This series consists of carefully selected volumes of significant original papers on predefined themes, normally growing out of a conference supported by a Mind Association Major Conference Grant. The Association nominates an editor or editors for each collection, and may cooperate with other bodies in promoting conferences or other scholarly activities in connection with the preparation of particular volumes.
Director, Mind Association: Daniel Whiting Publications Officer: Eliot Michaelson
Recently published in the series
Quine: Structure and Ontology
Edited by Frederique Janssen-Lauret
In the Light of Experience
Edited by Johan Gersel, Rasmus Thybo Jensen, Morten S. Thaning, and Søren Overgaard
Evaluative Perception
Edited by Anna Bergqvist and Robert Cowan
Perceptual Ephemera
Edited by Thomas Crowther and Clare Mac Cumhaill
Common Sense in the Scottish Enlightenment
Edited by C. B. Bow
Art and Belief
Edited by Ema Sullivan-Bissett, Helen Bradley, and Paul Noordhof
The Actual and the Possible
Edited by Mark Sinclair
Thinking about the Emotions
Edited by Alix Cohen and Robert Stern
The Social and Political Philosophy of Mary Wollstonecraft
Edited by Sandrine Bergès and Alan Coffee
The Epistemic Life of Groups
Edited by Michael S. Brady and Miranda Fricker
Reality Making
Edited by Mark Jago
The Metaphysics of Relations
Edited by Anna Marmodoro and David Yates
The Language of Ontology
Edited by J. T. M. MILLER
1
Great Clarendon Street, Oxford, OX2 6DP, United Kingdom
Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries
The moral rights of the authors have been asserted
First Edition published in 2021
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above
You must not circulate this work in any other form and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data Data available
Library of Congress Control Number: 2021934452
ISBN 978–0–19–289533–2
DOI: 10.1093/oso/9780192895332.001.0001
Printed and bound in the UK by TJ Books Limited
Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.
5.
6.
8.
Acknowledgements
The papers in this volume were presented at ‘The Language of Ontology’, a conference held in September 2017 at Trinity College Dublin. I am very grateful for funding for that event from the Irish Research Council (New Foundations Award), the Mind Association (Major Conference Grant), the Trinity Arts and Social Sciences Benefactions Fund, and the Analysis Trust.
Huge thanks must, of course, also go to all the speakers and participants at that event who made it a success. I would also like to thank Una Campbell, Keith Begley, Kenneth Pearce, and Peter Larsen for their help and advice on various aspects of organizing the conference; and to Thomas Hughes, Sarah Sawyer, and Peter Momtchiloff, for their support and advice in bringing this volume together. Lastly, I would like to thank Anna Bortolan for her advice and support throughout the entire process, from initial thoughts about organizing a conference to the final production of this volume.
List of Contributors
Delia Belleri is a FCT Junior Researcher and a member of LanCog (Language, Mind and Cognition Research Group), at the Centre for Philosophy of the University of Lisbon. Her research interests include philosophy of language, metaontology, and metaphilosophy.
Matti Eklund is Chair Professor of Theoretical Philosophy at Uppsala University. His publications are mainly in metaphysics, philosophy of language, philosophy of logic, and metaethics.
Vera Flocke is an Assistant Professor in the Department of Philosophy at Indiana University, Bloomington. She studied philosophy in Berlin, Pittsburgh, and St Andrews before completing her PhD at New York University in 2019.
Eli Hirsch is Professor of Philosophy, Brandeis University. He has generally worked on metaphysics, but more recently wrote a book on epistemology, and has also been teaching courses on normative and metaethics.
Thomas Hofweber is Professor of Philosophy at the University of North Carolina at Chapel Hill. He works mostly in metaphysics, philosophy of language, and philosophy of mathematics, and is the author of Ontology and the Ambitions of Metaphysics (OUP, 2016), as well as numerous articles.
James (J. T. M.) Miller is Assistant Professor of Philosophy at Durham University. His work covers topics in metaphysics, philosophy of language, and linguistics, particularly questions at the intersection of those domains. He recently co-edited the Routledge Handbook of Metametaphysics (2020).
Friederike Moltmann is Research Director at the French Centre Nationale de la Recherche Scientifique and in recent years has been Visiting Researcher at New York University. Her research focuses on the interface between natural language semantics and philosophy, often in relation to generative syntax. One particular interest of hers is natural language ontology, the branch of metaphysics whose subject matter is the ontology implicit in natural language. She received a PhD in 1992 from MIT and taught both linguistics and philosophy at various universities in the US, the UK, France, and Italy. She is the author of Parts and Wholes in Semantics (OUP, 1997) and Abstract Objects and the Semantics of Natural Language (OUP, 2013).
Alessandro Torza is a Fellow of the Institute for Philosophical Research at the National Autonomous University of Mexico. He specializes in metaphysics and its interaction with logic and language. His published work covers such topics as modality, laws and counterfactuals, indeterminacy, metametaphysics, and assorted Lewisiana. He is editor-in-chief of Crítica: Revista Hispanoamericana de Filosofía
Richard Woodward works at the Humboldt University in Berlin. His research interests are located in metaphysics (especially modality, fictionalism, indeterminacy, and metaontology), the philosophy of language (especially vagueness, conditionals, and metasemantics), and aesthetics (photography, fiction, and the imagination).
Stephen Yablo is David W. Skinner Professor of Philosophy in the Department of Linguistics and Philosophy at MIT. He specializes in metaphysics, philosophy of math, and philosophy of mind and language.
Introduction The Language of Ontology
J. T. M. Miller
Metametaphysics and metaontology are not new domains of inquiry. For as long as there has been metaphysical theorizing, there has been debate about the scope, methods, and limits of metaphysics. However, in the last two decades, the previously often sporadic and isolated interest in metametaphysics has consolidated into a lively and growing domain which includes, but is not limited to, historical work on various key figures in the history of philosophy, the analysis of a number of foundational concepts of metaphysical inquiry, reflection on the relationship between metaphysics and science, and direct debate about whether or not metaphysics is a substantive domain.1
In recent years, it has also become common to talk about metaphysical and ontological theories as ‘languages’.2 Disputing metaphysicians are, under this way of framing the issue, engaged in a dispute about which ‘language’ is the ‘best’ (or ‘fundamental’ or ‘privileged’)—a debate about which language accurately describes reality (or whose terms ‘carve nature at its joints’ or are ‘reference magnets’). One advantage of this way of conceiving of metaphysics is that it allows the metaphysician to avoid concerns about how we ordinarily use language. For example, it does not matter whether ‘Tables exist’ is true in English as English is not the ontologically best language. What matters is whether ‘Tables exist’ is true in the best language (or ‘Ontologese’, or the language of the ontology room, or the privileged language). The job of the metaphysician therefore becomes one of working out what is true in the language of ontology—the language that only includes terms that carve nature at its joints.3
Understanding metaphysics in this way faces a number of issues. For example, it is not obvious how the variety of metaphors employed in the above characterization should be cashed out. What does it mean to say that a language is ‘privileged’, or that it ‘carves nature at its joints’?
1 See Bliss and Miller (2020) for an overview of the range of topics within contemporary metametaphysics.
2 Though there are older traditions that would accept this way of phrasing ontological talk, in my experience the recent trend towards using ‘language’ in this way often links back to Dorr (2005).
3 Or, perhaps more precisely, the language whose nonlogical terms carve nature at its joints; cf. Sider (2011).
Furthermore, the approach seems to do little to assuage concerns about our epistemic access to reality, or, put another way, our ability to know what is true in the metaphysically best language. How could we know that we are stating truths in the best language, and how do we know that we even could express truths in such a language? After all, it cannot simply be through an act of will that we come to speak Ontologese, just as much as stating that I am speaking Italian does mean that I am.
The chapters in this volume address issues related to these concerns and various others arising from a focus on the language of ontology. The chapters focus on questions about whether the nature of language itself restricts, shapes, or other wise influences the ontological theories and debates expressed in that language. It seems necessary for creatures such as us that ontological questions are expressed in language (be that natural languages such as English, French, or Swahili, or formal languages such as those of logic, mathematics, or ‘Ontologese’). The chapters in this volume (in various ways) aim to reflect on the nature of language as a way to reveal something about the nature of metaphysics itself, and potentially the substantivity (or nonsubstantivity) of ontological debates.
The centrality of questions about the nature of language when considering the substantivity of metaphysics can be seen by considering one prominent divide in the literature (and one that will arise in some form in many of the contributions to this volume). On one side are those that endorse a substantive (or metaontological realist) account of metaphysics wherein metaphysical questions and statements are substantive claims about the structure of the world. On the other are those that support a more deflationary view of metaphysics which takes those same metaphysical and ontological questions and statements as unable to live up to such rigorous demands.
To give a better sense of these positions, albeit far too briefly, consider the idea that is implicit in many, if not most, firstorder ontological disputes that certain ways of speaking, terms, or entire languages are ‘objectively better’ than others. This idea, for many, seems to imply that there could be a language that is not merely ‘better’, but is the ‘objectively best’ (or is the metaphysically ‘privileged’, or ‘fundamental’, language).
However, deflationary accounts argue that metaphysical statements are ultimately only (or merely) about our concepts, our language, or our view of the world. Metaphysics would then be about the ways that we contingently happen to speak and think about the world, not about reality itself. Such views draw inspiration from, in particular, the work of Carnap (1950, 1959) and Putnam (1975). They argue, for a variety of reasons, that there cannot be an objectively best language that accurately describes the objective similarities and differences in reality. Many would reject the idea that natural language could just happen to be the objectively best language, but the principle carries over to the more formal languages of the physicist, biologist, economist, and metaphysician, making allencompassing questions about existence, and the nature of those things that exist, nonsubstantive.
It is worth stressing, though, that for these arguments to run counter to a substantive account of metaphysics, there must be at least two of these different languages, or different sets of ontological expressions, that are expressively equal, or truthconditionally equivalent. Without this, the thesis would only be the claim that there are these different languages or sets of ontological expressions, and we could still hold that one could be privileged. The deflationist must admit of no hierarchy of expressive richness in relation to the world (as opposed to pragmatic reasons for choosing one language over another). A harmless (to substantive metaphysics) view that few would find controversial is that there are many different, potentially infinitely many, sets of ontological expressions, or languages within which to express metaphysical claims. The controversy lies in the denial of any privileged set amongst those sets of ontological expressions, and importantly, privileged in a metaphysical sense, rather than some pragmatic or other sense.
If the deflationists are right, metaphysical debates lack the significance that metaphysicians generally take them to have. Metaphysical debates are nonsubstantive, merely verbal, in some sense ‘easy’ to solve, or even nonsensical. Some of the most notable examples, and that will be discussed in the chapters in this volume, are arguments from ‘quantifier variance’ (see Hirsch 2011), ‘easy ontology’ (see Thomasson 2015) and the related neoFregean view (see Hale and Wright 2001, 2009), and (global) expressivism (see Price 2011).
This is just to give a taster of one strand of debate arising from the language of ontology. This volume brings together new work from established and emerging authors that discusses these and various related questions connected to the possibility and coherency of a privileged language, what the consequences for metaphysical debates are if we deny the existence of an objectively best language, and the relationship between natural language and any putative ‘objectively best’ language. Together, these collected essays illustrate the strong connection between the nature of language and metametaphysics, and will provide inspiration for future research on these and related topics. In what remains of this Introduction, I will briefly summarize the main themes and questions in each of the chapters.
In ‘Ontology by Stipulation’, Eli Hirsch builds upon his highly influential previous work on ‘quantifier variance’ through a new deflationary line of argument. Hirsch argues that in cases where we observe an ontological dispute, we can introduce by stipulation languages for each disputant in such a way that it allows us to express all of the facts that are relevant to that debate. In this way, Hirsch argues that those engaging in ontological disputes need not disagree about what facts hold, and explains how this is the case even in situations where we posit finegrained facts.
In ‘Are Ontological Questions Meaningless?’, Delia Belleri argues that even a Carnapian should accept that ontological statements are meaningful. Belleri argues that we can respond to a semantic critique of ontology through conceiving of Ontologese as a Carnapian framework itself. This move renders ontological
claims meaningful, even to the Carnapian. However, Belleri goes on to argue that this may not vindicate ontology in the way that most metaontological realists want, and that ontology may still face serious deflationary concerns, such as from epistemic critiques.
In ‘Collapse and the Varieties of Quantifier Variance’, Matti Eklund distinguishes between the different theses that might all be called ‘quantifier variance’, and considers the significance of these different theses. Eklund goes on to consider how these different versions of quantifier variance fare with respect to the collapse argument, and the EklundHawthorne argument. Eklund also suggests the need to distinguish between variance theses that concern actual languages, and those that concern possible languages, and considers how this distinction might affect how we interpret those variance claims.
In ‘Ontological Expressivism’, Vera Flocke outlines the thesis of ‘ontological expressivism’ as the view that ontological existence claims express noncognitive states of mind. Flocke’s aim is not to defend ontological expressivism, but to show how there is a coherent version of expressivism to be applied to ontological discourse. Under this view, speakers do not make a factual mistake by asserting the truth or falsity of ontological existence propositions, but instead express a noncognitive disposition to assess the truth of propositions by considering those worlds at which numbers exist.
In ‘Why Our Natural Languages Are Ideal Languages for Metaphysics’, Thomas Hofweber argues that, contra how they are often supposed, natural languages are perfectly matched to the facts and thus ideally suited to describe reality. Importantly, Hofweber argues that we have reason to think that natural languages perfectly fit reality without first knowing what reality is like. Central to Hofweber’s view is the argument that if internalism is true for our own talk about facts, then there cannot be ineffable facts. This means that our natural language can represent all of the facts within the domain of metaphysics, and is ideal for metaphysics.
In ‘What Counts as a “Good” Metaphysical Language?’, J. T. M. Miller focuses on the deflationist claim that there are ‘equally good’ languages for describing reality, and on how we might understand the normativity of ‘good’ in this context. Miller argues that for languages to be compared to see if one is better (or that they are equally good), they must share the same semantic purpose. After distinguishing coarse and finegrained semantic purposes for languages, Miller argues that we can identify a semantic purpose against which we can compare putative ‘privileged’ metaphysical languages, and that, contra deflationism, claims that languages are equally good relative to this purpose rely on substantive firstorder metaphysical claims.
In ‘The Questions of Ontology’, Richard Woodward responds to the recent critique of Quine’s metaontological views from Kit Fine. Woodward argues that a Quinean approach can, contra Fine, both accommodate the apparent triviality of many existential questions and secure the autonomy of ontological existence questions from ordinary language. In this way, Woodward argues that the
Quinean ontological inquiry can continue beyond the point at which we have answered existence questions in ordinary language.
In ‘What “X Does Not Exist” Says About We Who Do Exist’, Stephen Yablo takes up the question of what it means to say that something does not exist, taking as his starting point a worry about the overdetermination of why negative existentials are true, and Kripke’s discussion of negative existential statements such as Holmes does not exist. Yablo considers various ways in which we might account for the cognitive content of these sorts of sentences, in particular stressing that we need a view that allows us to adequately distinguish between why Holmes does not exist is distinct from other true negative existentials, such as Vulcan does not exist The solution, for Yablo, is to hold that singular nonexistence claims, when true, are true because of facts like: ∀x (x is not Holmes, even if Holmes exists).
In ‘Structural Pluralism’, Alessandro Torza outlines and defends the position of ‘structural pluralism’—the view that there is a plurality of primitive notions of ‘structural’. This view is particularly important in the context of Sider’s defence of the notion of structure as central to metaontological realism. Torza argues that structural pluralism can avoid various problems associated with structural monism, and leads to a novel form of ontological deflationism—pluralistic quantifier variance. However, Torza argues that this form of deflationism, whilst compatible with the possibility of nonsubstantive ontological disagreement, is more moderate than other forms of deflationism derived from quantifier variance.
In ‘Levels of Ontology and Natural Language: The Case of the Ontology of Parts and Wholes’, Friederike Moltmann argues for the need to recognize a level of ontology—‘languagedriven ontology’—that is distinct from both the ontology of fundamental entities and the ontology of ordinary objects. The argument for this level of ontology comes from a linguistic analysis of how we talk about parts and wholes. This creates a new way that ontology and language might be connected, with this languagedriven ontology being tied to the functional part of language, not the semantic part. Moltmann argues that this new level of ontology has significant consequences for our notion of unity, especially given the (in Moltmann’s view) mandatory acceptance of languagedriven ontology when we use language.
References
Bliss, R. and Miller, J. T. M. (eds.) 2020. The Routledge Handbook of Metametaphysics London: Routledge.
Carnap, R. 1950. ‘Empiricism, Semantics, and Ontology’, Revue Internationale de Philosophie, 4: 20–40. Reprinted in the Supplement to Meaning and Necessity: A Study in Semantics and Modal Logic, 1956, University of Chicago Press.
Carnap, R. 1959. ‘The Elimination of Metaphysics Through Logical Analysis of Language’, trans. A. Pap. In A. J. Ayer (ed.), Logical Positivism, 60–81. Glencoe, IL: Free Press.
Dorr, C. 2005. ‘What We Disagree About When We Disagree About Ontology’. In M. E. Kalderon (ed.), Fictionalism in Metaphysics, 234–286. Oxford: Oxford University Press.
Hale, B. and Wright, C. 2001. The Reason’s Proper Study: Essays Towards a Neo-Fregean Philosophy of Mathematics. Oxford: Oxford University Press.
Hale, B. and Wright, C. 2009. ‘The Metaontology of Abstraction’. In D. J. Chalmers, D. Manley, and R. Wasserman (eds.), Metametaphysics: New Essays on the Foundations of Ontology, 178–212. Oxford: Oxford University Press.
Hirsch, E. 2011. Quantifier Variance and Realism: Essays in Metaontology. New York: Oxford University Press.
Price, H. 2011. Naturalism Without Mirrors. New York: Oxford University Press.
Putnam, H. 1975. Mind, Language and Reality: Philosophical Papers, vol. 2. Cambridge: Cambridge University Press.
Sider, T. 2011. Writing the Book of the World. New York: Oxford University Press.
Thomasson, A. L. 2015. Ontology Made Easy. Oxford: Oxford University Press.
1 Ontology by Stipulation
Eli Hirsch
In previous work I suggested that many ontological disputes can be viewed as merely verbal, in that each side can be charitably interpreted as speaking the truth in its own language. Critics have objected that it is more plausible to view the disputants as speaking the same language, perhaps even a special philosophy-room language, sometimes called Ontologese.1 In this chapter I want to suggest a different kind of deflationary move, in a way more extreme (possibly more Carnapian) than my previous suggestion. Let’s suppose we encounter an ontological dispute between two sides, the A-side and the B-side, and we assume that they are speaking the same language so that (at least) one of them is mistaken (perhaps the common language is Ontologese). My suggestion is that we can introduce by stipulation two languages, one for each side, such that in speaking the A-side stipulated language we capture whatever facts might be expressed in the A-side’s position, and in speaking the B-side stipulated language we capture whatever facts might be expressed in the B-side’s position. In this way we get whatever facts there might be in this ontological area without risking falsehood. This can be accomplished even if we believe in fine-grained facts. Let me try to explain how this works.
1.1
Suppose that language L1 and language L2 have the same sentences (phonetically individuated) and any sentence, relative to any context of utterance, has the same truth conditions in L1 and L2 (where sameness of truth conditions is sameness of truth-value with respect to any possible world).2 I’ll then say that L1 and L2 are truth-conditionally indiscernible. And let’s say that L1 and L2 are referentially indiscernible if any expression (relative to any context of utterance) has the same
1 See especially Dorr (2005); Sider (2011), p. 171.
2 Since the ontological disputes do not directly concern either meta-level sentences or sentences attributing intentional states, let it be understood that such sentences are excluded from the definitions and principles that I formulate in this section. I think they could be accommodated, but they introduce irrelevant complications.
intension—the same reference with respect to any possible world—in L1 and L2. Now consider the following thesis:
Referential Supervenience. It is necessarily the case that if L1 and L2 are truth-conditionally indiscernible, then L1 and L2 are referentially indiscernible.
The thesis of referential supervenience makes a claim about a necessary relationship between truth and reference. It should not be taken to say anything about whether reference or truth is in some sense more fundamental.
The thesis of the inscrutability of reference entails the thesis of referential supervenience, but the converse entailment does not hold. A reference scheme for a anguage is an assignment of (n-tuples of) objects to the expressions of the language in a way that generates the correct truth conditions of the sentences of the language. It’s well known that there must be an indefinite number of different reference schemes for any language. The inscrutability thesis says that, amongst these reference schemes, there isn’t a privileged one that is the real reference relation; any reference scheme can count as the reference relation.3 Referential supervenience trivially follows, because it is necessarily the case that if L1 and L2 are truth-conditionally indiscernible, then L1 and L2 have the same reference schemes.
Inscrutability, however, is not at all entailed by referential supervenience. The latter can perfectly well allow that there is for any language a privileged reference scheme, determined perhaps by considerations of causality or property naturalness, but that the truth conditions of the sentences settle which is the privileged reference scheme for the language.
To be clear, referential supervenience does not imply that it is impossible for a sentence to belong to two languages, and to have the same truth conditions in both languages, but to contain an expression that has a different reference in the different languages. Imagine a version of English, English*, in which the word ‘seller’ refers to any buyer and the word ‘buyer’ refers to any seller (where I’m here taking sellers and buyers to be people who, respectively, succeed in selling and buying). Since the English sentence ‘There exists a seller’ is truth-conditionally equivalent to the English sentence ‘There exists a buyer’, the sentence ‘There exists a seller’ has the same truth conditions in both English and English*. But the word ‘seller’ does not have the same reference in English and English*. This is evidenced by the fact that numerous other sentences containing the word ‘seller’ do not have the same truth conditions in English and English*. That kind of evidence of a difference of reference is absent where languages are truth-conditionally indiscernible. And absent that kind of evidence, it seems difficult to make any sense of the idea that there is a difference of reference. What could explain a
3 This view seems to be held at times by Quine (1960) and Davidson (1984).
difference of reference that has no connection to truth conditions? What could such a difference amount to?
Given these clarifications, I would expect referential supervenience to be acceptable to many philosophers. And here, finally, is the thesis on which my stipulation maneuver will depend:
Propositional Supervenience It is necessarily the case that if L1 and L2 are ruth-conditionally indiscernible, then any sentence (relative to any context of utterance) expresses the same proposition in L1 and L2.
Discussions of propositions in the literature sometimes lean inward toward cognitive states and meanings-in-the-head and sometimes outwards toward non-linguistic and non-mental facts in the world. It is always the latter sense that should be understood in the present discussion, since ontological controversy, at least in the sorts of examples I want to consider, is thought by the ontologists engaged in the controversy to be aiming for the objective mind-independent facts or truths; these are not supposed to be controversies about our cognitive states. (I will say something more about this in section IV.) Propositional supervenience is the thesis that true sentences in L1 and L2 express the same objective facts, where a fact can be roughly identified with a true proposition (and any proposition might be roughly identified with a fact that may or may not obtain).
There are two main views about the nature of propositions in this objective-fact sense. On the coarse-grained view, championed by Lewis and Stalnaker, propositions are identified with truth conditions or with possible worlds. On this view, propositional supervenience is trivially correct. It follows, I think, that, on this view, the stipulation maneuver that I’m going to describe is also trivially correct. But my interest here is in satisfying philosophers who believe in fine-grained (hyperintensional) propositions.4
The fine-grained view of objective propositions has many exotic twists and spins, but one quite central version holds that a proposition is a structured item whose structure matches the structure of some sentence that expresses it, and whose constituents are the referents of the terms of the sentence.5 Fine-grained theorists who reject the letter of the structured-proposition version will often accept the following principle: if two sentences have the same logical form, and corresponding words refer (at any context of utterance) to the same referents (with respect to any world), then the sentences express the same proposition (the same fact).6 The clause ‘at any context of utterance’ is intended to leave room for
4 There is, most assuredly, nothing in this chapter to discourage skepticism about the notion of objective fine-grained propositions. My aim, however, is to argue that, even if one accepts this notion, the stipulation maneuver may succeed.
5 Seminal formulations are in Russell (1903), Kaplan (1989), and Salmon (1986).
6 An alternative to the above principle might be to hold that in order for two sentences to express the same proposition it is required not only that corresponding words, as used in any context of
arguing that co-extensional proper names or kind-names do not have the same reference in other possible contexts of utterance. (And note that the principle says ‘if’, not ‘only if’: the stated conditions are sufficient but may not be necessary.) It immediately follows from this principle that referential supervenience implies propositional supervenience, since, if L1 and L2 are referentially indiscernible, any sentence in L1 and that same sentence in L2 trivially satisfy the conditions in the principle. I would therefore expect that many fine-grained theorists who are thinking of propositions in the objective-fact sense will be attracted to propositional supervenience.
It is certainly not my aim here to elaborate, let alone defend, some specific conception of fine-grained objective propositions or facts. I’ve given a broad and flexible formulation that seems to cover many familiar versions of the conception, and from which propositional supervenience seems to follow.
It should be understood that referential and propositional supervenience do not necessarily imply that truth-conditionally indiscernible languages L1 and L2 are one and the same language. There is, first, the obvious point that some utterances are instances of speech acts that do not have truth conditions, and L1 and L2 may differ in the meanings of such utterances. But, also, there may be some kinds of hyperintensional differences in the cognitive states expressed in L1 and L2 beyond the objective propositions or facts expressed. The only assumption required in what follows is that, if L1 and L2 are truth-conditionally indiscernible, any sentence expresses the same objective proposition or fact in L1 and L2.
1.2
Let me now begin to explain the stipulation maneuver. Suppose we come across the organicist van Inwagen debating the four-dimensionalist Lewis. In the background of their dispute is a large body of uncontroversial sentences—sentences that they would both acknowledge to be outside the range of their dispute. These would include many sentences about simples, about living things, and about sets and perhaps other abstract things. One kind of uncontroversial sentence is, ‘There is a set of table-wise interrelated simples’. A controversial sentence is, ‘There is something composed of a set of table-wise interrelated simples’.
For my purposes here I’ll count a sentence S as controversial iff it satisfies one of the following three conditions: (a) there is disagreement about the truth-value of S (where disagreement might consist in one side being committed to a certain truth-value of S while the other side is agnostic); (b) there is disagreement about
utterance, have the same reference in any possible world, but also that the words stand in some putatively stronger relation of synonymy. I will here assume, however, that an account of propositions that aims outwards toward non-linguistic and non-mental facts in the world will not appeal to synonymy (insofar as the latter involves sameness of cognitive states).
whether S is contingent; or (c) S is a compound sentence one of whose components is a sentence that satisfies (a) or (b).7
This dispute between Lewis and van Inwagen satisfies the following equivalence condition: for any controversial sentence C, each disputant is prepared to specify an uncontroversial sentence U such that this disputant claims that C and U are truth-conditionally equivalent; that is, C and U, relative to any context of utterance, hold true in the same worlds. What the disputants disagree about is which controversial and uncontroversial sentences are truth-conditionally equivalent to each other.
I will now stipulate into existence two versions of English that I’ll call ‘S-Inwagen’ and ‘S-Lewis’. ‘S’ is for ‘stipulation’. These are my stipulated languages; I’m not at this point suggesting that either of them is used by van Inwagen or Lewis. It’s stipulated that in both S-Inwagen and S-Lewis the uncontroversial sentences function in the manner agreed upon by van Inwagen and Lewis. In S-Inwagen it’s stipulated that the truth conditions of the controversial sentences stand to the truth conditions of the uncontroversial sentences in just the way that is claimed by van Inwagen; and in S-Lewis it’s stipulated that the truth conditions of the controversial sentences stand to the truth conditions of the uncontroversial sentences in just the way that is claimed by Lewis. For example, in S-Inwagen it’s stipulated that the controversial sentence ‘There is something composed of a set of table-wise interrelated simples’ is truth-conditionally equivalent to the uncontroversially false sentence ‘There is a living thing that is composed of a set of table-wise interrelated simples’; and in S-Lewis it’s stipulated that the controversial sentence ‘There is something composed of a set of table-wise interrelated simples’ is truth-conditionally equivalent to the uncontroversially true sentence ‘There is a set of table-wise interrelated simples’.8
Suppose now that van Inwagen is right about which uncontroversial and controversial sentences are truth-conditionally equivalent to each other; in the language he is using his claims about the equivalences are correct. Given that van Inwagen is right in his claims about the equivalences, it follows that, since his claims about the equivalences are stipulated to be correct in my S-Inwagen language, the language he is using is truth-conditionally indiscernible from my S-Inwagen language. Therefore, if I assert in S-Inwagen the same sentences van Inwagen asserts, I am asserting the same facts he is asserting. So, if he is right, I am right, and right about the same facts he is right about. And if he is wrong (the equivalences do not hold in his language), I’m still right, since in my S-Inwagen language the equivalences hold by stipulation.9
7 A more inclusive definition of ‘controversial’ might require some changes of formulations in what follows but would not, I think, make any substantive difference.
8 I am here ignoring the possibility of gunk, which would complicate the equivalences but not alter the essential point.
9 To be more accurate, I am right on the assumption that the ‘uncontroversial sentences’ agreed on by the disputants are true. Let that assumption be understood in all that follows.
The corresponding point holds for Lewis’s position. If he is right in what he asserts about the equivalences, then my corresponding assertions in my S-Lewis language express the same facts he asserts. And if he is wrong, I am anyway right. This is for me a win-win situation. If either of these philosophers have the facts, I have those same facts in one of my languages, and I suffer no risk of error.10
Of course, I have not determined which, if either, of these philosophers is right. And I have not ruled out that they are in fact speaking different languages and both are right. But these questions, though they might understandably be of interest to the philosophers involved, do not seem to be ontological questions about the facts, about objective reality. As regards objective reality, it seems that I have by way of stipulation secured whatever facts in this area these philosophers might be asserting.
The success of the stipulation maneuver doesn’t seem threatened by the possibility that both philosophers are using Ontologese, the special philosophy-room language. If either one of them is asserting facts in Ontologese, then I am asserting those facts in one of my languages, and that language, I suppose, would have a claim to being itself Ontologese.
But, it may be objected, isn’t it a matter of philosophical importance which language is Ontologese? The facts expressed in Ontologese are supposed to be in some sense fundamental; they are supposed to be in some sense the ground of facts expressed in lesser languages. Insofar as I don’t know which, if either, of my two stipulated languages is Ontologese, the objection goes, I don’t know which facts are fundamental.
My answer is that questions about fundamentality are just further grist for the stipulative maneuver. Let’s imagine a kind of organicist O who agrees with a four-dimensionalist F that there are such things as (non-living) tables but holds that facts about the existence of non-living composite things are not fundamental, whereas F holds that such facts are fundamental. Their dispute, then, concerns sentences like ‘It is fundamental that there exists a (non-living) table’.11 I assume that O will consider this sentence (relative to any context of utterance) to be not
10 It might be objected that there cannot be a stipulation on behalf of the side that is wrong. If Lewis, say, is wrong and van Inwagen is right, then there is no structured fact that could play the role of corresponding to the structure of the allegedly true sentence ‘This table is brown’ in S-Lewis. I will suggest three answers to this objection. The third is the hardest but the one that I think goes to the heart of the matter. (1) It must be possible to stipulate truth conditions for sentences. The stipulation can at least allow ‘This table is brown’ to be true by virtue of correctly expressing a coarse-grained fact. (2) Even if it were the case that one of my stipulations does not come off, I have ‘no risk of error’ at least in the sense of ‘no risk of making any false assertion (and no risk of missing out on any facts)’. (3) What counts as a ‘structured fact’ varies from language to language, and the question whether the S-Lewis sentence ‘This table is brown’ expresses a structured fact can itself be answered by stipulation. In the S-Lewis language it is correct to say, ‘The sentence “This table is brown” expresses the structured fact that this table is brown’. There is no linguistically neutral standpoint from which to declare what are the ‘structured facts’.
11 This seems to be the sort of ontological dispute envisioned in Dorr (2005).
just false, but necessarily false, and therefore truth-conditionally equivalent to such uncontroversial impossibilities as ‘Something is not identical with itself’, whereas F will hold the sentence to be truth-conditionally equivalent to the uncontroversially true ‘There exists a (non-living) table’.12 The stipulation maneuver, therefore, applies to this dispute: in my stipulated language on behalf of O, I can without risk of error assert whatever facts O may be asserting; and in my stipulated language on behalf of F, I can without risk of error assert whatever facts F may be asserting.
Let’s look at another example related to fundamentality. Fine and Sider seem to agree that it is fundamental that there exists an electron, but they disagree about whether it is fundamental that either there exists an electron or there exists an electron.13 I take it that Fine claims that such disjunctions are necessarily non-fundamental; so Fine holds that the sentence ‘It is fundamental that either there exists an electron or there exists an electron’ is truth-conditionally equivalent to some uncontroversial impossibility. Sider, on the other hand, presumably holds that the sentence is truth-conditionally equivalent to the uncontroversial sentence ‘It is fundamental that there exists an electron’. I can therefore easily stipulate that in my S-Fine language the equivalence held by Fine holds, whereas in my S-Sider language the equivalence held by Sider holds. Although the details would have to be carefully examined, I think it is plausible to suppose that the same maneuver can work for any other dispute between Fine and Sider about what is fundamental. Then, if I assert in my S-Fine language the sentences Fine asserts and I assert in my S-Sider language the sentences Sider asserts, I have asserted whatever facts about fundamentality these philosophers might have asserted; and I do so without any risk of error. Bringing fundamentality into the story does not clearly change anything.
A different kind of example concerns issues of modal metaphysics. In Williamson’s book on that topic he defends the doctrine of necessitism, which says that whatever exists, exists by necessity.14 Kripke, throughout all of his work, rejects necessitism.15 Here the stipulation maneuver seems to apply straightforwardly, at least on the plausible assumption that Williamson presents his doctrine as being necessarily true and Kripke rejects it as being necessarily false. In my S-Williamson language the doctrine of necessitism is stipulated to be truth-conditionally equivalent to an uncontroversial necessity, whereas in my S-Kripke language the doctrine is stipulated to be truth-conditionally equivalent to an uncontroversial impossibility. My risk-free assertions in these languages will then deliver whatever facts there might be about necessitism.
12 In order for ‘p’ to be truth-conditionally equivalent to ‘q’, it is not enough that the sentence, ‘Necessarily p iff q’ is true; it must be true relative to any possible context of utterance. In what follows, for ease of exposition, I will often omit mention of contexts of utterance.
13 Sider (2011), pp. 148ff.; Fine (2013). 14 Williamson (2013).
15 For example, Kripke (1980).
It must be understood that the stipulation maneuver is meant to undermine the significance of certain metaphysical disputes; it does not undermine the significance of any statements made on either side of the dispute, let alone any of the uncontroversial statements. (Each stipulated language may in some sense embody a different and potentially interesting ‘framework’.) Moreover, it is meant to apply to certain actual disputes that have occurred within philosophy. I claim, for example, that the particular dispute between Sider and Fine about disjunctive facts is vulnerable to the stipulation maneuver. And I suggest that any other dispute between them about fundamentality can in all likelihood be treated in the same way. And, I would say, the same holds for disputes about fundamentality between Fine and Schaffer or between Sider and Schaffer, or between the various other philosophers who have been arguing about fundamentality. But I am not able to rule out the possibility that there could be a dispute about fundamentality that cannot be treated in this way. Everything depends on there being uncontroversial sentences in terms of which each side can stipulate the equivalences required to construct a language for the other side. It is most emphatically not being claimed that the concept of fundamentality can be analyzed or defined in other terms. The same points hold for all of the other examples I discuss in this chapter.
In all of these examples I could have directly stipulated that in the S-languages certain sentences have certain truth conditions, instead of doing this indirectly by stipulating that certain equivalences hold. It’s seems, however, a bit easier to start out with certain uncontroversial sentences whose truth conditions are taken as understood, and then to stipulate that certain other sentences are truth-conditionally equivalent to the ones whose truth conditions are understood. Of course, there are an indefinite number of relevant equivalences, and it’s not possible to simply list them all in introducing my different S-language. I’m not going to take that problem very seriously, however. The equivalences can be roughly indicated, as I have done. That should be enough, I think, to introduce the different S-languages. Alternatively, one can think of the equivalences as not being directly stipulated, but rather as deriving from other stipulations. In the example of the dispute between organicists and four-dimensionalists, one can arrive at the two S-languages by stipulating differences in the way the quantifier operates in the languages.16 (The S-Inwagen and S-Lewis languages will certainly exhibit some form of quantifier variance regardless of where the focus of the stipulations is.) In the examples pertaining to fundamentality the stipulation might be to different intensions of the word ‘fundamental’, and here it might be possible to give a finitary list of the different basic categories of truths that Fine or Sider count as fundamental. In the modal example the stipulation might again be
16 The stipulations might be accomplished in the manner of Dorr (2005), Hirsch (2011), pp. 234–243, and especially Hirsch and Warren (2019).
about the quantifiers. Having mentioned this problem, I will ignore it in what follows, and talk about stipulating equivalences.
Something needs to be said about the notion of stipulation itself. Certainly, stipulating that one is using language in a certain way amounts to nothing unless one actually goes on to use the language in that way. This is why the relevant stipulations could not be of the general form, for example, ‘Let the sentences of the S-Inwagen language have the truth conditions corresponding to van Inwagen’s beliefs’; the stipulations must specify certain sentences and stipulate what their truth conditions are in a manner that would allow the sentences actually to be used in accordance with those truth conditions. It might then be asked what the semantic force is of making a stipulation: one’s use of the language will determine by interpretive charity what one means, with or without the stipulation. I think, however, that stipulating specific equivalences has some significant force. For example, suppose you stipulate in the introduction to a book, ‘I will always use “fish” as equivalent to “aquatic creature that has a fish-like shape” ’. Given this stipulation, the statement ‘Whales are big fish’ is true. Considerations of reference magnetism, or of Burgean communal pressures, do not cancel the stipulation, so long as your use of ‘fish’ accords with the stipulation. If later in the book you start saying, ‘It’s necessary for a fish to have a certain kind of internal structure’, I don’t think we should say that because of ‘charity to retraction’17 you were wrong when you initially said, ‘Whales are big fish’. Rather, we should say that you have (carelessly, inadvertently) changed your language and given up the stipulation, but that your sentence ‘Whales are big fish’ was true in your previous stipulated language. I don’t know if ‘charity to stipulation’ brings something essentially new into the story besides ‘use’, but I think it can count significantly. The example I just mentioned shows that stipulation can render as ineffectual appeals (a) to future charity to retraction, (b) to reference magnetism, and (c) to Burgean considerations about the community language. In this new stipulation maneuver that I’m trying to explain, I hope to satisfy some of the people who have rejected my previous appeal to charity as showing that each side speaks the truth in its own language. These critics have often appealed to consideration of (a), (b), or (c). I think even these critics should agree that the assertions I make in one of my S-languages are true by stipulation.
If it seems shocking that I can arrive by stipulation at the same true propositions that ontologists arrive at with intuitive insights and arguments, it must be borne in mind that stipulating is not waving a magic wand that produces facts; it is rather a speech act that tilts the force of interpretive charity in a certain direction. When the ontologists assert their favored equivalences and the various ontological sentences coming out of those equivalences, there is a charitable presumption that
