Post on 31-Oct-2021
The University of Manchester Research
Freedom, Fairness and Responsibility
Link to publication record in Manchester Research Explorer
Citation for published version (APA):O'Neill, M. (2009). Freedom, Fairness and Responsibility.
Citing this paperPlease note that where the full-text provided on Manchester Research Explorer is the Author Accepted Manuscriptor Proof version this may differ from the final Published version. If citing, it is advised that you check and use thepublisher's definitive version.
General rightsCopyright and moral rights for the publications made accessible in the Research Explorer are retained by theauthors and/or other copyright owners and it is a condition of accessing publications that users recognise andabide by the legal requirements associated with these rights.
Takedown policyIf you believe that this document breaches copyright please refer to the University of Manchester’s TakedownProcedures [http://man.ac.uk/04Y6Bo] or contact uml.scholarlycommunications@manchester.ac.uk providingrelevant details, so we can investigate your claim.
Download date:31. Oct. 2021
Freedom, Fairness and Responsibility
A dissertation presented
by
Martin Patrick O’Neill
to
The Department of Philosophy
in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
in the subject of
Philosophy
Harvard University Cambridge, Massachusetts
January 2009
© 2009 – Martin Patrick O’Neill
All rights reserved.
iii
Advisor: Professor Thomas M. Scanlon, Jr Martin Patrick O’Neill
Freedom, Fairness and Responsibility
Abstract
Philosophical problems of freedom and responsibility are among the most
recalcitrant philosophical problems that we have, and are connected to a range of
important issues in our understanding of agency, autonomy, blame, and the grounds of
moral assessment. Freedom, Fairness and Responsibility argues for a ‘Hybrid View’ on
these issues: that is, a view that combines the insights of more traditional compatibilist
and incompatibilist theories into a unified solution to this set of closely connected
philosophical problems.
Part One, ‘Agency, Autonomy and Desire: Or, Rescuing the Rational
Wanton’, is based around an examination of the views of Harry Frankfurt. In this part of
the dissertation, I argue that the sort of “volitional hierarchy” described in Frankfurt’s
work can generate neither a necessary nor a sufficient condition for personhood or for
freedom of the will. I also critically examine Frankfurt’s conception of desire, and his
account of the conditions for moral responsibility. This Part concludes that the
shortcomings of Frankfurt’s view should lead us to seek an alternative ‘pluralist’ account
of the conditions for autonomy.
iv
Part Two, ‘Freedom without Resentment: Responsibility and the Reactive
Attitudes’, undertakes a careful investigation of the influential views of P. F. Strawson.
Here, I critically investigate Strawson’s account of the relationship between the moral
attitudes and reactive interpersonal attitudes such as resentment, and give reasons why we
should reject the sort of ‘naturalistic compatibilism’ that Strawson’s approach embodies.
I conclude that Strawson has failed to show that our practices of using desert-entailing
reactive and moral attitudes are outside the scope of rational criticism.
Part Three, ‘Freedom, Fairness, Responsibility and Blame: A Hybrid
View’ presents and defends my own positive view regarding freedom and responsibility.
I argue that the standards of fairness that govern ‘responsibility-as-blameworthiness’
differ significantly from those standards of fairness that govern ‘responsibility-as-
assessability’. I conclude that we should therefore endorse a view that is broadly
incompatibilist about the former kind of responsibility, but compatibilist with regard to
the latter variety, and I further support this Hybrid view by appeal to some general
considerations of philosophical methodology.
v
Table of Contents
Acknowledgements viii
Dedication xiii
Epigraph xiv
Part One:
Agency, Autonomy and Desire – Or, Rescuing the Rational Wanton
1
1. Introduction 2
2. Second-Order Desires and the Concept of a Person 5
3. Volitional Hierarchy, Judgement-Sensitivity, Agency and Personhood 10
4. Rescuing the Rational Wanton: Personhood, Rationality and Desire 20
5. Personhood, Agency and Freedom of the Will 31
6. Decisiveness, Commitment and Higher-Order Desires 40
7. Hierarchy, Volition and Responsibility 49
8. Coda: Towards a Pluralist Account of Autonomy 65
Part Two:
Freedom without Resentment: Responsibility and the Reactive Attitudes
70
1. Strawson’s View: Practical Compatibilism as a Form of Naturalism 71
2. Relationships and Reactive Attitudes 78
a. What are the reactive attitudes? – Broad and narrow conceptions 78
b. Excuses and Exemptions – ‘Quality of Will’ and ‘Capacity’ 84
c. Making Sense of the ‘Objective Attitude’ 89
d. In Praise of the Objective Attitude:
On Expectations, Sympathy and Engagement
104
vi
e. Taking Stock: The Varieties of Reactive and Objective Attitudes 112
f. The Reactive Attitudes and the Varieties of Incompatibilism 115
g. The Objective Attitude and the Inescapability of ‘the Human Commitment’ 126
h. Rationality, Objectivity, and ‘The General Framework of Human Life’ 136
i. Rationality, the Reactive Attitudes, and “the Gains and Losses to Human Life” 144
(1) The Desirability Objection 145
(2) The Anti-Consequentialist Objection 147
j. Conclusion: Strawson’s First Question Answered 149
3. Moral Judgement, Responsibility and the Reactive Attitudes 151
a. “A More Usual Area of Debate” 151
b. “Kindred Attitudes”?
The Reactive Attitudes and their Structural Transformations
154
c. Strawson on Disapprobation, Resentment and the Contours of the Moral 166
d. Strawson’s Conception of the Moral:
Responsibility, Disapprobation and Resentment
173
e. The Moral Excuses and Exemptions, Moral Judgements,
and the Objective Attitude
184
f. On the Cultivation of the “Purely Objective View” 198
g. Strawson’s Second Question: Moral Attitudes and The Human Commitment 204
h. Naturalism, Practice and “the Consequence of the General Thesis” 208
i. Rationality, Responsibility and the Moral Attitudes 215
j. “A Crude Opposition of Phrase Where We Have a Great Intricacy of Phenomena” 223
k. Optimism, Pessimism, Moral Responsibility and the Objective Attitude 231
l. Optimism, Pessimism and the “General Framework” of Human Attitudes 245
m. Conclusion: Strawson’s Second Question Answered 254
4. Beyond Optimism and Pessimism: Or, Freedom without Resentment 255
Part Three:
Freedom, Fairness, Responsibility and Blame: A Hybrid View
257
1. A Real Reconciliation between Compatibilism and Incompatibilism 258
2. Fairness and Responsibility 266
vii
3. Conceptions of Responsibility and Blame 282
4. Fairness, Responsibility and Blame: The Case for Restricted Incompatibilism 294
5. Putting the Hybridity into the Hybrid View:
Circumscribed Incompatibilism and the Conditions for Moral Assessment
318
6. The ‘Meta Argument’ for the Hybrid View:
On the Enduringness and Recalcitrance of Disagreements about Responsibility
333
7. The Place of Optimism 338
8. On the Apparent Insolubility of the Problem of Responsibility
341
Bibliography 343
List of Figures:
Figure 1. Cartesian Desire 15
viii
Acknowledgements
[Oxton, 19 January 2009] During the time that I have been thinking and writing on these
issues, I have benefitted from the help of many people, and a number of institutions. I’m pleased
to acknowledge many of them here, and apologize to anyone whom I’ve forgotten.
It has been a rare pleasure and privilege to work on my dissertation under the supervision of Tim
Scanlon and Derek Parfit. As will be clear to anyone who reads my work, the influence of Tim
and Derek on my philosophical outlook has been profound and pervasive. I am grateful to both of
them for their guidance, enthusiasm and support.
In the case of Derek Parfit, my gratitude for his help dates back all the way to my time as a
B.Phil. student in Oxford, where he co-supervised my B.Phil. thesis. I have grown used to
Derek’s boundless philosophical enthusiasm, his thoughtfulness, and his wonderful comments on
my work. But familiarity with Derek’s qualities as an advisor and supervisor has never blunted
my awareness of how exceptionally fortunate I have been to work with him.
Tim Scanlon was the primary academic reason that I came to Harvard. I have learned an
enormous amount from his seminars, from working with him as a Teaching Fellow in one of his
undergraduate courses, and from many conversations with him over the years. I have come to
regard Tim as the very embodiment of the virtue of good philosophical judgement, and cannot
imagine a better philosophical advisor. I thank him, especially, for his ongoing, and only
moderately successful, attempts to curb my tendency towards writing with perhaps a tad too
much rhetorical enthusiasm.
As well as Tim and Derek, I am grateful to a number of other philosophers at Harvard. Dick
Moran, the third member of my dissertation committee, has been a thoughtful and insightful
interlocutor. This dissertation benefitted significantly from his influence, especially during its
early stages. Niko Kolodny, before his return to Berkeley, served for a year as the fourth member
of my committee. I learned a great deal from talking through my work with Niko, and my
dissertation certainly gained from his influence. Although Chris Korsgaard was not directly
involved in supervising my work, it is hard to over-estimate how much I have learned from her.
My understanding of the history of moral philosophy, and of the prospects of Kantian approaches
ix
to ethics, has been transformed by Chris, and she was a never-less-than-challenging interlocutor
whenever I presented my work in departmental workshops.
A number of other faculty members helped to make Harvard such a stimulating intellectual
environment in which to work. I learned a great deal from acting as a Teaching Fellow for
Michael Sandel, Melissa Barry and Michael Blake, and from conversations with, and/or taking or
auditing the classes of Arthur Applbaum, Stanley Cavell, Josh Cohen, Norman Daniels, Warren
Goldfarb, Richard Heck, Frances Kamm, Ed McCann, Jim Pryor, Charles Parsons, Hilary
Putnam, Mathias Risse, Amartya Sen, Alison Simmons, Richard Tuck and Andrew Williams.
It was a pleasure to be a member of such a thriving community of graduate students working in
moral and political philosophy as exists at Harvard. Versions of each of the three parts of the
dissertation were presented at different times to the Workshop in Moral and Political Philosophy,
where they received searching and deep scrutiny. I am grateful to the student members of the
Workshop, especially Kyla Ebels Duggan, Doug Edwards, Chris Furlong, Louis-Philippe
Hodgson, Liz Harman, Waheed Hussain, Sara Olack, Japa Pallikkathayil, Simon Rippon and Pat
Shin; and also to the regular and visiting faculty members of the Workshop, especially (in
addition to Tim Scanlon and Chris Korsgaard) Melissa Barry, Michael Blake, Niko Kolodny,
Jimmy Lenman, Gisela Striker and Dave Sussman.
I made a great deal of intellectual progress during my year (2002-03) as a Graduate Fellow at the
Center for Ethics and the Professions (now the Edmond J. Safra Foundation Center for Ethics),
based at the John F. Kennedy School of Government. I am grateful to the other members of the
Graduate Fellows’ seminar, especially Sara Olack, Máximo Langer and Pat Shin, for their
comments and suggestions on my work, and for their intellectual companionship during our year
at the Ethics Center. I would also like to express my thanks to Arthur Applbaum and Michael
Blake for running our weekly seminar, and to Dennis Thompson, Jean McVeigh and the staff of
the Ethics Center for looking after us all so handsomely.
For funding my year at the Ethics Center, I am grateful to the Edmond J. Safra Foundation. I am
also grateful to the Harvard University Project on Justice, Welfare and Economics for Fellowship
support, which gave me a year away from teaching and funded a visiting semester (Fall 2003) at
the University of Cambridge; and to the Woodrow Wilson National Fellowship Foundation, for a
Charlotte W. Newcombe Doctoral Fellowship.
x
For comments and suggestions on various parts of this dissertation, in addition to many of those
thanked above, I would also like to register my thanks to friends and colleagues from Cambridge
(UK), especially Paul Bou-Habib, Serena Olsaretti, and Andrea Sangiovanni, with whom I
participated in a very productive “work in progress” workshop during 2005-06. I also had
enlightening discussions of Strawson’s views with Akeel Bilgrami (when I was visiting Columbia
as a prospective graduate student), with Arthur Ripstein (when I was visiting Toronto for the
same reason), with Paul Russell (during a visit to UBC in Vancouver), and with Donald Davidson
and Philip Pettit (when each of them came to speak to the Jowett Philosophical Society in Oxford,
which I ran in 1998-99). All had some influence on what was to turn into Part Two of the
dissertation. A version of Part One was presented to the Philosophy Department of the University
of British Columbia in 2003, and I am grateful to that audience for their comments.
I started thinking about freedom and responsibility, as an undergraduate and then as a graduate
student, at Balliol College, Oxford. From this time, I’d like to thank especially my first
philosophy teacher, Helen Steward, for whom I first wrote a paper on free will, and my B.Phil.
co-supervisors, Derek Parfit and Jim Griffin. I thank my B.Phil. thesis examiner, Jerry Cohen, for
unyielding but (ultimately) beneficial criticism, and also for encouraging words about where I had
got things right. I am also grateful to the Royal Institute of Philosophy Jacobson Fund and to the
British Academy, for the financial support that got me started as a graduate student in Philosophy.
For friendship, solidarity and moral support during my time at Harvard, I would like to register
my thanks to many of those mentioned above, and especially also to Jake Beck, Dave Gray and
Simon Rippon (both for philosophical discussion and for knowing when it’s time to take a break
from philosophy); to my former housemates Chris Etheridge, Niels Janssen and Dave White, for
our bastion of shabby but civilized living in Mason Street; to my fellow philosophers Lauren
Ashwell, Selim Berker, Amanda Green, Natasha Irani, Aaron James, Paul Katsafanas, Michael
Kessler, Doug Marshall, Michael Rescorla, Andrew Roche, Kranti Saran, Jiewuh Song, Thomas
Teufel and Bharath Vallabha (in addition to the aforementioned members of the M&P
Workshop); to Ian Dowker, Agnes Kan, Carolina Mallol, Victoria Martin, Tom Oakley, Ricken
Patel, Martin Sandbu and Thad Williamson; and especially to Chris Brooke, for showing me how
Widener Library works, for Thursday night beer, and for giving me all his furniture when he
moved back to Oxford in 2000. I’m also particularly grateful to Dave Gray and Simon Rippon for
putting me up on my various visits back to Harvard since 2004.
xi
My time at Harvard was greatly enriched by visits from a number of friends from Britain, and I
thank them all for coming over. I also welcomed the distraction of two memorable trips to New
Hampshire during the Presidential primaries: with Don Conklin, Dom Sandbrook and Joe Guinan
campaigning (unsuccessfully) for Bill Bradley in 2000; and with Joe, Andrew Clark and Simon
Hooper, campaigning (again unsuccessfully) for Howard Dean in 2004. It’s gratifying that, since
I’ve left the States, Americans have finally started electing the best candidate in Presidential
elections. Another wonderful and much-appreciated distraction was a hilarious trip to the
Dominican Republic with Dave Gray and Louis-Philippe Hodgson in 2003. Equally important
was the opportunity to escape Harvard for the occasional weekend’s decompression in New York
City, Montreal, or DC, and I’m grateful to my fellow co-escapees.
Quite a lot of the thinking, reading and writing for this dissertation was done, not in the dusty
corners of the Widener Library, but in the Diesel Café and the (much-missed) Someday Café in
Davis Square, Somerville; and later in Caffè Nero in Cambridge (having discovered that good
independent coffee shops are much harder to find in East Anglia than in Eastern Massachusetts).
My quality of life while at Harvard was also greatly enhanced by the existence of The Plough &
Stars (and its Saturday morning football), the Burren, and, most of all, the People’s Republik. I’m
glad that such fine places existed.
Since leaving Harvard in 2004, my research has been supported first by St John’s College,
Cambridge (2004-07), and then by the Hallsworth Fund of the University of Manchester (since
2007). While based at Cambridge and then at Manchester, I’ve mostly been engaged in research
projects in political philosophy, unrelated to this dissertation. Nevertheless, given that this
dissertation was completed during this time, I would like to thank The Master and Fellows of
St John’s College, Cambridge, for generous support and for providing a superlative working
environment; and the Hallsworth Fund, as well as the members of the Politics Discipline Area,
the Philosophy Discipline Area and, especially, the Centre for Political Theory at Manchester, for
further generous support and for a working environment that, while much less beautiful and less
luxurious than St John’s, is nevertheless every bit as intellectually stimulating.
I thank my wife, Mary Leng, for her patience, support, love and understanding. Mary lived in
Toronto during my first three years at Harvard, and I thank her flatmates Lee Churchman, Bernie
Morgan and Hani Kim for making me feel so welcome during my frequent visits. I’d also like to
thank Mary’s mum and dad, Paul and Gwynne Leng, for their endless hospitality and friendship.
xii
It’s been a great pleasure to have become part of their family during the time I’ve been working
on this dissertation. Moreover, I’d like to thank their next-door-neighbours in Willaston, Ralph
and Maureen Bennett, for inviting Mary and I to housesit for them while they were in Australia in
Spring 2007: versions of Part Two were written at their dining-room table during this time.
From time to time I’ve had the splendid opportunity to accompany Mary, in what I think of as the
“Denis Thatcher” role, on a number of academic trips. Parts of this dissertation were drafted
while we were the guests of the Peter Wall Institute for Advanced Study at the University of
British Columbia in July 2003, and while we were being hosted by the Department of Philosophy
of the Universidade de Santiago de Compostela in March 2007. I thank both institutions for
generously extending their excellent hospitality to Mary’s other half.
This dissertation is dedicated to my parents, Tom and Liz O’Neill. I can only imagine that I’d
have finished more quickly if I’d had to face the prospect of justifying my sporadic progress to
my dad, who was no man for excuses. I only wish he had still been around to tell me off for
slacking (not that I was, of course…). But I’m sure he would have been delighted to see his son
with a Harvard Ph.D., even if it did take me a while to get one. My mum, Liz O’Neill, has been
an incredible source of love and support for all of my life. I know it must have been difficult for
Mum to see her only child head off to the US on such an open-ended assignment as this, and I
thank her for being so positive and encouraging about the move. I thank her also for organizing
seemingly endless lifts to and from Heathrow for me, for which I also thank Mum’s neighbour,
Denis Keane and my loyal uncle, Tim Sheehan. This finished version is commended to Mum
with love and admiration.
It would have amused my former selves to know how long I’ve worked on this dissertation before
finally letting it go. In one respect, it has been nice to be able to return intermittently to these
thoughts over the last few years, while my primary research interests have focussed elsewhere;
nevertheless, it is a great pleasure to have brought the project to its conclusion.
My final thanks are to my son, Thomas Paul O’Neill. Tommy’s immanent arrival was just the
world-changing impetus that I needed to get myself into gear to finish. Without even trying,
Tommy managed, even before his birth, to be a fantastically good influence on his dad. I look
forward to the day when we can discuss the claims and arguments that follow. Heaven knows
what Tommy will make of his dad’s unusual views…
xiii
This dissertation is dedicated to my mother,
Elizabeth O’Neill,
with thanks for her constant love and support,
and in memory of my dearly-missed father,
Thomas O’Neill
(1931-1995)
xiv
Epigraph
I have sometimes criticized the views of other philosophers, dead or living. Such adverse criticism is a form of compliment. It is only the very best with whom it is worth while to differ.1
1 P. F. Strawson, (1998), “Intellectual Autobiography,” in Lewis Edwin Hahn, ed., The Philosophy of P. F. Strawson, (Peru, Illinois: Open Court Publishing), p. 21.
1
Freedom, Fairness and Responsibility
Part One
Agency, Autonomy and Desire:
Or, Rescuing the Rational Wanton
2
Freedom, Fairness and Responsibility – Part One
Agency, Autonomy & Desire: Or, Rescuing the Rational Wanton
1. Introduction
This first part of ‘Freedom, Fairness and Responsibility’ is a treatment of some of the
important issues in moral psychology and the philosophy of action that are raised in the
work of Harry Frankfurt. Its aim is to question a number of Frankfurt’s central and most
highly influential claims: on the relationship between desire and volition, on the
conditions for autonomy, freedom of the will and moral responsibility, and on the criteria
for personhood.
The focus of Frankfurt’s work shifts considerably between his earliest papers and his
more recent work. Earlier articles, such as ‘Alternate Possibilities and Moral
Responsibility’ and ‘Freedom of the Will and the Concept of a Person’2, are concerned
with laying down a “hierarchical” account of the nature of free agency, moral
responsibility and personal identity. Here, Frankfurt is seeking to explain which features
or properties agents need to have in order to be active with regard to their mental lives in
the way needed for autonomy and responsibility.
2 Harry G. Frankfurt, (1969), “Alternate Possibilities and Moral Responsibility,” Journal of Philosophy, LXVI, no. 23, 829-39; and (1971), “Freedom of the Will and the Concept of a Person,” Journal of Philosophy, LXVIII, no. 1, 5-20. Both are reprinted in Harry G. Frankfurt, (1988), The Importance of What We Care About, (Cambridge: CUP). References to the latter article will henceforth be to “FW&CP”; page references are to its reprinting in the 1988 collection.
3
In later articles, from ‘Identification and Wholeheartedness’3 to ‘Autonomy, Necessity
and Love’4, Frankfurt’s interest in mental activity takes on a much less ‘instrumental’
form. His central task changes from that of undertaking an examination of the grounds
for moral responsibility, to an investigation of the conditions for psychic integration and
for the avoidance of alienation and ambivalence, considered as good in themselves. As
Sarah Buss and Lee Overton put it, “he has shifted his focus from the self-control
sufficient for moral responsibility to the self-integration that, he claims, is an implicit
goal of every human action.”5 As T. M. Scanlon has it, Frankfurt’s work has undergone a
shift of emphasis “from concern with an agent’s “ownership” of his or her desires as a
precondition for moral appraisal to a concern with an ideal of psychic health.”6
Nevertheless, despite these differences in emphasis, there is a striking degree of
continuity in Frankfurt’s writings. For example, it is significant that Frankfurt’s
conception of the nature of desire remains constant. So too does his central concern with
self-alienation, in its different forms: whether the self-alienation which Frankfurt discerns
in the phenomenon of ambivalence, wherein we find ourselves identifying with mutually
3 Also in Harry G. Frankfurt, ibid., (1988). Originally published in Ferdinand David Schoeman, ed., (1987), Responsibility, Character and the Emotions: New Essays in Moral Psychology, (New York: Cambridge University Press).
4 In Harry G. Frankfurt, (1999), Necessity, Volition and Love, (Cambridge: CUP). Originally published in H. F. Fulda and R.-P. Horstmann, eds., (1994), Vernunftbegriffe in der Moderne: Stuttgarter Hegel-Kongress 1993, (Stuttgart: Klett-Cotta).
5 Sarah Buss and Lee Overton, (eds.), (2002), ‘Introduction’, p. xii, in their Contours of Agency: Essays on Themes from Harry Frankfurt, (Cambridge, MA: MIT Press).
6 T. M. Scanlon, (2002), ‘Reasons and Passions’, in Buss and Overton, (eds.), pp. 165-183.
4
incompatible goals, or the self-alienation of being gripped by unwelcome desires or
impulses. We can explain this fixity of underlying concern in Frankfurt’s work with
reference to his concern with understanding the nature of human autonomy. Frankfurt’s
central questions remain these: when are we active rather than passive? when should we
say that an agent truly identifies with his desires and actions, as opposed to it being the
case that his actions merely happen to him, and his desires merely assail him?
Frankfurt’s pursuit of this set of questions is unified around the aim of making sense of
what it is to be an accountable, responsible individual in terms of what it is to be a
coherent and autonomous agent. It is an account on which the specification of the
conditions for responsibility is conducted against the background of the specification of a
distinctive conception of autonomy. On Frankfurt’s view, we are responsible when and
only when we are active and autonomous, and we are autonomous when and only when
our will is unified, internally coherent and properly structured.
In this paper I will be arguing against some of the central claims of Frankfurt’s earlier
work, but in so doing I shall be discussing themes which recur and ramify throughout his
writings. Contrary to Frankfurt’s claims, I want to show the following:
(1) That a hierarchical volitional account does not capture the nature of
personhood, or of freedom of the will or of responsibility;
(2) That the possession of ‘2nd order volitions’ is neither necessary nor
sufficient for personhood, or for rationality, or for freedom or moral
accountability;
5
(3) That Frankfurt misconceives the nature of desire and, with it, both our
relationship to our own desires, and the relationship between desire and
reason.
Frankfurt’s strikingly stark and simple account of autonomy appears to offer us a key to
answering a set of recalcitrant philosophical problems about agency and accountability.
By showing the ways in which that striking account nevertheless fails to be true to the
complexity of the phenomena, I hope to do two things. The first is to clear the ground for
a more complex or pluralist account of autonomy, which is fairer to the complexities of
the facts about the various ways in which agents can be coherent, active and integrated.
The second aim, consequent to the first, is to derail the line of thought which takes us too
quickly from an account of autonomy to an account of responsibility, thereby severing
the too-close link to which these concepts have come to be subject. In examining the
some of the central claims made by Frankfurt in his influential treatment of these issues,
part of my aim is to set the stage for my examination of P. F. Strawson’s arguments in
Part Two of the dissertation, and for my presentation and defence of my own ‘Hybrid
View’ about responsibility and freedom in Part Three.
2. Second-Order Desires and the Concept of a Person:
Frankfurt begins ‘Freedom of the Will and the Concept of a Person’ by rejecting P.F.
Strawson’s analysis of the concept of a person as that of “a type of entity such that both
predicates ascribing mental states of consciousness and predicates ascribing corporeal
6
characteristics … are equally applicable to a single individual of that single type.”7 Such
a criterion for personhood is too wide, as it could plausibly catch members of animal
species which, though they had some (perhaps rather simple) mental life, we would
rightly be reluctant to talk of as ‘persons’. What is needed, Frankfurt tells us, is a rather
narrower criterion, given that “… what is essential to persons is a set of characteristics
that we genuinely suppose – whether rightly or wrongly – to be uniquely human.”8 The
narrower criterion upon which Frankfurt settles is connected with the “structure of a
person’s will”.
Only persons, and not other creatures which enjoy a semblance of mentality, are able to
form “second-order desires” or “desires of the second order”. Such “second-order
desires” are, simply, desires which take as their object some particular simple first-order
desire, or set of first order desires, that an individual might have. First-order desires, here,
are “simply desires to do or not to do one thing or another”.9 A “second-order volition”,
in Frankfurt’s terminology, is a particular kind of “second-order desire”, whose content is
not simply that we have a first-order desire of a particular sort, but that this particular
first-order desire moves us to action. A second-order desire to φ involves us desiring to
desire to φ. A second-order volition to φ involves us desiring to desire to φ, and, at the
same time, desiring that the desire to φ is “effective” in bringing us to φ.10
7 P. F. Strawson, (1959), Individuals, (London: Methuen), pp. 101-102.
8 FW&CP, p. 12.
9 FW&CP, p. 12.
10 See FW&CP, pp. 14-16. See also, Scanlon, (2002), pp. 174-175.
7
A person, then, on Frankfurt’s view, is an individual that has “the capacity for reflective
self-evaluation that is manifested in the formation of second-order desires.” Persons
“want to have (or not to have) certain desires and motives” as well as “wanting and
choosing and being moved to do this or that.”11 Let us then take this Frankfurtian claim
about personhood:
P1: X is a person iff X has the capacity to form second-order desires.
This initial proposed criterion of personhood runs up against the possibility of an agent
who possesses the capacity to form second-order desires, but who never forms any
second-order volitions. Such an individual could be a ‘Gourmet Connoisseur’ of first-
order desires, subjecting them to careful and compelling criticism, but this self-criticism
might never link up with his agency in the right way. But this is a danger to which
Frankfurt is alive, and he thus refines P1 so that it makes references to the idea of
“second-order volitions”, rather than simply being couched in terms of second-order
desires. Frankfurt thereby comes to the considered view that, “… it is having second-
order volitions, and not having second-order desires generally, that I regard as essential to
being a person.”12 So, we should replace P1 by P2, which takes care of the case of the
inefficacious connoisseur of first-order states.13 Thus, we have:
11 FW&CP, p. 12.
12 FW&CP, p. 16.
13 We might nevertheless question why, as on Frankfurt’s view, the ‘Gourmet Connoisseur’ should not properly be called a person. In response, we might at least say that the Gourmet Connoisseur would be an individual whose personhood was not made manifest through his agency. The somewhat peculiar and specialist sense in which Frankfurt uses the term ‘person’ is apparent in this case; it will be discussed in more detail in what follows.
8
P2: X is a person iff X has (the capacity to form) second-order volitions.
So, does P2 get things correct about the nature of personhood? The first thing to say is
that it does have a tremendous initial plausibility. It marks human beings apart from the
lower animals, and explains why the possession of mental attributes is not, in itself,
sufficient for personhood. Moreover, it helps us to make sense of the thought that,
although we are always persons (so long as we retain the ability or capacity to form
second-order volitions, or perhaps as long as we sometimes have such volitions), we also
sometimes fail to manifest our capacity for personhood, or fail to be guided by such
volitions (as when we fail to subject our first-order desires to critical scrutiny).
However, pressing problems arise when we question whether having (the capacity to
form) second-order volitions is really both a necessary and a sufficient condition for
attributions of personhood to individuals. One reason why we might think that it was not
in fact a sufficient condition for personhood arises when we consider the fact that it is
quite possible to form second-order volitions which have as their object the first-order
desires of others. Sonya wants Raskolnikhov to want to repent of his sins, and to want to
throw himself on the mercy of the authorities, and she wants for those desires to be
effective in generating actions. But this second-order volition does not thereby conjure up
a person, composed of the mereological aggregate of Sonya’s second-order volitions and
Raskolnikhov’s first-order desires, even if, as a matter of fact, Sonya’s second-order
9
volitions pertaining to Raskolnikhov are satisfied, and even if Raskolnikhov is
accordingly moved to act by the desire that Sonya wanted him to have.
This worry might seem somewhat exotic, but it is difficult to avoid insofar as Frankfurt’s
P2 is taken to be giving an account of the nature of personhood. Thus, in order to avoid
the Sonya-Raskolnikhov problem, it seems that what is needed, in addition to P2, is a
prior way of assigning desires to individuals, and thus a prior way of individuating
persons. Without it, we could be led to the unwelcome position of positing a new person
whenever any of us forms a desire about the effective desires of another. Or, even if this
danger could be avoided, we would at least be faced with the problem of calling X a
person even if all of X’s second-order volitions took as their objects the desires of others
(where X desires not only that others have some particular desires, but also that those
desires be effective in motivating the actions of those others). It is surely at least a
conceptual possibility that there could exist an individual who went in for intricate and
involved criticism of the desires of others, forming second-order volitions which took
those desires as their objects, without placing himself under the same sort of scrutiny. (In
fact, far from being a mere conceptual possibility, we might think that this is an all-too-
readily identifiable character type.) If such an individual really could exist, then it is far
from clear that Frankfurt would want us to call such an individual a person.
What these considerations point us towards is the thought that we need a conceptually
prior criterion (or set of criteria) for individuating persons before we can apply
Frankfurt’s test for ‘personhood’. Thus, it is not the case that Frankfurt is presenting an
10
account of the nature of personal identity which could act as a rival to the accounts of
personhood which we can find in the work of, for example, Parfit, Shoemaker, Wiggins
and Williams.14 Rather, he is presenting a narrower and more specialized conception of
‘person’, which finds its use only after the prior issue of personal individuation has been
settled. Given this, Frankfurt’s initial presentation of his project as taking its point of
departure from P. F. Strawson’s Individuals is somewhat misleading. Thus, we might best
understand Frankfurt’s project as suggesting a novel, replacement usage for the term
‘person’; that is, as suggesting a sort of ‘normativized’ conception of ‘personhood’ which
stands in distinction to the more descriptive usage of these other writers. In doing so,
Frankfurt is best seen, like Locke, as outlining a forensic conception of personhood –
connected to the justification of accountability and moral responsibility.15 But
furthermore, going beyond these forensic considerations, it is a conception of personhood
that connects to an independently significant idea of agency or activeness.16 This further
idea presents an idea of genuine agency of importance for the self-understanding of
agents as well as for their (forensic) assessment by others.
14 In, for example, Derek Parfit, (1984), Reasons and Persons, (Oxford: OUP); Sidney Shoemaker, (1963), Self-Knowledge and Self-Identity, (Ithaca, NY: Cornell UP); Bernard Williams, (1973), Problems of the Self, (Cambridge: CUP); David Wiggins, (1967), Identity and Spatio-Temporal Continuity, (Oxford: Blackwell).
15 For Locke’s understanding of “person” as a forensic term, see John Locke, (1975) [1689], ed. Peter H. Nidditch, An Essay Concerning Human Understanding, (Oxford: Clarendon Press), Book II, Chapter XXVII, §26, p. 346. As Locke puts it: “[Person]… is a Forensick Term appropriating Actions and their Merit; and so belongs only to intelligent Agents capable of a Law, and Happiness and Misery.”
16 For the role of ideas of personhood, identification and activity in Frankfurt, see the illuminating discussions in Scanlon, (2002), ibid., and in Richard Moran, (2002), “Frankfurt on Identification: Ambiguities of Activity in Mental Life,” also in Buss and Overton, eds., (2002), pp. 189-217. See also Frankfurt’s “Reply to T. M. Scanlon” and “Reply to Richard Moran,” both also in Buss and Overton, eds. (2002), pp. 184-188 and 218-225.
11
3. Volitional Hierarchy, Judgment-Sensitivity, Agency and Personhood
The most important problem with P2, though, is of a quite different sort. It is generated
by the implausibility of P2 as a necessary condition for personhood. This might seem like
a strange claim, given the initial intuitive plausibility of P2. It is nevertheless an
important claim, given the larger purposes of this discussion, and what I want to go on to
say about Frankfurt’s misconception of the nature of desire. To see why P2 might fail to
present a necessary condition for personhood, it would be helpful to consider a fictional
character, whom we’ll call Reasonable Rebecca. Unlike Sonya-Raskolnikov and the
Gourmet Connoisseur, whose role was simply to cast some light on the aims of
Frankfurt’s project, and his understanding of the term ‘person’, Rebecca’s role is more
significant, and connects more closely to the substance of Frankfurt’s view. Accordingly,
she will receive much more attention than these other motley characters in what follows.
Reasonable Rebecca. Rebecca is an individual who gives a great deal of time and
attention to deciding what to do. She has a substantive and carefully-worked out
conception of what counts as a good reason for action. She combines this with a keen
ability to perceive how the world which she inhabits generates considerations to act in
one way rather than another, in the light of her conception of what does and does not
count as a good reason. For example, Rebecca sees the venality and shoddiness of the
current administration in her country as a reason to vote against it, the significance of
health as a reason to take exercise, and so on. To borrow some terminology from T.M.
Scanlon, Rebecca has “judgment sensitive attitudes”17 which are richly nuanced,
17 For the notion of a ‘judgment sensitive attitude’, see T.M. Scanlon, (1998), What We Owe To Each Other, (Cambridge, Mass.: Harvard University Press), pp. 18-22. See also Scanlon, (2002), pp. 168-169.
12
coherent and reliable. When she is unsure what to do, she thinks long and hard about the
relevant features of the different courses of action open to her, and assesses the
considerations in favour of and against those courses of action with great care and
attention.18
If one takes any virtue which one might plausibly like to hold as being constitutive of a
rational agent, Rebecca has it. That is, unless that virtue involves some kind of self-
reflection: for this is something which Rebecca just doesn’t go in for. Her deliberation is
turned outwards to the world, rather than inwards to her own mental life. Her deliberative
processes, with regard to questions of what to do, involve careful weighings of the
considerations which speak for or against particular actions, but do not involve the
assessment of her own desires (indeed, they may not involve the assessment of any of
Rebecca’s own “judgment sensitive attitudes”). Rebecca is extremely reflective, but not
even minimally reflexive. She lacks the kind of volitional hierarchy which Frankfurt
takes to be constitutive of personhood.
Frankfurt seems often to fail to mark the distinction between the separate properties of
reflectiveness and reflexiveness. The first involves the capacity for deliberation and of the
mediation of action by careful consideration of what one should do. The second, more
specifically, involves self-assessment, or at least the assessment of some particular items
of one’s inner mental life. It involves the process of reflection being turned inwards,
18 Here, I deliberately echo T. M. Scanlon’s understanding of ‘desire’ as explicable in terms of the idea of taking some consideration to provide a reason. See his (1998), pp. 6-8, 18, 37-55. Rebecca sees many considerations as providing reasons, and thus, accordingly, can be seen as having the desires corresponding to each particular perception of a consideration as counting as a reason.
13
whereas many forms of unreflexive deliberation (that is, of reflection) can remain turned
resolutely outwards.
Now, perhaps Rebecca long ago decided that the best way to become a skilful and
reasonable agent was to train and hone her first-order reactions to the world, and her
second-order capacities have simply withered on the vine. Perhaps she had a particular
kind of view about the demands of psychic health, which emphasized the importance of
achieving ‘flow’, and decried the way in which self-criticism or excessive reflexivity
leads to bad psychic energy.19 Or, perhaps, through some fluke of nature or nurture,
Rebecca just never really had the right kind of second-order capacities. Very likely, the
biographical explanation of her current state will be of minimal importance.
My operative claim, which Frankfurt would have to reject, is that Rebecca has lost very
little of normative significance in being bereft of such second-order volitions, or even in
being bereft of the capacity to generate such second-order volitions. Certainly, she has
not lost enough to disqualify her from ‘personhood’, in any way in which the property of
being a ‘person’ might seem to be important. Reasonable Rebecca, I suggest, stands as a
stark counterexample to P2, as the possibility of her existence speaks against the claim
that having (the capacity to form) second order volitions could be a necessary condition
for ascriptions of personhood, in Frankfurt’s ‘normatively loaded’ sense of the term.
(That is, at least insofar as ascriptions of personhood are taken to be normatively
19 For the concept of ‘flow’ in positive psychology, see Mihályi Csíkszentmihályi, (1990), Flow: The Psychology of Optimal Experience, (New York: Harper and Row). See also Csíkszentmihályi, (2000), Beyond Boredom and Anxiety: Experiencing Flow in Work and Play, (San Francisco: Jossey-Bass).
14
significant from the point of view of making judgments about an individual as a practical
agent.)
Frankfurt would describe Rebecca as a “wanton”, making use of a concept which he
demarcates as a contrast class to that of “persons”. “Wantons”, in Frankfurt’s description,
are those individuals who “have first-order desires but who are not persons because,
whether or not they have desires of the second order, they have no second-order
volitions.”20 In other words, wantons lack the kind of volitional hierarchy which persons
enjoy. So, Rebecca certainly fits the bill, in that she has first-order desires and yet lacks
second-order volitions. But, if the labelling of an individual as a “wanton” is meant to
suggest the impoverishment of their practical rationality, or the failure of their agency, in
contradistinction to the “person”, who enjoys the kind of practical rationality worth
wanting, then Rebecca’s case makes it hard to see the force or interest of the distinction.
This will be especially clear when we compare Rebecca to another character, who we’ll
call Arbitrary Alex. Alex demonstrates another way – quite separate to the Raskolnikhov-
Sonya problem of trans-individual second-order volitions – in which P2 might appear to
be a failure as a sufficient condition for ascriptions of “personhood”. (Again, at least
where such ascriptions are taken to be significant in terms of an individual’s practical
rationality or his status as an agent.)
20 FW&CP, p. 16.
15
Arbitrary Alex: Alex is an unusual, annoying and unpleasant fellow. He has
extremely crude first-order desires, which generally present themselves to him as
particular sorts of quasi-physical urges. When Alex read Descartes as an undergraduate,
he was extremely impressed by Descartes’s definition, in The Passions of the Soul, of
desire as “an agitation of the soul, caused by the [animal] spirits, which disposes it to will
for the future the things it represents to itself to be suitable.”21 This kind of hydraulic
metaphor seemed to Alex to characterize almost exactly the nature and phenomenology
of desire. He was also much impressed by the line-drawing of Le Desir by the illustrator
employed by Descartes to sketch the manifestations of the different passions. This
picture, Alex thought, bore a very striking physical resemblance to what he himself
tended to look like when he found himself assailed by a desire. (see Figure 1)
21 See René Descartes, (1989) [1649], The Passions of the Soul, ed. Stephen Voss, (Indianapolis: Hackett), p. 66. Note that, with Descartes’s characterization of the nature of desire, we have an interesting hybrid of a quasi-hydraulic conception (“an agitation of the soul”), with the positing of an underlying mechanism whereby one represents to oneself a judgment of suitability (“which disposes it to will for the future the thing it represents to itself to be suitable.”)
16
Figure 1. Cartesian Desire 22
Now, as has been mentioned, Alex’s desires are crude and unsophisticated. They lack the
sensitivity and coherence of Rebecca’s judgment-sensitive attitudes, being quite
disconnected from any underlying conception of what does or does not constitute a good
reason for action. Alex’s perception of the world as a repository of considerations relating
to reasons to act is of the strangest, and seemingly most faulty nature. He is regularly
seized by the desire to empty bottles of delicious and pricy wine into the bath, just for the
sake of it, and finds himself frequently intending to vote for shoddy and venal politicians.
22 The line-drawing of Le Desir occurs at p. 68 of Descartes, (1989) [1649], ibid.
17
In his most extreme moments, he finds himself compelled to seek out a saucer of mud, or
to busy himself with the counting of blades of grass.23
Where Rebecca’s first-order desires were well-ordered and cohesive, Alex’s are random
and unpredictable. Thus, their first-order desires differ from one another’s in important
respects aside from their relative calmness or violence. It is true that Rebecca’s desires
are calm, whereas Alex’s are violent. But of greater significance are the facts that
Rebecca’s are generated by keen and careful judgment, whereas Alex’s are not;
Rebecca’s are characterized by rich and sophisticated interrelations and connections,
whereas Alex’s are disconnected from one another; and Rebecca’s are constant and
reliable over time, whilst Alex is prone to sudden reversals and lurches of interest.
Nevertheless, Alex has something that Rebecca lacks. He has second-order volitions. As
he watches the passing carnival of his first-order desires, he forms second-order volitions
with regard to which of those first-order desires he would like to be effective in
determining his will. We can remain, for the moment, uncommitted on the question of
how coherent or cogent Alex’s second-order volitions might be. Perhaps he manages, by
a Herculean feat of will and ingenuity, to impose order on the swirling storms of his first-
order desires. Or perhaps his second-order volitions are themselves mutually
contradictory, inconstant and bizarre. We can leave this matter aside because the mere
23 On wanting a saucer of mud, see Elizabeth Anscombe, (2000) [1957], Intention, (Cambridge, Mass.: Harvard University Press); on having a plan of life that gives a central place to counting blades of grass, see John Rawls, (1971), A Theory of Justice, (Cambridge, Mass.: Harvard University Press), §65, p. 379.
18
fact that he has second-order volitions is enough, in Frankfurtian terms, to make Alex
into a person (in Frankfurt’s sense of the term) by P2.
Now, given this result, it seems to me that we can go one of two ways. Firstly, we can
simply reject P2 as a criterion of personhood, on the grounds that the case of Reasonable
Rebecca shows it to fail as a necessary condition for personhood, while the case of
Arbitrary Alex (certainly if coupled with a similarly alarming story about his second-
order volitions) shows P2 to fail as a sufficient condition for personhood. Or, we can
allow that P2 is the correct criterion for determining judgements of personhood, while
simply concluding that such judgments lack any genuine normative significance. My own
view is that we should take the first of these two options, and reject the plausibility of P2
as a criterion for deciding whether individuals are persons or not.
Now, one issue which has hitherto been to some degree lurking in the background is that
of the relationship between personhood and agency. As we’ve seen, Frankfurt thinks that,
in order for an individual to be a person, that individual needs to be able to demonstrate
the sort of volitional hierarchy described in P2. Put simply, Frankfurt sees this sort of
volitional hierarchy as a necessary precondition which must be met if an agent is to be
considered as genuinely active. Only when an individual is acting in accord with a
successful second-order volition is he displaying the sort of activeness with respect to his
desires which is the hallmark of full-blown agency. When an individual lacks operative
second-order volitions, Frankfurt thinks that we must view that individual as being
merely passive with respect to her first-order desires. Such an individual may go in for
19
some measure of activity – she may, after all, in a loose way of talking, do certain things
– but she is not an agent.
But, again, there is a lack of plausibility to this claim. For a start, it seems to me
outlandish to deny that Rebecca is an agent. She seems, indeed, to be an altogether
exemplary kind of agent – active and unified, and undertaking careful and cohesive
action-directed deliberation. This leads us to the issue of what we should say about Alex.
Here, one is tempted to think, it would indeed all depend on Alex’s second-order
volitions. But, rather than depending simply on whether or not such volitions were
operative, as Frankfurt would have it, it might seem much more plausible to think that it
would instead largely depend on the content of these second-order volitions. Consider
these two cases:
Herculean Alex: This version of Alex has first-order desires – and, in
general, first-order judgment-sensitive attitudes – which display the faults discussed
above. But, by a Herculean feat of will and ingenuity, Alex imposes order on the swirling
storms of his first-order desires. His second-order volitions have an admirable
cohesiveness and systematicity, and he is able thereby to construct an admirably unified
life as an agent from the unpromising materials of his first-order mental life.
Utterly Arbitrary Alex: This version of Alex has second-order volitions which
display all the incoherence and flux of his first-order mental life. Take any of the
dimensions on which Alex’s first-order mental life might be held to be faulty (lack of
20
coherence, lack of reason-responsiveness, diachronic disunification, or whatever other
lamentable property) and that fault reoccurs at the second-order level. His second-order
volitions are reliably successful in causing actions, however. Alex’s will may be a mess,
but it’s a strong will. If he decides to do something, then he does it.
Now, it seems to be plausible for us to induct Herculean Alex into the hallowed hall of
genuine agents, to take his place alongside Rebecca. But, it seems to me quite
implausible to allow Utterly Arbitrary Alex to join them there, insofar as agency is taken
to involve being active rather than passive in the face of one’s desires. There’s a strange,
rather formal sense in which we might indeed see Utterly Arbitrary Alex as displaying the
sort of ‘activeness’ with respect to his first-order mental life, of a sort which could
qualify him as an agent. But, when looked at more carefully, it seems more plausible
simply to think of this Alex as a horrific mess, falling far short of the conditions for
genuine agency. In the bric-a-brac filled loft of Alex’s mental life, there is simply not
enough cohesion or unification. It is true that some of Alex’s volitional mental items take
others of his mental items as their objects, but it is far from obvious why this fact, on its
own, could have the kind of significance with which Frankfurt invests it.
One could imagine the most afflicted and demented individual imaginable whilst still
imagining someone who could form second-order volitions, and who could carry them
through to satisfaction with impressive regularity. Conversely, as in the case of Rebecca,
one can imagine a controlled and admirable agent whose deliberation is turned resolutely
outwards, rather than back onto the contents of her own mental life. This surely suggests
21
that Frankfurt has gone wrong in thinking that the possession of second-order volitions is
the unique and necessary hallmark of activeness, and the sort of (genuine) agency
characteristic of persons.
4. Rescuing the Rational Wanton: Personhood, Rationality and Desire
We shall now turn to some elements of Frankfurt’s discussion of rationality in ‘Freedom
of the Will and the Concept of a Person’, and of the rationality which can nevertheless be
displayed by the ‘wanton’. It is important to be clear that Frankfurt does not make the
claim that rationality is itself constituted – as is personhood and genuine agency – by the
capacity for volitional hierarchy. Having told us that “the essential character of a wanton
is that he does not care about his will,”24 Frankfurt makes clear his view about the
relationship between personhood, rationality and “the wanton” in this (very rich) passage:
The fact that a wanton has no second-order volitions does not mean that each of his first-order desires is translated heedlessly and at once into action. He may have no opportunity to act in accordance with some of his desires. Moreover, the translation of his desires into action may be delayed or precluded either by conflicting desires of the first order or by the intervention of deliberation. For a wanton may possess and employ rational faculties of a high order. Nothing in the concept of a wanton implies that he cannot reason or that he cannot deliberate concerning how to do what he wants to do. What distinguishes the rational wanton from other rational agents is that he is not concerned with the desirability of his desires themselves. He ignores the question of what his will is to be. Not only does he pursue whatever course of action he is most strongly inclined to pursue, but he does not care which of his inclinations is the strongest. Thus a rational creature, who reflects upon the suitability to his desires of one course of action or another, may nonetheless be a wanton.25 In maintaining that the essence of being a person lies not in reason but in the will, I am far from suggesting that a creature
24 FW&CP, p. 16.
25 So here, at any rate, Frankfurt does draw the important distinction between ‘reflection’ and ‘reflexion’.
22
without reason may be a person. For it is only in virtue of his rational capacities that a person is capable of becoming critically aware of his own will and of forming volitions of the second order. The structure of a person’s will presupposes, accordingly, that he is a rational person.26
This is a crucial part of Frankfurt’s overall argument because it is extremely revealing as
to how Frankfurt conceives of the nature of reason and desire, and of the relationship
between the two of them. Thus, it will hopefully prove useful to examine some of the
claims made here by Frankfurt against the background of our earlier introduction to
Reasonable Rebecca, a particularly complex and sympathetic variety of “wanton”.
Let’s start with the claim that “the essential character of a wanton is that he does not care
about his will”. Well, really, this isn’t true on Frankfurt’s own view. The essential
character of a wanton is that , whilst possessing first-order desires, the wanton does not
have (the capacity to form) second-order volitions. But it is not at all clear that this is
really equivalent to saying that the wanton does not care about his will. In order to assess
this claim of Frankfurt’s, taking Rebecca’s case as a testing ground, we need to ask the
question of whether we really want to say that she, as our paradigm case of a ‘wanton’,
really does not care about her will.
Now, one’s ‘will’, we should recall, is defined by Frankfurt to be “the notion of an
effective desire – one that moves (or will move) a person all the way to action.”27
(Although here we will have to take this reference to a ‘person’ as a slip, which should be
26 FW&CP, p. 17. My italics.
27 FW&CP, p. 14.
23
replaced by reference to an ‘individual’, to use a term which is suitably general, with
regard to covering both ‘persons’ and ‘wantons’.) For, if Frankfurt is really serious in
saying that a will is something that only a person and not a wanton can have, then it is
rather obscure as to what he could mean when he says that the wanton “doesn’t care
about his will”. If a will is something that, ex hypothesi, a wanton does not have, then it is
trivially true that the wanton does not care about his will. At any rate, Frankfurt’s idea
here is that “an agent’s will, then is identical with one or more of his first-order
desires”,28 and the specific first-order desire which it is identical to is that specific first-
order desire on which an agent acts in some particular case.
So, to return to Rebecca’s case, Frankfurt’s claim would have to be that Rebecca, because
she is a wanton, does not care about the first-order desire on which she acts. Now, I will
grant that Rebecca forms no judgment-sensitive attitudes that are explicitly about her
first-order states, and so it is quite true that she does not care about her will, de dicto. She
has no ‘cares’ that take as their object her own ‘will’, as such. Nevertheless, it would be
wrong to say that Rebecca did not care about her will. She cares enormously about her
perceptions of considerations as reasons, her assessment of these considerations, and her
choices of which considerations on which to act. The judgment-sensitive attitudes which
move Rebecca to act are something that she cares deeply about: it is because she cares so
deeply that she spends so much time on careful deliberation and reflection. Her care for
her will is something that is demonstrated by the way in which she governs herself, and
28 FW&CP, p. 14.
24
the way in which she conducts her deliberations. It is not, in itself, simply an additional
mental item, as in Frankfurt’s view.
Rebecca’s care and concern about her will is a feature of her active relationship to her
reasons for action, in general, and not a ‘mental object’ which can be pointed to within
some kind of volitional inner space. To look for Rebecca’s ‘care about her will’ as a
specific desire or attitude among others is to commit the same kind of category mistake
as Ryle’s tourist in Oxford, who, having visited the colleges, libraries and lecture halls,
asked where he might find the University.29 Or, to put things another way, we might
instead say that, while Rebecca doesn’t care about her will “as her will”, it would be false
to say that she did not nevertheless care very much about her will, de re.
Thus, it would seem that we can confidently reject Frankfurt’s first claim about the
wanton, that he “does not care about his will”. In a similar way, we can call into question
three of Frankfurt’s other claims from the above passage. Specifically:
(1) “What distinguishes the rational wanton from other rational agents is that he is not
concerned with the desirability of his desires themselves.”
(2) “[The wanton] ignores the question of what his will is to be.”
(3) “[The wanton] does not care which of his inclinations is the strongest.”
29 Gilbert Ryle, (1949), The Concept of Mind, (London: Hutchinson).
25
If these claims are taken to be substantive claims about the cares and concerns of the
rational wanton (like Rebecca), and not construed narrowly as de dicto claims about the
content of specific second-order volitions, then they are all false. Rebecca, contra
Frankfurt, is concerned with the “desirability of her desires”, does not “ignore the
question of what her will is to be” and does, indeed, “care which of her inclinations is the
strongest”. Alex, on the other hand, at least in his more arbitrary manifestations, does not
really care about the “desirability of his desires”, other than opting for one over the other
as a brute act of will; he does, in an important sense, “ignore the question of what his will
is to be”, insofar as he has no interest in that question as a central normative concern; and
he does not care – not really – about “which of his inclinations are the strongest”. So,
again, Frankfurt’s vision of the volitional and agential impoverishment of the ‘wanton’
must be seen as hyperbolic.
So much for the normatively significant perils of wantonness. I want now to turn to what
Frankfurt has to say about reason, and rationality. As has already been mentioned,
Frankfurt does not take having (the capacity to form) second-order desires to be a
necessary condition for some exercises of rationality. As he allows, “nothing in the
concept of a wanton implies that he cannot reason or that he cannot deliberate concerning
how to do what he wants to do.” This might initially lead us to think that Frankfurt is
working with a purely instrumental conception of practical rationality. For example, we
might say that a heroin addict, even if he was a pure wanton with regard to his desire for
heroin, could be rational insofar as he went about securing the means to his end of
finding the next fix, or perhaps even if he went about organizing his heroin use activities
26
diachronically. If Frankfurt’s conception of rationality really were purely instrumental,
then the question of whether some particular individual was rational would be perfectly
orthogonal to the question of whether that individual was a ‘person’ or a ‘wanton’.
Rationality, on this view, is simply an executive capacity connected with taking the best
route to fulfil one’s operative desires, regardless of whether those desires had been
endorsed by a second-order volition.
However, as the second of the above-excerpted paragraph shows, Frankfurt’s conception
of rationality is not so straightforward. The interesting twist comes here: “I am far from
suggesting that a creature without reason may be a person. For it is only in virtue of his
rational capacities that a person is capable of becoming critically aware of his own will
and of forming volitions of the second order. The structure of a person’s will
presupposes, accordingly, that he is a rational person.”30 Thus, we see that Frankfurt’s
conception of an individual’s “rational capacities” covers not just the instrumental
executive ability which gets the heroin addict to his next fix by the most efficient possible
route, but includes also (what one might call) a ‘recognitional capacity’, whereby one can
become “critically aware” of one’s own will. This aspect of Frankfurt’s view may seem
surprising, insofar as it suggests that he is operating with a much less instrumental
conception of rationality than one might have initially supposed him to have.
Nevertheless, if we do not interpret Frankfurt as allowing that an individual’s rational
capacities include some kind of ‘recognitional capacity’ of this kind, then it is difficult to
see how one could make sense of Frankfurt’s talk of the role of rational capacities in
30 FW&CP, p. 17. My italics.
27
creating critical awareness of an agent’s own first-order desires, where that critical
awareness presumably includes some insight into what is and what is not reason-giving
within the content of those first-order desires.
This claim that rationality involves certain ‘recognitional capacities’ may seem strange,
on the face of it. For we might think that the ability to recognize that one has first-order
desires is a function of one’s capacity for self-knowledge rather than requiring any kind
of rational faculty. And merely to see that one has such first-order states is, of course, not
yet to see them as invested with any kind of normative significance. To make sense of
Frankfurt’s claim, we have to understand his conception of our ‘recognitional’ rational
capacities as involving an ability to see certain states of affairs as reason-giving, or as
normatively significant. The Frankfurtian rational ‘person’ turns this capacity towards his
first-order desires, and comes thereby to assess their relative claims on him. He takes
some of them to present him with genuine reasons for action, and others to fail to do so;
he judges their relative merits, and assesses which of them might be worthy of
endorsement. In short, the Frankfurtian person applies to his desires a capacity for
recognizing, perceiving, judging and assessing the claims of reasons.
However, it is precisely this kind of rational ‘recognitional’ capacity which, as we saw
earlier, Rebecca brings to bear directly on the world itself. So, one wonders, if having this
kind of rational ‘recognitional’ capacity is, for Frankfurt, constitutive of genuine
rationality, then why does the rational agent not bring this capacity to bear directly upon
the world, as Rebecca does? Why does he turn this recognitional capacity inward on
28
himself, in assessing the claims of his own desires, but never outwards, in assessing the
normative claims of external considerations? There seems to be no good reason as to why
an agent who genuinely possessed this sort of rational recognitional capacity need
restrain his or her use of it in this way. If one has such a capacity, then one could use it
reflectively, as well as reflexively. Given this, there seems to be a strange instability in
Frankfurt’s position. His ‘persons’ are possessed of exactly the capacity for the direct
recognition of reason-giving force which they might use to consider directly which
courses of action might be rationally compelling, and which might not. But, for some
strange reason, Frankfurt’s ‘persons’ have not yet learned to turn that capacity outwards
onto the world, or to use it directly in deciding how to act.
The explanation I would want to hazard as to why Frankfurt’s agents are in this way
constrained involves Frankfurt’s acceptance of a certain kind of explanatory picture. The
picture which Frankfurt accepts relates to the role of desire in the generation of action.
Frankfurt’s picture, it would seem, involves at least the following elements:
(1) Every action is motivated by some desire.
(2) Desire is basic in the explanation of action.
(3) Desires have two salient features: an object and a strength.
(4) Desires are ‘ours’ only in the “gross literal sense”.
(5) We are passive with respect to our desires.
(6) Action involves choosing between different candidate desires.
29
This picture of agency and desire is captured precisely by the ‘Cartesian’ model of desire
that was seen to be demonstrated by (and, in the example, also accepted by) Arbitrary
Alex. It is not the purpose of this paper to argue against this picture of the role and nature
of desire, but I will just say a few things with respect to it. Firstly, the elaboration of the
nature of Rebecca’s agency and deliberation, given above, should – I hope – count in
favour of rejecting this picture. If Rebecca-type (i.e. ‘non-reflexive’) agency is a
possibility, then this sort of picture of desire is misleading. This is because Rebecca does
not act by virtue of choosing between different candidate desires, but simply by
deliberating about what she has most reason to do. For agency of the type demonstrated
by Rebecca, it would be misleading to say that desire, as such, has any explanatory or
causal primacy. Secondly, though, it is important to note that one need not accept all that
much of an alternative account of desire (such as Scanlon’s account of desire as a
judgement-sensitive attitude whereby we take some consideration as providing a reason
for action) in order to think that Frankfurt has fallen into error here. One need only think
that a decision to act could have its genesis in one’s application of one’s ‘recognitional’
rational capacities directly upon the world, unmediated by any motivationally basic
straightforward desire to act in such a way.
If Frankfurt’s ‘Cartesian’ view of desire and agency still seems to be plausible, then the
spell of this picture can probably be broken by making use of Thomas Nagel’s distinction
between ‘motivated’ and ‘unmotivated’ desires. This tremendously helpful distinction
can be found early on in Nagel’s The Possibility of Altruism:
30
The assumption that a motivating desire underlies every intentional act depends, I believe, on a confusion between two sorts of desires, motivated and unmotivated. It has been pointed out before31 that many desires, like many beliefs, are arrived at by decision and after deliberation. They need not simply assail us,32 though there are certain desires that do, like the appetites and in certain cases the emotions. […] The desires which simply come to us are unmotivated, though they can be explained. Hunger is produced by lack of food, but is not motivated thereby. A desire to shop for groceries, after discovering nothing appetizing in the refrigerator, is on the other hand motivated by hunger. Rational or motivational explanation is just as much in order for that desire as for the action itself. The claim that a desire underlies every act is true only if desires are taken to include motivated as well as unmotivated desires, and it is true only in the sense that whatever may be the motivation for someone’s intentional pursuit of a goal, it becomes in virtue of his pursuit ipso facto appropriate to ascribe to him a desire for that goal. But if the desire is a motivated one, the explanation of it will be the same as the explanation of his pursuit, and it is by no means obvious that a desire must enter into this further explanation. … If the likelihood that an act will promote my future happiness motivates me to perform it now, then it is appropriate to ascribe to me a desire for my own future happiness. But nothing follows about the role of the desire as a condition of contributing to the motivational efficacy of these considerations. It is a necessary condition of their efficacy, to be sure, but only a logically necessary condition. It is not necessary either as a contributing influence, or as a causal condition.33
Nagel here gets things right, and shows how we might come to appreciate the alternatives
to the misleading picture presented above. What the ‘Cartesian picture’ (embraced by and
evinced by Arbitrary Alex, and seemingly standing behind Frankfurt’s view) holds is that
all first-order desires are unmotivated. But as Nagel and Scanlon have argued, and as the
example of Rebecca shows, this is implausible. Not all instances of ‘desiring’, when
desire is understood in a more expansive way, need involve the sort of process which
seems to be at work in Descartes’s characterization of ‘Le Desir’ in The Passions of the
31 “For example, by Aristotle, Nicomachean Ethics, Book III, Chapter 3.” (Nagel’s footnote.)
32 Note here the suggestion, carried by the notion of being “assailed”, that one need be passive in the face of desires.
33 Thomas Nagel, (1970), The Possibility of Altruism, (Princeton, NJ: Princeton University Press), pp. 29-30.
31
Soul (see Figure 1 above). What the ‘Cartesian picture’ also supposes is that every action
can be explained in terms of the prior existence of an unmotivated desire. This
supposition is also at work in Frankfurt’s talk of the will being identical with “an
effective desire – one that moves (or would move) a person all the way to action.”34 For
Frankfurt, there can be no willing without there also being an underlying (unmotivated)
first-order desire. Again, once we bear in mind the possibility of Rebecca-type agency,
this seems to be an important mistake.
The characterization of the nature of desire is, of course, a difficult and knotty
philosophical problem, and it is beyond the scope of this discussion to engage in
sustained argument on the side of the conception favoured by Nagel and Scanlon, as
against that held by Descartes or Frankfurt. The important point for our present purposes,
rather, is this more limited one: Frankfurt’s particular conception of the ‘recognitional’
rational capacities – that is, of our ability to see considerations as generating reasons for
action – is plausible only if one accepts Frankfurt’s picture of the nature of desire, rather
than that of Nagel and Scanlon. For, otherwise, it is difficult to see why agents should
have the capacity to subject their first order desires to rational scrutiny, without also
being able to turn that capacity for critical scrutiny outwards to the world. If the
Cartesian-Frankfurtian conception of desire is mistaken, and all actions need not issue
from some antecedent unmotivated desire, then there is no reason to think that the
rational recognitional capacities which Frankfurt allows with respect to the assessment of
34 FW&CP, p. 14
32
desires, might not also be used to guide action through their application directly to the
world.
Only if Frankfurt is right about the nature of desire could he also be right about the
impossibility of substantively rational agency of the sort which, I suggest, Rebecca
demonstrates. For, it is only if action is impossible without motivationally primary
‘unmotivated’ desires that it would be impossible for an agent to act rationally on the
basis of purely first-order judgments. Given that, it seems to me, Nagel and Scanlon
present a more plausible account of the nature of desire, we should therefore reject
Frankfurt’s strangely unstable conception of practical rationality, along with his
essentially Cartesian conception of the nature of desire.
5. Personhood, Agency and Freedom of the Will:
As the title of Frankfurt’s famous paper – ‘Freedom of the Will and the Concept of a
Person’ –suggests, his goal is to present an account of the conditions for freedom of the
will, as well as an account of the concept of a person. As we might expect, his account of
freedom of the will is closely analogous to his account of personhood, and suffers from
parallel problems. As Frankfurt puts it, “…there is a very close relationship between the
capacity for forming second-order volitions and another capacity that is essential to
persons – one that has often been considered a distinguishing mark of the human
33
condition. It is only because a person has volitions of the second order that he is capable
both of enjoying and of lacking freedom of the will.”35
Frankfurt’s idea here is roughly this: freedom of the will is connected to the relationship
between a person’s second-order volitions and the content of his will (that is, whichever
particular first-order desire successfully brings him to act). ‘Freedom of the will’, on this
account, is to be distinguished from mere ‘freedom of action’, which simply characterizes
being able to do as one pleases, or of “doing what one wants to do”.36 Whilst an animal
might enjoy freedom of action, insofar as it “may be free to run in whatever direction it
wants”, this is not enough for freedom of the will.37 Similarly, for a wanton, for whom no
such relation between first-order desires and second-order volitions exists (simply
because he has no second-order volitions), there is no problem of ‘freedom of the will’. It
can be true that a wanton might find himself enjoying “freedom of action” – that is, the
capacity to do as he wishes – but the matter of “freedom of the will” simply doesn’t arise
for him. One cannot have freedom of the will unless one has the requisite second-order
volitional states. Moreover, one could have freedom of the will without freedom of
action, as in cases where one tries to act in accordance with one’s second-order volitions,
but is prevented from doing so by some external obstacle (one is, say, chained up or
imprisoned or restrained in some way). Thus, on Frankfurt’s view, freedom of the will
35 FW&CP, p. 19.
36 FW&CP, p. 19.
37 FW&CP, p. 20.
34
and freedom of action are doubly dissociable: the first is neither a necessary nor sufficient
condition for the latter.
Now, clearly, Frankfurt makes a mistake in calling the freedom to do as one wants, which
may be enjoyed by animals as well as by wantons, “freedom of action”. For surely, on his
view, neither animals nor wantons can genuinely be said to act – they are, after all,
passive in the face of their first-order desires. Nevertheless, Frankfurt’s view is clear
enough if one replaces the idea of “freedom of action” with the more cumbersome, but
more felicitous “freedom to do what one wants to do”, or, as Frankfurt later puts it,
“freedom to do as one pleases.”
“Freedom of the will”, though, is a different matter. Frankfurt tells us that “… the
statement that a person enjoys freedom of the will means (also roughly) that he is free to
want what he wants to want. More precisely, it means that he is free to will what he wants
to will, or to have the will he wants. […] It is in securing the conformity of his will to his
second-order volitions, then, that a person exercises freedom of the will.”38 Thus, an
initial formulation of Frankfurt’s view on freedom of the will would have this form:
F1: X enjoys freedom of the will iff X’s will conforms to X’s second-order volitions.
Frankfurt would allow that X “acts freely” when X’s will conforms to his second-order
volitions. But there is more to freedom of the will, on Frankfurt’s view, than just acting
38 FW&CP, p. 20
35
freely. Thus, F1 is not the whole truth about Frankfurt’s conception of free will, for he
allows that there might be a case where X’s will conformed to X’s second order volitions,
but where this conformity was a matter of sheer luck. As he says, “it is in the discrepancy
between his will and the second-order volitions, or in his awareness that their
coincidence is not his own doing but only a happy chance, that a person who does not
have this freedom feels its lack.”39 This leads us to a revised Frankfurtian criterion for
freedom of the will:
F2: X enjoys freedom of the will iff [X’s will conforms to X’s second-order volitions
and the conformity of X’s will and his second-order volitions is not a mere
coincidence, but is instead X’s “own doing”].
F1 has the advantage over F2 of providing a clear criterion for determination of the
freedom of an individual’s will. The problem with F2 comes with its further condition
that the conformity of X’s will to his second-order volitions is itself of X’s “own doing”.
It is very difficult to know how to make sense of this. After all, we are talking about a
process that goes on within the confines of X’s mental life, and so the reference to a
relation being forged between two of X’s mental items “through X’s own doing” is very
difficult to understand.
In a slightly different formulation of this same thought, Frankfurt says that it is “in
securing the conformity of his will to his second-order volitions, then, that a person
39 FW&CP, pp. 20-21.
36
exercises freedom of the will”40 But this talk of securing makes matters no more clear. If
X secures a certain outcome then we might see that this is an outcome that he actively
brings about. But the question of whether it is appropriate to speak of X as being active,
or being a genuine agent at all, is precisely what is at stake when we address these
questions about freedom of the will. Thus, it seems obscure as to how we might make
sense of talk of either X “securing” an outcome, or of an outcome being of X’s “own
doing”, until we have a clear answer to the question of when it is that X acts freely. But
this is precisely the question for which we are trying to find an answer, and hence a rather
dangerous circularity threatens.
Matters are complicated further when we bear in mind that Frankfurt’s account of
“freedom of action”, or of the “freedom of doing as one pleases”, makes it clear that he
regards actions which spring solely from the first-order desires of an animal or wanton as
nevertheless being that individual’s “own doing”. Given this, we would have to conclude
that, even if the conformity between an agent’s will and his second-order volitions was
secured as a result of that particular first-order desire which constituted his will itself
causing the formation of a corresponding second-order volition, this would nevertheless
still be a conformity which was the agent’s “own doing” (it certainly wouldn’t be anyone
else’s doing!). This would lead us to the deeply counter-intuitive view that X enjoyed
freedom of the will in the case where his will conformed to his second-order volitions,
even if his second-order volitions were directly caused by his operative first-order desires.
If second-order volitions are not themselves causally independent of first-order desires,
40 FW&CP, p. 20
37
then any claims for their rational authority would appear to be deeply eccentric. They
would become no more than a sort of epiphenomenal rubber stamp for first-order desires.
Other readings of “by X’s own doing” seem to fare no better, though. Thus, we are left in
a dilemma. We can either employ a reading that is itself dependent on a prior standard of
free agency, in which case we are left with a deadening circularity. Or else we can use an
overly-permissive reading, which fails to distinguish genuine freedom of the will from
situations in which a mere simulacrum of freedom of the will is generated by the causal
effects of one’s first-order desires upon one’s higher-order volitions. Clearly, an
alternative approach is needed in order to make the best possible sense of Frankfurt’s
view.
Let us assume that we need to rule out the case where X’s first-order desires themselves
cause the formation of corresponding second-order volitions. One way to do this might be
to specify that the creation of the conformity between X’s will and X’s second-order
volitions need not be just “by X’s own doing”, but “by X’s doing in accordance with X’s
free will”. But this would be a false start, as we are once more threatened by a vicious
circularity. After all, it is the characterization of X’s freedom of the will that is itself up
for analysis. So, the best way forward in creating a sympathetic finessing of Frankfurt’s
F1 is surely therefore simply to do away with talk of X’s “own doing” in this sort of
(‘interior’, mental) context.
38
One sympathetic way of reconstructing F2, although it does not have a direct basis in
Frankfurt’s text, could be this:
F3: X enjoys freedom of the will iff [X’s will conforms to X’s second-order volitions
and the conformity of X’s will to his second-order volitions is not a mere
coincidence, but is instead causally dependent on X’s forming that particular
second-order volition].
This might seem to get nearer to the intuitive idea at which Frankfurt would appear to be
aiming. Furthermore, it might do so in a way which banishes the deadening obscurity,
within this ‘interior’ domain, of reference to the agent’s “own doing”. Nevertheless, for
reasons structurally identical to those given for the rejection of P2 as a criterion of
personhood, F3 seems to fail as a criterion for determining whether some agent enjoys
freedom of the will.
Reasonable Rebecca, lacking second-order volitions, is not even a candidate for the
enjoyment of “freedom of the will”. As Frankfurt tells us, “the wanton … neither has the
will he wants nor has a will that differs from the will he wants. Since he has no volitions
of the second order, the freedom of his will cannot be a problem for him. He lacks it, so
to speak, by default.”41 This seems wrongheaded, given that Rebecca’s robust form of
agency, her ability to govern her own conduct, and the sensitivity of her attitudes and
intentions to her judgments about the nature of the world and the reasons which it
41 FW&CP, p. 21.
39
presents to her, all mark out Rebecca as a paradigmatically free agent. Although Rebecca
fails the test set by F3, it seems strange to think that she does not have the qualities in
which we might be interested when we think of what it might be to have freedom of the
will, and when we consider why having this kind of freedom might be valuable. Rebecca,
in short, appears to be an free and autonomous agent, a proper candidate for moral
assessment, and someone with whom one could engage in the process of giving and
taking reasons for action. Rebecca surely has the kinds of freedom and autonomy worth
wanting. And so, it surely seems implausible to think that F3 could present even a
necessary condition for freedom of the will.
Conversely, in the case of Utterly Arbitrary Alex, we have a strong correlation between
the content of his will and that of his second-order volitions. We can further assume, of
course, that this conformity between first-order and second-order states is not a matter of
contingent chance or any kind of happy accident, but is due to Alex’s exercise of his will.
That is to say, whichever of the random and bizarre first-order desires turns out to be
operative as Alex’s will, it so turns out precisely because that is the desire that has been
endorsed by his (equally chaotic and ill-ordered) second-order volitions. So, Alex passes
the test of F3, whilst lacking valuable qualities of coherence, stability and normative
competence that we might plausibly think should be constitutive of any plausible
conception of freedom of the will worth wanting. This suggests that F3 also fails at
presenting us with a sufficient condition for freedom of the will.
40
Now, Frankfurt could allow that, in many ways, it might be better to be Rebecca than to
be Alex, whilst holding on to his view that it was nevertheless the case that Alex had a
free will, where Rebecca did not. But it seems to me that this remains unfair to the facts.
It is not only through packing too much generalized normative significance into the idea
of ‘freedom of the will’, or through supposing that all good things must go together, that
one would come to the view that Rebecca enjoyed a free will where Alex did not. One
would come to this conclusion, more directly, through thinking about the nature and
significance of this attribution of freedom.
The free agent is self-governing or self-directing in an important way: something that
Rebecca is, but Alex is not. And, whilst we might ultimately think that it was fair enough
to subject both Alex and Rebecca to moral assessment, the practice of moral assessment
seems to make sense with respect to Rebecca in a way that it does not with respect to
Alex. Certainly, one can imagine giving reasons to Rebecca, and accepting them back
from her. With Alex, no kind of moral dialogue seems possible. We would, in all
probability, either just attempt to avoid him, or instead just treat him, somewhat
amusedly, as an entertaining passing spectacle. In either case, we wouldn’t see him as a
candidate for moral co-deliberation. In conclusion, then, insofar as the idea of ‘freedom
of the will’ relates to plausible notions of self-government and reason-responsiveness,
and insofar as it links up with our understanding of the nature and purpose of moral
assessment, Frankfurt’s F3 would seem to fail as a criterion for ascribing freedom of the
will to an individual.
41
Thus, if we draw in the strands of the discussion in this and the foregoing two sections,
we may conclude that Frankfurt’s hierarchical accounts of (i) personhood, and (ii)
freedom of the will, as encapsulated in P2 and F3, are not a success. The possibility of
both Alex and Rebecca tell against them, and their rejection uncovers problems with
Frankfurt’s conception of reason and desire. And, whilst perhaps there is no human being
who reliably and constantly displays the special qualities we have ascribed to Rebecca,
the important point is that Rebecca’s form of rational deliberation is an intelligible and,
indeed, a normatively admirable one. Moreover, we can all manage, more or less and
from time to time, to reason as Rebecca does, giving our attention wholly to the true
objects of our practical deliberation, without the reflexive turn towards consideration of
our own first-order volitional states.
Given this, my hope is that we have rescued the “rational wanton” from the unwarranted
opprobrium in which Frankfurt holds her. I hope, also, that it is now clear that the
difficult truths about the conditions for personhood and for freedom of the will cannot be
found simply by appeal to the existence of a particular structure of volitional hierarchy,
of the kind that Frankfurt offers us. The truth is surely much more complex than that.
6. Decisiveness, Commitment and Higher-Order Desires:
Before moving on to a discussion of Frankfurt’s account of responsibility, it will be
useful to say something about Frankfurt’s treatment of desires and volitions of an order
higher than the second order. Frankfurt admits that we can easily conceive of third- or
42
fourth- (or higher) order desires of the form “I want to want to want that x” or “I want to
want to want to want to x” and so on. As Frankfurt puts it, “there is no theoretical limit to
the length of the series of desires of higher and higher orders; nothing expect common
sense and, perhaps, a saving fatigue prevents an individual from obsessively refusing to
identify himself with any of his desires until he forms a desire of the next higher order.”42
Indeed, it is a curious feature of bringing our second order desires and volitions into focus
that this process is one which naturally can lead us towards forming desires about those
second-order states. We could imagine, for example, a committed alcoholic who, in a
brief moment of clarity, comes to realize that his problem is not so much his strong desire
to drink, but the fact that he is a willing alcoholic, who wants to want to drink. Once this
piece of self-knowledge is in place, he laments that he lacks the right kind of second-
order desire – the kind of desire that would countervail against his first-order addictive
desires. And thus he forms the third-order desire – he desires that he had the desire not to
want to drink. Perhaps that third-order desire then fades, only to be replaced by the
fourth-order desire that it return, and so on. Inertia and indecision, whether of an
unfortunate drunk or of a dithering Hamlet, can often be accompanied by these kinds of
long trains of ever higher-level volitions, each withering on the vine, only to be replaced
by another which takes its predecessor as its object. These volitions are, perhaps, among
the more exotic fauna of our mental lives, but they are readily recognizable states
nonetheless.
42 FW&CP, p. 21.
43
My reason for raising Frankfurt’s treatment of desires and volitions of an order higher
than the second is not to belabour the point well made by Watson and others,43 that
appreciation of the existence of such mental states naturally leads one to come to think of
Frankfurt’s concentration on volitions of the second order (rather than any higher order)
as being under-motivated and arbitrary. Whilst this worry about arbitrariness is a salient
one, what I want instead to concentrate upon is the way in which Frankfurt seeks to
close-off the source of this kind of concern. It is my contention that, if we take
Frankfurt’s response to the challenge of higher-order desires seriously, we can show that
it can actually be used in order to defend the normative significance of Rebecca-style
(‘non-reflexive’) practical deliberation against the charge that it cannot but fail to embody
some lack of freedom or autonomy.
Frankfurt’s response to the problem of higher-order desires is to invoke the idea of a
decisive commitment. Rather than it being arbitrary to stop at the second-order level,
Frankfurt argues, we should notice that “it is possible … to terminate such a series of acts
[of forming ever higher-order desires, conceived as being needed in order to establish
genuine identification with a first-order state] without cutting it off arbitrarily. When a
person identifies himself decisively with one of his first-order desires, this commitment
“resounds” throughout the potentially endless array of higher order desires.”44 Now, one
might here point out that this would solve the problem of the ‘arbitrariness of second-
order volitions’ only if all second-order volitions involved decisive commitments of this
43 See, for example, Gary Watson, “Free Agency,” in Gary Watson, (ed.), (2003), Free Will (second edition), (Oxford: Oxford University Press), pp. 337-351.
44 FW&CP, p. 21.
44
sort, and that Frankfurt nowhere suggests that all second-order volitions have this
property. But our central concern here is not the assessment of this arbitrariness
objection, as such, but rather the implications of Frankfurt’s strategy for foreclosing this
line of objection. Thus, we can set aside the question of whether all, or only some,
second-order volitions really do involve decisive commitment, and concentrate instead on
the implications of Frankfurt’s elaboration of this idea of “decisive commitment” itself.
My contention is that, on examination, it turns out to be a Trojan horse, capable of
decisively destabilizing his view of autonomy and free agency from the inside.
Now, when Frankfurt talks of ‘decisive commitment’, we should bear in mind that the
mechanism for creating ‘decisive commitment’ is, on Frankfurt’s view, simply the
creation of a clear and commanding second-order volition. When we have such a
volition, we are said, therefore, fully to identify with the relevant first-order desire.
Decisive commitment is not generated by some further, as yet unelaborated, feature of
Frankfurt’s theory. Rather, it is something that can be displayed by second-order volitions
of precisely the kind that we have already been discussing. What this means, in effect, is
that Frankfurt allows that second-order volitions can have a special property of being
‘resounding’ or ‘decisive’45 – such that a ‘resounding’ second-order volition closes-off
any space in which it might subsequently be called into question. Unless Frankfurt can
defend the claim that certain second-order desires have this special property of being
resounding, then it looks as if there can be no principled way of granting authority to a
45 The two seem to be used interchangeably by Frankfurt.
45
second-order desire without having to have recourse to an underlying third- or higher-
order desire by virtue of which it is endorsed.
Frankfurt defends the existence of this resoundingness or decisiveness property for
desires by use of a rather straightforward example, in which the phenomenology of
desiring is brought into focus. It will be instructive to reproduce Frankfurt’s presentation
of how this resoundingness property manifests itself in experience:
Consider a person who, without reservation or conflict, wants to be motivated by the desire to concentrate on his work. The fact that his second-order volition to be moved by this desire is a decisive one means that there is no room for questions concerning the pertinence of desires or volitions of higher orders. Suppose the person is asked whether he wants to want to want to concentrate on his work. He can properly insist that the question concerning a third order desire does not arise. It would be a mistake to claim that, because he has not considered the second order volition he has formed, he is indifferent to the question of whether it is with this volition or with some other that he wants his will to accord. The decisiveness of the commitment he has made means that he has decided that no further questions about his second-order volition, at any higher order, remain to be asked. It is relatively unimportant whether we explain this by saying that this commitment implicitly generates an endless series of confirming desires of higher orders46, or by saying that the commitment is tantamount to a dissolution of the pointedness of all questions concerning higher orders of desire.47
So, what do we learn here about the property of resoundingness (or, as seems in
Frankfurt’s view to be the same thing, the property of decisiveness)? A resounding or
decisive volition of the second order (i) forecloses on any question of the pertinence of
46 It is hard to see how this would explain anything, given that an “endless series” of confirming desires would, indeed, be endless, and so would seem to lack the authoritative stamp of final confirmation. Any desire in such a series would still be open to confirmation or disconfirmation at the next level, and would thereby lack the requisite properties of finality or authority. Thus, only the second of Frankfurt’s proposed explanations (“commitment [as] … tantamount to a dissolution…”) could even hope to succeed. Frankfurt is therefore mistaken in being indifferent between these two proposed explanatory strategies, as when he says that “it is relatively unimportant” which explanation we accept.
47 FW&CP, pp. 21-22.
46
higher-level volitions, (ii) carries with it a sense of being decisively endorsed, without
the need for the agent in question to form any higher-order desire with respect to it. These
resounding, decisive second-order volitions are describable as such because (to use the
terms used in §4) they carry with them their own de re endorsement, thereby foreclosing
on the need for any ‘explicit’ de dicto endorsement, as would be achieved by their
approval by a volition of a higher order. Given that these second-order mental states are,
as it were, self-endorsing (in a de re sense) it follows that the existence (or otherwise) of
any higher-order mental state which could be described as involving a de dicto
endorsement of that decisive volition is rendered otiose. In short, resounding or decisive
volitions do not require any higher-level endorsement in order for them to be said to
represent the agent’s considered and autonomous judgment.
Now, if first-order desires could also have this property of resoundingness or
decisiveness, then we would have to say (even within the framework of Frankfurt’s
theory), that ‘non-reflexive’ agency of the sort associated with Rebecca could be fully
free and autonomous, insofar as it involved resounding or decisive first-order volitional
states. This, therefore, raises a deeper question as to what the real distinction is between
first-order and second-order volitional states. In Frankfurt’s presentation, it would seem
that the only difference between first-order desires and second-order desires is contained
in their different objects. The former take as their objects plans of action or states of the
world, whereas the latter take as their objects the first-order volitional states of agents. As
we have seen, Frankfurt’s account of desire is a rather minimalist one, and so it is no
doubt a deliberate feature of his theory that the difference between first-order desires and
47
second-order volitions can be characterized purely with regards to the objects taken by
the two kinds of mental state. Indeed, the power and attractiveness of Frankfurt’s theory
surely consists in the fact that it can treat all desires as being fundamentally of the same
kind, without distinguishing between, say, motivated and unmotivated desires (as Nagel
does), or offering some other kind of distinction between judgmental desires and ‘raw’
desires. The impressive central promise of Frankfurt’s view is that it promises to account
for difficult problems regarding freedom, personhood, autonomy and responsibility
whilst positing only a minimalist ontology of simple desires – desires that differ from one
another only in intensity and in the nature of their objects, and hence in their place in the
volitional hierarchy.
Nevertheless, if the only difference between first-order and second-order desires, as two
general classes of desire, is their place in the volitional hierarchy, rather than any
internally constituted phenomenological features that they may have, then it is unclear as
to why first-order volitional states should not also exhibit the property of resoundingness
or decisiveness, which their second-order relatives are able to display. As has been
mentioned, if they can display this property, then there is no reason for us to think that
‘wantons’ like Rebecca cannot display the full range of rational and autonomous
volitional behaviour. Frankfurt’s view therefore seems to face a dilemma here. On the
one hand, he can insist that the only real difference between first-order and second-order
desires resides in their objects, in which case he seems to have no good reason to resist
the plausible claim that first-order desires can therefore be just as resounding and
decisive as second-order volitions. On the other hand, he could posit a different way in
48
which first-order and second-order desires differ from one another, over and above their
different objects, in which case the distinctive structure of Frankfurt’s view would be
lost.
Let us start with the first kind of response. If first and second order volitional states differ
only in their objects, then there seems to be no reason why we could not re-tell the story
about decisiveness, or resoundingness, with regard to the first order desires of a
thoughtful and reflective agent like Rebecca, in the following way. Our story here utilizes
the structure of Frankfurt’s discussion of the nature and phenomenology of decisive
commitment, but simply moves the action down by one level in the volitional hierarchy –
from the second-order of reflexive desire, to the first-order states that characterize
Rebecca’s process of deliberation:
Consider Rebecca, who, without reservation or conflict, wants to concentrate on her work. The fact that her first-order desire to work is a decisive one means that there is no room for questions concerning the pertinence of desires or volitions of the second-order, or of higher orders. Suppose that Rebecca is asked whether she wants to want to concentrate on her work. She can properly insist that the question concerning a second order desire does not arise. It would be a mistake to claim that, because she has not considered the first order desire she has formed, she is indifferent to the question of whether it is with this desire or with some other that she wants her will to accord.48 The decisiveness of the commitment she has made means that she has decided that no further questions about her first-order desire, at any higher order, remains to be asked. It is relatively unimportant whether we explain this by saying that this commitment implicitly generates an endless series of confirming desires of higher orders, or by saying that the commitment is tantamount to a dissolution of the pointedness of all questions concerning second-order or higher orders of desire.
48 See, for example, the discussion at §4 above, wherein the question of whether Rebecca can be said to care about the content of her will is discussed, and where it is concluded that she displays a level of de re care about the content of her will, which is displayed by general features of her deliberative practice, even if her lack of second-order states means that it would be accurate to say that she lacks a second-order state of de dicto care about her first-order desires.
49
This story seems a plausible one to tell about a careful and decisive agent like Rebecca,
but obviously it is not a story which can be told without heralding the internal collapse of
Frankfurt’s view about agency and desire.
Grasping the opposite horn of the dilemma, Frankfurt could instead insist that there is, in
fact, a deeper distinction between first- and second- order volitional states, for example
that the first are ‘raw’ and are something that the agent undergoes, whereas the second
involve judgement, and are something that the agent actively authors or creates. But, if
Frankfurt chooses this second horn of the dilemma, then the distinctive structure of his
view has been lost, and the distinction between first- and second- order states would now
have to be made with regard to their origin, or more plausibly perhaps, with regard to
their content, rather than simply with regard to their objects. And, if the property that a
desire needs in order to be decisive or resounding is that the agent has come by it
actively, or that it is a desire that is robustly judgement-sensitive, then it is quite clear that
these properties can be enjoyed by desires that have courses of action and states of their
world as their objects (i.e. first-order desires), just as much as by desires that take other
desires as their objects.
Thus, we must conclude that one would be preparing a Procrustean bed if one were to
insist that only second-order volitions can be decisive or resounding, by virtue of
properties that mark out those desires as being distinctively second-order. To return to
our exemplar of good first-order reasoning, we recall that Rebecca’s first-order desires
are robustly judgement-sensitive, and are generated through an active process of
50
engagement with the normatively salient features of the world, and yet – as we have seen
– they stand at the bottom level of the Frankfurtian ‘volitional hierarchy’. Unless we want
to do violence to what it is for a volitional state to be of the first-order, then we must
admit that Rebecca’s complex, judgement-sensitive states fall under this description.
In either case, therefore, I would contend that the only way in which Frankfurt can defend
his solution to the puzzle of how second-order commitments can be authoritative and
non-arbitrary is to admit that the normative authority of agential identification can attach
also to desires of the first-order. But, if this is done, then Rebecca, our Rational Wanton,
is again clearly rendered as an exemplary agent. Whilst being an ill-considered, passive
and indecisive wanton may be a terrible thing, it is nevertheless the case that a thoughtful,
reflective, active and decisive wanton like Rebecca is a fine thing to be. Rebecca, once
again, passes muster as a free, rational and autonomous agent.
7. Hierarchy, Volition and Responsibility
In this section, I want to turn to Frankfurt’s treatment of responsibility in the fourth
section of ‘Freedom of the Will and the Concept of a Person’. It is in this section that the
full normative consequences of Frankfurt’s view really come to the fore. As we shall see,
Frankfurt’s account of responsibility is more permissive than his account of freedom of
the will, in that being a legitimate subject of moral responsibility is a necessary but not a
sufficient condition for exhibiting personhood and freedom of the will. Frankfurt,
51
therefore, defends the rather unusual claim that freedom of the will is not a necessary
condition for moral responsibility.
Nevertheless, although his account of responsibility comes apart from his account of
freedom of the will in certain respects, Frankfurt’s account of moral responsibility – and,
hence, of the conditions under which blame and punishment are rendered appropriate –
has its basis in his hierarchical account of agency. The conditions under which an
individual agent can legitimately be held responsible are such that, for Frankfurt, there
can be no question of the ‘wanton’ agent (such as Rebecca) being responsible; in
contrast, an agent like Alex, despite his pathologies of reasoning and his lack of agential
cohesion, can nevertheless legitimately be held responsible. It is the aim of this part of
my discussion to show the implausibility of Frankfurt’s account of responsibility, not
least because it can be shown to be much too ready to hold individuals responsible for
their conduct.
Now, as has been mentioned, Frankfurt believes that one may be morally responsible
even when one’s will is not free. Frankfurt thinks that we can be held responsible for
what we do when we act in accord with our second-order volitions, even if we would
have still acted in that way even if our second-order volitions had been different. In other
words, Frankfurt thinks that we are responsible for our second-order volitions, whether or
not we have freedom of the will (i.e. whether or not we are reliably able to realize those
second-order volitions through the determination of our (first-order) will). The structure
of Frankfurt’s view about responsibility will be put under careful scrutiny later in this
52
section, but, for now, the salient point to make is simply this Frankfurt thinks that
freedom of the will and moral responsibility can come apart from one another. Thus,
whereas most theorists of free will and moral responsibility accept what one might call
the ‘Equivalence Claim’ – that is, the claim that agents are properly held morally
responsible for their actions when and only when they act of their own free will –
Frankfurt is noteworthy in his rejection of that claim. As Frankfurt puts it, “The most
common recent approach to the problem of understanding the freedom of the will has
been, indeed, to inquire what is entailed by the assumption that someone is morally
responsible for what he has done. In my view, however, the relation between moral
responsibility and the freedom of the will has been very widely misunderstood. It is not
true that a person is morally responsible for what he has done only if his will was free
when he did it. He may be morally responsible for having done it even though his will
was not free after all.”49 As we see, Frankfurt undercuts the Equivalence Claim in only
one direction here. Although he asserts that we can be morally responsible even when our
will is not free, he does not make the suggestion that one can enjoy freedom of the will in
situations in which one is not morally responsible.50
Now, before addressing Frankfurt’s account of moral responsibility as such, we first need
a brief detour to examine the conception of freedom of the will at work in §4 of ‘Freedom
of the Will and the Concept of a Person’. This brief detour will examine the seemingly
49 FW&CP, pp. 23-24.
50 In this respect, Frankfurt’s view points forward to that of J. M. Fischer, and his defence of ‘semi-compatibilism’. See J. M. Fischer, (1995), The Metaphysics of Responsibility: An Essay on Control, (Oxford: Blackwell); and J. M. Fischer and Mark Ravizza, (1998), Responsibility and Control: A Theory of Moral Responsibility, (Cambridge: Cambridge University Press); and J. M. Fischer, (2007), My Way: Essays on Moral Responsibility, (Oxford: Oxford University Press).
53
inconsistent formulations of the conditions for freedom of the will that occur in
Frankfurt’s arguments, before returning to the central question regarding Frankfurt’s
account of the structure of responsibility. Recall Frankfurt’s earlier claim regarding
freedom of the will, as discussed above in Section 5 of our discussion:
F1: X enjoys freedom of the will iff X’s will conforms to X’s second-order volitions.
As we saw, F1 was subject to significant objections, not least with regard to situations
where the conformity of an individual’s will to his second-order volitions was a matter
not of his will being causally dependent on the operation of his second-order volitions but
where, for example, the causal dependence ran in the opposite direction (or, for example,
when this conformity was due to a common external cause). Our discussion in Section 5
then suggested that the most sympathetic reconstruction of Frankfurt’s view leads us to
the following more finessed statement of his position regarding freedom of the will:
F3: X enjoys freedom of the will iff [X’s will conforms to X’s second-order volitions
and the conformity of X’s will and his second-order volitions is not a mere
coincidence, but is instead causally dependent on X’s forming that particular
second-order volition].
That F3, or something very close to it, is the best interpretation of Frankfurt’s account of
free will seems well-confirmed when Frankfurt initially characterizes the nature of
freedom of the will (in contradistinction to moral responsibility). He tells us: “A person’s
54
will is free only if he is free to have the will he wants. This means that, with regard to any
of his first-order desires, he is free either to make that desire his will or to make some
other first-order desire his will instead.”51 So far so good, one might plausibly think, with
regard to Frankfurt’s view being captured by F3. However, the plausibility of F3 as the
most felicitous reconstruction of Frankfurt’s account of freedom of the will comes under
threat when we consider the way in which his view here seems to slide into new territory,
insofar as he is concerned to distinguish between the conditions needed to be satisfied in
order to give freedom of the will, as against those needed for moral responsibility.
Frankfurt continues:
Whatever his will, then, the will of the person whose will is free could have been otherwise; he could have done otherwise than to constitute his will as he did. It is a vexed question just how ‘could have done otherwise’ is to be understood in contexts such as this one. But although this question is important to the theory of freedom, it has no bearing on the theory of moral responsibility. For the assumption that a person is morally responsible for what he has done does not entail that the person was in a position to have whatever will he wanted.52
Here we see that a ‘could have done otherwise’ condition (or, as Frankfurt calls this
elsewhere, with regard to moral responsibility rather than freedom of the will, a
‘Principle of Alternate Possibilities’ or PAP53) has emerged in Frankfurt’s account, and
become attached to the specification of the circumstances where an agent is free. But, as
can be seen from an re-examination of F3, the earlier formulation of the Frankfurtian
conception of freedom of the will54 does not involve a ‘could have done otherwise’ or
51 FW&CP, p. 24.
52 FW&CP, p. 24. (First two sets of italics mine; final italics Frankfurt’s)
53 See Frankfurt, “Alternative Possibilities and Moral Responsibility,” in Frankfurt (1988), p. 1.
54 That is, the conception that emerges in §3 of FW&CP (i.e. pp. 19-22).
55
‘alternate possibilities’ condition at all. One could exhibit freedom of the will, by F3, so
long as one’s will conformed to, and was causally dependent on, one’s second-order
volitions. But this could be quite consistent with one’s second-order volitions themselves
being causally determined. According to F3, one could exhibit freedom of the will even if
there was only one possible course of action that one’s second-order volitions would
cause one to want to take, and even if one’s set of second-order volitions were themselves
determined, fixed and ‘outside of one’s control’. (For, we should remember, for our
second-order volitions themselves to be ‘under our control’ would surely be to call for
some ever-higher order of volitional identification.) Thus, although Frankfurt’s criterion
of freedom of the will requires an additional layer of freedom, as it were, over and above
that required for mere freedom of action, given that it requires the determination of one’s
first-order desires by whatever second-order desires one has), this earlier account is
nevertheless still a “hypothetical” account of freedom, that falls short of requiring
alternative possibilities.55
Now, as we’ve seen above, in §5 of our discussion, there are problems with F3 as a
principle for determining when an agent has freedom of the will, not least because of the
seemingly faulty judgements it renders in the cases of Alex and Rebecca. But, it must be
admitted, F3 nevertheless manages to introduce an initially plausible and forceful
compatibilist conception of freedom of the will, whereby there need not be any deep
metaphysical opposition between an individual’s being determined and that individual’s
having a free will. This advantage is lost in the move made above by Frankfurt, wherein a
55 Thanks to T. M. Scanlon for suggesting this way of formulating this point.
56
metaphysically problematic ‘could have done otherwise’ or ‘alternate possibilities’
condition finds its way into the specification of the conditions under which an individual
can exhibit freedom of the will. As Frankfurt himself admits, interpretation of such
counterfactual conditions is terribly difficult in this domain – as he puts it, “it is a vexed
question just how ‘could have done otherwise’ is to be understood.”56
Given the problems with moving his view in this direction, this modulation of Frankfurt’s
view seems to be unmotivated and puzzling. Furthermore, it is wholly corrosive of the
straightforwardly compatibilist credentials of F3. One should conclude, therefore, that
this modulation of Frankfurt’s view is at best a mere slip. The most sympathetic account
of his considered position on freedom of the will should remain the straightforwardly
compatibilist and metaphysically unpuzzling F3.
Why it is that F3 should be left in place should be especially clear when we consider
Frankfurt’s later discussion of what he calls volitional necessities. A volitional necessity
is, in Frankfurt’s terms, the kind of unavoidable second-order volitional commitment
displayed by Martin Luther when he (freely, and with freedom of the will) nails his theses
to the door, pronouncing that “Hier steh ich, ich kann nicht anders.”57 Volitional
necessities, Frankfurt argues in later papers, have the feature that they can be experienced
56 FW&CP, p. 24.
57 “Here I stand: I can do no other”, as Luther is supposed to have claimed in 1546 when he nailed his 95 Theses to the door of the Castle Church in Wittenberg. See Michael A. Mullett, (2004), Martin Luther, (New York: Routledge). For an amusingly deflationary account of these kinds of alleged “volitional necessities”, with reference to Luther, see Rogers Albritton, (1985), “Freedom of the Will and Freedom of Action,” reprinted in Watson, ed., (2003), ibid., pp. 408-23, at p. 419.
57
as wholly constraining, yet without being external or alien.58 They are constitutive of an
individual’s moral personality, and as such they can be integrated into the life of an
autonomous agent, rather than leading to a condition that must exclude an individual
from the possibility of autonomy. Martin Luther, though acting from necessity, acts of a
volitional necessity that constitutes his identity, and hence acts freely and with freedom
of the will.
By F3, insofar as Luther’s will – the content of which is that he wishes to go to
Wittenberg and nail his theses to the church door – is caused by his underlying second-
order volitional necessities, Luther has freedom of the will. But Luther could not have
done otherwise – not, of course, because he was a compulsive or an addict – but because
his second-order volitional state could not have been otherwise, given the volitional
necessities to which he was subject. Or, at least, his volitional state could not have been
otherwise without cost to his identity as a distinctive autonomous agent. By the view that
Frankfurt elaborates in §4, then, Martin Luther’s will was not free, precisely because it
could not have been otherwise. It could not have been otherwise because, ex hypothesi,
the content of Luther’s will was causally dependent on the content of his second-order
volitions, and those second-order volitions themselves could not have been otherwise. On
the view of §3 (and of F3), Luther’s will was free, because it was causally dependent on
his second-order volitions, even though those second-order volitions could not have been
otherwise. It is puzzling that Frankfurt’s paper seems to contain defences of these two
58 See his “Necessity and Desire,” (1984), Philosophy and Phenomenological Research, XLV, I, pp. 1-14, reprinted in Frankfurt, (1988), ibid.; and also his “Autonomy, Necessity and Love,” in Frankfurt, (1999), ibid. See also Harry G. Frankfurt, (2004), The Reasons of Love, (Princeton, NJ: Princeton University Press).
58
inconsistent views but, at any rate, this latter view, as captured by F3, is much to be
preferred.
Let us return from our excursion into the precise delineation of Frankfurt’s conception of
freedom of the will. The question that I now want to return to addressing regards the
plausibility of Frankfurt’s conception of moral responsibility, and the interrelation of
freedom of the will with moral responsibility on Frankfurt’s view.59 As has been
mentioned, the two come apart for Frankfurt. Indeed, one might even conjecture that the
reason that Frankfurt comes to elaborate this unfortunately modulated version of his
account of freedom of the will is connected with his concern to distinguish freedom of
the will from moral responsibility. Frankfurt is concerned to demonstrate that moral
responsibility does not require any reliance on a ‘could have done otherwise’ or ‘alternate
possibilities’ condition, which perhaps leads him too quickly to think that freedom of the
will, as such, should rely on such a condition. But, as has been suggested above, freedom
of the will in the sense that interests Frankfurt (i.e. freedom of the will according to F3)
need not rely on a condition of this kind.
59 It is a feature of Frankfurt’s discussion of moral responsibility that he never really explains with what kind of sense of responsibility he is operating. (Different senses of responsibility will be dealt with in much more detail in Part Three of this dissertation – see especially section 3 of Part Three.) In a very revealing footnote, in “Alternate Possibilities and Moral Responsibility,” Frankfurt admits that “The two main concepts employed in the principle of alternate possibilities are “morally responsible” and “could have done otherwise”. To discuss the principle without analyzing either of these concepts may well seem like an attempt at piracy. The reader should take notice that my Jolly Roger is now unfurled.” (Frankfurt, (1988), ibid., p. 6). Unfurled it very certainly is. But, in any event, piracy or not, one can presumably safely assume that Frankfurt’s conception of moral responsibility is either of an idea of justifiable liability to some form of moral blame, or of justifiable liability to the imposition of some kind of cost or burden.
59
So, how should we make sense of Frankfurt’s rejection of the Equivalence Claim, and his
separation of freedom of the will from moral responsibility? This is best understood if we
consider cases where the two are taken by Frankfurt to come apart. One famous case is
Frankfurt’s ‘Willing Addict’, who lacks freedom of the will, but who is nevertheless
morally responsible for his drug taking.60 The Willing Addict is an addict in the sense
that his desire to take some drug will prove effective no matter what he might want at the
second-order level; he is willing insofar as, as a matter of fact, he nevertheless has a
second-order volition to take that drug. As Frankfurt puts it, “… his addiction has the
same physiological basis and the same irresistible thrust as the addictions of the unwilling
and wanton addicts, but … he is altogether delighted with his condition. He is a willing
addict, who would not have things any other way.”61 He lacks freedom of the will
because “his desire to take the drug will be effective [i.e. will lead him to action] whether
or not he wants this desire to constitute his will”.62
Given that this necessitation of his will works independently of any causal mechanism
that flows through his second-order volitions, it is clear that the Willing Addict is unfree
both on the (questionable) ‘could have done otherwise’ conception of freedom of the will
of Frankfurt’s §4, as well as on the more plausible principle of F3 (given that the content
of his will is not, ex hypothesi, causally dependent on the content of his second-order
volitions, as it has an independent causal source (i.e. his addiction)). But, although the
60 FW&CP, pp. 24-5.
61 FW&CP, p. 24.
62 FW&CP, pp. 24-5.
60
addict’s will is not free, Frankfurt nevertheless thinks it possible that he could be morally
responsible for drug-taking behaviour, because “by his second-order desire that his desire
for the drug should be effective, he has made this will his own”.63 The Willing Addict
identifies with his addiction by virtue of his second-order volitions, and hence this
identification renders him responsible for his drug-taking behaviour, notwithstanding the
causal impotence of his second-order volitions with regard to the determination of his
first-order desires (which are, ex hypothesi, constituted determinately by his addiction).
The case of the willing addict allows us to reconstruct the content of Frankfurt’s account
of moral responsibility. (For it is a curious fact that Frankfurt is much more forthcoming
on the question of what moral responsibility does not involve than on the question of
what, exactly, the conditions of its application might be. We are told, for example, that it
does not rely upon the principle of alternate possibilities,64 and we are told that it does
not require freedom of the will.65 But he is not as explicit as he might be in saying what it
does involve.) It seems plausible to conclude that, for Frankfurt, an agent can be held
morally responsible for what he does when the content of his will corresponds to the
content of his second-order volitions. So, for an individual with a second-order volition to
φ, this would mean that, if the individual in question did have freedom of the will, and
therefore had the capacity to determine his will in accordance with his second order
volitions, then the content of his will would be a desire to φ. But, even in the absence of
63 FW&CP, p. 25.
64 Frankfurt, “Alternate Possibilities and Moral Responsibility,” passim.
65 FW&CP, §4.
61
freedom of the will, that individual can legitimately be held morally responsible for φ-
ing, even if he φ-ed because of some addiction, compulsion or other defect of the will, as
long as his second-order volition was itself constituted by a desire that he φ. This, surely,
is the natural reading of the Willing Addict discussion.
We can thus formulate Frankfurt’s account of moral responsibility like so:
R1: An individual, X, is morally responsible for φ-ing iff his will to φ is accompanied
by a second-order volition to φ.
As we may recall, this makes Frankfurt’s account of responsibility surprisingly similar to
his original, tentative claim about freedom of the will, as embodied in F1.
F1: X enjoys freedom of the will iff X’s will conforms to X’s second-order volitions.
One should immediately notice here that what distinguishes the conditions needed for
freedom of the will (on F3, the replacement view for F1) from the conditions for moral
responsibility need not be anything to do with a ‘could have done otherwise’ or ‘alternate
possibilities’ condition. By F3, the Willing Addict does not have freedom of the will,
because, although his will to take the drug is accompanied by a second-order volition to
take the drug, it is not causally dependent on that second-order volition. Hence, we can
say that his will is not free without having to invoke any claims about whether alternate
possibilities were open to him, and without asking whether his second-order volitions
62
might have been other than they were. The Willing Addict, though unfree, is responsible,
on Frankfurt’s view, because he fulfils the condition laid out in R1 – he acts in a way that
is in accord with his second-order volitions.
Now, my central claim in this section is that Frankfurt’s principle of moral responsibility,
R1, is excessively permissive, in the sense of licensing too many ascriptions of moral
responsibility, and it should accordingly be rejected. To see one way in which it is too
quick to condemn, consider these two kinds of drugs.66
1. Our first drug, we can call ‘nicotine’. When one is addicted to ‘nicotine’ one has
an overwhelming first-order desire to take that drug. But the effects of exposure to
nicotine are comparatively moderate, and it does not infect one’s higher-order
desires at all. Addicts of ‘nicotine’ have, let us assume, a compelling first-order
desire to take it, but, because of their knowledge of ‘nicotine’s’ harmful effects,
they generally also have a second-order volition to not want to want ‘nicotine’.
2. Let’s call our second drug ‘heroin’. ‘Heroin’ is much more powerful than
‘nicotine’, and really gets hold of its users in a much more comprehensive way. It
causes not just an overwhelming first-order desire to take the drug; it also, as it
were, causes its users to fall in love with it and its effects. Users of ‘heroin’, let us
66 In what follows, the names ‘nicotine’ and ‘heroin’ are just convenient placeholders. I do not mean to make any empirical claims about the properties of either drug, or about the phenomenology of addiction to either of them. I mean only to suggest that there are alternatives to Frankfurt’s particular ‘simple model’ of addiction. For a perceptive and insightful discussion of the phenomenology of “taking pleasure” in drug-taking, see the discussion in Richard Moran, (2002), ibid., pp. 211-2.
63
claim, develop not only a powerful first order desire to take it, they also develop
an overwhelming second-order volition such that they approve of their wanting it.
To further emphasize the power of this drug, let us assume that ‘heroin’
furthermore creates second-order volitions that are ‘decisive’ or ‘resounding’ in
Frankfurt’s sense. So great is its attraction that it closes out any possible space in
which its users might come to doubt, question or revise their second-order
volition to take it. Once an individual has exposed himself the delights of
‘heroin’, he ineluctably finds himself with a robust and decisive second-order
volition to seek out and take more of it.
Now, by Frankfurt’s R1, the user of ‘nicotine’ insofar as he may be an ‘Unwilling
Addict’ or a ‘Wanton Addict’, whose will to take that drug is unaccompanied by a
corresponding second-order desire to take it, is not morally responsible for seeking out
and taking ‘nicotine’. But the users of ‘heroin’ find that their will to take the drug is
always and in every case accompanied by a second-order volition to take that drug.
Hence, by Frankfurt’s R1, users of ‘heroin’ are always and everywhere responsible for
acting on their addiction. One might, of course, suggest that this result is of little
consequence, because addiction is, in real life, more like the ‘nicotine’ case, where an
agent’s ‘deeper’ volitions can remain unengaged by his addiction. But this would be, I
suggest, implausible and unfair to the phenomenological experience of addiction. When
addicts lose themselves in a drug, it is not always only because their higher-order
responses are ‘short-circuited’: with regards to the most terrifying and shocking forms of
addiction and compulsion, the higher-order responses themselves become co-opted
64
towards the addictive or compulsive goal. It seems the height of perversity to say that one
is not responsible for one’s behaviour when one’s addictions and compulsions are low-
level and leave one’s rational faculties in place, whereas one is responsible for one’s
addictive and compulsive behaviour when one’s addictions and compulsions colonize
one’s faculties of judgement and endorsement. Yet this is exactly what R1 would have us
judge in these cases. This is reason enough, I suggest, to reject R1 as a principle of
determining moral responsibility. As a principle it is simply too ‘judgmental’, and too
ready to ascribe full moral responsibility in cases where we should be much slower and
more circumspect about holding individuals responsible for what they have done.
Leaving aside cases of addiction and compulsion of the kind associated with ‘heroin’, it
is worth noting that the problem raised above is really just a special case of a more
general sort of worry about R1. Recall that, in rejecting F1, the most significant cases that
delivered counter-examples were ones where the agent’s will was not causally dependent
upon the agent’s higher-order volitions, but rather where their higher-order volitions were
themselves caused by their lower-order desires. It is a strange consequence of R1 that it
holds agents responsible even when their wills have this kind of internal structure,
whereby their ‘decisive commitments’ are driven by their desires, and not vice versa.
It may seem rather puzzling that Frankfurt should end up defending a position of this
kind, but I want to attempt a diagnosis of how and why he is led into this position. I
would suggest that Frankfurt’s account of the gap between responsibility and freedom of
the will only makes sense on the assumption that second order volitions are beyond the
65
reach of external causal factors. Otherwise, it seems terribly arbitrary to hold people
responsible simply on the basis of the content of their second-order volitional states, no
matter how those states came to arise. In cases where they are caused by the effect of
some drug, or in cases where they are caused by the effects of lower-order desires, it
seems absurd to judge that agents are morally responsible just because, as a matter of
fact, the content of their wills accords with the content of their second-order volitions.
Only by granting a particular kind of special kind of status and authority to second-order
volitions, in all cases, is it plausible to think that individuals can always legitimately be
held responsible for those second-order mental states. One might say that Frankfurt’s
claims about moral responsibility gain a certain plausibility by virtue of their smuggling
in certain assumptions about the free and authoritative status of second-order volitions.67
But second-order volitions can come about through ‘improper’ methods (as we have
seen), and are neither always free nor always authoritative. And so we must conclude that
R1 is much too permissive in holding people to standards of responsibility.
At this point, lest he be thought forgotten, I’d like to bring Arbitrary Alex back in. Alex
isn’t an addict or a compulsive, and let us assume that his second-order volitions are not
generated by the effect of any drug, or by the effects of his first-order desires.
Nevertheless, as we have seen, Alex is a chaotic mess, subject to massive lurches of
interest, with attenuated and faulty faculties of normative judgement. He is diachronically
disintegrated, random and more than a little crazy. But his second-order volitions always
give rise to his first-order desire and so, a fortiori, they certainly always correspond to
67 The “smuggling” metaphor was suggested to me by T. M. Scanlon.
66
his first-order desires. By R1, then, Alex is always morally responsible for what he does.
Yet this, surely, seems an eccentric conclusion. It would be an impermissibly harsh, and
unfair, principle of moral responsibility that always held Alex to account, and which
likewise held the ‘heroin’ addict accountable. Yet this is what R1 does. Our conclusion,
then, surely must be to reject R1, and with it Frankfurt’s account of responsibility.68
Drawing together the discussion of this and the foregoing sections, I conclude that we can
securely reject the accounts of personhood, autonomy, freedom and responsibility offered
by the ‘volitional hierarchy’ account of Frankfurt. Even on its own terms, with its
excessively simple moral psychology, this approach fails. Moreover, on closer inspection,
the conception of agency and desire that is embedded in this view has been shown to be
flawed. To understand autonomy, freedom and responsibility, we must look elsewhere.
Volitional hierarchy – an enticing and prima facie powerful approach – cannot make
good on its lustre of promise.
8. Coda: Towards a Pluralist Account of Autonomy
I want to end with a brief coda, pointing towards the right direction with regards to how
best to understand agency, freedom and autonomy. Frankfurt, we recall, suggests in the
fourth section of ‘Freedom of the Will and the Concept of a Person’ that “A person who
is free to do what he wants” [i.e. who has freedom of action, on Frankfurt’s view] and
“who has the will that he wants” [i.e. who has freedom of the will, on Frankfurt’s view]
68 What would count as a plausible account of moral responsibility will be examined in greater detail in the following Parts Two and Three.
67
“has, in that case, all the freedom it is possible to desire or to conceive.” He goes on to
suggest that, whilst “… there are other good things in life, and he may not possess some
of them” it is nevertheless the case that “there is nothing in the way of freedom that he
lacks.”69 Such an individual is, on Frankfurt’s view, responsible, free and autonomous.
As has been seen, Alex has (Frankfurtian) freedom of action and (Frankfurtian) freedom
of the will, and is therefore (on Frankfurt’s account) morally responsible. Yet it is
implausible to think of Alex as free or responsible. He is surely, I suggest, lacking in
autonomy, and therefore an unsuitable individual for us to hold fully accountable for
what he does.70
So, what, exactly does Alex lack? What does Rebecca, our ‘Rational Wanton’, have that
Alex does not? If Rebecca looks like a paradigm of autonomy, and Alex does not, how
are we to explain that difference? Well, to recall our description from Section 3 above,
“Where Rebecca’s first-order desires were well-ordered and cohesive, Alex’s are random
and unpredictable. Thus, their first-order desires differ from one another’s in important
respects aside from their relative calmness or violence. It is true that Rebecca’s desires
are calm, whereas Alex’s are violent. But of greater significance are the facts that
Rebecca’s are generated by keen and careful judgment, whereas Alex’s are not;
Rebecca’s are characterized by rich and sophisticated interrelations and connections,
69 FW&CP, pp. 22-23.
70 The role of autonomy as a precondition for accountability is discussed in more detail in Section 5 of Part Three.
68
where Alex’s are disconnected from one another; and Rebecca’s are constant and reliable
over time, whilst Alex is prone to sudden reversals and lurches of interest.”71
When we are interested in an agent’s autonomy, I conjecture, we are interested in all of
these sorts of properties, mentioned above, which an agent may manage to instantiate.
We care about normative competence, diachronic stability, coherence and
sophistication.72 Our concern with an agent’s autonomy, when viewed with a synoptic
eye, is a complex concern with a multiplicity of different virtues of judgement and
practical reason. Rebecca is not better than Alex in one way, but in many. Her agency is
more excellent than his in a number of irreducible dimensions, not by virtue of
possessing one single property that his does not.
Whilst Frankfurt’s search for a ‘silver-bullet’, a single criterion for deciding on an agent’s
status as free and autonomous, was a noble philosophical endeavour, we should learn
from it by drawing the most fundamental lessons of its failure, and not by heading off on
a quest for an alternative single criterion. Rather, we should learn from it by seeing that to
treat agential autonomy as an either/or property is to ride roughshod over the remarkable
complexity of agency, volition and practical reason. Autonomy does not seem to be
something that one simply has or lacks, it is something of which one has more or less,
according to a complex set of measures, over a number of separate dimensions.
71 See §3 above.
72 I do not claim this as an exhaustive list. Indeed, it may not be possible, ahead of time, to specify the total set of normatively significant properties in which we might reasonably be interested when we assess an agent’s autonomy.
69
Our own aims, as agents who value autonomy, should always be understood with this
complexity in clear view. Similarly, our practice of morally assessing the actions and
wills of others needs always to be sensitive to the ways in which their agency succeeds
(or fails) to embody genuine autonomy, and needs also to be sensitive to the ways in
which their agency is constrained or encumbered. Thus, if we are to avoid excessive
harshness and judgmentalism in our ethical judgements, we must keep the complexity of
autonomy in view. The lesson to be learnt from the foregoing investigation is, thus, a
lesson about the need for a pluralist account of autonomy. Such an account, in its full
elaboration, would explain all of the dimensions in which human agency can be more or
less excellent. It would help us to finesse our own goals as practical agents in the world,
and would help us to avoid the pitfalls of unconstrained judgmentalism, which would be
too eager to hold responsible those with only a highly limited degree of agential
autonomy. Such an alternative theory is likely to be, in a sense, frustratingly complex.
But in its complexity it would be doing nothing less than being fair to the complexity of
the normative facts.
70
Freedom, Fairness and Responsibility
Part Two
Freedom without Resentment:
Responsibility and
the Reactive Attitudes
71
Freedom, Fairness and Responsibility – Part Two
Freedom without Resentment: Responsibility and the Reactive Attitudes
1. Strawson’s View: Practical Compatibilism as a Form of Naturalism
The subject matter of my discussion in the second part of this dissertation is P. F.
Strawson’s highly influential essay, ‘Freedom and Resentment’,73 and the picture that it
paints of the relationship between moral judgement, responsibility and human reactive
attitudes and emotions. In what follows, I intend to explore a number of respects in
which Strawson’s picture might be misleading and, in so doing, to suggest some
alternative ways in which we might think about these issues. Strawson’s pioneering essay
has propelled debate about moral responsibility in fruitful directions, and has helped to
advance our understanding of an important range of issues concerning our moral and
emotional lives. My hope is that, through showing the fault-lines in Strawson’s view – its
lapses, elisions and lacunae – we can advance our understanding of these issues still
further.
As has often been pointed out, Strawson’s ‘Freedom and Resentment’ changed the
course of the free will debate, and moved it into hitherto uncharted territory. Strawson’s
intervention in that debate can essentially be seen as an attempt to recast the central
73 P. F. Strawson, (1962), “Freedom and Resentment,” Proceedings of the British Academy, 48, pp. 1-25, reprinted in Gary Watson, (ed.), (2003), Free Will, 2nd edition (Oxford: Oxford University Press), pp. 72-93. All page references to “Freedom and Resentment” (henceforth: “F & R”) give the pagination from the Watson volume.
72
problem of free will and moral responsibility as a practical rather than a theoretical
issue. In so doing, his avowed goal was to effect a ‘reconciliation’ between what he calls
the ‘optimist’ (i.e. compatibilist) and ‘pessimist’ (i.e. incompatibilist) positions. Such a
characterization is in some ways odd, as Strawson’s argument offers us nothing that
looks like a reconciliation, as such (and certainly not a reconciliation that the warring
parties could be thought likely to accept).74 But we might put things another way, and
say that Strawson’s naturalistic appeal to the real texture and structure of human
practices and attitudes seeks to transcend (or perhaps to short-circuit) the traditional
metaphysical stand-off between the competing parties in these debates.
Strawson’s naturalistic approach involves highlighting the intimate link that connects the
practice of holding someone to be responsible or accountable with what he calls the
‘reactive attitudes’. By presenting both judgements of moral responsibility and practices
of holding people responsible as being inseparable from deeply entrenched and
ineradicable human attitudes and reactions, Strawson presents a way of thinking about
moral responsibility that seeks to avoid traditional metaphysical disputes. He thereby
presents a non-standard, ‘practical’ compatibilism, by standing the usual order of
argument on its head: that is, by starting, not with highly abstract philosophical
considerations concerning the metaphysics of the will or the meanings of disputed
74 A truly ‘reconciling’ strategy, one might think, would take what is true in both kinds of view, and combine them into a single ‘hybrid’ view. (For such an approach, Part Three of the this dissertation.) Strawson’s strategy, on the other hand, is really closer to calling a plague on both their houses – he tells both the compatibilist and the incompatibilist that they miss the point through “over-intellectualizing” the free will debate. They are seeking, he tells us, to answer a “useless” question (F & R, p. 87).
73
theoretical terms, but with considerations rooted in our actual behaviour, practices and
psychology.
Strawson’s essay is dense and wide-ranging, and touches on a striking diversity of issues
concerning our emotional lives as personally inter-related individuals. Interestingly,
discussion of moral responsibility, as such, arises only in the latter half of his essay, for
he starts first with an examination of our ‘personal’ reactive attitudes: that is, the sorts of
attitudes and emotions that we display in ordinary interpersonal human relationships (the
attitudes, as he puts it, of “offended parties and beneficiaries”75). Strawson’s strategy is
to show, prior to moving on to the moral case, that we are non-rationally (n.b., not
irrationally) committed, as a basic matter of our psychological constitution as human
beings, to these interpersonal attitudes.
Strawson’s strategy is then to argue by extension from the interpersonal case to the case
of the ‘vicarious’ or ‘moral’ analogues of the interpersonal reactive attitudes. On
Strawson’s view, moral judgements should be understood as involving ‘impersonal’ (or
‘vicarious’) attitudes that are essentially continuous with the analogous attitudes that we
experience in direct interpersonal contexts. Moral judgement, on this view, involves the
expression of a reactive emotion (moral ‘indignation’) that is the vicarious analogue of
the attitude of resentment for harm in the simple interpersonal case. Given this
conception of moral judgement and of ascriptions of moral responsibility, as essentially
expressive of reactive emotions, Strawson argues that the same results can be obtained in
75 F & R, p. 83.
74
the case of our moral attitudes and judgements as he had obtained for the case of the
interpersonal attitudes. Thus, as in the case of the interpersonal reactive attitudes,
Strawson argues that we are non-rationally and inextricably committed, as a matter of
natural fact, to the full range of reactive moral attitudes, and thereby to making
ascriptions of moral responsibility.
On Strawson’s view, then, we have a natural commitment to the full set of reactive
attitudes, both (i) interpersonal and, by extension, (ii) moral. Thus, the question of the
conceptual ‘compatibility’ of free will and moral responsibility with an understanding of
human action as determined (or simply as a natural phenomenon like any other) is
rendered otiose. Even to raise such a theoretical question, Strawson thinks, would show
that we have “utterly failed to grasp … the fact of our natural human commitment to
ordinary interpersonal attitudes”.76 As he states his position in a later defence of his view,
“our proneness to reactive attitudes is a natural fact, woven into the fabric of our lives,
given with the fact of human society as we know it, neither calling for nor permitting
general ‘rational’ justification.”77 Hence, the dispute between the compatibilist and the
incompatibilist, if not exactly reconciled, is perhaps better seen as being dissolved.78
76 F & R, p. 83.
77 See P. F. Strawson, “Reply to Ayer and Bennett,” in Zak van Straaten, ed., (1980), Philosophical Subjects: Essays Presented to P. F. Strawson, (Oxford: Clarendon Press), at p. 265.
78 The Wittgensteinian echo here is quite deliberate, as there seems to be a quite significant Wittgensteinian strand in Strawson’s work (and, especially, in “Freedom and Resentment”.) See for example Strawson’s comment at F&R, p. 75 that “What I have to say consists largely of commonplaces”. This is reminiscent of a number of Wittgenstein’s claims, as when Wittgenstein tells us that “The work of the philosopher consists in assembling reminders for a particular purpose”, or shortly afterwards, when he says that “the aspects of things that are most important for us are hidden because of their simplicity and familiarity.” (Ludwig Wittgenstein, (1953), Philosophical Investigations, (Oxford: Blackwell), §§127, 129.) Or again, where Wittgenstein claims that (slightly earlier in the same sequence of remarks): “It [philosophy] leaves everything as it is.” (Wittgenstein, ibid., § 124) Strawson himself says, in a 1995 statement of “My
75
It is thus a feature of Strawson’s view that it defends a form of ‘practical’ compatibilism
that is (at least in its conception) structurally immune to philosophical assault from any
incompatibilist or sceptical source. Strawson’s conception of responsibility and the
reactive attitudes thus acts as a general device for forestalling revisionism or scepticism
about moral responsibility. It does this whether with regard to the thoroughgoing varieties
of ‘revisionism’ that involve a wholesale abandonment of the domain of the moral (as,
perhaps, with certain Nietzschean critiques of morality, or as in Bernard Williams’s
undermining of “the morality system”79), and also with regard to more modest varieties
of revisionism, which perhaps involve scepticism only about blame and desert-entailing
attitudes, rather than the entire moral domain. Whatever the revisionist strategy being
pursued on the question of moral responsibility – whether, for example, the standard
incompatibilism of Peter van Inwagen or the more direct scepticism about responsibility
associated with Galen Strawson – acceptance of P. F. Strawson’s account of
responsibility would have the effect of foreclosing any such argument before it could get
going.80
Philosophy” that “I must mention Wittgenstein; not, I think, because I directly derive any particular doctrine or view from him – far from it – but because I think there must be a kind of general, non-specific influence there, in so far as I sometimes feel my own work to be in one way, though certainly not in another, quite Wittgensteinian in spirit.” See Pranab Kumar Sen and Roop Rekha Verma, eds., (1995), The Philosophy of P F Strawson, (New Delhi: Indian Council of Philosophical Research), p. 18.
79 See Friedrich Nietzsche, (2006) [1887], On the Genealogy of Morality, 2nd edition, ed. Keith Ansell Pearson, trans. Carol Diethe, (Cambridge: Cambridge University Press) and, for a more recent view, influenced by Nietzsche, see Bernard Williams, (1985), Ethics and the Limits of Philosophy, (London: Fontana).
80 See for example, Peter van Inwagen, (1986), An Essay on Free Will, (Oxford: Oxford University Press); and Galen Strawson, (1986), Freedom and Belief, (Oxford: Clarendon Press) and his (1994), “The Impossibility of Moral Responsibility”, Philosophical Studies, 75: 5-24.
76
Strawson’s strategy thus involves a wholesale neutralization of any revisionist argument
about responsibility. In effect, he takes on no one specific revisionist argument, but rather
seeks to take them all on, to draw the sting from them all. That this is so can be seen right
from the start of ‘Freedom and Resentment’, where Strawson addresses himself not only
to the ‘pessimist’ (i.e. the incompatibilist), but also to the “the genuine moral sceptics”,
who hold that “the notions of moral guilt, of blame, of moral responsibility are inherently
confused…”.81 If Strawson’s strategy works, then all revisionist arguments must fail, and
the (merely) theoretical discussion of philosophical revisionism about moral
responsibility is misplaced and redundant. If Strawson is right, then we simply have to
take the commonplace phenomenology of holding individuals responsible fairly much as
we find it.
Strawson’s view is marked by both the originality of its conception and the attractiveness
of its conclusions, not least insofar as it suggests that one can pre-empt sceptical or
incompatibilist worries about moral responsibility without having to engage with sceptics
and incompatibilists on their own argumentative terrain. This conclusion is attractive,
moreover, not only because it avoids the interminable squabbling of metaphysical
disputation, but also because it leaves everything as it is in our moral lives, and avoids
any threat of a need for difficult or disturbing changes in our moral practices. Perhaps due
in part to these attractive features, Strawson’s view has proved highly influential since its
publication, and has inspired and provoked a very large number of its readers, winning
81 F & R, p. 72. It is worth pointing out that one could think that ideas of guilt, blame and responsibility were inherently confused without being a comprehensive moral sceptic. As will be seen, it is significant that Strawson appears to see no such possibility.
77
over a strikingly high proportion of them. In contemporary thinking about responsibility,
whether in ethics, political philosophy, or philosophy of law, Strawson’s view looms
large. In order to discover whether Strawson’s anti-revisionist approach can hold the
court in all these domains, we must examine how successfully his view is supported by
his arguments, and whether problems can be found with his ambitious naturalization of
the problem of moral responsibility. If Strawson’s view can be shown to fail, then the
possibility of revisionist or sceptical approaches to responsibility would again open up in
all of the domains mentioned here – across a broad spectrum, from the interpersonal, to
the moral, to the political.
Having laid out the nature and context of Strawson’s view, my own strategy will have
two main parts. Firstly, I shall start by examining Strawson’s account of the interpersonal
reactive attitudes, and interrogating his method for securing these interpersonal attitudes
from any kind of revisionist attack. I will lay special emphasis on problems with certain
destabilizing ambiguities that lurk within Strawson’s characterization of the reactive
attitudes, and that also lurk within his idea of an ‘objective attitude’. Secondly, I will
move on, as Strawson does, from the interpersonal case to the case of our moral attitudes
and judgements, and examine whether Strawson’s extension of his method to the moral
case is plausible. Here, I would like to suggest that, setting aside problems with
Strawson’s approach that emerge even in the interpersonal case, there are special
problems that face his naturalistic account of moral responsibility and of the nature of
moral judgement. I will argue that, even if we were reasonably minded to accept
Strawson’s approach with regard to non-moral interpersonal attitudes, we might
78
nevertheless have sufficiently good reasons to resist the extension of his strategy to the
domain of the moral.
2. Relationships and Reactive Attitudes
The first stage of Strawson’s argument is to describe the ‘reactive attitudes’ which occur
in ordinary interpersonal relationships, and then to attempt to show that such reactive
attitudes, whilst they may be suspended in certain special circumstances, are nevertheless
immune to any general form of critique from the sceptic or incompatibilist. In this part of
my essay, I want first of all to examine Strawson’s characterization of these ‘reactive
attitudes’ and then to move on to an assessment of the plausibility of his attempt to
insulate these attitudes from sceptical or incompatibilist attack.
A. What are the reactive attitudes? – Broad and narrow conceptions
Strawson’s paper has been influential in a number of ways, not least in coining the
philosophical term-of-art ‘the reactive attitudes’. At this point it is worth getting as clear
as possible on what exactly is meant by Strawson when he refers to these ‘reactive
attitudes’, as it is striking that, despite the subsequent popularity of the phrase, it is often
used without a completely clear sense of the range of attitudes that it covers. In
Strawson’s essay, as well as subsequently, the phrase seems to get used in both a narrow
and a broad way. When used more narrowly, the ‘reactive attitudes’ seem to be
essentially ‘desert-entailing’ – which is to say that they are a range of attitudes that are
not merely responsive to the behaviour of others, but which involve the withdrawal of
79
good will or the presence of ill will towards the person to whom the attitude is directed.82
Resentment is thus the paradigmatic reactive attitude on this ‘narrow’ reading, as it
essentially involves the withdrawal of good will. This is made most clear when, in the
latter part of his essay, Strawson stresses the continuity between the reactive attitudes and
certain moral attitudes, which he takes to be desert-entailing in the same way in which he
understands the attitude of resentment to be. Strawson tells us that:
[Moral] Indignation, disapprobation, like resentment, tend to inhibit or at least to limit our goodwill towards the objects of these attitudes, tend to promote an at least partial and temporary withdrawal of goodwill: they do so in proportion as they are strong; and their strength is in general proportioned to what is felt to be the magnitude of the injury and to the degree to which the agent’s will is identified with, or indifferent to, it. (These, of course, are not contingent connections.)83
Given that Strawson stresses that the connection is not a contingent one, we would
reasonably conclude that he means us to understand that a necessary, conceptual
connection holds between reactive attitudes such as resentment and the suspension of
good will (or activation of ill will) towards the object of that attitude. This is, I take it, the
standard understanding of the nature of resentment as an attitude. That the attitudes in
question are, on Strawson’s view, desert-entailing in this way, is confirmed by what he
goes on to say later in the same passage:
The partial withdrawal of goodwill which these attitudes entail, the modification they entail of the general demand that another should, if possible, be spared suffering, is,
82 Like others who have used the term since, I take the term ‘desert-entailing’ from Galen Strawson, Freedom and Belief, ibid. The characterization given above covers attitudes of negative desert; but there is also the positive analogue of such attitudes, involving the amplification or extension of good (rather than ill) will towards the person to whom the attitude is directed, as is the case with attitudes such as gratitude.
83 F & R, p. 90. (My italics.)
80
rather, the consequence of continuing to view him as a member of the moral community.84
The connection between the reactive attitudes and the (‘desert-entailing’) belief that they
carry with them in the appropriateness of suffering or punishment is further stressed by
Strawson when he moves on to the case of the ‘self-reactive’, or reflexive, analogues of
the reactive attitudes:
Just as the other-reactive attitudes are associated with a readiness to acquiesce in the infliction of suffering on an offender, within the ‘institution’ of punishment, so the self-reactive attitudes are associated with a readiness on the part of the offender to acquiesce in such infliction without developing the reactions (e.g. of resentment) which he would normally develop to the infliction of personal injury upon him: i.e. with a readiness, as we say, to accept punishment as ‘his due’ or as ‘just’.85
Therefore, we see that the attitude of resentment is taken by Strawson to go alongside the
idea that it would be appropriate that its object undergo some suffering.86 Thus,
resentment is a ‘desert-entailing’ attitude whether one has a teleological or an attitudinal
84 F & R, p. 90. (Strawson’s italics.)
85 F & R,. pp. 90-1.
86 This would seem to many to be the natural understanding of what we mean when we talk of ‘resentment’, but it is striking that T. M. Scanlon demurs: “On my view, when moral criticism applies this makes various reactive attitudes such as guilt, resentment and indignation appropriate. But these attitudes do not, as I understand them, entail the thought that it would be a good thing, or not a bad thing, if the person to whom they are directed should suffer in some way.” (See Scanlon, (1998), What We Owe to Each Other, (Cambridge, Mass.: Harvard University Press), p. 276). As against Scanlon’s understanding of resentment, I myself find it hard to understand how an attitude could be an attitude of resentment, as opposed to being some weaker attitude of disapproval, if it were not desert-entailing in this way. Scanlon’s view thus raises the question of how resentment could be a distinctive attitude if it were not desert-entailing. Such difficulties are avoided if, following Strawson, and fitting with standard usage, we understand resentment to be an essentially desert-entailing attitude. (The OED has resentment as being “a strong feeling of ill-will or anger against the author or authors of a wrong or affront”, which seems to be about right.) (Scanlon presents a modulated view of the nature of blame and (some of) the reactive attitudes in his recent book, (2008), Moral Dimensions: Permissibility, Meaning and Blame, (Cambridge, Mass.: Harvard University Press), but a careful discussion of Scanlon’s revised view is beyond the scope of the current discussion.)
81
understanding of desert.87 What we can also see is that Strawson moves freely between
identifying this ‘desert-entailing’ feature as being peculiar of resentment, specifically,
and identifying it as being a general feature of the reactive attitudes in general. Thus, on
this ‘narrow’ reading of the reactive attitudes, what marks them out is that they are
essentially ‘desert-entailing’ attitudes. Nevertheless, one should be careful to note that
not all reactive attitudes need involve a negative desert claim: for example, the attitude of
gratitude (another attitude that Strawson often uses in examples) and, presumably, (what
Strawson calls) its “vicarious analogue”, praise, involve a positive desert claim, to the
effect that the agent at whom the attitude is directed deserves some positive outcome, or
is the appropriate object of good will.
Many of those who have followed Strawson have adopted this narrow understanding of
the reactive attitudes as essentially desert-entailing, and it is certainly a reading of
Strawson’s view that seems entirely natural given what he says in the latter stages of his
essay.88 However, a very different, and much broader, understanding of the reactive
attitudes is suggested by other elements of Strawson’s discussion. This is especially
evident when the idea of a reactive attitude is first mentioned. When Strawson initially
introduces the idea of a reactive attitude, he does so by means of listing an extremely
87 Some explanation of this distinction will be useful here. Following Derek Parfit, I define a teleological desert-claim with regard to some agent, X, as being of the form “It is good that X suffers” (strong) or “It matters less than it otherwise would if X suffers” (weak); an attitudinal desert claim entails either that we are prima facie justified in withdrawing good will or (more strongly) in extending ill will towards X.
88 Although Scanlon rejects a desert-entailing account of reactive attitudes, he nevertheless takes Strawson himself to be defending a ‘desert-entailing’ conception of those attitudes. See fn. 17, p. 400 of What We Owe to Each Other: “Here my understanding of these attitudes [the reactive attitudes, specifically including “guilt, resentment, and indignation” (p. 276)] differs from Peter Strawson’s, since he takes them to entail “a partial withdrawal of goodwill” and a “modification of the general demand that another should, if possible, be spared suffering.” … Thus, on his understanding, but not mine, these are desert-entailing notions.”
82
heterogeneous group of attitudes, some of which are ‘desert-entailing’ in the above
senses (i.e. whether attitudinal or teleological), and others of which are not. As he puts it,
“I want to speak… of the non-detached attitudes and reactions of people directly involved
in transactions with each other; of the attitudes and reactions of offended parties and
beneficiaries; of such things as gratitude, resentment, forgiveness, love and hurt
feelings.”89 Whilst gratitude and resentment, and perhaps forgiveness, are plausibly seen
as ‘desert-entailing’ attitudes, this is much less plausible in the case of love, and even less
plausibly the case with regard to “hurt feelings”. Hurt feelings might involve one being
saddened, or shocked, or upset, but they have no conceptual connection to the belief that
the individual who caused those hurt feelings should be the object of ill will, or that it
would be appropriate for he or she to suffer.
Strawson’s first explicit use of the term “reactive attitudes” thus suggests that he means
to identify a very broad class of attitudes, covering any attitude whatsoever that takes as
its object the attitude towards oneself of another, within the context of a relationship (of
whatever kind) between oneself and that other agent. As he puts it: “Then we should
think … of the kind of importance we attach to the attitudes and intentions towards us of
those who stand in those relationships to us, and of the kinds of reactive attitudes and
feelings to which we ourselves are prone. In general, we demand some degree of
goodwill or regard on the part of those who stand in these relationships to us, though the
forms we require it to take vary widely in different connections. The range and intensity
of our reactive attitudes towards goodwill, its absence or its opposite vary no less
89 F & R, p. 75.
83
widely.”90 Here again it would seem that the reactive attitudes can be of any kind, as long
as they are reactive, in the sense of responding to another’s will or behaviour. What
marks them out is not their content (i.e. that they are desert-entailing), but merely their
object (the will of another individual, with whom one stands in some kind of
relationship).
Therefore, it would seem that the ‘official’ Strawsonian characterization of the reactive
attitudes is of the broad rather than the narrow kind. It is perhaps due to Strawson’s
concentration on a narrow subset of reactive attitudes – specifically, the recurring triad of
resentment, gratitude and forgiveness – later in his essay, that it can be tempting to think
that the category of reactive attitudes identifies only a narrow set of desert-entailing
attitudes, but this would be to read back the later parts of Strawson’s argument into the
initial introduction of the idea of the ‘reactive attitudes’. Given this, in what follows, it
will be important to bear in mind that the reactive attitudes can be of a wide range of
different kinds, some desert-entailing (in either the teleological and/or attitudinal senses),
such as resentment itself, and some most certainly not desert-entailing, such as ‘hurt
feelings’.
This definitional or conceptual point will be seen to be significant as my critique of
Strawson’s position progresses, for some of Strawson’s claims are plausible only on a
narrow reading of the ‘reactive attitudes’ and others are plausible only on the broad
reading; and his overall view earns unwarranted plausibility by means of some
90 F & R, pp. 76-77.
84
equivocation between the two readings. For example, the power of Strawson’s claim that
we are unable to give up on the reactive attitudes, as a whole, is much stronger when we
understand ‘the reactive attitudes’ to cover all of our attitudes towards others that result
from interpersonal relationships; it is considerably less plausible if it is taken to be a
claim only about a specific, desert-entailing, subset of our overall set of available
attitudes. That Strawson is sometimes rather cavalier with his concepts, and hence
equivocal between these two ways of understanding what is meant by ‘the reactive
attitudes’, undermines the cogency of his view. This should become clear as my
discussion develops, but I want first to examine the positive account of the reactive
attitudes that Strawson offers, together with his account of when and why we actually do
consider it appropriate to suspend these attitudes.
B. Excuses and Exemptions – ‘Quality of Will’ and ‘Capacity’
Perhaps the most outstandingly insightful element of Strawson’s account of the reactive
attitudes is his theory of when and why we take it to be appropriate to suspend or curtail
those attitudes. Here, Strawson does a tremendous job of diagnosing some deep features
of the day-to-day phenomenology of interpersonal relationships. He points out that our
criteria for suspension of the reactive attitudes are basically twofold: one kind of criterion
is concerned with the nature of the agent, and the other with the relationship between the
action to which we are reacting and the will and intentions of the agent who performed it.
Whatever view one takes of Strawson’s ‘naturalistic compatibilism’, this descriptive
presentation of the phenomenology of interpersonal relationships and reactions should be
acknowledged as displaying tremendous insight.
85
The first of Strawson’s criteria is concerned with whether the action at stake can truly be
seen as issuing from the intentions of the agent who caused it. Strawson points out that
we are concerned by the behaviour of others to the extent that it manifests certain
attitudes towards us, and so the nature of the reactive attitude which we hold towards a
given agent for a given action will depend on what we take to be the underlying volitional
state that stands behind that action. As T. M. Scanlon puts it, what Strawson tells us is
important with regard to interpersonal relationships is the ‘Quality of Will’ which the
behaviour of another instantiates.91
Thus, we can call this first criterion the Quality of Will Criterion. X will prima facie tend
to resent any action of Y’s which causes X harm of some kind, but that resentment is
defeasible in the face of considerations such as “Y didn’t mean to do that”, or (to some
extent) “Y did not realize that her action would have those effects for X”, or any other
excuse which shows that Y’s will was not antagonistic towards X. If I push you roughly
to the ground, we might generally think that it makes sense for you to resent me for doing
so; but, on learning that I was pushing you out of the way of a runaway trolley car, and
was acting out of a keen concern for your welfare or survival, your resentment would
quickly and naturally disappear (although not, perhaps, if I seemed nevertheless to relish
this occasion to give you a good shove, and had thus pushed much more violently than
the situation strictly demanded.) Any elaboration of my behaviour which gives a fuller
91 See T.M. Scanlon, (1986), “The Significance of Choice,” in Sterling M. McMurrin (ed.), The Tanner Lectures on Human Values, Volume 8, (Salt Lake City, Utah: University of Utah Press), reprinted in Stephen Darwall (ed.), (1995), Equal Freedom: Selected Tanner Lectures on Human Value (Ann Arbor: University of Michigan Press), pp. 39-104.
86
understanding of the quality of my will towards you can be used as the basis for an
excusing condition, insofar as your reactive attitudes towards me take as their object not
simply my behaviour, as such, but the way in which that behaviour evinces a particular
attitude or will towards you.
Strawson’s own list of excusing conditions which defeat the reactive attitudes for
‘Quality of Will’ reasons include: “‘He couldn’t help it’, when this is supported by such
phrases as ‘He was pushed’, ‘He had to do it’, or ‘It was the only way’”.92 (It should be
noted that Strawson would not see ‘He couldn’t help it’ as an excusing condition if it was
taken in a general, or one might say metaphysical sense, as meaning something of the
same sort as “He was causally determined”, or if it was otherwise unsupported by more
specific elaborations of the type that he gives above.) These ‘excuses’ do not diminish the
standing of the agent, but only redescribe or recharacterize his or her behaviour or, more
accurately, the intentions or volitions that lie behind that behaviour. As Strawson puts it:
They [the excuses] do not invite us to view the agent as one in respect of whom these attitudes are in any way inappropriate. They invite us to view the injury as one in respect of which a particular one of these attitudes is inappropriate. They do not invite us to see the agent as other than a fully responsible agent. They invite us to see the injury as one for which he was not fully, or at all, responsible. They do not suggest that the agent is in any way an inappropriate object of that kind of demand for goodwill or regard which is reflected in our ordinary reactive attitudes.93
Alongside these sorts of cases, there are also cases where we suspend the reactive
attitudes precisely because we see that the object to whom we are directing our attitudes
92 F & R, p. 77.
93 F & R, p. 77-78.
87
are not agents of a suitable kind. Thus, the second criterion is given by the fact that we do
not think the reactive attitudes appropriate with regard to anything other than normal,
adult, human agents. This gives us what we might call the Capacity Criterion. As the
reactive attitudes get their value, and find their role, in normal, adult relationships, it is
inappropriate to have such an attitude towards someone who is not a normal adult human
and thereby a competent partner or interlocutor in adult relationships (of whatever kind).
An individual may fall short of the standards of full, adult agential capacity either, as we
may say, structurally, through some general lack of ability, or temporarily, as with the
agent disabled, or at least frazzled, by, for example, grief, stress or drink. Thus, we
suspend our reactive attitudes ‘structurally’ when we learn that the person to whom we
are applying them is mentally incapacitated, or too young, and ‘temporarily’ when we
learn that he is, perhaps, suffering from extreme stress, is hypnotized, or, in some other
way, lacks the capacities of the ‘normal’ or fully competent agent.94
So, on Strawson’s view, the reactive attitudes are appropriately suspended when they do
not respond accurately to the ‘Quality of Will’ evinced by the behaviour of a fellow,
competent agent with the full ‘Capacities’ of a normal adult human. Taken in a limited
way, as an account of many central cases when we appropriately suspend our reactive
attitudes, Strawson’s picture is a compelling one. It fits in with the nature of our
94 This is not to say that, on Strawson’s view, such attitudes are never appropriate with regards to children, but only that there is a great deal of vagueness and difficulty in judging the appropriateness of such attitudes with respect to them, or with respect to ‘non-normal’ adult humans. (For an interesting and thought-provoking neo-Kantian discussion of questions of the normative status of children as agents, see Tamar Schapiro, (1999), ‘What is a Child?’, Ethics, 109: 715-38.) As a matter of natural fact, we also seem often to resent computers, machinery, the weather, and so on. These cases are more straightforward, as we can all quite readily see how they are never genuinely appropriate. A malfunctioning laptop is not a ‘competent’ agent or co-participant in an adult relationship.
88
interpersonal emotional lives that what exercises us is not simply what happens in our
relationships, but how what happens flows from the way in which we are regarded and
treated by others. We care less about someone’s bare physical movements than we do
about the way that they regard us and the nature of their will towards us. Accordingly, we
step back from attitudes such as resentment when we realize that the person with whom
we are dealing lacks (temporarily or structurally) the capacities of full moral agency.
These claims seem unarguable.
What is much more problematic, though, is what one might call the generalized extension
(or ‘Strong’ reading) of Strawson’s claims, taken as a determinate and complete picture
of when the reactive attitudes are appropriate and when they are not. We could agree that
Strawson gives us two plausible sufficient criteria for when it is appropriate to suspend
the reactive attitudes (let us call us this the ‘Weak’ reading of Strawson’s remarks about
‘Capacity’ and ‘Quality of Will’), but we need not take the view that he has given us a
full and complete account of when the reactive attitudes are appropriate, as we may think
that there are other kinds of possible excusing or excluding conditions for some, or all, of
the reactive attitudes. In other words, we could read Strawson as having provided us with
a good account of two sufficient conditions for suspending the reactive attitudes without
thinking that he has provided us with (disjunctive) necessary conditions for when the
reactive attitudes might properly be suspended.
On a ‘Strong’ reading of Strawson’s remarks, the full set of reactive attitudes are always
fully appropriate when the ‘quality of will’ and ‘capacity’ criteria are both met. On a
89
‘Weak’ reading, there may be other reasons for considering such attitudes inappropriate,
even when both criteria are met. It is important not to slide from acceptance of the
‘Weak’ version of Strawson’s view here to acceptance of this ‘Strong’ version; although,
as we shall see, this slide is something of which Strawson himself appears in places to be
guilty. But before moving on to address the question of whether Strawson does enough to
support a ‘Strong’ reading of his view, and whether there may be an illegitimate slide
from the ‘Weak’ to the ‘Strong’ version, it would first be useful to dissect the ‘attitudinal
alternative’ to the reactive attitudes which Strawson takes to be available: that is, the so-
called ‘objective attitude’. For only once we have a clear view of the nature and structure
of the ‘objective attitude’, which Strawson takes to be the only available alternative to
individuals if they are to avoid expression of the reactive attitudes, can we then fully
assess the plausibility of Strawson’s views on the personal reactive attitudes and
emotions.
C. Making Sense of the ‘Objective Attitude’
Strawson introduces the Objective Attitude with reference to our attitudes in cases in
which an individual fails (structurally) to satisfy the Capacity criterion. Excluding
someone from being the subject of a reactive attitude on ‘Capacity’ grounds involves a
much more significant move than excluding someone on the basis of a particular
judgement about their Quality of Will. It involves excluding them, as Strawson sees it,
from the whole ambit of reactive attitudes and emotions, rather than just suggesting the
inappropriateness of one particular kind of attitude. It places them outside the realm of
interpersonal relations, and thus marks them out as someone with regard to whom it is
90
inappropriate, as Strawson has it, to have any reactive attitudes at all (either temporarily,
in some cases, or permanently, in cases of structural incapacity). With regard to such
individuals, Strawson describes the attitude that we take as the ‘Objective attitude’,
understood as an alternative stance to the ‘Participant attitude’ that stands behind the set
of multifarious, particular reactive attitudes. He introduces his discussion of the
‘Objective attitude’ in this very rich passage:
When we see someone in such a light as this [i.e. as failing to meet the ‘Capacity’ criterion], all our reactive attitudes tend to be profoundly modified. I must deal here in crude dichotomies and ignore the ever-interesting and ever-illuminating varieties of case. What I want to contrast is the attitude (or range of attitudes) of involvement or participation in a human relationship, on the one hand, and what might be called the objective attitude (or range of attitudes) to another human being, on the other. Even in the same situation, I must add, they are not altogether exclusive of each other; but they are, profoundly, opposed to each other. To adopt the objective attitude to another human being is to see him, perhaps, as an object of social policy; as a subject for what, in a wide range of sense, might be called treatment; as something certainly to be taken account, perhaps precautionary account, of; to be managed or handled or cured or trained; perhaps simply to be avoided, though this gerundive is not particular to cases of objectivity of attitude.95
It is significant here that, just as he introduces the idea of the ‘objective attitude’,
Strawson makes an explicit admission that he is here dealing in crude dichotomies that
fail to capture the variety of real cases. This is just an aside in the passage, of course, and
presumably Strawson’s considered view was that the picture which one could derive from
the starkest distinctions between the objective and participant attitudes contained the
central truths about our reactive attitudes, with the intermediate ‘varieties of case’
involving no more than complex amalgams of distinct attitudes. But, what I want to
suggest is that the importance of avoiding ‘crude dichotomies’ in this area is greater than
95 F & R, p. 79.
91
Strawson seems to allow, and that, far from being marginal or unimportant, close
attention to the ‘intermediate’ varieties of case allow us to see that this field contains
more complexity, of a philosophically significant kind, than Strawson’s more schematic
discussion would allow.
On Strawson’s view here, the ‘objective’ and the ‘involved’ (or ‘participant’) attitudes
are “profoundly opposed” to one another, even though they can, as a matter of
psychological fact, exist alongside each other, such that he allows that they are not fully
“exclusive”. One might say that their “profound opposition” is due to the fact that they
involve thinking about other people in two fundamentally incompatible ways, as we
might think about light either as a wave or as a particle. The fact that, in another sense,
they are not “exclusive”, is meant to suggest that these two opposed modes of thought
can exist simultaneously in the same person with regard to their attitude towards any
other particular individual – just as we may think of light as being, simultaneously, both
wave and particle.
As Strawson describes it, this ‘objective’ attitude involves seeing another person in
purely instrumental terms – as someone to be “treated”, “managed or handled or cured or
trained”,96 and not as someone whose volitional or behavioural states should be taken
seriously as those of a fellow, rational human being. On this understanding, the ‘objective
attitude’ is very stark indeed – it seems to involve nothing less than the eradication of any
kind of attitude of human involvement, in favour of a purely instrumental or technocratic
96 F&R, p. 79.
92
mode of treating another individual, as if he or she was no different, in essence, to a
machine, an animal, or some kind of natural obstacle. This leads one to think that the
‘objective attitude’ therefore involves the suspension of any genuine reactive attitude
towards another. It is the attitude of the coldly utilitarian social planner, or of the
Orwellian state towards its ‘proles’, or of the similarly dystopian figures of the purely
technocratic and instrumentalizing businessman or scientist. The objective attitude, on
this ‘Stark’ reading, is purely bureaucratic, calculating, emotionally neutral and
colourless; it involves seeing others as means rather than ends – or, as it were, of taking
towards one individual the kind of attitude that the psychopath takes towards all other
individuals.
This ‘Stark’ reading of the objective attitude gives us a clear understanding of why
widespread adoption of such an attitude is incompatible with living a genuinely human
life, and it buttresses Strawson’s claim that it is inconceivable that we should adopt the
objective attitude, for merely ‘theoretical’ reasons, towards everyone else all the time.
(As would, be demanded (one might presumably think) by certain incompatibilist or
sceptical (“pessimist”) views.) Taking such a course would, on this ‘Stark’ reading, be
nothing less than opting to immerse ourselves deeply within a view of others that is
nothing less than a dehumanizing variety of the psychopathological. If this is really what
is involved in the adoption of the objective attitude, then it is surely unsurprising that
Strawson is so wary of that attitude. Moreover, if this were the right way of
understanding the objective attitude, then any philosophical view that commits us to the
93
adoption of such an attitude would surely appear as a view that we may have good reason
to attempt to avoid.
However, the ‘Stark’ reading is not the only way of making sense of the ‘objective
attitude’, for Strawson’s text also suggests a much weaker, ‘Moderate’ reading. For
example, immediately after the passage quoted above, we are told that:
The objective attitude may be emotionally toned in many ways, but not in all ways: it may include repulsion or fear, it may include pity or even love, though not all kinds of love.97
Clearly, this weakened version of the ‘objective attitude’ is rather different to the ‘Stark’
version sketched above. The ‘Stark’ version involves the suspension of the totality of
human reactive attitudes, on the ‘broad’ reading laid out in section (2A) above, such that
the behaviour and will of others prompts no kind of reactive attitudinal response
whatever from us. But, on this weakened conception of the objective attitude, we see that
it may include a variety of reactive attitudes, such that full objectivity of attitude is not
taken to preclude a broad variety of reactive emotional responses. Far from involving an
affectless instrumentalism in one’s dealing with others, the ‘weakened objective attitude’
can encompass, on Strawson’s own admission, “repulsion”, “fear”, “pity” and “love”,
presumably among other kinds of attitudes.
Thus, it is important to note that what is often understood to be the standard way of
understanding the nature of the ‘objective attitude’, as involving the exclusion of all
97 F & R, p. 79.
94
‘reactive attitudes’, must be mistaken if we are to make sense of this presentation of a
‘weakened’ version of the objective attitude. Adopting an objectivity of attitude cannot
involve the complete exclusion of all reactive attitudes, on the ‘broad’ conception of the
reactive attitudes which (as has been shown in Section 2(A) above) should be attributed
to Strawson. So, at most, the ‘objective attitude’ can involve the exclusion of some, but
not all, of the reactive attitudes. This leaves us with the question of how, with regard to
the ‘objective attitude’, the division between the excluded and permitted reactive attitudes
is to be drawn. Strawson’s next remarks suggest a way forward:
But it [the objective attitude] cannot include the range of feelings and attitudes which belong to involvement or participation with others in inter-personal human relationships; it cannot include resentment, gratitude, forgiveness, anger, or the sort of love which two adults can sometimes be said to feel reciprocally for each other.98
Strawson’s suggestion seems to be that, whilst the objective attitude does not exclude the
full set of ‘broad’ reactive attitudes, it does exclude a subset of those attitudes which, he
says, are those that “belong to involvement or participation with others” in inter-personal
relationships. This is not fully perspicuous as it stands. For, on the ‘broad’ understanding
of the reactive attitudes, what marks out those attitudes is simply that they take as their
objects the will or behaviour of others in the context of inter-personal relationships; they
are not otherwise marked out by their content. But it is clear that reactive attitudes of fear,
pity, repulsion or love can, and do, all occur within the context of inter-personal
relationships. Every single member of the full set of ‘broad’ reactive attitudes, insofar as
they are, indeed, reactions to the actions, will or behaviour of other people, by definition
belong to our “involvement or participation with others”. Therefore, on the face of it,
98 F & R, p. 79. (My italics.)
95
Strawson’s explicit restriction does not seem to have marked out any kind of sub-set of
the reactive attitudes, as all of the reactive attitudes are attitudes that belong to the
domain of interpersonal involvement and participation. So, Strawson must have in mind
something more specific.
One suggestion for what Strawson may have in mind here might be that he intends to
refer to those reactive attitudes which occur only in the context of inter-personal human
relationships. It is to be admitted, of course, that repulsion and fear can occur with regard
to a multiplicity of different kinds of objects, not all of which are other people, whereas
gratitude and forgiveness, for example, happen only in the context of interpersonal
relationships. But then the inclusion of ‘anger’ on Strawson’s list of reactive attitudes that
are excluded by the objective attitude looks rather odd, for we can be angry about all
sorts of things (malfunctioning computers, the weather, ourselves) outside of the context
of interpersonal reactive attitudes. So this suggestion is also unable to account for
Strawson’s distinction between those attitudes which are compatible with the objective
attitude and those which are incompatible with that attitude.
The best solution that would seem to be available, with regard to making sense of the
distinction between those reactive attitudes which are compatible with the objective
attitude and those that are not, would seem to involve a return to the distinction made in
Section 2(A) above, between the ‘broad’ and ‘narrow’ characterizations of the reactive
attitudes. In other words, the reactive attitudes which are incompatible with the objective
attitude actually seem to be those that fall under the scope of the ‘narrow’ set of desert-
96
entailing attitudes. This is suggested, in one significant way, because Strawson again
invokes the familiar desert-entailing triad of “resentment, gratitude and forgiveness”
when he lists those reactive attitudes which cannot operate from within the objective
attitude. The cut between ‘broad’ and ‘narrow’ reactive attitudes would thereby explain
why attitudes like resentment, gratitude and forgiveness should be incompatible with the
objective attitude, whilst attitudes like fear and repulsion are not.
Moreover, that the crucial cut is between desert-entailing and non-desert-entailing
attitudes also offers one plausible and sympathetic elaboration of Strawson’s claim that
the reactive attitudes that are excluded by the objective attitude are those that belong to
“involvement or participation with others in inter-personal relationships”, insofar as
desert-entailing attitudes do not have any plausible application other than with regard to
normal, competent human beings. For, insofar as a desert-entailing attitude involves a
belief that its object should suffer some misfortune, or that the welfare of its object is
thereby of lessened importance, it must apply, at any rate, to an object that has interests,
is involved in living a life, and for whom things can go better or worse. We cannot resent
or be grateful towards the weather, other than in a knowingly ‘degenerate’ or faulty
sense, even though we can be saddened by or afraid of the weather. Thus my suggestion,
by way of offering a maximally sympathetic reconstruction of Strawson’s
characterization of the objective attitude, is that the objective attitude is incompatible
with the full set of desert-entailing reactive attitudes (the reactive attitudes conceived
‘narrowly’), whilst not excluding the broader set of non-desert-entailing reactive attitudes
(that is, the reactive attitudes in the ‘broad’ sense).
97
Before moving on to consider where this leaves the idea of an ‘objective attitude’, we
need first to deal with one or two conspicuous outliers on Strawson’s list of the reactive
attitudes that he takes to be incompatible with the objective attitude. The exclusion of
resentment, forgiveness and gratitude makes perfect sense, as these are archetypal desert-
entailing attitudes. But the exclusion of “anger” and of “the sort of love which two adults
can sometimes be said to feel reciprocally, for each other”99 is harder to fathom. It may
well prove to be the case that Strawson’s list of excluded attitudes will need to be pruned
or reconfigured if we are to make the best possible sense of the contours of the objective
attitude.
As has been suggested above, anger is not the kind of attitude that is restricted in its
application only to other people, and so is implausibly characterized as being, in the
general case, a desert-entailing attitude. It is revealing, though, that anger seems at least
to border upon being a desert-entailing attitude. This becomes clearer if we consider the
difference between being “angry about x” and being “angry with x”. One can be angry
about all sorts of things (the state of American politics, global warming, or whatever
else), but it sounds odd to talk of being angry with such things. When we are angry with
someone (and it always does have to be with someone in order for the attitude to seem
fully in place), we typically also resent them, and hence we often speak of being ‘angry
with x’ as something like a shorthand way of describing our having a more fully
articulated desert-entailing attitude (of anger and resentment) towards x.
99 F & R, p. 79.
98
There are still cases, of course, of being angry with malfunctioning equipment, or indeed
of being angry with individuals who do not fulfil Strawson’s capacity criterion (the
young, the mentally ill, etc.), but we tend, at the same time, to recognize that such anger
has a certain inappropriateness in such cases (in much the same respect as we can see the
degeneracy of resenting a broken-down car100, or of being grateful towards a beautiful,
sunny day). So, I want to suggest that, although the inclusion of “anger” on the list of
attitudes that are incompatible with the objective attitude is somewhat problematic, we
can deal with this problem in one way through the assimilation of anger (in the sense of
“anger with x”) to a related type of desert-entailing attitude (i.e. we can place ‘anger’ in
the vicinity of resentment itself). This can be done whilst marking the distinction between
anger in this more specific sense and anger simpliciter (“anger about x”). With regard to
anger simpliciter, we can identify it as an attitude that is a mere impostor on this list of
“non-objective” attitudes – as a cuckoo in the nest of the desert-entailing reactive
attitudes.
So, what should we say about “the sort of love which two adults can sometimes be said to
feel reciprocally, for each other”?101 This is a more complex case. Clearly, there are
forms of love which involve no sort of desert-entailing attitude – love of certain
activities, love of pets, love of life, and so on. But reciprocal adult love, of the sort that
100 It is the “degeneracy” of demonstrating this kind of reactive attitude that makes us laugh so much at the sight of John Cleese as Basil Fawlty, thrashing his broken-down Austin 1100 car with a branch, having angrily warned it to start or face the consequences: BASIL: “Come on, start, will you!? Start, you vicious bastard!! … Right! Well…this is it! I'm going to give you a damn good thrashing!” See the “Gourmet Night” episode of Fawlty Towers, (BBC Television: 1975).
101 F & R, p. 79.
99
obtains (at least in the ideal case) between life-partners, might more plausibly be seen as a
desert-entailing attitude. It might plausibly be thought to involve a conception of its
object as the kind of agent who, by virtue of their characteristics or shared history,
deserved to flourish or, more likely, it would at least involve the thought that the
flourishing of this individual was something that mattered more (and ‘deservedly so’) to
the person who loved them. There is a great deal that could be said on this topic, but it
would be a diversion to pursue it much further here.102 Suffice to say that it is, at least,
quite intelligible to think of reciprocal adult-love as a desert-entailing attitude, and
thereby as incompatible with the ‘objective attitude’, even if, on closer inspection, it
might turn out that we should to some degree recharacterize our conception of this kind
of love so as to remove it from the domain of the desert-entailing.
We can now ask the question of where this leaves us with regard to characterizing the
objective attitude. First and foremost, our examination of exactly what is and what is not
excluded, in the way of reactive attitudes, by adoption of the objective attitude, should
102 Galen Strawson addresses the question of whether loving some individual need commit us to regarding that individual as a “truly responsible” agent. This question is related to the one I am here raising, but is, so to speak, a prior question. We may think that it is only appropriate to hold a desert-entailing attitude towards some agent, X, if that agent is “truly responsible” for the volitional states, actions, or characteristics to which that desert-entailing attitude is directed. Galen Strawson is drawn to the view that such a belief in “true responsibility” is an important aspect of our feelings of love, even though, on his view, such “true responsibility” is an impossibility. He finds himself, therefore, with the possibility of a kind of cognitive tragedy – our understanding showing us that we simply cannot have what we most want, in terms of reactive attitudes and human relations. I am rather more sanguine about the prospects for love in the absence of “true responsibility”, not least because “love” is the name for a very complex set of emotional reactions, only some of which might be thought to implicate any kind of view about responsibility or desert, and so it is difficult to see how love, as a general kind of emotional reaction, could be inconsistent with the ‘moderate’ objective attitude. (On my view, if we want to search for the reactive attitude that we should most regret losing if we were to adopt the ‘moderate’ objective attitude, then that attitude is not love but gratitude, which, like resentment, makes no sense whatsoever if we are committed to the principled exclusion of desert-entailing attitudes.) For Galen Strawson’s view on love, see his (1986) Freedom and Belief, (Oxford: OUP), chapter 16, ‘Antinomy and Truth’, and, esp. pp. 309-310.
100
lead us towards adumbrating a more carefully nuanced conception of the objective
attitude than one commonly finds described. In contrast to the implausibly ‘Stark’
reading of the objective attitude, which Strawson himself explicitly rejects, we should
instead prefer a ‘Moderate’ conception of the objective attitude, under which only desert-
entailing reactive attitudes are excluded. As should be clear, adoption of this ‘Moderate’
objective attitude would not separate us from meaningful human interaction, and would
not involve us in taking a purely instrumental or technocratic attitude towards our fellow
human beings. Indeed, it would permit a broad range of possible reactive attitudes to
obtain, allowing complex and sophisticated interpersonal relationships to flourish. It
would require us to excise only a narrow and specific range of desert-entailing attitudes –
such as resentment, gratitude, and forgiveness – from our interpersonal attitudinal
‘vocabulary’, leaving plenty of space for anger, pity, fear, repulsion, appreciation,
fellowship, (most forms of) love, solidarity, sympathy, and most of the broad range of
interpersonal reactions and emotions.
What should be clear, therefore, is that a wholesale adoption of the objective attitude
would not commit us to any kind of thoroughgoing psychopathy, or to a lonely alienation
from humanity. Adopting the objective attitude towards a fellow individual does not
preclude the possibility of having a full, nuanced and rewarding interpersonal relationship
with that individual. Moreover, adopting the objective attitude, even towards everyone
and all the time, would not banish us from the domain of personal relationships. We
could, so to speak, embrace the objective attitude without being expelled from the garden
of the truly human. So, whilst on the rejected ‘Stark’ reading, the objective attitude is
101
every bit the caricature of a purely bureaucratic state of (non-)involvement, quite alien
from what it is to be an emotionally engaged human individual, this is certainly not the
case with the preferred ‘Moderate’ reading of that attitude.
Given this result, one must take issue with some puzzling remarks that Strawson makes
about the objective attitude. They are rendered all the more peculiar by the fact that they
come immediately after the point in his discussion where he happily admits that the
objective attitude need not involve full emotional disengagement, and need not exclude
all of the reactive attitudes. Strawson suggests that:
If your attitude towards someone is wholly objective, then though you may fight with him, you cannot quarrel with him, and though you may talk to him, even negotiate with him, you cannot reason with him. You can at most pretend to quarrel, or to reason, with him.103
Strawson’s claim here is that neither ‘quarrelling’ nor ‘reasoning’ with someone is
compatible with adopting the objective attitude towards that person. This claim seems
particularly implausible in the case of reasoning, but is also, I would suggest, just as
implausible in the case of quarrelling. To take reasoning first, it is unclear why one
should think that engaging in a process of reasoning with someone need involve any kind
of desert-entailing emotional involvement. Indeed, we can surely reason with another
individual without having any kind of emotional response or reactive attitude towards
them. “Reasoning with”, in the sense of thinking through a problem, or in the sense of
offering considerations in support of some belief or some course of action, seems like the
paradigm case of a human activity which can be undertaken in the absence of any
103 F & R, p. 79.
102
emotional colouration. We can ‘reason with’ people in our professional lives, in the
absence of any emotional involvement or attitudinal stance; this is something done daily,
one would imagine, by mathematicians and philosophers, as well as by therapists,
counsellors and physicians. We can reason with any other individual that seems capable
of understanding and accepting reasons. The limitations on whom we can reason with
therefore seem to be cognitive rather than attitudinal, by which I mean that the
preconditions for two individuals to ‘reason with’ one another are given by the capacity
of both individuals to understand and give reasons, rather than by any aspect of either
party’s reactive attitudes.
A fictional example perhaps might shed some illumination: one can consider the
interesting character of Mr Spock in Star Trek, who is incapable of emotion (and hence
does not have the usual range of reactive attitudes) but is committed to clear, logical
reasoning. We might, perhaps, think that he strikes us an implausible agent. We may
think that such an individual would be incapable of action, and subject to motivational
paralysis. But, minimally, he would at least seem to be someone with whom we could
reason, certainly at least insofar as we were engaged with him in theoretical rather than
practical reasoning. And, even if Spock was an unlikely agent, what he lacks is surely not
the fact that he does not have desert-entailing emotional attitudes, but simply that he may
have no plausible motivational states at all. There seems no reason to think that Spock
could not, in principle, ‘reason with’ his associates, nor they with him, even though such
a character is committed to the adoption of an ongoing ‘objective attitude’ that is much
closer to the Stark than to the Moderate variant (i.e. Spock’s emotional repertoire is
103
excised of many or all non-desert-entailing reactive attitudes, as well as all the desert-
entailing attitudes).
So, with regard to the matter of our “reasoning with” others, it would seem that Strawson
is overselling the strangeness or unacceptability of the objective attitude. If the objective
attitude precluded us from reasoning with each other, that would certainly be a
consideration in favour of avoiding such an attitude, and might also buttress Strawson’s
view of a thoroughgoing objective attitude as humanly unattainable. But, if the adoption
of the objective attitude does nothing to stop us from reasoning with one another, given
that reasoning with one another does not involve the adoption of any desert-entailing
attitudes, then we can again present the objective attitude as (perhaps surprisingly)
neither alienating, nor paralyzing, nor particularly unwelcome. Moreover, we would have
less reason to think that the objective attitude is unattainable for human beings.
‘Quarrelling with’ may be a trickier case, and its solution may be analogous to that
suggested above for ‘anger’. It may well be that ‘quarrelling’ actually covers a fairly
broad range of kinds of interactions – from simple ‘desert-free’ conflict, expressed
through explicit disagreement, through to a darker and more robust variety of conflictual
interaction that involves the expression of a desert-entailing attitude akin to resentment.
We may quarrel with someone when we simply contend with them over some point of
dispute, and we may do this whilst retaining an objective attitude (in the ‘moderate’
sense), even if our dispute is one which is ‘emotionally coloured’ in one way or another.
On the other hand, quarrelling is frequently something that goes hand-in-hand with the
104
withdrawal of good will (and hence with a desert-entailing attitudinal stance), or with the
acceptance of a teleological desert claim.104 The important point, though, is simply one
of accuracy with regard to what we have to give up when we adopt the objective attitude,
and what we may retain. If the objective attitude rendered fighting, contention and
disputation impossible for us, then this again would be to its disadvantage, and would fit
well with the Strawsonian thought that a broad adoption of the objective attitude would
be both impossible and unwelcome. But we should not think that adoption of the
objective attitude need preclude us from the questionable delights of any of these forms
of human interaction.
By subjecting Strawson’s remarks about the objective attitude to this kind of careful
scrutiny, the contours of the objective attitude should now be much clearer. The key
claim to be made is that adopting this attitude does not involve the unmanageable and
unconscionable loss that Strawson suggests. But it is also worth making the stronger
point, in addition, that, in fact, it is from within the ‘objective attitude’ that many of the
most valuable and significant human relationships take place. It is to this stronger point,
regarding the significant attractions, and easily-overlooked importance, of the objective
view, that I now turn.
D. In Praise of the Objective Attitude: On Expectations, Sympathy and Engagement
As has been shown in the foregoing discussion, the costs of the objective attitude are not
as great as Strawson suggests. But, setting aside the downside of the objective attitude, it
104 For the relevant distinctions among varieties of desert, see Section 2A above.
105
is worth instead turning to its advantages, which are far greater than the Strawsonian
view countenances. Perhaps the most striking example of where significant human
relationships are deeply ‘objective’ (in the moderate sense) is given by the familiar case
of the relationships that obtain between parents and their (young) children. We do not, in
general, take desert-entailing attitudes towards our children, and, when we do, we
generally recognize the inappropriateness of such attitudes. What this means, in effect, is
that parental love and regard is not accompanied, at least within reasonable bounds, by
the latent ‘threat’ of the withdrawal of good will, or of the activation of a (negative)
desert-entailing attitude. There is, of course, an expectation of certain standards of
behaviour from the child, but nevertheless the parent’s good will is, at base,
unconditional.
Children are held to normative expectations, once their general capacities as agents start
to develop, but these are not expectations which are backed up by the threat of some kind
of reactive attitudinal sanction. Yet this preclusion of desert-entailing attitudes does not
mean that parent-child relationships are not rich, complex and rewarding. And it is not
that they have their characteristic qualities in spite of the preclusion of desert-entailing
attitudes; on the contrary, it is the unconditionality of the exclusion of desert-entailing
attitudes that in part makes personal relationships within families such a potential source
of human joy. Of course, it is true that there are many aspects or intra-familial
relationships that make them unsuitable for a model for other kinds of human relationship
– especially with regard to the distinctive kind of denial of autonomy that is involved in
paternalism. But my suggestion is that the impressive qualities which we associate with
106
relationships that have dispensed with the possibility of coming to involve desert-
entailing attitudes are qualities which can obtain just as much in the absence of
paternalism. Forbearance with regard to desert-entailing attitudes is a distinctive, and
attractive, quality of some of the most admirable forms of human interaction.105
We can set aside the particular case of intra-familial relationships in considering other
cases where rich human reactive attitudes can flourish in the absence of (or with the
exclusion of) the desert-entailing attitudes. There is, one might suggest, nothing special
with regard to the desert-entailing attitudes at all – they simply constitute one kind of
response we can have to each other, or one kind of region within the space of emotional
attitudes and responses. If this is so, then one need not think that giving up on this region
within the space of our interpersonal attitudes need involve giving up on anything more.
We can kick away this part of our attitudinal repertoire without the whole structure of our
interpersonal relationships falling to the ground.
Alongside the example of our ‘moderately’ objective attitudes to children (and often
towards other friends and family members), there are many cases in fiction that might be
used to show how the rejection of the desert-entailing reactive attitudes need not involve
us in emotional alienation or disengagement. There is a magnificent moment, for
example, in Graham Greene’s novel The Power and the Glory, in which the (anti-)hero,
an alcoholic priest in 1930s Mexico known throughout only as ‘the Whisky priest’, who
is being hunted down by the anti-clerical authorities during a period of religious
105 There is a fuller discussion of interpersonal and moral attitudes in parent-child relationships in Section 3(J) below.
107
persecution, has a moment of clarity whilst imprisoned in a police cell. It is a powerful
and striking exemplar of what a shift to a full-scale rejection of desert-entailing attitudes
(and thereby an adoption of a ‘moderate’ variety of the objective attitude) might look
like. Moreover, it displays vividly that such an attitudinal shift need not go hand-in-hand
with any kind of emotional disengagement. Indeed, it shows quite the opposite – that
adopting the objective attitude can actually go together with a kind of heightened
emotional connection towards others.
In this passage from Greene’s novel, the Whisky priest has been betrayed to the police by
an informant (the “half-caste” referred to below). He is imprisoned overnight in an
overcrowded cell, in the company of a number of the poor and the wretched of Mexican
society, along with one particular pious middle-aged bourgeois woman, who has been
arrested for keeping religious pictures in her home. The pious woman approaches him to
hear her confession, but they fall into a quarrel about their fellow prisoners, whom the
pious woman despises for their immorality and ‘beastliness’ (they have overheard the
muffled sounds of two of their fellow inmates’ lovemaking from a corner of the cell), but
with whom the Whisky priest sympathises. Outraged by his lack of propriety, and his
resistance to the sort of platitudinous or conventional moral judgements that she might
have expected from a priest, the pious woman threatens to report the Whisky priest to his
bishop.
The priest’s initial reaction to this confrontation is one of contempt and, indeed, of
resentment, towards the pious woman. She engages his hostility, and he finds himself
108
moving towards a robustly desert-entailing attitude towards her. But, as can be seen
below, this hostility soon dissolves in the face of the Whisky priest’s adoption of a
‘moderate’ variety of the objective attitude:
He couldn’t help laughing: she had no sense of how life had changed. He said: ‘If he [the Bishop] gets the letter he’ll be interested – to hear I’m alive.’ But again he became serious. It was more difficult to feel pity for her than for the half-caste who a week ago had tagged him through the forest; but her case might be worse. The others [i.e. the poor and wretched sharing their prison] had so much excuse – poverty and fever and innumerable humiliations. He said, ‘Try not to be angry. Pray for me instead.’ ‘The sooner you are dead the better.’ He couldn’t see her in the darkness, but there were plenty of faces he could remember from the old days which fitted the voice. When you visualized a man or woman carefully, you could always begin to feel pity – that was a quality God’s image carried with it. When you saw the lines at the corners of the eyes, the shape of the mouth, how the hair grew, it was impossible to hate. Hate was just a failure of imagination. He began to feel an overwhelming responsibility for this pious woman. ‘You and Father José,’ she said. ‘It’s people like you who make people mock - at real religion.’ She had, after all, as many excuses as the half-caste. He saw the kind of salon in which she spent her days, with the rocking-chair and the family photographs, meeting no one.106
The turning-point in the priest’s attitude is nicely caught by Greene’s “but again he
became serious”. This seriousness, so to speak, has the effect of alienating him from his
easy, initial move towards a desert-entailing, resentful hostility. How this happens is
captured by the idea that “when you visualized a man or woman carefully, you could
always begin to feel pity” – the thought that close attention to the particularities of
someone else’s situation can at the same time both undermine any desert-entailing
attitude, and provide the basis for feeling sympathy and compassion for them. Greene
106 Graham Greene, (2003) [1940], The Power and the Glory, (London: Penguin), p. 131. In his autobiography, A Sort of Life, (1974), (London: Penguin), Greene describes the prison dialogue in The Power and the Glory, of which the above passage forms part, as one of the few passages in his novels that gave him a sense of satisfaction. John Updike, in his ‘Introduction’ to The Power and the Glory, contends that the prison scene in that book “in which the priest, at the nadir of his abasement and peril, sits up all night in a dark crowded dark cell listening to the varied voices – the disembodied souls – of the other inmates is, in its depth, directness and strange comedy, worthy of Dostoevsky, another problematical believer.” (Ibid, p. viii)
109
paints this as an essentially religious attitude (“that was a quality God’s image carried
with it”), but it seems to me that it need not be so. All that this attitudinal shift need
involve is an attentional shift from the particular attitude that is being evinced by another,
to the peculiarities of background, history or personality that might explain that current
attitude. This absolutely involves taking a more ‘objective’ attitude towards another
individual, in a number of senses: firstly, it involves suspension of the current desert-
entailing attitude; secondly, it is more ‘objective’ in that it involves shifting from a prima
facie emotional response to a focus on the peculiarities of circumstance and
psychological constitution of the person to whom one is reacting; thirdly, and relatedly,
this shift in a sense involves seeing the person to whom one is reacting more as a part of
the natural world of causes and effects, and less as an inscrutable source of presently
existing attitudes. Yet none of this is to suggest that taking the “objective attitude” here is
in any way distancing or alienating.
Now, Strawson himself allows that the objective attitude can remain open to us through
something like an act of will. He allows that it need not always be generated simply by
seeing that the Capacity or Quality of Will criteria for suspending the desert-entailing
reactive attitudes has been met. But Strawson does not seem to see how, as in the Whisky
priest’s case, a move to the objective attitude can be both (i) motivated by a general
commitment to view others in a particular way; and (ii) can involve emotional
engagement with other people, rather than instrumentalization or avoidance. As Strawson
puts it:
110
The objective attitude is not only something we naturally tend to fall into in cases like these, where participant attitudes are partially or wholly inhibited by abnormalities or by immaturity. It is also something that which is available as a resource in other cases too. We look with an objective eye on the compulsive behaviour of the neurotic or the tiresome behaviour of a very young child, thinking in terms of treatment or training. But we can sometimes look with something like the same eye on the behaviour of the normal and the mature. We have this resource and can sometimes use it: as a refuge, say, from the strains of involvement; or as an aid to policy; or simply out of intellectual curiosity.107
What the Whisky priest’s move from easy resentment to a more serious (and ‘objective’)
compassionate engagement shows is that the objective attitude can be seen in a very
different light to its presentation here by Strawson. Significantly, it need not function as a
refuge “from the strains of involvement”, and can, indeed, be exactly the opposite. The
Whisky priest’s initial contempt and resentment are, so to speak, a very easy and costless
kind of involvement – they allow him to keep the pious woman at arm’s length, and to
free himself of any sense of connection with her. It is his adoption of a moderate
‘objective attitude’, together with use of his powers of intelligence and sympathetic
imagination in coming more accurately to see what the pious woman’s life was like, that
actually brings him into emotional engagement with her. As Greene has it: “Hate was just
a failure of imagination. He began to feel an overwhelming responsibility for this pious
woman.”108 If the Whisky priest was driven by a concern to find an easy refuge from the
strains of involvement, then his optimal strategy would be to leave his initial reactive
attitudes unexamined. But the adoption of the objective attitude makes this refuge
unavailable: it shows this refuge for what it is – a sensuous collapse into easy emotion, at
the expense of integrating one’s understanding and imagination into a more considered,
107 F & R, pp. 79-80.
108 See Greene, (2003), ibid. p. 131.
111
fair and sophisticated emotional reaction. Perhaps this does not fit very well with
Strawson’s overly schematic picture of the reactive attitudes, but the Whisky priest’s
move from hostility to a compassionate sense of “responsibility for” the pious woman
shows that adoption of the objective attitude will often involve the refinement rather than
the rejection of our interpersonal reactive attitudes.
The more nuanced picture that can be extracted from considering situations such as that
described above by Greene should lead us to be wary of the way in which Strawson
wants to draw a clear distinction between ‘objective’ and ‘participant’ attitudes. For
example, Strawson claims that:
But what is above all interesting is the tension there is, in us, between the participant attitude and the objective attitude. One is tempted to say: between our humanity and our intelligence. But to say this would be to distort both notions.
Strawson simply goes wrong in positing these two attitudes – the ‘objective’ and the
‘participant’ – as the stable pillars between which we are stretched in tension. Given that
the objective attitude need not be emotion-free, it is not clear that he has marked out two
distinctive attitudes at all. In fact, Strawson’s participant/objective distinction looks very
much like a false dichotomy, as there is nothing in the objective attitude, understood
‘moderately’, that prohibits participation.
If we understand the objective attitude in its ‘moderate’ rather than its ‘stark’
manifestation, there need be no such tension between objectivity of attitude and the
“participant attitude”, unless by “the participant attitude” we mean to characterize only
112
those kinds of rough, defeasible and unelaborated desert-entailing emotional reactions
that may assail us before we have considered the true structure of some interpersonal
situation. But our “humanity” certainly does not consist in prizing these sorts of
emotionally blunt reactions at the price of thoughtful, considered and robust interpersonal
engagement. So, whilst Strawson realizes that it would be somewhat distorting to
characterize our humanity, as such, as being in tension with our intelligence (where our
intelligence might lead us to the objective attitude), he does not realize quite what a
comprehensive distortion this kind of picture really involves.
E. Taking Stock: The Varieties of Reactive and Objective Attitudes
It is hopefully now clear that the conceptual terrain in this domain is rather more complex
than the Strawsonian picture might initially have suggested. Most importantly, before
moving on to consider the main strands of Strawson’s argument for his form of
‘naturalistic’ or ‘practical’ compatibilism, these two results of the foregoing discussion
should be borne in mind:
(a) the ‘reactive attitudes’ can be understood in different ways, ‘broadly’ and
‘narrowly’, and that only on the ‘narrow’ reading need they be desert-
entailing; and
(b) accordingly, there are ‘stark’ and ‘moderate’ ways of understanding the
‘objective attitude’, and the rejection of desert-entailing reactive attitudes
commits us only to a ‘moderate’ objective attitude, which is itself consistent
113
with (and may even encourage) deeply engaged forms of emotional inter-
relation.
With this in mind, I want now to return to the question of the excuses and exemptions, by
virtue of which we properly come to suspend our reactive attitudes, and of the account in
terms of ‘Capacity’ and ‘Quality of Will’ which Strawson gives of the excuses and
exemptions. As we have seen above, there are different ways of reading Strawson’s
account – a ‘strong’ reading whereby the ‘Capacity’ and ‘Quality of Will’ criteria
together give necessary and sufficient conditions for when it might be appropriate to
suspend a reactive attitude, and a ‘weak’ reading whereby they give us only sufficient but
not necessary conditions for the appropriateness of our suspending a particular attitude.
As we saw in Section 2(b), on a ‘Strong’ reading of Strawson’s view, the full set of
reactive attitudes are always fully appropriate when the ‘quality of will’ and ‘capacity’
criteria are met. On a ‘Weak’ reading, there may be other reasons for considering such
attitudes inappropriate, even when both criteria are met. On a ‘strong’ reading, it is
impossible for any general, theoretical or philosophical considerations to give us
independent reasons for revising or curtailing any of our reactive attitudes. On a ‘weak’
reading, we can grant the insight of Strawson’s account in terms of ‘quality of will’ and
‘capacity’, whilst granting that this may not be the full story about when the reactive
attitudes are appropriate and when they are not. It is my contention that Strawson slides
from a ‘weak’ to a ‘strong’ version of his view due, at least in part, to a failure to keep
114
apart the ‘stark’ and ‘moderate’ versions of the objective attitude; and my aim is to
substantiate this charge in sections 2(G) to 2 (J) below.
Put simply, we should be careful to note that, on the one hand, we may have reasons for
rejecting the ‘narrow’ reactive attitudes that are not reasons for rejecting the full, broad
set of reactive attitudes in toto. Moreover, some of the reasons we may have for not
rejecting the full, broad set of reactive attitudes do not speak against the rejection of a
subset of the reactive attitudes, such as the ‘narrow’, desert-entailing reactive attitudes.
Hence, even if the ‘strong’ version of the Strawsonian position may appear to have at
least a prima facie plausibility with regard to the reactive attitudes considered broadly, it
may lose ground to the ‘weakened’ version when we see how the domain of the reactive
attitudes may be subdivided through attending to the ‘narrow’ set of desert-entailing
reactive attitudes.
At any rate, it should be clear that different standards should be appropriate for the
adoption of the objective attitude, depending on whether we are operating with the stark
or the moderate version. For the purposes of the current discussion, it is the ‘moderate’
version of the objective attitude which should be the primary object of our attention, as it
is this version of the objective attitude which provides the most interesting test case for
the ‘strong’ version of Strawson’s view – that is, for the claim that no general, theoretical
or philosophical view could or should lead us towards the objective attitude.
115
F. The Reactive Attitudes and the Varieties of Incompatibilism
With the foregoing background in place, and with the conceptual territory mapped out
with greater clarity, we can now turn to the central question around which the entire
argument of “Freedom and Resentment” turns. At base, this question asks whether we
should accept a ‘Strong’ reading of Strawson’s account of when the reactive attitudes are
appropriate and when they are not; that is, whether it really is the case that general
philosophical considerations should have no bearing on the appropriateness of our
reactive attitudes. Strawson asks:
What effect would, or should, the acceptance of the truth of a general thesis of determinism have upon these reactive attitudes? More specifically, would, or should, the acceptance of the truth of the thesis lead to the decay or the repudiation of all such attitudes? Would, or should, it mean the end of gratitude, resentment, and forgiveness; of all reciprocated adult loves; of all the essentially personal antagonisms?109
It is no surprise, given Strawson’s own ‘Strong’ reading of his account of the excuses and
exemptions, that his own answer is a negative one. He holds that the full set of reactive
attitudes are legitimately in place whenever the Quality of Will and Capacity criteria are
met. But is Strawson’s answer to this question the right answer? Or, adopting a ‘Weak’
reading of the Strawsonian account of the excuses and exemptions, should we instead
allow that the acceptance of some distinctively philosophical view (for example, assent to
the truth of a thesis of determinism) could also serve to undermine the legitimacy of (at
least some of ) the reactive attitudes? This is the central question in assessing the success
or failure of the Strawsonian project.
109 F & R, p. 80.
116
Before answering Strawson’s question, one should notice here that Strawson is not asking
the question that has most commonly been contested by compatibilists and
incompatibilists over the course of debates about determinism and freedom of the will.
He does not ask whether freedom of the will is compatible with determinism, and neither
does he ask whether the use of moral terms, or the formation of moral judgements, are
consistent with determinism. Instead, Strawson asks simply whether acceptance of the
truth of determinism ought to (or even could) lead us to suspend our reactive attitudes.
This is not, as yet, a question about moral ideas at all, although, in the development of
Strawson’s argument, this question will come to be linked to moral ideas, and to the more
common questions about the compatibilism of free will and determinism.110
One should also notice here that, by the lights of the distinctions introduced above,
Strawson may be running together some separate questions. Firstly, we see that Strawson
seems to elide questions of what effect the acceptance of a thesis of determinism would
have with questions of what effect it should have. This elision is a striking symptom of
Strawson’s naturalism, which will be given more attention in what follows, although we
can set this aside for now. More significant, for our present purposes, is the issue of
whether Strawson’s question should be seen to apply to the reactive attitudes taken as a
broad or a narrow set. The invocation of the familiar triad of resentment, gratitude and
forgiveness in the formulation of the question suggests that it should be read as limited
specifically to the narrow set of desert-entailing attitudes; but, on the other hand,
Strawson’s reference to “these attitudes” is to “what I have called the participant reactive
110 This linking move in Strawson’s argument will be discussed fully in Section 3, below.
117
attitudes” that are “essentially natural human reactions to the good or ill will of others, as
displayed in their attitudes and actions”,111 and this suggests that he is referring to the
reactive attitudes understood broadly.
What should be emphasized at this stage is that, depending on whether we take the
question at hand to refer to the broad or narrow reactive attitudes, we find ourselves in
two very different dialectical positions. To elaborate, we should consider the differences
between, on the one hand, philosophical views which hold that acceptance of the truth of
determinism would undermine the whole set of reactive attitudes, and, on the other hand,
those views which hold that the challenge is only to the narrow set. Depending on which
of these two kinds of view are in question, the position against which Strawson’s own
view should be contrasted will be of a different kind. Consider the difference between
these two families of views, both of which could be used as the basis for a claim that we
should reject (at least some of) the reactive attitudes on the basis of general, theoretical
philosophical considerations:
Type-1. Incompatibilism about Agency, as such
Views of this kind hold that the possibility of an individual genuinely being an
agent, as opposed to being a passive conduit for the play of natural forces,
depends on the falsity of the thesis of determinism (or, perhaps, the falsity of
111 F & R, 80.
118
some other hypothesis about the physically or temporally instantiated nature of
human agency).112
There are also related views – which we may describe as ‘Scepticism about
Agency, as such’ – that hold that humans can never be agents, whether or not
determinism is true.
Type-2. Incompatibilism about the Grounds for Claims of Desert
Views of this kind hold that, if determinism is true, then individuals cannot, by
virtue of how they act, legitimately deserve to suffer some harm or bear some
good. Typically, this is because such views hold that the kind of agency needed
in order for individuals to deserve some harm or some good by virtue of their
actions is itself impossible if determinism is true. Notice that such views need
not claim that agency, as such, is necessarily incompatible with determinism.
On such a view, one might think that some weaker form of agency was
compatible with determinism, but that a more robust form of agency, of a kind
that would be needed in order for claims of desert to be justifiable, was not
compatible with determinism. Therefore, one can accept a view of this type
(type-2) whilst rejecting the wholesale incompatibilism of the first type of view
of (type-1).
112 Kant was perhaps the most illustrious proponent of such a view. See inter alia Immanuel Kant, Part I, “Concerning the indwelling of the evil principle alongside the good or Of he radical evil in human nature” of the Religion within the Boundaries of Mere Reason (Ak. 6:19 – 6: 53); the “Critical Elucidation of the Analytic of Pure Practical Reason” in The Critique of Practical Reason (Ak. 5: 89 – 5: 106); Section III, “Transition from metaphysics of morals to the critique of pure practical reason” of The Groundwork of the Metaphysics of Morals (Ak. 4: 447 – 4: 463), and the “Third Conflict of the Transcendental Ideas” of “The Antinomy of Pure Reason” in The Critique of Pure Reason (A444/B472-A451/B479).
119
There are related views – which we may describe as ‘Scepticism about the
grounds for desert’ – that hold that humans can never deserve some benefit or
burden by virtue of how they have acted, whether or not determinism is true.
It important to stress that one could be a compatibilist with regard to agency, as such (and
so reject views of type-1), whilst nevertheless being an incompatibilist (or sceptic) about
the kind of agency that could legitimate ‘desert-entailing’ reactive attitudes or
judgements (and so accept a view of type-2). The obverse view, of course, cannot
plausibly be held: – if one is a sceptic or incompatibilists about the possibility of agency,
as such, then one surely must also be a sceptic or incompatibilist about the kind of agency
that could underlie or justify claims of desert. I shall name a view that rejects
incompatibilism about agency, as such, but which embraces incompatibilism (or
scepticism) about ‘desert-entailing’ agency (by which I mean agency of a kind that could
ground legitimate claims of desert), the ‘Hybrid View’. The ‘Hybrid View’ stands in
contrast to what I shall describe as the ‘Standard Incompatibilist View’ (i.e. views of
type-1), which further also embraces the position of full incompatibilism about agency, as
such. The ‘Hybrid View’ is, therefore, a restricted (type-2) form of incompatibilism, that
rejects the standard, unrestricted incompatibilism conclusions of type-1 views.113
113 Part Three of this dissertation is largely devoted to an argument for why one should accept a ‘Hybrid View’ or Type-2 form of incompatibilism. At this stage, I want only to make clear that such a view is a possibility; its elaboration and defence will come later.
120
It makes a great deal of difference to Strawson’s dialectical position whether he is
arguing against a Standard Incompatibilist or against a proponent of some version of the
Hybrid View. Proponents of a Standard Incompatibilist view would answer Strawson’s
question by saying that acceptance of the truth of determinism should lead to a
repudiation of the full set of reactive attitudes, on a broad construal of those reactive
attitudes. They would hence reject the truth of the ‘Strong’ reading of Strawson’s account
of the exemptions and excuses, for the full set of reactive attitudes.
On the other hand, proponents of some variety of the Hybrid View can deny that the
broad set of reactive attitudes should be repudiated if one accepts the truth of some thesis
of determinism, given that those attitudes are nevertheless accurate when they respond to
the actions of another individual, subject to the Capacity and Quality of Will constraints.
Hence, the Hybrid View is compatible with an acceptance of the truth of the Strong
reading of Strawson’s account of the exemptions and excuses, outside of the domain of
the (narrow) desert-entailing reactive attitudes. But the proponent of the Hybrid View
must nevertheless hold the view that the truth of a thesis of determinism should lead to
the repudiation of any reactive attitude which is itself desert-entailing, insofar as such a
view is incompatibilist (or sceptical) about the aspects of individual agency that might
ground a justifiable claim of desert. Thus, the Hybrid View claims that acceptance of the
truth of determinism should lead to a repudiation of the narrow set of reactive attitudes
(that is, those reactive attitudes that are themselves desert-entailing), for distinctively
philosophical reasons, but not to the repudiation of the broad set of all reactive attitudes.
Given this, the Hybrid View entails the repudiation of the Strong reading of Strawson’s
121
account of the exemptions and excuses, for the narrow reactive attitudes, whilst
nevertheless allowing that the Strong reading may be appropriate outside of this narrow
domain, with regard to the broader set of the reactive attitudes.114
With this in mind, we can now return to Strawson’s question. On the Standard
Incompatibilist view, we should, indeed, reject all reactive attitudes – they are, after all,
reactions to the putative agency of others, and they are rendered senseless if nobody
really can be an agent at all. But on the Hybrid View, we should reject only the reactive
attitudes of the narrow, desert-entailing kind. So, where Strawson asks whether
acceptance of the truth of determinism should “mean the end of gratitude, resentment and
forgiveness”,115 the proponent of the Hybrid View (alongside the Standard
Incompatibilist) answers ‘yes’. But, when Strawson goes on to ask whether this should
also mean the end of “reciprocated adult loves [and] of all the essentially personal
antagonisms”,116 the proponent of the Hybrid View answers ‘no’, thereby parting
company from the proponent of Standard (or, one might say, ‘crude’) versions of
Incompatibilism.
114 I have here merely introduced the Hybrid View, without attempting to motivate it fully, or to argue for its acceptance. This is quite deliberate. As mentioned in the previous note, my concern here is simply to identify the possibility of such a position within the available conceptual space, and to suggest that Strawson’s dialectical position against such a possible view is especially revealing as regards the possible shortcomings of the Strawson’s overall argumentative strategy. A fuller examination of the attractions of, and reasons for accepting, the Hybrid View is pursued in Part Three.
115 F & R, p. 80.
116 F & R, p. 80.
122
If it is true, as I suspect that it is, that there need be nothing ‘desert-entailing’ about
reciprocated adult love, then there is no reason for it to be imperilled, on the Hybrid
View, by acceptance of the truth of determinism. Moreover, personal antagonism can
surely legitimately exist whenever there is genuine opposition or contention between one
individual and another. Opposition or contention counts as personal whenever it exists as
a reaction to features of the personality or ‘quality of will’ of the individual with whom
one finds oneself in an antagonistic relation. It is only on the crudest (or, at any rate, the
starkest) forms of incompatibilism that we need think that determinism rules out the
possibility of personality or of willing.117 On the Hybrid View, antagonisms can be as
personal as one would like, insofar as they are reactions of opposition or contention
fuelled by, and focused upon, the will and personality of others, and need not remotely be
imperilled by the truth of determinism. Personal antagonisms would fall foul of the
Hybrid View only where those antagonisms shaded over into stronger, desert-entailing
attitudes of antagonistic resentment. But this would be a stronger, modulated version of
personal antagonism: it is no necessary part of personal antagonism, per se, that it
subsume a related reaction of interpersonal resentment.
The foregoing clarifications and distinctions help to illustrate the important point that
Strawson’s most interesting philosophical opponent is actually the proponent of the
Hybrid View, rather than the Standard (or ‘crude’) Incompatibilist. This is because the
Hybrid View is, as it were, a more modest view than is Standard Incompatibilism.
Strawson’s Incompatibilist opponent would need to show that all reactive attitudes
117 For a convincing and exhaustive argument that willing is compatible with determinism, see Brian O’Shaughnessy, (1980), The Will: A Dual Aspect Theory, (Cambridge: Cambridge University Press).
123
(considered broadly) are illegitimate, whereas the proponent of the Hybrid View need
argue only that the narrow, desert-entailing attitudes are under threat, whilst registering a
verdict of ‘no contest’ with regard to Strawson’s claims that the entire set of reactive
attitudes could not and should not be threatened by acceptance of a thesis of determinism.
The Standard Incompatibilist needs to reject the ‘Strong’ version of Strawson’s account
of the exemptions and excuses within the entire domain of the reactive attitudes. The
advocate of the Hybrid View need reject the ‘Strong’ version of Strawson’s view only
within the narrowly delimited domain of the desert-entailing reactive attitudes. And yet it
is sufficient to reject the Strong version of Strawson’s view if it can be shown to fail
within this more narrowly circumscribed domain. It is more difficult for Strawson,
therefore, to argue against the Hybrid View than against Standard Incompatibilism.
So, although Strawson’s use of ‘reactive attitudes’ is often very broad indeed, the most
interesting philosophical opposition to Strawson’s view involves a narrower reading of
the term. From the standpoint of the Hybrid View, the pertinent philosophical question is
whether Strawson’s strategy of a wholly general rescue of the reactive attitudes from
philosophical worries (involving the ‘Strong’ reading of Strawson’s position on the
exemptions and excuses, as described in 2(b) above) can work. To show its failure, one
would need only to show that the narrow desert-entailing reactive attitudes could not be
saved from sceptical or incompatibilist worries. In so doing, one need not show that a
wholesale adoption of a Stark objective attitude is possible, desirable, or philosophically
unavoidable, but can instead limit one’s examination to the question of the adoption of a
Moderate objective attitude. From the argument of Section 2(c), above, it should be clear
124
that the defence of the Moderate objective attitude will be a much easier task than the
defence of the Stark objective attitude.
Defending Strawson’s view becomes much more difficult if his attempt at insulating the
reactive attitudes from philosophical attack is seen (as it properly should be) to need to
succeed against ‘Hybrid’ or partial versions of incompatibilism, as well as against the
simpler, across-the-board versions. Thus, in what follows, the success of Strawson’s
argument will be judged against this more demanding standard. Only if Strawson
manages to argue successfully against the partial or restricted form of incompatibilism
involved in the Hybrid View, alongside its advocacy of a limited, ‘Moderate’ objective
attitude, can his overall argumentative strategy be held to have succeeded. Having lodged
this methodological decision, along with the interesting shift in the burden of
philosophical proof which it brings to the dialectic, we can now turn to Strawson’s
considered, central question:
We can sometimes and in part, I have remarked, look on the normal (those we rate as ‘normal’) in the objective way [i.e. with the objective attitude] in which we have learned to look on certain classified cases of abnormality. And our question reduces to this: could, or should, the acceptance of the determinist thesis lead us always to look on everyone exclusively in this way?118
Strawson’s answer to this question is a resounding ‘no’. He denies absolutely that the
acceptance of a thesis of determinism (or, for that matter, any general philosophical
position) should (or even could) lead us to a wholesale adoption of the Objective View.
Strawson’s presentation of his view that no general revisionist philosophical theory can
118 F & R, p. 81.
125
lead us to reject the legitimate applicability of the reactive attitudes can be captured in
two lines of argument, which we might call the ‘Appeal to the Inescapability of the
“Human Commitment”’ and the ‘Appeal to the “Gains and Losses to Human Life”’. In
the ‘Inescapability of the “Human Commitment”’ line of argument, Strawson denies that
we could adopt the objective attitude in the general case; in the ‘Gains and Losses to
Human Life’ argument he denies that we should generally adopt the objective attitude,
even if (as he denies) that were possible for us. The ‘Inescapability of the “Human
Commitment”’ argument represents Strawson’s ‘official’ position, and his first line of
defence; the ‘Gains and Losses to Human Life’ argument is a ‘fallback’ position, which
is strictly unnecessary by the lights of Strawson’s own view, but which is nevertheless
extremely revealing about the way in which he understands the issues that are here at
stake.
Strawson’s argument for this position, using these twin tracks to defuse philosophical
views that would call into question the appropriateness or legitimacy of the reactive
attitudes, is the core of his distinctive “naturalistic compatibilism”. I shall sketch
Strawson’s argumentative strategy below, and give reasons for why it should be
considered unsuccessful with regard to its attempt at heading off the limited or restricted
incompatibilist attack of the Hybrid View. Moreover, it will be seen that Strawson’s view
does not generate satisfactory strategies for circumventing the force the reasons that we
may have, if we accept a restricted incompatibilist or Hybrid view, for rejecting the
narrow reactive attitudes, and for adopting the Moderate objective attitude.
126
G. The Objective Attitude and the Inescapability of ‘the Human Commitment’
Strawson’s strategy, with regard to the protection of the reactive attitudes from
philosophical rejection (whether incompatibilist or sceptical with regard to the grounds
for those attitudes) is to follow a strategy of stressing the ‘inescapability of the human
commitment’ to those reactive attitudes. Strawson admits that the objective attitude is
generally accessible for us although, on his view, this move to the objective attitude is
always “as a refuge” or simply out of intellectual curiosity, rather than being capable of
motivation by principled, systematic philosophical reasons. Hence, he allows that it is at
least logically possible that we might adopt such an attitude all the time. But Strawson
nevertheless questions whether the wholesale adoption of the objective attitude would
really be an open possibility for normal human beings. As he puts it:
It does not seem to be self-contradictory that this [i.e. the adoption of the Objective Attitude, such that we “always… look on everyone exclusively in this way”] might happen. So I suppose we must say that it is not absolutely inconceivable that it should happen. But I am strongly inclined to think that it is, for us as we are, practically inconceivable. The human commitment to participation in ordinary inter-personal relationships is, I think, too thoroughgoing and deeply rooted for us to take seriously the thought that a general theoretical framework might so change our world that, in it, there were no longer such things as ordinary inter-personal relationships as we normally understand them; and being involved in inter-personal relationships as we normally understand them precisely is being exposed to the range of reactive attitudes and feelings that is in question.119
Now, we can grant Strawson the truth of most of what he says here, with regard to the
strength and depth of the human commitment to interpersonal relationships, whilst
nevertheless denying him the plausibility of the conclusion that he would draw from
these claims. The key to seeing where Strawson goes wrong here is to recall the
119 F & R, p. 81. [My italics.]
127
distinctions between the broad and narrow reactive attitudes, and between the moderate
and stark objective attitudes. Let us grant that there is a human commitment to
participation in inter-personal relationships, that is thoroughgoing and deep-rooted. But
this ‘human commitment’ could survive the adoption of the objective attitude in its
‘moderate’ variant, as has been argued in Section 2(c) above. We do not condemn
ourselves to isolation from the domain of human relationships by adopting the objective
attitude, insofar as the attitude that one thereby adopts rules-out only desert-entailing
reactive attitudes.
Similarly, we can grant to Strawson the intriguing and insightful suggestion that being
involved in inter-personal relationships just is being exposed to a full range of reactive
attitudes and feelings. But the reactive attitudes and feelings in question need not,
specifically, be the narrow reactive attitudes, but instead are the full, wide set of attitudes
which infuse and perhaps even constitute human relationships. It would be implausible to
claim that we were no longer properly said to be involved in human relationships just
because we excised some particular minor corner of our repertoire of attitudes (such as
the narrow reactive attitudes) from our overall emotional and attitudinal life. We stand in
rich relationships to one another to the degree that those relationships are structured by a
rich multiplicity of inter-related reactive attitudes, and there seems to be no reason to
think that any particular attitude (or small set of attitudes) has any special status in being
a sine qua non for the existence of bona fide human relationships. In short, we can accept
Strawson’s claim about “the human commitment”, whilst nevertheless denying that the
human commitment need rule out the rejection of a subset of reactive attitudes (such as
128
the narrow reactive attitudes), or rule out the thoroughgoing adoption of the moderate
variant of the objective view. Thus, we can retain “the human commitment” whilst
nevertheless simultaneously fully accepting the Hybrid View, and fully acting upon it
through the ‘theoretically-driven’ (or, more precisely, ‘philosophically-driven’)
banishment of the desert-entailing attitudes from what we might call our ‘reactive
vocabulary’.
As his argument continues, Strawson’s strategy is to make use of this putative “human
commitment” as a kind of (anti-)philosophical inoculation against the bacilli of
incompatibilism or of scepticism about moral responsibility:
This, then, is part of the reply to our question. A sustained objectivity of inter-personal attitude, and the human isolation which this would entail, does not seem to be something of which human beings would be capable, even if some general truth were a theoretical ground for it.120
Now, this line of argument is a questionable one, and one may want to take issue with
Strawson’s focus on what we are capable of doing, rather than upon what we really ought
to be doing. But, for the sake of argument, we can happily here grant Strawson’s claim.
What we should notice, though, is that Strawson’s claim, especially insofar as it relates to
the prospect of “human isolation”, works only with regard to the starkest reading of the
nature of the objective attitude. Adoption of the moderate objective attitude involves no
such human isolation, and hence we have no good reason to think that it need be
something of which most human beings would be constitutionally incapable – or, if this
120 F & R, p. 81.
129
is a different idea, of which they would be incapable by virtue of their “human
commitment” to interpersonal relationships.121
Strawson is led into a disappointingly schematic approach to these issues by his failure to
see that there might be different gradations of the ‘objective attitude’. This is shown by
some remarks that come slightly later in his discussion. It will be useful at this point to
dispel the particular misunderstanding of the objective attitude (at least in its moderate
variant) which these remarks embody:
For reasons of policy or self-protection we may have occasion, perhaps temporary, to adopt a fundamentally similar attitude [i.e. the ‘objective attitude’, as held towards the ‘abnormal’] to a ‘normal’ human being; to concentrate, that is, on ‘how he works’, with a view to determining our policy accordingly or to finding in that very understanding a relief from the strains of involvement.122
As the passage from Greene’s The Power and the Glory beautifully (see section 2(C)
above) illustrates, the move to the objective attitude need not be motivated by either
‘policy’ or ‘self-protection’. But, more significantly, it is important to note that adoption
of the objective attitude need not involve a shift, as Strawson seems here to suggest, from
the realm of reasons to the realm of causes. It is certainly true, of course, that paying
careful attention to the real situation of the person to whom one is reacting is likely to
provide an understanding of their predicament that crowds-out, or defuses, the cruder
kinds of desert-entailing reactive attitude. That is certainly one of the features of the
121 Perhaps this strategy is too concessive towards Strawson’s position. Many people over human history have either survived isolation from human interaction (prisoners, explorers, mountaineers, shipwrecked “Robinson Crusoes”, etc.) or positively embraced it (orders of silent monks, religious hermits, etc.) It would be news to a Carthusian monk that withdrawal from ordinary human relationships was either beyond human capacities or too terrible a prospect to contemplate.
122 F & R, p. 82.
130
emotional arc traced by the Whisky priest in his jail cell. But the adoption of a moderate
objective attitude need not involve viewing another individual simply as a nexus of
cause-and-effect, or as somehow less than a reasoning, autonomous agent. To withdraw
some of the desert-entailing reactive attitudes is not thereby to view an individual as a
mere causal being, rather than a rational or moral one. Thus, Strawson is jumping over
some significant conceptual gaps here. On the Hybrid View, we may of course still see
our fellow humans as autonomous agents who act for reasons, and whose actions (and
not their mere behaviour) can be the grounds for our (non-desert-entailing) reactive
attitudes. It is just that we do not believe that the narrow, desert-entailing reactive
attitudes would be appropriate with regards to such agents.
Strawson’s appeal to the ‘inescapability of the human commitment’ in effect has two
parts. There is the ‘negative’ part, which we have just examined, whereby he suggests
that we are incapable of rejecting our reactive attitudes. As we have seen, this strategy is
unsuccessful, as it could work only by virtue of an elision of the important differences
between the broad and narrow reactive attitudes, and between the stark and moderate
objective attitudes. But there is also, as we shall see, the ‘positive’ part, which is, so to
speak, a kind of appeal to the authority of our current practice with regard to exemptions
and excuses. Strawson here claims to diagnose a deep commitment to a way of dealing
with exemptions and excuses which instantiates, in our unquestioned and deep-seated
behaviour, the standards associated with a ‘Strong’ reading of his account in terms of
‘capacity’ and ‘quality of will’ (see section 2(b) above). This ‘positive’ account, in terms
131
of the structure of current practice, provides a second part to Strawson’s appeal to “the
human commitment”. As he continues:
Neither in the case of the normal, then, nor in the case of the abnormal, is it true that, when we adopt an objective attitude, we do so because we hold such a belief. So my answer has two parts. The first is that we cannot, as we are, seriously envisage ourselves adopting a thoroughgoing objectivity of attitude to others as a result of theoretical conviction of the truth of determinism; and the second is that when we do, in fact, adopt such an attitude in a particular case, our doing so is not the consequence of a theoretical conviction which might be expressed as “Determinism in this case”, but is a consequence of our abandoning, for different reasons in different cases, the ordinary inter-personal attitudes.123
We have seen adequate reasons above for why it is that the ‘negative’ aspect of this
strategy is unsuccessful. But what of the positive side of the strategy? Is it really the case
that the adoption of an objective attitude, which precludes desert-entailing reactive
attitudes, is always due to particular, specific excuses and exemptions, rather than being
generated by overarching theoretical or philosophical considerations? It is particularly
striking that Strawson gives us no argument for this claim, but merely presents a bald,
empirical generalization (i.e. “… our doing so is not the consequence of a theoretical
conviction…”). Thus, if one were to say, as against this ‘positive’ move of Strawson’s,
that one could envisage adopting an attitude of thoroughgoing (though moderate)
objectivity, it is not at all clear what Strawson can say in response. Indeed, if one points
out that certain Buddhist (and indeed, Christian) religious thinkers have claimed to have
attained such an attitude on the basis of reasons of a general kind, then it would appear
(insofar as we can regard these thinkers as reliable) that Strawson’s merely empirical
claim here faces a straightforwardly empirical refutation.
123 F & R, p. 82.
132
Perhaps Strawson would have us believe that the Dalai Lama actually resents the Chinese
for the invasion of Tibet but, in the absence of a good argument for why this must be the
case, it would seem rather unfair to think so, in the face of his assertion that he does not
allow himself to hate, or blame, or resent, on the basis of his entirely general view about
the appropriateness of compassion with regard to all living things, and his scepticism
about the existence of the kind of ‘selves’ that could be appropriate targets for blame.124
Now, if it really were the case that only certain Christian or Buddhist saints could achieve
such an objectivity of attitude on the basis of ‘theoretical’ considerations, this might still
be a significant result – for it could be that Strawson’s claim could still be true with
regard to the rest of us.
However, it seems that, if it really is the case that certain individuals can attain this sort
of ‘objectivity’ of attitude, on the basis of a general view, rather than as the result of a
particular excuse or exemption, and at all times, this does at least seem to suggest that
we might expect that the rest of us could attain such objectivity some of the time, insofar
as we assented to a theoretical or philosophical view that held such a suspension of
reactive attitudes to be appropriate or mandated. Even this would seem to be sufficient to
counter Strawson’s claim, as his claim is the extremely strong one that we can never
attain such ‘objectivity’ on the basis of general or theoretical considerations.
124 For the Dalai Lama’s claim to feel no ill will towards the Chinese, see The Dalai Lama of Tibet, (1998), Freedom in Exile, new edition, (New York: Abacus); for the Buddhist theory of “no-self”, see Steven Collins, (1982), Selfless Persons: Image and Thought in Theravāda Buddhism, (Cambridge: Cambridge University Press).
133
The main problem here, at base, is that Strawson’s ‘positive’ strategy is oddly question-
begging. What Strawson really needs to show is that there is no general theoretical
conviction that ought to lead us to the objective attitude. He does not do this, as his
concern here is with how we do act, rather than with how we should act. But, even on this
‘naturalistic’ territory, relating to what we actually do, Strawson offers no argument for
why we can, as a matter of fact, only suspend reactive attitudes for the ‘particular’
reasons captured by his account of the exemptions and excuses, rather than for general
philosophical reasons. He simply asserts that this is the case. But, what Strawson should
here be defending is the legitimacy of a ‘strong’ reading of his (undoubtedly very
insightful) account of the exemptions and excuses in terms of Quality of Will and
Capacity, as opposed to a ‘weak’ reading which took that account to give only sufficient,
but not necessary, conditions for the suspension of the (desert-entailing) narrow reactive
attitudes. No such argument is forthcoming; instead, Strawson simply allows himself the
larcenous assumption of the truth of the view which he needs to defend.
In order to destabilize Strawson’s positive strategy, one need only think of the different
kinds of general or theoretical accounts that individuals have found to justify a general
adoption of the (moderate) objective view. These need not be based on incompatibilist
worries about determinism, as such (as Strawson himself seems to think: “determinism in
this case”), but might be constituted by scepticism about the kind of ultimate control that
might be thought necessary for action that could really merit an individual being subject
to ‘desert-entailing’ reactive attitudes. Alternatively, such general views might be
religious in character, or prompted by concerns about the fairness (in the general case) of
134
desert-entailing attitudes or judgements. That sceptical and incompatibilist views have
held such an enduring and deep hold on people, in many different times and places, not
simply as fanciful abstractions, but as genuine challenges to our usual ways of going on
with our relationships, emotions and judgements, itself gives the lie to Strawson’s claim
that such philosophical or theoretical views never stand behind the revision of reactive
attitudes. Indeed, Strawson’s ‘positive’ appeal to “the human commitment” could be
successful only if nobody, as a matter of fact, was ever persuaded by incompatibilist,
sceptical or Hybrid views. In other words, this element of Strawson’s argument could be
successful only if it were redundant and unnecessary, as it would be in circumstances
where there was nobody to whom it could be addressed, because there was nobody who
took a view opposed to Strawson’s own. Insofar as the positive appeal to “the human
commitment” has real work to do (e.g. when it can be addressed to the many who find
themselves convinced of incompatibilism, or of some Hybrid View), it simply cannot
work.
To strike a personal note, given my belief in the kind of restricted incompatibilism
captured by the Hybrid View, I often suspend, or try to suspend, desert-entailing reactive
attitudes precisely for theoretical reasons springing from this belief. My belief is that I
should do so in all relevant cases, but this, can, admittedly be a difficult goal to achieve.
Nevertheless, it is enough to act as a counter-example to Strawson’s claims about “what
we do” that, by introspection, I can see that the claim that reactive attitudes are never
suspended as the result of a theoretical conviction is false. Needless to say, I would
certainly not pretend to claim some special ability in this respect – in fact, quite the
135
opposite: my belief would rather be that, if I can manage to do this, fairly much anyone
can. Thus, by introspection, or otherwise, by paying attention to the practice of many
other people, Strawson’s empirical claim about how people think, and what their reasons
for suspending the reactive attitudes can be, can be seen to be demonstrably false.
In order for his sweeping claim about our current practice to stand, Strawson would need
both an argument both for why his description of our current practice really must be an
accurate one, and for why my objection, based on nothing more complicated than either
simple introspection, or else attention to the claims of others about their reasons for
action and belief, must embody a mistake or a misunderstanding. It is difficult to see how
Strawson could begin to pursue either line of argument, in part because he has moved the
debate onto the territory of examining, empirically speaking, how people are and how
they think. One cannot hope to appeal successfully to the authority of current practice, as
Strawson would like to do, if current practice is actually more diverse and contested than
he imagines. This question of how people are, and what their understanding of their
reasons for action happens to be, surely is one area where general philosophical
arguments will have difficulty getting purchase. The facts are, simply, the facts, and no
amount of philosophical disputation can make those facts otherwise. But, contra
Strawson, general philosophical arguments, whether sceptical or incompatibilist, can get
purchase with regard to individuals’ reasons for moving towards the adoption of the
(moderate) objective attitude. It is because they can gain such purchase that such
arguments have been felt so deeply, and have seemed so enduringly challenging and
136
disturbing. If they really were as innocuous as Strawson would have us believe, then it
would be difficult to understand their ongoing power to unsettle us. 125
H. Rationality, Objectivity, and ‘The General Framework of Human Life’
The foregoing section has suggested a number of reasons why Strawson’s appeal to the
‘Inescapability of the Human Commitment’ is ultimately unsuccessful. But the main
problem with Strawson’s entire approach is that it fails to engage the real question – and
this question is a normative one about what we ought to do, rather than a descriptive
question about our current commitments. Now, Strawson both acknowledges this
potential problem, and offers a suggestion for why it might not, after all, really be a
problem at all. Strawson’s strategy here is to try to forestall the criticism that his
approach is too ‘naturalistic’ by showing how such a line of criticism fails to take
seriously “the general framework of human life”. As he puts it:
For the real question is not a question about what we actually do, or why we do it. It is not even a question about what we would in fact do if a certain theoretical conviction gained theoretical acceptance. It is a question about what it would be rational to do if determinism were true, a question about the rational justification of ordinary inter-personal attitudes in general. To this I shall reply, first, that such a question (about what it would be rational to do) could seem real only to one who had utterly failed to grasp the purport of the preceding answer, the fact of our natural human commitment to ordinary interpersonal attitudes. This commitment is part of the general framework of human life, not something that can come up for review within this general framework.126
Later, Strawson says of this “general framework of human life” that:
125 On the recalcitrance and enduringness of incompatibilist intuitions, see Part Three, and especially Section 6.
126 F & R, p. 83.
137
The existence of the general framework of attitudes itself is something we are given with the facts of human society. As a whole, it neither calls for, nor permits, an external ‘rational’ justification.127
Before moving on to Strawson’s (‘fallback’) argumentative strategy of the ‘Appeal to the
“Gains and Losses to Human Life”’, we should first examine whether this appeal to “the
general framework of human life” can do anything to rescue his ‘Appeal to the
Inescapability of “the Human Commitment”’. There are (at least) two lines of response to
Strawson’s approach here. Firstly, we can simply dig-in our anti-naturalistic heels, as it
were, and again make the point that even if a certain commitment to a framework were
completely unavoidable, this would not show that this commitment really was immune to
rational criticism. Secondly, we could allow Strawson’s claim that we are committed to
some particular “framework of human life”, but question exactly what the composition
and extent of that framework might be.
If the commitment is simply to a life that involves human relationships and the adoption
of interpersonal reactive attitudes, then the proponent of the Hybrid View, who advocates
the adoption of only a moderate objective attitude, can present his proposed revisions to
the reactive attitudes (that is, the eradication of desert-entailing attitudes) as being wholly
innocuous for Strawson’s purported ‘general framework’. For, if “the general
framework” is plausibly general, then it cannot have too much in the way of determinate
content. It would be an error to build into the structure of “the general framework of
human life” a commitment to sets of reactive attitudes that were unnecessary for human
society, human interaction or human flourishing. So, if the framework in question really
127 F & R., p. 91.
138
is a general one, then there seems to be no obvious reason why it cannot survive the
banishment of the narrow reactive attitudes, so long as the broad reactive attitudes that
make human relationships possible are not banished alongside them. Similarly, if the
commitment is simply to a framework that allows human interaction, it can be quite
untroubled by the adoption of a moderate objective attitude, which does nothing to
preclude valuable forms of human involvement. The fact that we have reactive attitudes
in the context of human relationships may be a structural part of the “general framework
of human life”, but the existence of any particular narrow subset of those reactive
attitudes are just a few elements that might fit within that general framework; they are
not, themselves, an essential element of any necessary framework for meaningful human
life.
Alternatively, we might want to admit that it may indeed be part of our ‘human nature’,
or part ‘the general framework of human life’, that we are predisposed towards a
traditional set of reactive attitudes, including the desert-entailing narrow reactive
attitudes such as resentment, gratitude and forgiveness. But, we should also see our
commitment to making sure that we make good sense of the world, to keeping coherence
between our beliefs and attitudes, and to having a special regard for the truth, as
similarly being aspects of our humanity and of our commitment to a certain general
‘framework’. One can, in fact, see this framework as a framework of a more fundamental
kind, than a commitment to a particular subset of reactive attitudes. As Galen Strawson
puts it: “The fact that the incompatibilist intuition has such power for us is as much a
natural fact about cognitive beings like ourselves as is the fact of our unreflective
139
commitment to the reactive attitudes.”128 We might say that this natural fact is itself a
straightforward consequence of a general framework of natural human commitments. The
philosophical worries that might drive us to abandon the narrow reactive attitudes are not
alien extrusions from some place beyond human nature – they are simply expressions of
deep features of our cognitive commitment to bringing coherence to our beliefs, and to
our parallel commitment to living with a secure self-understanding.
Thus, the insight that there is an inconsistency between our employing the narrow
reactive attitudes and our assenting to the truth of certain incompatibilist, sceptical or
Hybrid views, should not be regarded as being in tension with some supposed
commitment to a ‘human framework’. Instead, such an insight should be seen as issuing
directly from just such a general commitment to a ‘framework’ of overall rational
coherence. P. F. Strawson’s selective invocation of an appeal to ‘humanity’ is therefore
rather misleading, and illegitimately treats our commitments to maintaining some
reflective coherence between our attitudes and beliefs as somehow not involved in any
“framework of human life”. As against Strawson’s view, we might believe that hoping to
maintain some degree of reflective control over the ‘reactive vocabulary’ that we deploy
in our interpersonal relationships, rather than unreflectively resigning ourselves to
whatever group of emotional reactions seem easiest or which present themselves in the
most obvious manner, is deeply characteristic of humanity, and fits perfectly with our
general commitment to “the general framework of human life.”
128 See Galen Strawson, (1993), “On ‘Freedom and Resentment,’’ in J. M. Fischer and Mark Ravizza (eds.), (1993), Perspectives on Moral Responsibility, (Ithaca, NY: Cornell University Press), p. 70.
140
Strawson’s argument for the inescapability of our commitment to certain reactive
attitudes as a precondition for the possibility of meaningful human agency is of the same
sort, and is every bit as unsuccessful, as would be an a priori argument in favour of the
inescapability of our commitment, as a prior condition to the possibility of sensory
perception, to the Newtonian nature of space. We can accept that space is not really
Newtonian (or Euclidean) even though we tend to experience it as such, just as we can
accept that our traditional, ‘natural’ set of emotional reactions and attitudes is misguided
and should be revised, even though we might be inclined naturally towards accepting
them.
Strawson goes on, later in his essay, to make a very revealing analogy between the
justification of induction and the justification of the “general framework of human life”.
The failure of this putative analogy is especially significant insofar as it strikingly
illuminates how implausible the narrow reactive attitudes are as candidates for essential
elements of such a ‘general framework’:
Compare [our commitment to the reactive attitudes within “the general framework of human life”] the question of the justification of induction. The human commitment to inductive belief-formation is original, natural, non-rational (not irrational), in no way something we choose or could give up. Yet rational criticism and reflection can refine standards and their application, supply ‘rules for judging of cause and effect’. Ever since the facts were made clear by Hume, people have been resisting acceptance of them.129
Strawson’s characterization of our commitment to a “general framework of attitudes”130
which is inescapably given, and the analogy with induction, are misleading. Strawson’s
129 F & R, p. 92, fn. 7.
130 F & R, p. 91
141
move here is to invite us to accept Hume’s naturalistic justification of induction as giving
us a general model that is also applicable in the case of the reactive attitudes as well. But,
this analogy is flawed. In the case of induction, what is at issue is our commitment to a
practice (or set of practices) that we cannot show to rest on any foundations. In the case
of the narrow reactive attitudes, on the other hand, what is at issue is our commitment to
a practice of (for example) resenting which we can show (or so we think) to be actively
incompatible with certain of our theoretical commitments (such as belief in some
incompatibilist or Hybrid view). Thus, insofar as the analogy with induction is meant to
provide some kind of argumentative backing for Strawson’s appeal to “the general
framework of human life”, it fails, given the fact that the supposedly analogous relation
to which Strawson points is actually one of significant disanalogy.131 It is also worth
noting that, whereas the commitment to induction seems universal and unchanging,
commitment to any particular subset of the reactive attitudes may well be much more
culturally and historically local.132
Rather than the supposed analogy of the reactive attitudes to induction, a better analogy
might be to belief in the existence of God, and the relationship of this belief to the
existence of religious attitudes. In many previous times, and in many human societies,
certain religious attitudes might have been seen as inescapable features of day-to-day life
or as “part of the general framework of human life”. But widespread rejection of the
131 For a related treatment of problems with Strawson’s analogy to induction, see Galen Strawson, (1986), Freedom and Belief, pp. 89-90; and Galen Strawson, (1993), ‘On “Freedom and Resentment”’, in Fischer and Ravizza, (eds.), Perspectives on Moral Responsibility, pp. 72-3.
132 On blaming and shaming among the Greeks, see Bernard Williams, (1993), Shame and Necessity, (Berkeley, CA: University of California Press), and also Arthur W. H. Adkins, (1960), Merit and Responsibility: A Study in Greek Values, (Oxford: Oxford University Press).
142
belief in the existence of God has acted (quite properly) to undermine the legitimacy of
these religious attitudes. One might want to imagine, back in a time of near-ubiquitous
religious belief, that there were philosophers who told us that it was inconceivable that
we should give up our use of ideas like ‘sin’, ‘grace’ and ‘redemption’ on the basis of a
‘merely theoretical’ conviction as to the non-existence of God. After all, using ideas like
‘sin’ and ‘grace’ is not just a matter of theoretical belief, but provides (at least some
aspects of) a form of life and a way of orienting one’s self-understanding. Yet we should
surely want to say that it would have been a mistake to try to put these ideas or religious
attitudes beyond the scope of rational criticism, no matter how pervasive, powerful or
central they may have been in structuring the lives and self-conceptions of those who
made use of, and lived through, those ideas and attitudes. We would surely conclude that
anyone who argued that ideas like ‘sin’ are immune to any influence from changing
theoretical beliefs is being rather misleading. In short, one should not slide from the
plausible claim that such ideas and attitudes can be of central practical importance, to the
false claim that such ideas and attitudes, because of their practical importance, are
beyond the scope of rational criticism.
Returning to the analogy with induction, a much more revealing way of presenting a case
that is structurally similar to the case of the reactive attitudes is available if we mark the
difference between questioning the whole set of reactive attitudes, and questioning only
the narrow set of desert-entailing reactive attitudes. Given that what is in question is not
our giving up all of our reactive attitudes, after all, but rather just a particular set of them,
a better analogy is not to induction, in toto, but to particular, narrow patterns of inductive
143
belief formation. For example, consider the way in which human beings seem to have an
original and ‘natural’ tendency to posit supernatural explanations for particular natural
events. For example, one group might say that an earthquake happened because the gods
were angry. Their process of belief-formation, in coming to this view, might be (at least)
quasi-inductive in character, in that it links an antecedent condition which has been
believed to hold in the past (the anger of the gods) to a resultant condition (the
earthquake) which was typically seen to have followed it. Nevertheless, we now
understand that such quasi-induction was faulty as an example of successful theoretical
reasoning. When rival ways of looking at the world supplanted worldviews that
promiscuously posited supernatural entities, these sorts of beliefs and patterns of
induction were also abandoned. So, similarly, if we come to adopt a general beliefs about
human agency of an incompatibilist or Hybrid type, we should naturally hope that our
commitment to narrow, desert-entailing reactive attitudes would in consequence be
abandoned.
Therefore, Strawson’s strategy of appealing to the “inescapability of the human
commitment” is misguided, both because it misses the central question of what it would
be rational for us to do, rather than what we tend “naturally” to do, and because it
illegitimately places the narrow reactive attitudes at the centre of “the human
commitment”. Moreover, that strategy cannot be rescued by appeal to a “general
framework of human life” that is outside the scope of rational criticism, for, even if there
is such a ‘framework’ (which we may doubt), it is implausible to see the (narrow)
reactive attitudes as constituting an inescapable element of a framework of general
144
human commitments. If Strawson wants to save the reactive attitudes from philosophical
challenges, then he needs to pursue a different kind of strategy.
I. Rationality, the Reactive Attitudes, and “the Gains and Losses to Human Life”
Strawson’s secondary fallback strategy is to move from an emphasis on the
‘inescapability of the “human commitment”’ to an appeal to the “gains and losses to
human life”. As he himself puts it:
And I shall reply, second, that if we can imagine what we cannot have, viz. a choice in this matter, then we could choose rationally only in the light of an assessment of the gains and losses to human life, its enrichment and impoverishment; and the truth or falsity of a general thesis of determinism would not bear on the rationality of this choice.133
And, in a related footnote, he continues:
But there is another question which should be raised … Quite apart from the issue of determinism might it not be said that we should be nearer to being purely rational creatures in proportion as our relation to others was in fact dominated by the objective attitude? I think this might be said; only it would have to be added, once more, that if such a choice were possible, it would not necessarily be rational to choose to be more purely rational than we are.134
Strawson seems to claim here that the only criteria for deciding whether or not we should
reject the desert-entailing reactive attitudes are a set of consequentialist considerations
concerned with the effect that this choice would have on our lives generally. Our
theoretical commitments, Strawson tells us, are of no relevance to the question of the
rational choice which we can make with respect to the retention or rejection of these
133 F & R, p. 83.
134 F & R, p. 83, fn. 4.
145
reactive attitudes. This rational choice, Strawson suggests, is a matter of assessing
consequences – hence the appeal to “the gains and losses” and to “enrichment and
impoverishment”. Given this claim, Strawson thinks that we must retain our belief in the
legitimacy of the reactive attitudes, or, at any rate, retain our uses for these attitudes,
given that the rational cost-benefit analysis which he enjoins us to undertake would tell us
that our lives would be impoverished by the absence of those attitudes.
I have two objections at this point to Strawson’s appeal to “gains and losses”. These two
objections operate at different levels of generality: each rejects a different aspect of
Strawson’s underlying assumptions. Therefore, my two objections give us two separate
and coherent points at which we can reject Strawson’s appeal to “the gains and losses of
human life”. Given our previous rejection of his appeal to “the inescapability of the
human commitment”, this gives us sufficient reason to reject his entire argument for the
‘quarantine’ of the reactive attitudes from philosophical criticism.
(1) The Desirability Objection
My first objection is that it may well be the case that we would be better off without the
desert-entailing reactive attitudes. One can easily imagine that a world without guilt,
blame and resentment might be an altogether more positive place to live than our current
world of interpersonal relationships. As with other matters, this ‘Desirability Objection’
seems more plausible, and hence Strawson’s view seems less attractive, when we bear in
mind the distinction between the broad and narrow reactive attitudes. There is no doubt
146
that a life without any reactive attitudes would be unconscionable; but a life without the
narrow reactive attitudes might be curiously delightful.
We would still, in our desert-free world, have praise and encouragement, even if we
would have to jettison gratitude. (Gratitude would, no doubt, be the attitude whose
absence we should feel most keenly if we were to adopt the moderate objective attitude.)
We would be able to rid ourselves of pride, which might be no great loss. Our
interpersonal relationship and lives of human engagement would not need to be
abandoned, but only that segment of our lives which employ the desert-entailing
attitudes. Moreover, as I have suggested above, a rejection of desert-entailing reactive
attitudes would not preclude friendship or genuine love. What would be precluded are
some of the most negative, destructive and unpleasant of human attitudes, resentment
among them. Strawson assumes that this would be an impoverishment, but it may well be
that it would lead to an enormously positive transformation in human affairs. Indeed,
Derek Parfit has suggested that Strawson has got his nomenclature precisely wrong: he
suggests that it is the incompatibilist who should be described as the ‘optimist’, and the
compatibilist as the ‘pessimist’.135 Parfit’s claim is all the more plausible when we bear
in mind the salient restriction to the narrow set of desert-entailing attitudes. It seems more
than plausible to think those who think we have good theoretical reasons to eradicate
attitudes like resentment are, in an important respect, optimists; whilst those, like
Strawson, who think we must take our emotional and reactive ‘vocabulary’ pretty much
as we find it, by contrast seem pessimistic.
135 Derek Parfit, in conversation.
147
If one can at least suggest that Strawson’s claim about the “gains and losses to human
life” is contestable, then it is robbed of its essential force. The question at hand, of the
consequential rationality of rejecting the desert-entailing reactive attitudes, then turns into
something much more disputable, and much less cut-and-dry than Strawson would have
us believe. Moreover, an abandonment of the desert-entailing reactive attitudes would
lead to an enrichment of human life, then, if we wish to follow Strawson’s strategy of
assessing the “gains and losses”, then we ought rationally to reject such attitudes.
(2) The Anti-Consequentialist Objection
Strawson regards the rationality of our choice with regard to the desert-entailing reactive
attitudes as a matter of consequentialist practical rationality. The previous ‘desirability’
objection allowed Strawson that assumption, but I now want to question Strawson’s
assumption that the rationality of this choice must be determinable from consequentialist
considerations, that is “only in the light of an assessment of the gains and losses to human
life.”136 The conception of rationality with which Strawson seems to be working is akin
to what we might describe as a ‘maximizing conception’ of rationality – that is, a
conception of rationality that says that we should always act in such a way as to bring
about what would most promote our overall wellbeing.137 Strawson is telling us that we
136 F & R, p. 83.
137 Or perhaps, on a more impersonal or impartial reading, one should read it instead as a utilitarian standard of rationality. “The gains and losses to human life” seems to be ambiguous between ‘first-personal’ and ‘impartial’ readings.
148
should act in such a way that the effects of our action are the ones which would provide
the most favourable balance of gains over losses.
However, it is by no means clear that this is a plausible account of the demands of
rationality in this context. For, we might well believe that the fundamental principle of
practical rationality is that we should always act in accord with some deontological
principle. We might imagine that there exists a contractualist constraint on our actions
and attitudes, such that we can only legitimately take up a given reactive attitude towards
another, or act on account of that attitude, if that attitude is something which that other
agent could not reasonably reject, or if it would be fair to bring that attitude to bear on the
agent in question. And one might think that a very clear case of an attitude which could
be reasonably rejected by its object would be one which relied for its support on a false
belief.138
Thus, even if we accept Strawson’s conception of the question as one in terms of
practical, rather than theoretical, rationality, his conclusion does not follow from it (even
setting aside the Desirability Objection), as his conception of practical rationality is
highly contestable. For, we could well believe that the highest principle of practical
rationality is to act always in such a way that one’s actions and attitudes display no patent
failure of coherence with one’s beliefs (such as one’s belief in the Hybrid View). On this
conception, we would have another kind of objection to Strawson’s appeal to “the gains
138 In Sections 4, 5 and 6 of Part Three, I examine the way in which it may be unfair to subject anyone to desert-entailing reactive attitudes.
149
and losses to human life”, and another resource for resisting his conclusion that the full
set of reactive attitudes must be left in place.
J. Conclusion: Strawson’s First Question Answered
So, let us now return to Strawson’s question, in its two central formulations:
What effect would, or should, the acceptance of the truth of a general thesis of determinism have upon these reactive attitudes?139 Could, or should, the acceptance of the determinist thesis lead us always to look on everyone exclusively in this way [i.e. with the objective attitude]?140
If we accept a version of the Hybrid View, or any other kind of incompatibilist or
sceptical view which involves a restricted incompatibilism, limited to the grounds for
claims of desert, then we should answer these questions by saying that we should
repudiate and excise the narrow, desert-entailing, reactive attitudes from our lives. This is
not impossible for us, because we can jettison that small region of our emotional
vocabulary populated by the narrow desert-entailing attitudes without thereby rejecting
reactive attitudes, as such, and without falling foul of any general “human commitment”
or “framework of human life”. Moreover, it is not at all clear that this would involve a
loss, as opposed to an intriguing gain, for the quality of human life, even if such
consequentialist considerations were taken to be decisive (which they need not be).
139 F & R, p. 80.
140 F & R, p. 81.
150
Similarly, if we accept a Hybrid View (or other ‘restricted incompatibilist’ views), we
should “look on everyone exclusively” through the lens of the objective attitude,
understood specifically in its moderate variant. Whilst adoption of a stark objective
attitude might be impossible, or at least deeply unpalatable, no such problems attend
adoption of the more modest, moderate version of that attitude. Moreover, to adopt the
stark objective attitude is to throw the baby out with the bath water, whilst adoption of
the moderate objective attitude does just the right amount to reshape our reactive
attitudes to the demands of coherence with a limited, Hybrid revisionist view. No “human
commitment” should or would stop us from embracing the moderate objective attitude;
and such an attitude cannot be avoided through a consequentialist appeal to the “gains
and losses of human life”.
Given these answers to the two formulations of Strawson’s question, we have no reason
to slide from appreciation of the undoubted elegance and insight of his account of the
exemptions and excuses in terms of ‘Capacity’ and ‘Quality of Will’, to the ‘strong’ view
that these two criteria exhaust the necessary and sufficient conditions for the applicability
of the reactive attitudes. We can, and should, suspend some of the reactive attitudes on
the basis of general, philosophical commitments.
None of this, of course, is to show that we should adopt a Hybrid View about agency and
claims of desert.141 But it is to say that, no matter what Hybrid, incompatibilist or
sceptical view we may hold about the nature of human agency, there is no reason to think
141 In Part Three, however, (especially Sections 4 and 6), I argue that we should adopt such a view.
151
that such a general philosophical view could, or should, be without practical effect as
regards the legitimacy or appropriateness of our reactive attitudes. If we want to leave our
day-to-day reactive ‘vocabulary’ unrevised and untroubled, then we need to show that no
such general, revisionary philosophical view is correct. What we cannot do, with any
hope of success, is to take refuge instead in the alternative Strawsonian strategy of
naturalistic avoidance. My hope is that this has been conclusively demonstrated by the
argument of this foregoing part of my discussion.
3. Moral Judgement, Responsibility and the Reactive Attitudes
A. “A More Usual Area of Debate”
Strawson’s strategy, in ‘Freedom and Resentment’, is a strategy of transferral. His aim is
first to establish a result regarding the reactive attitudes – that is, that they are immune to
any attempt to undermine them, whether by means of appeal to incompatibilist or
sceptical philosophical theories. Then, his intention is to transfer this result to the related
case of the moral attitudes of indignation and disapprobation. His treatment of the
‘participant’ reactive attitudes like resentment, gratitude and forgiveness is a precursor
for a similar treatment of moral blame, praise and responsibility. Indeed, it was because
of their purported relevance for the moral case, and their possible use as the basis for a
transferable argumentative strategy, that the non-moral reactive attitudes engaged
Strawson’s attention in the first place. As Strawson has it:
The point of the discussion of the reactive attitudes in their relation – or lack of it – to the thesis of determinism was to bring us, if possible, nearer to a position of compromise in a more usual area of debate. We are not now to discuss reactive attitudes which are
152
essentially those of offended parties or beneficiaries. We are to discuss reactive attitudes which are essentially not those, or only incidentally are those, of offended parties or beneficiaries, but are nevertheless, I shall claim, kindred attitudes to those I have discussed. I put resentment in the centre of the previous discussion. I shall put moral indignation – or, more weakly, moral disapprobation – in the centre of this one.142
The argument of Section 2 has, I hope, shown that Strawson fails to inoculate the narrow,
desert-entailing reactive attitudes from the challenge of scepticism or incompatibilism,
because he gives us no good reason not to move towards a moderate version of the
objective attitude, which is itself inconsistent with any desert-entailing reactive attitude.
Therefore, his strategy of transferral cannot work for any other kind of desert-entailing
attitude, including the “kindred attitudes” that are Strawson’s concern now that he has
moved his argument into the “more usual area” of debate about moral attitudes. The
moral attitudes of responsibility, guilt and blame seem like the very epitome of desert-
entailing attitudes, and as such would be just as impermissible by the lights of the Hybrid
View, or from within the moderate objective attitude, as would be attitudes like
resentment. Thus, if we have reason to think that we should accept the Hybrid View, or
some other view that enjoins us to adopt a version of the moderate objective attitude, then
we have no reason not to reject desert-entailing moral attitudes in just the same way as
we would have no reason not to reject desert-entailing non-moral attitudes.
Strawson himself certainly sees these moral attitudes of, for example, moral indignation,
blame and condemnation, as desert-entailing, as we can clearly see both from his
presentation of them as “kindred attitudes”, and from this passage towards the end of his
paper:
142 F & R, p. 83.
153
The concepts we are concerned with are those of responsibility and guilt, qualified as ‘moral’, on the one hand – together with membership of a moral community; of demand, indignation, disapprobation and condemnation, qualified as ‘moral’, on the other hand – together with that of punishment. Indignation, disapprobation, like resentment, tend to inhibit or at least to limit our goodwill towards the objects of these attitudes, tend to promote an at least partial and temporary withdrawal of goodwill: they do so in proportion as they are strong; and their strength is in general proportioned to what is felt to be the magnitude of the injury and to the degree to which the agent’s will is identified with, or indifferent to, it. (These, of course, are not contingent connections.)143
If the best understanding of the moral attitudes is as desert-entailing, and if Strawson’s
strategy of showing the philosophical ‘immunity’ of non-moral desert-entailing attitudes
has been shown to have demonstrably failed, then the ‘transferral strategy’ for rescuing
the moral attitudes from incompatibilist or sceptical worries must also fail. So, in one
sense, we can give very short shrift to Strawson’s hopes for extending his argument to
“the more usual area of debate”, in covering moral reactive attitudes. We might say that
his aim, in the second half of his paper, is to build on the foundations established in the
first half. But the argument of Section 2 has shown that those foundations are illusory,
and so the second half of Strawson’s strategy is therefore bereft of the argumentative
support on which it would need to rely in order for it to work successfully.
Nevertheless, it is worthwhile to discuss in more detail Strawson’s treatment of the moral
reactive attitudes, for a number of reasons. Firstly, some readers may remain unpersuaded
by the arguments of Section 2. If so, it will be useful to ask the question of whether
Strawson’s strategy of transferral would be successful if his position, regarding the
immunity of the non-moral reactive attitudes from philosophical worries, were to be
143 F & R, p. 90.
154
accepted. That is, one can ask whether, if we grant Strawson the foundations that he takes
himself to have earned for the second half of his argument, that latter part of the
argument is itself successful. Secondly, we may want to interrogate Strawson’s
conception of the nature of blame, responsibility and moral judgement, in order to see
whether it accurately represents the nature and shape of these ideas. We can, at the very
least, use Strawson’s discussion as providing a context for examining the striking claim
that moral indignation and disapprobation are “kindred attitudes” to the non-moral
reactive attitudes. Thirdly, we can show what effect the adoption of the moderate
objective attitude should have with regard to responsibility, blame and moral judgement,
through juxtaposing those effects against the discussion that Strawson gives us of what
the effects would be of adopting the stark variant of the objective attitude. In so doing,
this will allow us to see what the adoption of a view such as the Hybrid View should be
with regard to moral concepts, judgements and attitudes.
B. “Kindred Attitudes”? The Reactive Attitudes and their Structural Transformations
Strawson’s ‘strategy of transferral’ involves treating the moral attitudes in the same way
– in viewing them as “kindred attitudes” – to non-moral attitudes such as resentment. One
respect in which Strawson takes the two sorts of attitudes to be kindred to one another is
insofar as he takes both to be desert-entailing. But he takes their similarity to run deeper.
Intriguingly, Strawson regards the moral attitudes as transformations of interpersonal
reactive attitudes from the simple two-person case, whereby reactive attitudes take as
their objects the attitudes of others towards oneself, to the more complicated multi-
person case, whereby the moral reactive attitudes take as their objects the attitudes of
155
others towards third-parties. This account of the moral attitudes as “structural
transformations” of familiar interpersonal reactive attitudes is a fascinating one; if
successful, it would certainly demonstrate the “kindred” nature of the two sets of
attitudes. Moreover, if this account of the nature of the moral attitudes were successful,
then there could be little doubt that the legitimacy of Strawson’s ‘strategy of transferral’
would be beyond question.
However, there are good reasons to question this “structural transformation” account of
the moral attitudes, and their relation to the interpersonal reactive attitudes. Here is
Strawson’s statement of that account:
The reactive attitudes I have now to discuss might be described as the sympathetic or vicarious or impersonal or disinterested or generalized analogues of the reactive attitudes I have already discussed. They are reactions to the qualities of others’ wills, not towards ourselves, but towards others. Because of this impersonal or vicarious character, we give them different names. Thus one who experiences the vicarious analogue of resentment is said to be indignant or disapproving, or morally indignant or disapproving. What we have here is, as it were, resentment on behalf of another, where one’s own interest and dignity are not involved; and it is this impersonal or vicarious character of the attitude, added to its others, which entitle it to the qualification ‘moral’.144
This is certainly an arresting and elegant account of the nature of the moral attitudes.
Moreover, it is strikingly economical, as we need not posit any new variety of attitudinal
content in order to move from the interpersonal to the moral; rather, we can view the
domain of the moral as straightforwardly being nothing more than a ‘structural
transformation’ of the domain of interpersonal reactive emotions. Nevertheless, this is not
144 F & R, pp. 83-4. [My italics.]
156
a plausible account of what it might be for an attitude to be moral, and it gets Strawson’s
strategy of transferral off to an unpromising start.
To see why this account is implausible, one need only concentrate upon the way in which
the content of interpersonal attitudes can actually cut completely across the ‘structural’
features of such attitudes. We can feel moral indignation in the two-person case,
regarding the attitude of someone else towards us, just as much as we can experience
(non-moral) resentment in such cases. Similarly, we can straightforwardly resent the
attitude of some agent towards a third-party, without that resentment shading off into a
moral attitude. In order to illustrate these points, we can consider some examples:
(1) Liar’s Poker. You financially decimate me in a game of poker, having
deliberately misled me by means of a sophisticated and brilliantly executed
bluff. I may naturally resent you in this case, feeling considerable ill will and
anger towards you, related to a feeling of being comprehensively beaten and
humiliated by your skilful strategic manoeuvrings. You’ve shown me for the
flat-footed fool that I am, and have deeply hurt both my dignity and my
wallet. Nevertheless, I feel no moral attitude towards what you’ve done to me
– after all, we’re all engaged in a fair game, in which we all knew the rules
ahead of time, and in which deception is all part of the (in this case dubious)
fun.
157
(2) Jewel Heist. We plan a daring jewel-heist together, but at the last minute you
back out of our plan, citing your moral qualms. I resent you for your decision,
which harms my interests, and which puts all my weeks of hard work and
logistical planning to waste. I find myself feeling massive ill will towards you.
At the same time, though, I find your decision to be morally admirable. My
resentment of you is leavened by the merest sprinkling of moral admiration;
this has the effect of giving even greater piquancy to my deep resentment of
you, tinged as it now is my a more acute sense of my own moral degradations.
(3) Dot Com Deceit. You financially decimate me again, this time not in a card
game, but by inducing me to invest in your internet start-up company, having
systematically misled me about its prospects. Although the company
collapses, you make away with a considerable personal financial gain. In this
case, I may naturally both resent you, feeling considerable ill will and anger
towards you for exploiting my friendship for your personal benefit, and also
feel an attitude of moral indignation and disapprobation towards you, as you
have violated important norms of honesty and friendship. In this case, my
resentment is coloured by, and amplified by, my sense of having been morally
wronged by you.
(4) Trolley Car. You are driving a runaway trolley car, and must choose between
two forks in at a set of points on the track. If you direct the trolley car down
the leftward fork, you will hit and severely injure Green; if you direct the
158
trolley car down the rightward fork, you will hit and severely injure Blue. You
go left, and injure Green. I’m a good friend of Green’s, and feel a vicarious
sense of angry resentment towards you for causing his injury. I nevertheless
have no attitude of moral indignation or disapprobation towards you – I
realize that what you did was morally permissible, given the difficult
circumstances in which you found yourself.
(5) Heartbreaker Missionary. You break my daughter’s heart by jilting her, in
order to move to the developing world and devote your life to charitable and
missionary work. I feel a vicarious sense of angry resentment towards you, on
my daughter’s behalf. Nevertheless, I feel no moral disapprobation toward
you – I realize that you have selflessly subordinated the happiness and
interests of two affluent, comfortable Westerners (my daughter and yourself),
so that you can instead promote the interests of the genuinely needy and
disadvantaged. Indeed, my resentment is tinged with moral admiration for
you; this admiration softens my resentment, but it certainly does not
completely dislodge it.
(6) Exploitative Boss. You operate a garment factory in the third world, in which
children labour, in horrific conditions and for miniscule wages, to produce
sportswear. Your callous exploitation of the weak and powerless occasions
both my vicarious resentment and my moral disapprobation. In this case, my
159
two kinds of attitude towards you feed each other, and mutually reinforce one
another.
What do these examples show? Well, (1) ‘Liar’s Poker’, (2) ‘Jewel Heist’ and (3) ‘Dot
Com Deceit’ are cases in which the salient relationship is between just two agents: me
and you. Here, the individual whose attitudes are under examination is the individual who
is the ‘victim’ of the behaviour by the person to whom those attitudes are directed. In the
latter three cases, (4) ‘Trolley Car’, (5) ‘Heartbreaker Missionary’ and (6) ‘Exploitative
Boss’, the situation is more complicated. Here, we have three parties rather than just two.
Firstly, there is the party (“you”) whose behaviour is in question; secondly, we have the
party (“me”), whose attitudes towards your behaviour are in question; lastly, we have the
third parties or victims (Green, my daughter, the Third World child workers) who are
directly affected by your behaviour. ‘Liar’s Poker’ and ‘Trolley Car’ are analogous cases
to one another, insofar as they are situations in which resentment exists, but where there
is no simultaneous moral attitude, either existing alongside that resentment, or colouring
it. In ‘Jewel Heist’ and ‘Heartbreaker Missionary’ (another analogous pair of cases), by
contrast, we have the existence of resentment alongside a simultaneous moral attitude:
but, it is a moral attitude not of indignation or disapprobation, but, rather, of admiration
(or, if one prefers, of ‘approbation’). Lastly, in ‘Dot Com Deceit’ and ‘Exploitative Boss’
(our final related pair of cases), we have the simultaneous existence of resentment and of
moral disapprobation.
160
These sorts of cases suggest that Strawson’s ‘structural transformation’ account of moral
attitudes such as disapprobation and indignation is much too straightforward, and does
not do justice to the real complexity of these attitudinal phenomena. Moral indignation
cannot be, as Strawson claims it is, the ‘vicarious analogue’ of resentment, because, as
the ‘Trolley Car’ and ‘Heartbreaker Missionary’ cases show, one can feel vicarious
resentment without feeling any moral indignation at all (as in Trolley Car) or even whilst
one feels the contrary attitude of moral admiration or approbation (as in Heartbreaker
Missionary).
What Strawson’s ‘structural’ account leaves out of the picture is what is distinctive of the
content of moral attitudes. Rather than just being vicarious forms of resentment, moral
attitudes of, for example, indignation are in place only when there is a prior moral
judgement that the person to whom that attitude is directed has done something morally
wrong. Where one judges that there is no wrong, the moral indignation will either be
absent or, if it does emerge, it will be irredeemably ‘faulty’, degenerate or inappropriate.
Even in cases where an instance of vicarious resentment coincides with an instance of
moral disapprobation (as in the case of ‘Exploitative Boss’), the latter attitude is not
appropriately in place simply because of the presence of the former. Instead, something
further is needed, and that additional something is the moral judgement that there has
been morally deplorable action in this situation. We morally disapprove of the
‘Exploitative Boss’ not just because we resent him, vicariously, on behalf of his workers,
but because we further judge that he has wronged those third parties. By contrast, in the
‘Trolley Car’ and ‘Heartbreaker Missionary’ cases, our resentment on the part of a third
161
party is unaccompanied by the moral judgement that someone has been wronged, and
hence an attitude of moral disapprobation would not be legitimately in place.
We can thus suggest that the ‘structural transformation’ account of the moral attitudes
errs in that it is simply too aggressively reductive in seeking to given an account of
distinctively moral attitudes in terms of non-moral attitudes of resentment. What is
missing is the role that substantive moral judgements must have in creating and
legitimizing attitudes of moral appropriation and disapprobation. So, one regard in which
Strawson’s account of moral attitudes fails is in gliding over the ways in which moral
disapprobation is separable from, and something quite different from, vicarious
resentment. We do not give these two kinds of attitudes “different names… because of
this impersonal or vicarious character”145, but because of their very different internal
content. One might say that, in giving an account of moral indignation, Strawson over-
emphasizes the significance of the ‘indignation’, at the cost of paying attention to the
‘moral’ aspect of the attitude.
Another way in which Strawson’s account is problematic is in its suggestion that moral
judgements are, at root, vicarious (that is, finding their role only in the three-person case)
whereas reactive attitudes such as resentment are essentially inter-personal, that is,
essentially at home in the simple two-person case. The existence of straightforward
interpersonal moral disapprobation (as in ‘Dot Com Deceit’) shows the implausibility of
this way of marking out moral from non-moral reactive attitudes. Here, though, Strawson
145 F & R, p. 83.
162
seems to see that an overly-stark presentation of the ‘structural’ differences between
moral and non-moral attitudes is likely to lead him into conceptual trouble, and he
thereby attempts to provide his account with a welcome degree of additional nuance. As
he puts it:
Both my description of, and my name for, these attitudes [i.e. the moral or ‘vicarious’ attitudes] are, in one important sense, a little misleading. It is not that these attitudes are essentially vicarious – one can feel indignation on one’s own account – but that they are essentially capable of being vicarious. But I shall retain the name for the sake of its suggestiveness; and I hope that what is misleading about it will be corrected in what follows.146
The nuancing of Strawson’s account is just this: that, backing away from the excessively
stark claim that the moral attitudes are essentially vicarious, he moves to the weaker
claim that such attitudes are essentially capable of being vicarious. But the weaker
position to which Strawson moves here is also indefensible. For it does not mark out the
domain of the moral from the non-moral to say that moral attitudes are “essentially
capable of being vicarious” whereas non-moral attitudes are not. Cases such as ‘Trolley
Car’, ‘Heartbreaker Missionary’ and ‘Exploitative Boss’ show that non-moral attitudes of
resentment are just as capable of finding expression in ‘vicarious’, or third-person, cases
as are moral attitudes of disapprobation. Moreover, even in cases (like ‘Exploitative
Boss’) where vicarious resentment can exist alongside moral disapprobation, it is not
itself thereby transformed into moral disapprobation, but continues to exist as a
distinctive attitude alongside that disapprobation. And, as ‘Heartbreaker Missionary’ and
‘Trolley Car’ demonstrate, vicarious attitudes of resentment can exist in the absence of
moral disapprobation (and, indeed, even in the presence of moral approbation).
146 F & R, p. 84. [My italics]
163
Strawson’s suggestion that the domain of moral attitudes is to be considered distinct from
the domain of non-moral attitudes because the former, but not the latter, are “essentially
capable of being vicarious” is therefore to be rejected. But, once this suggestion has been
rejected, there is a substantial lacuna in Strawson’s account of how we might distinguish
the moral from the non-moral. For, if both moral and non-moral attitudes can be vicarious
(as demonstrated by cases (4), (5) and (6)), then we are presumably owed some other
account of how we might distinguish these two kinds of attitudes. If only moral attitudes
were capable of being vicarious, we would always know that all ‘third-personal’ attitudes
were moral in nature. But we have seen that non-moral judgements can also be vicarious.
In the two person case, if Strawson was right that only moral attitudes are capable of
being vicarious, then we could always know whether a particular attitude was a moral
attitude or not by considering whether that attitude could also be one that might be had on
our behalf by a third-party, or whether it could make sense only as a direct attitude of the
particular ‘offended party’ in question. But, again, if non-moral attitudes like resentment
can also be had “on our behalf” by third parties, then this test will not deliver any
determinate results as regards distinguishing moral from non-moral attitudes. On
Strawson’s view, therefore, it becomes obscure as to how we are to distinguish moral
from non-moral attitudes.
So, the question that needs to be answered is this: given that non-moral attitudes, just as
much as moral attitudes, are capable of being vicarious, what is it that makes moral
attitudes distinctive? My suggestion would be that we need to appeal to the content of the
164
substantive judgements, regarding the existence of moral wrongdoing, that generate and
legitimize moral attitudes such as indignation and disapprobation , but which are missing
in the case of non-moral attitudes. If this correct, though, then there is a sense in which
moral attitudes cannot reasonably be seen as “kindred attitudes” to reactive attitudes such
as resentment, in the way in which Strawson suggests (and on which his argumentative
strategy relies). For moral attitudes are not, therefore, to be seen as mere ‘structural
transformations’ of attitudes such as resentment. Instead, they are a special class of
attitudes which are, so to speak, ‘conceptually downstream’ of substantive moral
judgements. They rely for their existence (except in degenerate or faulty cases) and
legitimation on the existence of prior beliefs about the moral status of the actions of
others. In this way, such attitudes are hostage to the fortunes of such moral beliefs – for
example, if we come to the belief that some action is not properly to be regarded as
morally wrong, then an attitude of moral disapprobation becomes inappropriate in such a
case. To put things another way, moral attitudes are policed by different, and stricter,
norms than are non-moral reactive attitudes. They are answerable to our moral beliefs in
a way that connects them to judgements about how things are, morally speaking.
This puts Strawson’s treatment of the moral attitudes into considerable difficulty. If his
‘structural transformation’ account of the moral attitudes had worked, then it would be
legitimate to treat moral attitudes in the same way as he has treated non-moral reactive
attitudes. But, if the distinctiveness of moral attitudes resides in their ‘responsibility to’
antecedent moral beliefs, then we find ourselves dealing with attitudes that are not
relevantly “kindred” to the non-moral reactive attitudes. For, if moral attitudes depend on
165
our antecedent moral beliefs, then the appropriateness of some moral attitude will rely
upon the truth of that underlying belief. That means that the moral attitudes face different
normative standards with regard to their appropriateness and legitimacy than do non-
moral attitudes. Moreover, it opens the door for an additional dimension in which moral
attitudes can come under threat from general philosophical considerations. If, for
example, we hold that a some particular class of action is not properly to be understood
as morally wrong, then it follows from this immediately that attitudes of moral
disapprobation with regard to the perpetrators of such actions are not properly in place.
Strawson, in his structural account of the moral attitudes, would have us view moral
disapprobation as simply being that particular form of resentment which can be generated
from third-person perspectives. But, in moving beyond the failed ‘structural’ account of
the moral attitudes, we must jettison an understanding of such attitudes as
straightforwardly “kindred” to attitudes such as resentment. Moral attitudes are
normatively connected to moral beliefs in a way that attitudes such as resentment are not.
Given this, they face additional standards of appropriateness, not faced by attitudes such
as resentment. Moreover, once one brings into question this supposed “kindredness”
between moral attitudes and reactive attitudes such as resentment, an additional question
is brought into focus. Specifically, as has been mentioned in section 3(A) above,
Strawson’s ‘structural’ account of the kindredness of the two kinds of attitudes smuggles
in unquestioned the assumption that the moral attitudes are themselves desert-entailing
attitudes in the same way as attitudes such as resentment are desert-entailing. That is to
say that Strawson here assumes that the reactive attitudes to which the moral attitudes are
166
taken to be kindred are of the narrow, desert-entailing kind (such as resentment itself).
But this is a large assumption that itself needs to be subjected to careful scrutiny.
My hope is that the foregoing discussion has brought into question Strawson’s overly-
schematic account of the kindredness between resentment and the moral attitudes. In the
next section, I shall delve deeper in examination of the account of the moral which is
embedded in Strawson’s discussion of the moral attitudes, with the aim of rendering more
perspicuous a range of relevant issues, among them: whether moral attitudes are best seen
as kinds of reactive attitudes at all and whether, if we can view them in this way, they are
best seen as reactive attitudes of a broad or narrow kind.
C. Strawson on Disapprobation, Resentment and the Contours of the Moral
I have suggested that Strawson is somewhat cavalier in his characterization of what
makes an attitude a moral attitude; this suggests that there may be some problematic and
unexamined assumptions lurking within his account of the nature of morality and of
moral attitudes. With this in mind, it will be useful to turn to Strawson’s account of the
nature of the demands of morality, as a prelude to a fuller examination of his
understanding of the moral domain:
The personal reactive attitudes rest on, and reflect, an expectation of, and demand for, the manifestation of a certain degree of goodwill or regard on the part of other human beings towards ourselves; or at least on the expectation of, and demand for, an absence of the manifestation of active ill will or indifferent disregard. (What will, in particular cases, count as manifestations of good or ill will or disregard will vary in accordance with the particular relationship in which we stand to another human being.) The generalized or vicarious analogues of the personal reactive attitudes rest on, and reflect, exactly the same expectation or demand in a generalized form; they rest on, or reflect, that is, the demand for the manifestation of a reasonable degree of goodwill or regard, on the part of
167
others, not simply towards oneself, but towards all those on whose behalf moral indignation may be felt, i.e. as we now think, towards all men.147
We should here note the connection that Strawson takes to obtain between the reactive
attitudes (both moral and non-moral) and certain kinds of normative expectations. The
normative expectations in question relate to “the manifestation of a certain degree of
goodwill or regard” by others. The model that Strawson has in mind would seem to be
this: when some individual fails to manifest this goodwill or regard towards us, and
thereby fails to meet this normative expectation in the simple, interpersonal case, we
resent him or her; when that same individual fails to manifest this goodwill or regard
towards others, or towards humanity in general, and thereby fails to meet this normative
expectation in its more general case, we feel the vicarious analogue of resentment – i.e.
moral disapprobation or indignation – towards that individual.
This account of the moral, as we have seen in the foregoing section, is rather problematic.
For a start, the moral demands to which we hold others can be much richer and more
complex than the simple expectation that others manifest a general goodwill or regard
towards others. Moral demands do not straightforwardly reduce to expectations of the
goodwill of others; the moral standards to which we hold others are richly specified by a
variety of moral principles, that cannot satisfactorily be captured by a simple injunction
to evince good will towards others.
147 F & R, p. 84.
168
Let us consider two already familiar examples. If the criminal in ‘Jewel Heist’ had acted
out of his good will towards his fellow criminals, he would have gone ahead with the
robbery, and thereby acted wrongly. It is therefore implausible to think that acting with a
good will towards others is a sufficient condition for acting morally. On the other hand,
let us imagine that you have a very low regard for me, and view me with an ill will; you
therefore have no qualms about setting me up for a drop in our poker game (as described
in ‘Liar’s Poker’). You regard me without good will, and yet you do not wrong me in
such a situation, and your behaviour towards me is not itself morally problematic.
Therefore, it is implausible to think that acting with a good will towards others need be a
necessary condition for acting morally. In both sorts of cases, the normative expectations
associated with distinctively moral standards come apart from the simpler normative
expectation that others evince goodwill towards us.
Moreover, as has been already examined, it is inaccurate to characterize the vicarious
variants of reactive attitudes such as resentment as themselves being moral attitudes. On
the contrary, vicarious resentment (as in ‘Trolley Car’ and ‘Heartbreaker Missionary’) is
just vicarious resentment – which is to say, it is simply resentment felt on the part of
another individual rather than on one’s own account. It is not thereby a moral attitude.
For an attitude to be moral it needs to be connected in the right way with moral
judgements and moral beliefs; one thereby needs recourse to the content of distinctively
moral expectations in order to know when we should speak of attitudes as moral in
nature. The content of those distinctively moral expectations will be more highly
169
articulated and more complex than Strawson’s general reference to an expectation for a
certain degree of goodwill.
However, as has already been mentioned tangentially, perhaps the most troubling aspect
of Strawson’s account is the way in which it is permeated by a ‘desert-entailing’
conception of both the reactive and moral attitudes. That this might be a problem is
already suggested by the attempted reduction of the attitude of moral disapprobation to
the vicarious analogue of the specific attitude of resentment, given resentment’s status as
the archetypal narrow, desert-entailing attitude. But the strongly desert-entailing
conception of moral demands and attitudes is evident throughout Strawson’s discussion
of these issues. For example, recall that Strawson emphasizes the way in which the
reactive and moral attitudes “rest on and reflect”148 certain kinds of expectations or
demands for a certain ‘quality of will’ (i.e. for the good will of others). We are told that
personal reactive attitudes such as resentment “rest on and reflect” a demand for good
will towards ourselves; whilst the moral attitudes “rest on and reflect” a demand for the
good will of others “towards all men”. The use of “reflect”, in addition to “rest on” is
very significant here. Strawson’s idea seems to be not only that personal and moral
reactive attitudes are occasioned by a failure to meet this demand for good will, but also
further that such attitudes have a kind of isomorphism with (i.e. “reflect”) the failure of
good will (or presence of ill will) that provide them with their target. In other words, we
react in the personal case to a lack of good will or to the presence of ill will from some
individual, x, by virtue of a reaction (i.e. resentment) that involves a projection of ill will
148 F & R, p. 84.
170
towards (or withdrawal of good will from) that same individual, x. The demand that has
been violated, and the reaction to that demand, take the same form – i.e. a withdrawal of
good will (or projection of ill will). This is what makes attitudes such as resentment
distinctively desert-entailing.149 Such attitudes are desert-entailing because they have a
sort of ‘structural isomorphism’ to (i.e. they “rest upon and reflect”) acts of volition that
embody (in their ‘quality of will’) violations of the general demand for the presence of
good will, and for the absence of ill will.
What goes for the personal reactive attitudes, such as resentment, goes also, for Strawson,
for the moral attitudes. Just as resentment is desert-entailing because of its “reflection” of
an instance of another’s failure to meet the demand for good will towards me, so too
moral indignation is, on Strawson’s account, desert-entailing because it is in the same
way a “reflection” of an instance of another’s failure to meet the generalized demand for
good will (i.e. the demand for good will “towards all men”150).
For Strawson, it is of the very essence of the moral attitudes, just as it is of the very
essence of resentment, that such attitudes are desert-entailing. This is perhaps the most
significant dimension on which Strawson takes such attitudes to be mutually “kindred”.
This can be seen perhaps most clearly here:
149 See also the discussion of different ways in which an attitude may be desert-entailing in section 2(a).
150 F & R, p. 84.
171
The generalized and non-generalized forms of demand, and the vicarious and personal reactive attitudes which rest upon, and reflect, them are connected not merely logically. They are connected humanly; and not merely with each other.151
It is a stark picture of moral attitudes – as “logically and humanly” isomorphic to the
content of a generalized demand for good will, and hence as themselves internally
involving the withdrawal of good will. This particularly stark picture of the moral
attitudes is, however, rather implausible. To see why, let us return to the two cases
discussed above in which there was a violation of the expectation that another should act
morally – ‘Dot Com Deceit’ and ‘Exploitative Boss’. In both cases, resentment existed
alongside moral disapprobation although this did not mean that there was only one
combined attitude present in each case. Let us now imagine variants of each case, which
we’ll call ‘Dot Com Deceit*’ and ‘Exploitative Boss*’, in each of which cases, we have
only moral disapprobation, but without resentment:
(7) Dot Com Deceit*. You financially decimate me by inducing me to invest in
your internet start-up company, having systematically misled me about its
prospects; although the company collapses, you make away with a
considerable personal financial gain. In this case, I feel an attitude of moral
disapprobation towards you, as you have violated important norms of honesty
and friendship. Although I’m tempted to resent you for doing this to me, I find
myself feeling no such reactive attitude – instead, my feeling of moral
disapprobation is accompanied by a range of (wide) reactive attitudes of
151 F & R, p. 84.
172
sorrow, hurt and disappointment, but no narrow (desert-entailing) reactive
attitudes whatever.
(8) Exploitative Boss*. You operate a garment factory in the third world, in which
children labour, in horrific conditions and for miniscule wages, to produce
sportswear. Your callous exploitation of the weak and powerless occasions
my moral disapprobation. I do not resent you, though, as I think such attitudes
are a waste of energy and lead to a misdirection of our attention. Though I
morally disapprove of what you’ve done, I think it more important to change
economic structures so that you can’t go on doing this in the future, rather
than directing ill will towards you now.
Both (7) and (8) present scenarios that are perfectly intelligible. They are both cases in
which moral disapprobation exists, quite unaccompanied by any feeling of ill will. Now,
if Strawson’s view about the nature of moral attitudes were correct, then both (7) and (8)
would be “logically and humanly”152 impossible. It simply would not make sense to
describe a situation where we have moral disapprobation, but no withdrawal of good will,
or extension of ill will. After all, if the moral attitudes are structurally isomorphic to the
withdrawal of good will which occasions their existence, then there is no way in which
one might hold an attitude of moral condemnation that was not, at the same time, a
withdrawal of good will (that is, a form of vicarious resentment).
152 F & R, p. 84.
173
The possibility, and ready intelligibility, of cases such as Dot Com Deceit* and
Exploitative Boss* show the problems in Strawson’s overly schematic conception of the
nature of moral attitudes. By failing to give any independent weight to the content of
moral attitudes, and in treating those attitudes simply as vicarious forms of resentment,
Strawson has failed to see the connections that link moral attitudes with substantive
moral judgements and beliefs. At the same time, by pressing his account of moral
attitudes as “logically and humanly” desert-entailing, Strawson has failed to see how
attitudes of moral disapprobation can exist quite separately from any (narrow) desert-
entailing reactive attitudes such as resentment. In short, Strawson thinks that we cannot
have moral disapprobation without resentment; moreover, he really sees moral
disapprobation as nothing more than a form of resentment. We have shown, however, that
this picture is misguided. Moral disapprobation is very different from, and can exist in the
absence of, reactive attitudes such as resentment.
D. Strawson’s Conception of the Moral: Responsibility, Disapprobation and Resentment
Having highlighted some of the problems with Strawson’s conception of the moral
attitudes, and with their relation to the (narrow) reactive attitudes, we shall now turn to
examine the underlying conception of morality, and of moral responsibility, with which
Strawson is operating. After all, the central question about which philosophers have
struggled most consistently in this broad domain, is the question of whether moral
responsibility, as such, can be compatible with the causal determination of action (or,
more broadly, with an understanding of human agency as a natural phenomenon). And it
was a concern to present a ‘reconciliation’ between rival, ‘optimist’ and ‘pessimist’,
174
views about moral responsibility that led Strawson to his argumentative strategy via a
discussion of the reactive attitudes. Thus, we must find what we should best say about the
Strawsonian strategy, as regards its treatment of moral responsibility and its relation to
other elements of moral life, given the foregoing discussion of Strawson’s account of the
moral attitudes.
The key passages as regards Strawson’s conception of moral responsibility comes late on
his essay, in the passage where he most explicitly commits himself to a conception of the
moral and reactive attitudes as fundamentally desert-entailing:
The concepts we are concerned with are those of responsibility and guilt, qualified as ‘moral’, on the one hand – together with that of membership of a moral community; of demand, indignation, disapprobation, and condemnation, qualified as ‘moral’, on the other hand – together with that of punishment. Indignation, disapprobation, like resentment, tend to inhibit or at least to limit our goodwill towards the objects of these attitudes, tend to promote an at least partial and temporary withdrawal of goodwill: they do so in proportion as they are strong; and their strength is in general proportioned to what is felt to be the magnitude of the injury and to the degree to which the agent’s will is identified with, or indifferent to, it. (These, of course, are not contingent connections.)153
Here, the picture suggested above by Strawson’s talk of the moral attitudes “resting on”
and “reflecting” the demand for good will is given full and clear expression. Strawson’s
picture is that, in seeing others as members of a shared moral community, we hold them
to normative expectations in terms of their evincing a certain standard of good will (or
absence of ill will). This is the demand that also stands behind our standards for feeling
narrow reactive attitudes such as resentment, but here transformed into the general rather
than simple interpersonal case. When we view others as members of our moral
153 F & R, p. 90.
175
community, and hence as subject to these demands of good will, we thereby hold them
morally responsible. This, on Strawson’s view, is what the idea of moral responsibility
consists in – the idea of holding others to expectations regarding their volition states (the
‘quality’ of their ‘wills’) towards us and towards third parties.
This account of moral responsibility as a pattern of normative expectations is one half of
Strawson’s picture. We might call this the ‘formal’ aspect of Strawson’s picture of the
moral domain. The second element of Strawson’s picture is then given by the
‘substantive’ specification of the attitudes and practices that are legitimated by our
holding some individual to a standard of moral responsibility. This ‘substantive’ element
of the specification of moral expectations is given by Strawson’s list of the attitudes of
“demand, indignation, disapprobation, and condemnation”,154 together with the idea of
punishment. Strawson’s view would seem to have this form: in viewing you as a member
of our moral community, as someone whom we hold morally responsible, we thereby
view you as someone who can be held to moral demands, who can be subject to
indignation, disapprobation and condemnation, and who can be punished. But, given the
‘practical orientation’ of Strawson’s view, it is not that we should see your ‘moral
responsibility’ as some kind of precondition for the making of moral demands of you, or
for disapproving of, condemning or punishing you. Rather, it is your liability to
disapproval, condemnation and punishment that constitute your membership of our moral
community, and thereby your status as morally responsible.
154 F & R, p. 90.
176
Given that the moral attitudes are, on Strawson’s description, themselves essentially
desert-entailing, it follows that, in viewing you as a member of our moral community, we
thereby view you as liable to desert-entailing attitudes and, by extension, as also liable to
punishment. If we see you as ‘one of us’, and thereby as morally responsible, we see you
as a proper object for these varieties of desert-entailing attitudes and treatment that are
themselves a deep-seated internal aspect of the moral domain. To ask an additional
question about whether it is appropriate to blame you or punish you, or to ask whether
you “really are” morally responsible is therefore, on this view, to make a sort of category
mistake. Once we are within the bounds of a shared moral community, such questions
simply do not come up for debate.
Thus, on Strawson’s view, once morality is off the ground at all, it brings with it a view
of others as proper objects for desert-entailing attitudes and forms of treatment. There is
no gap to be filled, on Strawson’s view, between what T. M. Scanlon has called questions
of “attributive responsibility” and what he calls “substantive responsibility”.155 One may
naturally think that there is some further question to be asked, once one takes the view
that some individual has failed to act in accord with some moral demand, as to whether
that individual can be condemned or punished for acting in that way. But Strawson
simply denies the existence of an intelligible further question. Rather, seeing another
individual as subject to moral demands simply is to view that individual as morally
responsible, and this simply is to view that individual as properly liable to blame,
condemnation or punishment for his moral failings. Thus, there is no ‘additional’
155 On the ideas of attributive and substantive responsibility, see T. M. Scanlon, (1998), What We Owe to Each Other, (Cambridge, Mass: Harvard University Press), Chapter 6 on ‘Responsibility’.
177
question about substantive responsibility, as standards of condemnation and punishment
are internal to the moral domain and essential to it, and hence are directly settled by
questions of attribution.
This last significant point, about the internal connection between (i) viewing another
individual as subject to moral demands, (ii) seeing that individual as “morally
responsible” and (iii) seeing that individual as a proper subject for blame, condemnation
and punishment, goes to the very heart of the Strawsonian understanding of the nature
and contours of the moral domain. His earlier talk of the moral attitudes “resting upon
and reflecting”156 moral demands for good will pointed towards this stark and arresting
picture, but its fullest expression comes only in this later part of Strawson’s discussion.
Strawson is at his clearest here:
But these attitudes of disapprobation and indignation are precisely the correlates of the moral demand in the case where the demand is felt to be disregarded. The making of the demand is the proneness to such attitudes. The holding of them does not, as the holding of objective attitudes does, involve as a part of itself viewing their object as other than a member of the moral community.157
Here Strawson lays out with great clarity the claim regarding ‘structural isomorphism’
that was examined above in Section 3(C). Moral attitudes are seen as correlated to moral
demands. Moral demands are demands for good will, or the absence of ill will. When
those demands are not met, the appropriate moral attitudes are those that embody this
‘failure’ of good will – that is, they are attitudes that internally involve the withdrawal of
156 F & R, p. 84.
157 F & R, p. 90.
178
good will, or the presence of ill will. Hence, making moral demands involves, as an
internal, conceptual matter, subjecting the objects of those demands (that is, the fellow
members of one’s moral community) to desert-entailing attitudes, whenever those
demands are not met. This is simply what it is to view others as morally responsible. As
Strawson puts it, “the making of the moral demand is the proneness to such attitudes”.158
It may be that many of Strawson’s readers, who are perhaps eager to appropriate his
highly amenable conclusions for their own purposes, skip over this arresting presentation
of the nature of moral demands and of moral community. But it would be difficult to
overemphasize the importance of Strawson’s conception of the moral for the success of
his project in assimilating the defence of the moral attitudes (and moral responsibility
itself) to the defence of the (narrow) desert-entailing attitudes such as resentment.
Strawson’s metaethics plays an essential role in tying the fates of the moral attitudes to
the fates of the simple interpersonal reactive attitudes.
Now, I have already argued (see Section 2) that Strawson’s attempt to preserve the
(narrow) reactive attitudes from the possibility of philosophical contestation is a failure.
If one accepts the conclusions of that part of my discussion, then, even if one accepts
Strawson’s rather stark metaethical position, one will nevertheless also be forced to
conclude that his attempt to put the moral attitudes beyond the reach of philosophical
critique will also fail. But my wish was to address myself also to those who think that
Strawson’s ‘inoculation’ of the personal reactive attitudes from philosophical danger was
158 F & R, p. 90.
179
a success. If one takes this view, then one can only also accept the extension of
Strawson’s strategy to the case of moral responsibility if one is prepared to accept the
tight connection which Strawson draws between the ‘reactive’ and ‘moral’ cases, and if
one is prepared to accept Strawson’s metaethical claims, which present desert-entailing
attitudes and practices as being at the very centre of the domain of the moral.
If one does not accept Strawson’s account of the nature and structure of morality, then
one might look to separate the broad question of when we can make moral demands of
others, from the narrower questions of when we can hold them responsible, condemn
them or punish them. If one sees these two kinds of questions as separable, pace
Strawson, then one will again have scope to part company from his attempt to ‘cordon
off’ questions of moral responsibility from the range of contestation by ‘external’
philosophical considerations.
Thus, especially for those who may be unconvinced by the arguments of Section 2
(above), a great deal turns on the plausibility of the purported internal connections that
Strawson draws between moral demands, desert-entailing attitudes, condemnation and
punishment. We must therefore ask whether we have sufficient reason to reject the
picture of the moral domain which makes these connections. We must therefore consider
whether to agree with Strawson when he makes claims such as these:
The partial withdrawal of goodwill which these attitudes [i.e. the moral attitudes of indignation and disapprobation] entail, the modification they entail of the general demand that another should, if possible, be spared suffering is, rather, the consequence of continuing to view him as a member of the moral community; only as one who has offended against its demands. So the preparedness to acquiesce in that infliction of
180
suffering on the offender which is an essential part of punishment is all of a piece with this whole range of attitudes of which I have been speaking.159
The conceptual claims about morality that we need to subject to assessment should now
all be in view. The following claims, at least, are essential elements of this overall
Strawsonian picture:
Claim (α) Moral demands essentially entail a proneness to certain kinds of affective
attitudes.
Claim (β) The attitudes in question, within the domain of the moral, are essentially
desert-entailing – that is, they involve the withdrawal of good will, or the
extension of ill-will, or “the modification of the general demand that another
should, if possible, be spared suffering”.
Claim (χ) To view another individual as a member of one’s moral community is thereby
to see him as morally responsible, which is thereby to see him as a proper
target for condemnation, punishment, and other desert-entailing attitudes and
practices. It is to see him as subject to the desert-entailing attitudes described
in (β).
In fact, this picture is mistaken. The cases of (7) Dot Com Deceit* and (8) Exploitative
Boss* show that we have reason to reject Claim (β). There are many cases where moral
disapprobation is entirely unaccompanied by any kind of withdrawal of good will or
activation of ill will. To deny the possibility of negative moral assessment in the absence
159 F & R, p. 90.
181
of some negative desert-entailing attitude is to ride roughshod over the genuine
complexity of the phenomenology of moral judgement; moreover, it is to fail to see that
moral attitudes can exist without being versions of, and even without being accompanied
by, the narrow reactive attitudes such as resentment.
Moreover, Strawson’s Claim (α) involves a significant conceptual slide. Let us grant to
Strawson, for the sake of dialectical simplicity, the simplifying (although mistaken)
assumption that moral demands are reducible to demands on others for good will.
Nevertheless, demanding good will of you, and withdrawing my own good will in its
absence, are not the same thing, as Strawson claims they are. If I demand that you are
considerate towards me, then I make a normative claim on your future volitional states,
whether or not I am likely to, or prone to, withdraw my good will towards you if you do
prove to be inconsiderate towards me. It is difficult to understand why Strawson should
have deliberately made the claim that a demand that a certain expectation be met just is a
proneness to a certain kind of reaction if that demand is not met.160 The two are entirely
different states; and so any claim of their mutual identification is patently unsupportable.
Perhaps, reading Strawson charitably, we might instead take his claim to be that the
making of a moral demand, and the proneness to certain affective reactions if the demand
are not met, are causally related, or always go together for normal human beings. This
claim at least could be true, whereas the conceptual claim that Strawson actually makes is
straightforwardly false. But even the causal or ‘behavioural’ claim is false. Imagine an
160 See F & R, p. 90.
182
individual, Cynical Cid, who was very pessimistic about the moral performance of others,
and who had reached a state of resignation to the generality and pervasiveness of moral
failure. This individual might nevertheless go on making moral demands of others, but
might have no emotional or affective response whatever when those demands were not
met. Or, even if Cynical Cid did in general have certain kinds of affective response to the
failures to meet moral demands, they might be responses of wry amusement, vague
disappointment or patronizing condescension, rather than the desert-entailing attitudes,
involving the withdrawal of good will, described by Strawson. Moreover, as things are all
the time for Cynical Cid, so they often are for the rest of us. Sometimes the violation of
moral demands earns our ill will, other times we’re just resigned, disappointed or
saddened. That is to say that, even if many cases of the violation of moral demands
provide the occasion for desert-entailing reactive attitudes, it would be inaccurate to the
phenomenology of real human experience to claim that they all do.
Given our rejection of Claims (α) and (β), the implausibility of Claim (χ) should be
equally evident. We can allow that we sometimes make moral demands of others without
thereby seeing them as proper objects for condemnation. Instead, we may see them as
candidates for our assistance, or even our pity. That we are ready to make moral demands
of some individual is, at least in some sense, to see that individual as a member of our
moral community. Nevertheless, pace Strawson, even once we see an individual as a
member of our moral community, there is always a further question of how and whether
we are to hold them responsible, or blameworthy, for their moral failings.161 Blame,
161 For discussion of conceptions of responsibility and blameworthiness, see Section 3 of Part 3.
183
condemnation and punishment do not come along with the making of moral demands,
without further elaboration. That we see an individual as a member of our moral
community and thereby subject to such moral demands, is at best a necessary and not a
sufficient condition for viewing that individual as a proper object of desert-entailing
attitudes or practices.
We are now in a position to draw some conclusions regarding (what we may call)
Strawson’s metaethics. We should reject Strawson’s account of the nature of the moral
domain, and of the attitudinal and practical consequences of making moral demands on
others. Given this, even if we were minded to accept his ‘inoculation’ of the interpersonal
reactive attitudes from philosophical critique, we should reject his account of the
relationship between the reactive attitudes and the moral attitudes. Moral demands,
expectations and attitudes need not, pace Strawson, have the same kind of conceptual
structure as attitudes such as resentment; they need not be ‘desert-entailing’ like the
narrow reactive attitudes; they are more intimately linked to beliefs and judgements than
are the non-moral reactive attitudes; in short, they are not ‘kindred’ attitudes to attitudes
such as resentment.
Thus, when Strawson claims that (a) the full range of moral attitudes are all of one piece,
and that (b) the full set of moral attitudes must ‘stand or fall’ with reactive attitudes such
as resentment, we can remain resolutely unconvinced. The case of the moral attitudes is
quite separate from the case of resentment, and needs to be treated carefully in accord
with a sensitivity to the particular issues that are at stake in the moral domain.
184
E. The Moral Excuses and Exemptions, Moral Judgements, and the Objective Attitude
The foregoing discussion has cast doubt on Strawson’s assimilation of the case of the
moral attitudes to the case of the interpersonal reactive attitudes. It has thereby opened up
the space for us to consider Strawson’s account of the moral excuses and exemptions, and
his general attempt to ‘inoculate’ ideas of moral responsibility from philosophical
critique, in a way that will be unencumbered by the somewhat eccentric commitments of
Strawson’s distinctive metaethical views. That is to say, we can assess the success of
Strawson’s account of the ‘immunity’ of moral responsibility from sceptical or
incompatibilist worries while bracketing the less plausible aspects of his account of the
nature of moral attitudes and of the domain of the moral. The rest of our discussion will
thereby keep the distinctiveness of the moral attitudes in view, and will from that position
take issue with Strawson’s account of how the moral attitudes, and the idea of moral
responsibility, should be moved to a privileged position beyond the reach of the
disputations of compatibilists and incompatibilists.
We recall from Section 2(B) that, with regard to the reactive attitudes, Strawson presents
an account of the exemptions and excuses in terms of ‘Capacity’ and ‘Quality of Will’.
We may also recall that this account of the exemptions and excuses is open to a Strong
and a Weak construal. On the Weak construal, Strawson may be read as advancing only
the uncontroversial thesis that the ‘Capacity’ and ‘Quality of Will’ criteria provide
sufficient conditions for the suspension of reactive attitudes. On the Strong construal,
which is the one which Strawson actually holds, these twin criteria provide not just
185
independently sufficient conditions for suspending the reactive attitudes, but also provide
disjunctively necessary conditions for the legitimate suspension of such attitudes.
As one would expect, Strawson advances an account of the excuses and exemptions, with
regard to moral responsibility, that is analogous to his treatment of the excuses and
exemptions in the case of interpersonal reactive attitudes. As with the earlier case,
Strawson advances a ‘strong’ version of the account of exemption and excuse in terms of
‘Capacity’ and ‘Quality of Will’. Thus, on Strawson’s view, the only reasons we may
have for considering an individual to not be morally responsible for his action in a
particular case is if (a) that individual is not a normally function adult human (i.e. does
not meet the ‘Capacity’ criterion), or if (b) that individual’s action in this particular case
does not reflect upon his volitional state, as when ‘he was pushed’, ‘he didn’t mean to do
it’, ‘he was over-tired’, and so on (i.e. when the agent’s ‘Quality of Will’ is not itself in
violation of the moral demands to which that agent is subject).
If the Strong construal is correct, then it can never be appropriate to exempt individuals
from attitudes of moral disapprobation or blame on general philosophical grounds. In
other words, if the Strong construal of the exemptions and excuses stands, then sceptical
or incompatibilist worries about moral responsibility are otiose. Hence, it is incumbent on
us to investigate whether the Strong construal is more plausible with regard to the moral
attitudes than it was shown to be with regard to the interpersonal reactive attitudes.
186
As we may recall from Section 2(B), excusing someone from a particular reactive attitude
on ‘Quality of Will’ grounds shows only that the putative reactive attitude was based on a
mischaracterization of that other agent’s will. This is as much the case with excusing an
individual from moral attitudes as it is with regard to excusing them from ordinary
interpersonal reactive attitudes. This means that ‘Quality of Will’ excuses do not really
bring the nature of the individual’s agency into question; they simply tell us that certain
attitudes to which we might have been drawn with regard to that agent are inappropriate
given an accurate characterization of that agent’s will. As Strawson puts it:
Considerations of this class operate in just the same way, for just the same reasons, in connection with moral disapprobation or indignation; they inhibit indignation without in any way inhibiting the sort of demand on the agent of which indignation can be an expression, the range of attitudes towards him to which it belongs. But in this connection, we may express the facts with a new emphasis. We may say, stressing the moral, the generalized aspect of the demand, considerations of this group have no tendency to make us see the agent as other than a morally responsible agent; they simply make us see the injury as one for which he was not morally responsible. The offering and acceptance of such exculpatory pleas as are here in question in no way detracts in our eyes from the agent’s status as a term of moral relationships. On the contrary, since things go wrong and situations are complicated, it is an essential part of the life of such relationships.162
Strawson’s claims here are plausible. Thus, the real philosophical action, with regard to
questions of moral responsibility, is located more with regard to the ‘exempting’
Capacity conditions, than with regard to the ‘excusing’ Quality of Will conditions. As
was pointed out in Section 2(B), excluding someone from being the subject of a reactive
attitude on ‘Capacity’ grounds involves a much more significant move – it excludes
them, as Strawson sees it, from the whole ambit of moral attitudes and emotions, rather
than just suggesting the inappropriateness of one particular kind of attitude. It places
162 F & R, p. 85.
187
them outside the realm of moral relations, and thus marks them out as someone with
regard to whom it is inappropriateto have any moral attitudes at all (either temporarily, in
some cases, or permanently, in cases of structural incapacity).163
As was the case with regard to exemptions from the interpersonal reactive attitudes,
Strawson describes the attitude that we take towards such individuals as the ‘Objective
attitude’. This attitude is to be understood as an alternative stance to the alternative view
of others as belonging to our moral community (that is, the moral analogue of the
‘participant attitude’), that stands behind the set of moral attitudes. As Strawson puts it:
Seeing an agent in such a light as this tends, I said, to inhibit resentment in a wholly different way. It tends to inhibit resentment because it tends to inhibit ordinary inter-personal attitudes in general, and the kind of demand and expectation which those attitudes involve: and tends to promote instead the purely objective view of the agent as one posing problems simply of intellectual understanding, management, treatment and control. Again the parallel holds for those generalized or moral attitudes towards the agent which we are now concerned with. The same abnormal light which shows the agent to us as one in respect of whom the personal attitudes, the personal demand, are to be suspended, shows him to us also as one in respect of whom the impersonal attitudes, the generalized demand, are to be suspended. Only, abstracting now from direct personal interest, we may express the facts with a new emphasis. We may say: to the extent to which the agent is seen in this light, he is not seen as one on whom demands and expectations lie in that particular way in which we think of them as lying when we speak of moral obligation; he is not, to that event, seen as a morally responsible agent, as a term of moral relationships, as a member of the moral community.164
Here we see Strawson’s peculiar view of the moral domain to the fore. For Strawson, the
making of moral demands essentially involves our proneness to take up certain kinds of
reactive (moral) attitudes if those demands are not fulfilled. The possibility of making
163 On the difference between ‘structural’ and ‘temporary’ invocations of the Capacity criterion, see Section 2(b) above.
164 F & R, p. 86. [My italics.]
188
moral demands in an ‘affectless’ register, as in, say, (8) Exploitative Boss*, or as things
are with Cynical Cid, do not even enter as conceptual possibilities on the Strawsonian
view. As has been argued above, this itself is a serious mistake.
But let us set this issue aside and instead raise the question of whether Strawson is right
in making these central claims, Claim (φ) and Claim (γ):
Claim(φ) The Strong construal of the Capacity criterion is the correct one. That is, we
should see the Capacity criterion as providing necessary and sufficient
conditions for when an agent should be exempted from any of the full set of
the moral attitudes. (As opposed to where, in particular cases, he should be
excused, in accordance with the Quality of Will criterion).
Claim (γ) If we take the ‘objective attitude’ towards some agent, insofar as we suspend
(certain) moral attitudes towards him, we thereby see that agent as not being
subject to moral demands and obligations. That is, taking the ‘objective
attitude’ towards some agent is to see him as outside the bounds of our moral
community, as not being “a term of moral relationships”.165
If Claim (φ) is true, then general philosophical considerations (be they Incompatibilist,
Sceptical or Hybrid) cannot legitimate a suspension of the moral attitudes, which is to say
that there can be no philosophical threat to everyday moral responsibility. Now, it would
165 F & R, pp. 85-6.
189
seem that Claim (γ) does much to bolster the plausibility of Claim (φ). For if the only way
in which we can look upon an individual as less than fully morally responsible is to see
them as standing entirely outside the scope of moral demands, then allowing that there
may be general philosophical reasons for scepticism about moral responsibility is to
allow that such general philosophical reasons would also threaten the existence of the
whole moral domain. In other words, Claim (γ) suggests a ‘Stark’ reading of the objective
attitude in the moral case, such that the only way to view other individuals ‘objectively’,
rather than accepting the whole range of moral attitudes towards them, is to view them in
a way that is entirely outside the bounds of morality.
Let us, though, pause, and recall the analogous dialectical situation from Section 2(C).
Here, the Strong construal of the excuses and exemptions was shown to be plausible only
on a similarly ‘Stark’ reading of the objective attitude. But, it was shown that the ‘Stark’
reading of the objective attitude was not the only version of that attitude that was
available in the case of interpersonal reactive attitudes. Whilst the ‘Stark’ objective
attitude ruled out all reactive attitudes tout court, the ‘moderate’ objective attitude ruled
out only the ‘desert-entailing’ reactive attitudes, such as resentment. On a ‘moderate’
objective attitude, it was shown than we could look at other individuals ‘objectively’
whilst nevertheless still viewing them as agents who were participants in real human
relationships. Thus, it was shown that Strawson overplayed his hand in emphasizing the
costs and difficulties of moving towards some version of the objective attitude, such that
the plausibility of his ‘Strong’ account of the exemptions and excuses, and his resistance
190
to philosophical sources for reasons to modify our reactive attitudes, could both be
overcome.
In the argument of Section 2, a great deal followed, ultimately, from our ability to
distinguish between the broad and the narrow (i.e. desert-entailing) reactive attitudes, and
it was shown that much of the surface plausibility of Strawson’s own position relied on
an equivocation between an understanding of the reactive attitudes in a ‘broad’ and a
‘narrow’ way. Once one can mark the difference between the broad and narrow reactive
attitudes, one is then able to distinguish between ‘Stark’ and ‘Moderate’ versions of the
objective attitude; and once the possibilities of the moderate objective attitude are laid
out, the exaggerations of Strawson’s approach become apparent, and the Strong construal
of the Strawsonian account of the exemptions and excuses thereby comes to seem
implausible. The question, then, is whether an analogous strategy can be pursued here
against Strawson in the moral case.
If Strawson were right about the nature of the moral domain, and the internal structural
relationships between, on the one hand, making moral demands and, on the other hand,
proneness to certain kinds of desert-entailing moral attitudes, then no such analogous
strategy would be open to us, as there would be no distinction to be made between
‘broad’ and ‘narrow’ moral attitudes. In other words, if Strawson’s claims (α), (β) and
(χ) all held true, then there would be no way of pursuing a parallel argumentative strategy
against Strawson’s position in the moral case. However, Strawson’s account of the nature
of moral demands and attitudes has been shown to bowdlerize the variety and content of
191
moral attitudes. We therefore have good reason to reject claims (α), (β) and (χ).166
Hence, we have good reason to think, pace Strawson, that there is, indeed, an analogous
distinction to be made in the moral domain between ‘broad’ moral attitudes and ‘narrow’
(i.e. desert-entailing) moral attitudes.
Let us recall, for example, the differences between the moral attitudes described in some
of the cases given above in Sections 3(C) and 3(D). Some of the cases, such as (3) Dot
Com Deceit, and, especially, (6) Exploitative Boss, can be seen as involving the
expression of a full-blooded attitude of moral blame, or moral condemnation, considered
as the kind of robust, desert-entailing attitudes upon which Strawson focuses. Here we
really do have moral attitudes of the kind that Strawson himself describes – forms of
moral attitude that do, indeed, involve an “at least partial and temporary withdrawal of
goodwill”.167 But contrast this with cases (7) and (8), the variant cases of Dot Com
Deceit* and Exploitative Boss*, where we have ‘weaker’ moral attitudes of
disapprobation or indignation, generated by a moral judgement of having been wronged,
but which occur without any stronger attitude of moral blame or condemnation. This
moral attitude of disapprobation can, of course, be accompanied by other attitudes,
whether of disappointment, resignation, “hurt feelings”, or whatever else, but they need
not be accompanied by (and nor do they ‘internally’ entail, nor essentially involve) any
withdrawal of good will or activation of ill will. We can thus conclude that, just as was
166 See the discussion of these claims in Section 3(D) above.
167 F & R, p. 90.
192
the case with regard to the personal reactive attitudes, we can speak of both ‘broad’ and
‘narrow’ (i.e. ‘desert-entailing) moral attitudes.
It would seem, therefore, that the usual range of moral attitudes straddle the border
between the ‘broad’ and the ‘narrow’: we often slide between moral attitudes that are
desert-entailing and those that are not. But it is nevertheless significant that that the two
types of attitude can be kept apart conceptually. Returning to Strawson’s list, we recall
that he sees the entire set of moral attitudes, of “demand, indignation, disapprobation,
condemnation”168 as standing or falling together, as an undifferentiated set. But that
suggestion of undifferentiation is precisely what is so misleading about his picture.
The question of the conditions for making moral demands is, as we have seen, a
conceptually separate matter from the question of what attitude should be in place if that
demand has not been met.169 Moreover, the simple attitude of “moral disapprobation”,
which one can take to be the most basic moral attitude, is not in itself a narrow, desert-
entailing attitude. As shown by cases like (8) Exploitative Boss*, one can feel the attitude
of moral disapprobation without feeling any further attitude of ill will or any form of
moral blame considered as a form of sanction. Hence we can reject Strawson’s Claim (β),
which states that all moral attitudes are essentially desert-entailing.
168 F & R, p. 90.
169 See the discussion and rejection of Claim (α) in Section 3(D).
193
Now, perhaps attitudes of moral ‘indignation’ may be seen as an interesting halfway case,
in which different gradations of attitude can be identified. Some moral indignation is
simply a form of “moral anger” occasioned by the fact of moral wrongdoing, but not
necessarily directed as a form of ill will towards the wrongdoer. We should, perhaps,
therefore, distinguish between (wide) attitudes of “indignation at” and (narrow) attitudes
of “indignation with”, as in the case of the distinction between the wide attitude of anger
simpliciter and the narrower attitude of “anger at”.170 With moral condemnation, which I
take to be an attitude of the same kind as blame, we come closer to the expression of a
full, narrow, desert-entailing attitude: to a form of moral disapprobation which carries
with it the expression of moral censure, and a concomitant withdrawal of good will (or
activation of ill will) towards the moral transgressor. To put things in one potentially
illuminating way: moral disapprobation is merely the attitudinal correlate of a bald moral
judgement – it is straightforwardly the opposite of endorsement; whereas, on the other
hand, moral blame and moral condemnation are attitudes that are in an important way
targeted at others, and involve an “at least partial and temporary” withdrawal of good
will towards them.
Strawson is not wrong when he identifies the existence of a form of moral attitude which
is deeply and inextricably desert-entailing, for such moral attitudes certainly do exist.
Where Strawson goes wrong is in thinking that such attitudes exhaust the totality of the
possible moral attitudes. His Claim (β) is not so wide of the mark as to see desert-
entailing attitudes where none exist; rather, this claim goes wrong insofar as it describes
170 See the discussion of anger in Section 2(C).
194
all, rather than only some, moral attitudes as desert-entailing. He overlooks the existence
of non-desert-entailing moral attitudes such as simple moral disapprobation. Whereas
Strawson paints all moral attitudes as belonging to the same (desert-entailing) species, a
more nuanced examination shows that they can be of a much broader variety of kinds.
Given that moral attitudes can come in ‘broad’ and ‘narrow’ varieties, we thereby find
ourselves in an analogous dialectical position to that which we faced in Section 2, with
regard to the inter-personal reactive attitudes. We therefore need first to understand what
would be involved in taking up a moderate version of the objective attitude, with regards
to its consequences for the moral attitudes. Insofar as the moderate objective attitude
involves only the preclusion of desert-entailing attitudes, and not of broader forms of
reactive attitude, it need not preclude broad moral attitudes of appraisal, approbation and
disapprobation, or indignation. It certainly need not preclude making moral demands
upon others, or of assessing them against the background of certain kinds of normative
expectations. What the moderate objective attitude does rule out is those varieties of the
moral attitudes that are desert-entailing; it thereby precludes attitudes such as moral
blame and moral condemnation, insofar as those attitudes involve the withdrawal of good
will or the activation of ill will.171
171 It is a large question as to whether blame really does essentially involve the adoption of a ‘desert-entailing’ attitude. For the purposes of the current discussion, blame will always be understood to refer to a desert-entailing attitude. I share the view of Gideon Rosen, that “[a]s is commonly observed, moral blame is a sort of sanction. Even when it is not expressed, it is a form of adverse treatment: a form of psychic punishment.” [My emphasis] in his (2003), “Culpability and Moral Ignorance,” Proceedings of the Aristotelian Society, CIII, I, 61-84. To a large degree, disagreement on this point is merely terminological: even those who insist that blame, as such, is not desert-entailing must surely allow that there is a particular variety of moral attitude that is desert-entailing. Let anyone who takes issue with my use of the term “blame” call this desert-entailing moral attitude by whatever name they prefer; at any rate, it is this variety of moral attitude, which really is the “moral analogue of resentment”, so to speak, which I am here discussing.
195
Is the moderate objective attitude consistent with viewing other individuals as morally
responsible? The answer to this question will depend, not surprisingly, on a more detailed
specification of what we mean by ‘moral responsibility’. Strawson, with his commitment
to the conceptual linkages laid out in Claim (χ), uses the term in a very broad way,
making no distinction between, on the one hand, moral responsibility in the sense of
being held to moral standards, or being seen as an appropriate subject for moral
assessment, and, on the other hand, moral responsibility as being blamed, sanctioned, or
made to bear certain costs or forms of hard treatment. Although the distinction that I have
in mind here is not quite identical to Scanlon’s way of cutting the relevant conceptual
territory, it would not be too far wide of the mark to say that Strawson makes no
distinction between ideas of attributive and substantive responsibility.
Making such a distinction between different senses of responsibility will be essential,
however, if we are to make sense of the ways in which the moderate objective attitude is
and is not consistent with holding individuals morally responsible. If by ‘moral
responsibility’ we mean only the idea of being held to certain moral standards, and
subjected to moral assessment for one’s success or failure in meeting those standards,
then attributions of moral responsibility are quite fully consistent with holding the
moderate version of the objective attitude. If, on the other hand, by ‘moral responsibility’
we mean being held liable to blame, punishment or other forms of hard treatment
(including the withdrawal of good will or the activation of ill will), then we find that the
196
moderate objective attitude is not consistent with treating other individuals as morally
responsible.
Thus, the question of whether the moderate objective attitude precludes moral
responsibility is in part terminological; in fact, giving a clear answer to the question
inextricably involves one in careful delineation of different salient senses of ‘moral
responsibility’. But, setting aside the terminological point, the underlying substantive
issues should be clear – taking the moderate objective attitude towards some agent need
not involve seeing him as outside the ambit of moral demands; it need not involve giving
up on the moral assessment of that individual; but it does require that one give up on the
thought that that individual can, simply by virtue of having been subjected to negative
moral assessment, thereby legitimately be blamed, condemned, punished or ill-treated.
We can now return to assessment of Strawson’s Claim (γ). Claim (γ) held that to take the
objective attitude towards some individual is thereby to see him as not subject to any
moral demands; as outside the scope of our moral community; and as not being a
potential “term of moral relationships” or “member of the moral community”.172 This
may be an accurate characterization of the commitments of the Stark objective attitude,
but it certainly is not an accurate characterization of the commitments or consequences of
the objective attitude considered in its moderate variant. An understanding of our fellow
agents as subject to moral demands and moral assessment can coexist with a strict
172 F & R, p. 86.
197
application of the moderate variant of the objective attitude. Therefore, we have sufficient
reason to reject Strawson’s claim (γ).
Now, the aim of the rest of this essay is, in effect, to adjudicate on the plausibility of
Strawson’s claim (φ): that is, to ask whether it is only application of the Capacity
criterion that can give us reason to suspend any of the moral attitudes towards some
agent, or whether we might also have reason to do so on the basis of general
philosophical considerations (of an incompatibilist, sceptical or Hybrid kind). This
question is, of course, not yet settled by the rejection of Claim (γ). But Strawson’s
dialectical strategy is such that, having rejected Claim (γ), we are much less likely to be
convinced by many of his arguments; for Strawson’s strategy is to emphasize either (i)
the impossibility or (ii) the horrific costs of adopting the objective attitude. Yet, with the
moderate version of the objective attitude in view, the impression of impossibility and/or
normative costliness begins, as we shall see, largely to disappear.
Our rejection of Claim (γ) shows one way in which the costs of the (moderate) objective
attitude are not as great as Strawson believes: for we can adopt the objective attitude, and
with it reject certain narrow (desert-entailing) moral attitudes, without being forced to
abandon the moral domain in toto. Thus, just as the argumentative strategy of Section 2
was essentially to show how we could give up on resentment without giving up on human
relationships of a rich and rewarding kind, so, similarly, we can here emphasize that we
can give up on the desert-entailing attitudes of blame and condemnation without giving
up on the domain of the moral.
198
F. On the Cultivation of the “Purely Objective View”
Strawson is not so rash as to make the claim that the objective view is accessible only
through an application, in a particular case, of the Capacity criterion in exempting an
individual from the full glare of the moral attitudes. He allows, as he did in the
straightforward interpersonal case, that the objective view is also accessible to us via an
act of will. He allows that the objective view is one that we can occupy if we choose to
do so, at least to some degree and for a limited amount of time; but he takes issue with
the claim that we can occupy that view for long, and with the claim that we ever have
good reason to occupy that view (other than by virtue of applying the Capacity criterion
in some particular case). Strawson’s discussion of the “cultivation of the purely objective
view” is highly revealing, though, and is a good place to start our assessment of the
plausibility of Claim (φ). Here is Strawson on the possibility of voluntarily cultivating the
objective view:
I have remarked also that the suspension of ordinary inter-personal attitudes and the cultivation of a purely objective view is sometimes possible even when we have no such reasons for it [i.e. via the application of the capacity criterion] as I have just mentioned. Is this possible also in the case of the moral reactive attitudes? I think so; and perhaps it is easier. But the motives for a total suspension of moral reactive attitudes are fewer, and perhaps weaker: fewer, because only where there is antecedent personal involvement can there be the motive for seeking refuge from the strains of such involvement; perhaps weaker, because the tension between objectivity of view and the moral reactive attitudes is perhaps less than the tension between objectivity of view and the personal reactive attitudes, so that we can in the case of the moral reactive attitudes more easily secure the speculative or political gains of objectivity of view by a kind of setting on one side, rather than a total suspension, of those attitudes. These last remarks are uncertain; but also, for the present purpose, unimportant.173
173 F & R, p. 86.
199
There are three things worth noting about this passage: the first involves a conceptual
clarification; the second involves our understanding of the reasons we might have for
coming to occupy the objective view; and the third concerns the potentially revealing
admission by Strawson that seems to be lurking within his discussion at this point.
Firstly, then, let us start with the conceptual clarification. Strawson is here discussing
“the cultivation of the purely objective view”. What might it mean for a view to be purely
objective? And, insofar as Strawson’s discussion is of a pure version of the objective
view, might this mean that my invocation of a ‘moderate’ version of that view is just, in
effect, to change the subject? This is a worry that should certainly be addressed, but it
does not amount to a substantial objection. A “pure” objective view is, I take it, an
objective view that is unadulterated by co-existing alongside any less-than-objective
view. It is to be contrasted, therefore, with a “mixed” view, that is partially objective and
partially non-objective. This distinction is quite orthogonal to the distinction between the
“stark” and “moderate” objective views. The distinction between the stark and moderate
variants of the objective view is a distinction with regard to what the contours and
constitution of the objective view actually are. The stark and objective variants present
answers to the question of what the objective view demands; they thereby also provide
answers to the question of what it is for the objective view to exist, full and
unadulterated. Whether the best understanding of the objective view is in its moderate or
stark variants is, therefore, a prior question to the question of whether that view can or
does exist in a pure or mixed form in a particular case. Hence, there is no problem with
postulating the existence of a version of the objective view that is both “moderate” and
200
“pure” (just as we could similarly imagine a view that was “stark”, though “mixed” or
“adulterated”). The two distinctions are orthogonal, cutting fully across one another.
Secondly, we can turn to the reasons that we might have for coming to occupy the
objective view. Strawson, as we see here, sees our reasons for adopting the objective
view (other than via the application of the capacity criterion) as being exhausted by the
motive of “seeking refuge from the strains of involvement”.174 But, as has been
demonstrated using the case of Graham Greene’s Whisky Priest in The Power and the
Glory, it need not be that the adoption of the moderate objective attitude grants relief
from involvement in the interpersonal case. Indeed, adoption of the objective attitude can
go hand-in-hand with a heightened or enhanced degree of human involvement.175 The
same applies in the moral case. Adoption of the moderate objective attitude of course
precludes the accessibility of easy attitudes of moral blame and condemnation. But the
adoption of the moderate objective attitude need not involve moral disengagement from
others. Rather, it can provide the occasion for a richer and more nuanced appreciation of
the moral standing of others. Moral attitudes conceived within the bounds of the
moderate objective attitude may be the very paradigm of sensitive moral engagement, not
least because it forecloses the possibility of easy recourse to the most visceral and
unthinking forms of blame and condemnation.
174 F & R, p. 80.
175 On this claim, see the discussion of Graham Greene’s The Power and the Glory at Section 2(D) above.
201
Thirdly, we can turn to the interesting semi-revelation which is buried in this part of
Strawson’s argument. Strawson here allows that “the tension between objectivity of view
and the moral reactive attitudes is perhaps less than the tension between objectivity of
view and the personal reactive attitudes.”176 This admission is rather more important than
it would, on its surface, appear. Indeed, further reflection reveals this to be a surprising
admission for Strawson to make, given his other commitments. For, if Strawson were
right in his claims about the nature of moral attitudes – if it really were the case that the
moral attitudes simply are “structural transformations” of interpersonal reactive attitudes
– then it would be difficult to see how this admission could be correct. Yet the admission
that the tension between objectivity of view and the moral reactive attitudes does not
seem particularly great is surely correct. It is correct, though, only because, as has been
argued above, there is no real tension between a moderate version of the objective
attitude and the simultaneous holding of (broad) moral attitudes. Insofar as Strawson
allows that moral reactive attitudes do not seem to be in great tension with the objective
view, he is making a very plausible claim about the phenomenology of moral attitudes.
But it is a claim that gains its plausibility if we understand moral attitudes as taking their
content from prior moral judgements, and if we see such attitudes as capable of existence
in the absence of desert-entailing reactive attitudes. That is to say, Strawson’s admission
here is an admission that is plausible only if his Claims (α), (β), (χ) and (γ) are
themselves false.
176 F & R, p. 86.
202
Hence, Strawson’s admission is a potentially destabilizing one. If moral attitudes really
can coexist with the objective attitude, the moral attitudes cannot be desert-entailing in a
way that any variant of the objective attitude would disallow (hence, we should reject
claims (β) and (χ)), and it cannot be the case that the objective attitude, as an internal
conceptual matter, itself disallows the existence of any kind of moral attitude (and hence
claim (γ) must be false). We may say that, in this instance, Strawson’s (highly plausible)
admission about moral phenomenology is itself in deep tension with his account of the
nature of moral attitudes.
This tension is all the more evident when Strawson makes the further point that “we can
in the case of the moral reactive attitudes more easily secure the … gains of objectivity of
view by a kind of setting on one side, rather than a total suspension, of those attitudes.”177
If moral attitudes really were essentially desert-entailing, in the manner suggested by
Strawson’s claims (β) and (χ), then it is difficult to see how we might, at one and the
same time, both hold that a moral attitude was appropriate (i.e. not “suspend” it), whilst
nevertheless “setting it upon one side”.
In talking setting a moral attitude “upon one side” Strawson would seem to have in mind
the position of simultaneously holding a moral attitude, whilst nevertheless not seeing it
as the basis for the kind of reactive interpersonal attitude which would itself conflict with
the objective attitude. But such a position is fully intelligible only if we distinguish
between, on the one hand, the broad moral attitudes such as disapprobation (which are
177 F & R, p. 86.
203
most closely related to moral assessment and moral judgement, as such) and, on the other
hand, the “affectively loaded” narrow moral reactive attitudes of blame and
condemnation. Once this distinction is allowed, Strawson’s remarks about “setting upon
one side” certain attitudes within the moral domain are unproblematic. But, unless we do
allow this kind of internal complexity in the domain of the moral attitudes (and thereby
reject claims (β) and (χ)), the more nuanced moral attitudes that Strawson is here
describing simply cannot be conceptually coherent.
Thus, either Strawson’s conception of those attitudes is mistaken or his admissions here
must be rejected out of hand. Needless to say, what should be jettisoned here is the
conception of the moral attitudes that creates this needless tension with an abundantly
plausible account of the real complexity of the phenomenology of moral attitudes.
Indeed, the existence of this tension should itself provide good reason to reassess the
conception of the nature of the moral attitudes which is embedded at the centre of
Strawson’s approach.
Strawson himself seems to detect that he has here got himself into a dialectical pickle,
and that he has ended up, through a piece of admirable felicity to the real phenomenology
of moral attitudes, in undermining the background plausibility of his underlying theory of
the nature of those attitudes. That he recognizes the tricky position into which he has
backed himself is, I take it, what drives him to the sheepish semi-throwaway claim that
“these last remarks are uncertain… and unimportant”.178 For, if he allowed that his
178 F & R, p. 86
204
remarks were true, then he would have begun to unravel the account of the moral
attitudes on which the second ‘movement’ of his overall argumentative strategy relies.
G. Strawson’s Second Question: Moral Attitudes and The Human Commitment
Strawson’s remarks about our ability to “set aside” the moral attitudes provide an
interesting moment at which he seems to have a partial intimation of the problems
generated by his account of the moral attitudes. But, given that he puts no ultimate weight
on these worries, let us, like Strawson, put them aside as “uncertain and unimportant” and
move on to the main remaining elements of Strawson’s argument. We may recall, from
Section 2(F), that the first part of Strawson’s argument was concerned to give an answer
to the question of whether “the acceptance of the truth of the thesis of determinism …
would or should lead to the decay or repudiation of … the reactive attitudes.”179 His
answer, as we know, was that it would not. Thus, in answering this question in the
negative, Strawson sought to buttress the plausibility of a strong reading of his account of
the exemptions and excuses with regard to the personal reactive attitudes.
In the second movement of his argument, Strawson now turns to ask the analogous
question with regard, this time, to moral (or, as he has it, ‘vicarious’) reactive attitudes:
What concerns us now is to inquire, as previously in connection with the personal reactive attitudes, what relevance any general thesis of determinism might have to their vicarious analogues.180
179 F & R, p. 80.
180 F & R, pp. 86-7.
205
Again, as in the case of the personal reactive attitudes, Strawson’s claim is that the
“general thesis of determinism” (or, for that matter, any kind of general theoretical or
philosophical view) could and should have no effect on the existence and legitimacy of
the moral attitudes. In arguing for this second conclusion, Strawson is thereby seeking to
give the case in support of his Claim (φ) – that is, the claim that we should give a strong
reading of the role of the Capacity criterion as a condition for granting an exemption
from any of the set of moral attitudes, such that the Capacity criterion should be seen as
giving necessary and sufficient conditions for when an individual might legitimately be
exempted from subjection to any of the moral attitudes.181
Strawson’s argumentative strategy in answering his second question is almost identical to
the strategy which he pursued in answering his first main question. In other words, he
thinks that he can support Claim (φ), with regard to the capacity criterion giving
necessary and sufficient conditions for exemption from any of the moral attitudes, for the
same sorts of reasons that he previously thought would work in support of the claim that
the capacity criterion gave necessary and sufficient conditions for legitimate exemption
from any of the personal reactive attitudes. We may recall, from Section 2, that
Strawson’s strategy in answering his first question in the negative had two distinctive
parts, as follows:
(A) The appeal to “the Inescapability of the Human Commitment” (see the
discussion of Section 2(G)), backed up by a related appeal to the
181 For Claim (φ), see Section 2(E) above.
206
ineliminable features of “the General Framework of Human Life” (see the
discussion of Section 2(H)). This element of Strawson’s strategy stresses
the human impossibility of giving up the reactive attitudes.
(B) The Appeal to “the Gains and Losses to Human Life” (see the discussion
of Section 2(I)). This element of Strawson’s strategy stresses the
undesirability and/or irrationality of giving up the reactive attitudes.
Strawson’s strategy with regards to the moral attitudes has the same elements, albeit that
they are presented in a different sequence: as he puts it, “the answers once more are
parallel; though I shall take them in a slightly different order.”182 In this case, he begins
with an analysis of the rationality of suspending the moral attitudes, and then turns to a
fuller examination of the possibility of suspending such attitudes. Here, though, as with
his arguments regarding the personal reactive attitudes, the arguments in terms of the
rationality of suspending the relevant attitudes are really ‘fallback’, secondary arguments,
officially rendered unnecessary by Strawson’s ‘official’ position, which stresses the
impossibility of giving up such attitudes.
In Section 2(F), we distinguished between different varieties of incompatibilist or
sceptical views with which Strawson’s view might be seen as standing in opposition. On
fully incompatibilist or sceptical views, all reactive and moral attitudes are
182 F & R, p. 87.
207
inappropriate.183 But, on views which we may describe as ‘Hybrid views’, only the sort
of agency required for claims of desert, and not agency per se, is held to be inconsistent
with an understanding of human agency as a natural phenomenon.184 Hence, Hybrid
views rule out only the narrow, desert-entailing moral attitudes. Such views are not
inconsistent with moral appraisal, and nor with the broad moral attitudes that go
alongside it. Given this, as in the case of the personal reactive attitudes, the dialectically
most significant opponent for Strawson’s position is the advocate of Hybrid, or ‘Type-2’,
positions. For, in order to counter such views, Strawson must show that his appeals to the
impossibility and/or the irrationality of the rejection of the moral attitudes holds up when
the moral attitudes in question are only the narrow, desert-entailing attitudes, rather than
the whole set of broad moral attitudes. Unless his arguments can work in the case of the
narrow moral attitudes, Strawson will be without resources to counter the implications of
Hybrid views for the case of moral attitudes.
Thus, in assessing the plausibility of Strawson’s claim (φ), we should take Strawson’s
opponent to be the advocate of some variant of the Hybrid View. The Hybrid View
presents a clear and selective threat to the plausibility of Strawson’s claim (φ), thereby
undermining the negative answer which Strawson would like to give to his ‘Second
Question’. On such a view, there are, indeed, pace Strawson, consequences of general
theoretical views for the survival and legitimacy of certain moral attitudes.
183 See the discussion of ‘Type-1’ views in Section 2(F).
184 See the discussion of ‘Type-2’ views in Section 2(F).
208
H. Naturalism, Practice and “the Consequence of the General Thesis”
Before moving on to his claims about the irrationality and impossibility of rejecting (any
of) the moral attitudes on the basis of general philosophical considerations, Strawson first
makes a number of claims about the actual consequences of general philosophical views,
and about the reasons that people actually have for suspending certain attitudes.
Strawson’s approach here is a highly naturalistic one – he wants to interrogate the reality
of the attitudinal phenomena, so as to defuse the philosophical threat to moral attitudes.
This is all of a piece with his general “naturalistic compatibilist” approach. However,
some of the claims that Strawson makes here are rather implausible. For example,
Strawson tells us that:
First, we must note, as before, that when the suspension of such an attitude or such attitudes occurs in a particular case, it is never the consequence of a belief that the piece of behaviour in question was determined in a sense such that all behaviour might be, and, if determinism is true, all behaviour is, determined in that sense.185
This claim of Strawson’s is unsupported, and is surely false if we take seriously the
reports of many incompatibilists who say that they do suspend certain moral attitudes for
such reasons. In effect, Strawson here is suggesting that there are no honest
incompatibilists or sceptics. It is an unattractive philosophical methodology to assume
that one’s opponent cannot mean what he says, but Strawson here suggests that no
incompatibilist really ever does suspend (any of) his moral attitudes for general or
theoretical reasons. At any rate, Strawson offers no argument for why moral attitudes
cannot, as a matter of fact, ever be suspended as a consequence of the application of the
capacity criterion. Indeed, to even suggest that this is so sits very oddly with his earlier
185 F & R, p. 87.
209
admission that “the cultivation of the purely objective view” is possible as an act of
will.186 If it is open to us through sheer volition, then surely this makes us all the more
able to suspend certain reactive moral attitudes on the basis of deep-seated and
fundamental general philosophical beliefs.
Strawson goes on to make an intriguing claim that, whilst unsupported, might
nevertheless be read (perhaps mischievously) as suggesting that Strawson himself should
not be unamenable to a Hybrid variant of an incompatibilist position:
For it is not a consequence of any general thesis of determinism which might be true that nobody knows what he’s doing or that everybody’s behaviour is unintelligible in terms of conscious purposes or that everybody lives in a world of delusion or that nobody has a moral sense, i.e. is susceptible of self-reactive attitudes. In fact, no such sense of ‘determined’ as would be required for a general thesis of determinism is ever relevant to our actual suspension of moral reactive attitudes.187 There are a number of distinct claims bundled together here. Even the ‘generalized’ or
‘crude’ incompatibilist (i.e. the holder of a Type-1 view) need not deny that individuals
lack certain kinds of self-knowledge (that “nobody knows what he’s doing”), and nor
need he claim that individual behaviour is unintelligible in terms of conscious purposes,
nor that nobody has a moral sense (in Strawson’s sense). What the general incompatibilist
does hold, though, is that, in an important sense, we do live in a world of delusion,
because, although we may be aware of what we are doing, and although we have self-
reactive attitudes, we are not genuinely agents at all. What we have with regard to our
‘agency’ are appearances without the associated underlying phenomena (that is, genuine
186 F & R, p. 86. See the discussion of Section 3(F).
187 F & R, p. 87.
210
agency of the kind that makes us proper objects for moral assessment and moral attitudes)
with which we associate those appearances.
Nevertheless, setting aside the position of the ‘Type-1’ theorist, the interesting thing in
this regard about the Hybrid or ‘Type-2’ theorist is that he can agree with all of
Strawson’s claims here. She can allow that we (often) know what we’re doing, and that
our behaviour is intelligible in terms of conscious purposes, and that we have a moral
sense. She can also deny that these appearances are in any way delusional, and allow that
they are enough to get genuine agency, and along with it the conditions for moral
appraisal, off the ground. What the Hybrid theorist denies is that this is enough for the
legitimation of the narrow, desert-entailing moral reactive attitudes. Where Strawson is
surely guilty of a “non sequitur of numbing grossness”,188 is in moving from the claim
that the truth of determinism would not make us deluded, or incapable of acting for
reasons, or destroy our moral sense, or make our actions unintelligible, to the entirely
separate claim that “no such sense of ‘determined’ … is ever relevant for our actual
suspensions of moral reactive attitudes.”189 This latter claim simply does not follow, for,
as we have seen, the proponent of the Hybrid view can, without making an error of
reasoning, agree with Strawson that the truth of determinism would not activate any of
these familiar, capacity-based ‘exempting conditions’, while nevertheless denying that
this shows that determinism is therefore irrelevant with regard to the suspension of the
moral attitudes.
188 To use a phrase that Strawson uses of Kant in P. F. Strawson, (1966), The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason, (London: Routledge), at p. 137.
189 F & R, p. 87.
211
In fact, Strawson is here guilty of begging the question against his opponent. What he
needs to show is that only the familiar exempting conditions, associated with the Capacity
criterion, are relevant for the suspension of the moral reactive attitudes. For it is not that
the proponent of the Hybrid view (nor the sophisticated Incompatibilist) need think that
the truth of determinism undermines the appropriateness of certain attitudes via its effect
on our general capacities. If this was the view of the Hybrid theorist, then the proponent
of the Hybrid view would not, in fact, be denying claim (φ) at all, rather than giving an
account of one particular way in which the capacity criterion might fail to be met. But the
proponent of the Hybrid view, insofar as he rejects claim (φ), disagrees with Strawson’s
assumption that the only grounds for exemption from (any of) the moral attitudes is via
activation of the capacity criterion. Thus, Strawson here, in arguing against his opponent,
assumes the truth of precisely the conclusion for which he needs to argue. He is guilty
here, therefore, of an unfortunate exercise in circular reasoning.
In fairness to Strawson, he later came to acknowledge the question-begging nature of this
part of his argument. In replying to the arguments of Rajendra Prasad, in his ‘Reactive
Attitudes, Rationality and Determinism’,190 Strawson was admirably concessive:
For example, he [Prasad] shows conclusively that one of my arguments against the relevance of the thesis of determinism to the question of moral assessment and responsibility is either invalid or question-begging. I argue that the truth of determinism would not entail the presence in any particular case of one or more of the normal ‘inhibitors’ of moral reactions. But do I contend that such presence is a merely sufficient
190 Rajendra Prasad, (1995), ‘Reactive Attitudes, Rationality and Determinism’, in Sen and Verma (eds.), (1995), ibid., pp. 346-76.
212
or also a necessary condition for the inhibition? If I contend only the former, the argument is inconclusive; if the latter, it begs the question.191
Strawson is surely right to say that determinism poses no threat to the intelligibility of our
actions, to the existence of our moral sense, or to our self-knowledge. He is so keen to
avoid giving ground to any “thesis of determinism”, because the kind of harsh and crude
incompatibilism that he seems so eager to avoid is a view that itself denies these plausible
claims of Strawson’s. Perhaps the possibility of a Hybrid version of Incompatibilism had
not occurred to Strawson. If it had, he could have seen that one can admit the salience of
general philosophical views for the potential suspension of (some of) the moral reactive
attitudes without thereby thinking that agency, intelligibility, moral sense and self-
knowledge are themselves under threat. The Hybrid view allows us to acknowledge the
plausibility of Strawson’s positive claims here, without going all the way to embracing
his unsupported conclusion. The Hybrid view allows us to avoid throwing out the baby
with the bathwater; but, unlike Strawson’s view, it does allow us to throw out the
bathwater.
Strawson continues his discussion by invoking a comparison of the inter-personal with
the moral case, with regard to the possible “decay or repudiation” of the reactive
attitudes:
Second, suppose it granted, as I have already argued, that we cannot take seriously the thought that theoretical conviction of such a general thesis would lead to the total decay of the personal reactive attitudes. Can we then take seriously the thought that such a conviction – a conviction, after all, that many have held or said they held – would
191 P. F. Strawson, (1995), ‘Replies’, in Sen and Verma, ibid., p. 430.
213
nevertheless lead to the total decay or repudiation of the vicarious analogues of these attitudes?192
Given that I have argued in Section 2 that we can take seriously the thought that that
general theoretical convictions could lead to the repudiation of the personal reactive
attitudes, we might here wish not to grant Strawson his first claim. But let us, for the sake
of argument, grant him at least this. Nevertheless, even if we grant him this concession,
we can resist the extension to the moral case. Given our arguments against Strawson’s
account of the structural similarities of the moral and interpersonal cases (in Sections
3(B)-(D)), there is no reason to think that it might not be possible that moral attitudes,
even if not personal reactive attitudes, could come under threat from general
philosophical views.
Let us suppose, as on some variant of the Hybrid view, we could show that it was always
unfair to bring some desert-entailing attitude to bear on an individual.193 With regard to
the personal reactive attitudes, we may perhaps simply wish to conclude that this is just
so much the worse for ‘fairness’ in this case. We may conclude that it is simply the case
that we often face seemingly unfair reactive attitudes, and that this does not matter, since
the domain of reactive attitudes is in a sense outside of the ambit of norms of fairness.
But we may nevertheless not be able to make an analogous move in the moral case. If a
moral attitude is unfair, then this surely means that we have reason to reject it. Moral
attitudes, connected as they are with moral judgements, may be held to stricter standards
192 F & R, p. 87.
193 I offer an account of why this might be so Part Three, especially Section 4.
214
of appropriateness than are personal reactive attitudes. Even if we allow that
interpersonal reactive attitudes are just “part of human life”, outside of the scope of
normative requirements, it is surely not something that we can also say about moral
attitudes, insofar and precisely because such attitudes are normatively significant
attitudes of moral assessment.
Hence, we have reason to resist the further claims, relating to the relation between the
moral and non-moral reactive attitudes, made here by Strawson in this regard:
I think that the change in our social world which would leave us exposed to the personal reactive attitudes but not at all to their vicarious analogues, the generalization of abnormal egocentricity which this would entail, is perhaps even harder for us to envisage than the decay of both kinds of attitude together.194
There are three points to be made here. Firstly, adoption of a Hybrid view, and its
associated moderate objective attitude, would not involve giving up on the realm of moral
judgements and attitudes altogether, but only on the desert-entailing subset. Nor would it
involve giving up on moral demands. Given our rejection of claims (α), (β), (χ), and (γ),
we can resist the characterization of the moral whereby the repudiation or decay of the
narrow reactive attitudes could only possibly be achieved by a simultaneous repudiation
of moral demands, as such, and of the whole moral domain.
Secondly, and consequently, one need not think that the “change in our social world”
which such an adoption of the moderate objective attitude would enjoin would involve
anything even approaching “abnormal egocentricity”. Consider Graham Greene’s
215
Whisky Priest, languishing in his jail cell – egocentricity might have been better served
by an easy recourse to the familiar and self-justifying attractions of easy moral blame,
rather than the more difficult and potentially personally costly adoption of the moderate
objective attitude.
Thirdly, I have here been concerned only to show the implausibility of Strawson’s claim
that the moral and personal varieties of reactive attitude must always stand or fall
together, rooted as this claim is in Strawson’s problematic metaethics. But this concern
with conceptual clarification, and with the explication of how moral attitudes may face
legitimizing conditions not faced by interpersonal attitudes, does not mean that, as a
matter of fact, I disagree fundamentally with Strawson’s final claim. The consequences of
the Hybrid view are, indeed, that narrow, desert-entailing attitudes all lapse together,
whether or not those attitudes are distinctively moral in character.
I. Rationality, Responsibility and the Moral Attitudes
As ever when he is addressing the question of the rationality of the suspension of the
reactive attitudes, whether those attitudes are moral or merely ‘personal’, Strawson
begins by highlighting the way in which he takes such a question to be essentially
hypothetical. As in the case of the personal reactive attitudes, Strawson thinks that we are
addressing the question of the rationality of an unreal and unattainable choice:
Finally, to the further question whether it would not be rational, given a general theoretical conviction of the truth of determinism, so to change our world that in it all these attitudes were wholly suspended, I must answer, as before, that one who presses 194 F & R, p. 87.
216
this question has wholly failed to grasp the purport of the preceding answer, the nature of the human commitment that is here involved: it is useless to ask whether it would be rational for us to do what it is not in our nature to (be able) to do.195
It would seem that Strawson here holds that the foregoing remarks about our practices in
suspending the reactive attitudes, as discussed above in section 3(H), constitute an appeal
to “the nature of the human commitment”. As we have seen, though, however one wishes
to describe this line of argument, we cannot regard it as successful – it is, as Strawson
himself allows, either “invalid or question begging”.196 Moreover, following the
argument of Section 2(G), it is implausible to think that, even if there might be a human
commitment to avoiding the stark objective attitude, that there is any kind of “human
commitment” to always being prepared to deploy the full range of reactive attitudes,
including the desert-entailing moral attitudes.
However, setting these issues aside, it is worth addressing the question of whether
Strawson really is correct in saying that it is useless to question the rationality of doing
“what is not in our nature to (be able to) do”.197 Let us therefore grant for the sake of
argument that, due to some putative “human commitment”, it is not in our nature to (be
able to) adopt the moderate objective attitude. Would the question of the rationality or
desirability of that attitude therefore be useless or otiose? I suggest that it would not.
There may be some things that are not “in our nature” to do, but which we may
nevertheless achieve given sustained practice and training. It is not in my nature to pick
195 F & R, p. 87. [Strawson’s italics.]
196 Strawson, (1995), ibid., p. 430.
197 F & R, p. 87.
217
up foreign languages easily, but that does not mean that it is useless for me to address the
question of whether it might be rational or desirable for me to learn Spanish or German.
However, Strawson may have something stronger in mind – that is, courses of action
which are not just “not in our nature” in this weak sense, but which are quite completely
beyond our powers. No human being has travelled beyond the solar system or lived for a
thousand years, and it seems quite beyond our powers to do either. Nevertheless, it is
surely an intelligible question to address the desirability of achieving these goals. If we
were to decide that it would be rational for us to pursue these goals, then we might now
be able to start on courses of action (research into the technology of propulsion, or the
biochemistry of aging, for example) that could in the future begin to bring us close to
achieving these ultimate goals. So too with our attitudes: even if we cannot change them
all of a sudden, and all at once, human history is full of examples where people have
endeavoured to change their “attitudinal repertoire” through careful and gradual
processes of self-transformation. We see such processes at work in the spiritual exercises
of St Benedict or St Ignatius Loyola, or in Pascal’s commitment to the production of
religious belief in his later self, as a consequence of accepting his ‘wager’.198 Only if
Strawson thinks that we have no significant capacity for self-transformation of this kind
can he hold fast to his claim that it is useless to question the rationality or desirability of
adopting an attitudinal stance that is, as a matter of fact, currently beyond us.
198 On the use of forms of training in order to inculcate particular beliefs or attitudes in oneself, see, in a Christian context, St. Benedict, (1986) [6th century], The Rule of St Benedict, (Liturgical Press); St. Ignatius of Loyola, (1996) [1524], “Spiritual Exercises,” in his Personal Writings, (London: Penguin); and Blaise Pascal, (1995) [1670], Pensées, (London: Penguin). On earlier, Greek technologies for the self-transformation of one’s “attitudinal repertoire”, see William Vernon Harris (2002), Restraining Rage: The Ideology of Anger Control in Classical Antiquity, (Cambridge: Mass.: Harvard University Press).
218
Moreover, questions of the rationality and desirability of some course of action can be
usefully pursued even when some course of action is utterly unattainable. Even in
situations where some courses of action are wholly beyond our powers, even through
working gradually towards a distant goal, we may still nevertheless take an interest in
making our behaviour approximate to some unattainable target. Let us assume that it is
beyond human reason to avoid all logical errors in reasoning over the course of a human
life. Nevertheless, we may take it as a valuable goal for us to come as close to this limit
as is possible for us. It can be useful to know what would be maximally rational for us, if
only as a way of orienting our efforts within the range of possibilities that really are open
to us.
Finally, with Kant, we may think that one of the tasks of philosophy is to tell us what we
may hope for, as well as the tasks of telling us what we should believe and what we
should do.199 Questioning the rationality and desirability of certain courses of action can,
at the very least, orient our hopes, even if it cannot inform our action due to our
compromised capacities. It is surely not idle to address the question of the rationality of
our attitudes, even if we cannot, as we are, shake them off; or, if addressing such
questions is, indeed, idle, then it is idle only in the same way that hope is idle.
199 As with Kant’s delineation of the three central questions of philosophy in the “Canon of Pure Reason” of his (1999), [1781/1787], Critique of Pure Reason, ed. and trans. Paul Guyer and Allen Wood, (Cambridge: Cambridge University Press).
219
Thus, if we grant – as we should – that there is a real and useful question to be answered
regarding the rationality of suspending the moral reactive attitudes, how should we
answer that question? Here is Strawson’s proposal, which is parallel to that which he
offered in the case of the reactive attitudes:200
To this I must add, as before, that if there were, say, for a moment open to us the possibility of such a godlike choice, the rationality of making or refusing it would be determined by quite other considerations than the truth or falsity of the general theoretical doctrine in question. The latter would be simply irrelevant; and this becomes ironically clear when we remember that for those convinced that the truth of determinism nevertheless really would make the one choice rational, there has always been the insuperable difficulty of explaining in intelligible terms how its falsity would make the opposite choice rational.201
Here, Strawson lays out one of the most puzzling aspects of his view. If we hold some
general view, ψ, then surely we should also hold that its consequences are true. Let our ψ
in this case be any general philosophical view, whether Incompatibilist or Hybrid, that
denies that human beings have got the kind of capacities that could legitimate claims of
desert, or make desert-entailing attitudes appropriate. Thus, if we hold ψ, we cannot at
the same time hold any belief that entails that human beings can be objects of claims of
desert, or be legitimately subjected to desert-entailing attitudes. It presumably thereby
becomes irrational for anyone who holds ψ to be true to hold any desert-entailing
attitudes, insofar as such attitudes make sense only if their objects are proper objects of
claims of desert.
200 Strawson’s earlier treatment of the rationality of suspending reactive attitudes is discussed above in Section 2(I), on ‘Rationality, the Reactive Attitudes and “the Gains and Losses to Human Life”’.
201 F & R, pp. 87-8.
220
We must therefore enquire as to what Strawson can have in mind when he speaks of
“quite other considerations than the truth or falsity of the general theoretical doctrine
[(e.g., a doctrine such as ψ)] in question”.202 There seems to be no available construal of
this claim other than one that takes Strawson to be claiming that we should decide upon
which attitudes to evince on the basis of considerations of consequential or prudential
rationality. Now, this view would seem particularly implausible if one thought that it
could be applied to individual attitudes, case by case, in a piecemeal way. A
consequentialist calculus of individual attitudes would render the landscape of our
emotional lives quite unrecognizable; indeed, such a view would seemingly rob different
kinds of attitudes of their distinctive identities, for, rather than being appropriate or
inappropriate responses to different sets of circumstances, all such attitudes would be no
more than mental performances designed to further our goals or increase our prudential
advantage.
Strawson’s position would be much more plausible if it was instead understood as
relating, not to individual attitudes considered case-by-case, but rather to our decision to
engage with, or abandon, an integrated framework of reactive and moral attitudes,
considered as a unified system. One could answer the question of whether one should
engage with others at all – be that personally, emotionally or morally – in terms of
consequential considerations without thereby undermining the coherence or specificity of
the particular attitudes within that general framework. Strawson’s appeal to
consequentialist rationality could, thereby, be seen as a plausible response to the sparse,
202 F & R, p. 87.
221
desert landscape that might be offered up by the cruder versions of Incompatibilism
(i.e.‘Type-1 views’), which would suggest removing human action and human
engagement from our picture of the world altogether, through the adoption of a stark
objective view.
Nevertheless, a view of this kind gets no real purchase against Hybrid versions of
incompatibilism. For what the Hybrid view itself entails is the possibility of the adoption
of a moderate objective view. And the possibility of the objective view existing in a
moderate variant itself means that we need not treat human attitudes as an indivisible
totality that can only be accepted or rejected as a unified system. In denial of this
systematicity, Hybrid versions of incompatibilism suggest that we should retain the
greater parts of our attitudinal and moral landscape, excising only those aspects of our
beliefs and attitudes which clash directly with the denial of the legitimacy of claims of
desert. Against a view of this kind, Strawson’s appeal to consequential rationality seem
unmotivated and implausible.203
We can, therefore, reject Strawson’s somewhat cryptic claim that “quite other
considerations” should lead us to retain the narrow moral attitudes, even if we were to
endorse the truth of some general Incompatibilist or Hybrid view. If we hold that the
Hybrid view is true, then it would be rational for us to excise the desert-entailing moral
attitudes from our attitudinal repertoire, in the same way as it would be rational for us to
excise the desert-entailing personal reactive attitudes.
203 One may also plausibly think that the (1) Desirability and (2) Anti-Consequentialist objections outlined in Section 2(I) with regard to the personal reactive attitudes apply also in the moral case.
222
Before setting issues of the rationality of our attitudes aside, though, and to pause in the
almost unrelenting criticism of Strawson’s arguments, we should recall the
extraordinarily insightful remarks to which Strawson turns immediately after these
comments about rationality. Strawson tells us that: “for those convinced that the truth of
determinism nevertheless really would make the one choice [i.e. the suspension of
(certain) moral attitudes] rational, there has always been the insuperable difficulty of
explaining in intelligible terms how its falsity would make the opposite choice
rational.”204 One can, perhaps surprisingly, even whilst endorsing the Hybrid view, fully
endorse Strawson’s insightful point here, at least with regard to the narrow, desert-
entailing reactive attitudes. The Hybrid view denies that claims of desert are ever
appropriate, given the truth of determinism, or at least given the belief that human agency
should be seen as a natural phenomenon like any other. But this is not to say that, if
determinism were false, or if human agency were somehow a purely rational
phenomenon, existing, as Kant held in the Third Antinomy,205 outside of the bounds of
time and space, that claims of desert would, somehow, be appropriate or legitimate.
It is indeed difficult to see how claims of desert could ever be legitimate, even if human
agency were not a process in the natural world. Thus, one might wish to conclude that,
whether or not determinism is true, and whether or not human agency is a natural
phenomenon, we should reject the desert-entailing moral and reactive attitudes. The
204 F & R, p. 87-8.
205 See the “Third Conflict of the Transcendental Ideas” of “The Antinomy of Pure Reason” in Kant, ibid., The Critique of Pure Reason (A444/B472-A451/B479).
223
moderate objective attitude may be the only rationally justifiable attitude for us to adopt
towards ourselves and others, whatever the proper understanding of the metaphysics of
agency turns out to be.206
J. “A Crude Opposition of Phrase Where We Have a Great Intricacy of Phenomena”
Towards the end of the penultimate section of his discussion, Strawson has a wonderful
semi-confessional, and partially concessive, moment:
I am aware that in presenting the argument as I have done, neglecting the ever-interesting varieties of case, I have presented nothing more than a schema, using sometimes a crude opposition of phrase where we have a great intricacy of phenomena. In particular the simple opposition of objective attitudes on the one hand and the various contrasted attitudes which I have opposed to them must seem as grossly crude as it is central.207
This comment can stand as an excellent summary of some of the criticisms that have here
been levelled at Strawson’s view. What Strawson misses, in pursuing his schematic
divisions between, on the one hand, the objective and participant attitudes and, on the
other hand, the objective and moral attitudes, is the existence of the moderate variant of
the objective attitude. Moreover, the damage done to the plausibility of Strawson’s view,
with regard to the crudeness of the ‘objective attitude’, is actually rather greater than
Strawson allows. The ways in which this “crudeness” undermines Strawson’s position
has been a constant recurring theme of this discussion, and I need not enumerate the
manifestations of this problem here again. What is striking, though, is that Strawson has
206 For present purposes, I simply offer up this suggestion, without developing or defending it. For a fuller discussion, see Part Three, especially Sections 4 and 6.
207 F & R, p. 88.
224
an inkling of the source of deep philosophical problems faced by his theory, even if he
seems not to be aware of the severity of those problems.
Strawson’s treatment of two kinds of ‘intermediate case’ are quite revealing. These two
kinds of intermediate cases involve, firstly, the relationship between parents and their
children and, secondly, the relationship between psychiatrists and their patients. I shall
pass over the second case for our present purposes, not least because the case of
psychotherapy brings with it a full raft of its own peculiarities and difficulties, and
instead concentrate on Strawson’s treatment of the central and illuminating case of
parent-child relationships. Strawson’s treatment of this issue shows some particular
problems with his view, and is also interesting on its own terms:
Let me pause to mitigate this crudity a little, and also to strengthen one of my central contentions, by mentioning some things which straddle these contrasted kinds of attitudes. Thus parents and others concerned with the care and upbringing of young children cannot have to their charges either kind of attitude in pure or unqualified form. They are dealing with creatures who are potentially and increasingly capable both of holding, and being objects of, the full range of human and moral attitudes, but are not yet truly capable of either. The treatment of such creatures must therefore represent a kind of compromise, constantly shifting in one direction, between objectivity of attitude and developed human attitudes. Rehearsals insensibly modulate towards true performances. The punishment of a child is both like and unlike the punishment of an adult.208
Now, I have no wish to deny that sometimes people treat their children in a way that
involves looking on those children as deserving subjects of punishment. Some parents
take up desert-entailing (narrow) reactive attitudes towards their children, viewing them
as individuals capable of being the proper objects of such attitudes. In such cases,
children may be blamed or punished in much the same way as adults are often blamed
208 F & R, p. 88.
225
and punished. I also have no wish to deny that sometimes parents (or other adults) treat
their children in the second way that Strawson suggests – as a sort of pantomime of
desert-entailing treatment, as a ‘rehearsal’ for the ‘real’ thing. In this ‘simulation’ of full,
robust, adult treatment, what we have is an instrumental attempt to produce certain kinds
of future response and behaviour in those children: a form of training. This kind of
instrumentalizing, forward-looking treatment is of the same kind as the conception of
blame advanced by instrumentalist compatibilists like Moritz Schlick and J. J. C.
Smart.209 Given that both kinds of treatment exist, Strawson is surely right that often, in
parent-child relationships, what we have is a compromise between the two, or an almost
imperceptible modulation between one and the other.
What I want to suggest, though, is that these options do not exhaust the phenomenology
of what goes on in parent-child relationships, when parents hold their children to certain
normative expectations. Strawson leaves out the most interesting kind of case, which fits
under neither the description of full ‘desert-entailing’ treatment, nor the description of
forward-looking ‘training’. In this other kind of case, which is equally familiar to us,
children are held to normative expectations in a genuine rather than simulated way, but
without the existence of (or threat of being subject to) desert-entailing attitudes if those
normative standards are not met. What Strawson overlooks is the possibility of requiring
a certain standard of behaviour from others, but without the latent threat of the
withdrawal of good will (or activation of ill will) if that standard of behaviour is not met.
209 I shall discuss the views of Schlick and Smart about blame more fully in the following section, (3(K)).
226
Parent-child relationships can often be clear exemplars of the existence of the moderate
variety of the objective attitude. The parent’s underlying commitment to the child may
place even the temporary withdrawal of good will outside of the range of possibility; but
that parent may nevertheless regard their child (once it has reached a certain level of
development) as a moral being, and thereby as properly subject to moral demands and
expectations, and properly open to moral assessment in light of the child’s success or
failure in living up to those normative expectations. Needless to say, when there is a
failure to live up to such expectations, that may generate a reaction of disappointment or
sadness, but it need not generate the kind of narrow, desert-entailing reactive attitudes
that Strawson focuses upon.
Rather than seeing intra-familial relationships as a compromise between, on the one hand,
being open to attitudes such as resentment and, on the other, indulging simply in a
process of training, we may instead view them as (at least sometimes) showing how to
move beyond this false dichotomy. Our choices, in such relationships, are not between
being open to moral considerations (and thereby also open to desert-entailing reactive
attitudes) or excluding such reactive attitudes (and thereby also stepping beyond the
bounds of morality). Instead, we can have moral engagement without being open to the
narrow subset of desert-entailing reactive attitudes. To see such a possibility in action we
need not think there is anything unusual or exotic about the attitudes of Graham Greene’s
Whisky priest, having his epiphanic moment in his Mexican prison cell.210 Other
examples are more quotidian, and closer to home. The exemplar of one variety of intra-
210 See the discussion of The Power and The Glory in Section 2(D).
227
familial relationships shows that we can all, so to speak, have morality without
resentment.
Thus, Strawson’s presentation of the available options as either ‘true performance’ or
‘rehearsal’ is importantly misleading. For Strawson, we have a ‘true performance’ of
holding another individual to normative standards only when we are open to the narrow
desert-entailing reactive attitudes. But this is an inappropriate standard for what is to
count as a ‘true performance’ of moral or emotional engagement with others. So, when
we exclude narrow reactive attitudes, this does not mean that what we’re therefore doing
is merely rehearsing or pretending to do something. Our performance may be a genuine
one within the moral domain, without measuring up to Strawson’s (mistaken) standard
for what is to count as a ‘true performance’. Once one allows that the price for
abandoning the full set of reactive attitudes need not be complete human isolation or the
abandonment of the domain of the moral, then the power of Strawson’s position, which is
dependent on a simple dichotomy between the objective attitude and ‘participant’
attitudes, evaporates. This is why, in the face of “a great intricacy of phenomena”, a
“crude opposition of phrase”211 can be so very misleading.
The important point about Strawson’s treatment of the case of children, and our attitudes
towards them, is that he here misses a case where the possibilities of the moderate
objective attitude are clearly dramatized. But he also, in discussing the case of children,
goes on to a discussion of determinism, which itself makes an odd kind of mistake about
211 F & R, p. 88.
228
the lessons that we should draw from the phenomenology of family relationships, and
from the ongoing modulation of parental attitudes towards children as those children
develop:
Suppose we try to relate this progressive emergence of the child as a responsible being, as an object of non-objective attitudes, to that sense of ‘determined’ in which, if determinism is a possibly true thesis, all behaviour may be determined, and in which, if it is a true thesis, all behaviour is determined. What bearing could such a sense of ‘determined’ have upon the progressive modification of attitudes towards the child? Would it not be grotesque to think of the development of the child as a progressive or patchy emergence from an area in which its behaviour is in this sense determined into an area in which it isn’t? Whatever sense of ‘determined’ is required for stating the thesis of determinism, it can scarcely be such as to allow of compromise, borderline-style answers to the question, ‘Is this bit of behaviour determined or isn’t it?’ But in this matter of young children, it is essentially a borderline, penumbral area that we move in.212
Strawson’s argument here seems to have something like this form:
(1) We tend to be readier to deploy the full range of reactive attitudes towards
children as they develop further. When they are at early stage of development, we
are more inclined to suspend many of the reactive attitudes.
(2) It would be absurd (“grotesque”) to think of the development of a child as
essentially involving a change from being ‘determined’ to not being ‘determined’
in his or her behaviour.
Therefore,
(3) The truth or otherwise of the ‘thesis of determinism’ has no bearing upon the
appropriateness (or otherwise) of the reactive attitudes.
212 F & R, p. 88.
229
This is an illicit form of argument, as it relies on what one might describe as a ‘bait and
switch’ technique. In effect, Strawson is arguing against a position that nobody has any
plausible reason to hold. In attempting to demonstrate (3), Strawson makes the
assumption that the truth (or otherwise) of the thesis of determinism could only be
relevant for the appropriateness (or otherwise) of the reactive attitudes if it gave a
necessary and sufficient condition for the appropriateness of such attitudes. That is, only
if one thought that the falsity of determinism was a necessary and sufficient condition for
the appropriateness of reactive attitudes towards some individual would one think that it
was appropriate to suspend such attitudes when and only when determinism did not
obtain.
Only on a deeply implausible view of this kind (let us call this the ‘Lock-Step View’)
would one have any reason to think that our tendency to suspend certain reactive attitudes
towards children (but not towards adults) showed that we have reason to believe that
children’s agency is determined, while that of adults is not. Thus, only on this
implausible ‘Lock-Step’ view would one have any reason to be disturbed by premise (2)
of Strawson’s argument. Strawson’s conclusion (3) thus holds only for this kind of
implausible ‘Lock-Step’ account of the relevance of determinism for the reactive
attitudes.
But most of those who hold an Incompatibilist or Hybrid view reject this ‘Lock-Step’
view. They hold that the truth of determinism is a sufficient condition for the
inappropriateness of certain reactive attitudes, but they also hold that there can be other
230
considerations that make (some or all) reactive attitudes inappropriate in certain
circumstances. Any plausible form of an Incompatibilist or Hybrid view can allow that
sometimes it is appropriate to suspend reactive or moral attitudes for reasons that have
nothing to do with the truth of determinism – for example, in cases where the agents in
question lack the requisite capacities for autonomous or reflective agency, as in the case
of young children. In short, such views can accept the weak reading of Strawson’s
capacity criterion, whilst rejecting its strong reading. On these more plausible views,
there is no need to accept the claim that the suspension of reactive attitudes towards
children could only plausibly be motivated by acceptance of the “grotesque” claim (2).
And only on the implausible assumptions of the straw-man ‘Lock-Step’ view could one
believe that Strawson’s conclusion (3) could possibly follow from his premises (1) and
(2). We should therefore conclude that this argument of Strawson’s is conspicuously
unsuccessful.
Moreover, Strawson’s argument is question-begging in a second way. If one accepts the
Hybrid view, or some form of Incompatibilism, one will typically accept that our current
practices, as regards the deployment and suspension of the reactive attitudes, may not be
in good order. Such views are, after all, revisionist with regard to current practice.
Therefore, one cannot take the fact that certain attitudes are currently viewed as
appropriate towards adults (and older children), but not towards younger children, as
evidence that such attitudes really are appropriate towards adults and older children. On
the Hybrid view, for example, we hold that all (narrow) desert-entailing attitudes are
inappropriate in all cases. But Strawson’s argument here suggests that we can draw
231
implications from a description of current norms of appropriateness to judgements of all-
things-considered normative appropriateness. But this is precisely what is denied by
revisionist views such as the Hybrid view. Thus, Strawson’s argument has no prospect of
convincing his philosophical opponents, given that he begs the question against them in
this significant respect.
K. Optimism, Pessimism, Moral Responsibility and the Objective Attitude
Let us turn now to Strawson’s concluding section, and the way in which he draws
together the strands of his argument. Before assessing the ultimate success or failure of
Strawson’s arguments for claim (φ), it is first incumbent upon us to address some
conceptual and terminological issues regarding the relationships between compatibilism,
incompatibilism, ‘optimism’, ‘pessimism’, and Strawson’s own distinctive position. The
Incompatibilist, or “pessimist” as Strawson calls him, denies the truth of Strawson’s
claim (φ), as the pessimist believes that we may have reason to exempt individuals from
(certain) moral attitudes on the basis of general philosophical considerations. The
proponent of the Hybrid view, by contrast, subscribes to the particular version of
incompatibilism that holds that the moral attitudes from which we should exempt
individuals, on general philosophical grounds, are all and only all of the narrow, desert-
entailing reactive attitudes. The “optimist”, by comparison, can (at least presumably)
accept that Strawson’s account of the Capacity criterion gives us an exhaustive set of
reasons for when and why we can exempt individuals from any of the set of moral
attitudes. The optimist, therefore, can accept claim (φ), as Strawson does.
232
Now, this raises the puzzling question of why it is that Strawson himself not a self-
described optimist. After all, Strawson is, in an important way, a kind of compatibilist,
albeit of a naturalistic or ‘practice-based’, rather than of a ‘theoretical’ kind. Why, then,
should he not describe his own view as a species of ‘optimism’, rather than as a purported
‘reconciliation’ between optimist and pessimist positions? The answer, which is to some
degree surprising, comes in this passage, where Strawson pulls together the different
strands of his position, and contrasts his view with the view of others:
And now we can try to fill in the lacuna which the pessimist finds in the optimist’s account of the concept of moral responsibility, and of the bases of moral condemnation and punishment; and to fill it in from the facts as we know them. For, as I have already remarked, when the pessimist himself seeks to fill it in, he rushes beyond the facts as we know them and proclaims that it cannot be filled in at all unless determinism is false. Yet a partial sense of the facts as we know them is certainly present to the pessimist’s mind. When his opponent, the optimist, undertakes to show that the truth of determinism would not shake the foundations of the concept of moral responsibility and of the practices of moral condemnation and punishment, he typically refers, in a more or less elaborated way, to the efficacy of these practise in regulating behaviour in socially desirable ways. These practices are represented solely as instruments of policy, as methods of individual treatment and social control. The pessimist recoils from this picture; and in this recoil there is, typically, an element of emotional shock. He is apt to say, among much else, that the humanity of the offender himself is offended by this picture of his condemnation and punishment.213
What is most striking here is Strawson’s description of the optimist as one whose view
about the compatibility of moral responsibility with determinism appeals to the social
efficacy of practices of blame and punishment. Strawson’s optimist takes a view whereby
we are to construe such attitudes and practices on a purely instrumental or
consequentialist basis. This description does capture the content of the views of a small
number of philosophers who would characterize their views as forms of compatibilism,
213 F & R, p. 89.
233
but only a rather small proportion of the total compatibilist population. The sort of
theorists who fall under Strawson’s description of “optimism” are the instrumental,
broadly utilitarian theorists of blame and punishment, such as Moritz Schlick and J.J.C.
Smart.214 But most compatibilist positions do not appeal to utilitarian or consequentialist
considerations in this way.
Most compatibilists, from Hobbes and Hume onwards, defend their position simply by
denying the incompatibilist claim that practices of condemnation or punishment
presuppose the existence of some kind of agential capacity which individuals could not
possess if determinism were true.215 So, despite the earlier identification of optimism
with compatibilism, the two positions are not the same. Standard compatibilism is a very
different position to the view that Strawson calls ‘optimism’; the latter view is, in fact, a
very specific kind of forward-looking, instrumentalist compatibilism. Or, to put things in
a different way, one might say that ‘optimism’, in Strawson’s sense, is a form of highly
non-standard compatibilism
This difference between optimism and compatibilism is a revealing one. Hume, whom we
shall treat as the paradigm case of a standard compatibilist, contrasts the “liberty of
214 See Moritz Schlick, (2002) [1930], Fragen der Ethik, (Frankfurt: Suhrkamp); and J. J. C. Smart, (1961), “Free Will, Praise and Blame,” Mind, 70, 291-306, reprinted in Watson, ed., (2003), ibid., pp. 58-71. See also Richard J. Arneson, (2003), “The Smart Theory of Moral Responsibility and Desert,” in Serena Olsaretti, ed., Desert and Justice, (Oxford: Clarendon Press), pp. 233-58.
215 For classic compatibilist views, see Thomas Hobbes, (1996)[1651], Leviathan, ed. Richard Tuck, (Cambridge: Cambridge University Press); and David Hume, (2000)[1739-40], A Treatise of Human Nature, ed. David Fate Norton and Mary J. Norton, (Oxford: Oxford University Press). See also James Harris, (2005), Of Liberty and Necessity: The Free Will Debate in 18th Century British Philosophy, (Oxford: Oxford University Press) and Paul Russell, (1995), Freedom and Moral Sentiment: Hume’s Way of Naturalizing Responsibility, (Oxford: Oxford University Press).
234
spontaneity” with the “non-existent liberty of indifference”, claiming that only the first,
and not the second, are needed in order for individuals to be viewed and treated as
morally responsible.216 Hume needed to make no appeal to the efficacy or desirability of
treating people as if they were responsible for, on his view, they really are responsible.
Indeed, the sort of consequentialist defence of moral responsibility mounted by
philosophers like Schlick and Smart is hard to motivate unless one is prepared to give a
certain amount of prima facie ground to the incompatibilist position. If there were no
truth in incompatibilism, then it is hard to see why one would not rest content with a form
of standard, Humean compatibilism. Only if moral attitudes and practices seem under
intense threat from determinism is there good reason to look for some alternative form of
defence for those attitudes and practices. Consequentialist compatibilism, or ‘optimism’,
gets off the ground only as a defensive move against the threat of incompatibilism. If
incompatibilism has no punch (because, for example, Hume was right that only the
compatibilist “liberty of spontaneity” were required for moral responsibility), then there
is no need for ‘optimism’.
There is, therefore, something rather curious about the way that Strawson draws the
distinction between optimism and pessimism. If one thinks, as Strawson seems to, that
these two are the main existing philosophical options, then one is ex ante writing-off the
plausibility of standard Humean compatibilism. If anything, optimism and pessimism
represent one’s options after one has given up on standard Humean compatibilism, or
once one has conceded that there is a large truth in incompatibilism. Once ground has
216 See David Hume, (2000) [1739-40], A Treatise of Human Nature, ibid., Book II, “Of the Passions”, Part 3, “Of the will and direct passions”, esp. Sections 1-2: “Of liberty and necessity”, pp. 257-65.
235
been given to incompatibilism one’s options are, so to speak, either to admit defeat (i.e.
‘pessimism’) or else to mount a rearguard action by virtue of appeal to a consequentialist
or instrumental account of how responsibility can still be saved (i.e. ‘optimism’). One
interpretation of what is going on here, then, is that, in a sense, Strawson does accept the
(partial) truth of incompatibilism. This would explain why he does not even consider the
plausibility of a standard Humean version of compatibilism. It would also place him in
the same position as Kant, who, after all, rejected compatibilism out of hand as giving us
(in Hume’s “liberty of spontaneity”) no more than “the freedom of a turnspit”,217 whilst
nevertheless rejecting incompatibilism in favour of a kind of practical compatibilism.218
The other possible hypothesis which would explain Strawson’s peculiarly idiosyncratic
conceptual delineations is, in effect, the very opposite of the supposition that Strawson is
a sort of closet incompatibilist. This alternative proposal is that Strawson’s view is, in
fact, much closer to standard compatibilism than he presents it as being in ‘Freedom and
Resentment’. By setting up the dialectical situation as involving an irresoluble stand-off
between optimist and pessimist, rather than the more usual stand-off between
compatibilist and incompatibilist, Strawson is able to ride to the rescue of the dialectical
stalemate by presenting a novel position which is neither optimist nor pessimist, in his
217 Kant claims in Chapter 3 (“On the Incentives of Pure Practical Reason”) of The Critique of Practical Reason [1788], reprinted in Mary Gregor, (ed.), (1996), Immanuel Kant: Practical Philosophy, that “… if the freedom of our will were merely the latter (psychological and comparative but not also transcendental, i.e. absolute), then it would at bottom be nothing better than the freedom of a turnspit, which, once wound up, also accomplishes its motions of itself” (Ak. 5: 957), p. 218.
218 For illuminating discussion of Kant’s particular variety of “practical compatibilism”, see Henry E. Allison, (1990), Kant’s Theory of Freedom, (Cambridge: Cambridge University Press); and also Christine M. Korsgaard, “Morality and Freedom” and “Creating the Kingdom of Ends: Reciprocity and responsibility in personal relations,” both in her (1996), Creating the Kingdom of Ends, (Cambridge: Cambridge University Press).
236
terms. But, as I have argued, standard Humean compatibilism is itself neither a version of
‘optimism’ nor a version of ‘pessimism’. And Strawson’s view, on close inspection, is
really perhaps not all that far from this well-worn view, albeit that he defends it in places
by means of appeal to unusual argumentative strategies (‘the Inescapability of the Human
Commitment’, ‘the Gains and Losses to Human Life’, etc.). That Strawson at least came
to be a standard compatibilist of the Humean kind, even if that was not his view at the
time of ‘Freedom and Resentment’, is something that he happily admitted over thirty
years after the original article, in the ‘Reply to Prasad’ (which is surely (as we have
already seen in Section 3(H) above) Strawson’s most honestly concessive piece of
writing on these issues):
Let me concede that we rightly regard an action as an appropriate subject for moral judgement only if we regard the agent as one who could have acted, and chosen to act, otherwise than he did; only if he acted freely. The question is: what does this mean? It certainly means that he was not subjected to overwhelming external or internal compulsion to act as he did. It means more: that he did not lack the material and mental resources (he was rich or strong or intelligent enough) to do otherwise; and that he had the opportunity to do otherwise (there were no obviously insuperable obstacles in the way of doing so). Besides these somewhat negative considerations, it means also, ideally, that he knew what he was doing, was aware of other possibilities and chose to do what he did in the light of his beliefs about the facts and his attitudes (including moral attitudes) and preferences. All this seems to be about as full a statement as could be required of what it is to act freely, and hence, in the relevant sense, though not perhaps in every sense, responsibly. But I find in none of this an explicit or implicit denial of a thesis of determinism, stated, for example, in such simple and familiar terms as “Every event has a cause”. […] If, in saying the above, I have dwindled into a mere compatibilist – in the company, say, of Hume – I am content with that.219
Here, we have Strawson arguing, very straightforwardly, that we are morally responsible
if we act freely; and we act freely if we act in accord with Hume’s “liberty of
spontaneity” – that is, “in the light of [our] beliefs about the facts and [our] attitudes
219 See Strawson, (1995), “Reply to Prasad,” in Sen and Verma (eds.), ibid., pp. 431-2.
237
(including moral attitudes) and preferences”. Such a view may be to some degree
plausible, but it certainly is not the kind of move beyond, or transcendence of, the
discourse of compatibilism and incompatibilism, as Strawson’s position is often taken to
be. It is simply a mainstream form of compatibilism. If we really are looking for a view
that acknowledges the partial truth of both compatibilism and incompatibilism, and in so
doing moves beyond both, then Strawson’s view is not a candidate for that job. Instead,
we would have to seek something more in the region of a Hybrid View.220
Now that the relationship between optimism, pessimism, compatibilism and
incompatibilism has been set out in its unexpected complexity, we are in a good position
to make better sense of Strawson’s rejection of “optimism” (that is, his rejection of
consequentialist or instrumental forms of compatibilism). As we shall see, this rejection
has two strands – one ‘conceptual’, the other ‘emotional’. He begins with the conceptual
element:
The reasons for this recoil [i.e. recoil from the optimist’s picture] – the explanation of the sense of an emotional as well as a conceptual, shock – we have already before us. The picture painted by the optimists is painted in a style appropriate to a situation envisaged as wholly dominated by objectivity of attitude. The only operative notions invoked in this picture are such as those of policy, treatment, control. But a thoroughgoing objectivity of attitude, excluding as it does the moral reactive attitudes, excludes at the same time essential elements in the concepts of moral condemnation and moral responsibility. This is the reason for the conceptual shock.221
220 In Part Three, I suggest that such a view does indeed manage to integrate and move beyond compatibilism and incompatibilism in just this way. See especially Section 6.
221 F & R, p. 89.
238
Strawson is surely correct here to accuse ‘optimism’ of the Smart or Schlick variety of
being in a sense guilty of a conceptual confusion. Such views save moral attitudes and
practices only by converting them into something very different from what they really are
– into a series of performances, either mental (in the case of attitudes) or actual (in the
case of practices such as blaming and punishing), designed to achieve particular kinds of
results, but without their internal moral character left intact. Indeed, as I argued above at
Section 3(I), a consequentialist calculus of individual attitudes would render the
landscape of our emotional lives quite unrecognizable. A view such as Schlick’s or
Smart’s would rob different kinds of attitudes of their distinctive identities. They would
no longer be appropriate or inappropriate responses to particular sets of circumstances.
Instead, all attitudes would be no more than mental performances designed to further our
goals or increase our prudential advantage.
On Strawson’s view, given his metaethical commitments, such attitudes would no longer
be moral because they would no longer have the right kind of internal relation to narrow
desert-entailing attitudes that are the impersonal analogue of resentment. On my view,
such attitudes would no longer be moral attitudes because they would no longer have the
right kind of internal relation to moral judgements. For, once we allow that the right
norms for policing our attitudes are based on purely instrumental or consequentialist
considerations, we sever the relationship that exists between moral judgements and those
moral attitudes. Nevertheless, I can agree with Strawson in thinking that Smart and
Schlick simply change the subject: theirs cannot be a defence of moral attitudes because,
whether on Strawson’s account or mine, moral attitudes are simply not that calculating.
239
Thus, we should be broadly sympathetic to Strawson’s conceptual hostility to the
optimist strategy, even if Strawson goes wrong in his precise specification of the way in
which the Schlickian approach involves a conceptual misunderstanding of the nature of
distinctively moral attitudes. There are, though, two points to be made about Strawson’s
conceptual hostility to ‘optimism’. Firstly, given that optimism and compatibilism are
two distinct kinds of view, Strawson’s rejection of optimism on grounds of conceptual
confusion does not extend to other forms of compatibilist views. A reader of Strawson’s
who made the common identification of optimism with compatibilism might think that
Strawson is making a broader claim here; but this argument against the optimist functions
only against the kind of instrumentalist or consequentialist view advocated by Schlick
and Smart.
Secondly, whilst one should endorse the general approach of Strawson’s conceptual
objection to optimism, it is worth noting that it sits very uncomfortably with other
elements of his position. When he appeals to “the gains and losses to human life” in
arguing for the retention of the full set of moral and reactive attitudes, Strawson makes
use of precisely the kind of instrumentalizing, consequentialist reasoning to which he is
so hostile when it is deployed by the optimist.222 But the charge of conceptual
misunderstanding can be deployed against any line of argument that would seek to rescue
a judgement-dependent attitude (such as a moral attitude) on consequentialist grounds, in
cases where the judgement on which that attitude depends has itself come to be rejected.
222 See the discussions in Sections 2(I) and 3(I).
240
Therefore, if we judge that narrow desert-entailing attitudes are inappropriate on general
philosophical grounds, we cannot then seek to rescue those attitudes, as Strawson does in
the appeal to “the gains and losses to human life”, by virtue of appeal to consequentialist
considerations, without thereby being guilty of a conceptual mangling of the attitude in
question. Thus, Strawson’s highly plausible rejection of the consequentialist reasoning of
‘optimists’ could itself be deployed against the consequentialist strand in his own positive
view. He is right to accuse the optimist of changing the subject in a conceptually
confused way, but this charge sticks just as effectively against those, like Strawson, who
would argue for the retention of otherwise unsupportable judgement-dependent attitudes
on the basis of a rational calculus of the “gains and losses to human life”.
Alongside his (broadly) successful conceptual challenge to instrumental forms of
compatibilism (i.e. ‘optimism’), Strawson also mounts a parallel, and much more
problematic line of objection. This second line of objection appeals to the ‘emotional’
shock of the optimist’s proposal, as opposed to its ‘conceptual’ shock. Whilst the first
line of argument is successful (albeit that it generates internal tensions for other elements
of Strawson’s view), the second is not, and it demonstrates some of the main problems as
regards Strawson’s thinking about attitudinal objectivity:
The deeper emotional shock is a reaction, not simply to an inadequate conceptual analysis, but to the suggestion of a change in our world. I have remarked that it is possible to cultivate an exclusive objectivity of attitude in some cases, and for some reasons, where the object of the attitude is not set aside from developed inter-personal and moral attitudes by immaturity or abnormality. And the suggestion which seems to be contained in the optimist’s account is that such an attitude should be universally adopted to all offenders. This is shocking enough in the pessimist’s eyes. But, sharpened by shock, his eyes see further. It would be hard to make this division in our natures. If to all
241
offenders, then to all mankind. Moreover, to whom could this recommendation be, in any real sense, addressed? Only to the powerful, the authorities. So abysses seem to open.223
Strawson’s ‘emotional’ objection to the optimist’s position is based on the familiar point
that the adoption of a thoroughgoing objective attitude would involve an unsustainable
(or, perhaps, sustainable yet horrific) alienation from human involvement and human
relationships. This ‘emotional’ point has been thoroughly discussed in the foregoing
sections of this paper, and I need not rehearse its problems in any further detail here.224
Suffice to say that this ‘emotional’ point makes sense only if we ignore the possibility of
the moderate objective attitude, and if we thereby make the mistake of assuming that the
objective attitude must be stark, and thereby emotionally shocking.
Thus, even if this line of objection did function successfully against a crude version of
‘optimism’, which held that all moral attitudes and practices would have to be replaced
by forward-looking, consequentialist replacements, it would not work against a restricted
version of ‘optimism’ that was informed by a version of the Hybrid view. That is to say, a
restricted version of ‘optimism’, which looked to remove and replace only the narrow,
desert-entailing attitudes and practices, would not thereby recommend a level of
emotional alienation or estrangement sufficient to generate the “emotional shock” that
here exercises Strawson.225
223 F & R, p. 90.
224 See the discussion of the objective attitude in Sections 3(E)-(G).
225 I discuss the motivation for, and prospects of, such a ‘restricted’ version of ‘optimism’ in Sections 5-8 of Part Three.
242
Strawson’s worries about the emotional shock of the adoption of the objective attitude
lead him to the rhetorical flourish that constitutes perhaps the oddest moment in his
discussion of these issues. “To whom,” asks Strawson “could this recommendation [for
an adoption of the objective view towards all mankind] be, in any real sense, addressed?
Only to the powerful, the authorities. So abysses seem to open.”226 Strawson’s worries
here are tricky to locate, but seem to be roughly (so to speak) Orwellian.227
His thought seems here to have something like this form: (1) let us assume that we should
individually adopt the objective attitude towards everyone; (2) yet this is emotionally
impossible for normal human beings; (3) thus, the recommendation that we should adopt
the objective attitude can only be addressed to “the powerful, the authorities”; so (4)
given (1), (2) and (3), the impersonal agency of state authority should be exercised in
accordance with the restrictions imposed upon it by the objective attitude. But (5) power
and authority exercised in accord with the objective attitude would be a vista too terrible
to contemplate: indeed, “abysses seem to open”. The form of the objection would seem,
so to speak, to be a form of modus tollens – because we have (consequentialist?) reasons
to avoid advocating (4), we should reject the premise (1) that leads us towards there. But
this seems like a form of wishful thinking. If the truth of (1) and (2) leads us to (4), then
so much the worse for us; we cannot reject some truth simply because we do not like its
consequences.
226 F & R, p. 90.
227 By which I have in mind the George Orwell of Orwell, (1949), Nineteen Eighty-Four, (London: Secker and Warburg).
243
Thankfully, though, the consequences of power and authority being exercised in accord
with the objective view are not so terrible as Strawson believes. As I’ve mentioned, I
assume that Strawson’s worries are broadly Orwellian. That is, his concern seems to be
that the ‘objective’ exercise of authority leads to a kind of bureaucratic
instrumentalization of those who live under that authority, and that authority and power
exercised ‘objectively’ leads to the dehumanization of its subjects. But one does not slide
all the way to bureaucratic disregard for the individual simply from adoption of the
objective view, especially if that objective view is of a modest rather than a stark kind.
Indeed, the policies of “the powerful” and “the authorities” could benefit greatly from
being conducted in a way that excises the emotional tinge of interpersonal relationships.
One might plausibly think that what politics allowed us to do was to take a step back
from the interpersonal reactive attitudes of face-to-face interactions, and to interpose a
layer of cold rational thinking. So, even if Strawson was right that the objective view was
impossible in day to day interactions, he would surely admit that government policy can
be formed without the need to express narrow reactive emotions of this kind. Such
policies might, for example, be more interested in the deterrence and rehabilitation of
criminals rather than their condemnation and punishment. This surely would not be
particularly emotionally shocking.
The state (or some state agency, or some other impersonal agent of “the powerful” or “the
authorities”) may act in such a way that its actions accord with certain reactive attitudes
which we have, but it certainly does not actually have those attitudes itself. Thus, it would
244
be quite consistent for us to engineer political arrangements which attempted to be
faithful to the insights of, say, the Hybrid View, and which looked to be consistent with
the moderate objective attitude, even if, for whatever reason Strawson might give us, we
could not rid ourselves of those attitudes in our day to day dealings with each other. Even
if I cannot suspend resentment towards the man who steals my wallet, I do not have to
think that the state should treat that man in a way which displays some analogue of my
resentment. Rather, I can make a general commitment to supporting political institutions
which manifest judgements and practices that I can rationally endorse, even if they
happen to be in tension with emotionally charged (narrow) reactive attitudes that I find
myself unable to excise from my personal emotional repertoire.
So, whilst it is true that an Orwellian state takes a view of its citizens that is in some ways
‘objective’, it is not a necessary consequence of taking an objective view that we find
ourselves on the path to ‘dehumanizing’ forms of authority. On the contrary, we can,
through our political processes, authorize and endorse principles for the impersonal
agency of our political institutions which are responsive to our underlying beliefs and
commitments. In this way, the adoption of a degree of objectivity of view by the state and
its agencies can be a precondition for a more rational and considered approach to policy
making. It is a crude and inappropriately personalized vision of politics that would see
the role of our collective institutions as consisting in some kind of re-enactment or
amplification of our own reactive attitudes of blaming, condemning and resenting. It is
this kind of atavistic view of politics that is more likely to cause ‘abysses to open’. By
comparison, addressing the injunction to our collective political institutions to stick by
245
the constraints of the moderate objective view suggests an approach that might allow
public policy, and the exercise of power and authority, to rise above the level of the
abysmal.
L. Optimism, Pessimism and the “General Framework” of Human Attitudes
We may recall that Strawson’s arguments for Claim (φ) come in the opposite order to his
earlier arguments regarding the capacity criterion in the case of the interpersonal reactive
attitudes. 228 In Section 3(I) we addressed Strawson’s arguments regarding the rationality
(or, more accurately, the irrationality) of adopting the objective view, at the expense of
(some of) the moral reactive attitudes. In this section, we shall turn to examine
Strawson’s arguments regarding the impossibility of giving up some of the moral reactive
attitudes, and of thereby moving towards some generalized version of the objective view.
Our previous discussion has shown that the adoption of such a view need not be
irrational. If it can be demonstrated that the adoption of such a view is also not
impossible, then Strawson will have lost both of the argumentative planks upon which he
might have relied for the support of claim (φ), that is, the claim that the Capacity criterion
provides necessary and sufficient conditions for when (any of) the moral reactive
attitudes may be appropriately suspended. If it can be shown that Strawson’s
impossibility argument is as unsuccessful in this case as his irrationality argument, then
we must conclude that, bereft of support, claim (φ) cannot but fall when challenged by
228 For discussion of Strawson’s analogous arguments in the case of the interpersonal reactive attitudes, see Sections 2(G)-(H).
246
the Hybrid View, or by any independently plausible revisionist or (partially)
incompatibilist view.
Strawson thinks that the mistake made by revisionist, incompatibilist views
(“pessimism”) is related to a similar kind of error that is perpetrated by instrumentalist
compatibilists (“optimists”). The pessimists call for a fundamental change in our
conceptualization of moral responsibility on the basis of the appreciation of general facts
about the nature of human agency. But this, Strawson claims, is to indulge in the error of
“over-intellectualization”. Once we resist the blandishments of this kind of “over-
intellectualization”, thinks Strawson, then the impossibility of revisionism emerges:
Optimists and pessimists misconstrue the facts in very different styles. But in a profound sense there is something in common to their misunderstandings. Both seek, in different ways, to overintellectualize the facts. Inside the general structure or web of human attitudes and feelings of which I have been speaking, there is endless room for modification, redirection, criticism, and justification. But questions of justification are internal to the structure or relate to modifications internal to it. The existence of the general framework of attitudes itself is something we are given with the fact of human society. As a whole, it neither calls for, nor permits, an external ‘rational’ justification. Pessimist and optimist alike show themselves, in different ways, unable to accept this.229
This hopeful appeal to the “general framework of human life” should be familiar to us
from its first deployment by Strawson in the case of the non-moral reactive attitudes, and
has been discussed above.230 That appeal is as unsuccessful here as it was in the former
case. For a start, views that are incompatibilist about the desert-entailing moral attitudes
do not threaten any kind of “general framework” of human interaction; they merely seek
229 F & R, p. 91-2.
230 See Section 2(H) above.
247
to undermine a particular, and rather narrow, sub-set of reactive moral attitudes.
Strawson’s over-inflation of the incompatibilist threat leads him to misidentify the scope
of the challenge posed by revisionist views.231
Moreover, Strawson gives us no reason to accept his characterization of the
incompatibilist or Hybrid threat to (some of) the reactive moral attitudes as being
launched from some ‘external’ standpoint outside of the general framework of human
life. The aims of self-understanding and internal coherence stand at the heart of any
plausible general structure of human attitudes, and it is considerations such as these that
drive us towards Hybrid or incompatibilist views – not some unmotivated and mysterious
‘external’ threat, coming from outside of the intelligible bounds of human life. As
Thomas Nagel accurately puts it, “the push to objectivity is after all a part of the
framework of human life.”232 Given this, the characterization of the revisionist
philosophical threat as ‘external’ is implausible.
Even if we set aside this source of implausibility in Strawson’s claims, though, we find
another such source lurking in the background. When Strawson accuses his opponents of
231 One may also here have the independent worry that Strawson has mistaken some particular and culturally local varieties of human attitude for “humanly necessary” forms of attitude. See the related discussion above, at Section 2(H).
232 Thomas Nagel, (1986), The View from Nowhere, (Oxford: OUP), p. 126. I would also heartily endorse Nagel’s comments on Strawson’s attempt to defend his view by means of an appeal to an external/internal distinction, at ibid. p. 125: “I believe this position [i.e. Strawson’s] is incorrect because there is no way of preventing the slide from internal to external criticism once we are capable of an external view. It needs nothing more than the ordinary idea of responsibility. The problem of free will, like the problem of skepticism, does not arise because of a philosophically imposed demand for external justification of the entire system of ordinary judgements and attitudes. It arises because there is a continuity between familiar “internal” criticism of the reactive attitudes on the basis of specific facts, and philosophical criticisms on the basis of supposed general facts.”
248
“over-intellectualization” there is a sense in which he means exactly what he says. He is
accusing them of treating the question of whether certain human reactive attitudes are
properly regarded as appropriate, in the general case, as a matter of intellectual
judgement, rather than of blind commitment. For Strawson, any form of
‘intellectualization’ of these questions counts as over-intellectualization.
Now, it is certainly true that the views of Strawson’s opponents are philosophical views,
designed to be supported by reasons, and thereby appealing to the judgement, and not to
the non-rational commitments or passions, of those who encounter those views and the
arguments for them. But if Strawson would wish to make his own view at all plausible,
then he cannot but appeal to the intellects and theoretical reason of his readers. That, after
all, is the very task in which his essay is engaged. But, given this, the charge of “over-
intellectualization”, when levelled at his opponents, is difficult to take seriously. If one
really wants to counter revisionist philosophical views by accusing them of “over-
intellectualization”, then one should have the courage of one’s convictions, and abandon
reason in favour of faith, authority, passion or dogma within this whole domain.
Strawson, of course, does no such thing, and so his selective invocation of a charge of
“over-intellectualization” cannot be viewed as having any plausible force against the
arguments of his opponents.233
233 Strawson goes on to attempt to buttress this part of his argument by drawing an analogy with the justification of induction. The failure of this analogy has been discussed extensively above, for which see Section 2(H).
249
Strawson’s naturalistic opposition to even acknowledging the genuine existence of a
philosophical problem in this domain comes to the fore even more strongly in what he
has to say about the putative mistakes made by “the genuine moral sceptic”:234
Even the moral sceptic is not immune from his own form of the wish to over-intellectualize such notions as those of moral responsibility, guilt, and blame. He sees that the optimist’s account is inadequate and the pessimist’s libertarian alternative inane; and finds no resource except to declare that the notions in question are inherently confused, that ‘blame is metaphysical’. But the metaphysics was in the eye of the metaphysician. It is a pity that talk of the moral sentiments has fallen out of favour. The phrase would be quite a good name for that network of human attitudes in acknowledging the character and place of which we find, I suggest, the only possibility of reconciling these disputants to each other and the facts.235
Despite the striking rhetorical force of this passage, Strawson has given us no reason to
accept his underlying claim that treating the questions of the appropriateness or
justifiability of (certain) moral reactive attitudes as a matter for rational investigation or
theoretical judgement is somehow to ‘over-intellectualize’ those questions. In opposition
to Strawson’s view, I would instead claim that treating these questions as a matter on
which we can come to rational judgement is a minimal condition for having anything
defensible to say on these questions at all. Strawson’s approach, insofar as it is avowedly
concerned to avoid ‘over-intellectualization’, in fact amounts to a form of anti-
intellectualization. In appealing to the authority of existing practices and attitudes,
Strawson is thereby committed, by methodological fiat, to not being prepared to credit
the thought that an ‘everyday’ practice can be unjustifiable or inappropriate. But such
234 In discussing the position of “the genuine moral sceptic”, Strawson here interestingly pre-empts the view that was to be adopted by his son, Galen Strawson. I shall discuss the view of the younger Strawson, especially with regard to the claim that there is a kind of inherent confusion in the idea of moral responsibility, at greater length in Part Three, and especially in Section 4.
235 F & R, p. 92.
250
practices surely can be unjustifiable or inappropriate, insofar as they rest on false beliefs,
or involve irrational or unjustifiable judgements or attitudes. Sometimes normal practice
can go wrong, be irrational, or generate genuine and irreducibly philosophical problems.
To think, ex ante, that this must be impossible is indefensibly anti-intellectual.
It has been a recurring claim of this discussion that Strawson often errs in over-estimating
the internal coherence and indivisibility of the various kinds of human attitudes. By
failing to see the possibility of moderate versions of the objective view, he takes it that
any threat to the desert-entailing reactive or moral attitudes must also be a threat to the
possibility of all human reactive or moral attitudes in general. This tendency is
Strawson’s thought is again to the fore when he downplays the significance (as a way of
buttressing the plausibility of certain kinds of revisionist views) of the insights of an
elevated historical and anthropological awareness of the specificity and contingency of
any particular set of reactive attitudes. As can be seen, Strawson aims to pass over this
worry rather too quickly:
One factor of comparatively minor importance is an increased historical and anthropological awareness of the great variety of forms which these human attitudes may take at different times and in different cultures. This makes one rightly chary of claiming as essential features of the concept of morality in general, forms of those attitudes which may have a local and temporary features of our own culture. But an awareness of variety of forms should not prevent us from acknowledging also that in the absence of any forms of these attitudes it is doubtful whether we should have anything that we could find intelligible as a system of human relationships, as human society.236
Far from being of only “minor importance”, it would instead seem that the malleability
and diversity of the human emotional repertoire is actually of great potential significance.
236 F & R, p. 92-3.
251
For an awareness of historical and anthropological differences between different cultures
and groups goes some way towards undermining the thought that the reactive attitudes
are a unified and integrated set, which must, so to speak, always stand or fall together. If
they do not form such an integrated set, then the excision of the desert-entailing attitudes
begins to seem much less like an impossibility. As has been argued in Sections 2(G) and
2(H), there is no reason to think that the excision of the desert-entailing reactive (or
moral) attitudes need destabilize the broader set of human attitudes. Awareness of the
malleability of our attitudinal repertoire can only serve to bolster these claims, and with it
to lend further plausibility to the rejection of claim (γ) in Section 3(E).237 As against
Strawson, we need not think that the existence of a system of human relationships
requires the preservation of any of the attitudes that are considered illegitimate or
unjustifiable under the terms of the Hybrid View.
If Strawson were only to jettison the misplaced essentialism of his view – which sees the
reactive and moral attitudes, including especially the desert-entailing attitudes, as forming
a unified, indivisible and indispensable set of human attitudes, necessary for the
preservation of human relationships – then he would have no reason not to entertain
moderate revisionist views such as the Hybrid View. For Strawson allows, in the abstract,
that the modification of human attitudes may be both possible and desirable:
Finally, perhaps the most important factor of all is the prestige of these theoretical studies themselves. That prestige is great, and is apt to make us forget that in philosophy, though it is also a theoretical study, we have to take account of the facts in all their bearings; we 237 That is, where the rejection of Claim (γ) involves the rejection of the claim that taking the moderate objective attitude towards some agent thereby involves seeing him or her as being wholly outside the scope of moral relationships.
252
are not to suppose that we are required, or permitted, as philosophers, to regard ourselves, as human beings, as detached from the attitudes which, as scientists, we study with detachment. This is in no way to deny the possibility and desirability of redirection and modification of our human attitudes in the light of these studies. But we may reasonably think it unlikely that our progressively greater understanding of certain aspects of ourselves will lead to the total disappearance of those aspects. Perhaps it is not inconceivable that it should; and perhaps, then, the dreams of some philosophers will be realized.238
Once we jettison Strawson’s excessively ‘integrated’ or unified account of the reactive
attitudes, we can come to very different conclusions to Strawson’s about where the
growth of our theoretical self-understanding could come to lead us. It has been argued in
this section that we have no reason to think that the disappearance of the desert-entailing
moral attitudes should be an impossibility. In previous sections, it has already been
argued that this disappearance would, moreover, be desirable, not least insofar as it is
rationally required of us. Strawson’s admission that it is possible that “the dreams of
some philosophers will be realized” is made rather grudgingly, and with evident regret.
But we need have no such ambivalence about the realization of such philosophical
dreams.
Strawson’s essay ends with a moderate redescription of his view as a form of modified
optimism: that is, a form of modified compatibilism, of a kind that keeps one eye on
consequentialist considerations. This is, of course, all of a piece with his claim, which has
been disputed in this essay, that his view represents a compromise between compatibilist
and incompatibilist views. As Strawson puts things:
238 F & R, p. 93.
253
If we sufficiently, that is radically, modify the view of the optimist, his view is the right one. It is far from wrong to emphasize the efficacy of all those practices which express or manifest our moral attitudes, in regulating behaviour in ways considered desirable; or to add that when certain of our beliefs about the efficacy of some of these practices turn out to be false, then we may have good reason for dropping or modifying those practices. What is wrong is to forget that these practices, and their reception, the reactions to them, really are expressions of our moral attitudes and not merely devices we calculatingly employ for regulative purposes. Our practices do not merely exploit our natures, they express them. Indeed the very understanding of the kind of efficacy these expressions of our attitudes have turns on our remembering this. When we do remember this, and modify the optimist’s position accordingly, we simultaneously correct its conceptual deficiencies and ward off the dangers it seems to entail, without recourse to the obscure and panicky metaphysics of libertarianism.239
Strawson is absolutely right that standard ‘optimism’ fails: as has been argued above, it
succeeds only in changing the subject under consideration.240 He is surely further correct
that we should use all our philosophical effort to avoid what he describes, wonderfully
and justly, as “the panicky metaphysics of libertarianism”. And one can only applaud his
attempt to come to a considered view which integrates the insights of earlier, ultimately
unsuccessful, ‘optimist’ and ‘pessimist’ views.241 Moreover, it would be eccentric to
demur from his claim, sympathetic to the optimist, that it would not be wrong to
emphasize the efficacy of our expressions of our moral attitudes.
Most of all, Strawson is correct in saying that our practices and attitudes do not merely
exploit our natures, but express them. But our nature is not simply that of creatures
blindly committed to an endless exchange of desert-entailing attitudes. Our natures,
239 F & R, p. 93.
240 See the argument of Section 3(K), above.
241 In my Part Three, I present an alternative way of integrating the insights of these different kinds of views, as well as offering some general considerations regarding why a successful treatment of the problems of freedom and responsibility must be sensitive to the entrenched and opposed philosophical positions which exist in this domain (see especially Section 6).
254
expressed in our practices and attitudes, are the natures of rational creatures, committed
to coming to self-understanding as doers and knowers, and committed to the constraints
of rationality on what we believe and how we act. Our natures, understood in this richer
way, are, indeed, expressed clearly and distinctly in our disavowal of attitudes which we
have good reason to reject. What Strawson has left out is the important fact that our
moral attitudes – that is, the moral attitudes of rational creatures – are themselves
something more than expressions of unavoidable pre-rational emotions. Rather, they fall
within the bounds of reason.
M. Conclusion: Strawson’s Second Question Answered
We are now in a position to answer Strawson’s second question, in which he addressed
the question of what effect any general thesis of determinism might have on our moral
reactive attitudes. As he puts the question:
What concerns us now is to inquire, as previously in connection with the personal reactive attitudes, what relevance any general thesis of determinism might have to their vicarious analogues.242
Strawson’s answer to this question ran parallel to his answer to his first question,243 and
mine will do likewise.244 Firstly, though, we must dispute Strawson’s formulation of this
question for, as has been argued above, it is already to make a number of significant
errors to assume that moral attitudes like disapprobation or blame are themselves best
242 F & R, pp. 86-7.
243 F & R, p. 87. (“The answers once more are parallel.”)
244 For Strawson’s ‘First Question’, see F & R, p. 90. For my answer to that first question, see Section 2(J) above.
255
understood as “vicarious analogues” of desert-entailing reactive attitudes such as
resentment.245 Nevertheless, my parallel answer remains appropriate in this case.
Adoption of a Hybrid View, and with it the moderate objective attitude, involves the
repudiation of all desert-entailing reactive attitudes, whether those attitudes are moral
attitudes or not. This repudiation is not impossible for us, and nor would it be irrational,
nor necessarily undesirable.246
Accordingly, we should reject Strawson’s Claim (φ), according to which the Capacity
Criterion gives us necessary and sufficient conditions for when we should exempt any
individual from any of the moral attitudes. We should reject it because adoption of a view
such as the Hybrid View would itself give us sufficient reason to exempt individuals from
the narrow, desert-entailing moral attitudes, such as moral blame.247 In short, Strawson’s
strategy of ‘naturalistic avoidance’ fails, and therefore the moral attitudes, every bit as
much as the interpersonal attitudes such as resentment, are not immune from rejection on
the basis of general philosophical considerations.
4. Beyond Optimism and Pessimism: or, Freedom without Resentment
The foregoing discussion has provided us with sufficient reason to reject Strawson’s
practical (or naturalistic) compatibilism about moral responsibility. Moreover, it has been
245 See the argument above, especially Sections 3(B)-3(D).
246 See the arguments above of Sections 3(G)-3(J).
247 I here treat blame as a desert-entailing attitude. For further discussion of blame, see Part Three below, and especially Section 3.
256
shown how Strawson’s view would seem to lack the argumentative resources to argue
against a ‘Hybrid View’, which rejects only the narrow reactive and moral attitudes, and
enjoins adoption of a moderate version of the objective attitude.
The question therefore now arises of whether such a ‘Hybrid View’ can be motivated
successfully, and presented in a way that makes it a compelling solution to the problems
of freedom and responsibility, rather than just one option among others in logical space.
Such a view might be able to make good on Strawson’s suggestion that the best way
forward would involve going beyond the standard oppositions between compatibilism
and incompatibilism, and between ‘optimism and ‘pessimism’, and might involve
effecting a reconciliation (of sorts) between these opposed positions. It is to the task of
presenting and defending such a view, and effecting such a reconciliation, that I now turn
in Part Three.
257
Freedom, Fairness and Responsibility
Part Three
Freedom, Fairness,
Responsibility and Blame:
A Hybrid View
258
Freedom, Fairness and Responsibility – Part Three
Freedom, Fairness, Responsibility and Blame: A Hybrid View
1. A Real Reconciliation between Compatibilism and Incompatibilism
In “Freedom and Resentment”, P. F. Strawson purports to offer a ‘reconciliation’
between compatibilist and incompatibilist, or between ‘optimist’ and ‘pessimist’,
positions on freedom of the will. In reality, as we have seen in the preceding Part of this
dissertation, he does no such thing. Strawson, in fact, calls a plague on both their houses.
He accuses the proponents of both kinds of position of “over-intellectualization,”248 and
suggests that they are attempting to answer a “useless” question.249 So, this isn’t really
very much of a reconciliation. In what follows, I want to explore what a real
reconciliation between these two positions would actually look like. One thing that a real
reconciliation might involve is taking seriously the two sorts of views on their own terms;
believing, as it were, that they are addressing themselves to a real question, rather than
holding, as Strawson does, that both kinds of views are mired deep in error.
248 F & R, p. 91.
249 F & R, p. 87.
259
My suggestion is that a genuine ‘reconciling strategy’ would involve isolating what is
true from what is false in rival compatibilist and incompatibilist views, and combining
those true elements into a coherent ‘Hybrid’ view. Such a Hybrid View has been lightly
sketched in the previous discussion (and especially here and there in Part 2). The
foregoing discussion also showed that, whilst Strawson’s argumentative strategies may,
in part, be successful against ‘crude’ or ‘stark’ forms of incompatibilism, they are
nevertheless ineffective against such a ‘Hybrid’ position. However, other than by virtue
of demonstrating the immunity of such a Hybrid view to the reach of Strawson’s
naturalistic lines of argument, whether regarding the ‘Inescapability of the Human
Commitment’ or the more instrumental appeal to ‘The Gains and Losses to Human Life’,
the positive case for adopting such a view has not yet been made. It is the task of the
present discussion to make the positive case for the adoption of such a view.
Before making the case for such a Hybrid view, it would be useful to start by laying out
more precisely what the content of such a ‘reconciling’ view must be. The view which I
shall defend rejects incompatibilism about the conditions for agency, as such, but
embraces incompatibilism (or even thoroughgoing scepticism) about the possibility of the
kind of agency that might ground claims of desert. In the terms of the previous
discussion, this is a ‘type-2’ version of ‘restricted incompatibilism’, as opposed to the
sort of (‘type-1’) standard or unrestricted incompatibilism, which deals with the
possibility of agency, as such, rather than more narrowly with the possibility of agency of
a kind that could ground legitimate claims of desert.250 To put things simply, such a
250 On the distinction between ‘Type-1’ and ‘Type-2’ views, see Section 2(F) of Part Two, above.
260
‘Hybrid’ view does not deny that robust and morally significant forms of agency are
compatible with causal determination. But such a view does deny that (either if
determinism is true, or, alternatively, simpliciter) individual agents can deserve to suffer
some harm or bear some good.
The Hybrid view that I shall be defending, therefore, entails that we should suspend and
repudiate some, but not all, of our broad repertoire of reactive attitudes. The reactive
attitudes which should be suspended and repudiated are the narrow set of desert-entailing
reactive attitudes, such as Strawson’s recurring triad of resentment, gratitude and
forgiveness. Endorsement of the Hybrid View therefore involves adopting the moderate
(but certainly not the stark) version of the ‘objective attitude’.251 The moderate objective
attitude, of the kind adopted by Graham Greene’s Whisky Priest towards the pious
bourgeois woman who shared his dark, crowded prison cell, is the general attitudinal
stance that is consistent with acceptance the Hybrid View.252 (Although it should be
pointed out that, whilst adoption of the moderate objective attitude is consistent with, and
may even be especially propitious for, the kind of deeply charitable attitude displayed by
Greene’s Whisky Priest, it is not that adoption of the moderate objective attitude need
involve this kind of supererogatory degree of sympathy.)
As has been extensively discussed, adoption of the Hybrid View involves rejection of
Strawson’s central claim (φ). Claim (φ) is the claim that the strong construal of
251 On the moderate and stark variants of the Objective Attitude, see Section 2(C) of Part 2.
252 On the example of Greene’s Whisky Priest, as a case that demonstrates that the moderate objective attitude need not be inconsistent with emotional engagement, see Section 2(C)-(D) of Part 2.
261
Strawson’s ‘Capacity criterion’253 is the correct one. That is, it is the claim that we should
see the Capacity criterion as providing necessary and sufficient conditions for when an
agent should be exempted from any of the full set of the moral attitudes. The Hybrid view
involves the rejection of claim (φ), as it postulates general philosophical reasons for
exempting all individuals from the narrow, desert-entailing reactive attitudes, even in
cases where those individuals meet the standards of rational agency described by the
Capacity criterion.
Despite holding that there are general reasons for exempting individuals from the narrow,
desert-entailing reactive attitudes, though, the Hybrid View is fully consistent with
holding that the Capacity criterion does give necessary and sufficient conditions for the
appropriate suspension of the remaining, broad class of non-desert-entailing reactive
attitudes. It is also fully consistent with the view that, given its rejection of standard
unrestricted compatibilism, we should still regard normal adult human agents as being
fully morally assessable participants in moral relationships – or, as we may put it, full
members of the moral community.254
253 Where the ‘Capacity Criterion’ is the claim it is inappropriate to have a reactive attitude towards someone who is not a normal adult human and thereby a competent partner or interlocutor in adult relationships. Claim (φ) is thus the more particular claim that we should in general exempt an individual from any of the reactive attitudes only when they fail to exhibit the relevant rational or agential capacities.
254 Thereby rejecting also Strawson’s Claim (γ). Claim (γ) holds that, if we take the ‘objective attitude’ towards some agent, insofar as we suspend (certain) moral attitudes towards him, we thereby see that agent as not being subject to moral demands and obligations. That is, taking the ‘objective attitude’ towards some agent is to see him as outside the bounds of our moral community, as not being “a term of moral relationships
262
I have suggested that a Hybrid View of this kind might serve as a genuine reconciliation
between more familiar compatibilist and incompatibilist positions. What I have in mind is
this: the Hybrid View is able to grant that there is some truth in both standard
compatibilism and in standard incompatibilism.
Incompatibilism is to some degree correct, insofar as the totality of our moral practice,
including the deployment of the desert-entailing reactive attitudes, is not in good order
when held up against our best understanding of the nature of agency. Incompatibilism is,
then, right in holding that our current moral practice calls for revision. Where
incompatibilism goes wrong is in thinking that this need for revision applies without
restriction of domain. Instead of this unrestricted version of incompatibilism, we should
see that the truth in incompatibilism is restricted to a particular sub-domain of our current
practice – that is, that it is restricted to that sub-domain involving the desert-entailing
attitudes. Thus, incompatibilism is part of the truth, but not the whole truth.
On the other hand, there is also some truth in standard compatibilism. The sort of
autonomous agency that is a precondition for full moral assessment is itself compatible
with the truth of determinism. Any of the properties that we properly care about when we
care about an agent’s freedom – from normative competence, to diachronic stability, to
coherence and sophistication, are demonstrably realizable, whether or not determinism
(or any related doctrine) happens to be true.255 Any of the properties of agency which
255 I do not claim this as an exhaustive list, but it captures some of the robust features of agential autonomy displayed by our character of ‘Reasonable Rebecca’ in the first part of this dissertation. On the need for such a ‘multipolar’ or pluralist account of autonomy, see Part 1, Section 8.
263
mark out an individual as free, or autonomous, as a proper subject for moral assessment,
or as a co-participant in the moral community are, as the compatibilist holds, secure from
‘metaphysical’ threat. But the compatibilist goes wrong in thinking that it follows from
this that our current moral practice is in good order. The compatibilist goes wrong in
sliding from the correct view about freedom, autonomy and moral community to a related
but erroneous view about desert, (desert-entailing) blame and (desert-entailing)
responsibility. Like the incompatibilist, the compatibilist is too ambitious in thinking that
the true parts of his doctrine apply without restriction of domain. On the view I defend,
we can show that the compatibilist should fight shy of extending the scope of his view to
encompass the desert-entailing attitudes and the desert-entailing regions of our current
moral practice.
My claim, then, is that the incompatibilist is too alarmist and the compatibilist too
sanguine. But nevertheless, they both get things partially correct. With regard to our
current practice, there are elements we should keep unchanged, and elements that are ripe
for rejection and supersession. The compatibilist and incompatibilist are like two of the
blind men who, in the fable, examine different parts of the elephant. One, feeling the
tusk, reports that elephants are hard, smooth and cold. Another, examining the elephant’s
side, pronounces the beast to be soft, rough and warm. Neither is completely wrong.
Their error is an error only of over-generalization.
In granting the partial truth of each doctrine, we are able to have some sense of why the
dispute between the two views has been so enduring and recalcitrant. (Compare this to a
264
view that holds that one of either compatibilism or incompatibilism gets things
completely correct about free will and moral responsibility. On any such view, the depth
and enduringness of this most venerable of all philosophical disputes can only seem
mysterious and aberrant.) The dispute is enduring because the participants in the dispute
all get at part of the truth. This is not to say that there might not be views about freedom
and responsibility that are simply wrong (I share Strawson’s views on the “obscure and
panicky metaphysics of libertarianism”256), but it is to say that some of the central
thoughts that stand at the core of rival compatibilist and incompatibilist views are equally
defensible, and equally directed towards the truth of the matter. Moreover, through the
reconciliation offered by the Hybrid View, one can come to see the core elements of
these putatively rival views as being unproblematically mutually consistent.257
Perhaps it is still too ambitious to speak of the Hybrid View as offering a reconciliation
between the competing doctrines of compatibilism and incompatibilism. What is
certainly not being offered is a reconciliation in which both parties can keep everything
that they might want or hope to keep. Both must give ground, in acknowledging the
partial error of their view. Indeed, the two parties may even disagree on which of them is
being expected to give up the most, for a Hybrid view of the kind I am defending may be
viewed, alternatively, as either a slightly amended version of compatibilism or as a
moderately concessionary form of incompatibilism.
256 F & R, p. 93.
257 I discuss the enduringness and recalcitrance of philosophical disputes about freedom of the will and responsibility in more depth in Section 6 below, where I also discuss the sense in which the Hybrid View can be seen as reconciling earlier views on the subject, and the general methodological considerations that speak in favour of a view that is able to grant at least the partial truth of popular and enduring rival philosophical positions.
265
One could not, of course, even begin to imagine the view that would give compatibilists
and incompatibilists all that they might want or hope to keep as regards the contents of
their views, whilst nevertheless representing those two views as mutually consistent. But
my contention is that the Hybrid View offers the most plausible reconciliation available,
and one which leaves both of the traditional warring camps with much of their honour
preserved. For, in defending the Hybrid View, it is clear that both compatibilist and
incompatibilist views involve significant philosophical insights of power and importance.
If this is enough for a “real reconciliation”, then it is so much the better for the prospects
for acceptance of a version of the Hybrid View. If this still falls short of a “real
reconciliation”, then one may have two replies. Firstly, it’s a stronger and much more
genuine effort at reconciliation than that associated with Strawson’s naturalistic
compatibilism. Secondly, if this attempted reconciliation is not good enough, then
perhaps no fully satisfactory ‘reconciliation’ is possible, or even desirable.
Strawson suggests (misleadingly, as regards the content of his own view) that
reconciliation between the opposed parties on this issue might involve “a formal
withdrawal on one side in exchange for a substantial concession on the other”.258 It was a
presciently good suggestion, even though it was not one on which Strawson himself made
good. In what follows, I argue for exactly such an exchange: for why the incompatibilists
should undertake a “formal withdrawal” in exchange for substantial concessions from the
258 F & R, pp. 72-3.
266
compatibilists.259 It is to the substantive case in favour of the Hybrid View that I
therefore now turn.
2. Fairness and Responsibility
In order to make sense of the way in which we may have good reason to adopt restricted
or partial versions of both compatibilism and incompatibilism, I want to make use of a
general framework for thinking about claims regarding moral responsibility that has been
developed by R. Jay Wallace. Wallace’s own positive position is a form of Strawsonian
compatibilism that holds that individuals are properly held morally responsible whenever
they have the general capacity for “reflective self-control”. Wallace thus endorses a view
that is a (fairly) standard form of compatibilism, consistent with holding the truth of (a
version of) Strawson’s claim (φ). Clearly, this is a very different (and less
‘concessionary’) view to the restricted compatibilism of the Hybrid View. So, it is worth
emphasizing at the outset that, in following the general methodology suggested by
Wallace, I am not thereby endorsing his substantive position regarding moral
responsibility. Nevertheless, it seems to me that Wallace has identified the best general
way of thinking about what makes claims about responsibility true or false, and that his
framework for approaching these issues is therefore a particularly useful and illuminating
one. The fact that he makes use of a particular general approach to arrive at (what I shall
259 The language of “formal withdrawals” and “concessions” is illuminating and suggestive, but I am not wedded to this particular way of describing the orientation of the dialectical situation. One might just as plausibly suggest that what is involved is a “formal withdrawal” from the compatibilists in exchange for substantial concessions from the incompatibilists.
267
argue to be) an incorrect conclusion does not speak against the usefulness or significance
of that general approach.
Wallace’s methodological suggestion is that we should understand the question of what it
is for an agent to be morally responsible as being an irreducibly normative, rather than in
some sense ‘metaphysical’, question. As he describes the structure of his own approach:
I propose that we see the issue of what it is to be morally responsible as in the first issue a normative question. To be responsible (for x), according to this proposal, is to be someone whom it would be fair to hold responsible (for x). This framework makes the question of the conditions of responsibility turn crucially on our moral norms of fairness. Equipped with a realistic understanding of the stance of holding someone morally responsible, we need to consider carefully the conditions under which it would be fair (in the twin senses of reasonable and deserved) to adopt this stance towards people. My hope is that by casting the issue in these normative terms, we may engage more directly than other approaches have succeeded in doing with the intuitions that have persistently fed the incompatibilist tradition. 260
In following Wallace’s suggestion, I am thereby treating the question of the ascription of
moral responsibility as a normative question that is continuous with, and internal to, our
moral discourse. The question of the appropriateness of judgements of responsibility
becomes, on this approach, itself a moral question. Therefore this approach guards
against the danger, repeatedly stressed by Strawson, of treating worries about
responsibility as excessively abstract, ‘external’ questions, seen as illegitimately
threatening the coherence of our moral practices from the outside. If, following Wallace’s
methodological approach, it is nevertheless found that there is good reason to adopt a
Hybrid (or ‘restrictively incompatibilist’) position, then such a result will have been
260 R. Jay Wallace, (2002), “Précis of Responsibility and the Moral Sentiments”, Philosophy and Phenomenological Research, LXIV: 3, p. 681.
268
obtained without the need to appeal top the sort of ‘external’ metaphysical considerations
of which Strawson is so wary. Thus, if following this approach nevertheless leads to
conclusions that are in some way critical of our current practices of ascribing
responsibility, those results will be all the more robust when considered from any
viewpoint which is itself more suspicious of ‘external’ than of ‘internal’ challenges to our
moral beliefs and practices.
Wallace’s approach has a number of advantages. Firstly, as has been mentioned, it is a
good way of motivating discussion of moral responsibility whilst avoiding any
(potentially controversial) need to invoke ‘external’ metaphysical considerations.
Secondly, and relatedly, Wallace’s approach gives us the most promising way of making
sense of the intuitions that stand behind the different varieties of incompatibilism. When
we are convinced by an incompatibilist view, it presents itself to us not as a ‘mere’
metaphysical puzzle about how it could be appropriate to hold some individual
responsible for some piece of behaviour, given that the facts about the nature of agency
are thus and so. Rather, when we are convinced of an incompatibilist view, we typically
hold that it would be wrong to hold individuals responsible in some regard. Wallace’s
normative approach helps us to make sense of the fact that being convinced of different
views as regards the ascription of responsibility is phenomenologically of the same kind
as being convinced of some first-order moral view. It is quite unlike being exercised by
an abstract metaphysical puzzle – rather, our moral sense as well as our intellectual
faculties are exercised and tested when we come to take a view on these issues.
269
Thirdly, and following on from these consideration, the understanding of the debate as
essentially normative helps to explain one of the central and most intriguing features of
the philosophical debate about responsibility: that is, the enduringness of incompatibilist
intuitions. That such intuitions have an essentially moral motivation helps to explain why
they are so resistant to being extinguished by even the most sophisticated and developed
compatibilist views.
This ‘normative approach’ thereby generates the following general schema for
understanding the conditions that make a person responsible for some action, x:
(N) S is morally responsible (for action x) if and only if it would be appropriate to
hold S morally responsible for action x.261 As it stands, (N) presents only a formal condition on ascriptions of responsibility. In
order to move to a more substantive interpretation of (N), we need two things. The first of
these is a fuller interpretation of the operative sense of ‘appropriate’. The second is a
fuller specification of what it is to hold someone morally responsible for some action. On
the first question, I am broadly in agreement with Wallace (although, as we shall see,
with some significant differences): we should understand the operative idea of
‘appropriateness’ in terms of our norms of fairness. Where I part company from Wallace
is with regard to the second issue: the fuller specification of what it is to hold someone
morally responsible. I shall turn to this latter issue in the following section, but we should
start by an examination of the idea of appropriateness as fairness.
261 R. Jay Wallace, (1994), Responsibility and the Moral Sentiments (Cambridge, MA: Harvard University Press), p. 91. (Henceforth, references to this book will be as ‘R & MS’.)
270
Wallace suggests that, when we appeal to moral norms in deciding on the appropriateness
of ascriptions of responsibility, the salient norms to which we must appeal are norms
concerning fairness. As he puts it:
Incompatibilists typically contend that it would be unfair to hold people responsible if they lack strong freedom of the will, whereas compatibilists deny this; to settle the issue, it is necessary to take a step into normative moral theory, determining the content of the principles of fairness to which we are committed, to see what those principles imply about the fairness of holding people morally responsible.262 […] moral norms of fairness have a privileged position in determining what it is to be a responsible agent: they set the standards of appropriateness in terms of which schema (N) is to be interpreted.263
It is highly plausible to think that incompatibilist intuitions are driven by moral
considerations, as this supposition seems to fit well with the phenomenology of such
intuitions. But the important question is that of which particular moral norms drive those
intuitions. When we put some pressure on those intuitions, the suggestion that they are
connected with norms of fairness seems correct. When we find ourselves convinced of an
incompatibilist view, we typically think that it would be wrong to hold some individual
responsible with regards to some action. More specifically, we think that it would be
wrong to hold such an individual responsible in some regard precisely because we
262 R & MS, p. 85.
263 R & MS, p. 94. Wallace suggests, in an interesting footnote (fn. 16, p. 94) that his view also has affinities with that developed by H. L. A. Hart, in his discussion of “moral liability responsibility” in his “Postscript: Responsibility and Retribution,” in his (1968), Punishment and Responsibility: Essays in the Philosophy of Law (Oxford: Clarendon Press), pp. 210-37. Hart’s claim is that people are properly held responsible, in his sense of “moral liability responsibility” when and only when they deserve blame. If we view notions of desert as claims about the fairness of impositions of suffering or hard treatment, then Hart’s view is analogous to Wallace’s claim that norms of fairness grant the standards of appropriateness for holding individuals morally responsible (at least in the sense of holding them liable for their actions).
271
believe that this would involve treating them unfairly. Wallace is correct when he claims
that, “though they are seldom self-conscious about the fact, incompatibilists often appeal
to our distinctively moral interest in fairness, suggesting that it would be unfair to hold
people if determinism is true.”264 Therefore, Wallace’s suggestion that the proper
conception of ‘appropriateness’ in judging ascriptions of responsibility is one that
answers to moral norms of fairness is convincing.
The question then arises, however, of how we are best to make sense of the norm of
fairness that is connected to the appropriateness of judgements of responsibility. More
needs to be said by way of interpretation of this norm. Wallace’s suggestion is to
decompose it into two constituent conceptions of fairness, each leading to an independent
principle governing ascriptions of responsibility. Wallace’s two conceptions are of
fairness as desert and fairness as reasonableness. As Wallace sees it, the concern of
fairness in terms of desert is connected to the accuracy of our judgements of the
responsibility of some agent for some particular act, whereas the concern for fairness in
terms of reasonableness is connected to our assessment of the capacities of the agent in
question.265
As an example of the first kind of judgement, of fairness-as-desert, Wallace gives this
kind of case:
264 R & MS, pp. 93-4.
265 R & MS, pp. 106-9.
272
Take, for example, a case of forcible constraint, where an agent is physically prevented from carrying out some moral obligation that we accept. The incompatibilist might say that it would be unfair to hold the agent responsible in this case, because the agent does not deserve to be blamed or sanctioned…266
Wallace’s idea of ‘fairness as desert’ thus involves the thought that no one deserves to be
held responsible for outcomes that do not accurately reflect upon the intentions of the
agent in question. As can be seen, this is closely related to the ‘Quality of Will’ account
of the excuses developed by Strawson.267 Strawson holds that the reactive attitudes are
not in place when they are not a reaction to the actual ‘quality of will’ evinced by the
agent to whom they are directed. Wallace’s example of the physically constrained agent
presents just such a case – such an agent is not properly held morally responsible for his
omission, because his omission does not reflect upon the quality of his will, but instead is
a direct result of his being physically prevented from undertaking his duty.
Wallace makes this Strawsonian point within the context of his normative conception of
responsibility by saying that it is unfair to hold such an agent responsible in this sort of
case, in the sense that such an agent does not deserve to be held responsible where his
action or omission does not reflect his quality of will. The idea is simple: it is unfair –
because undeserved – to hold an agent responsible in cases where that agent has done
nothing wrong (where assessments of right and wrong apply to the quality of will evinced
by an agent, rather than applying to his ‘mere’ physical actions).
266 R & MS, pp. 106-7.
267 See the discussion of Strawson’s ‘Quality of Will’ criterion above, in Section 2(B) of Part 2.
273
Later in his book, Wallace codifies a general principle based on this idea of fairness-as-
desert:
The Principle of No Blameworthiness without Fault: It is fair to hold some agent, S, responsible for doing x only if, in doing x, S really did something wrong.268
The Principle of No Blameworthiness without Fault is, in effect, a slightly more formal
restatement of Strawson’s ‘Quality of Will’ account of when judgements of responsibility
are appropriate. It is best seen as a principle concerned with the accuracy of judgements
of responsibility. A judgement of responsibility is inaccurate when it does not track the
moral properties of the actions of the agent in question, and it will fail to track those
moral properties when it is not concerned with that agent’s ‘quality of will’. The core
intuition here is clear: It is unfair to hold an agent (morally) responsible for some action if
one’s assessment of the agent’s ‘quality of will’ is inaccurate. For the sake of punchiness,
therefore, I want to re-label Wallace’s Principle of No Blameworthiness without Fault as
the Principle of Accuracy. It elaborates a deep and uncontroversial norm of fairness,
which we should fully endorse. It is, no doubt, (at least) part of the truth about when
ascriptions of responsibility are fair and when they are unfair. What is in issue, as shall be
seen, is how large a part of the truth is given by the Principle of Accuracy.
Turning to the other aspect of Wallace’s elaboration of fairness, we recall that the other
sense of fairness under consideration involves an idea of fairness-as-reasonableness. In
268 This paraphrases the content of the principle of fairness developed by Wallace at R & MS, pp. 135.
274
order to give a sense of what Wallace has in mind when he speaks of ‘fairness-as-
reasonableness’, he gives an example of the following kind:
Consider an example: a young child does something morally wrong – lies to her parents, say, about whether she has cleaned her room.269 There may well be good reason to scold or punish the child in this situation, but I take it we would think it unfair to hold the child fully responsible for her deed, in the way we would ordinarily hold morally responsible an adult who lies for personal advantage.270
The idea here is that it is reasonable to hold some other agent responsible for their
actions only when that agent has the full set of normal adult capacities. It is clear here
that, again developing the parallelism between Wallace’s account of these matters and
that of Strawson, Wallace’s idea of fairness-as-reasonableness basically expresses the
same idea as that involved in Strawson’s Capacity criterion for when it is in order to hold
some other agent responsible. According to Strawson’s Capacity criterion, it is
inappropriate to hold responsible someone who is not a normal adult human, and thereby
a competent partner or interlocutor in adult relationships (of whatever kind). This is
exactly the idea that Wallace also has in mind here in the example of the misbehaving –
but non-responsible – child. It is not that it would be inaccurate to describe the child’s
actions as an instance of wrongdoing (let us suppose), and so there is therefore no
violation of the Principle of Accuracy. But it would nevertheless be unreasonable, and
therefore unfair, to hold the child – who does not possess the full set of adult agential
capacities – responsible for this moral infraction.
269 Let us grant Wallace the benefit of the doubt that misleading someone about this sort of triviality, for no discernable gain, really does count as even a prima facie instance of moral wrongdoing, rather than a sort of sub-moral peccadillo.
270 R & MS, p. 108.
275
Just as Wallace codifies his version of the ‘Quality of Will’ criterion for holding agents
responsible into a general principle of ‘fairness-as-desert’ (thereby generating the
Principle of No Blameworthiness without Fault, or what I have called the Principle of
Accuracy) so too he has also codified his version of the ‘Capacity’ criterion into a general
principle of ‘fairness-as-reasonableness’. This leads to the following general principle
regarding when it is fair to hold individuals responsible for their actions:
Principle of No Blameworthiness without the General Capacity for Reflective Self-Control: It is fair to hold an agent, S, responsible for doing x only if, at the time of the action, S possessed the powers of rational self-control: the general capacity to appreciate the moral reasons for doing otherwise and to regulate his behaviour in light of them.271
As Wallace describes this principle, it makes sense of the idea of fairness-as-
reasonableness that stands alongside the idea of fairness-as-desert that is captured by the
Principle of Accuracy. As Wallace puts it:
… it would be unfair to hold someone to moral obligations one accepts, in the absence of the powers of reflective self-control, in the sense that it would be unreasonable to hold the person to moral obligations under those conditions. To make this proposal is, in effect, to postulate a moral principle of reasonableness, namely that it is unreasonable to demand that people do something – in a way that potentially exposes them to the harms of moral sanction – if they lack the general power to grasp and comply with the reasons that support the demand. I take this to be a principle of reasonableness that any competent moral judge would endorse, on reflection; like the principle of desert applied to [above] (“no blameworthiness without fault” [or, in my terms, the Principle of Accuracy]), it expresses an abstract moral conviction in which we have the utmost confidence.272
271 I am here largely following the formulation of the principle to be found in Gideon Rosen, (2002), “The Case for Incompatibilism,” Philosophy and Phenomenological Research, LXIV. 3, pp. 699-706.
272 R & MS, p. 161
276
Given the content of Wallace’s second principle, and its similarity to Strawson’s
‘capacity criterion’, and for the sake of making for an easier discussion, I shall here re-
christen it the Capacity Principle. Wallace is correct in stating that, like the Principle of
Accuracy, the Capacity Principle expresses a deep and firmly held moral truth. Again,
the more interesting question is not so much whether the Capacity Principle provides one
necessary condition for the fairness of blame (among other conditions), but the more
significant question of whether, as Wallace goes on to claim (echoing Strawson, and his
Claim (φ)273), the Capacity and Accuracy principles together constitute jointly sufficient
conditions for fair ascriptions of blame.
What is also much more questionable than the plausibility of the Capacity Principle (as
expressing simply just a necessary condition for the fairness of blame), though, is
Wallace’s characterization of his first principle as a principle of desert, as against this
second principle being a principle of reasonableness. This distinction actually runs no
deeper than the particular way in which Wallace has chosen to express each principle. It
is, in fact, a distinction without a (significant) difference. For consider how easy it would
be to express either principle in either the language of ‘desert’ or of ‘reasonableness’.
Wallace suggests that it is unfair to hold responsible the individual who has done nothing
wrong, because such an individual does not deserve to be held responsible. This is of
course true, but it is just as natural to say that it would be unreasonable to hold that
individual responsible, given that he has done nothing wrong.
273 See Part 2, especially Section 3(E) above, for Strawson’s claim (φ).
277
Similarly, Wallace suggests that, in the case of his lying young child, it is unfair to hold
that child responsible because it would be unreasonable to do so. This, again, is
reasonable enough insofar as it goes, but it is equally plausible to say that it would be
unfair to hold the child responsible because she does not (by virtue of her undeveloped
capacities) deserve to be held responsible. Wallace’s claim that scolding or punishing the
lying child “would not be unfair … in the sense that the child does not deserve to be
punished or blamed”274is therefore quite jarring. Indeed, the claim is simply false, as we
would, indeed, say that, given that the child lacks the requisite capacities, it would be
unfair to punish or scold her, in the sense that she does not deserve to be scolded or
punished.
In both cases, appealing to considerations of ‘reasonableness’ and ‘desert’ actually does
no more than to refer the question back, one level further down as it were, to the
underlying considerations that explain the undeservingness or unreasonableness in each
case. We would therefore be much better off, in terms of clarity, were we simply to
jettison Wallace’s overdrawn distinction between fairness-as-reasonableness and
fairness-as-desert. What matters in making sense of these issues are actually the more
fundamental principles of fairness to which these considerations of ‘reasonableness’
and/or ‘desert’ themselves refer.
In the case of the first principle, the underlying consideration, to which Wallace’s idea of
‘desert’ refers the matter in question, is the concern with the accuracy of our judgements.
274 R & MS,. p. 108.
278
This is why I have renamed his first principle as the Principle of Accuracy. It is
uncontroversial that inaccurate ascriptions of responsibility are unfair, and that they are
unfair in virtue of their inaccuracy. Indeed, the Principle of Accuracy should be seen as a
wholly uncontroversial moral norm. By contrast, it is somewhat more difficult to describe
the consideration, or set of considerations, to which Wallace’s idea of ‘fairness-as-
reasonableness’ refers in the second kind of case. Contra Wallace, and as I have remarked
above with regard to the case of the lying child, it is very natural to describe this
consideration as a consideration relating to desert, just as it is also natural to speak in
terms of reasonableness. What is in question is simply the core issue of when it is fair to
hold someone responsible for some act, in the sense of when it is reasonable to view them
as deserving to be held responsible with regard to that act. (As against, with the Principle
of Accuracy, the more tightly defined question of whether or not it is even accurate to
link the assessment of that agent with that action.)
Wallace’s Capacity Principle can best be seen as a particular interpretation of this more
general consideration, regarding when it is reasonable to view an individual as deserving
to be held responsible for some act. I therefore want to distinguish the content of
Wallace’s particular principle (the Capacity Principle) from the broader issue of the
general set of considerations to which it is designed to offer an interpretation. For the
sake of terminological devilment, I’m going partially to reverse Wallace’s labels, and
refer to the general set of considerations relevant for deciding on the fairness of holding
an agent responsible as considerations of Reasonable Desert. My general conceptual
claim is that, accepting the independent force of the Principle of Accuracy in identifying
279
a necessary condition for when it is fair to hold an individual responsible, the remaining
sufficient conditions that tell us when it is fair to hold an individual responsible for some
action are determined by the content of our Norms of Reasonable Desert. Thus, to apply
Wallace’s Schema (N) in deciding when an individual is responsible for some act, is to
come to a settled view about the content of our Norms of Reasonable Desert, and thereby
to arrive at a fully considered judgement as to what is fair or unfair in holding individuals
responsible.
As we have seen, Wallace’s view is that the Capacity Principle provides a full
interpretation of the demands of our Norms of Reasonable Desert with regards to holding
agents responsible for particular acts. On Wallace’s view, therefore, the Principle of
Accuracy and the Capacity Principle constitute individually necessary and jointly
sufficient conditions for the fairness of holding agents responsible. On this view, it is
appropriate (because fair) to hold an individual S responsible for some act, x, just when
the Principle of Accuracy and the Capacity Principle are jointly satisfied with regard to
that agent, S, performing action x. Therefore, Wallace has provided us with more or less
a direct transposition of the central claims of Strawson’s view into the language of a
fairness-based normative interpretation of the conditions of responsibility. Wallace’s
view, therefore, can be captured in the following principle:
Principle of Joint Sufficiency (of Accuracy and Capacity):275 It is fair to hold an agent S responsible for doing x if and only if [S really did something wrong in doing x and at the time of the action, S possessed the powers of rational self-control] 275 This is a modified version of the “Sufficiency Thesis” identified by Gideon Rosen in his, (2002), ibid., p. 700.
280
One can think that Wallace’s Schema (N) provides us with the correct way of thinking
about responsibility without thinking that the Principle of Joint Sufficiency gives the
proper interpretation of the ideas of appropriateness or fairness which that schema
involves. Let us assume that it is uncontroversial that the Principle of Accuracy gives us a
necessary condition for the appropriateness of holding an agent responsible for some act.
It is plausible, thereby, to hold (as Wallace does) that the set of individually necessary
(and jointly sufficient) conditions will be given by the conjunction of the Principle of
Accuracy with whatever principle (or principles) serves as the best interpretation of the
Norms of Reasonable Desert (insofar as those norms concern the fairness of holding an
agent responsible over-and-above the basic necessary condition of accuracy).
Nevertheless, one may disagree with Wallace insofar as one might hold that the Capacity
Principle is only a partial, rather than an exhaustive, interpretation of the demands of our
Norms of Reasonable Desert.
While the Capacity Principle uncontroversially provides one of the necessary conditions
for the fairness of holding an agent responsible, it is not (as we shall see in the following
sections) plausibly viewed as providing, together with the Principle of Accuracy, jointly
sufficient conditions of fairness for holding agents responsible. This claim will, of course,
require careful defence, but for my present purposes all that I want to emphasize is that it
is fully coherent to accept Wallace’s normative conception of responsibility, together
with the general schema (N) for making judgments of when ascriptions of responsibility
281
are in order, whilst nevertheless rejecting Wallace’s specific interpretation of when that
schema (N) is satisfied, as captured by the Principle of Joint Sufficiency.
This puts us in a similar position (unsurprisingly, perhaps, given the similarity of their
views) as in the discussion of Strawson above in Part Two. I there accepted that
Strawson’s Capacity criterion provided us with a necessary condition for judging agents
to be responsible. But I rejected Strawson’s stronger Claim (φ) – that is, the claim that we
should give a strong reading of the role of the Capacity criterion as a condition for
granting an exemption from any of the set of moral reactive attitudes. (Under such a
strong reading, the Capacity criterion is seen as giving necessary and sufficient
conditions for when an individual might legitimately be exempted from subjection to any
of the moral attitudes, including the desert-entailing moral attitudes.276)
The rejection of Strawson’s Claim (φ) is parallel to the rejection of Wallace’s Principle of
Joint Sufficiency. To put things into Strawsonian language, we may say that it is
uncontroversial that the Principle of Accuracy provides a full account of the excuses from
being held responsible.277 What is more controversial is finding the proper account of the
exemptions from being held responsible. Strawson’s Claim (φ) is that the capacity
criterion gives necessary and sufficient conditions for exemptions from (any of) the moral
reactive attitudes. In Wallace’s terms, the analogous claim is that it is unfair to hold an
individual responsible for some act, by virtue of fairness demanding that they be
276 On Strawson’s Claim (φ), see especially Sections 3(E)-3(M) of Part 2.
277 For the distinction between exemptions and excuses, see Part 2, Section 2(B).
282
exempted (as against excused) from being held responsible, only when that individual
fails to satisfy the Capacity Principle. In making the case for rejecting Wallace’s
Principle of Joint Sufficiency, we would therefore simultaneously make the positive case
for rejection of Strawson’s Claim (φ).
My claim, therefore, is that we should accept Wallace’s schema (N), but reject his
Principle of Joint Sufficiency. We should reject this principle because, as we shall see,
there is good reason to think that the Capacity Principle does not exhaust our Norms of
Reasonable Desert, and therefore does not provide us (alongside the uncontroversial
Principle of Accuracy) with a full account of the demands of fairness in holding agents
responsible. But in order to make good on this claim, we need first to turn back to the
other main question regarding the interpretation of schema (N): that is, the question of
spelling out a fuller specification of what it is to hold someone morally responsible for
some action. For only with a fully developed account of what it is to hold someone
morally responsible will we be able to specify exactly what the proper standards of
fairness for holding people responsible should be. It is therefore to this other question that
I shall now turn.
3. Conceptions of Responsibility and Blame
It is possible to give a full account of the conditions of fairness for holding individuals
responsible only once we know what it is to hold an individual responsible. For,
depending on the nature of responsibility, the conditions for its fair application could vary
283
significantly. There are, indeed, a number of possibilities with regard to how we should
understand what it is to hold someone responsible: the idea is far from self-interpreting,
and thus requires some careful analysis.
Wallace’s account of what is involved in holding someone responsible is rich and
plausible. I shall suggest that, as with his account of the content of norms of fairness, it
gets at least part of the truth, albeit only a part. Wallace suggests that we should make
sense of the stance of holding someone responsible in terms of the (desert-entailing)
reactive attitudes of resentment, guilt and indignation. Wallace’s suggestion is that “to
hold people morally responsible is to hold them to a particular subclass of demands or
expectations – namely, the moral demands that one accepts – in [a] way distinctively
connected to the reactive emotions.”278 His suggestion is therefore that we understand
responsibility in terms of blame, or, more properly, in terms of blameworthiness (the
distinction will become clear presently). I shall therefore give the name responsibility-as-
blameworthiness to the view of responsibility that Wallace develops. Perhaps
unsurprisingly, given the close allegiance of the two views, Wallace’s account of
responsibility can be seen as a more explicit version of the account of responsibility that
lies in the background of Strawson’s view.
It is one thing to explain the stance of holding someone responsible in terms of its
relation to blame, or to blameworthiness. But this then of course raises the prior question
of how blame, and blameworthiness, are themselves to be understood. Blame for Wallace
278 Wallace, (2002), ibid., pp. 680-81.
284
(and, one might add, for Strawson) involves the subjection of the blamed party to the
reactive emotions and, in particular, to the narrow reactive attitudes such as resentment.
We blame another individual when, in response to some transgression of theirs, we
subject them to one of the (narrow) reactive emotions. As Wallace puts it, “[to] blame
someone is to be subject to one of the reactive emotions in terms of which the stance of
holding people responsible is essentially defined, and these emotions are expressed by the
sanctioning behaviour to which the stance of holding people responsible inclines us.”279
More fully, then, the picture of how responsibility is related to blame and to sanction on
this view (which Wallace calls the “reactive account of moral responsibility”280) is this:
People who are morally responsible may be made to answer for their actions, in the sense that their actions render them liable to certain kinds of distinctively moral responses.281 These responses include most saliently the response of moral blame, which is called for when the responsible agent has done something morally wrong,282 but they extend beyond simple blame to include a range of sanctioning responses as well, such as avoidance, reproach, scolding denunciation, remonstration and (at the limit) punishment. Moral blame and moral sanction may thus be thought of as the special kinds of appraisal to which morally responsible agents are open, and the hallmark of holding people responsible will be a tendency to respond to them with these forms of appraisal.283
The connection here between responsibility, blame and forms of sanction is particularly
significant. The ‘reactive account’ of responsibility, with its emphasis on the way in
279 R & MS, p. 52.
280 R & MS, p. 66.
281 Here, as elsewhere, Wallace’s inspiration (other than Strawson) seems to be Hart’s idea of “moral liability responsibility” in his Punishment and Responsibility, ibid.
282 That is, when the Principle of Accuracy has been satisfied.
283 R & MS, p. 54.
285
which the desert-entailing reactive attitudes are constitutive of the stance of holding
someone responsible, makes sense of the connection between holding someone
responsible and being disposed towards these various forms of sanctioning behaviour. It
is not that holding someone responsible necessarily involves actively treating them in
these sorts of ways, but rather that, in holding them responsible, we at least hold that such
forms of sanctioning behaviour may be appropriate. (In the terminology of Part 2 of this
dissertation, we may say that the reactive account of responsibility is at least weakly
desert-entailing, even if not a strongly desert-entailing). The relationship between
responsibility, blame and sanction on the “reactive account” of responsibility is further
clarified by Wallace here:
Once blame is understood in terms of the reactive emotions, […] we also have a natural and appealing explanation to hand of what unifies the sanctioning responses to which the stance of holding people responsible disposes us (such as avoidance, censure, denunciation, reproach, and scolding). These can all plausibly be understood as forms of behaviour that serve to express the reactive emotions to which we are subject when we blame people for their moral failings.284 What is essential to the harmful moral sanctions, on the reactive account, is their function of expressing the moral emotions of resentment, indignation and guilt; this is the real point of such responses as avoidance, denunciation reproach, censure and the like, and what holds them together as a class. Sanctioning behaviour belongs to the syndrome of responses to which the reactive emotions disposes those who are subject to them, because the connection with reactive emotions is part of the conventional meaning of such behaviour.285 ([Wallace adds in a footnote:] “Perhaps this is what P. F. Strawson means when he writes that a preparedness to acquiesce in the infliction of suffering is “all of a piece” with the reactive emotions.”286)
284 R & MS, p. 67.
285 R & MS, p. 68.
286 R & MS, p. 68, fn. 25.
286
Wallace has here succeeded in making explicit the picture of responsibility, blame and
sanction that lies much more implicitly at the heart of Strawson’s view. Moreover, this
account of what blame is captures very precisely our best understanding of that idea, at
least in the central sense of what we might call ‘moral blame’. Wallace delineates this
conception whilst drawing back from an excessively or unconvincingly desert-entailing
understanding of blame, on which that attitude would be understood as involving the
belief that it was good that the blamed individual should suffer (i.e. an understanding of
blame as necessarily strongly desert-entailing).287 The weakly desert-entailing conception
of blame, on which blame involves at least a broad disposition towards certain forms of
sanctioning behaviour, and the belief at least that such forms of sanction would not be
inappropriate, seems to get things right. Wallace’s understanding of blame makes good
sense of the ways in which blame is desert-entailing, without falling into caricature or
exaggeration.
Wallace’s ‘reactive account’ has to escape another possible pitfall, in that it needs to be
able to make sense of the fact that we can hold an agent responsible without engaging
ourselves in even the mildest version of the attitude of blame (let alone without engaging
in the sort of sanctioning behaviour which expresses that attitude). Wallace’s way around
this potential problem is to stress the distinction between blame and blameworthiness. If I
hold you blameworthy, then I consider you to be an appropriate subject for an attitude of
blame, whether or not I then actually evince any negative reactive attitude towards you,
287 Thus, although this account of blame certainly characterizes blame as being desert-entailing, it need not involve accepting the very strong version of desert as capture by what Scanlon calls ‘The Desert Thesis’, whereby we hold that it is (morally) good when some individual suffers. See Scanlon, (1998), What We Owe to Each Other, ibid., pp. 274-77.
287
or avoid doing so. This is why the “reactive account of responsibility” should be
understood, in my terms, as an account of responsibility-as-blameworthiness rather than,
so to speak, responsibility-as-blame. As Wallace makes the distinction:
On the reactive account, blame requires that you actually are subject to a reactive emotion, but an emotional response of this sort is not necessarily required for you to hold your colleague morally blameworthy. It suffices for you to believe that indignation or resentment would be fitting responses on your part, and that they would be fitting because your colleague has done something morally wrong. This seems to correspond to our ordinary judgements of moral responsibility quite exactly.288
With this final finessing of Wallace’s position now in view, we have a full account of
what it is to hold someone responsible, in terms of responsibility-as-blameworthiness.
This account has a disjunctive structure, given that, when we hold another individual to
be responsible, in the sense of blameworthy, it may either be the case that we thereby
blame them (i.e. subject them to one of the desert-entailing reactive attitudes), or we may
instead forswear the attitude of blame itself, but nevertheless hold that such an attitude
would nevertheless be appropriate. This gives us a conception of responsibility with the
following content:
Responsibility as Blameworthiness: To hold S responsible for some act, x, is to hold that S is blameworthy for doing x. This means that we either blame S for doing x, or else at least hold that it would be appropriate to blame S for doing x.
Wallace’s confidence in the precision of his own view (recall that he tells us that it
“seems to correspond to our ordinary judgements of moral responsibility quite
288 R & MS, pp. 76-7.
288
exactly”289) is to a large degree well placed. The ‘reactive account’ of responsibility-as-
blameworthiness does, indeed, capture the nature and structure of our ordinary
judgements of moral responsibility. What is more controversial is the question of whether
this is the only significant sense of responsibility with which we need to concern
ourselves when developing a substantive interpretation of the formal schema (N), for the
legitimate ascription of judgements of responsibility.
My suggestion is that Wallace’s description of responsibility-as-blameworthiness does
not exhaust all aspects of the range of moral phenomena with which we are concerned
when we are concerned with moral responsibility. I shall therefore now turn to a rival
conception of moral responsibility which is, I believe, just as important and central to our
moral lives as the notion of responsibility-as-blameworthiness. This rival conception is a
conception of responsibility-as-assessability, and, as we shall see, it is not best captured
by the approach involved in the reactive account. This ‘thinner’ conception of
responsibility often operates alongside or, so to speak, in the background of, the ‘thicker’
version of responsibility-as-blameworthiness. But it is vitally important to keep these two
versions of responsibility conceptually distinct from one another.
This thinner idea of responsibility-as-assessability is a very straightforward one. It is
simply the idea that an agent is in a sense responsible for some action, x, when he can
properly be assessed in terms of his moral character with regard to the performance of
that action. This is not the thicker notion of responsibility-as-blameworthiness, according
289 R & MS, p. 77.
289
to which, for an agent to be responsible, it must both be the case that he can properly be
assessed, morally speaking, and moreover that he is, in virtue of that moral assessment,
an appropriate target of blame. Rather, a judgement of responsibility-as-assessability
endorses only the first half of this conjunction – i.e. that an agent is properly open to
moral assessment with regard to some action – whilst remaining resolutely agnostic on
the second element, that is, the further question of whether blame, and the sanctions
associated with it, can legitimately be directed at that agent in connection with that
action.
This thin notion of responsibility-as-assessability was frequently deployed in my
discussion of Strawson in the previous part of the dissertation,290 often so as to contrast
simple judgements of moral character from more emotionally loaded forms of assessment
that involve the deployment of the (desert-entailing) reactive attitudes. Negative moral
judgments are, of course, forms of disapprobation, but such judgements may be made
without the involvement of any desert-entailing moral emotion, such as resentment or
indignation. But, whilst holding an agent responsible in the sense of responsibility-as-
blameworthiness is always thereby also to hold them responsible in the sense of
responsibility-as-assessability, the converse is not the case.
These two forms of responsibility might, erroneously, be taken always to stand or fall
together, and this view has a certain degree of initial plausibility, given that so many
290 See for example the discussion in Section 3C.
290
cases of real-world moral judgement involve both forms of responsibility.291 But,
nevertheless, the two forms are quite distinct from one another, as becomes clear when
we consider cases where a judgement of moral ‘faultiness’, even if it is accompanied by
certain moral emotions (such as sadness, regret, or so on) is nevertheless not
accompanied either by the occurrence of a desert-entailing reactive attitude, or by the
judgement that a desert-entailing reactive attitude would be appropriate in that case. We
should judge, therefore, that it is possible to hold an agent responsible in the sense of
responsibility-as-assessability without thereby also holding him or her responsible in the
sense of responsibility-as-blameworthiness.
It is worth pausing for a moment at this point to consider the terminology that I am here
using. I have spoken of both the ideas of blameworthiness and of assessability as forms of
the core idea of responsibility. Indeed, it is on this basis that I am here criticizing Wallace
for assimilating all judgements of responsibility to the sub-type of judgements of
responsibility-as-blameworthiness. But one might just as easily say that, in fact, the idea
of moral assessability is not itself a notion of responsibility, as such, at all but rather a
somewhat distinct idea that is, so to speak, conceptually upstream of the idea of
responsibility. We might say that, in fact, such judgements of assessability are
judgements regarding whether agents (in connection with particular actions) are even
candidates for being held responsible. To say that an agent was in this sense (morally)
291 Both Strawson and Wallace take the view that, when we hold someone responsible, it is always in the sense of responsibility-as-blameworthiness as well as in the sense of responsibility-as-assessability. I have argued against this at length in my discussion of Strawson, and will not rehearse the points I made earlier in detail here.
291
assessable would then be a precondition for making a judgement that such an agent was
responsible, rather than a particular kind or variety of judgement regarding responsibility.
I have some sympathy with this line of criticism. Nevertheless, in an important sense, this
terminological dispute is of only marginal importance. What matters is simply that we
should be careful to distinguish two separate ideas: the first being the idea that an agent is
a proper object of moral assessment with regard to a particular action, and the second
being the idea that an agent is a proper object of blame – in Wallace’s sense of subjection
to (at least) the desert-entailing reactive emotions (and, derivatively, to reactions such as
scolding, punishing, and so on). We can acknowledge the point that, in a sense, it is the
second of these two ideas that fits best with our usual idea of moral responsibility, while
the first idea might seem to be, so to speak, too thin a notion by comparison. I shall
nevertheless refer to them both as varieties of responsibility, on the grounds that, firstly,
this does no violence to normal usage; secondly, that this fits with many aspects of recent
philosophical usage;292 and thirdly, for the sake of terminological tidiness, as this makes
it easier to think about both notions in terms of Schema (N).
My claim is absolutely not that these two different ideas of responsibility are in some
sense rivals as interpretations of the ‘correct’ idea of responsibility. Rather, both of these
kinds of responsibility judgements are important and central features of our moral lives. It
is not a question of choosing between them, but rather of making sense of the conditions
of application of each of them. For our current purposes, what is most important to
292 For example, my idea of responsibility-as-assessability is significantly similar to T. M Scanlon’s idea of “attributive responsibility”. (See Scanlon, What We Owe to Each Other, (1998), ibid., Chapter 6.)
292
realize is that the operative substantive standard of fairness in relation to which a
judgement of responsibility should be held, under Schema (N), to be fair, will vary
depending on whether we are dealing with an idea of responsibility-as-blameworthiness
or an idea of responsibility-as-assessability. For what it is to hold someone responsible
will make a difference to what makes it fair to hold someone responsible in any particular
way. We should not be surprised, for example, if it turns out (as I will argue that it does)
that the standards for when it is fair to hold someone responsible, in the sense of
responsibility-as-blameworthiness, are much more stringent and much more difficult to
satisfy than are the standards for when it is fair to hold someone responsible in the sense
of responsibility-as-assessability.
In his forthcoming discussion of blame, T. M. Scanlon makes the following illuminating
claim:
Blame is a familiar aspect of moral experience, but it is surprisingly unclear exactly what it involves. Accounts of blame tend towards two ideas. The first is essentially evaluative: to blame someone is to arrive at a negative assessment of his or her character. The second is punitive: blame is a kind of sanction, a milder form of punishment.293
I say that this claim is illuminating, but not because I would want to phrase things in quite
this way. For the purposes of my argument, I am content to accept Wallace’s
characterization of blame as involving subjection to the desert-entailing moral emotions,
which puts the notion squarely and unambiguously on the ‘punitive’ side of Scanlon’s
distinction, whereby blame is a “milder form of punishment”. But I think that Scanlon’s
distinction, if modulated from being about blame to being about responsibility, gets
293 T. M. Scanlon, (2008), “Blame”, (unpublished manuscript version), p. 125.
293
things right. The idea of responsibility-as-assessability is simply the idea of an agent
being responsible insofar as his character is a proper target for moral assessment or
evaluation. It is, in Scanlon’s terms, “essentially evaluative”. (One might rechristen this
notion “responsibility as aptness for evaluation” if this were not so very cumbersome.)
By contrast, the idea of responsibility-as-blameworthiness, given the sense of
blameworthiness which I have adopted by way of Wallace and Strawson, fits the bill of
the rival kind of “punitive” conception. To judge that an agent is responsible, in the sense
of responsibility-as-blameworthiness, is indeed to engage in “a milder form of
punishment” (or at least, more weakly, to hold that “a milder form of punishment” would
not be inappropriate in this case).
Scanlon rejects both the evaluative and the punitive conceptions, presenting instead a rich
and complex account of blame and blameworthiness, which might be seen as falling
somewhere in the conceptual space in between the two. My own strategy is, in a sense,
more modest, and less revisionist. It is to take both the evaluative and the punitive
conceptions of responsibility at face value, and then to determine the appropriate
standards of fairness that are involved in the application of each variety of responsibility.
Scanlon suggests that blame really is neither the “essentially evaluative” nor the punitive
variants described above. On my view, like Strawson’s and Wallace’s, blame is properly
understood to be just this punitive notion, or as “a milder form of punishment”. But
responsibility, on my view, comes in both variants – there is both an evaluative and a
punitive sense of responsibility, and each variant is subject to its own distinctive
normative standards.
294
In the following sections, therefore, I want to consider how we should best come to a full
account of the standards of fairness implicated in the application of each of these two
kinds of responsibility. To put things more simply, I want to discover when it is fair to
evaluate others, or to hold them responsible in the sense of responsibility-as-assessability
and, at the same time, I want also to address the separate question of when it is fair to
hold others responsible in the (‘punitive’) sense of responsibility-as-blameworthiness. In
order to undertake this task, it will be necessary to investigate the norms of fairness that
are implicated in holding others responsible and, especially, to give a substantive account
of the content of the Norms of Reasonable Desert with regards to both assessment and
blameworthiness. It is to these questions that I now turn.
4. Fairness, Responsibility and Blame: The Case for Restricted Incompatibilism
I shall start with the more difficult of the two questions facing us, which is the question of
when it is fair to hold someone responsible in the sense of responsibility-as-
blameworthiness. Once this question has been addressed, I shall turn, in later sections, to
the (more straightforward) question of when it is fair to hold someone responsible in the
sense of responsibility-as-assessability. As we have seen above, Wallace (following
Strawson), endorses the Principle of Joint Sufficiency, whereby he holds it to be fair to
blame some agent when and only when that agent satisfies both the Principle of Accuracy
and the Capacity Principle. The suggestion for which I shall argue is that Wallace is
wrong in holding that the Principle of Joint Sufficiency is a plausible account of when it
295
is fair to hold some agent to be responsible, in the sense of responsibility-as-
blameworthiness, with regard to some action. I allow that the Principle of Accuracy and
the Capacity Principle each give us necessary conditions for when it is fair to hold
someone blameworthy, but I deny that they are jointly sufficient conditions. In
considering the Norms of Reasonable Desert, with regard to blameworthiness, we need
something over and above the satisfaction of these two principles in order for an agent to
be fairly held to be blameworthy. Or so I shall argue.
I’ll begin with a case, borrowed from Galen Strawson, which is perhaps the central and
most basic story available to us when we come to think about fair conditions for
blameworthiness. It is the story of Heaven and Hell:
An old story is very helpful in clarifying this question [of clarifying the meaning of what Galen Strawson calls ‘true’ moral responsibility]. This is the story of heaven and hell. As I understand it, true moral responsibility is responsibility of such a kind that, if we have it, then it makes sense, at least, to suppose that it could be just to punish some of us with (eternal) torment in hell and reward others with (eternal) bliss in heaven. The stress on the words ‘makes sense’ is important, for one certainly does not have to believe in any version of the story of heaven and hell in order to understand the notion of true moral responsibility that it is being used to illustrate Nor does one have to believe in any version of the story of heaven and hell in order to believe in the existence of true moral responsibility. On the contrary: many atheists have believed in the existence of true moral responsibility.294
When Galen Strawson talks about “true moral responsibility” here, what he seems to
have in mind is the notion of full-blooded responsibility-as-blameworthiness, rather than
any thinner notion of mere moral assessability, or any revisionist replacement for the
standard, strong version of the idea of blameworthiness. What is in question is the desert-
294 Galen Strawson, (1994), “The Impossibility of Moral Responsibility,” Philosophical Studies, 75: 5-24, pp. 7-8.
296
entailing conception of responsibility, whereby a judgement of responsibility is, at
minimum, a judgement that certain desert-entailing moral emotions can appropriately be
directed towards some agent for some action. When Strawson talks about what would
need to be the case for the story of Heaven and Hell even to “make sense”, I take it that
what he has in mind is (to put things in Wallace’s terms) is the question of when it could
even possibly be fair to punish some wrongdoer with eternal torment. Blaming someone,
of course, is not as bad as condemning them to the eternal fire, but both are forms of hard
or unwelcome treatment. Therefore, the case of Hell presents us with a particularly stark
example of the general question which we need to answer: that is, the question of when it
can be fair to impose some ‘punitive’ sanction on an individual in response to some
particular action of theirs (or, at any rate, when we can fairly hold that such a sanction
would be appropriate).
The case of Heaven and Hell can help to show how satisfaction of the demands of the
Accuracy and Capacity principles may not be enough for a judgement of responsibility-
as-blameworthiness to be in good order. To see why, imagine yourself in the position of a
judge who is attempting to justify to a hellbound individual (let’s call him Harry) his fate
in Hell. We can imagine that Harry has performed, deliberately and with malice of
forethought, some action or series of actions of sufficient despicableness as to merit the
eternal fire (given standard beliefs about sinfulness, and so on). We can imagine a
conversation with something like this structure:
297
Judge: You’ve behaved badly, and so it makes sense, in the sense of not
being unfair, for you to bear the consequences of that bad action. In this case, the
consequences are that you go to Hell. I take no pleasure in your fate, and I don’t see your
future suffering as a source of good, or as a teleological state at which I must aim. I’m not
making some extravagant claim of positive desert. It’s a shame that it’s come to this, but
Hell is a fair and appropriate response to what you’ve done. I fear, old chap, that you can
have no complaint given the circumstances.
Hellbound Harry: I admit that I did behave in a morally bad way, but how can it be
fair to punish me for what I did? After all, although my action was under my control, in
the sense that I did what I intended to do, the fact that I found myself in that particular
choice-situation when I acted was itself outside of my control. Also, the fact that,
psychologically speaking, I was set up in one way rather than another, and was bound to
react to that choice situation in the way in which I did react, was also something that is
beyond my control. This means that it is unfair to blame me, and a fortiori, it’s unfair to
punish me with eternal torment!
Judge: You don’t deny that what you did was, indeed, morally faulty, and
you don’t deny that you have the general capacities of reflective self-control that are
enjoyed by adult human agents? That is, you don’t deny that the judgement of eternal
damnation satisfies both the Accuracy Principle and the Capacity Principle?
298
Hellbound Harry: Not at all. I don’t deny these claims. Your judgement is, in a sense,
accurate. I did what I did, and it evinced a faulty, morally criticizable Quality of Will.
Moreover, I have all the capacities for self-control that any competent human agent might
ever want to have. But you’re begging the question if you assume that, if I satisfied these
two principles, I am thereby bereft of grounds for complaint against my punishment.
Whilst I don’t deny that I acted, freely and autonomously, in this
morally bad way, and that I can be morally assessed on that basis, I nevertheless hold that
it would be unfair to hold me responsible, in the sense of blameworthy, for acting in that
way. It would be unfair because I am subject to what Nagel calls constitutive luck,295 as
well as luck in the circumstances in which I found myself.296 To put things another way,
although I’m in a sense responsible for acting in the way I acted, in the sense that I’m
assessable for it, and in the sense that it does reflect what my character is like, I’m not
responsible for the fact that my character is that way, or for the fact that I found myself in
the situation in which I was placed.
My responsibility, such as it is, is shallow, rather than deep. It goes
back only so far, to facts about my character, but I am not responsible for that character
itself. It is too shallow to be the basis for a judgement of blameworthiness; too shallow to
be used in justifying my consignment to Hell. Although I did act in a particular way, the
295 See Thomas , “Moral Luck,” (1976) Proceedings of the Aristotelian Society, supplementary volume L, reprinted in Nagel, (1979), Mortal Questions, (Cambridge: Cambridge University Press), pp. 24-38. Nagel describes the phenomenon of “constitutive luck” in terms of luck regarding “the kind of person you are, where this is not just a question of what you deliberately do, but of your inclinations, capacities and temperament.” (p. 28)
296 See also Thomas Nagel, ibid., p. 28.
299
fact that I acted in that way is something for which I do not have deep responsibility, at
least insofar as we understand responsibility as being connected to liability to blame and
punishment. In a sense, what I have done, although it is in one (relatively shallow) sense
down to me, is more fundamentally due to matters that are themselves beyond my
control. And it is not fair to blame me for matters that are ultimately beyond my control.
As Bertrand Russell once put it, “you can act as you please, but you can’t please as you
please”.297 For myself, I did as I pleased, and I don’t deny it; but I did not please as I
please, for this is simply impossible.298 And so I do not deny that I did as I wanted to do,
but I do deny that I am, in turn, responsible for what I wanted. You cannot fairly punish
me for things that are ultimately beyond my control.
Judge: I’m not sure I understand your complaint. You don’t deny that you
acted with moral faultiness, and you don’t deny that you have the full set of capacities for
self-governing, autonomous agency. I’m trying to think what else you might want or
need, in order for a judgement of responsibility-as-blameworthiness to be justifiably, by
which I mean fairly, in place. Here’s the best I can come up with. Perhaps it would be
unfair to hold you responsible for what you have done if you did not have an adequate
opportunity to avoid this instance of blame or punishment.299 Or, to put a similar point in
a different way, perhaps it would be unfair to blame you if you did not have the ability to
297 Quoted by Galen Strawson in his (1986), Freedom and Belief, (Oxford: Clarendon Press), p. 49.
298 See John Locke, (1975) [1689], ibid., Book II, Chapter XXI, “Of Power”, §25: “For to ask whether a Man be at liberty to will … which he pleases, is to ask, whether a Man can will, what he wills; or be pleased with what he is pleased with. A Question, which, I think, needs no answer: and they, who can make a Question of it, must suppose one Will to determine the Acts of another, and another to determine that; and so on in infinitum.” (p. 247)
299 See T. M. Scanlon, “Blame,” ibid., pp. 179, 201-7.
300
exercise your general powers in the circumstances relating to this particular action.300 Do
you deny that you had an adequate opportunity to avoid Hell, or that you had the ability
to exercise your powers with regard to the actions that have condemned you to eternal
damnation?
Hellbound Harry: This talk of abilities and opportunities isn’t particularly
illuminating; indeed, quite the opposite. In a sense, I had all the abilities and
opportunities that I could want. If I’d chosen to act otherwise than I did, then I could have
acted otherwise than I did. But my point is deeper. It is that although I acted voluntarily
in a particular way, with all the abilities and opportunities I could possibly have for doing
otherwise, in the end the fact that I acted in one way rather than another is due to factors
that are beyond my control. I am not the sole ground of my action – it flows back beyond
me to facts about my environment and my mental constitution for which I am not, in turn,
responsible.
The deep sense of responsibility which could justify the imposition
of blame, punishment, or the desert-entailing reactive attitudes, demands that I be the sole
and final cause of my action. When you impose blame, punishment, or the desert-
entailing reactive attitudes, you are not just grading me for how I happen to be, but
imposing hard and unwelcome treatment upon me. You’re doing something unwelcome
and hurtful to me when you do this, and for that you need a special justification, if you
are not thereby to treat me unfairly and unjustifiably. But I am not the sole and final
300 R & MS, Chapter 7, “The Lure of Liberty”, pp. 195-225.
301
cause of my own actions, and therefore it is unfair to hold me responsible – in the sense
of blameworthy – for those actions. You simply cannot justify this kind of treatment to
me, or to anyone else who is to suffer it.
What should we make of Harry’s protestations? Some will find the deep norm of fairness
with regard to judgements of responsibility, which Hellbound Harry grasps towards here,
to be broadly plausible; others may find it excessive, or to involve an implausible kind of
conceptual overstretch. The core thought involved in this norm of fairness is a powerful
one, although it is curiously difficult to pin down. One way of expressing it is to say that
it is unfair to blame someone (i.e. to see them as fairly being subjected to forms of hard
or unwelcome treatment) simply for how things are in the world, as opposed to the
contrasting fairness of blaming them for features of their actions and characters that are
ultimately due to facts only about themselves. A certain seductive idea of agency
imagines a contrast between (a) how things happen to be, and (b) the actions that people
perform. As Nagel puts it, “something in the idea of agency is incompatible with actions
being events, or people being things.”301 But any plausible understanding of the nature of
human agency involves seeing (b) as a subset of (a), rather than as a genuine contrast
class.
If the seductive idea of agency were sustainable – that is, if actions were not events
among others – then it might conceivably be fair to hold people responsible (in the sense
of blameworthy) for the actions they performed (subject only to the Accuracy and
301 Thomas Nagel, “Moral Luck”, in his (1979), Mortal Questions, ibid., p. 37.
302
Capacity principles). In other words, if the ‘seductive idea’ of agency were to be
sustainable, then Harry’s complaints might themselves be unsustainable. Perhaps then he
could not appeal to constitutive luck, or to the dependence of his character on things other
than his character – and so we could justify his punishment to him without difficulty.302 It
would (or at least could) be fair to hold Harry blameworthy. But this ‘seductive idea’ of
agency is not remotely plausible, given a naturalistic understanding of human agency,
and this generates our predicament. Actions are events among others, and, given that it is
unfair to hold individuals blameworthy for how things simply happen to be – to hold
them responsible for the unfolding of ‘mere events’ – it is unfair to hold them
blameworthy for their actions, when those actions are themselves also ‘mere events’,
unfolding as the result of factors beyond the scope of individual responsibility.
The unfairness in blaming Harry derives from the fact that Harry’s actions are not
grounded and explained only by facts about Harry, but reach beyond him, backwards in
their origins. Hence, Harry has an intelligible complaint against the fairness of blame. In
a strange sense, he is both victim and perpetrator of his bad actions. He performs those
acts, and can fairly be held morally assessable for those acts, but he is also the ‘site’ at
which his bad character formed, the unavoidably passive victim of bad luck of different
kinds. We can assess Harry-as-agent, morally speaking, without any injustice to Harry-
as-victim. But when we blame or punish Harry-as-agent, we cannot do so without also
302 Not to endorse the claim that any particular punishment, even as much as Hell, was appropriate in Harry’s case (whatever its details might be, and whatever his transgressions). Rather, it is just to say that it would at least make sense, in Galen Strawson’s terms, that the punishment of Hell could be appropriate in his case. We could at least begin to understand how it might possibly be fair to hold Harry blameworthy in this way.
303
blaming or punishing Harry-as-victim, and this is something that we could not hope to
justify to him. It is a paradigm of unfair treatment. It fails to take seriously the regressive
character of responsibility-as-blameworthiness.
This regressive aspect to our ordinary idea of responsibility-as-blameworthiness seems to
be what drives, and what explains, the recalcitrant judgement that there is some residual
unfairness in Harry’s case. We can allow that Harry has all the abilities, capacities and
opportunities that any agent might hope to have. He did what he did, and for reasons that,
we may suppose, are robustly and unproblematically his own. There is no failure of free
or autonomous agency in Harry’s case, and he can be properly evaluated for his conduct
by moral standards. And yet the question of his responsibility, in the sense of
responsibility-as-blameworthiness, remains resolutely unsettled.
Somehow, Harry still retains grounds for complaint, undermining our tendency to blame
him for what he has done. Those grounds for complaint are generated by the regressive
character of responsibility. Knowing that we can impute Harry’s actions to his character
and intentions, we are driven naturally to ask whether Harry’s character and intentions
can also be imputed to him, or whether they may result from some exogenous source. We
are driven back to the grounds of the grounds of his action, and to the causes of its
causes: eventually, we reach the point where it makes no sense to impute responsibility to
Harry. The justification for blaming runs out as we track back through this regression.
Blaming becomes unfair; Harry’s complaint seems wholly justified. Without changing
our views as to Harry’s moral character, his virtues or vices, we nevertheless are forced
304
to view him as undeserving of our blame. In an important sense, Harry – like all of us – is
not responsible for what he does, at least in one sense (that of responsibility-as-
blameworthiness) of our two central senses of responsibility.
Harry’s protestations, and their power and plausibility, suggest that the Norms of
Reasonable Desert are not, after all, exhausted by the Capacity Criterion. We therefore
have reason to reject the Principle of Joint Sufficiency. This raises the question of how we
are to identify exactly what sorts of conditions would have to be met, in addition to the
satisfaction of the Capacity Criterion, in order for the Norms of Reasonable Desert to be
satisfied with regard to the ascription of a judgement of responsibility-as-
blameworthiness. Our exploration of Harry’s grounds for complaint suggests the
following additional principle, which, as part of the Norms of Reasonable Desert, stands
alongside the Capacity criterion as a necessary condition which must be met in order for
ascriptions of responsibility-as-blameworthiness to be justifiable. We can call this
principle the Principle of No Blameworthiness without Regressive Responsibility:
Principle of No Blameworthiness without Regressive Responsibility It is fair to hold some agent, S, responsible for doing x only if S is also responsible for the conditions on which S’s doing x depends.303
This principle provides a clear interpretation of the deep incompatibilist intuition that is
an important and central part of our thinking about agency and responsibility, and
303 We recall that the full name for the Capacity Principle is the “Principle of No Blameworthiness without the General Capacity for Reflective Self-Control”; we might, by contract, rechristen this new principle the “Principle of No Blameworthiness without the General Capacity for Regressive Self-Control”. My suggestion is thus that blame could be fair only when an individual’s control over his actions had a property of (unattainable) depth: that the control be not only reflective, but also regressive.
305
provides a plausible account of that intuition as being, at base, a normative intuition about
fairness. It is to a principle of this kind that Hellbound Harry appeals when he argues for
the unfairness of being blamed or punished with Hell, by pointing to the facts of his
constitutive luck, his lack of control over the origins and development of his own
character, and his lack of control over the particular set of choice situations in which he
happened to find himself. Let us call this principle, for convenience, the Regression
Principle. My claim is that, when it comes to the case of responsibility-as-
blameworthiness, the Norms of Reasonable Desert entail that both the Capacity principle
and the Regression principle must be satisfied in order for an individual properly to be
held responsible for some action under Schema (N). It is difficult to provide arguments,
as such, for the Regression principle (after all, to what would one appeal in justifying
such a basic norm?); but it nevertheless provides an elaboration of some of our deepest
and most secure intuitions about the fairness of blame.
The Regression principle can, moreover, be used to explain the attraction of a number of
familiar incompatibilist positions. Standard incompatibilism, of the sort associated with
Peter van Inwagen, holds roughly that: “If determinism is true, then our acts are the
consequences of laws of nature and events in the remote past. But it is not up to us what
went on before we were born, and neither is it up to us what the laws of nature are.
Therefore the consequences of these things (including our present acts) are not up to
us.”304 Van Inwagen would further endorse the additional premise that, if our present acts
are not “up to us”, then we are not responsible for our present acts. The Regression
304 Peter van Inwagen, (1983), ibid., p. 56.
306
principle can be seen as standing in the background of this standard incompatibilist
picture. It is unfair to blame an agent, S, for action x, if it would be unfair to blame that
agent for the conditions on which S’s doing x depends. And, if determinism is true, the
conditions on which S’s doing x depends are given by the set composed of the laws of
nature, on the one hand and, on the other hand, some earlier state of the world, obtaining
long before S’s birth. Now, it is clear that it would be unfair to blame S for either the
laws of nature or for some ancient state of the world. Therefore, it is unfair to blame S for
his doing x.
The Regression principle can explain why we might be convinced by a standard
incompatibilist argument of this kind. If we rejected the Regression principle, and
thought that the Norms of Reasonable Desert were exhausted by the Capacity principle,
then such incompatibilist arguments would seem toothless and beside the point. Such
arguments would seem misdirected because there is nothing in the truth of determinism
that would rob any agent, like Harry, of the full set of capacities for reflective self-
control. Thus, the power of incompatibilist arguments of this kind can be largely
explained by the independent plausibility of the Regression principle.
Nevertheless, the significance of the Regression principle is not restricted to its role in
standard incompatibilist arguments. It can also be shown that, whether or not
determinism is true, we cannot conceive of the sort of agency that could satisfy the
demands of the Regression principle, and thereby legitimate the fairness of blame. This
307
task has been performed by Galen Strawson, by way of his well-known ‘Basic
Argument’. Here is Strawson’s argument in full:
(1) Interested in free action, we are particularly interested in actions that are performed
for a reason (as opposed to ‘reflex’ actions or mindlessly habitual actions). (2) When one acts for a reason, what one does is a function of how one is, mentally
speaking. (It is also a function of other things – but the mental factors are crucial when moral responsibility is in question.)
(3) So if one is to be truly responsible for how one acts, one must be truly responsible for how one is, mentally speaking – at least in certain respects.
(4) But to be truly responsible for how one is, mentally speaking, in certain respects, one must have brought it about that one is the way one is, mentally speaking, in certain respects. […] One must have consciously and explicitly chosen to be the way one is, mentally speaking, in certain respects, and one must have succeeded in bringing it about that one is that way.
(5) But one cannot reasonably be said to choose, in a conscious, reasoned, fashion, to be the way one is, mentally speaking, in any respect at all, unless one already exists, mentally speaking, already equipped with some principles of choice, P1 – preferences, values, pro-attitudes, ideals – in the light of which one chooses how to be.
(6) But then to be truly responsible, on account of having chosen to be the way one is, mentally speaking, in certain respects, one must be truly responsible for one’s having the principles of choice P1, in the light of which one chose how to be.
(7) But for this to be the case, one would must have chosen P1, in a reasoned, conscious, intentional fashion.
(8) But for this, i.e. (7), to be so one must already have some principles of choice P2, in the light of which one chose P1.
(9) And so on. Here we settle out on a regress that we cannot stop. True self-determination is impossible because it requires the actual completion of an infinite series of choices of principles of choice.
(10) So true moral responsibility is impossible, because it requires true self-determination, as noted in (3). [Which, by (9), we see to be impossible.]305
Is Galen Strawson’s argument successful? It seems that Strawson’s argument certainly
demonstrates something, although perhaps not all that it might be designed to
demonstrate. What Strawson’s argument shows is that, whether or not determinism is
true, the Regression principle can never be satisfied. Thus, if an agent can properly be
305 Galen Strawson, (1994), ibid., pp. 6-7.
308
held responsible only if his actions satisfy the Regression principle, then an agent can
never, after all, be fairly held responsible. Strawson’s argument can be see as appealing
to the Regression principle at stage (3), and again at stages (4), (6), and (8). Indeed, the
idea of responsibility as having a regressive character is what drives this argument.
The reason that one might think that this argument does not achieve as much as Strawson
hopes for it is that there is a deep ambiguity in the term “truly responsible”. If by “truly
responsible”, Strawson means responsible in the sense of blameworthy, as I would
assume he does mean, then the argument is successful. But I would not want to allow that
responsibility-as-assessability is not a “true” sense of responsibility; and, if “truly
responsible” is given that reading, then the argument fails. Claim (3) seems overblown
and unjustifiable with regard to this other sense of responsibility. I shall discuss in more
detail in the next two sections why this should be.
Although I hold that we should endorse Strawson’s argument, at least in the case of
responsibility-as-blameworthiness, it nevertheless has some potential points of weakness,
and would benefit from some modification. Stage (4) is difficult to defend, even if one
endorses the Regression Principle. For even if we think that some agent S, is responsible
(in the sense of blameworthy) for doing action x only if S is also responsible (in the sense
of blameworthy) for the conditions on which S’s doing x depends, one need not think that
the only mechanism for transmitting responsibility from prior conditions to current acts is
the mechanism of “conscious and explicit” choice. We might, for example, hold that S
309
was responsible for his doing x in a case where S was recklessly responsible for the
conditions on which his doing x depends.
Nevertheless, the overall structure of Strawson’s argument is still secure, and its main
point still stands. The infinite regress described at Stage (9) cannot be completed, no
matter how broadly or narrowly we delineate the possible mechanisms for transmitting
responsibility from the antecedent conditions of an agent’s acting in some way to the
agent’s action itself. The essential point stands that self-determination (or, as Strawson
puts it, following Nietzsche, the self being a “causa sui”306) is an impossibility. It is a
simple and important point, and it has been made many times by many philosophers, not
least (as we have already seen) John Locke in the discussion “Of Power” in Book II of An
Essay Concerning Human Understanding.307 But, if we accept the Regression Principle,
then responsibility-as-blameworthiness is in order only with regard to self-determining
agents. And so, as Galen Strawson shows, responsibility-as-blameworthiness is never
justifiably in order.
What is perhaps most interesting about Galen Strawson’s argument is that it shows that,
once we endorse the Regression principle, there are a number of ways of coming to see
that individuals have to be given general absolution from subjection to being held
306 See Galen Strawson, (1994), ibid., p. 15. Strawson quotes Nietzsche’s wonderful remark that “The causa sui is the best self-contradiction that has been conceived so far, it is a sort of rape and perversion of logic. […] the desire to bear the entire and ultimate responsibility for one’s actions oneself, and to absolve God, the world, ancestors, chance, and society involves nothing less than to be precisely this causa sui and, with more than Baron Münchausen’s audacity, to pull oneself into existence by the hair, out of the swamps of nothingness.” (See Friedrich Nietzsche, (1966) [1886], Beyond Good and Evil, trans. Walter Kaufman, (New York: Random House)).
307 John Locke, (1975) [1689], ibid., Book II, Chapter XXI, “Of Power”, §25, p. 247.
310
responsible, in the sense of responsibility-as-blameworthiness. This is because we can
run a reductio argument of the sort here deployed by Strawson in both the “space of
reasons” and in the “space of causes”. The causal reading of Strawson’s argument makes
it not unlike a variety of standard incompatibilist argument. The regression principle tells
us that we can hold individuals responsible only if they are also responsible for the
conditions on which their acting in a particular way depends. But, even if they are
responsible for some set of proximate causal conditions for their action, pursuing causal
chains with sufficient vigour brings us back to a more distant causal point at which such
individuals can have no responsibility (for example because, as in the van Inwagen case,
we have reached a point long before their birth).
But the same kind of regressive procedure also works with regards to the justifications, as
well as the causes, of action. An agent may act in a particular way on the basis of
principles for action that she endorses, but, by the Regression principle, she is fairly held
responsible for so acting only if she is also responsible for the conditions on which her
holding those background principles depends. This, in turn, may depend, in some
proximate way, on a set of more general background normative principles that the agent
endorses; but these principles, in turn, are something for which we can fairly hold the
agent responsible only if she is responsible for the background principles by virtue of
which she has come to hold them. And so a justificatory regress presents itself here,
structurally identical to the causal regress.
311
If it is a “brute fact” that an agent, S, endorses principles x, y, or z, then there is an
unfairness in holding her responsible for acting on the basis of such principles, no
different in type, in a basic sense, to the unfairness that would be involved in holding
someone responsible for the size of their feet. And so we must ask on what basis some
agent comes to endorse particular principles of action. We are thereby led towards an
interminable and unsatisfiable justificatory regress with the same structure, and bringing
us to the same conclusion, as the more familiar causal regress that is more often invoked
in ‘incompatibilist’ thinking.
Our conclusion, therefore, is a very general one. It is that there is no such thing as agency
of a kind that could warrant blame, and hence no agent is ever responsible in the sense of
being blameworthy. Standard incompatibilist arguments can establish this result more
narrowly, for physically instantiated agents, like human beings, whose agency is
embedded in a causal order that is independent of them, and which reaches back before
the time of their first existence. Galen Strawson’s more general argument, though,
generalizes this conclusion from the ‘special case’ of flesh-and-blood (or, at any rate,
spatiotemporal) agents, and shows how we can draw the same conclusions in the entirely
general case, applying to any conceivable agent who acts on the basis of reasons. We
might say that, not only can no embodied agent ever deserve blame, or fairly be held
responsible (in the sense of blameworthy), but neither could the angels.308
308 By “angels” here, I have in mind the idea of a (potentially) autonomous agent that exists (somehow) outside of time and space. In some traditions, angels are no more than emanations or representations of the will of God, in which case one might perhaps be reluctant even to accord them the status of being distinct agents.
312
This conclusion may seem extravagant, perhaps, or simply too radical to accept.
Although some of its strangeness and seeming unacceptability can be drawn away once
we bear in mind that this is a judgement about blameworthiness, as such, and not about
moral evaluation or the possibility of agency. It is not the same as the conclusion that
there are no agents, or that we cannot make moral judgements about the agents that do
exist; rather, it is a conclusion purely and simply about the general and unavoidable
unfairness of blame. But many might still want to resist this conclusion as being simply
too radical. In resisting it, one need only find some way of rejecting the Regression
principle. For we should recall that neither the arguments of van Inwagen nor of Galen
Strawson make an independent case for that principle; rather, both of them, more or less
explicitly, rely on the plausibility of such a principle. And my own positive case for the
Regression principle may seem somewhat thin: it consists in no more than elaborating the
seemingly plausible complaints that we could imagine coming from the lips of a
condemned man like Hellbound Harry.
The dialectical situation is clear enough, though. If we accept the Regression principle,
then blame is always unfair. Some would make the argument that, by modus tollens,
since many cases of blame are manifestly are fair (for example, because they satisfy the
Accuracy and Capacity principles), we therefore have good reason to reject the
Regression principle. (This, to put it only somewhat crudely, is in some ways the central
argument of parts of Jay Wallace’s book, Responsibility and the Moral Sentiments.309)
But such an argument begs the question. On the other hand, arguments in favour of the
309 See, for example, R & MS, Chapter 7, “The Lure of Liberty”, pp. 195-225.
313
Regression principle can seem similarly unsatisfying, for what can they rely upon? The
Regression principle is an expression of some deep seated and recalcitrant normative
intuitions, as expressed by the protestations of our Hellbound Harry. It is hard to find
anything more normatively fundamental that might be offered in support of this principle,
just as it is hard to find any sound basis for undermining the principle.
One thing that can be stressed, though, is that there is nothing outlandish or idiosyncratic
about endorsing the Regression principle, and thereby coming to the view that there is an
inherent unfairness in all judgements of responsibility-as-blameworthiness. On certain
accounts, Kant certainly shared this view, and one may view his strategy of constructing
the idea of noumenal agency as his attempt to find room for the possibility of moral
responsibility, given the apparent impossibility of self-determining agency. If self-
determining agency were a phenomenal possibility, then Kant would have had no need to
restrict ascriptions of full responsibility to noumenal agents. Clearly, Kant thought that
the possibility of self-determining agency had to be established if the possibility of
responsible, blameworthy, agency were to be saved. For Kant also seems to have thought
that ascriptions of responsibility in the absence self-determination were illicit. Kant’s
view, famously, was that standard compatibilism (and, with it, one might think, the
rejection of the Regression principle) was a “wretched subterfuge” and a “petty word-
jugglery”, offering no more than “the freedom of the turnspit”.310 (Indeed, the seriousness
310 On Kant’s views on freedom, responsibility and self-determination, see inter alia Part I, “Concerning the indwelling of the evil principle alongside the good or Of he radical evil in human nature” of the Religion within the Boundaries of Mere Reason (Ak. 6:19 – 6: 53); the “Critical Elucidation of the Analytic of Pure Practical Reason” in The Critique of Practical Reason (Ak. 5: 89 – 5: 106); Section III, “Transition from metaphysics of morals to the critique of pure practical reason” of The Groundwork of the Metaphysics of Morals (Ak. 4: 447 – 4: 463), and the “Third Conflict of the Transcendental Ideas” of “The Antinomy of Pure Reason” in The Critique of Pure Reason (A444/B472-A451/B479).
314
with which Kant took the threat to responsibility from the impossibility of self-
determination is something of which Kant’s latter day followers sometimes lose sight.) I
mean here to do no more than gesture towards Kant’s view, and will not examine it in
detail within the bounds of the present discussion. My wish is only to make it clear that
the acceptance of the Regression principle is a position that has a significant pedigree in
the history of philosophy.
Nevertheless, it would not be plausible to claim that endorsement of the Regression
principle, and the consequent rejection of the fairness of blame and of responsibility-as-
blameworthiness, would not be in some ways a radical step. It would entail a severe
revision of our moral outlook, which commonly takes for granted the fairness of blame,
at least in a prima facie way. In another sense, though, the revision is not so severe for, as
I have been at pains to explain, there is nothing “external” or alien about admitting the
truth of the Regression principle. As Nagel puts it, the collapse of responsibility under
threats like that posed by the Regression principle is not some kind of outside threat, or a
“philosophically imposed demand for external justification” of our ordinary ideas and
practices, rather “it needs nothing more than the ordinary idea of responsibility”.311 Our
conclusion that responsibility-as-blameworthiness is fully and exceptionlessly
undermined by the Regression principle is simply a consequence of carefully working
through an aspect of our common moral thought that has always been there; or, to put
things another way, of germinating a germ of an idea that already exists within the broad
set of our normative commitments.
311 Thomas Nagel, (1986), “Freedom,” in A View from Nowhere, (Oxford: Oxford University Press), p. 125.
315
Nevertheless, there is much in our common moral traditions, and in much (although
certainly not all) religious thought that is inconsistent with accepting the Regression
principle. As Derek Parfit has memorably put it to me, once we reject a belief in the
propriety of desert-entailing blame, “paintings of the Crucifixion wholly fail to achieve
their intended effect … because the Bad Thief steals the picture.”312 And so a certain
reorientation in sensibility is certainly a necessary consequence of following through our
acceptance of the Regression principle, and the rejection of the fairness of blame. But it is
probably a price well worth paying. Indeed, it may be no price at all.
Nevertheless, I remain acutely aware that the Regression principle may still strike many
readers as lacking support. I would not expect all my readers to accept this normative
principle, and we seem to be running short of things that can be said for or against it.
How, then, should we make sense of the dialectical situation at which we have thereby
arrived? Have we reached normative bedrock here? Gideon Rosen supposes that we have.
He offers an incompatibilist response to Wallace’s more comforting Strawsonian
arguments, and ends on this very plausible note:
It is a familiar fact that when we step back from our immersed view of human action and view the morally charged event in all its detail as the necessary upshot of prior causes, there is a powerful inclination to suppose that blame is out of order. My suggestion is that […] there is a residual thought, namely, that it is unfair to blame someone for doing what, through no fault of his own, it was settled in advance he would do. But perhaps you disagree. Perhaps you find your self altogether unmoved by clear-headed reflection on the facts about what has been settled or determined in advance. Perhaps you can look Judas in the eye and hold him responsible for his betrayal, knowing 312 Derek Parfit, private correspondence.
316
full well that while he satisfies Wallace’s conditions, it was settled from the beginning, as a matter of necessity, that he would act badly in the circumstances. Well then, we disagree about what is fair. Could this be bedrock? That would certainly be disappointing. Unfortunately, I see no way to advance the debate in the absence of a more articulate conception of the fundamental idea of moral fairness.313
This gets things largely correct. It is hard to deny the power of the intuition or inclination
that Rosen here identifies, which is itself associated with endorsement of the Regression
principle (although the story told by Rosen, involving the existence of prior causes that
have settled some action in advance is, as we have seen, only a special case of the more
general phenomenon of the impossibility of responsibility for prior conditions for action).
The one point where I would want to part company from Rosen here is in his suggestion
that it is in some way disappointing to reach bedrock. Far from it. Part of the reason for
digging is to discover where it is that the bedrock lies. Moreover, Rosen’s hope in some
future “more articulate conception of the fundamental idea of moral fairness” may well
be misplaced. It is very difficult to imagine how such a conception could serve to
undermine our deep commitment to the idea of fairness embodied in the Regression
principle; on the contrary, the Regression principle has the appearance of bedrock
because that is what it is. It is not plausible to believe that this principle could be
undermined by an elaboration of the fundamental idea of moral fairness, because that
principle is one of the fixed points with which an elaboration of this fundamental idea
would have to contend, and which it would have to accommodate in order to be
successful.
313 Gideon Rosen, (2002), “The Case for Incompatibilism”, Philosophy and Phenomenological Research, LXIV: 3, pp. 699-706, at p. 706.
317
There is an important sense, then, in which the Regression principle does represent
normative bedrock. Little can be said to undermine it, just as little can be done to support
it on the basis of other normative principles. Nevertheless, even without visible means of
external support, I contend that we have overwhelming reason to endorse the Regression
principle; it expresses a deep and important feature of our thinking about agency and
responsibility. Given acceptance of the Regression Principle, we should conclude that it
is never appropriate, because it is never fair, to hold individuals responsible in the sense
of blameworthy.
With this claim in mind, let us recall Schema (N):
(N) S is morally responsible (for action x) if and only if it would be appropriate to
hold S morally responsible for action x.314
If we accept the Regression principle, then, in those cases where we mean blameworthy
when we talk of morally responsible, and where, as discussed above in Section 2, and
following Wallace, appropriateness is understood in terms of fairness, our conclusion
should be that:
(N1) S is morally responsible (i.e. blameworthy) (for action x) if and only if, with
regard to S and x, the Accuracy, Capacity and Regression principles are all satisfied.
As we have seen, the Regression principle is never satisfied. Therefore, by Schema (N),
no agent, S, is ever morally responsible for any action, x.
314 R & MS, p. 91. [My italics]
318
In Section 6 below I will attempt to say a little more in defence of this conclusion, and
offer a methodological argument for why it is inappropriate to reject the Regression
principle. But, I shall now turn to the task of delineating the conditions of fairness for the
other sense of responsibility, that of responsibility-as-assessability.
5. Putting the Hybridity into the Hybrid View: Circumscribed Incompatibilism and the Conditions for Moral Assessment
The previous section was concerned with determining the proper interpretation of
Schema (N) for one of our two central senses of responsibility, that of responsibility-as-
blameworthiness. The function of this section is to consider the best way to fill-out
Schema (N) for the other sense of responsibility which we have delineated, that of
responsibility-as-assessability. We should recall some of the difference between these
two ideas of responsibility. Whereas the idea of responsibility-as-blameworthiness is a
fully (albeit negatively) desert-entailing notion, this is not the case with responsibility-as-
assessability. The conditions under which an agent is responsible, in the sense of morally
assessable, are in a sense far less demanding than the conditions under which that same
agent is properly the object of blame. For an agent to be morally assessable, it need only
be the case that judgements about that agent’s moral character can appropriately be made,
where this is seen as implying nothing about the further question of whether that agent
can, on the basis of that moral assessment, appropriately be blamed, punished or in any
other way treated as responsible in the sense of blameworthy.
319
My answer to the question of when it is appropriate to hold an individual responsible, in
the sense of morally assessable, for some action, is a very familiar one, for it is the
standard sort of answer that has been given by compatibilists throughout the history of
debates on moral responsibility and freedom of the will. More specifically, it is exactly
the sort of answer to the question of the grounds of responsibility that has been delineated
by P.F. Strawson and R. Jay Wallace. An agent can be morally assessed on the basis of
some action, x, when that action really does reflect upon that agent’s character or Quality
of Will (i.e. where the Accuracy criterion or Accuracy principle is met) and where that
agent has the normal set of capacities for autonomous self-direction characteristic of
normal adult human beings (i.e. where the Capacity criterion or Capacity principle is
satisfied). In the case of responsibility-as-assessability, the compatibilist account
developed by Strawson and Wallace hits the mark very accurately. Where Wallace and
Strawson go wrong is simply in thinking that their account applies without difficulty to
the idea of responsibility in general, rather than being restricted to the particular domain
of the narrower idea of responsibility-as-assessability.
This sort of compatibilist account gets things right in the case of moral assessment
because a different standard of fairness is in operation in the case of this sense of
responsibility than was in operation in the case of responsibility-as-blameworthiness.
This is because, when we blame someone, or hold it that they may appropriately be
blamed, we hold that it is appropriate, because it is fair, that they suffer some hard or
unwelcome treatment (insofar as, as discussed in Section 2, blameworthiness should be
seen as a desert-entailing notion). By contrast, mere moral assessment, although it may in
320
a sense be unwelcome when it is negative in character, is not itself a form of hard or
unwelcome treatment, because it is not a form of treatment at all. It is not something that
we do to each other. Rather, it is merely a process of coming to a judgement on the moral
character of others.
In morally assessing another human being, I simply determine what it is that they are
like, as in J. J. C. Smart’s characterization of moral praise and ‘dispraise’ as involving
simply “telling people what people are like”.315 In this regard, it is no different in kind to
assessing your sporting or musical ability. If I judge that you are mean and vindictive, my
judgement is fair just insofar as you really are mean and vindictive, just as my judgement
that you have a tin ear or two left feet is in place just in case that it is accurate. I do not
have to ask anything about the conditions under which you came to be that way, as my
judgement is fair, and therefore appropriate, just in case it is accurate.
Moral assessment, so seen, is indeed simply a form of “grading”. Some may object that
“grading” of this kind is pointless.316 But it is a mistake to think this, as if moral
assessment could be significant or important only if it could be given some ‘external’
purpose. The significance of the moral domain itself explains the significance of our
understanding of the moral properties that are demonstrated by (ourselves and) those
around us. We have, at any rate, a significant interest in understanding the moral
315 See J. J. C. Smart, (1961), “Free Will, Praise and Blame,” Mind, 70, 291-306, reprinted in Watson, ed., (2003), ibid., pp. 58-71.
316 On the alleged pointlessness of ‘mere’ moral “grading” see, for example, the discussion in T. M. Scanlon, ibid., “Blame”.
321
character of those with whom we interact, where that interest has both epistemic and
practical dimensions. But, even if does have a practical dimension that connects it to our
reasons for engaging in certain forms of behaviour (e.g. shunning the dishonest), this is
not to say that the judgement of moral assessment is itself a form of treatment, or that it
involves anything more than a careful judgement as to “what someone is like” in certain
important normative dimensions.
So, the question arises, why should it be fair to hold an individual responsible in the sense
of assessable when it is not fair to hold them responsible in the sense of blameworthy?
The main reason, as has been suggested, is that moral assessment is simply a form of
judgement, guided by the truth at which it aims, rather than something that we do to one
another. Given this, responsibility-as-assessability is not subject to the same norms of
fairness as is responsibility-as-blameworthiness. If we recall Hellbound Harry’s
complaints from the previous section, we can see that they would fall completely flat
when transposed from the case of blameworthiness to the case of moral assessability.
Imagine the stalling way in which Harry’s attempt at complaint would here fail:
Hellbound Harry: I admit that I did behave in a morally bad way, but how can it be
fair to judge that what I did reflects badly on my character? After all, although my action
was under my control, in the sense that I did what I intended to do, the fact that I found
myself in that particular choice-situation when I acted was itself outside of my control.
Also, the fact that, psychologically speaking, I was set up in one way rather than another,
and was bound to react to that choice situation in the way in which I did react, was also
322
something that is beyond my control. This means that it is unfair to judge me, morally
speaking or to subject me to moral assessment. I’m just a product of my environment,
and of a multiplicity of conditions beyond my control, and so it is simply unfair to treat
me as a location at which moral assessment can legitimately be directed.
Judge: Now, now, Harry! You’ve already admitted to me, in our earlier
conversation, that what you did was, indeed, morally faulty, and you don’t deny that you
have the general capacities of reflective self-control that are enjoyed by normal adult
human agents. That is, you don’t deny that the negative assessment of your moral
character satisfies both the Accuracy Principle and the Capacity Principle. So, here you
are asking me to admit that my moral judgement of you needs also to satisfy the
Regression principle – that you are not really bad or, indeed, a candidate for moral
assessment, unless you are responsible for the full set of conditions that have formed your
character to be as it is. But, whatever your case with regard to the fairness or
appropriateness of blame, Harry, you have no case here, when we are speaking of
straightforward moral assessment.
I admit that your character itself results from conditions extending
backwards in time, and far beyond your control. But that is just how you are, Harry, as a
spatiotemporal being, like any other, existing in a universe that has a history. You still are
what you are, your character is still your character, and your morally bad action is still
your morally bad action, at least in those cases (i.e. where the Capacity and Quality of
Will criteria are satisfied) where that action connects in the right way with your
323
intentions, desires and plans. My moral assessment of you is, I admit, the moral
assessment of a contingent human individual, who might not have existed, and who
might have had different properties if history had been different. But this is just to say
that you, Harry, are a contingent being, and might have been otherwise than you are. But,
as things have turned out, Harry, you are as you are, and the moral properties that I
ascribe to your character when I make an accurate moral judgement simply are those
properties which your character happens to display.
I leave open the question of whether it would be fair to blame you,
punish you, shun you, or damn you, on the basis of your having these moral properties
which, as you stress and as I admit, you have in a sense only contingently. But this is
where the difficult normative issue concerning the fairness of holding you responsible
really resides: it resides in the question of how it would be fair to treat you, given certain
truths about your moral character. But for that question to even get off the ground, Harry,
we have first to admit that your moral character, as evidenced by your actions, is vicious
in certain significant normative respects. It is a pity that you are like this, and perhaps, as
you have argued, nothing follows from this about how it would be fair or unfair for me to
treat you. But I do you no disservice, and do not treat you unfairly, when I tell the truth
about what you are like. And, in holding you responsible in the sense of responsibility-
as-assessability, all I am doing is forming a judgement of what you are like, on the basis
of how you act.
324
My obligation of fairness to you, in forming moral judgements
about your character, has just two simple elements. These elements are, on the one hand,
a restriction of accuracy, and, on the other hand, a restriction that shows me that you are
properly to be morally assessed only the basis of your autonomous action. This second
restriction applies because autonomous action forges links between your character and
your actions. Hence I do not hold you responsible on the basis of non-autonomous
inclinations or spasms, which may tell me nothing of what you are really like.
In the first of our dialogues between Harry and his Judge, Harry surely had the better of
things. In this second meeting, we must by contrast admit that the judge gets things right.
Harry’s protestations, the first time out, seemed to get at an important, if difficult to
delineate, normative principle of fairness. How could we blame him for his actions when
his actions were dependent on conditions outside his control? One must admit at least a
certain prima facie plausibility to the sense that we would be doing something unfair to
Harry in this case, even if one is not, in the end, convinced by my earlier defence of the
Regression principle. But, on this second occasion, Harry’s complaints seem to have lost
the foothold that was provided to them by the special problem of justifying a desert-
entailing attitude like blame. We can justify mere judgements of how Harry happens to
be, when they are considered on their own terms, and not bundled in with any
assumptions about what follows in terms of Harry’s blameworthiness, simply when those
judgements are accurate, and describe things as they really are. To ask for any more, as
Harry does here, seems extravagant and undermotivated.
325
The failure of Harry’s protestations in this example therefore serves as a demonstration of
how and why standard incompatibilism (or thoroughgoing scepticism about
responsibility) goes too far. Such views develop a line of thought that makes good sense
with respect to responsibility-as-blameworthiness, and then illicitly reapply it in a
context, regarding the norms governing responsibility-as-assessibility, where it is
inappropriate. Incompatibilists and responsibility-sceptics are not, of course, confused
from the outset, because the Regression principle does, indeed, govern the
appropriateness of certain kinds of ascription of responsibility. But should nevertheless
conclude that their error is simply to go too far in applying the Regression principle
beyond its proper scope of application.
Therefore, there is no need to invoke an additional normative principle such as the
Regression principle when we attempt to find the conditions for the fairness of moral
assessment. The twin conditions generated by having to satisfy the Accuracy and
Capacity criteria give us all that we need. Indeed, if there is any challenge to this view of
when moral assessment is appropriate then it comes, so to speak, from the opposite
direction. The worry is not so much that we would need to satisfy the Regression
principle, but that we might not even have to satisfy the Accuracy and Capacity
principles in order for it to be appropriate to hold some agent responsible (in the sense of
assessable) for some action. In order to see where this threat might come from, consider
these remarks of Nagel’s:
A person may be greedy, envious, cowardly, cold, ungenerous, unkind, vain, or conceited, but behave perfectly by a monumental act of will. To possess these vices is to be unable to help having certain feelings under certain circumstances, and to have strong
326
spontaneous impulses to act badly. Even if one controls the impulses, one still has the vice. An envious person hates the greater success of others. He can be morally condemned as envious even if he congratulates them cordially and does nothing to denigrate or spoil their success. Conceit, likewise, need not be displayed. It is fully present in someone who cannot help dwelling with secret satisfaction on the superiority of his own achievements, talents, beauty, intelligence or virtue. […] Yet people are morally condemned for such qualities, and esteemed for others equally beyond control of the will: they are assessed for what they are like.317
Nagel here points to the fact that we can sometimes seemingly assess others morally on
the basis of facts about them other than their Quality of Will. I may judge you badly for
your conceitedness, even though you don’t want to be conceited (i.e. you do not have a
conceited ‘Quality of Will’) and even if you do not act in a conceited way. An individual
like this is, of course, less than fully autonomous, in that there is a lack of coherence
among his desires, character traits and inclinations. But such disconnections, and internal
tensions within people’s psychological states are, of course, the rule rather than the
exception. Schema (N), we recall, talks about the appropriateness of judgements of
responsibility with regard to an agent, S’s, performance of an action, x. But Nagel’s
example reminds us that not all assessments of particular agents are assessments of their
qualities in having performed any particular action, x or y. Sometimes we assess agents
directly, in virtue of certain traits of character, independently of their actions.
Nagel’s example suggests others in which we have a departure from the ‘standard case’,
where we assess a competent agent, with the capacity for reflective self-control, in light
of his autonomous performance of some action, x. For example, there is the complication
of cases where we have severely non-autonomous agents (recall Arbitrary Alex from Part
317 Nagel, “Moral Luck”, in his (1979), Mortal Questions, ibid., pp. 32-3.
327
One) whose actions do not reflect any settled ‘Quality of Will’ and who lack the capacity
for reflective self-control. If such an agent performs a terrible and vicious crime, it is very
tempting to say that we are entitled to assess his actions in performing that crime, by use
of moral criteria. Thus, in this case, we appear to have moral assessment without the
satisfaction of either the Accuracy principle (for the agent’s behaviour does not reflect a
settled Quality of Will but was, let us suppose, carried out on a transient whim) or of the
Capacity principle (for, again, ex hypothesi, the agent lacks the requisite capacities of
reflective self-control).
These sorts of cases point to the vast complexity of the territory.318 Schema (N) is
addressed to standards of responsibility for the assessment of agents in light of their
performance of particular actions, and our investigation thus far has been conducted in
these terms. Yet, we should not lose sight of the fact that not all moral assessment
(broadly speaking) is assessment of an agent in light of some particular action or set of
actions. Some assessment, like that described above by Nagel, is of the agent’s character
itself, independent of the agent’s actions. Similarly, some moral assessment is of an
agent’s performing some particular action (such as Arbitrary Alex’s crime committed on
a whim), even when we accept that the action in question is not linked in any robust way
to stable features of that agent’s character.
318 For the purposes of the present discussion, we will not spend too much time exploring the fascinating question of when character, as such, can be a proper object of moral assessment. For an illuminating discussion that does delve deeper into these issues, see Angela Smith, (2005), “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics 115:2, pp. 236-271.
328
So, should we conclude that holding an agent responsible, in the sense of treating them as
being morally assessable, does not require the satisfaction of the Accuracy and Capacity
criteria, just as it also does not require the satisfaction of the Regression criterion? The
best answer to this question has to be a careful and nuanced one. We should admit, as
Nagel’s example and the example of Arbitrary Alex show, that it is not the case that
absolutely all forms of moral assessment require the satisfaction of the Accuracy and
Capacity criteria.
We could take a hard line here, and say that genuine moral assessment is always of
persons, in light of some action (or set of actions), and that we have a person only in
cases where the Capacity criterion is satisfied, and a genuine action (as opposed to a sub-
agential instance of behaviour) only when there is the right relationship between the
agent’s Quality of Will and what it is that he actually does. This would involve biting two
kinds of bullet: the first is accepting the claim that Alex is not, in fact, a person after all;
the second is denying Nagel’s claim that moral assessment can ever properly be directed
at unexpressed character traits, reserving the domain of the moral to cases of action for a
reason. But taking this hard line may be rather excessive, and one would end up holding
fast to some very counter-intuitive claims in order to preserve the significance and special
status of the Accuracy and Capacity principles.
Nevertheless, even having made this sort of concession to the existence of other species
of moral assessment, we can still keep an important place for the Accuracy and Capacity
criteria, especially as regards providing a full interpretation of Schema (N) for the idea of
329
responsibility-as-assessability. If we set aside the acknowledged complexities of the
assessment of attitudes and character, especially where those attitudes or character traits
are unexpressed, we nevertheless still have to answer the question of when some agent, S,
can be morally assessed for performing some action, x. If the question in view is how or
whether S can himself be assessed in light of his doing x, then the right kind of
relationship between S and x has to obtain – i.e. x has to reflect S’s Quality of Will, and
the Accuracy criterion has to be satisfied. Otherwise, the performance of x itself tells us
nothing much about S himself. This is how the question of how we would morally assess
S for doing x differs from the question of how we would morally assess S’s doing x. For
example, we could say that Alex’s cruel jibe at Rachel was a bad thing to do, even if we
allowed that Alex is not himself morally assessable for doing this, given that his action
did not reflect any stable features of his character. Thus, in terms of the question that we
have here set ourselves – that is, the interpretation of Schema (N) for the case of moral
assessment – the Accuracy principle retains its significance.
Following this line of thought further, we come also to an explanation of the special
significance of the Capacity principle for moral assessment. For, even if we can, without
conceptual confusion, go in for the moral assessment of impulsive or thoughtless actions,
we are especially interested in the moral assessment of autonomous actions. This is part
of the moral significance of autonomy. We are not only especially interested, morally
speaking, in actions that reflect agents’ quality of will, but within the set of such actions,
we are especially interested in those which reflect the deep and abiding features of an
agent’s attitudes and character.
330
In Part One, I argued that there is no simple or straightforward account that can capture
all that we are concerned with when we address ourselves to the idea of autonomy, and
that we should instead prefer a pluralist account of the various dimensions of
autonomy.319 Among the elements of autonomy, so understood, will be normative
competence, diachronic stability, coherence and sophistication in an agent’s attitudes and
desires. Accordingly, on this view, autonomy is not something that one simply has or
lacks, but it is something of which one has more or less, according to a complex set of
measures, over a number of separate dimensions. My suggestion is that, the more
autonomous an agent’s actions are, the more significant the moral assessment of those
actions will be, and the more those actions will tell us about what that agent is really like.
This is the somewhat more complex truth to which the Capacity criterion provides a
rough approximation. For, unless an agent has the capacity for reflective self-control,
which is itself one of the essential elements of agential autonomy, assessment of that
agent’s actions will tell us ‘shallow’ rather than ‘deep’ facts about what that agent is
really like, morally speaking. Moral assessment of non-autonomous agents runs the risk
of a certain kind of unpleasant judgementalism – of applying strict moral standards of
appraisal without taking note of the various ways in which someone’s agency may be
restricted or encumbered. But, in the case where our assessment is of the actions of a
robustly autonomous agent, this danger is avoided. The proper and appropriate target of
strict moral assessment is the agent whose actions display his quality of will (thereby
319 See Part One, Section 8, above.
331
satisfying the Accuracy principle) and whose quality of will is itself born from reflection,
deliberation and the application of the full set of capacities associated with autonomy
(thereby satisfying the Capacity principle). Thus, we should conclude that the Capacity
criterion does, indeed, retain its significance, albeit that, if we wish to be more precise,
we should understand it as being elliptical for a more fully elaborated set of the capacities
associated with autonomous agency.
I am aware that the foregoing discussion has skirted some issues of great complexity and
considerable delicacy, and much more can be said: on the questions of the moral
assessment of attitudes and character; on the relationship between autonomy and moral
significance and between autonomy and personhood; and, indeed, about the distinction
between what is active and what is passive in our mental lives. Nevertheless, the two
essential points of conclusion for this section are now clear. The first is that, even if some
forms of moral assessment are in some ways appropriate in cases where the Accuracy and
Capacity principles remain unsatisfied, those two principles still retain their special
significance with regard to the assessment of agents for their actions, such that
assessment of that kind is most fully appropriate in cases where those two principles are
fully satisfied. Secondly, and this is the most important point for our present discussion,
the appropriateness of responsibility-as-assessability does not depend on satisfying the
Regression principle (as it does so depend in the case of responsibility-as-
blameworthiness).
332
Given this, and bracketing for present purposes our caveats about the complexities
underlying the Capacity principle, we can provide a full interpretation of the content of
Schema (N) in the case of moral assessment. For this ‘thin’ version of non-desert-
entailing responsibility, the schema reads as follows:
(N2) S is morally responsible (i.e. (fully) morally assessable) (for action x) if and only
if, with regard to S and x, the Accuracy and Capacity principles are both satisfied.
Here, then, we now have a full elaboration of the Hybrid View. The Hybrid View is the
view that combines (N1), which is an incompatibilist or sceptical view about
blameworthiness, with (N2), a compatibilist view about this thinner notion of moral
responsibility as assessability. If the Hybrid View is correct, then traditional views about
the conditions for moral responsibility go wrong in going too far, each in their own way.
Nevertheless, both kinds of traditional views contain an important kernel of truth, and by
splitting apart our examination of the two separate senses of responsibility, we have been
able to see how these two kernels of truth can peacefully co-exist alongside one another,
avoiding the apparent conflict to which such differing views have traditionally been
subject. In Strawson’s words, we have “a formal withdrawal on one side in exchange for
a substantial concession on the other,”320 and have thereby come as near as can be
imagined to a genuine reconciliation of compatibilist and incompatibilist views about
moral responsibility.
320 F & R, pp. 72-3.
333
6. The ‘Meta Argument’ for the Hybrid View: On the Enduringness and Recalcitrance of Disagreements about Responsibility
One of the frustrating features of enduring philosophical debates is that views become
very entrenched, and consequently very difficult to shift. Once one has been convinced
of, for example, Humean compatibilism, it is easy to harden one’s heart to the appeal of
incompatibilism, or the plausibility of the normative principle of fairness expressed by
the Regression principle. Consequently, it is very difficult to convince people of a new or
different view about moral responsibility, and I labour under no misapprehension that the
considerations presented in Section 4 would undermine the conviction of a convinced
compatibilist, any more than the discussion of Section 5 would change the mind of an
entrenched incompatibilist. The compatibilist will remain convinced that the Regression
principle is too extravagant and outlandish to be worth considering, while the
incompatibilist will see my circumscription of the area of application for the Regression
principle to the case of blameworthiness as (something like) a piece of intellectual
cowardice, insisting that, if it applies, then it applies to the case of responsibility-as-
assessability as well.
These are deep disagreements, without much in the way of shared background to which
one can appeal. Indeed, normative bedrock may already have been reached. Nevertheless,
I think that there are arguments available that point towards acceptance of the Hybrid
View, or of something like it. The line of argument available points to the very
enduringness and recalcitrance of debates over moral responsibility, and tries to turn this
feature of the debate to its advantage. Let us call this the ‘Meta’ Argument for the Hybrid
view. It is not a deductive argument; but, even so, it provides a very useful way to think
334
about what might be going on in debates over moral responsibility, and how we might
hope to come to a plausible conclusion about the conditions for moral responsibility.
The Meta argument runs like this:
(1) If standard compatibilism is correct, incompatibilists make a gross normative
error in according any significance to the Regression principle.
(2) If standard incompatibilism is correct, then compatibilists are guilty of a kind of
normative tone-deafness in failing to see the significance of the considerations of
fairness that generate the Regression principle.
(3) Hence, by (1) and (2), if either standard view is correct, a large number of
disputants in these debates have been making fundamental errors, the absurdities
of which are patently obvious to proponents of the opposite view.
(4) If either of the standard views were correct, then we would reasonably expect
competent reasoners, thinking and debating about these issues over a very long
time, to see the absurdity of the errors involved in the incorrect view. We would
therefore have expected to see much more in the way of convergence on versions
of the most plausible view, and a wholesale rejection of views that involve this
kind of gross error.
(5) But the debate about responsibility is utterly interminable and recalcitrant.
(6) So, by (4) and (5), perhaps neither standard view is correct.
(7) If neither standard view is correct, perhaps the correct view is one that draws
elements from both the main types of views on this issue.
335
(8) If such a view could explain why the standard views get at part of the truth, and
why both kind of views involve rather subtle rather than gross errors, then we
would have a view that would be much more likely to be correct, given the
enduringness of the debate, and the ability of both compatibilist and
incompatibilist views to go on attracting adherents.
What is distinctive about the Hybrid view is that it offers an explanation of why standard
compatibilist and incompatibilist views have each seemed to be convincing: both sorts of
views get at part of the truth about how we should understand the conditions under which
judgements of responsibility are fair, and therefore appropriate. Thus, there are the
resources contained within the Hybrid view for explaining why debates about the proper
conditions for ascriptions of responsibility should be so difficult to settle, and so
enduring. The debate has been so enduring simply because both kinds of standards views
have made the same small mistake – they have failed to see that their accounts apply only
to one particular sense of ‘responsibility’, rather than giving the correct conditions for
ascriptions of some wholly general idea of responsibility.
From the standpoint of the Hybrid view, the error that has mired the debate about
responsibility thus far is a conceptual error, born of a lack of care in keeping apart
different notions of responsibility. This is not the same sort of gross error of completely
failing to see to what sort of norms the application of ascriptions of responsibility should
be responsive. Those who simply defend or reject out of hand the Regression principle
may seem to talk past each other, or not to share a common conceptual terrain with one
336
another. The resources of the Hybrid view can help to show how their disagreement is, in
fact, not so profound, but is simply born out of a failure to delineate carefully enough the
conception of responsibility that is under investigation. In short, the Hybrid view is able
to offer a solution to the problem of responsibility that preserves the honour of both
parties in the traditional dispute, showing how neither party had a wholly faulty grasp of
the normative and conceptual terrain, but how instead both had got at something true and
important. Thus, one might suggest that, in order for a more standard view to be as
convincing as the Hybrid view, it would need to be able to offer as good an explanation
of why the debate had gone on for so long, and with such little in the way of palpable
progress. More traditional views are unlikely to be able to do so.
I do not for a moment take this ‘Meta’ argument to be decisive. But, nevertheless, it
provides a strong additional reason for why we should take the Hybrid view seriously.
We are unlikely to be able to attain more than this in defending the Hybrid View. After
all, the Hybrid view is a position regarding a very deep and difficult philosophical
problem, within a debate characterized by deep disagreements about the conditions for
responsibility and, indeed, about the structure of the very notion of responsibility itself. In
a philosophical debate of this kind, we are simply not dealing with the sort of issue that
can be settled beyond doubt by means of a formal argument. Galen Strawson makes the
point beautifully about how, methodologically speaking, philosophy can make progress.
His remarks are especially apposite when we attempt to get to grips with the really
difficult questions, and with those questions where there is deep disagreement and a long
history of dispute:
337
It’s often said that argument is the heart of philosophy, and especially of analytic philosophy, but I’m sure that’s not true, if argument is thought of as primarily a matter of formally arrayed premises and conclusions. Argument in this sense is the handmaiden of philosophy, an underlabourer (the head underlabourer), to be summoned as necessary. All arguments have premisses, after all, and not all premisses can be argued for on pain of never getting started. The fundamental philosophical activity, I think, is a kind of open, investigative dwelling on ideas. It may well make use of formal argument, but it need not, and it is at its hear an essentially looser matter of redescribing things, putting them in other ways, spreading them out descriptively, telling stories that articulate and animate them. These are the instruments and the experiments of philosophy. It is, as a science, a suasive art, a mixture of plain speaking and the “arduous invention which is the very eye of research” (George Eliot, Middlemarch, ch. 16).321
This gets things exactly right. One cannot hope to establish a view like the Hybrid View
by wholly formal methods, and one cannot expect to arrive at a knockdown argument in
its favour. But one can tell a story, or stories, that makes it more and more plausible. As
we have seen, one can make the ‘first-order’ case for the component elements of the
Hybrid View (as has been done above in sections 4 and 5), and one can then also make
the ‘second-order’ (or ‘meta’) case for why the view so constructed has structural
advantages unattainable by its opponents (as has been the job of this section).
The combination of these two accounts, at different levels, may be enough to perform the
task of persuasion. If not, then it is difficult to see what more could be done. My
contention, and hope, is that an “open and investigative dwelling” on the ideas
marshalled within the Hybrid view will lead one to acceptance of the truth of that view.
321 Galen Strawson, “Introduction”, in his (2008), Real Materialism and other essays, (Oxford: Oxford University Press), p. 3.
338
7. The Place of Optimism
There remain a small number of loose threads that need to be gathered up. The most
important of these contains the place of what P. F. Strawson called ‘optimism’: that is,
the view that the practice of holding people responsible can be justified by virtue of its
effects, rather than by means of its independent fairness or appropriateness. If we accept
(N1) and (N2), and thereby hold that the Hybrid View gets to the truth about the
conditions for ascriptions of responsibility, then what should one say about these kinds of
instrumentalist justifications of the practice of holding people responsible for their
actions?
Wallace, whose normative account of responsibility we have followed, despite diverting
from his particular conclusions, has this to say:
Of course, it would be possible to bring other practical norms to bear in assessing the stance of holding people responsible. Such norms might even support a verdict at odds with the conclusion that is grounded in the moral considerations of fairness. For instance, it might turn out that there are good strategic reasons to hold responsible a class of people whom it would clearly be unfair to treat in that way – perhaps such treatment would materially advance our interests in security, or happiness, or the eventual attainment of a more egalitarian social world. About such a case, we would surely not say that it is indeterminate whether the people in question are responsible agents, but that there are strategic reasons for treating as responsible a class of people who are not responsible in fact. This shows that moral norms of fairness have a privileged position in determining what it is to be a responsible agent: they set the standards of appropriateness in terms of which schema (N) is to be interpreted.322
This analysis of “optimist” or “instrumentalist” accounts of responsibility hits the nail on
the head. The Hybrid View, with its specification of interpretations (N1) and (N2) of
Schema (N), is the last world on when it is fair to hold agents responsible, in each of our
322 R & MS, p. 94.
339
two distinct senses. It thereby tells us when, in each regard, agents really are responsible.
It is a consequence of that account that agents are never blameworthy, and there is never
a straightforward justification, relying simply on the facts about an agent’s actions, for
any of the desert-entailing attitudes (or actions) that we might deploy. According to (N1),
given the impossibility of satisfying the Regression principle, it is never appropriate,
because it is never fair, to blame, resent or punish.
But this does not mean that, in coming to all things considered judgements about what we
– individually or collectively – have reason to do, we must therefore prejudge that we
never have reason to display any desert-entailing reactive attitude or to take part in any of
the blaming behaviours of reproach, scorn, or punishment. Sometimes we may have very
good reason to act as if we did believe that some group or individual were responsible (in
the sense of blameworthy) for what they have done, even though we do not believe that
they are responsible, and even though we believe that we thereby treat them in one way
unfairly.
Accepting the Hybrid View does not mean that we have to empty all the prisons, as it
were, even though there is no doubt that many people who are in prison should not be
there, and, indeed, even though there is no doubt that many of our prisons should not
exist. Acceptance of the Hybrid View does not mean that all instances of punishment are
wrong, all-things-considered, but it does mean that there is a real and ineradicable
unfairness involved in every instance of our punishing, blaming or resenting. There is no
paradox in holding the view that, on the one hand, no one is really responsible, in the
340
sense of blameworthy, but that, on the other hand, and acknowledging the unavoidable
unfairness involved, we often have reason to impose forms of hard and unwelcome
treatment on those who have acted badly.
None of this should be surprising. We have a plethora of reasons for action that run
beyond our pro tanto obligation to respect the norms of fairness that govern ascriptions of
responsibility, and determine (one aspect of) the fairness of blaming behaviour, and of
the imposition of hard and unwelcome treatment on others. Often, these other reasons
win out, and we thereby have decisive overall reason to act in one way unfairly, and to
treat ‘as responsible’ those whom we know to be free of responsibility.
All of these claims can be accommodated without difficulty by the Hybrid view, which
accepts Wallace’s Schema (N) as providing a general account of what it is to hold some
agent responsible for some action. If we accept Schema (N), then we understand the
claim that some agent, S, is responsible for some action x as being elliptical for the claim
that it is fair to hold that agent responsible for action x. This opens up a difference from
any view of responsibility that holds that some agent, S, is responsible for some action x
when and only when we have final, all-things-considered reasons to impose liability for
costs, or unwelcome treatment, on S in virtue of his having done x. On a view of this
latter kind, we simply cannot say that, although S is not responsible for x, we
nevertheless have reason to treat him as if he is, for, on such views, if we have decisive
reasons to treat x as if he were responsible, then we simply have to say that this is what it
is for x to be responsible. By contrast, the Hybrid view, or any view that accepts Schema
341
(N), is able to make sense of the possibility of cases in which it might be justifiable, all-
things-considered, to impose some cost, attitude, treatment or punishment on S in virtue
of his having done x, while nevertheless accepting that, strictly speaking, S is not
blameworthy for having done x.
The Hybrid view can find conceptual space for (what we might call) justifiable but unfair
blaming behaviour. By contract, accounts that treat responsibility as an ‘all-things-
considered’ idea lose sight of the real moral loss that can be involved in some cases,
where we may have good reason to impose liability or costs on some individual in light
of some action of his, even though there is a deep and unavoidable unfairness in our
doing so. The Hybrid view has the virtue of not sweeping this potential difficulty under
the carpet. But it does insist that, whenever we do treat people as blameworthy or in some
other way liable for the unwelcome costs of their actions, a small residue of unfairness
remains. The truth in “optimism” is that, rather than being a mere conceptual possibility,
there will often be real cases in which unfair blaming behaviour will indeed be
justifiable, when the full range of relevant reasons are taken into consideration.
8. On the Apparent Insolubility of the Problem of Responsibility
Thomas Nagel has claimed that “the problem of responsibility may well be insoluble”.323
In a sense, Nagel is right. There are deep and competing intuitions on the subject, such
that no single interpretation of Schema (N) could ever be satisfactory. No single account
323 Thomas Nagel, (1986), The View from Nowhere, (Oxford: Oxford University Press), Chapter 7 on ‘Freedom’, p. 120
342
of the conditions for responsibility could ever provide a convincing solution. The only
hope for a plausible solution is to reframe the question, and splitting responsibility into its
two component senses is one good way of doing this.324
The problem of responsibility is insoluble, but that is because there is no such thing as the
problem of responsibility. Instead, as we have seen, there are (at least) two such
problems, and these, we can at least hope, may be soluble. I have, at any rate, suggested
one way of solving them.
324 This general strategy has been followed by a number of others, notably by T. M. Scanlon in his (1998), What We Owe to Each Other, (Cambridge, Mass.: Harvard University Press), Chapter 6 on ‘Responsibility’ (although, as we have seen, Scanlon’s way of distinguishing between two senses of responsibility is importantly different to the distinction between responsibility-as-assessability and responsibility-as-blameworthiness explored above).
Bibliography
Arthur W. H. Adkins, (1960), Merit and Responsibility: A Study in Greek Values, (Oxford: Oxford University Press).
Rogers Albritton, (1985), “Freedom of the Will and Freedom of Action,” The Proceedings and Addresses of the American Philosophical Association, 59: 2, 239-51; reprinted in Watson, ed., (2003), pp. 408-23.
Henry E. Allison, (1990), Kant’s Theory of Freedom, (Cambridge: Cambridge University Press).
G. E. M. Anscombe, (2000) [1957], Intention, (Cambridge, Mass.: Harvard University Press).
Aristotle, Nicomachean Ethics, (2000), ed. and trans. Terence Irwin, (Indianapolis: Hackett).
Richard J. Arneson, (2003), “The Smart Theory of Moral Responsibility and Desert,” in Serena Olsaretti, ed., (2003), pp. 233-58.
St. Benedict, (1986) [6th century], The Rule of St Benedict, (Collegeville, MN: Liturgical Press).
Sarah Buss and Lee Overton, eds., (2002), Contours of Agency: Essays on Themes from Harry Frankfurt, (Cambridge, MA: MIT Press).
Steven Collins, (1982), Selfless Persons: Image and Thought in Theravāda Buddhism, (Cambridge: Cambridge University Press).
Mihályi Csíkszentmihályi, (1990), Flow: The Psychology of Optimal Experience, (New York: Harper and Row)
__________, (2000), Beyond Boredom and Anxiety: Experiencing Flow in Work and Play, (San Francisco: Jossey-Bass).
Dalai Lama of Tibet, (1998), Freedom in Exile, new edition, (New York: Abacus).
Stephen Darwall, ed., (1995), Equal Freedom: Selected Tanner Lectures on Human Value (Ann Arbor, Michigan: University of Michigan Press).
René Descartes, (1989) [1649], The Passions of the Soul, ed. Stephen Voss, (Indianapolis: Hackett).
J. M. Fischer, (1995), The Metaphysics of Responsibility: An Essay on Control, (Oxford: Blackwell).
343
344
__________, (2007), My Way: Essays on Moral Responsibility, (Oxford: Oxford University Press).
J. M. Fischer and Mark Ravizza, eds., (1993), Perspectives on Moral Responsibility, (Ithaca, NY: Cornell University Press).
__________, (1998), Responsibility and Control: A Theory of Moral Responsibility, (Cambridge: Cambridge University Press).
Harry G. Frankfurt, (1969), “Alternate Possibilities and Moral Responsibility,” Journal of Philosophy, LXVI, no. 23, 829-39; reprinted in Frankfurt (1988), pp 1-10.
__________, (1971), “Freedom of the Will and the Concept of a Person,” Journal of Philosophy, LXVIII, no. 1, 5-20; reprinted in Frankfurt (1988), pp 11-25. [Referred to in the text as “FW&CP”; page references are to Frankfurt (1988).]
__________, (1984), “Necessity and Desire,” Philosophy and Phenomenological Research, XLV, I, pp. 1-14; reprinted in Frankfurt (1988), pp. 104-16.
__________, (1987), “Identification and Wholeheartedness,” in Ferdinand David Schoeman, ed., (1987); reprinted in Frankfurt (1988), pp. 159-76.
__________, (1988), The Importance of What We Care About, (Cambridge: Cambridge University Press).
__________, (1994), “Autonomy, Necessity and Love,” in H. F. Fulda and R.-P. Horstmann, eds., (1994); reprinted in Frankfurt (1999), pp. 129-41.
__________, (1999), Necessity, Volition and Love, (Cambridge: Cambridge University Press).
__________, (2002), “Reply to T. M. Scanlon,” in Buss and Overton, eds., (2002), pp. 184-8.
__________, (2002), “Reply to Richard Moran,” in Buss and Overton, eds., (2002), pp. 218-25.
__________, (2004), The Reasons of Love, (Princeton, NJ: Princeton University Press).
H. F. Fulda and R.-P. Horstmann, eds., (1994), Vernunftbegriffe in der Moderne: Stuttgarter Hegel-Kongress 1993, (Stuttgart: Klett-Cotta).
Graham Greene, (1974), A Sort of Life, (London: Penguin).
__________, (2003) [1940], The Power and the Glory, (London: Penguin).
Lewis Edwin Hahn, ed., (1998). The Philosophy of P. F. Strawson, (Peru, Illinois: Open Court Publishing).
345
James Harris, (2005), Of Liberty and Necessity: The Free Will Debate in Eighteenth Century British Philosophy, (Oxford: Oxford University Press).
William Vernon Harris (2002), Restraining Rage: The Ideology of Anger Control in Classical Antiquity, (Cambridge: Mass.: Harvard University Press).
H. L. A. Hart, (1968), Punishment and Responsibility: Essays in the Philosophy of Law, (Oxford: Clarendon Press).
Thomas Hobbes, (1996) [1651], Leviathan, ed. Richard Tuck, (Cambridge: Cambridge University Press).
David Hume, (2000) [1739-40], A Treatise of Human Nature, ed. David Fate Norton and Mary J. Norton, (Oxford: Oxford University Press).
St. Ignatius of Loyola, (1996) [1524], “Spiritual Exercises,” in his Personal Writings, (London: Penguin).
Immanuel Kant, (1996) [1788], The Critique of Practical Reason, reprinted in Mary Gregor, ed., (1996), Immanuel Kant: Practical Philosophy, (Cambridge: Cambridge University Press).
__________, (1998) [1785], The Groundwork of the Metaphysics of Morals, ed. Mary Gregor, (Cambridge: Cambridge University Press).
__________, (1998) [1793], Religion within the Boundaries of Mere Reason, ed. and trans. Allen Wood and George di Giovanni, (Cambridge: Cambridge University Press).
__________, (1999) [1781/1787], The Critique of Pure Reason (1999), ed. and trans. Paul Guyer and Allen Wood, (Cambridge: Cambridge University Press).
Christine M. Korsgaard, (1996), Creating the Kingdom of Ends, (Cambridge: Cambridge University Press).
John Locke, (1975) [1689], ed. Peter H. Nidditch, An Essay Concerning Human Understanding, (Oxford: Oxford University Press).
Sterling M. McMurrin, ed. (1988), The Tanner Lectures on Human Values, Volume 8, (Salt Lake City, Utah: University of Utah Press).
Richard Moran, (2002), “Frankfurt on Identification: Ambiguities of Activity in Mental Life,” in Buss and Overton, eds., (2002), pp. 189-217.
Michael A. Mullett, (2004), Martin Luther, (New York: Routledge).
Thomas Nagel, (1970), The Possibility of Altruism, (Princeton, NJ: Princeton University Press).
346
__________, (1976), “Moral Luck,” Proceedings of the Aristotelian Society, supplementary volume L; reprinted in Nagel, (1979), pp. 24-38.
__________, (1979), Mortal Questions, (Cambridge: Cambridge University Press).
__________, (1986), The View from Nowhere, (Oxford: OUP).
Friedrich Nietzsche, (1966)[1886], Beyond Good and Evil, trans. Walter Kaufman, (New York: Random House).
__________, (2006) [1887], On the Genealogy of Morality, 2nd edition, ed. Keith Ansell Pearson, trans. Carol Diethe, (Cambridge: Cambridge University Press).
Serena Olsaretti, ed., (2003), Desert and Justice, (Oxford: Clarendon Press).
George Orwell, (1949), Nineteen Eighty-Four, (London: Secker and Warburg).
Brian O’Shaughnessy, (1980), The Will: A Dual Aspect Theory, (Cambridge: Cambridge University Press).
Derek Parfit, (1984), Reasons and Persons, (Oxford: Oxford University Press).
Blaise Pascal, (1995) [1670], Pensées, (London: Penguin).
Rajendra Prasad, (1995), “Reactive Attitudes, Rationality and Determinism,” in Sen and Verma, eds., (1995), pp. 346-76.
John Rawls, (1971), A Theory of Justice, (Cambridge, Mass.: Harvard University Press).
Gideon Rosen, (2002), “The Case for Incompatibilism,” Philosophy and Phenomenological Research, LXIV. 3, 699-706.
__________, (2003), “Culpability and Moral Ignorance,” Proceedings of the Aristotelian Society, CIII, I, 61-84.
Paul Russell, (1995), Freedom and Moral Sentiment: Hume’s Way of Naturalizing Responsibility, (Oxford: Oxford University Press).
Gilbert Ryle, (1949), The Concept of Mind, (London: Hutchinson).
T. M. Scanlon, (1986), “The Significance of Choice,” in Sterling M. McMurrin, ed. (1988), pp. 149-216; reprinted in Stephen Darwall, (ed.), (1995), pp. 39-104.
__________, (1998), What We Owe To Each Other, (Cambridge, Mass.: Harvard University Press).
__________, (2002), “Reasons and Passions,” in Buss and Overton, eds., (2002), pp. 165-183.
347
__________, (2008), “Blame”, unpublished manuscript version, (Harvard University: Department of Philosophy).
__________, (2008), Moral Dimensions: Permissibility, Meaning and Blame, (Cambridge, Mass.: Harvard University Press).
Moritz Schlick, (2002) [1930], Fragen der Ethik, (Frankfurt: Suhrkamp).
Sidney Shoemaker, (1963), Self-Knowledge and Self-Identity, (Ithaca, NY: Cornell University Press).
Tamar Schapiro, (1999), “What is a Child?”, Ethics, 109: 715-38.
Ferdinand David Schoeman, ed., (1987), Responsibility, Character and the Emotions: New Essays in Moral Psychology, (New York: Cambridge University Press).
Pranab Kumar Sen and Roop Rekha Verma, (eds.), (1995), The Philosophy of P. F. Strawson, (New Delhi: Indian Council of Philosophical Research).
J. J. C. Smart, (1961), “Free Will, Praise and Blame,” Mind, 70, 291-306; reprinted in Watson, ed., (2003), pp. 58-71.
Angela Smith, (2005), “Responsibility for Attitudes: Activity and Passivity in Mental Life,” Ethics 115:2, 236-271.
Galen Strawson, (1986), Freedom and Belief, (Oxford: Clarendon Press).
__________, (1993), “On ‘Freedom and Resentment’,” in Fischer and Ravizza, eds., (1993), pp. 67-100.
__________, (1994), “The Impossibility of Moral Responsibility,” Philosophical Studies, 75: 5-24.
__________, (2008), Real Materialism and Other Essays, (Oxford: Oxford University Press).
P. F. Strawson, (1959), Individuals, (London: Methuen).
__________, (1962), “Freedom and Resentment,” Proceedings of the British Academy, 48, pp. 1-25; reprinted in Gary Watson, ed., (2003), pp. 72-93. [Referred to in the text as “F & R”; page references are to the Watson volume.]
__________, (1966), The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason, (London: Routledge).
__________, (1980), “Reply to Ayer and Bennett,” in van Straaten, ed., (1980), pp. 260-6.
__________, (1995), “Replies,” in Sen and Verma, eds., (1995), pp. 398-433.
348
__________, (1998), “Intellectual Autobiography of P. F. Strawson,” in Lewis Edwin Hahn, ed., (1998), pp 1-21.
John Updike, “Introduction,” in Greene (2003).
Peter van Inwagen, (1986), An Essay on Free Will, (Oxford: Oxford University Press).
Zak van Straaten, ed., (1980), Philosophical Subjects: Essays Presented to P. F. Strawson, (Oxford: Clarendon Press).
R. Jay Wallace, (1994), Responsibility and the Moral Sentiments, (Cambridge, Mass.: Harvard University Press). [Referred to in the text as “R & MS”.]
__________, (2002), “Précis of Responsibility and the Moral Sentiments”, Philosophy and Phenomenological Research, LXIV: 3, 680-1.
Gary Watson, (1975), “Free Agency,” Journal of Philosophy, 72: 8, 205-20; reprinted in Watson, (ed.), (2003), pp. 337-51.
__________, ed., (2003), Free Will, 2nd edition, (Oxford: Oxford University Press).
David Wiggins, (1967), Identity and Spatio-Temporal Continuity, (Oxford: Blackwell).
Bernard Williams, (1973), Problems of the Self, (Cambridge: CUP).
__________, (1985), Ethics and the Limits of Philosophy, (London: Fontana).
__________, (1993), Shame and Necessity, (Berkeley, California: University of California Press).
Ludwig Wittgenstein, (1953), Philosophical Investigations, (Oxford: Blackwell).