BOOK REVIEW

Jonathan Dancy (ed.), Reading Parfit. Blackwell Publishers, Oxford, 1997.

As Jonathan Dancy tells us in his preface, the essays in Reading Parfit are
published seven years late. The papers were collected by 1990. Dancy’s
original plan was to include responses by Parfit but those responses had
reached the length of three books by 1996 – a shift of proportion that made
independent publication necessary. So by now the essays have become a
part of an ongoing discussion and their publication is useful especially for
those of us who did not have access to them before. One aim of the book
is to provide a critical commentary to Derek Parfit’s Reasons and Persons
(Clarendon Press, Oxford, 1984, 2nd ed., 1987), to help readers and teach-
ers alike to follow its complex and winding ways. The essays appear in an
order corresponding to that of the book. Parfit’s book is complicated, and
so are the critical essays. In giving a brief summary of the main points of
each article I assume some aquaintance with Parfit’s book.
The first five essays center on rationality. In ‘Parfit and Indirectly Self-
defeating Theories’ Jonathan Dancy takes up Parfit’s claims that both prac-
tical rationality and consequentialism are indirectly self-defeating. One
who always tries to do what is best for him will end up worse than possible.
One who always tries to do what makes the outcome best will end up with
less than optimal outcomes. Dancy argues that Parfit’s versions of practical
rationality and consequentialism are more self-destructive than just indi-
rectly self-defeating, namely self-refuting. At least in some instances these
theories both tell us to pursue some aim and not to pursue it. Distinctions
between act and agent, and between outcome and motive do not help to
solve the contradiction.
In an essay entitled ’Rationality and The Rational Aim’ David Gau-
thier, unlike Dancy, accepts that practical rationality is ‘just’ indirectly
self-defeating. But that, he thinks, should make us wonder whether Parfit’s
account of rationality is correct. He suggests a formal interpretation of the
rational aim that one’s life should go as well as possible. Specification is
needed concerning the various substantive aims an agent wants to realize
in his life. Once these are known it is rational for the agent to adopt a set

Erkenntnis 49: 237–242, 1998.

238 BOOK REVIEW

of coherent dispositions to act which are likely to make his life go as well
as possible. There might be a set of actions that if performed would make
his life go better, but there is no coherent set of dispositions an agent could
adopt to realize this set. So in a sense his life is going as well as possible
even though some desirable actions are not an option.
In ‘Which Effects?’ Frank Jackson is concerned with consequentialism
and overdetermination. How can consequentialists disapprove of an act
that does not make any difference since a second person simultaneously
does something to the same effect? The main example in the literature is
someone’s being shot by both X and Y , either shot being sufficient for
murder. Parfit says we should consider both acts collectively and both are
wrong since together they make things worse. Jackson says we should con-
sider them separately and neither did something wrong in consequentialist
terms since neither makes things worse. But that does not mean there is
nothing morally wrong about X and Y in consequentialist terms. Both have
bad characters since they would have killed in a counterfactual single shot
situation, they were wrong in planning to kill together or individually, and
since neither could rely in advance on the others success each of them is
wrong in raising the subjective probability of murder.
‘Parfit and the Time of Value’ by Michael Stocker deals with the ratio-
nality of preferences concerning the time of an event. In Parfit’s examples
for pure time preferences the time of an event is clearly irrelevant for
its evaluation (‘future tuesdays’). Parfit is attracted to a normative view
according to which it should not make a difference to us if for instance
we feel some particular pain in the past, future or presence. In his very
inspiring essay Stocker lists and connects various ways in which time does
matter for ethics. It might not matter so much when we try to answer the
question what the right thing to do is. But it is crucial when we want to
know who we are and whether our life has value, questions that are central
to a good life. As for pain he argues that in most instances past pains should
not be cared about.
In their joint essay Philip Pettit and Michael Smith highlight a contrast
they find in Parfit between foreground theories and background theories
of practical rationality. Background theories specify to which reasons for
action it is rational to appeal, foreground theories are instruments to judge
the rationality of reasons objectively. It may be rational to act on a certain
desire while it is irrational to have that desire (because it is undesirable to
have it). Within the area of rationality they are particularly interested in
the view that my best reasons to act are my present desires. They discuss
which members of the class of Parfit’s present aim theories, ‘Parfit’s P ’,
are candidates for a foreground or a background theory of rationality. They
 BOOK REVIEW 239

radicalize Parfit’s critique of prudential reasoning, Parfit’s S. To give the

same weight to future aims and present aims in their view is not just an
unstable strategy, it is absurd.
The next six essays treat personal identity and related issues. David O.
Brink in his ‘Rational Egoism and the Separateness of Persons’ supports
Parfit in his reductionist account. Psychophysical continuity and connect-
edness are what matters, not identity of any sort. But contrary to Parfit
this is for Brink no barrier for rational egoism. Direct intertemporal com-
pensation is still possible. Together with Sidgwick’s argument for temporal
neutrality and the separateness of persons continuity is a plausible rationale
for rational egoism. Rational egoism is furthermore (and contrary to Pet-
tit/Smith) not an arbitrary position. Finally, since continuity makes direct
compensation over time possible, the welfare of continuants in fission and
fusion cases is within the limits of a person’s egoistic concern.
‘Parfit on Identity’ by Sidney Shoemaker is an abridged version of his
well known review essay on part 3 of Reasons and Persons in Mind (94
(1985), pp. 443–53). Shoemaker agrees with the reductionist account in
terms of psychological continuity and connectedness, but finds fault with a
number of details. Firstly, mental entities cannot be referred to without
assuming that they are functional parts of the mental life of a person,
they require subjects. Secondly, it is not the discovery of some empirical
fact that would decide between reductionism and non-reductionism. Even
after the ‘discovery’ of Cartesian egos reductionism about personal iden-
tity would still be an option. Thirdly, reductionism and non-reductionism
cannot be distinguished by the fact that non-reductionism treats personal
identity as being all-or-nothing while reductionism admits of degrees. If
there are no borderline cases like the Combined Spectrum, reductionism
may admit degrees of connectedness while the question of identity may al-
ways have a determinate answer. Finally, a reductionist account of personal
identity does not help to weaken the Self-interest theory of rationality.
Shoemakers last point is taken up by Mark Johnston in ‘Human Con-
cerns without Superlative Selves’. Psychological continuity and connect-
edness justify concern for the future. ’Justify’ should be put in quotation
marks since Johnston thinks that in metaphysics there is nothing anyway
that matters for self concern, so literally speaking any relation might be
taken to justify concern; a position that he takes to be strengthened by
reductionism. Parfit’s revisionist claims are weakened by several alterna-
tive ways of adjusting our expectations concerning self concern. John-
ston defends a reductionist position in which connectedness of the right
sort ‘constitutes’ personal identity, accepts fission and fusion but takes the
240 BOOK REVIEW

extravagant line to reject teletransportation as a case of preservation of

identity.
Simon Blackburn discusses problems specific for any account of per-
sonal identity from the third person perspective. It is in the first person
perspective that we deliberate what to do. In this perspective I cannot but
consider the future as open and as open to me, not just to someone closely
connected to me. I envisage myself as doing something and facing the
consequences, even if it is prudent to take into account that I might change
during the process. Blackburn calls this phenomenon with reference to
Kant’s theory of pure apperception the ’Unity Reaction’. This reaction
constitutes a difference between what I will do and what someone else
however dear to me will do. It might be only a difference for me, but
I cannot stop making this difference even if I become an adherent of a
bundle theory of the self. Since the argument if correct would not show
that Parfit’s account of personal identity must be false, Blackburn’s title
question ‘Has Kant Refuted Parfit?’ remains unanswered.
Judith Jarvis Thomson in ‘People and their Bodies’ shows just how
much of a mess our intuitions about personal identity are. With no under-
standing of what personal identity is in sight and far away from the simple
understanding that people just are their bodies, philosophers discuss crite-
ria of personal identity, physical and psychological. Both kinds of criteria
have their problems in solving hypothetical cases of survival, fission and
fusion, cases designed to put the who is who in doubt. Thomson is brilliant
in pointing out just how large these problems are. Her diagnosis concerning
the deepest source of these problems is: We cannot tolerate indeterminacy
in matters of personal identity. Confronted with thought experiments we
do not only expect a determinate answer to the question who the emerging
person(s) is (are). We also think there must be a determinate answer to
the question what the person subjected to the thought experiment feels
afterwards. Why this is so is unclear and requires a better understanding of
personal identity itself.
John McDowell questions in ‘Reductionism and the First Person’ a
presupposition of Parfit’s reductionism, namely, the separability of mental
states. McDowell contends that this kind of account of personal identity is
possible only if the content of mental states is separable from the person
whose states they are. But there are no identity-neutral mental contents.
Each content figures as related to ‘the same thinking thing’, the subject
as extended through time. A memory is always a memory from the first
person point of view, a memory of something the person herself experi-
enced and now remembers. Even a quasi-memory carries the illusion to be
an experience of the same experiencing subject. McDowells claims further
 BOOK REVIEW 241

that Parfit himself is very much a Cartesian. This has to do with Parfit’s
view that a simple view of personal identity involves belief in a ‘further
fact’ and McDowell’s conviction that it need not. In addition McDowell
thinks that even a simple view of personal identity is compatible with the
claim that it might be indeterminate if some continuation of someone’s life
is the continuation of one and the same person. I found the explanation of
these two theses and their interrelation rather opaque.
Robert Merrihew Adams titles his review essay (first published in The
Philosophical Review 98 (1989), pp. 439–84) ‘Should Ethics Be More Im-
personal?’. It is one of only two essays relating to part 4 of Parfit’s book, on
future generations. On Adam’s view ethics should contrary to Parfit rather
not become more impersonal. Firstly, we should not give up common sense
morality with its bias towards the agent’s nearest and dearest in favour of
a theory close to consequentialism. Common sense morality does not have
always optimal consequences but it is not self-defeating in the way Parfit
thinks it is, and well known bad consequences of consequentialism cannot
be avoided. Secondly, we should not stop caring differently about ourselves
and other people because of reductionism about personal identity. Like
Johnston (and in line with Susan Wolf) Adams argues that our caring is an
element of a form of life with no need of justification. Reductionism also
may not be taken as permission to lessen the requirements of desert and
compensation. We might not be entitled to punish after fission, but as long
as the causal basis for psychophysical continuity is normal it is sufficient
for desert. Parfit’s examples, designed to show that it is not, can be inter-
preted in a different way. Thirdly, an account of our obligations towards
future generations need not be impersonal. Since which people live in the
future depends on what we do now we have to consider their prospects of
life without regard to who they are and how their situation came about.
Instead Adams suggests we should consider the future as a joint project
of mankind that includes interest in the relation future people will have to
us. In a brief discussion of the Repugnant conclusion he questions Parfit’s
ranking of various possible worlds as better or worse independently of how
they came about.
This discussion is carried further by Larry S. Temkin in ‘Rethinking
the Good, Moral Ideals and the Nature of Practical Reasoning’. From
the Repugnant Conclusion we can learn that utility is not everything, that
gains in utility might be outweighed by losses in other moral ideals (per-
fectionism, happiness, equality, and alleviating the lot of the worst off),
and that numbers may count for other moral ideals like equality as well.
From the Mere Addition Paradox we can learn something important about
our moral ideals. Firstly, possible worlds cannot be judged without taking
242 BOOK REVIEW

their history into account. Secondly, they cannot be measured on one scale.
A (transitive) moral ranking of possible worlds is impossible. It might
be possible to compare states of affairs pairwise only, because too many
of our all-things-considered judgements involve ‘essentially comparative’
principles. Furthermore some of our principles cannot be applied to every
pair of possible worlds. Temkin tries to rebut the impression of irrationality
that accompanies his theses.
Reading Parfit covers most of the topics of Reasons and Persons. I
missed comments on identity and morality not limited to the ethics of
individuals. Reading Parfit is a real help if one is puzzled by Parfit and
looks for guidance. But I doubt that average students will be able to use
the book on their own without quite a lot of instruction.

Universität Bielefeld MARTINA HERRMANN

Abteilung Philosophie
Postfach 100131
33501 Bielefeld