Getting Things Right: Fittingness,

Reasons, and Value Conor Mchugh

Visit to download the full and correct content document:
https://ebookmass.com/product/getting-things-right-fittingness-reasons-and-value-con
or-mchugh/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Getting multi-channel distribution right Second Edition

Ailawadi

https://ebookmass.com/product/getting-multi-channel-distribution-
right-second-edition-ailawadi/

Writing for Publication in Nursing and Healthcare:

Getting it Right, 2nd Edition Karen Holland

https://ebookmass.com/product/writing-for-publication-in-nursing-
and-healthcare-getting-it-right-2nd-edition-karen-holland/

Programming the Photon: Getting Started with the

Internet of Things Rush Christopher.

https://ebookmass.com/product/programming-the-photon-getting-
started-with-the-internet-of-things-rush-christopher/

Programming the Photon: Getting Started with the

Internet of Things (Tab) Christopher Rush

https://ebookmass.com/product/programming-the-photon-getting-
started-with-the-internet-of-things-tab-christopher-rush/
A Useful History of Britain: The Politics of Getting
Things Done 1st Edition Michael Braddick

https://ebookmass.com/product/a-useful-history-of-britain-the-
politics-of-getting-things-done-1st-edition-michael-braddick/

Reasons First Mark Schroeder

https://ebookmass.com/product/reasons-first-mark-schroeder/

Metaepistemology First Edition. Edition Mchugh

https://ebookmass.com/product/metaepistemology-first-edition-
edition-mchugh/

Normativity : epistemic and practical First Edition.

Edition Mchugh

https://ebookmass.com/product/normativity-epistemic-and-
practical-first-edition-edition-mchugh/

Constructing Practical Reasons Andreas Muller

https://ebookmass.com/product/constructing-practical-reasons-
andreas-muller/
Getting Things Right
 Getting Things Right
Fittingness, Reasons, and Value

C O N O R M C H U G H A N D J O N AT H A N WAY
 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
© Conor McHugh and Jonathan Way 2022
The moral rights of the authors have been asserted
First Edition published in 2022
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: 2022937301
ISBN 978–0–19–881032–2
ebook ISBN 978–0–19–253827–7
DOI: 10.1093/oso/9780198810322.001.0001
Printed and bound in the UK by
Clays Ltd, Elcograf S.p.A.
Links to third party websites are provided by Oxford in good faith and for information only.
Oxford disclaims any responsibility for the materials contained in any third party website
referenced in this work.
 Contents

Acknowledgements
Introduction
1 Reasons
2 Fittingness First
3 Preliminaries
3.1 Strategy and Limits
3.2 Ontology of Reasons
3.3 Normativity
3.4 Constitutive Accounts
3.5 Plan of the Book

1. Reasons
1 Introduction
2 Roles and Features of Reasons: Desiderata for a
Constitutive Account
3 Some Theories of Reasons
3.1 Ought-Based Theories
3.1.1 Reasons as Explanations of Oughts
3.1.2 Reasons as Evidence of Oughts
3.2 Value-Based Theories
3.3 Primitivism
4 Reasons as Premises of Good Reasoning
4.1 The Account
4.2 Attractions
4.3 Developing the Account
 4.3.1 Conative and Epistemic Conditions
4.3.2 Objective and Subjective Reasons
5 Challenges
5.1 Shouldn’t Good Reasoning Be Understood in Terms of
Reasons?
5.2 Further Challenges
6 Conclusion

2. Good Reasoning
1 Introduction
2 Clarifications and Desiderata
3 The Fittingness View
4 Other Things Equal and Normality
5 Objections
5.1 Reasoning to the Belief that a Fittingness Condition
Obtains
5.2 Reasoning from Believing a Means is not Choiceworthy
to Believing It’s Not Necessary
5.3 Necessarily Fitting Responses, Opaque Necessary
Connections
6 Back to Reasons
6.1 First Objection: Very Weak Reasons for Action
6.2 Second Objection: Irrelevant Considerations
6.3 Third Objection: Bare Testimony
6.4 First Implication: The Guidance Condition
6.5 Second Implication: Incentives for Attitudes
7 Conclusion
Appendix 1Modelling Fittingness-Preservation
Appendix 2Structural Features of Reasons

3. Fittingness
1 Introduction
2 Examples of Fittingness
 3 Fittingness and Deontic Status
3.1 Clarifications
3.2 Four Marks
3.2.1 Insensitivity to Incentives
3.2.2 Strength
3.2.3 Absences
3.2.4 Objectivity
4 Fittingness and Reasons
4.1 Reasons Accounts of Fittingness
4.2 Incentives
4.3 Further Problems
4.4 Fittingness and the Balance of Reasons
5 What Makes Responses Fitting?
5.1 Fittingness Conditions
5.2 Fitting Belief
5.3 Fitting Intention
5.4 Fitting Desires and Emotions
6 Conclusion

4. Value
1 Introduction
2 Preliminaries
3 Attractions of Norm-Attitude Accounts
3.1 Between Primitivism and Subjectivism
3.2 Generalizing: Specific Values, Attributive Good, Good-
For
3.3 Norms and Values
4 The Wrong Kind of Reason Problem
4.1 Wrong-Kind Reasons and Buck-Passing
4.2 Wrong-Kind Reasons and the Fitting-Attitude Account
5 The Partiality Problem
5.1 The Problem
 5.2 The Buck-Passers’ Response
5.3 A Fitting-Attitude Solution to the Partiality Problem
5.4 Related Challenges
6 Conclusion

5. The Explanatory Role of Reasons I: The Weights of Reasons

1 Introduction
2 The Simple Approach
3 The Defeasibility Approach
3.1 Introducing the Approach
3.2 Weights and Defeasible Reasoning: Outweighing
3.3 Weights and Defeasible Reasoning: Generalizing
4 Placing Weight
5 Applying the Account
5.1 Simple Outweighing
5.2 Incomparability and Equal Weight
5.3 Failure of Additivity
5.4 Holism I: Attenuating and Disabling
5.5 Holism II: Intensifying and Enabling
5.6 Are Conditions and Modifiers Reasons? Are Disabled
Reasons Reasons?
6 Conclusion
6. The Explanatory Role of Reasons II: From Weights to Deontic
Status
1 Introduction
2 Overall Strength and Reasons Against
2.1 An Account of Overall Strength, and a Worry
2.2 Reasons Against as Reasons for Absences
2.2.1 Reasons for Absences
2.2.2 Absences, Fittingness, and Reasoning
2.3 Reasons Against as Reasons for Alternatives
3 Strength Amended
 4 From Overall Strength to Deontic Status
5 The Balancing Asymmetry
6 Pointless Responses
7 Reasons’ Explanatory Role
8 Conclusion

7. Reasons for Belief, Action, and Emotions

1 Introduction
2 Reasons for Belief
2.1 Perceptual Reasons
2.2 Weak Reasons, Lottery Reasons
2.3 Reasons for Credences
3 Reasons for Action
3.1 The Problem of Reasons for Action and a Solution
3.2 Two Objections
3.3 The Response Condition
4 Reasons for Emotions
4.1 Emotions and Reasoning
4.2 A Challenge to Reasons for Emotions
5 Conclusion
Conclusion

Bibliography
Index
 Acknowledgements

The ideas that were to become this book were first formulated in the
summer of 2012. During the long journey from those first inchoate
ideas, we have benefited from the feedback and encouragement of
many people, to all of whom we are very grateful. They include
Sophie Archer, Sam Asarnow, Olivia Bailey, Selim Berker, Davor
Bodrožić, John Brunero, Richard Chappell, Jean-Marie Chevalier,
Justin D’Arms, Tyler Doggett, Fabian Dorsch (1974–2017), Tom
Dougherty, Julien Dutant, Luke Elson, Pascal Engel, Davide Fassio,
Giulia Felappi, Suki Finn, Guy Fletcher, Daniel Fogal, Magnus Frei,
Jan Gertken, Javier González de Prado Salas, Marie Guillot, Grace
Helton, Frank Hoffman, Brad Hooker, Ulf Hlobil, Chris Howard,
Thomas Hurka, Benjamin Kiesewetter, Uriah Kriegel, St.John
Lambert, Dave Langlois, Woo Ram Lee, Yair Levy, Clayton Littlejohn,
Arturs Logins, Errol Lord, Susanne Mantel, Beri Marušić, Miriam
McCormick, Anne Meylan, Jonas Olson, Hille Paakkunainen, David
Plunkett, Meret Polzer, Franziska Poprawe, Joëlle Proust, Wlodek
Rabinowicz, Andrew Reisner, Susanna Rinard, Debbie Roberts, Simon
Robertson, Rich Rowland, Nils Säfström, Eva Schmidt, Thomas
Schmidt, Mark Schroeder, Kieran Setiya, David Shoemaker, Matthew
Silverstein, Martin Smith, Daniel Star, Philip Stratton-Lake, Christine
Tappolet, Naomi Thompson, Josefa Toribio, Lee Walters, Ralph
Wedgwood, and Daniel Wodak.
A version of the whole manuscript was discussed by the
Normativity Reading Group at Southampton in 2019–20, and by a
Princeton-Humboldt Graduate Seminar co-organized by Thomas
Schmidt and Michael Smith in 2021. In each case we received
extensive, searching and constructive feedback. We also received
detailed and helpful written comments on the whole manuscript
from Philip Fox and—as a then-anonymous reviewer for OUP—Justin
Snedegar. These comments all led to significant improvements.
Material from the book has been presented in various places, too
numerous to list, over the years. The book has benefited a great
deal from the helpful comments and discussion on these occasions.
There are doubtless many others besides those mentioned above
who have provided feedback. We are no less grateful to them, and
we apologize for their omission.
Special thanks are due to our Southampton colleagues Alex
Gregory, Brian McElwee, Kurt Sylvan, and Daniel Whiting, with whom
we have discussed material and issues from this book over many
years. Without such insightful and generous interlocutors we would
have produced a much worse book. More generally, our department
at Southampton has been a tremendously stimulating, supportive,
and collegial environment for this research. We are grateful to all our
colleagues for making that the case.
Our editor at OUP, Peter Momtchiloff, was supportive and patient
throughout the long development of this project. Many thanks to
him. Thanks also to Saranya Ravi, our production manager, and Atus
Mariqueo-Russell, who prepared the index.
Our research has been supported by several institutions and
funders. The Arts and Humanities Research Council funded the
project ‘Normativity: Epistemic and Practical’ (2014–16, grant
number AH/K008188/1) at the University of Southampton, during
which central parts of the research for the book took shape. We
were also lucky enough to spend periods of research leave at other
institutions. Conor McHugh spent part of 2014–15 as a visiting
researcher with the Logos group at the University of Barcelona.
Jonathan Way spent two weeks in June 2018 at the Humboldt
University of Berlin, funded by the Erasumus+ programme, and
2018–19 as a Murphy Fellow at Tulane University. We are grateful to
these institutions for these invaluable opportunities.
Parts of the book include or are based on material from the
following previously published works:
 Fittingness First. Ethics, 126, 575–606 (2016). Copyright The
University of Chicago, DOI 10.1086/684712.
What is Good Reasoning? Philosophy and Phenomenological
Research, 96, 153–174 (2018). DOI 10.1111/phpr.12299.
All Reasons Are Fundamentally for Attitudes. Journal of Ethics and
Social Philosophy, 21, 151–174 (2022). DOI
10.26556/jesp.v21i2.1341.
Finally, we each want to express thanks of a different kind to
those without whose personal support the writing of this book would
have been, at least, harder and drearier.
Conor McHugh has been helped more than she knows by Marie
Guillot’s inexhaustible kindness, patience, and encouragement. The
arrival of Éilis Guillot McHugh may not have accelerated the
completion of the book, but it certainly made things more fun. It is
Conor’s great good fortune, for which he is humbly grateful, to share
his life with them both. His greatest debt is to his parents, Brianne
and Reggie McHugh. He dedicates his contribution to this book to
them, with love.
Jonathan Way would like to thank Hannah Young. Her love and
support have made the final years of this project—a period of
lockdowns as well as the challenges of completing a book—much
easier, and very much happier. Above all, he is grateful to his family,
and especially his parents, Frances Pitt-Brooke and David Way. He
dedicates his contribution to this book to them, with love.
 Introduction

This book has two main aims. The first, narrower, aim is to develop
and defend a constitutive account of normative reasons as premises
of good reasoning. The second, broader, aim is to develop and
defend a fittingness-first account of the structure of the normative
domain, of which the account of normative reasons is a part. In this
Introduction we explain these aims and how we plan to execute
them. In §1 and §2 we explain the first and second aims
respectively. In §3 we make some points about our strategy, set out
some assumptions, and provide a plan for the rest of the book.

1 Reasons
It’s been a long day, and now it’s nearly 6 p.m. The match starts at 6
p.m. and you arranged to watch it with your friend, at the bar round
the corner. Its being nearly 6 p.m. is thus a reason for you to believe
that the match is about to start, leave the office, and be glad the
day is over.
That it’s nearly 6 p.m. is here a normative reason. In the familiar
gloss, it is a consideration that counts in favour of certain responses.
We might also say it supports or helps justify or makes a case for
these responses. Normative reasons should be distinguished from
motivating reasons: considerations for which, or in light of which, we
respond. Motivating reasons can be normative reasons—you might
decide to leave the office for the reason that it’s nearly 6 p.m. But
they needn’t be—your reason for believing that the coin will come up
heads might be that it came up tails last time, but that is no reason
to believe this. Normative reasons must also be distinguished from
explanatory reasons: considerations that explain an outcome. Your
low blood sugar levels might help explain why you snapped at your
partner, without justifying your doing so. Hereafter, by ‘reasons’ we
mean normative reasons unless we indicate otherwise.1
Reasons have had a lot of attention recently. This makes sense,
since they play significant roles. Reasons help determine deontic
status—what you ought and may do: it might be that you ought to
leave the office, despite the work you have to finish, given your
arrangement with your friend. Reasons are relevant to reasoning: in
deciding whether to leave the office, you should take into account
that it’s nearly 6 p.m., and thus that the match is about to start. And
reasons are connected to value: insofar as there’s reason to be glad
your day is over, it is good that your day is over. Understanding
reasons thus seems important for understanding all of these things.
The first aim of the book, we said, is to develop a constitutive
account of reasons. This is an account of what it is for a
consideration to be a reason.2 More precisely, it is an account of
what it is for the relation of being a reason to obtain between a
consideration (such as that it’s nearly 6 p.m.), a type of response
(such as believing that the match is about to start), and an agent
(such as you).3 We will normally express this more simply by talking
about what reasons are.
Constitutive accounts should be distinguished from substantive
accounts. A substantive account tells us which considerations are
reasons for what, and why. For instance, a broadly utilitarian
substantive account of reasons for action might say that all reasons
to act are considerations which help explain why acting in some way
would promote wellbeing. A constitutive account may have
substantive implications, but it may also leave substantive questions
largely open. Suppose, for instance, that to be a reason is to be a
consideration that helps explain why you ought to respond in a
certain way (Broome 2004, 2013). This doesn’t yet tell us which
considerations are reasons for what. It depends, among other
things, on what you ought to do. We say more about constitutive
accounts below (§3.4).
Our account says that reasons are premises of good reasoning. A
little more precisely, for a consideration to be a reason to respond in
a certain way is for that consideration to be a (potential) premise of
good reasoning to that response. For example, for the consideration
that it’s nearly 6 p.m. to be a reason to believe that the match is
about to start, and to leave the office, is for it to be good reasoning
to think ‘it’s nearly 6pm, so the match is about to start, so I’ll leave
the office’. On our account, then, the role of reasons in reasoning is
central to their nature: to be a reason is to be suited to play a
certain role in reasoning.
This requires much unpacking, which will come in Chapter 1. As
we explain there, this account has significant attractions. It offers a
unified account of reasons for different responses—most obviously,
beliefs and actions. It explains notable features of reasons, such as
why only certain things—e.g. beliefs and actions, but not headaches
or height—are subject to reasons. It allows that substantive
accounts of reasons for different responses may be very different.
And it straightforwardly explains the role that reasons play in
reasoning.
The account also raises many questions. First, good reasoning
from mistaken beliefs or intentions does not correspond to reasons.
‘It’s nearly 6 p.m., so the match is about to start’ is good reasoning
regardless of whether it’s nearly 6 p.m. But if it’s not nearly 6 p.m.—
perhaps you forgot that the clocks went back—then its being nearly
6 p.m. is not a reason to leave the office now. So, what constraints
does the account need to place on the premises of good reasoning?
Second, what is good reasoning, anyway? Third, how does the
account explain the role that reasons play in determining deontic
status? Fourth, how does the account explain the connections
between reasons and value? Fifth, how does it apply to responses
which might seem not to feature in reasoning, such as emotions?
The key to answering many of these questions, we will suggest,
is the notion of a fitting attitude: an attitude that gets things right.
And this brings us to our second aim.

2 Fittingness First
That it’s nearly 6 p.m. is a reason to believe the match is about to
start, leave the office, and be glad the day is over. Perhaps because
of this, you ought to respond in these ways, or would be rational to
do so. It’s also good that the day is over. The bar you’re heading to
might be a good one for watching the match. The friend you’re
meeting might be a delightful fellow.
These are all normative claims. If true, they state normative facts,
involving normative properties—for example, being a reason, being
what you ought to do, being rational, being good, being a good bar
to watch the match in, being delightful. It’s hard to explain what
normativity is, or what distinguishes the normative from the non-
normative. But we have a clear enough sense that there’s an
important distinction between claims like the above and claims like
that it’s nearly 6 p.m., or that your friend is waiting at the bar, or
that what you’ll soon be drinking contains alcohol.4 It’s enough for
our purposes to pick out the normative by reference to some of
these paradigmatic cases. (This is not to deny, of course, that it can
be unclear whether something is normative.) We also have a sense
of some distinctions within the normative domain. For instance, the
good and the delightful seem to group together as evaluative
properties; being what you ought to do and being what you may do
group together as deontic properties.5
Our second aim is to give an account of the structure of the
normative domain: of how the normative fits together. This is an
account of the dependence relations within the normative domain.
There are lots of possibilities here. It could be that no normative
property depends on any other—the normative domain is ‘flat’. Or
perhaps it could be that all or some normative properties are
mutually dependent—the normative domain is circular, or an
interlocking cluster.6 The most common idea, though, is that the
normative domain has a foundational structure: there are some
normative properties on which all others depend, and which do not
depend on other normative properties. These properties are
normatively basic: they come ‘first’ in the normative domain. (They
may not be basic simpliciter—that is, they may depend on non-
normative properties.) Our view will be of this form. The kind of
dependence we will be most interested in is that associated with
constitutive accounts (more on this in §3.4).
It is, we think, an intrinsically interesting question how the
normative domain is structured. (It is an analogue of familiar
philosophical questions about how the physical, or mental, or
mathematical, or modal domains are structured, and a generalization
of familiar questions about how the epistemic and moral domains
are structured—for instance, is the good prior to the right or
conversely?) It is also important for its bearing on questions about
how the normative fits with other domains. Metanormative views
often proceed by arguing that some normative properties are
normatively foundational, and then arguing that these foundational
properties can in turn be accounted for in non-normative terms.7
This can work only if the structure of the normative domain is
correctly identified. Tracing the structure of the normative domain
has important side benefits too. For instance, it requires us to
carefully distinguish different normative properties, which is often
crucial for assessing arguments and theories in normative
philosophy.
On the view we develop, there is a single normative property,
fittingness, which is normatively basic, and on which all other
normative properties depend. Thus, fittingness is first in the
normative domain. (Again, this does not imply that fittingness is
basic simpliciter; we take no stand on this issue.) We develop this
view by defending constitutive accounts of several normative
properties whose only normative element is fittingness.8
What is fittingness? This question will be addressed in detail later
(especially Chapter 3). But the idea is familiar enough. When you
admire, fear, intend, or believe, your attitude can be fitting or
unfitting, depending on its object. It seems fitting to admire
Mandela, fear an onrushing tiger, intend to phone your mother on
her birthday, and believe that the Seine flows through Paris. As we
will also sometimes put it, in having these attitudes you would be
getting things right. It is not fitting to admire Idi Amin, fear an
onrushing kitten, intend to ignore your mother on her birthday, or
believe that the Thames flows through Paris. In having these
attitudes you would be getting things wrong. Fittingness—this way
of getting things right in your responses—is, we will argue, distinct
from doing what you ought or may do, or what there is reason to
do. It can thus form the basis for constitutive accounts of these
properties.
We can now see, in outline, how our aims align. Reasons, we say,
are premises of good reasoning. But not all premises of good
reasoning are reasons: good reasoning from false beliefs or mistaken
intentions does not correspond to reasons. How should this be
accommodated? And what is good reasoning, anyway? We answer
by appeal to fittingness. Reasons are premises of good reasoning
from fitting responses—e.g. from true beliefs and intentions for the
choiceworthy. Good reasoning is, roughly, reasoning which preserves
fittingness, other things equal. That is, good reasoning is such as to
take you to a fitting response, if you begin from fitting responses, at
least other things equal. We thus end up with an account of reasons
in terms of fittingness: roughly, reasons are premises of fittingness-
preserving reasoning from fitting responses (Chapters 1 and 2).
Why does this support fittingness-first? As noted, reasons help
determine the deontic status of a response—whether you ought or
may so respond. This makes it independently plausible to think that
deontic status should be understood in terms of reasons. We
develop a detailed account of the—surprisingly under-theorized—
relation between reasons and deontic status in Chapters 5 and 6. It
is also independently plausible to understand evaluative properties in
terms of fittingness. For something to be delightful, or admirable, or
loathsome, is for it to be fitting to delight in, admire, or loathe. For
something to be good is for it to be desirable—fitting to desire. We
develop such a view of value in Chapter 4.
When we put these views together, we thus end up with the
following picture (where the arrows mean ‘depends on’).9

Finally for this section, we note that our two aims need not stand or
fall together. It is possible to develop a constitutive account of
reasons as premises of good reasoning without accepting fittingness-
first, and conversely.10 Moreover, the way in which we develop our
constitutive account of reasons does not require fittingness-first—for
instance, one might accept that account while denying that
evaluative properties are to be understood in terms of fittingness.
Nonetheless, we hope to show that combining the two aims is a
natural, plausible, and fruitful approach.

3 Preliminaries
Before going further, some preliminary matters. In §3.1 we make
some remarks about our strategy and ambitions. In §3.2–3.4 we
highlight some assumptions. (Readers less interested in the fine
detail could safely skip these subsections.) §3.5 sets out the plan for
the book.

3.1 Strategy and Limits

As explained, our aim is to develop and defend accounts of reasons
and the structure of the normative domain. But in respect of both
development and defence, our aim should be qualified. We neither
present comprehensive accounts of reasons and the normative
domain nor try to conclusively establish what we do offer.
There are various normative properties we won’t give any account
of. These include rationality, responsibility, rights and obligations,
and thick evaluative properties, such as courage, justice, and
crudeness. There are others we discuss only briefly, such as specific
evaluative properties, like the admirable, the regrettable, and the
amusing (Chapter 4). In this way, we shall offer only the beginnings
of a fittingness-first account of the normative domain.
Our account is also incomplete in that we focus primarily on
reasons for belief and action, which we take to be equivalent to
reasons to intend so to act. The idea that reasons are premises of
good reasoning is probably most tempting in this context. It thus
makes sense to develop such a view by focusing on these cases.
However, we do address questions about reasons for other
responses at various points in the book, especially Chapter 7. We
also recognize that the presumed equivalence between reasons for
action and reasons for intention is contentious. We address this in
Chapter 7, §3.
Our defence of our views is also limited. Establishing them
conclusively, or even showing them to be the most plausible overall,
would require systematic examination of actual and possible rivals.
Our aim is more modest: to present our account as a serious
contender. We do this by explaining attractions of our views and
advantages over salient alternatives, for instance by showing how
our views fit with plausible general ideas and deliver intuitive
verdicts in a range of cases.
This will require some assumptions about the conditions under
which responses are fitting. We assume throughout that it’s fitting to
believe just what is true. As we explain in Chapter 3, §5.2, this
assumption is orthodox but not trivial or uncontentious. In other
cases, there is no such orthodoxy. In particular, it’s highly
contentious what makes intention fitting—or, as we can put it, what
makes a course of action choiceworthy.11 Rather than addressing
this question, we will make what we take to be reasonable
assumptions about the choiceworthiness of courses of action. We
take a similar approach to questions about the fittingness of desire,
preference, and emotions—as we did above, in suggesting that it’s
fitting to admire Mandela but not Amin, to fear an onrushing tiger
but not an onrushing kitten, and so on. These assumptions are made
for illustrative purposes; nothing substantive turns on them.

3.2 Ontology of Reasons

Following Scanlon (1998) and others, we have described reasons as
‘considerations’. We shall take these to be propositions. In ordinary
talk we can describe all sorts of entities as reasons. We can say that
its being nearly 6 p.m., or the match starting, or your friend, is a
reason to leave the office. But such claims can be rephrased, with
only loss of elegance, to cite propositions: a reason to leave the
office is that it’s nearly 6 p.m., or that the match is starting, or that
your friend is waiting. So taking reasons to be propositions allows us
to cover all these cases (Schroeder 2007: 20–21).
Further, we assume that reasons are true propositions or, as we
shall sometimes say, facts. (Nothing will turn on this use of ‘fact’.)
Falsehoods don’t genuinely support anything. Of course, if you
mistakenly think that it’s nearly 6 p.m., you might take this to
support leaving the office and you might reason on this basis. But
when you realize your mistake you’ll stop doing so.12
We also assume that propositions are individuated finely. For
example, the proposition that the match starts now is distinct from
the proposition that the match starts at 6 p.m., even though it’s now
6 p.m. It follows that reasons are also individuated finely. This is a
natural view, given reasons’ role in reasoning (Setiya 2015).
However, it will play no substantial role.13 Those who prefer to think
of reasons as coarse-grained Russellian propositions, or as states of
affairs,14 will be able to interpret or reformulate the discussion
accordingly (see Chapter 1, n. 26).
We will generally allow ourselves to freely and colloquially refer to
different sorts of entities as reasons. However, we will sometimes
need to be more precise. When helpful we will use angled brackets
to refer to propositions: ‘<it’s nearly 6 p.m.>’ refers to the
proposition that it’s nearly 6 p.m. This allows us to say things like
‘<it’s nearly 6 p.m.> is a reason to leave the office’.15

3.3 Normativity
We distinguished the normative by some paradigm cases and
contrasts. Here we want to say a bit more. In particular, we want to
distinguish the normative as we understand it—what is sometimes
called ‘authoritative’ or ‘substantive’ or ‘genuine’ normativity—from
the merely formally normative.
There are norms of etiquette, university codes, and composers’
directions. These can be expressed using normative language; they
might say, for instance, that you ought to place the fork on the left,
take 120 credits each year, or play staccato. They thus have the
form of normativity. Intuitively though, they do not have the same
force, significance, or authority. They don’t seem to matter in and of
themselves. It can be fine not to care about etiquette, for instance;
when it’s not, that is because of further considerations, such as that
violating etiquette would be disrespectful. By contrast, it seems to
matter in and of itself that you keep your promises, look after your
health, and follow the evidence. It’s not fine not to care about these
things.
Our concern in this book is with normativity, not formal
normativity. We take it to be a datum that there is such a distinction,
though it is hard to characterize precisely. What matters is just that
there is an intuitive distinction and that the central topics of this
book—normative reasons, deontic status, certain evaluative
properties, and fittingness—fall on one side of it.16
But is fittingness normative? We talk of responses being fitting
and allow that this can equally be put in terms of responses being
correct. But, as several writers have noted, correctness can be
merely formally normative. There are standards of correctness
governing chess play, knot tying, and cooking cacio e pepe.17 These
standards are merely formally normative. Why not understand the
claims that it is correct to believe the truth, admire the admirable,
and fear the fearsome in the same way?
It is not clear that there is a special problem about fittingness, or
correctness, here. As far as we can tell, all the terms we have used
to express normative notions can also be used to express merely
formally normative notions. We can talk about the reasons to
manipulate the rope, in tying a knot, just as we can talk about
correct knot tying (Lord and Sylvan 2019a). But these reasons don’t
have force or significance or authority in themselves. If there is a
problem here, it must be because talk of fitting responses can only
be understood as referring to formal normativity.
That does not seem to be the case. Claims about the fittingness
of believing, admiring, fearing, and the like seem to us to be
naturally heard as normative. It does not seem fine to not care
about whether you are getting things right in having these attitudes.
Moreover, such claims are closely related to other claims that are
clearly normative. For instance, the claim that it is fitting to admire
an object is closely related to the clearly normative claim that the
object is admirable.18
 Of course, without a more definitive way to characterize the
normative our point here is hard to establish. Intuitively though,
getting things right in your attitudes seems to matter in a way that
correct chess play and knot tying, like etiquette and composers’
directions needn’t. We thus see no reason to deny that fittingness is
normative.
Finally, even if you insist that fittingness is not normative, you can
agree with much of what we argue in this book. So long as you
agree that responses can be fitting, and that fittingness is different
from deontic status and reasons, you can accept our accounts of
reasons and the structure of the normative domain. Indeed, if you
think that fittingness is not normative, but are on board with the rest
of our project, you can take us to have shown how to reduce the
normative to the non-normative. This isn’t how we understand our
project, but it’s not something we argue against.

3.4 Constitutive Accounts

A constitutive account of a property is an account of what it is for
that property to obtain. So constitutive accounts are not the same as
conceptual analyses, traditionally understood. A constitutive account
of Fs says what it is to be F, not what it is to think about Fs.
Constitutive accounts are also not the same as epistemic accounts:
they don’t tell us how to identify Fs. These points are easily seen
with standard examples: being water might be being H2O but you
can think about water, and identify it, without thinking about H2O.
Thus when we say, for example, that reasons are premises of good
reasoning, we are not claiming that you can’t think about reasons, or
identify them, without thinking about them in this way. However,
though we aim to defend constitutive accounts, we will spend some
time examining concepts, especially in our discussion of fittingness
in Chapter 3. Equally, we allow that we can learn things about what
we have reason to do by considering them in light of our account.
 Philosophy is full of constitutive accounts but it is controversial
how they are to be understood. They are variously claimed to state
property identities (Jackson 1998), structures (King 1998; Schroeder
2007: 61–72), essences (Fine 1994; Wedgwood 2007: 136–144),
and grounds (Rosen 2015a). We will not fuss about this: we talk
freely of properties e.g. being, being constituted by, being analysed
by, and being understood in terms of, other properties.
We do assume, however, that when what it is to be F is to be G,
the Fs are so because they are Gs, or because of the constituents of
being G. (For example, if what it is to be a triangle is to be a three-
sided closed figure, then the triangles are so because they are three-
sided closed figures, or because they are three-sided and because
they are closed figures.)19 Assuming that ‘because’ is transitive and
asymmetric, this has important consequences. Transitivity is
important because we want to claim, e.g. that, in analysing deontic
status in terms of reasons, and reasons in terms of fittingness, we
have thereby analysed deontic status in terms of fittingness.
Asymmetry is important because, combined with transitivity, it
implies that constitutive accounts cannot be circular. It cannot be,
for instance, that reasons are analysed in terms of deontic status
and deontic status is analysed in terms of reasons. This will be an
important constraint at various points.20
Finally, we should address an obvious question. We will argue
that fittingness is first in the normative domain. But why assume
that anything has this status? Why take normativity to have a
foundational structure? We make two points about this. First, we do
not assume foundationalism at the outset. As should be clear, we
arrive at our fittingness-first account via independently defended
views—that reasons are premises of good reasoning from fitting
responses, that good reasoning is fittingness-preserving reasoning,
that deontic statuses should be understood in terms of reasons, and
evaluative properties in terms of fitting attitudes.21 Together, these
views make a fittingness-first approach tempting, but our focus will
largely be on the parts, rather than the whole. Second, we do take
foundationalism about the normative, developed via constitutive
accounts, to have significant attractions. In general, accounts of a
property F in terms of another property G promise straightforward
and parsimonious explanations of features of F and of connections
between F and G. We discuss several examples of this in the book.
And since the normative domain seems highly interconnected, the
hypothesis that there is one basic normative property has significant
explanatory promise.

3.5 Plan of the Book

We end with an overview of the main claims of each chapter.
Chapter 1 sets out our core account of reasons, as premises of
good reasoning. We propose some desiderata for a constitutive
account of reasons, and argue that our view, unlike some prominent
alternatives, is well-placed to accommodate them. We introduce and
begin to discuss some challenges the view faces.
Chapter 2 sets out our account of good reasoning, as roughly,
reasoning which preserves fittingness, other things equal. We argue
that this account has important attractions, and then develop it in
light of queries and objections. We then show how to combine our
accounts of reasons and good reasoning, and discuss some
implications of doing so.
Chapter 3 addresses a crucial question raised by our accounts:
what is fittingness? Since we take fittingness to be normatively
basic, we do not try to answer this question by providing a
constitutive account. Rather, we point to intuitive examples of
fittingness, describe four distinctive marks of fittingness, and discuss
what makes responses fitting.
Chapters 4 puts fittingness to work in a further way. We defend
constitutive accounts of evaluative properties in terms of fittingness
—so-called ‘fitting-attitude’ accounts. This chapter is in some ways
independent of the rest of the book—in particular, it does not rely on
our accounts of reasons and good reasoning. However, it is a crucial
part of the case for a fittingness-first account, as well as providing
further support for our claim that fittingness is a distinctive and
significant property.
Chapters 5 and 6 turn to the central challenge left outstanding by
the first half of the book: explaining the role of reasons in
determining deontic status. Chapter 5 gives our account of what it is
for a reason to outweigh another. Very roughly, one reason
outweighs another when it’s good reasoning to respond to the
former and not the latter. Chapter 6 explains how to get from the
weights of reasons to deontic status. Again, very roughly, what you
ought to do is what there is the strongest overall case for doing,
where this is a matter of what the weightiest reasons support. We
show how these natural ideas can be explicated and defended within
our framework and discuss objections.
Chapter 7 turns from the overall picture to the specific. We
investigate several types of reasons that might seem to fit badly with
our account. We argue that, where these reasons are genuine, we
can accommodate them. Along the way we argue, among other
things, that reasons for action can be understood as reasons for
intention, and that emotions and related attitudes can feature in
reasoning.

Getting Things Right: Fittingness, Reasons, and Value. Conor McHugh and
Jonathan Way. Oxford University Press. © Conor McHugh and Jonathan Way 2022.
DOI: 10.1093/oso/9780198810322.003.0001

1 These distinctions are standard—see Alvarez 2016 for an overview. Note that
the glosses are not intended as synonyms. For instance, in Chapter 1, n. 21 we
register a doubt about the equivalence of ‘is a reason for’ and ‘counts in favour of’.
See also Broome 2013: 54.
2 Some philosophers distinguish between a consideration being a reason and
an agent having or possessing that reason (e.g. Schroeder (2007); Lord (2018)).
The latter requires, at least, that the reason be epistemically accessible to the
agent. In these terms, our account concerns the reasons there are, not the
reasons you possess. However, we use the ‘being’ and ‘having’ locutions
equivalently (though favouring the former). We discuss whether reasons must be
accessible in Chapter 1, §4.3.1.
3 There may be other parameters too, such as a time. As is hopefully clear, we
use ‘response’ as our cover-term for the kinds of things there can be reasons for,
which we take to be actions and attitudes. See Chapter 7. We use ‘to do’ and its
cognates broadly to cover actions and attitudes. ‘Consideration’ is Scanlon’s (1998)
useful cover term for the entities which are reasons. See §3.2.
4 Proviso: many of the crucial expressions in these claims are polysemous. For
instance, as we already noted, ‘reason’ can refer to motivating and explanatory
reasons. Our claim here is that there’s an intuitive distinction between these sets
of claims when used in the kinds of context we’re trying to make salient, e.g.
through the description of our examples.
5 Terminology varies. In particular, while some distinguish the normative and
the evaluative, we use ‘normative’ as our cover term, and treat the evaluative as a
subset of the normative (Dancy 2004: 24).
6 For something like this, see McDowell 1985; Wiggins 1987.
7 E.g. Gibbard 1990; Smith 1994; Schroeder 2007—though Gibbard and Smith
work in the first instance with concepts rather than properties.
8 Others to propose fittingness-first views of the normative—often concerning
concepts, rather than properties—include Brentano (1889/2009), Broad (1930),
Ewing (1939, 1948), and more recently Chappell (2012) and Howard (2019a). For
fittingness-based accounts (not always in these terms) of elements of the
normative domain, see, e.g., Brandt 1946; Chisholm 1986; Gibbard 1990; D’Arms
and Jacobson 2000a, 2000b; Danielsson and Olson 2007; Thomson 2008; Nye,
Plunkett, and Ku 2015; Cullity 2018a; Whiting 2018, 2021.
9 This is the most straightforward way to represent the view that emerges.
However, it faces a simple worry: isn’t being good reasoning itself an evaluative
property? We respond to this worry in Chapter 2, n. 24.
10 For examples, see n. 8 and Chapter 1, n. 24.
11 See Davidson 1978/2001; Wedgwood 2007: 24–5; Alvarez 2010: 15–16.
12 Two points. First, in taking reasons to be facts, we deny that they are mental
states. Your belief that it’s nearly 6 p.m. might represent a reason, but it is not
itself the reason. Facts about mental states might be reasons, as when the fact
that you believe you are being followed is a reason to see a psychiatrist (Dancy
2000; though contrast Gibbons 2010). Second, some hold that our remarks here
are true only of objective reasons and that there are also subjective reasons,
which can be false propositions. We discuss this in Chapter 1, §4.3.2.
13 One exception is the argument of Chapter 4, §5.3, at least as developed in
McHugh and Way 2022a.
14 Dancy 2000, 2018; Mantel 2018.
 15 In Chapter 3, we also use angled brackets to refer to concepts—thus ‘<6
p.m.>’ refers to the concept of 6 p.m. There we also use square brackets to refer
to properties.
16 For discussion of the distinction see Parfit 2011; Broome 2013; McPherson
2018; Woods 2018; Wodak 2019a; Rowland 2022. Of course, one problem is that
it is controversial what falls on the ‘normative side’. For example, Foot (1972)
holds that moral norms are merely formal, Maguire and Woods (2020) think
epistemic norms are merely formal, and Broome (2013) thinks rational norms
might be.
17 The latter examples are from Schroeder 2010 and Lord and Sylvan 2019a
respectively. The example of chess is ubiquitous.
18 For more on fittingness as normative, see Howard 2018; Rowland 2022.
19 See Rosen 2010, 2015a on the ‘grounding-reduction’ link. The ‘because’ here
is one of grounding. For overviews of the large literature on this topic, see Bliss
and Trogdon 2014; Raven 2015. Our formulation is intended to allow for the view
that constitutive accounts state identities, which seems inconsistent with the first
disjunct (Audi 2012; Dorr 2016). Note that it’s also plausible that when to be F is
to be G, the Fs are so fully because they are Gs. But we won’t need this further
claim.
20 Transitivity may not be necessary for these results. We might instead say
that fittingness is first because all else either depends on it, or depends on
something that depends on it, or…. However, the view that ‘because’ is transitive is
orthodox (e.g. Rosen 2010) and simplifies exposition, so we work with it here.
21 Thus we build up our account from what Wodak (2020) calls ‘local questions
of explanatory priority’ (50). We agree with Wodak that it is worth exploring other
ways of explaining connections between normative properties.
 1
Reasons

1 Introduction
The aims of this book, we said, are to develop and defend a
constitutive account of normative reasons as premises of good
reasoning, and a fittingness-first account of the structure of the
normative domain. In this chapter, we begin these tasks by
introducing and making an initial case for our account of reasons.
We begin, in §2, by proposing some desiderata for a constitutive
account of reasons. In §3 we discuss some prominent accounts of
reasons and observe some difficulties they face in accommodating
these desiderata. In §4, the heart of the chapter, we introduce our
account, explain its promise with respect to our desiderata, and
discuss some ways of developing the core account to allow for
certain further claims about reasons. In §5, we address one
important challenge to our account, and explain some further
challenges, which set the agenda for subsequent chapters.

2 Roles and Features of Reasons: Desiderata

for a Constitutive Account
That it’s nearly 6 p.m. might be a reason for you to believe that the
match is about to start, decide to leave the office, and be glad that
the day is over. What is it for this to be the case? In asking this
question we are seeking a constitutive account of a certain relation,
the reason-relation, that can obtain between a fact, a response-type,
and an agent (and perhaps other parameters such as a time).1
How are we to assess answers to this question? Beyond general
desiderata on theory choice—simplicity, extensional plausibility, non-
arbitrariness, and the like—we shall set five desiderata for an
answer: two roles that reasons play, and three features of reasons,
which a constitutive account should accommodate and illuminate.
The first is that reasons have a role in reasoning: they are
suitable for reasoning with. Reasoning, after all, is a matter of taking
considerations into account and responding in light of them. It
seems platitudinous that reasons—considerations that genuinely
support responses—are appropriate for doing this with. Indeed, it
seems platitudinous that reasons have a special relevance to
reasoning—they are not merely among the considerations you
should be taking into account when you reason, but should be the
central focus. We will call this role in reasoning the deliberative role
of reasons. An adequate account of reasons must accommodate and
help to make sense of this role.2
Another important role of reasons is that they determine a
response’s deontic status: that is, they determine whether an agent
ought, may, or ought not respond in a certain way. That it’s nearly 6
p.m. can make it the case that you may believe that the match is
about to start; that you have urgent work to finish can make it the
case that you ought to stay in the office and miss the match. While
deontic status is overall, reasons are paradigmatically pro tanto or,
as we shall normally say, contributory. They compete and combine in
order to determine deontic status. In ordinary talk we sometimes
use metaphors of weight and of strength to describe this. For
example, we might say that you ought to stay in the office because
your urgent work is a stronger reason than the match starting, or
that you ought to leave because the match together with a promise
to meet your friends outweighs the urgent work. There are difficult
questions, which we address later, about how to cash out these
metaphors and explain how reasons interact to determine deontic
status. For now, the point is just that they somehow do. We will call
this the explanatory role of reasons. The second desideratum on an
account of reasons, then, is that it should accommodate and
illuminate their explanatory role.3
Reasons have further features that a constitutive account should
capture. As our opening example suggests, there are reasons for a
variety of kinds of response. Though we will often focus on reasons
for intention, action, and belief, there are also reasons for other
attitudes too, including conative attitudes like desire and being glad,
evaluative attitudes like admiration and fear, reactive attitudes like
gratitude and blame, and further doxastic attitudes, such as levels of
confidence and suspension of judgment.4 Reasons for these different
responses look like species of a common genus. They seem to stand
in the very same relation—the relation of being a reason!—to the
response they favour. In support of this, note that there doesn’t
seem to be any ambiguity when we say that its being nearly 6 p.m.
is a reason to believe something, to decide to do something, and to
feel a certain way. Moreover, in each case we find reasons playing
their characteristic roles, such as being suitable for taking into
account in reasoning, and determining deontic status. We thus
assume that a constitutive account of reasons should be unified, in
the sense that it applies to reasons of these different kinds and
reveals what they have in common.5
Notwithstanding the variety of responses for which there are
reasons, there are also limits. There can be reasons to believe that
the match is about to start and to intend to leave the office, but not,
it seems, to have a headache, or to get older, or to win a fair lottery.
If support for these claims is needed, note that there seems no need
for anything to play the deliberative or explanatory roles when it
comes to these things: it doesn’t make sense to deliberate about
whether to have a headache, get older, or win a fair lottery, or to
claim that we ought, may, or ought not do these things. (It might
make sense to deliberate about whether to cause yourself to, e.g.,
have a headache, and it might make sense to claim that you ought
to do this—e.g. as part of a medical experiment, or to win a prize.
But a reason to cause yourself to have a headache is not the same
as a reason for a headache.) The general principle here, we suggest,
is that there can only be (normative) reasons for things that can be
done for (motivating) reasons. We call this the response condition. A
constitutive account of reasons should help explain why this
condition holds.6
Finally, although there is constitutive unity among reasons, there
is also substantive diversity. Many different sorts of considerations
can be reasons, and what is a reason for one response might not be
a reason for another. For instance, reasons for acting are very often
respects in which so acting would be good. Paradigmatic reasons for
belief, on the other hand, are evidence for what is believed, rather
than respects in which so believing would be good. An account of
reasons must accommodate this substantive diversity.
In sum, in assessing an account of reasons we should consider
how well it accommodates and makes sense of:

(i) The deliberative role of reasons

(ii) The explanatory role of reasons
(iii) The constitutive unity of reasons
(iv) The response condition
(v) The substantive diversity of reasons.

In the next section we consider how some prominent theories of

reasons fare against these desiderata, showing that they face
difficulties. We will not survey all actual or possible theories; nor do
we aim to refute those we do consider. Our aim is more modest: to
raise sufficient doubt about prominent theories for the development
of an alternative to seem worthwhile. The task of developing that
alternative will begin in the following section and continue for the
rest of the book.
 3 Some Theories of Reasons

3.1 Ought-Based Theories

A prominent trend in recent theorizing about reasons is to explain
reasons in terms of deontic status, and in particular in terms of facts
about what you ought to do—‘oughts’. We will consider the two main
versions of this idea, one according to which reasons explain oughts
and the other according to which reasons are evidence of oughts.7

3.1.1 Reasons as Explanations of Oughts

Broome (2004, 2013) has argued that a reason is, roughly, a fact
that explains what you ought to do. To be a reason is thus to be an
explanans for a certain kind of explanandum. That it’s nearly 6 p.m.
is a reason to believe that the match is about to start, or to decide
to leave the office, because it explains why—it makes it the case that
—you ought to do these things.
The view that reasons are explanations of oughts does well by
some of our desiderata. It promises to explain (iii) the constitutive
unity of reasons, (iv) the response condition, and (v) the substantive
diversity of reasons. Constitutive unity holds because all reasons
explain oughts. The response condition follows from an analogous,
and similarly plausible, condition on oughts—only things you can do
for reasons can be things you ought to do. Substantive diversity
holds because explanations of oughts might be different across
different responses: facts about what you ought to believe, for
instance, might require quite different explanations from facts about
what you ought to intend.
Unfortunately, this view does less well in making sense of the two
central roles of reasons. Consider (i) the deliberative role. It’s not
clear what connection there is between a consideration’s explaining
why you ought to respond in a certain way and its being suitable for
reasoning to so responding. Perhaps the point of deliberation is to
work out what you ought to do. But working out what you ought to
do typically requires you to first distinguish those considerations that
are relevant to this question—the reasons—from those that aren’t. If
the relevant considerations are just those that explain what you
ought to do, it’s not clear how you could distinguish them without
first knowing what you ought to do (Brunero 2013: 814). So it’s
unclear how you could have the capacity for successful deliberation,
on this picture.
Perhaps this difficulty can be overcome. But in any case, it’s not
clear why working out what you ought to do, supposing this is the
point of deliberation, should require paying attention to explanatory
considerations. Compare: if you aim to work out what’s legal, you’ll
attend to facts about what the law says or evidence for such facts,
not necessarily to the facts that explain why the law is as it is. More
generally, working out whether p is a different endeavour to
identifying explanations of why p. Broome’s view thus does not seem
a particularly good fit with the deliberative role of reasons.
The view that reasons are explanations of oughts might seem
tailor-made to capture (ii), the explanatory role of reasons. Doesn’t it
effectively identify reasons with this role? Not quite. The rough
version of the view, set out above, is too rough to capture this role.
In particular, it does not allow that there can be a reason for you to
do something you are merely permitted to do, or that you ought not
do. In these cases, nothing explains the fact that you ought, since
there is no such fact. Reasons therefore cannot be explanations of
oughts.
To accommodate this, Broome introduces the notion of a
‘weighing explanation’: an explanation of deontic status that takes
into account the various considerations bearing on a response and
how they weigh up. A reason for a response, Broome suggests, will
participate in this weighing explanation, even if the explanandum is
not that you ought to respond that way. What makes it a reason for
the response is the role it plays in this explanation:

A pro tanto reason for N to F is something that plays the for-F role in a
weighing explanation of why N ought to F, or in a weighing explanation of
 why N ought not to F, or in a weighing explanation of why it is not the case
that N ought to F and not the case that N ought not to F.8

Unfortunately, introducing the notion of a ‘for-F role in a weighing

explanation’ threatens to render the account circular. The ‘for-F’ role
is not simply the role of making it the case that you ought to F, since
a consideration can play this role even when it’s not the case that
you ought to F. It must therefore play a kind of defeasible favouring
role. But this sounds very much like the role of being a reason to F
(Brunero 2013).9 And a ‘weighing explanation’ sounds like an
explanation in which considerations supporting conflicting outcomes
interact in some way that depends on their weights, to yield an
overall outcome. If so, it presupposes the notions of support and
weights, which seem inseparable from the notion of a reason
(Gregory 2016). Thus, while Broome’s account does, in a sense,
make the explanatory role of reasons central to their nature, it
seems to fall into circularity in doing so.
A more general worry about Broome’s account, reflected in some
of these points, is that the order of explanation it proposes seems
backwards. As noted, when you ought to do something, this is often
because of the relevant reasons and how they weigh up. For
instance, if you ought to leave the office, that might be because its
being nearly 6 p.m. is a reason to leave the office which outweighs
any competing considerations. In this case, its being nearly 6 p.m.
explains why you ought to leave the office. Why is it that this fact
explains why you ought to leave the office? What is it about this fact
that is relevant? The answer is surely: that this fact is a reason, with
a certain weight. Generalizing, it seems that the facts which are
reasons explain oughts, when they do, because they are reasons.
Broome cannot accept this observation. On his view, that it’s
nearly 6 p.m. is a reason to leave the office because it explains why
you ought to leave the office.10 So, on pain of circularity, he cannot
accept that you ought to leave the office because this is a reason to
leave the office.11 So while Broome can accept that facts that are
reasons explain oughts, he cannot accept that they do so because
they are reasons. Thus, even if he can accept the explanatory role
as initially stated, he cannot accept this further claim it is natural to
accept alongside it.

3.1.2 Reasons as Evidence of Oughts

Kearns and Star (2009) propose a different way of understanding
reasons in terms of oughts. ‘A reason to φ’, they write, ‘is simply
evidence that one ought to φ’ (Kearns and Star 2009: 216). Thus,
that it’s nearly 6 p.m. is a reason for you to believe the match is
about to start, and to decide to leave the office, because it is
evidence that you ought to respond in these ways.12
This account shares some virtues with Broome’s. It accounts in a
broadly similar way for (iii) the constitutive unity and (v) the
substantive diversity of reasons: while reasons are unified by their
nature as evidence that you ought to do something, what constitutes
such evidence might vary depending on the response in question.
The account also has advantages over Broome’s. In particular, it
seems a better fit with (i) the deliberative role of reasons. If
deliberation aims at working out what you ought to do, then
evidence bearing on what you ought to do seems the right sort of
thing to take into account.
In other respects, however, the account seems less attractive
than Broome’s. Reasons-as-evidence lacks any straightforward
explanation of (iv) the response condition. Evidence of what you
ought to do is not subject to any such restriction. There can be
evidence that you ought to have a headache, or get older, or win the
lottery—just imagine reliable testimony to that effect. Of course, the
view could add a clause that only things which can be done for
reasons are subject to reasons. But this is not to illuminate why the
condition holds; it is to stipulate that it does.13
The remaining desideratum is (ii) the explanatory role of reasons.
Reasons-as-evidence faces difficulties capturing this role. The basic
problem is that, in general, evidence and explanations can come
apart: while some evidence helps explain what it is evidence for,
much evidence isn’t like this. So, if any evidence that you ought to φ
is a reason to φ, it’s not clear why reasons should generally
contribute to determining whether you ought to φ. We take this to
be the worry behind a well-known counterexample to the account
due to Broome (2008). According to Broome, that a reliable book
says you ought to eat cabbage is evidence that you ought to eat
cabbage, but what determines whether you ought to eat cabbage is
not the book’s say-so, but rather the nutritional benefits of
cabbage.14
In response, Kearns and Star (2008) distinguish between
‘fundamental’ and ‘nonfundamental’ explanation, and maintain that
the book’s testimony nonfundamentally explains why you ought to
eat cabbage. More generally, all reasons ‘make it the case, in at least
a nonfundamental sense’ that you ought to do whatever you ought
to do (Kearns and Star 2008: 47). However, it’s hard to see what
entitles Kearns and Star to this claim. Though a distinction between
fundamental and nonfundamental explanation can be made, it’s not
generally true that evidence of a fundamental explanation is itself an
explanation, nonfundamental or otherwise. A particle physicist’s
testimony might be evidence of the fundamental explanation of how
some particles behave without in any way making the particles
behave that way. So it’s hard to see why, on the reasons-as-evidence
view, reasons should have much of a role in determining what you
ought to do.15
This view also faces a version of the last worry we raised for
Broome’s view, regarding the order of explanation. Even if some
evidence that you ought to φ helps explain why you ought to φ, it
does not generally do so in virtue of being evidence that you ought
to φ. Kearns and Star thus cannot accept that, when a reason to φ
helps explain why you ought to φ, it does so because it is a reason
to φ. More generally, Kearns and Star must reject the plausible idea
that oughts are explained by reasons in virtue of being reasons.

3.2 Value-Based Theories

Often, reasons to do something depend on what would be valuable.
That you would enjoy watching the match seems like a reason to
watch it, which in turn explains why its being nearly 6 p.m. is a
reason to leave the office. The next family of views we consider take
this connection between reasons and value to be constitutive of
reasons: a reason for a response is a consideration that
appropriately links that response to the good. A representative such
account is given by Maguire:

to be a reason for an option is to be a fact about that option’s promoting

some state of affairs, on the condition that that state of affairs is valuable.16

How well does such an account fare by our desiderata, (i)–(v)? Not
very well, it seems to us. It does offer an explanation of (iii), the
constitutive unity of reasons: reasons for diverse responses are
unified by their connection to value. However, it seems less well
placed as regards the other desiderata.
In order for the account to help explain (i) the deliberative role of
reasons, there must be a tight connection between reasoning and
value. But it’s not clear there is such a connection. Certainly, in
reasoning about how to act it’s often appropriate to think about the
good features of the available options. But this doesn’t seem true of
reasoning to belief and other attitudes. For example, when
considering whether you’re immortal, it doesn’t seem appropriate to
take into account that believing yourself immortal would make you
happy (Way 2016). Similarly, that admiring a sculpture would make
you seem sophisticated is not an appropriate premise of reasoning
towards admiring it. You should instead be thinking about evidence
for or against your immortality, and about the aesthetically relevant
features of the sculpture.
For related reasons, the account is not well placed to capture (ii)
the explanatory role of reasons. While it is common to hold that the
deontic status of action is determined by value, few have wished to
extend this claim wholesale to the deontic status of belief and other
attitudes.17 Whether you ought to believe you’re immortal, or admire
the sculpture, seems to depend at least in part on the evidence
concerning your immortality, and on the aesthetic features of the
sculpture. These considerations seem to be reasons bearing on
belief and admiration independently of the value of holding these
attitudes. By the same token, the value-based account does poorly
on desideratum (iv), the substantive diversity of reasons. Reasons
for belief and admiration include the sorts of considerations just
mentioned—evidence and aesthetic features—which do not seem
plausibly explained by the value-based account.
These claims might seem to ignore the possibility that attitudes
can be non-instrumentally valuable. For sure, it may be said, reasons
for belief and admiration are not plausibly explained by the
instrumental value of believing and admiring—it is not always useful
to believe what there is evidence for, for example. But such attitudes
can also be non-instrumentally good. Perhaps, for instance, true
belief, or knowledge, is good for its own sake. And perhaps
admiration is non-instrumentally good when directed towards
something good. More generally, perhaps conative and evaluative
attitudes are non-instrumentally good when directed towards the
good (Hurka 2001: ch. 1). It might then seem plausible to explain
reasons for such attitudes in terms of such value.
However, even if we allow that attitudes can be valuable in this
way, our objections stand. First, even if some attitudes can be non-
instrumentally good, it is implausible that all attitudes for which
there can be reasons can be non-instrumentally good. Notoriously, it
is in fact not very plausible that true belief, or knowledge, is non-
instrumentally good: there seems nothing good in knowing how
many particles of dust there are on my office shelf, for example
(Whiting 2013; Côté-Bouchard 2017). Nor is it especially plausible
that envy, shame, or anger can be non-instrumentally good,
although there can be reasons for envy, shame, and anger (Howard
2018). Second, even if all attitudes for which there can be reasons
can be non-instrumentally good, the value-based theory as stated
requires a stronger claim: that whenever there is a reason for an
attitude, that attitude is good. That this remains implausible
becomes clear when we consider outweighed reasons. For instance,
there can be outweighed reasons to believe falsehoods. But false
belief need not be in any way valuable (Way 2013a).18
Finally, the value-based theory does not explain the response
condition. Value is not subject to any such restriction: it can be good
to have a headache, or get older, or win the lottery. Indeed, there
seems to be a deep difference here between value and reasons.
While reasons seem to be essentially for the guidance of our
conduct, and thus only apply to responses which can be done for
reasons, the good seems unconstrained by our ability to pursue it.
As with reasons-as-evidence, the value-based theory could add a
clause to incorporate the condition, but again this would be to
stipulate rather than explain.
We therefore don’t find the value-based approach promising. We
don’t deny that reasons are sometimes substantively value-based. It
may often be true that there are reasons to act because of the good
that so acting will promote. What we deny is that this is so because
promoting the good is what it is for there to be a reason for you to
do something.
Value-based accounts of reasons are structurally similar to desire-
based accounts. While value-based accounts analyse reasons in
terms of good outcomes, desire-based accounts analyse reasons in
terms of desired outcomes: for example, a reason is a consideration
that explains why a response promotes the satisfaction of some
desire.19 Analogues of the above problems affect the desire-based
account. The ways that believing you are immortal or admiring the
statue might get you what you want do not seem to be appropriate
premises of reasoning towards those attitudes; nor do they seem to
determine whether you ought to so believe or admire. Ordinary
evidential reasons for belief and aesthetic reasons for admiration
don’t seem to derive from our desires; no desire need be served by
your believing some proposition for which you have some
outweighed evidence, or by your admiring something that there is
some reason to admire but that is not admirable overall. And desire-
promotion isn’t subject to the response condition: having a
headache, getting older, and winning the lottery might promote
desires, although you can’t do them for reasons.

3.3 Primitivism
A number of theorists have claimed that reasons are primitive: what
it is to be a reason cannot be explained in more basic terms. As
Scanlon famously writes:

I will take the idea of a reason as primitive. Any attempt to explain what it is
to be a reason for something seems to me to lead back to the same idea: a
consideration that counts in favor of it. ‘Counts in favor how?’ one might
ask. ‘By providing a reason for it’ seems to be the only answer.20

On this view, no constitutive account of reasons can be given. This

doesn’t mean that nothing at all can be said about reasons. As
Scanlon’s quote illustrates, it might be possible to offer paraphrases,
such as ‘counting in favour’. But, whatever else they may achieve,
these paraphrases won’t explain the property of being a reason in
terms of more fundamental properties.21
When a view says so little, it is hard to provide direct objections.
For sure, primitivism does not seem promising when it comes to our
desiderata, (i)–(v). While it may satisfy (iii)—the constitutive unity of
reasons is explained by their standing in the primitive relation of
being a reason—it does little on its own to explain the others.
Primitivists will reply, though, that many theories will take some
properties as primitive: explanations must come to an end
somewhere. And the property of being a reason, they may add, is a
good candidate for being primitive. This, they may say, is because
other normative properties are best analysed in terms of reasons.
Moreover, accounts of these other normative properties will help
explain some of (i)–(v). For example, (i) the deliberative role of
reasons might be explained in terms of an account of good
reasoning: perhaps good reasoning is a matter of responding to
reasons. Similarly, (ii) the explanatory role of reasons can be
explained by an account of deontic status in terms of reasons:
deontic statuses are just certain sorts of configurations of reasons.22
These replies are reasonable. As previewed in the Introduction,
our own theory has a similar structure: we take fittingness as
normatively basic and explain other normative properties in terms of
fittingness. Ultimately, then, primitivism must be assessed from a
broader perspective. Since it explains nothing by itself, it is attractive
only when paired with a reasons-first approach to normativity, on
which at least some other normative properties are explained in
terms of reasons.23 Primitivism is thus hostage to the success of the
reasons-first programme. We’ll raise problems for aspects of this
programme in this and later chapters, and offer an alternative
picture that we’ll argue fares better. This will be our case against
primitivism.
There is plenty more to be said about all of these accounts of
reasons. Our goal, though, is not to refute them but to present and
explore an alternative. We take the foregoing points to be sufficient
to motivate this task. In the next section, we introduce our account
and note some of its attractions. Development and defence of the
account, and of the broader picture it is part of, is the task for the
rest of the book. Points of comparison with other accounts will come
up along the way.

4 Reasons as Premises of Good Reasoning

4.1 The Account

In a slogan, our account is: reasons are premises of good reasoning.
Slightly more fully: a reason for a response is a fact which is a
premise of good reasoning to that response.24 This requires
unpacking, but first, to get a feel for it, consider an example.
Suppose you think to yourself: ‘It’s nearly 6pm. So, the match is
about to start. So, I’ll leave the office.’ Here, the fact that it’s nearly
6 p.m. features as a premise in reasoning to the belief that the
match is about to start, and to the intention to leave the office. This
is, we can suppose, good (enthymematic) reasoning. On our
account, it is because the fact that it’s nearly 6 p.m. can feature in
this way that it is a reason to believe that the match is about to
start, and to intend to leave the office.
As this indicates, the account takes the deliberative role of
reasons as central to their nature. It is no accident that reasons are
suitable for reasoning with: rather, what makes some consideration a
reason is its suitability, in a certain way, for reasoning with.
We will now start to unpack the account with some clarifications
and refinements. First, we understand reasoning broadly, as a
certain kind of transition in which a set of responses, the premise-
responses, leads to some (further) response, the conclusion-
response. This transition is such that the conclusion-response is
thereby based on the premise-responses. The sorts of things that
count as responses include, at least paradigmatically, attitudes. For
example, theoretical reasoning might involve a transition from a set
of beliefs to a belief. Practical reasoning might involve a transition
from a belief and an intention to a further intention, or, perhaps, to
an action. Though reasoning is thus a psychological process, it will
sometimes be convenient to refer to a piece of reasoning by its
expression in thought or talk, as above.25
To illustrate, the belief that it’s nearly 6 p.m. and the belief that, if
so, the match is about to start, might be premise-responses of
theoretical reasoning to a conclusion-response consisting in the
belief that the match is about to start. This reasoning can be
expressed as ‘it’s nearly 6pm; if so, the match is about to start; so,
the match is about to start.’ Similarly, the intention to watch the
match and the belief that leaving the office now is necessary for
watching the match might be premise-responses of practical
reasoning to a conclusion-response consisting in the intention to
leave the office now. This reasoning can be expressed as ‘I’ll watch
the match; leaving the office now is necessary for doing so; so, I’ll
leave the office.’
Second, a proposition, <p>, is a premise of some reasoning if the
belief that p is among the premise-responses of that reasoning.
Thus, in the first example above, <it’s nearly 6 p.m.> and <if it’s
nearly 6 p.m., the match is about to start> are premises of the
reasoning; in the second example, <leaving now is necessary for
watching the match> is the premise. These propositions are thus
candidates for being reasons for the corresponding conclusion-
responses. By contrast, the contents of intentions do not count as
premises and as such are not candidates for being reasons.26
Third, when we say that reasons are premises of good reasoning,
we are not referring to token episodes of reasoning. We are referring
to pieces of reasoning understood as abstract types. These types
need not be instantiated in order to be assessable as good or bad
reasoning. For example, the move from believing that it’s nearly 6
p.m., and that if so the match is about to start, to believing that the
match is about to start, is good reasoning whether or not you
perform it. Thus subjects don’t need to reason in order to have
reasons. (If you do reason from your reasons, though, then they are
not only normative reasons but also motivating reasons. We take
this to follow from the connection noted above between reasoning
and basing.)
Fourth, reasons are for agents: as we put it above (§2), the
reason-relation obtains between a fact, a response-type, and an
agent. Our formulations so far leave this implicit but we can make it
explicit: a reason is a fact which is a premise of good reasoning
which is appropriately related to an agent. Different views can be
taken of what the appropriate relations might be. For example, it
may be that <p> is a reason for you only if you have some of the
premise-responses of the relevant piece of good reasoning. This
would be one way to allow for reasons to vary between agents. We
discuss this further below (§4.3, Chapter 2, §5.3).27
Fifth, it is a substantive question what makes reasoning good. We
shall discuss this question later (§5.1 and Chapter 2). For now, we
Another random document with
no related content on Scribd:
 CHAPTER III.

The Relation of Vapor and Air.

Sect. 1.—The Boylean Law of the Air’s Elasticity.

I N the Sixth Book (Chap. iv. Sect. 1.) we have already seen how
the conception on the laws of fluid equilibrium was, by Pascal and
others, extended to air, as well as water. But though air presses and
is pressed as water presses and is pressed, pressure produces upon
air an effect which it does not, in any obvious degree, produce upon
water. Air which is pressed is also compressed, or made to occupy a
smaller space; and is consequently also made more dense, or
condensed; and on the other hand, when the pressure upon a
portion of air is diminished, the air expands or is rarefied. These
broad facts are evident. They are expressed in a general way by
saying that air is an elastic fluid, yielding in a certain degree to
pressure, and recovering its previous dimensions when the pressure
is removed.

But when men had reached this point, the questions obviously
offered themselves, in what degree and according to what law air
yields to pressure; when it is compressed, what relation does the
density bear to the pressure? The use which had been made of
tubes containing columns of mercury, by which the pressure of
portions of air was varied and measured, suggested obvious modes
of devising experiments by which this question might be answered.
Such experiments accordingly were made by Boyle about 1650; and
the result at which he arrived was, that when air is thus compressed,
the density is as the pressure. Thus if the pressure of the
atmosphere in its common state be equivalent to 30 inches of
mercury, as shown by the barometer; if air included in a tube be
pressed by 30 additional inches of 164 mercury, its density will be
doubled, the air being compressed into one half the space. If the
pressure be increased threefold, the density is also trebled; and so
on. The same law was soon afterwards (in 1676) proved
experimentally by Mariotte. And this law of the air’s elasticity, that the
density is as the pressure, is sometimes called the Boylean Law, and
sometimes the Law of Boyle and Mariotte.

Air retains its aerial character permanently; but there are other
aerial substances which appear as such, and then disappear or
change into some other condition. Such are termed vapors. And the
discovery of their true relation to air was the result of a long course
of researches and speculations.

[2nd Ed.] [It was found by M. Cagniard de la Tour (in 1823), that at
a certain temperature, a liquid, under sufficient pressure, becomes
clear transparent vapor or gas, having the same bulk as the liquid.
This condition Dr. Faraday calls the Cagniard de la Tour state, (the
Tourian state?) It was also discovered by Dr. Faraday that carbonic-
acid gas, and many other gases, which were long conceived to be
permanently elastic, are really reducible to a liquid state by
pressure. 39 And in 1835, M. Thilorier found the means of reducing
liquid carbonic acid to a solid form, by means of the cold produced in
evaporation. More recently Dr. Faraday has added several
substances usually gaseous to the list of those which could
previously be shown in the liquid state, and has reduced others,
including ammonia, nitrous oxide, and sulphuretted hydrogen, to a
solid consistency. 40 After these discoveries, we may, I think,
reasonably doubt whether all bodies are not capable of existing in
the three consistencies of solid, liquid, and air.
39 Phil. Trans. 1823.

40 Ib. Pt. 1. 1845.

We may note that the law of Boyle and Mariotte is not exactly true
near the limit at which the air passes to the liquid state in such cases
as that just spoken of. The diminution of bulk is then more rapid than
the increase of pressure.

The transition of fluids from a liquid to an airy consistence appears

to be accompanied by other curious phenomena. See Prof. Forbes’s
papers on the Color of Steam under certain circumstances, and on
the Colors of the Atmosphere, in the Edin. Trans. vol. xiv.] 165

Sect. 2.—Prelude to Dalton’s Doctrine of Evaporation.

Visible clouds, smoke, distillation, gave the notion of Vapor; vapor

was at first conceived to be identical with air, as by Bacon. 41 It was
easily collected, that by heat, water might be converted into vapor. It
was thought that air was thus produced, in the instrument called the
æolipile, in which a powerful blast is caused by a boiling fluid; but
Wolfe showed that the fluid was not converted into air, by using
camphorated spirit of wine, and condensing the vapor after it had
been formed. We need not enumerate the doctrines (if very vague
hypotheses may be so termed) of Descartes, Dechales, Borelli. 42
The latter accounted for the rising of vapor by supposing it a mixture
of fire and water; and thus, fire being much lighter than air, the
mixture also was light. Boyle endeavored to show that vapors do not
permanently float in vacuo. He compared the mixture of vapor with
air to that of salt with water. He found that the pressure of the
atmosphere affected the heat of boiling water; a very important fact.
Boyle proved this by means of the air-pump; and he and his friends
were much surprised to find that when air was removed, water only
just warm boiled violently. Huyghens mentions an experiment of the
same kind made by Papin about 1673.
41 Bacon’s Hist. Nat. Cent. i. p. 27.

42They may be seen in Fischer, Geschichte der Physik, vol. ii. p.

175.

The ascent of vapor was explained in various ways in succession,

according to the changes which physical science underwent. It was a
problem distinctly treated of, at a period when hydrostatics had
accounted for many phenomena; and attempts were naturally made
to reduce this fact to hydrostatical principles. An obvious hypothesis,
which brought it under the dominion of these principles, was, to
suppose that the water, when converted into vapor, was divided into
small hollow globules;—thin pellicles including air or heat. Halley
gave such an explanation of evaporation; Leibnitz calculated the
dimensions of these little bubbles; Derham managed (as he
supposed) to examine them with a magnifying glass: Wolfe also
examined and calculated on the same subject. It is curious to see so
much confidence in so lame a theory; for if water became hollow
globules in order to rise as vapor, we require, in order to explain the
formation of these globules, new laws of nature, which are not even
hinted at by 166 the supporters of the doctrine, though they must be
far more complex than the hydrostatical law by which a hollow
sphere floats.
 Newton’s opinion was hardly more satisfactory; he 43 explained
evaporation by the repulsive power of heat; the parts of vapors,
according to him, being small, are easily affected by this force, and
thus become lighter than the atmosphere.
43 Opticks, Qu. 31.

Muschenbroek still adhered to the theory of globules, as the

explanation of evaporation; but he was manifestly discontented with
it; and reasonably apprehended that the pressure of the air would
destroy the frail texture of these bubbles. He called to his aid a
rotation of the globules (which Descartes also had assumed); and,
not satisfied with this, threw himself on electrical action as a reserve.
Electricity, indeed, was now in favor, as hydrostatics had been
before; and was naturally called in, in all cases of difficulty.
Desaguliers, also, uses this agent to account for the ascent of vapor,
introducing it into a kind of sexual system of clouds; according to
him, the male fire (heat) does a part, and the female fire (electricity)
performs the rest. These are speculations of small merit and no
value.

In the mean time, Chemistry made great progress in the

estimation of philosophers, and had its turn in the explanation of the
important facts of evaporation. Bouillet, who, in 1742, placed the
particles of water in the interstices of those of air, may be considered
as approaching to the chemical theory. In 1748, the Academy of
Sciences of Bourdeaux proposed the ascent of vapors as the subject
of a prize; which was adjudged in a manner very impartial as to the
choice of a theory; for it was divided between Kratzenstein, who
advocated the bubbles, (the coat of which he determined to be
50,000th of an inch thick,) and Hamberger, who maintained the truth
1⁄
to be the adhesion of particles of water to those of air and fire. The
latter doctrine had become much more distinct in the author’s mind
when seven years afterwards (1750) he published his Elementa
Physices. He then gave the explanation of evaporation in a phrase
which has since been adopted,—the solution of water in air; which
he conceived to be of the same kind as other chemical solutions.

This theory of solution was further advocated and developed by Le

Roi; 44 and in his hands assumed a form which has been extensively
adopted up to our times, and has, in many instances, tinged the
language commonly used. He conceived that air, like other solvents,
167 might be saturated; and that when the water was beyond the
amount required for saturation, it appeared in a visible form. The
saturating quantity was held to depend mainly on warmth and wind.
44 Ac. R. Sc. Paris, 1750.

This theory was by no means devoid of merit; for it brought

together many of the phenomena, and explained a number of the
experiments which Le Roi made. It explained the facts of the
transparency of vapor, (for perfect solutions are transparent,) the
precipitation of water by cooling, the disappearance of the visible
moisture by warming it again, the increased evaporation by rain and
wind; and other observed phenomena. So far, therefore, the
introduction of the notion of the chemical solution of water in air was
apparently very successful. But its defects are of a very fatal kind; for
it does not at all apply to the facts which take place when air is
excluded.

In Sweden, in the mean time, 45 the subject had been pursued in a

different, and in a more correct manner. Wallerius Ericsen had, by
various experiments, established the important fact, that water
evaporates in a vacuum. His experiments are clear and satisfactory;
and he inferred from them the falsity of the common explanation of
evaporation by the solution of water in air. His conclusions are drawn
in a very intelligent manner. He considers the question whether
water can be changed into air, and whether the atmosphere is, in
consequence, a mere collection of vapors; and on good reasons,
decides in the negative, and concludes the existence of
permanently-elastic air different from vapor. He judges, also, that
there are two causes concerned, one acting to produce the first
ascent of vapors, the other to support them afterwards. The first,
which acts in a vacuum, he conceives to be the mutual repulsion of
the particles; and since this force is independent of the presence of
other substances, this seems to be a sound induction. When the
vapors have once ascended into the air, it may readily be granted
that they are carried higher, and driven from side to side by the
currents of the atmosphere. Wallerius conceives that the vapor will
rise till it gets into air of the same density as itself, and being then in
equilibrium, will drift to and fro.
45 Fischer, Gesch. Phys. vol. v. p. 63.

The two rival theories of evaporation, that of chemical solution and

that of independent vapor, were, in various forms, advocated by the
next generation of philosophers. De Saussure may be considered as
the leader on one side, and De Luc on the other. The former
maintained the solution theory, with some modifications of his own.
De 168 Luc denied all solution, and held vapor to be a combination of
the particles of water with fire, by which they became lighter than air.
According to him, there is always fire enough present to produce this
combination, so that evaporation goes on at all temperatures.
 This mode of considering independent vapor as a combination of
fire with water, led the attention of those who adopted that opinion to
the thermometrical changes which take place when vapor is formed
and condensed. These changes are important, and their laws
curious. The laws belong to the induction of latent heat, of which we
have just spoken; but a knowledge of them is not absolutely
necessary in order to enable us to understand the manner in which
steam exists in air.

De Luc’s views led him 46 also to the consideration of the effect of

pressure on vapor. He explains the fact that pressure will condense
vapor, by supposing that it brings the particles within the distance at
which the repulsion arising from fire ceases. In this way, he also
explains the fact, that though external pressure does thus condense
steam, the mixture of a body of air, by which the pressure is equally
increased, will not produce the same effect; and therefore, vapors
can exist in the atmosphere. They make no fixed proportion of it; but
at the same temperature we have the same pressure arising from
them, whether they are in air or not. As the heat increases, vapor
becomes capable of supporting a greater and greater pressure, and
at the boiling heat, it can support the pressure of the atmosphere.
46 Fischer, vol. vii. p. 453. Nouvelles Idées sur la Météorologie,
1787.

De Luc also marked very precisely (as Wallerius had done) the
difference between vapor and air; the former being capable of
change of consistence by cold or pressure, the latter not so. Pictet,
in 1786, made a hygrometrical experiment, which appeared to him to
confirm De Luc’s views; and De Luc, in 1792, published a concluding
essay on the subject in the Philosophical Transactions. Pictet’s
Essay on Fire, in 1791, also demonstrated that “all the train of
hygrometrical phenomena takes place just as well, indeed rather
quicker, in a vacuum than in air, provided the same quantity of
moisture is present.” This essay, and De Luc’s paper, gave the
death-blow to the theory of the solution of water in air.

Yet this theory did not fall without an obstinate struggle. It was
taken up by the new school of French chemists, and connected with
their views of heat. Indeed, it long appears as the prevalent opinion.
169 Girtanner, 47 in his Grounds of the Antiphlogistic Theory, may be
considered as one of the principal expounders of this view of the
matter. Hube, of Warsaw, was, however, the strongest of the
defenders of the theory of solution, and published upon it repeatedly
about 1790. Yet he appears to have been somewhat embarrassed
with the increase of the air’s elasticity by vapor. Parrot, in 1801,
proposed another theory, maintaining that De Luc had by no means
successfully attacked that of solution, but only De Saussure’s
superfluous additions to it.
47 Fischer, vol. vii. 473.

It is difficult to see what prevented the general reception of the

doctrine of independent vapor; since it explained all the facts very
simply, and the agency of air was shown over and over again to be
unnecessary. Yet, even now, the solution of water in air is hardly
exploded. M. Gay Lussac, 48 in 1800, talks of the quantity of water
“held in solution” by the air; which, he says, varies according to its
temperature and density by a law which has not yet been
discovered. And Professor Robison, in the article “Steam,” in the
Encyclopædia Britannica (published about 1800), says, 49 “Many
philosophers imagine that spontaneous evaporation, at low
temperatures, is produced in this way (by elasticity alone). But we
cannot be of this opinion; and must still think that this kind of
evaporation is produced by the dissolving power of the air.” He then
gives some reasons for his opinion. “When moist air is suddenly
rarefied, there is always a precipitation of water. But by this new
doctrine the very contrary should happen, because the tendency of
water to appear in the elastic form is promoted by removing the
external pressure.” Another main difficulty in the way of the doctrine
of the mere mixture of vapor and air was supposed to be this; that if
they were so mixed, the heavier fluid would take the lower part, and
the lighter the higher part, of the space which they occupied.
48 Ann. Chim. tom. xliii.

49 Robison’s Works, ii. 37.

The former of these arguments was repelled by the consideration

that in the rarefaction of air, its specific heat is changed, and thus its
temperature reduced below the constituent temperature of the vapor
which it contains. The latter argument is answered by a reference to
Dalton’s law of the mixture of gases. We must consider the
establishment of this doctrine in a new section, as the most material
step to the true notion of evaporation. 170

Sect. 3.—Dalton’s Doctrine of Evaporation.

A portion of that which appears to be the true notion of evaporation

was known, with greater or less distinctness, to several of the
physical philosophers of whom we have spoken. They were aware
that the vapor which exists in air, in an invisible state, may be
condensed into water by cold: and they had noticed that, in any state
of the atmosphere, there is a certain temperature lower than that of
the atmosphere, to which, if we depress bodies, water forms upon
them in fine drops like dew; this temperature is thence called the
dew-point. The vapor of water which exists anywhere may be
reduced below the degree of heat which is necessary to constitute it
vapor, and thus it ceases to be vapor. Hence this temperature is also
called the constituent temperature. This was generally known to the
meteorological speculators of the last century, although, in England,
attention was principally called to it by Dr. Wells’s Essay on Dew, in
1814. This doctrine readily explains how the cold produced by
rarefaction of air, descending below the constituent temperature of
the contained vapor, may precipitate a dew; and thus, as we have
said, refutes one obvious objection to the theory of independent
vapor.

The other difficulty was first fully removed by Mr. Dalton. When his
attention was drawn to the subject of vapor, he saw insurmountable
objections to the doctrine of a chemical union of water and air. In
fact, this doctrine was a mere nominal explanation; for, on closer
examination, no chemical analogies supported it. After some
reflection, and in the sequel of other generalizations concerning
gases, he was led to the persuasion, that when air and steam are
mixed together, each follows its separate laws of equilibrium, the
particles of each being elastic with regard to those of their own kind
only: so that steam may be conceived as flowing among the particles
of air 50 “like a stream of water among pebbles;” and the resistance
which air offers to evaporation arises, not from its weight, but from
the inertia of its particles.
50 Manchester Memoirs, vol. v. p. 581.

It will be found that the theory of independent vapor, understood

with these conditions, will include all the facts of the case;—gradual
evaporation in air; sudden evaporation in a vacuum; the increase of
171 the air’s elasticity by vapor; condensation by its various causes;
and other phenomena.

But Mr. Dalton also made experiments to prove his fundamental

principle, that if two different gases communicate, they will diffuse
themselves through each other; 51 —slowly, if the opening of
communication be small. He observes also, that all the gases had
equal solvent powers for vapor, which could hardly have happened,
had chemical affinity been concerned. Nor does the density of the air
make any difference.
51 New System of Chemical Philosophy, vol. i. p. 151.

Taking all these circumstances into the account, Mr. Dalton

abandoned the idea of solution. “In the autumn of 1801,” he says, “I
hit upon an idea which seemed to be exactly calculated to explain
the phenomena of vapor: it gave rise to a great variety of
experiments,” which ended in fixing it in his mind as a true idea.
“But,” he adds, “the theory was almost universally misunderstood,
and consequently reprobated.”

Mr. Dalton answers various objections. Berthollet had urged that

we can hardly conceive the particles of an elastic substance added
to those of another, without increasing its elasticity. To this Mr. Dalton
replies by adducing the instance of magnets, which repel each other,
but do not repel other bodies. One of the most curious and ingenious
objections is that of M. Gough, who argues, that if each gas is elastic
with regard to itself alone, we should hear, produced by one stroke,
four sounds; namely, first, the sound through aqueous vapor;
second, the sound through azotic gas; third, the sound through
oxygen gas; fourth, the sound through carbonic acid. Mr. Dalton’s
answer is, that the difference of time at which these sounds would
come is very small; and that, in fact, we do hear, sounds double and
treble.

In his New System of Chemical Philosophy, Mr. Dalton considers

the objections of his opponents with singular candor and impartiality.
He there appears disposed to abandon that part of the theory which
negatives the mutual repulsion of the particles of the two gases, and
to attribute their diffusion through one another to the different size of
the particles, which would, he thinks, 52 produce the same effect.
52 New System, vol. i. p. 188.

In selecting, as of permanent importance, the really valuable part

of this theory, we must endeavor to leave out all that is doubtful or
unproved. I believe it will be found that in all theories hitherto 172
promulgated, all assertions respecting the properties of the particles
of bodies, their sizes, distances, attractions, and the like, are
insecure and superfluous. Passing over, then, such hypotheses, the
inductions which remain are these;—that two gases which are in
communication will, by the elasticity of each, diffuse themselves in
one another, quickly or slowly; and—that the quantity of steam
contained in a certain space of air is the same, whatever be the air,
whatever be its density, and even if there be a vacuum. These
propositions may be included together by saying, that one gas is
mechanically mixed with another; and we cannot but assent to what
Mr. Dalton says of the latter fact,—“this is certainly the touchstone of
the mechanical and chemical theories.” This doctrine of the
mechanical mixture of gases appears to supply answers to all the
difficulties opposed to it by Berthollet and others, as Mr. Dalton has
shown; 53 and we may, therefore, accept it as well established.
 53 New System, vol. i. p. 160, &c.

This doctrine, along with the principle of the constituent

temperature of steam, is applicable to a large series of
meteorological and other consequences. But before considering the
applications of theory to natural phenomena, which have been
made, it will be proper to speak of researches which were carried on,
in a great measure, in consequence of the use of steam in the arts: I
mean the laws which connect its elastic force with its constituent
temperature.

Sect. 4.—Determination of the Laws of the Elastic Force of Steam.

The expansion of aqueous vapor at different temperatures is

governed, like that of all other vapors, by the law of Dalton and Gay-
Lussac, already mentioned; and from this, its elasticity, when its
expansion is resisted, will be known by the law of Boyle and
Mariotte; namely, by the rule that the pressure of airy fluids is as the
condensation. But it is to be observed, that this process of
calculation goes on the supposition that the steam is cut off from
contact with water, so that no more steam can be generated; a case
quite different from the common one, in which the steam is more
abundant as the heat is greater. The examination of the force of
vapor, when it is in contact with water, must be briefly noticed.

During the period of which we have been speaking, the progress

of the investigation of the laws of aqueous vapor was much
accelerated 173 by the growing importance of the steam-engine, in
which those laws operated in a practical form. James Watts, the
main improver of that machine, was thus a great contributor to
speculative knowledge, as well as to practical power. Many of his
improvements depended on the laws which regulate the quantity of
heat which goes to the formation or condensation of steam; and the
observations which led to these improvements enter into the
induction of latent heat. Measurements of the force of steam, at all
temperatures, were made with the same view. Watts’s attention had
been drawn to the steam-engine in 1759, by Robison, the former
being then an instrument-maker, and the latter a student at the
University of Glasgow. 54 In 1761 or 1762, he tried some experiments
on the force of steam in a Papin’s Digester; 55 and formed a sort of
working model of a steam-engine, feeling already his vocation to
develope the powers of that invention. His knowledge was at that
time principally derived from Desaguliers and Belidor, but his own
experiments added to it rapidly. In 1764 and 1765, he made a more
systematical course of experiments, directed to ascertain the force of
steam. He tried this force, however, only at temperatures above the
boiling-point; and inferred it at lower degrees from the supposed
continuity of the law thus obtained. His friend Robison, also, was
soon after led, by reading the account of some experiments of Lord
Charles Cavendish, and some others of Mr. Nairne, to examine the
same subject. He made out a table of the correspondence of the
elasticity and the temperature of vapor, from thirty-two to two
hundred and eighty degrees of Fahrenheit’s thermometer. 56 The
thing here to be remarked, is the establishment of a law of the
pressure of steam, down to the freezing-point of water. Ziegler of
Basle, in 1769, and Achard of Berlin, in 1782, made similar
experiments. The latter examined also the elasticity of the vapor of
alcohol. Betancourt, in 1792, published his Memoir on the expansive
force of vapors; and his tables were for some time considered the
most exact. 174 Prony, in his Architecture Hydraulique (1796),
established a mathematical formula, 57 on the experiments of
Betancourt, who began his researches in the belief that he was first
in the field, although he afterwards found that he had been
anticipated by Ziegler. Gren compared the experiments of
Betancourt and De Luc with his own. He ascertained an important
fact, that when water boils, the elasticity of the steam is equal to that
of the atmosphere. Schmidt at Giessen endeavored to improve the
apparatus used by Betancourt; and Biker, of Rotterdam, in 1800,
made new trials for the same purpose.
54 Robison’s Works, vol. ii. p. 113.

55 Denis Papin, who made many of Boyle’s experiments for him,

had discovered that if the vapor be prevented from rising, the
water becomes hotter than the usual boiling-point; and had hence
invented the instrument called Papin’s Digester. It is described in
his book, La manière d’amolir les os et de faire cuire toutes sorts
de viandes en fort peu de temps et à peu de frais. Paris, 1682.

56 These were afterwards published in the Encyclopædia

Britannica; in the article “Steam,” written by Robison.

57 Architecture Hydraulique, Seconde Partie, p. 163.

In 1801, Mr. Dalton communicated to the Philosophical Society of

Manchester his investigations on this subject; observing truly, that
though the forces at high temperatures are most important when
steam is considered as a mechanical agent, the progress of
philosophy is more immediately interested in accurate observations
on the force at low temperatures. He also found that his elasticities
for equidistant temperatures resembled a geometrical progression,
but with a ratio constantly diminishing. Dr. Ure, in 1818, published in
the Philosophical Transactions of London, experiments of the same
kind, valuable from the high temperatures at which they were made,
and for the simplicity of his apparatus. The law which he thus
obtained approached, like Dalton’s, to a geometrical progression. Dr.
Ure says, that a formula proposed by M. Biot gives an error of near
nine inches out of seventy-five, at a temperature of 266 degrees.
This is very conceivable, for if the formula be wrong at all, the
geometrical progress rapidly inflames the error in the higher portions
of the scale. The elasticity of steam, at high temperatures, has also
been experimentally examined by Mr. Southern, of Soho, and Mr.
Sharpe, of Manchester. Mr. Dalton has attempted to deduce certain
general laws from Mr. Sharpe’s experiments; and other persons
have offered other rules, as those which govern the force of steam
with reference to the temperature: but no rule appears yet to have
assumed the character of an established scientific truth. Yet the law
of the expansive force of steam is not only required in order that the
steam-engine may be employed with safety and to the best
advantage; but must also be an important point in every consistent
thermotical theory.

[2nd Ed.] [To the experiments on steam made by private

physicists, are to be added the experiments made on a grand scale
by order of the governments of France and of America, with a view
to 175 legislation on the subject of steam-engines. The French
experiments were made in 1823, under the direction of a
commission consisting of some of the most distinguished members
of the Academy of Sciences; namely, MM. de Prony, Arago, Girard,
and Dulong. The American experiments were placed in the hands of
a committee of the Franklin Institute of the State of Pennsylvania,
consisting of Prof. Bache and others, in 1830. The French
experiments went as high as 435° of Fahrenheit’s thermometer,
corresponding to a pressure of 60 feet of mercury, or 24
atmospheres. The American experiments were made up to a
temperature of 346°, which corresponded to 274 inches of mercury,
more than 9 atmospheres. The extensive range of these
experiments affords great advantages for determining the law of the
expansive force. The French Academy found that their experiments
indicated an increase of the elastic force according to the fifth power
of a binominal 1 + mt, where t is the temperature. The American
Institute were led to a sixth power of a like binominal. Other
experimenters have expressed their results, not by powers of the
temperature, but by geometrical ratios. Dr. Dalton had supposed that
the expansion of mercury being as the square of the true
temperature above its freezing-point, the expansive force of steam
increases in geometrical ratio for equal increments of temperature.
And the author of the article Steam in the Seventh Edition of the
Encyclopædia Britannica (Mr. J. S. Russell), has found that the
experiments are best satisfied by supposing mercury, as well as
steam, to expand in a geometrical ratio for equal increments of the
true temperature.

It appears by such calculation, that while dry gas increases in the

ratio of 8 to 11, by an increase of temperature from freezing to
boiling water; steam in contact with water, by the same increase of
temperature above boiling water, has its expansive force increased
in the proportion of 1 to 12. By an equal increase of temperature,
mercury expands in about the ratio of 8 to 9.

Recently, MM. Magnus of Berlin, Holzmann and Regnault, have

made series of observations on the relation between temperature
and elasticity of steam. 58
58 See Taylor’s Scientific Memoirs, Aug. 1845, vol. iv. part xiv.,
and Ann. de Chimie.
 Prof. Magnus measured his temperatures by an air-thermometer;
a process which, I stated in the first edition, seemed to afford the
best promise of simplifying the law of expansion. His result is, that
the 176 elasticity proceeds in a geometric series when the
temperature proceeds in an arithmetical series nearly; the
differences of temperature for equal augmentations of the ratio of
elasticity being somewhat greater for the higher temperatures.

The forces of the vapors of other liquids in contact with their

liquids, determined by Dr. Faraday, as mentioned in Chap. ii. Sect. 1,
are analogous to the elasticity of steam here spoken of.]

~Additional material in the 3rd edition.~

Sect. 5.—Consequences of the Doctrine of Evaporation.—

Explanation of Rain, Dew, and Clouds.

The discoveries concerning the relations of heat and moisture which

were made during the last century, were principally suggested by
meteorological inquiries, and were applied to meteorology as fast as
they rose. Still there remains, on many points of this subject, so
much doubt and obscurity, that we cannot suppose the doctrines to
have assumed their final form; and therefore we are not here called
upon to trace their progress and connexion. The principles of
atmology are pretty well understood; but the difficulty of observing
the conditions under which they produce their effects in the
atmosphere is so great, that the precise theory of most
meteorological phenomena is still to be determined.

We have already considered the answers given to the question:

According to what rules does transparent aqueous vapor resume its
form of visible water? This question includes, not only the problems
of Rain and Dew, but also of Clouds; for clouds are not vapor, but
water, vapor being always invisible. An opinion which attracted much
notice in its time, was that of Hutton, who, in 1784, endeavored to
prove that if two masses of air saturated with transparent vapor at
different temperatures are mixed together, the precipitation of water
in the form either of cloud or of drops will take place. The reason he
assigned for the opinion was this: that the temperature of the mixture
is a mean between the two temperatures, but that the force of the
vapor in the mixture, which is the mean of the forces of the two
component vapors, will be greater than that which corresponds to
the mean temperature, since the force increases faster than the
temperature; 59 and hence some part of the vapor will be
precipitated. This doctrine, it will be seen, speaks of vapor as
“saturating” air, and is 177 therefore, in this form, inconsistent with
Dalton’s principle; but it is not difficult to modify the expression so as
to retain the essential part of the explanation.
59 Edin. Trans. vol. 1. p. 42.

Dew.—The principle of a “constituent temperature” of steam, and

the explanation of the “dew-point,” were known, as we have said
(chap. iii. sect. 3,) to the meteorologists of the last century; but we
perceive how incomplete their knowledge was, by the very gradual
manner in which the consequences of this principle were traced out.
We have already noticed, as one of the books which most drew
attention to the true doctrine, in this country at least, Dr. Wells’s
Essay on Dew, published in 1814. In this work the author gives an
account of the progress of his opinions; 60 “I was led,” he says, “in
the autumn of 1784, by the event of a rude experiment, to think it
probable that the formation of dew is attended with the production of