
Given these kinds of quotations it is tempting to think that there is just
one kind of pluralism that then is applied in various ways. In a moment
it will be clear why that is a mistake.
I am not going to go through all the specific instances of logical plural-
ism or aletheic pluralism. However, under the logical pluralists, I am
thinking of Beall and Restall of course but then also those like Roy Cook
and Stewart Shapiro who think of model theories as scientific models.
And there is Gillian Russell’s version of logical pluralism and Hartry
Field’s version as well.1 Under the aletheic pluralism banner is of course
Crispin Wright’s pioneering work and then different ways of developing
it by Edwards, Lynch, Pedersen, and Ulatowski.2
2 Contexts, Domains, and Pluralisms

I want to think about the relationship between these two patches of lit-
erature: aletheic pluralism and logical pluralism. The first comparison is
that in the aletheic pluralism literature emphasizes the difference between
Aletheic and Logical Pluralism 455
moderate pluralism and strong pluralism.3 A moderate aletheic pluralist

says that, in addition to the domain-specific truth properties, there is a
truth-as-such property; a strong aletheic pluralist denies that. This differ-
ence is not as significant in the logical pluralism literature, and one of the
things that I want to do is think about applying this distinction across
board.4 But before we can do that we have to get straight on one other
more important comparison.
The second comparison is domains versus contexts. This is a crucial
distinction that gets overlooked in each literature. There is a temptation
to think that this does not really matter. However, the distinction under-
lies two fundamentally different kinds of pluralism, and this is why these
are not just two applications of a single model of pluralism.
Thinking about domains is common in the truth pluralism literature,
while thinking about context is common in the logical pluralism litera-
ture.5 It is a standard in discussions of aletheic pluralism to posit distinct
domains of discourse and assume that each proposition belongs to only
one domain and that each domain has its own truth property. Propositions
might have multiple truth properties, but not in any interesting way. For
example, there is no sense in which there are multiple incompatible stan-
dards relevant to determining whether a particular proposition is true.
Hence, it could be that ‘the earth is round’ has multiple different truth
properties. It is correspondence true and it is also coherence true, and you
could think of other kinds of truth properties it might have as well. But
because that proposition belongs to the correspondence domain, there is
no sense in which the other truth properties really make much difference.
In particular, if we were going to define truth as such, the fact that ‘the
earth is round’ has some other truth property is not going to figure into
that proposition being in the extension of truth as such. What matters is
that the proposition has correspondence truth and it belongs to the cor-
respondence domain. Call this the domain approach to pluralism.
On the other hand, in discussions of logical pluralism, it is not stan-
dard to assume that there are domains or to assume the domain specific-
ity of arguments and consequence relations. (I am going to take validity,
entailment and logical consequence to be so tightly related that if one is
a pluralist about one of them then one is going to be a pluralist about the
others.) Although there have been a number of recent discussions of
domain-specific logical pluralism, it is a new development.6 In Beall and
456 K. Scharp
Restall (2006), there is no talk about domains (other than domains of

quantification). They never say that some argument is valid because it
belongs to a particular domain and there is a specific property of logical
consequence that is appropriate for that domain. That is not the view.
Their view is a global view. The idea is that logical consequence is unset-
tled in a certain sense. It can be sharpened up in various ways. We can in
one context of discussion sharpen it in one way, just like ‘rich’ or any
other unsettled notion. We can sharpen it in various ways in different
conversations for different purposes, and that I am going to call the con-
text approach.
You might be skeptical that the domain approach and the context
approach are really two distinct approaches, so I want to say a little bit to
bring out what I think are deep differences here. One way of bringing out
the contrast is to notice that according to a common aletheic pluralism,
when ‘P is true’ is uttered, the truth property ascribed depends on the
domain of P. On a common logical pluralism, however, when ‘A is valid’
is uttered, the extension of ‘valid’ depends on the context of utterance.
That is, the interests of those engaged in the conversation.
For the pluralist appealing to domains (common with respect to truth),
the relevant standard for an ascription is determined by the targets
domain, the target being whatever is called true. The target’s domain is
what determines the standard of truth. But for the context pluralist (com-
mon with respect to logic), the relevant standard for an ascription is
determined by the utterer’s context. Those are quite different. We can use
the analogy from a contemporary epistemology to clarify this point.
In contemporary epistemology, there’s a debate that continues to rage
about the semantic features of the word ‘knows’. In this debate, one fac-
tion calls themselves ‘subject sensitive invariantists’.7 They think that
when I say ‘S knows that P’, whether S knows that P might depend on
various standards in play but if I am a subject sensitive invariantist, the
standards that are in play are the ones that are in play for S, the subject
that I am talking about—the target of my ascription.
For an epistemological contextualist, on the other hand, there might
be multiple standards in play that determine whether a particular true
belief counts as knowledge.8 For an epistemological contextualist, when
someone asserts that S knows that P, the relevant epistemic standards are
those used in the conversation, that is, in the speaker’s context of utter-
ance. The domain approach is analogous to subject sensitive invariantism
in the epistemological debate, and the context approach is analogous to
contextualism in the epistemological debate. Keeping these to views sepa-
rate has been thought to be essential to progress in epistemology.
Hopefully those working on pluralism can see the domain vs context
divide as equally important.
It is striking because these two pluralist traditions have developed
somewhat independently of one another. For the most part, the truth
pluralists have been working with one model, the domain approach, and
the logical pluralists are working with a different model, the context
approach.
With that point made, we can return to the contrast between moderate
pluralism and strong pluralism as long as we keep the underlying domain/
context distinction in mind. If we adhere to the domain approach, we
can formulate a moderate logical pluralism in the following way: argu-
ment A is valid-as-such if and only if A belongs to domain 1 and A is
valid1 or A belongs to domain 2 and A is valid2 or … However, if we
adhere to the context approach, it is not obvious how to formulate a
moderate logical pluralism. One problem is that, on a context approach,
arguments are not confined to a unique context. A single argument might
be valid in one context and invalid in another.
For what it is worth, I think the context approach is preferable because
of the implausibility of disjoint domains for truth bearers. As anyone
who is familiar with the aletheic pluralism debate can attest, the literature
is filled with attempts to solve mixing problems, and that is a direct con-
sequence of assuming the domain approach.9 If one does not assume the
domain approach and instead assumes the context approach, then these
sorts of mixing problems never arise.
For another reason to prefer the context approach, consider a common
motivation for logical pluralism: paraconsistent logic is good for data-
bases that might have inconsistent information. This motivation might
go back to one of the important papers in that this tradition is Nuel
Belnap’s ‘How a Computer Should Think’, which advocates first-degree
entailments (FDE) for consequences in a database. However, if we take
the domain approach to logical pluralism, then what becomes of this
458 K. Scharp
motivation? Imagine there is a zoologist who is keeping track of various

data about ring-tailed lemurs in a database, and assume that classical
logic is the appropriate consequence property for the domain. Then,
according to domain-specific logical pluralism, we cannot say that para-
consistent logic is the appropriate one for this database. The topic of the
database is ring-tailed lemurs. That is a problem, but it goes away on the
context approach.
Do not take these to be knockdown objections; they are not. For what
it is worth, this is my view on the matter. I think that the context approach
is better than the domain approach, but I do not think that what I have
said here is conclusive by any means. It is just an indication of why I am
using the context approach in what follows instead of the domain
approach.
3 Replacing Truth and Aletheic Pluralism

The standard way of thinking about aletheic pluralism is to use the
domain approach. Forget about that for a moment, and instead use the
context approach to be aletheic pluralist. If we pick the right truth prop-
erties—namely, the ones that I have advocated in recent work to replace
the concept of truth—then one arrives at something very much like the
view that I have been defending.10 Hence, much to my surprise, I end up
being an aletheic pluralist of this sort based on my recent work. If you
take the context not to be the context of utterance but rather the context
of assessment because I am an assessment sensitivity theorist with respect
to the truth predicate of natural language. In short, a context approach to
pluralism about truth where there are only two truth properties—being
ascending true and being descending true—and the relevant context is
the context of assessment, looks a lot like the view I have been defending
under a different guise.
Nevertheless, my account suffers from deficiencies due to the weakness
of the resulting truth predicate. If the truth predicate of natural language
has the features that I claim it has, then that truth predicate is weaker
than it would be if it had instead the features that, say, Beall thinks it
has.11 Beall thinks the truth predicate of English obeys an unrestricted
T-schema. I think it does not; I think the T-schema has exceptions.
Wouldn’t it be nice to have a view that combines the best of both?

Wouldn’t it be nice if we could have a view that includes the benefits of
the view that I advocate and the view that the deflationists advocate and
none of the costs? Perhaps a more thoroughgoing pluralism can help! In
the remainder, I attempt to formulate such a view then evaluate it with
respect to current standards in the literature on the paradoxes associated
with truth.
4 oordinated Logical and Aletheic

C
Pluralism
Approaches to the aletheic paradoxes that advocate stronger truth predi-
cates tend to get paired with weaker logics. By ‘stronger’ I do not mean
‘has a bigger extension’; by ‘stronger’ I mean ‘satisfies more principles’.
Approaches that have weaker truth predicates get paired with stronger
logics. For example, my approach posits a weaker truth predicate but it is
compatible with classical logic, which is strong.12 Beall’s approach posits
a stronger truth predicate but it is not compatible classical logic. In the
coordinated logical and aletheic pluralism view—the current proposal—
the truth predicate that is relevant in a particular context of utterance and
the logical consequence predicate that is relevant in a particular context
of utterance are coordinated in a way that avoids triviality but provides
the optimal combination that you can have. Perhaps a coordinated logi-
cal and aletheic pluralist could get the best of both worlds! Or so it might
seem.
The proposal is to incorporate logical and aletheic pluralism into a
single approach to the aletheic paradoxes. In this view, the two kinds of
pluralism are coordinated because the logical pluralism and the aletheic
pluralism do not float free of one another. Instead, in a particular context
of utterance, there is a parameter that determines both a truth property
and a logical consequence property together.
What follows is not a familiar aletheic pluralism because the relevant
truth properties are inspired by approaches to the aletheic paradoxes, not
proposals for analyzing the nature of truth.13 Moreover, what follows is
not a familiar logical pluralism because the relevant validity properties are
460 K. Scharp
inspired by approaches to the aletheic paradoxes, not proposals for ana-

lyzing the nature of logical consequence.
The non-classical approaches to the aletheic paradoxes advocate logics
that are different from the standard logics at play in the discussion over
which is the right logic. The discussion about which is the right logic
typically is about classical logic, intuitionistic logic, or relevance logic
(R—the strong one). But none of those work for non-classical approaches
to the liar and the other paradoxes. One’s logic must be far weaker than
those if one is going to get a non-classical approach to paradoxes.14 So
when I talk about logical pluralism in this proposal, I am taking the
options for your logical consequence relation are going to include a wide
range of weak logics like BX or LP or substructural logics on where tran-
sitivity fails or contraction fails. It is hard to imagine any of these paradox-
motivated non-classical logics would be appealing without thinking of
them as solutions to the paradoxes.
Likewise, truth properties under consideration in this proposal are not
correspondence, supercoherence, superassertibility; none of that. Instead,
the truth properties are classical gappy truth, classical glutty truth, and
various transparent or deflationary kinds of truth. Similarly, it is hard to
imagine any of these paradox-motivated non-deflationary truth proper-
ties would be appealing without thinking of them as solutions to the
paradoxes.
Although these are not the only combinations one could include in the
coordinated pluralist view, the following are some examples:
• Classical logic and classical gappy truth

• Classical logic and classical glutty truth
• Classical logic and assessment-sensitive truth
• Weakly classical logic and revision theoretic truth
• Paracomplete logic and deflationary truth
• Paraconsistent logic and deflationary truth
• Non-transitive substructural logic and deflationary truth
• Non-contractive substructural logic and deflationary truth15, 16
In each case, ‘deflationary truth’ refers to a truth property that satisfies the
T-schema, where the biconditional involved in the T-sentences is the
obvious one from the logic in question (e.g., in paracomplete logic, the
T-sentences have the paracomplete biconditional; in BX, they have the
BX biconditional; etc.).
This coordinated pluralist theory being presented entails that all ale-
theic and logical expressions of natural language designate properties.
Which property they designate depends on some parameter or variable. I
use ‘context’ as a general term for what determines this parameter. One
might be more specific about this model in various ways. For example,
one might take the context to be a context of utterance and take it to deter-
mine the contents of the word ‘true’ and the logical expressions (e.g., ‘and’,
‘or’, ‘if and only if ’, ‘valid’, ‘entails’). This would be a standard contextual-
ist view. Instead, one might take the context to be a context of assessment
and take it to determine only the extensions of the truth predicate and the
logical expressions. This would be a semantic relativist view.17 Other vari-
ants are possible as well (e.g., non-indexical contextualism, and mixing
options for distinct expressions).
5 Discussion
Objection: Context dependence is incompatible with aletheic or logical
pluralism. These views are competitors that cannot both be true.
My reply: The essence of pluralism is that something’s extension or
reference is sensitive in some way to its domain or its context or whatever.
Contextualism is one way of interpreting any pluralism that utilizes the
context approach. Non-indexical contextualism is another option too, as
is assessment sensitivity. Each of these is a kind of extension-sensitivity.
Hence, contextualists about truth are pluralists about truth (a truth prop-
erty for every content of ‘true’). Pluralists need not be contextualists,
obviously; they could go with some other context-based semantic theory,
the domain approach, or some other way to link the plurality of truth
properties to the single word, ‘true’.
Objection: The coordinated pluralist view is formulated as a kind of
strong pluralism. What about moderate pluralisms? Could they have any
role in a coordinated view?
462 K. Scharp
Reply: I am not sure about this. We are dealing with paradoxes, so

generalizing is risky. The issue with moderate pluralism is whether there
is an ‘-as-such’ property that generalizes over the specific properties. The
coordinated view could accept that there are truth properties that include
many individuals covered by the specific truth properties.18
One question is whether there is an extension of ‘true’ that is the union
of all the other extensions of ‘true’. If we include classical glut views
(according to which T-Out fails), there might be lots of sentences in the
extension of that truth predicate that cannot be ‘released’ (to use Beall’s
terminology). They would be harmless. And that extension of ‘true’ might
include everything that is in any other extension.
For example, consider a context where a paracomplete logic is appro-
priate and a deflationary truth property. Then excluded middle is not
going to have the deflationary truth property in that context. Excluded
middle is not compatible with a paracomplete logic and a deflationary
notion of truth (assuming that liar sentences are in the language). So it
seems to me that, at best, some contexts might have an ‘-as-such’ property
for truth, and some might have an ‘-as-such’ property for logical conse-
quence, but no context will have both. And there might very well be
reasons to think that even this is not attainable.19
Objection: What determines which logic/truth pair is determined by a
context? Wouldn’t there be just two contexts: the safe context where the
paradoxes are not a threat and the dangerous context where the paradoxes
are a threat?
My reply: It depends on how you set up the coordinated theory. There
is no context in which both truth and logic behave intuitively. That would
give us triviality. Instead, in different contexts the boundaries of truth
and logic wander around the realm of the intuitive without ever encom-
passing all of it. One might take truth/logic combinations to be deter-
mined by the interests of those in the context. For example, if I want a
reason about truths in a certain way, then I need to insure I am in a
context where the truth property obeys the principle in question. If I
want a reason in a slightly different way, I might shift contexts and have
a different extension for my truth predicate so that it obeys some other
principle. The same goes for the logic. For example, say I want to use
reasoning by cases. Then I should not be in contexts in which certain

non-classical logics are operative. Instead, say I want to use modus
ponens. I should not be in one of the contexts that determines the logic
LP, which, notoriously, does not count modus ponens valid.
One final objection: Why would the logical and aletheic features be
aligned like this? It seems like some kind of preestablished harmony (in a
pejorative sense).
My reply: Yes, it does seem rather convenient. However, the objection
is only relevant if the coordinated view is a descriptive theory. It need not
be; we can be conceptual engineers, who offer prescriptive views. We
could say: forget about the way our language actually works and think of
this as a proposal for how to reform it.20
6 Revenge
I have presented coordinated logical and aletheic pluralism and defended
it from a couple of obvious objections. However, there is a serious prob-
lem with it. I think it ultimately fails, and it fails because it does not do
better on revenge paradox phenomena.
One of the most horrible things about working on the paradoxes
associated with truth is the revenge problem. When a theorist offers a
solution to the paradoxes, it turns out that the proposed solution itself
gives rise to new paradoxes that are structurally identical to the liar para-
dox. And these new paradoxes determine that the proposed solution is
inconsistent. The revenge phenomenon is extremely frustrating and it
affects almost everybody in this literature.21 I say ‘almost’ everybody
because I claim it does not affect my own view. Whether I am right or
not, if an approach to the aletheic paradoxes does not do better on the
revenge paradoxes, then that is a non-starter—one should not even
bother with it.
In the coordinated pluralist view just presented, there are lots of
revenge paradoxes. In particular, sentences like:
( 1) (1) is not just true

(2) (2) is Xnot true
464 K. Scharp
are associated with revenge paradoxes. Sentence (1) is a standard revenge

paradox for the paraconsistent dialetheist. Reasoning that is parallel to
the liar reasoning shows that if a paraconsistent language in question has
a sentence like (1), then it is trivial.22 Sentence (2) is a standard revenge
paradox for the paracomplete theorist (the word ‘Xnot’ expresses exclu-
sion negation, which is like kryptonite to the paracomplete theorist).
Reasoning parallel to the liar also shows that if a paracomplete language
has a sentence like (2), then it is trivial.
However, I do not think that either of these revenge paradoxes is a seri-
ous problem for the coordinated pluralist. Recall, the whole point of for-
mulating the coordinated pluralist view is to combine the best of these
approaches and avoid their faults. In the coordinated proposal, each con-
text behaves like an entire language, and the meanings of the logical
expressions and truth predicate shift around from context to context to
avoid triviality anywhere. Let context 1 be where a paraconsistent logic is
operative and context 2 be where a paracomplete logic is operative. In
context 1, there simply are no resources to construct a sentence like (1)—
the relevant expressions have contents that are determined in part so as to
avoid this kind of triviality. In particular, there is no expression that
would mean what ‘just’ would have to mean to derive a contradiction
from reasoning about (1). Likewise, in context 2, there are no resources
to construct a sentence like (2). In particular, there is no expression that
would mean what ‘Xnot’ would have to mean to derive a contradiction.
These kinds of expressive limitations are familiar ways that paraconsistent
theorists and paracomplete theorists have developed to avoid these
revenge paradoxes.
The whole point of the coordinated pluralist position is to avoid each
specific expressive limitation by confining it to a particular context. If
someone wants to use ‘just’ or ‘Xnot’, these resources are available in
certain contexts, just not contexts in which there is available an incom-
patible truth predicate (i.e., a truth predicate that obeys principles that
would lead to triviality). The coordinated pluralist can dodge worries
about these expressive limitations because the limitations are local, not
global. For any bit of vocabulary whatsoever, there is a context in which
it is available. What is not available for the coordinated pluralist (or
anyone else) are combinations of vocabularies that together lead to
triviality.
In sum, the effects of well-known revenge paradoxes from (1) and (2)
are mitigated by the coordinated pluralist view. That is exactly what it was
designed to do.
However, there are big problems. We have lots of potential contexts of
utterance and in each of them there is some non-trivial combination of
logical expressions and a truth predicate.
Or, more accurately, we have a single language with context-dependent
logical and aletheic vocabulary. In each context of utterance, these expres-
sions have particular contents, and in no context of utterance are the
contents such as to entail triviality. Or so it seems.
So far we have been ignoring the language’s ability to talk about what
goes in different contexts of utterance. Of course, we have to be able to
utilize these sorts of resources to even set up the coordinated pluralist
theory. However, once we focus on expressions like ‘context’ and ‘true in
a context’, we can see why they make trouble. For example, imagine in
Context 1, the truth predicate has the content of a classical gappy truth
predicate (i.e., it obeys T-Out, but not T-In) and the logical expressions
are classical. In Context 2, the truth predicate has the content of a defla-
tionary truth predicate and the logical expressions are paracomplete (e.g.,
Field’s logic). In a conversation with Context 1, someone might utter a
sentence:
(3) Sentence (4) is true.
Because we are dealing with multiple contexts, we can label this:

Sentence (4) is true in Context 1. Here, ‘true in Context 1’ is just another
name for the particular content ‘true’ has in Context 1.
In the same conversation, we might also get the following uttered:
(4) Sentence (4) is not true in Context 2.
Because sentence (4) is uttered in Context 1, if it has contained just a

bare truth predicate (‘true’), then this truth predicate would have had the
same content as the predicate in (3), namely, the classical gappy content.
But instead (4) has ‘true in Context 2’, which is not a classical gappy
truth predicate. It is instead a deflationary truth predicate.
466 K. Scharp
Now, still in Context 1, where classical logic is operative, someone

could reason in the following way. Assuming (4) is true, we can use the
T-schema for ‘true in Context 2’, together with facts about the identity of
sentence (4) to derive a contradiction in the usual liar style. Each of the
logical inferences is fine because Context 1 is classical.
What has gone wrong? We have used the resources needed to formu-
late the coordinated logical and aletheic pluralist theory to construct a
revenge paradox that the theory cannot handle.
In fact, we have shown how to ‘import’ a liar sentence from one con-
text where it is harmless into another context where it results in
triviality.
One might protest: just ban all terminology like ‘true in a context’!
This suggestion evidences two major confusions. First, the coordinated
theory is supposed to avoid expressive limitations like this, so this sort of
move isn’t remotely plausible given the aims of the coordinated theory.
Second, we are talking about natural language. You cannot ban parts of
natural language. If we want an approach to the paradoxes that isn’t just
a mildly interesting toy, then we need to take all the resources of natural
language seriously.
The revenge paradox associated with sentence (4) shows that, as it
stands, coordinated pluralism is trivial and so unacceptable. Thus, our
attempt to arrive at a better approach to the liar and other paradoxes
affecting truth by being more pluralist has backfired. What happened?
Why did it fail? The answer is that revenge paradoxes come in many
forms, but they all utilize some tool or feature of an attempt to solve the
aletheic paradoxes. Some revenge paradoxes depend on what the approach
says about the truth predicate. Others depend on what the approach says
about logic. Still others depend on other elements of an approach, for
example, the appeal to contexts. In the coordinated pluralist view I have
been considering, each context encapsulates a particular combination of
logic and truth. However, appealing to the tool of contexts to put all these
combinations together into a single coordinated pluralist view is its undo-
ing. The vocabulary used to describe the contexts, although not itself logi-
cal or aletheic, can be exploited by a revenge paradox. Hence, particular
revenge paradoxes for particular combinations of logic and truth predi-
cate (e.g., sentences (1) and (2)) are not much of a threat to the coordi-
nated pluralist. Instead, it is the revenge paradoxes that quantify over
contexts that pose the problem. To incorporate each one of the range of
approaches to the paradoxes into a single view, we have appealed to con-
texts. But in so doing, we expose the coordinated view to a new kind of
revenge paradox that is not a problem for any one of the particular
approaches incorporated. It is the incorporation tools (i.e., the frame-
work of contexts) that expose the coordinated view to revenge and, ulti-
mately, triviality.
Given the revenge susceptibility of coordinated logical and aletheic
pluralism, it is better to stick with an approach that does not give rise to
revenge paradoxes at all, even if that view has other costs related to the
truth predicate. The attempt to do better by combining the good aspects
of my view and the good aspects of other views like Beall’s fails.
One might object: If there are revenge paradoxes on the coordinated
view, why doesn’t my own approach fall prey to them? The reason is that
the two truth predicates that I advocate are specially designed to avoid
revenge paradoxes, and they have to be very carefully designed to avoid it.
For example, I have argued that ‘true’ is assessment-sensitive, and that
in situations where truth paradoxes abound, we ought to use new con-
cepts, ascending truth and descending truth, instead of the concept of
truth. Using these resources, we can construct sentences like
( 5) For all U, for all A, (4) is not ascending true in U from A.

(6) For all U, for all A, (5) is not descending true in U from A.
Here ‘U’ ranges over contexts of utterance and ‘A’ ranges over contexts of
assessment. It is easy to show that for all U and for all A, (5) and (6) are
ascending true in U from A and not descending true in U from A. There
is nothing paradoxical or contradictory about this result. Indeed, ascend-
ing truth and descending truth have been designed to work together
flawlessly in classical logic.23
When we appeal to contexts to coordinate truth and logic together, that
very careful design is lost, and the result is that we get the revenge paradox.
Appealing to contexts allows us to combine many logic/truth predicate
combinations. But this very mechanism introduces a way of generating
revenge paradoxes that need not be present in any of the specific contexts.
Perhaps there are revenge-immune ways of putting together pairs of logics
and truth predicates, but they don’t seem obvious if they exist.
468 K. Scharp
7 Conclusion
There are two different approaches to pluralism: the domain approach
and the context approach. I have argued against the domain approach
and in favor of the context approach. Moreover, the approach that I
advocate for the paradoxes associated with truth can be understood as a
kind of pluralism. Finally, trying to do better than my approach by merg-
ing many logic/truth combinations into a single pluralist view fails. The
problem is that the pluralist mechanism itself provides the materials for
revenge paradoxes.
Notes
1. See Cook (2002), Beall and Restall (2006), Shapiro (2006, 2014),
Russell (2008), and Field (2009).
2. Cr. Wright (1992, 2003), Lynch (2009), Pedersen (2010), Edwards
(2011), Ulatowski (2017), and the papers in Pedersen and Wright
(2013).
3. See Pedersen (2006), Cotnoir (2013b), Caret (2017), Yu (2017), and
Strollo (2018) for discussion.
4. However, see Berto (2015).
5. See Cr. Wright (1992) and Lynch (2009) for examples of aletheic plural-
ists appealing to domains, and see Beall and Restall (2006) and Shaprio
(2014) for examples of logical pluralists appealing to contexts. Note that
Beall and Restall are inconsistent on this matter: at (2006: 88) they deny
that they appeal to contexts as part of their pluralism, but they explicitly
do so throughout the book (2005, 69, 91, 94, 99, 116, 118). They also
say that there are multiple ‘senses’ of words like ‘valid’ (2005, 29), which
would suggest an ambiguity or context-dependence. See Hjortland
(2015) and Caret (2017) for discussion.
6. See Pedersen (2014) and Kouri and Shapiro (2017) for example.
7. See Fantl and McGrath (2002), Hawthorne (2003) and Stanley (2005).
8. See Cohen (1986), DeRose (1992), and Lewis (1996).
9. See Tappolet (1997).
10. See Scharp (2013).
11. Beall (2009).
12. Scharp (2013).

13. See Beall (2013b) and Cotnoir (2013a) for other examples.
14. For example, the T-Schema is inconsistent in intuitionistic logic (I) and
in the logic of relevant implication (R).
15. See Scharp (2013) for assessment-sensitive, Ripley (2014) for non-tran-
sitive, and Zardini (2011) for non-contractive. See Field (2008) for the
rest.
16. Remember that the deflationary truth predicates will have different
extensions in different logics. Each one will have exactly the logical
truths of that logic in its extension.
17. See MacFarlane (2014) on assessment sensitivity, Shapiro (2014) on
assessment-sensitive logical expressions and Scharp (2014) on assess-
ment-sensitive truth predicates.
18. See Cotnoir (2013a) for discussion.
19. Some combination views might include a classical glut property accord-
ing to which (T-Out) fails; that is, in which the move from 
to P fails. If so, we could stick a whole bunch of logical principles under
that glut property without any of them being released. If we really want
a context in which we could say that those logical principles are true, we
could do that.
20. See Cappelen (2018) and Scharp (2013) for discussion.
21. See Beall (2008), Scharp (2013, 2014), and Bacon (2015) for
discussion.
22. See Beall (2009, 2013a) and Scharp (2018).
23. See Scharp (2013: chap. 9).
References
Bacon, A. 2015. Can the Classical Logician Avoid the Revenge Paradoxes?
Philosophical Review 124 (3): 299–352.
Beall, J., ed. 2008. Revenge of the Liar. Oxford: Oxford University Press.
———. 2009. Spandrels of Truth. Oxford: Oxford University Press.
———. 2013a. Shrieking Against Gluts. Analysis 73: 438–445.
———. 2013b. Deflated Truth pluralism. In Truth and Pluralism: Current
Debates, ed. Nikolaj J.L.L. Pedersen and C.D. Wright. Oxford: Oxford
University Press.
Press.
470 K. Scharp
Berto, F. 2015. A Modality Called ‘Negation’. Mind 124: 761–793.

Cappelen, H. 2018. Fixing Language. Oxford: Oxford University Press.
Caret, C. 2017. The Collapse of Logical Pluralism has been Greatly Exaggerated.
Cohen, S. 1986. Knowledge and Context. Journal of Philosophy 83: 574–583.
Cook, R. 2002. Vagueness and Mathematical Precision. Mind 111: 225–247.
Philosophy Compass 5: 492–504.
Cotnoir, A. 2013a. Validity for Strong Pluralists. Philosophy and Phenomenological
Research 86: 563–579.
———. 2013b. Pluralism and Paradox. In Truth and Pluralism: Current Debates,
ed. Nikolaj J.L.L. Pedersen and C.D. Wright. Oxford: Oxford University
Press.
DeRose, K. 1992. Knowledge, Assertion, and Context. Philosophical Review
111: 167–203.
Edwards, D. 2011. Simplifying Aletheic Pluralism. Southern Journal of Philosophy
49: 28–48.
Fantl, J., and M. McGrath. 2002. Evidence, Pragmatics, and Justification.
Field, H. 2008. Saving Truth From Paradox. Oxford: Oxford University Press.
———. 2009. Pluralism in Logic. Review of Symbolic Logic 2: 342–359.
Hawthorne, J. 2003. Knowledge and Lotteries. Oxford: Oxford University Press.
Hjortland, O. 2015. Logical Pluralism, Meaning-Variance, and Verbal Disputes.
Australasian Journal of Philosophy 91: 355–373.
Kouri, T., and S. Shapiro. 2017. Logical Pluralism and Normativity. Inquiry.
https://doi.org/10.1080/0020174X.2017.1357495.
Lewis, D. 1996. Elusive Knowledge. Australasian Journal of Philosophy 74:
549–567.
Lynch, M. 2009. Truth as One and Many. Oxford: Oxford University Press.
MacFarlane, J. 2014. Assessment Sensitivity. Oxford: Oxford University Press.
Teach Us About Aletheic Pluralism? The Monist 89: 103–117.
———. 2010. Stabilizing Aletheic Pluralism. The Philosophical Quarterly 60
(238): 92–108.
———. 2012. Recent Work on Alethic Pluralism. Analysis 72: 588–607.
259–277.
Pedersen, Nikolaj J.L.L., and C.D. Wright, eds. 2013. Truth and Pluralism:
Current Debates. Oxford: Oxford University Press.
Ripley, D. 2014. Paradoxes and Failures of Cut. Australasian Journal of Philosophy

91: 139–164.
Russell, G. 2008. One True Logic? Journal of Philosophical Logic 37: 593–611.
———. 2016. Logical Pluralism. In The Stanford Encyclopedia of Philosophy
(Winter 2016 edition), ed. E.N. Zalta. https://plato.stanford.edu/archives/
win2016/entries/logical-pluralism/.
———. 2014. Truth, Revenge, and Internalizability. Erkenntnis 79: 597–645.
———. 2018. Shrieking in the Face of Vengeance. Analysis. https://doi.
org/10.1093/analys/anx163.
Shapiro, S. 2006. Vagueness in Context. Oxford: Oxford University Press.
Stanley, J. 2005. Knowledge and Practical Interests. Oxford: Oxford University
Press.
Strollo, A. 2018. A Simple Notion of Validity for Aletheic Pluralism. Synthese
195: 1529–1546.
Ulatowski, J. 2017. Commonsense Pluralism About Truth. New York: Palgrave
Macmillan.
Wright, C. 1992. Truth and Objectivity. Cambridge: Harvard University Press.
———. 2003. Saving the Differences. Cambridge: Harvard University Press.
Yu, A. 2017. Logic for Aletheic Pluralists. Journal of Philosophy 114: 277–302.
Zardini, E. 2011. Truth Without Contra(di)ction. Review of Symbolic Logic 4:
498–535.
Index1
B C
Beall, Jc, v, 4, 12–17, 20, 26n16, Carnap, Rudolp, 14, 218, 229,
27n22, 27n27, 65, 124n2, 230–231n2, 356
142n3, 190n7, 217–230, 238, Cases
240, 260n6, 302, 305–307, constructions, 13, 220, 305
312, 313, 321n2, 323n24, situations, 13, 219, 221, 225,
324n33, 324n34, 344, 345, 233n11, 305
366, 391, 392, 398, 402, 404n7, worlds, 13, 305, 393
425n8, 425n12, 430, 433, Circularity, epistemic, 185–188
449n8, 450n16, 451n22, 454, Closure, semantic, 310–313,
456, 458, 459, 462, 467, 468n5 318–320
Bivalence, 371, 375, 376, 378, 379, Commutativity, 15, 16, 241, 247,
384n15, 385n18, 397, 398, 262n7, 266n22
432, 433, 436, 442, 450n15 Consequence, logical, 14, 66, 102,
Blackburn, Simon, 9, 26n11, 59n14, 219, 238, 251–259, 301,
64, 66–72, 74, 76–78, 80, 331–357, 376, 388, 408, 430,
81n11–13 455
Note: Page numbers followed by ‘n’ refer to notes.

1

https://doi.org/10.1007/978-3-319-98346-2
474 Index
Content, 8, 9, 13, 21, 49, 52, 56, 161–164, 165n19, 165n21,

64–66, 69–71, 73–76, 78, 88, 165n24, 165n25, 181, 182,
96, 98–101, 116, 125n8, 214n4, 289
125n10, 143n12, 147, 150, Domain, 4, 35, 63, 85–103, 108,
158, 165n17, 187, 188, 308, 111–117, 131, 146, 189n7,
382, 392, 394, 395, 438, 446, 213, 223, 239, 289, 307, 333,
461, 464, 465 336, 365, 388, 407, 430,
Context, 20, 24, 51, 88, 101, 443–448, 454–458
148, 153, 161, 181, 182, Double-counting objection, 8, 9,
214n1, 229, 230, 238, 242, 26n11, 67, 69, 71, 74
247, 259n4, 259–260n5, Dummett, Michael, 19, 27n20, 36,
264n15, 266n22, 266n23, 48, 133, 166n30, 239, 287,
286, 288, 289, 293, 295n9, 317
322n12, 369, 384n13, 398,
430, 439, 440, 443–448,
450n16, 451n26, 458, 459, E
461–468, 468n5, 469n19 Edwards, Douglas, v, 10, 11, 81n7,
Contextualism, 27n22, 165n21, 100, 101, 103n1, 103n2,
456, 457, 461 103n4, 124n2, 124n5, 125n7,
Contraction, 15, 16, 241, 246, 313, 125n8, 135, 136, 138, 142n4,
460 143n9, 143n13, 166n43,
Cotnoir, Aaron, 17, 23, 27n27, 189–190n7, 383n2, 424n5,
81n7, 101, 124n2, 124n3, 424n6, 425n16, 454
124n5, 139, 142n4, 143n10, Epistemic constraint, 134, 372–374,
190n7, 322n9, 359n14, 378–380, 384n11
424n4–6, 425n8, 425n16, Excluded Middle, Law of (LEM),
433–438 206, 244, 245, 259n3, 261n7,
265n21, 279, 340, 342, 343,
351, 352, 384n13, 393, 398,
D 435, 450n13
Data, experimental, 178 Expressivism, 8, 9, 64, 65, 68,
Deflationism, 4, 25n2, 35, 37, 71–73, 76–78, 97, 104n16,
45–47, 51, 56, 59n26, 59n30, 104n18, 142n4, 165n21
60n32, 69, 80, 141, 143n12,
173, 185, 189n6, 273n49,
389, 459 F
Dependence, 42, 43, 157, 158, 255, Folk theory, 12, 171, 173, 174, 184,
284, 461 185
Disagreement, 11, 40, 58, 59n14, Form, logical, 51, 111, 114–117,
66, 143n11, 146, 149–156, 260n5, 278, 282, 286, 288,
Index 475
290, 294n6, 296n24, 377– 230, 231n3, 234n15, 239,

379, 392, 394 241, 251, 254, 261n6, 261n7,
Functionalism, conceptual, 13 270n37, 279, 288, 290–293,
Functionalism, truth (alethic), 12, 296n26, 305, 311, 316, 347,
171–175, 185, 188, 192n29, 354, 355, 369, 376, 384n9,
449n5 396–398, 400, 421, 429,
Fundamentality, 11, 136, 139 434–437, 439, 445, 446, 448,
451n19, 458–460, 466, 467
core, 15, 199–214, 397, 448
G formal, 17, 234n15, 278, 281,
Generalism, logical, 13–14, 16, 17, 282, 285, 291, 292, 294n3,
26n16, 277 296n25, 303, 317, 319, 320,
Generalized (Generalised) Tarski’s 333
Thesis (GTT), 12–14, 16, intuitionistic, 4, 13–15, 19, 20,
26n16, 219, 232n8, 232n9, 27n21, 200, 202, 203, 213,
233n11, 238–259, 305, 306, 217–219, 229, 230, 279, 287,
344, 346, 391, 402, 404n4, 305, 340, 369, 370, 373, 375,
430, 450n16 376, 383n6, 396, 421, 429,
Grounding, 109, 111–123, 124n6, 431, 436, 450n14, 451n19,
126n12, 252–254, 257, 460, 469n14
269n34, 269n36, 272n44–46, relevant, 4, 13, 27n21, 202,
273n48, 292 217–219, 227, 229, 232n9,
234n15, 244, 245, 272n46,
305, 446
H substructural, 15, 16, 26n18, 241,
Horwich, Paul, 25n2, 26n11, 47, 246, 248, 252, 254, 256, 258,
69, 146, 164n3, 166n30, 263n11, 266n22, 267n28,
166n43, 189n6 267n29, 272n46, 273n51,
Hyperintensionality, 115–117, 384n13, 460
125n10 Lynch, Michael, vi, 4, 8, 9, 12, 19,
21, 22, 26n8, 26n19, 27n21,
27n24, 27n27, 38–40, 42, 44,
L 45, 55, 56, 58n7, 64, 65, 80,
Limitations, expressive, 17, 303, 81n5, 81n6, 86, 100, 104n17,
309–318, 464, 466 104n18, 104n21, 108,
Logic 124n3–5, 125n7, 132,
classical, 4, 13–15, 19, 20, 22, 24, 137–140, 142n4, 143n11,
104n22, 142n4, 203, 219, 143n13, 146, 147, 158, 161,
476 Index
Lynch, Michael (cont.) N

166n30, 166n43, 169, 171, Names, empty, 18, 346–349, 351,
172, 174, 175, 189n5, 355–357, 439, 440
190n7–9, 192n28, 192n32, Negation
305, 307, 324n34, 366, 367, classical, 205–206, 209, 214n4,
371, 372, 374–376, 378, 379, 218, 221, 223, 231n2
381, 382, 383n5, 384n11, intuitionistic, 16, 224, 225, 227,
388–390, 392–395, 397–399, 229, 230
403, 404n2, 404n5, 424n3, relevant, 16, 218, 221, 223–225,
424n4, 424n6, 425n8, 227, 229, 230, 231n2
425n13, 431–433, 441, 442, Nihilism, logical, 17, 18, 278, 285,
448, 449n3–5, 449n10, 301–321, 331–357
450n15, 454, 468n5 Normativity, 13, 26n16, 166n33,
166n41, 232n8, 234n15, 238,
260–261n6, 302, 318, 392,
M 451n22
Meaning, 9, 14–16, 36, 37, 47–54, of truth, 11, 162, 163, 164n3
59n14, 59n19, 67, 88,
143n12, 217–230, 243,
264n18, 294n1, 308, 319, O
324n31, 333, 334, 336, 349, Overdetermination, 18
350, 352, 353, 355–357,
360n23, 442, 464
Mixed atomics, 7, 10, 26n10, 86, P
95–100, 102, 104n17, 125n8, Paradox
409–411 liar, 20, 21, 246
Mixed compounds, 7, 10, 19, revenge, 25, 311, 463–468
26n10, 27n21, 86, 100–102, semantic, 16, 20, 21, 24, 246,
109–112, 121–123, 124n5, 257, 263n11, 271n42, 312,
391 324n33
Mixed conjunctions, 7, 10, 39, 98, Particularism, logical, 16, 17,
110, 121–123 277–294
Mixed inferences, 8, 19, 22, 24, 66, Pedersen, Nikolaj, v, 10, 21, 26n7,
387–404, 432, 433, 436, 438, 26n8, 26n19, 55, 56, 81n7,
441, 447, 448, 449n10, 81n13, 96, 100, 104n24,
451n19, 451n20 124n2–5, 125n7, 125n8,
Monotonicity, of logical 126n12, 132, 138, 139,
consequence, 15, 16, 243–245, 142n4, 143n13, 167n43,
252, 262n7, 272n46 190n7, 193n32, 367, 372,
Index 477
374, 378, 379, 381, 382, 94–97, 104n21, 104n22,

383n2, 383n5, 385n20, 389, 108–109, 117, 123, 131–142,
393, 424n3, 424n4, 424n6, 142n3, 142n4, 145–164,
425n8, 425n13, 425n16, 432, 189–190n7, 190n9, 234n17,
449n3, 449n5, 450n16, 454 365–368, 370–372, 374, 375,
Pluralism, 407–424 378, 380–382, 383n2–4,
context-based, 454–458 387–404, 407–424, 429–448,
coordinated, 24, 25, 460, 461, 453–468
463–466 virtual logical, 22, 388, 395
domain-based, 19, 20, 305, 307, Predicates, 10, 43, 67, 86, 134, 210,
308 268n30, 310, 333, 350, 368,
global logical, 289, 290 389, 417, 432, 458
interlinguistic, 12, 175–184, incomplete, 18, 350–354, 356
191n20 Price, Huw, 9, 72–78, 81n13
intralinguistic, 12, 175–184, Properties
190n14, 191n19 abundant, 90–93
local logical, 278 sparse, 90–92
logical, v, 5, 12–24, 26n19,
27n22, 103n2, 190n7,
217–230, 238–240, 244, Q
262n7, 302, 304–309, 320, Quantification, 250, 251, 257,
322n11, 331, 332, 365–368, 267n29, 306, 314–317,
370, 372, 374, 378, 380–382, 323n22, 336, 338, 343,
383n2–4, 385n19, 387–404, 384n8, 456
407–424, 430, 431, 435, unrestricted, 314–316, 318, 319
438–442, 447, 448, 449n2, Quasi-realism, 9, 63–80
449n5, 450n16, 451n22,
453–468
moderate, 6, 26n8, 36–46, 48, R
59n31, 108, 126n12, 143n10, Ramsification, 171, 173, 186, 187,
143n11, 455, 457, 461, 462 190n8, 191n25
moderate truth, 139 Realism, 9, 19, 27n20, 47, 48, 56,
normative, 149 63, 68, 132–136, 138–142,
ontological, 23, 26n6, 48, 142n4, 389, 420, 421
103n2 Reflexivity, of logical consequence,
strong truth, 10, 108–123, 138, 15, 16, 262n7
139, 143n10 Relativism, 14, 26n15, 26n16,
truth (alethic), v, 4–12, 18–24, 27n24, 181, 240, 305, 308,
27n23, 27n24, 27n27, 35, 36, 366, 431, 450n16, 461
40, 48, 55, 59n14, 63–80, assessment-sensitive, 20, 27n24
478 Index
Representation, 9, 65, 74–77, 96, descending, 21, 458, 467

171, 173, 281, 282, 284–287, -reducing, 10, 109, 110, 113,
294, 295n11, 295n14, 319, 115–121, 123
321n1, 321n2, 407, 408 reductionism, 108, 109, 113
Restall, Greg, vi, 4, 12–17, 20, replacement theory of, 21
26n16, 27n22, 190n7, superassertibility, 114, 134, 135,
217–230, 238, 241, 260n6, 138, 143n7, 155, 431, 460
302, 305–307, 321n2, 344, superwarrant, 35, 38, 39, 41, 42,
345, 366, 391, 392, 402, 45, 46, 58n6, 117, 372, 374,
425n12, 430, 433, 450n16, 379, 389–391, 393, 431, 436
451n22, 454, 456, 468n5 Truth condition
maximal, 221, 223, 224, 231n5,
232n6, 232n7
S minimal, 221–223
Scope problem, 5, 11, 146, 154–156,
158, 161–164
Self-reference, 18, 342, 346–347 U
Sher, Gila, vi, 95, 124n2, 124n5, Ulatowski, Joseph, vi, 20, 27n23,
125n8, 307, 366, 417, 143n13, 167n43, 177–182,
424n3 191n19, 193n32, 454
Singular term, 10, 86, 87, 89,
92–97, 101, 268n30, 358n5
V
Vagueness, 16, 18, 24, 232n7, 245,
T 290, 317–320, 323n27,
Topic-neutrality, 307–308 324n33, 346, 347, 350–353,
Transitivity 356, 357, 360n19, 439–443,
of grounding, 118–120 448
of logical consequence, 15, 16, Validity
401 context-relative, 443–448
Truth domain-relative, 443
ascending, 21, 458, 467
coherence, 110, 113, 155, 407,
409–411, 455 W
correspondence, 35, 48, 110, 113, Williamson, Timothy, 18, 124n5,
138, 155, 372, 374, 379, 396, 164n14, 304, 317, 331, 337,
407, 409–411, 421, 455, 460 338, 347, 424n6
Index 479
Wright, Cory, vi, 81n13, 143n13, 192n32, 242, 319, 323n27,

167n43, 170, 185–187, 324n34, 367, 371, 387, 389,
190n9, 191n24, 192n28, 404n3, 407, 420, 421,
193n32, 322n14 425n13, 431, 449n3, 454,
Wright, Crispin, vi, 4, 5, 26n7, 35, 468n2, 468n5
36, 49, 63, 64, 81n6, Wyatt, Jeremy, v, 10–12, 25n2,
81n13, 92, 93, 96, 108, 27n24, 27n27, 60n34, 81n13,
124n2, 124n4, 124n5, 95, 96, 103n3, 104n24,
125n8, 132, 134–136, 124n2, 125n7, 125n8,
138–140, 142n4, 143n7, 125n10, 143n13, 165n24,
143n10, 143n11, 143n13, 167n43, 189n6, 424n3,
150, 165n20, 167n43, 451n18

Pluralisms in Truth and Logic, Jeremy Wyatt, Nikolaj J. L. L. Pedersen, Nathan Kellen (Ed.)

Uploaded by

Document Information

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

Pluralisms in Truth and Logic, Jeremy Wyatt, Nikolaj J. L. L. Pedersen, Nathan Kellen (Ed.)

Uploaded by

Copyright:

Available Formats

Pluralisms in Truth

More information about this series at

Palgrave Innovations in Philosophy

Library of Congress Control Number: 2018963837

© The Editor(s) (if applicable) and The Author(s) 2018

Cover illustration: Bashutskyy shutterstock.com

Pluralism has made inroads into a number of areas in philosophy. This

Incheon, South Korea Jeremy Wyatt

Part I Truth 1

Truth: One or Many or Both? 35

Truth Pluralism, Quasi-Realism, and the Problem of Double-­

The Metaphysics of Domains 85

Strong Truth Pluralism107

Methodological Pluralism About Truth131

Normative Alethic Pluralism145

Truth in English and Elsewhere: An Empirically-Informed

Part II Logic 197

Core Logic: A Conspectus199

Connective Meaning in Beall and Restall’s Logical Pluralism217

Generalised Tarski’s Thesis Hits Substructure237

Varieties of Logical Consequence by Their Resistance to

Part III Connections 363

Pluralism About Pluralisms365

A Plea for Immodesty: Alethic Pluralism, Logical Pluralism,

Logic for Alethic, Logical, and Ontological Pluralists407

Pluralisms: Logic, Truth and Domain-Specificity429

Aletheic and Logical Pluralism453

Dorit Bar-On is Professor of Philosophy, Department of Philosophy, and

Gillman Payette is Postdoctoral Fellow at the University of British Columbia.

Truth in English and Elsewhere: An Empirically-Informed

© The Author(s) 2018 3

2 Truth Pluralism (Part I)

Truth pluralists are engaged in critical debates on two fronts—one exter-

theory can be plausibly applied to empirical discourse, accounting for the

(mix-comp) Mt. Everest is extended in space and Bob’s drunk driving is

(mix-inf ) If Mt. Everest is extended in space, then Bob’s drunk driv-

〈π is beautiful〉 is a mixed atomic proposition. It features a mathemati-

Dorit Bar-On and Keith Simmons’ contribution “Truth: One or

pluralist to account for the “differential appeal of realist and anti-

(CU) All beliefs and indicative statements are subject to a single

Blackburn proposes to accommodate semantic diversity by appealing to

Douglas Edwards’ contribution “The Metaphysics of Domains” pro-

This is not to say that Wright is opposed to giving truth a “non-default”

about truth. Wyatt argues that a proper defense of alethic functional-

3 Logical Pluralism (Part II)

(GTT) An argument is validx if and only if in every casex in which

Legitimacy The standard of legitimacy for cases is given by three features

Logical Functionalism The network approach to conceptual analysis can

Logical Generalism To borrow a phrase from Hartry Field, Beall and

Logical Relativism Beall and Restall’s logical pluralism involves a form of

Structural Rules and Properties It is common to distinguish between the

(Monotonicity) If Γ ⊨ ϕ, then Γ, Δ ⊨ ϕ

(Transitivity) If Γ ⊨ ϕ and Δ, ϕ ⊨ ψ, then Δ, Γ ⊨ ψ

(Contraction) If Γ, ϕ, ϕ ⊨ ψ, then Γ, ϕ ⊨ ψ

(Commutativity) If Γ, ϕ, ψ, Δ ⊨ χ, then Γ, ψ, ϕ, Δ ⊨ χ

Reflexivity says that anything is a logical consequence of itself.

Neil Tennant’s contribution “Core Logic: A Conspectus” presents an

of logical pluralism. Unlike logical generalists, domain-based logical

case with a corresponding, case-specific type of validity. If validity is

4 Connections (Part III)

The most widely explored form of truth pluralism is domain-based. The

(1) If truth is epistemically constrained in one domain and epistemically

(2) is accepted by most (domain-based) truth pluralists. Many of them take

(mod) The logic of a mixed compound or inference that involves

Incheon, South Korea Jeremy Wyatt

Part I Truth 1

Truth: One or Many or Both? 35

Truth Pluralism, Quasi-Realism, and the Problem of Double-

The Metaphysics of Domains 85

Strong Truth Pluralism107

Methodological Pluralism About Truth131

Normative Alethic Pluralism145

Part II Logic 197

Core Logic: A Conspectus199

Connective Meaning in Beall and Restall’s Logical Pluralism217

Generalised Tarski’s Thesis Hits Substructure237

Part III Connections 363

Pluralism About Pluralisms365

Logic for Alethic, Logical, and Ontological Pluralists407

Pluralisms: Logic, Truth and Domain-Specificity429

Aletheic and Logical Pluralism453

2 Truth Pluralism (Part I)

3 Logical Pluralism (Part II)

4 Connections (Part III)

We have just reflected on the potential connections between domain-

1 Introduction: Truth—the One

3 oderate Pluralism and Truth

As a result, Quine’s “subject-matter first” point cannot be the double-

3 Content, Not Truth