Rawls and moral psychology

Thomas Baldwin
In his obituary of John Rawls Ben Rogers remarked that after completing A Theory of
Justice (TJ) Rawls intended to develop further his ideas on moral psychology.
In the
event the debates aroused by TJ kept Rawls fully occupied and he never wrote the
extended account of moral psychology that he had apparently intended to develop.
Nonetheless there are many discussions of moral psychology in his writings, and this is an
important aspect of his thought which, in my judgment, has not received the attention it
deserves (for example, the four volume collection of papers on Rawls edited by Chandran
Kukathas contains no papers directly on this theme)
I What is Moral Psychology?
The first issue to be addressed here is what it is that Rawls has in mind when he writes of
‘moral psychology’. We can readily get an extensional, reference-fixing, answer to this
question from chapter 8 of TJ where there is a long section (§75) called ‘The Principles of
Moral Psychology’. Rawls here summarises and comments on the account of moral
development he has presented in the preceding sections, in particular his account of the
development of a ‘sense of justice’, a disposition to act in accordance with the principles
of justice for their own sake and to feel guilt and shame when one recognises that one has
violated these principles. This already gives us one feature of moral psychology, namely
that it deals with feelings and judgments whose content is distinctively moral, and I shall
come back later to Rawls’s account of the psychological ‘laws’ which are supposed to
give rise to a sense of justice. But the point to note now, which Rawls himself picks out, is
that on this account our psychology is itself affected by moral features of the context in
which we grow up and live:
Perhaps the most striking feature of these laws (or tendencies) is that their formulation
refers to an institutional setting as being just, and in the last two as being publicly
see www.guardian.co.uk/obituaries/story/0,3604,848488,00.html
Thomas Baldwin: Rawls and moral psychology
known to be such. The principles of moral psychology have a place for a conception of
justice; ..... Thus some view of justice enters into the explanation of the corresponding
sentiment; hypotheses about this psychological process incorporate moral notions even
if these are understood only as part of the psychological theory. (TJ 430)
Rawls recognises that some theorists will regard this as odd:
No doubt some prefer that social theories avoid the use of moral notions (TJ p. 430).
But, he holds, this is a mistake:
The moral sentiments are a normal part of human life. One cannot do away with them
without at the same time dismantling the natural attitudes as well (TJ 428).
So in his use of the term ‘moral psychology’ Rawls implies that in some respects our
psychology is inherently ‘moral’, not only in respect of its content, but also its context.
Several issues are raised by this conception of moral psychology. First there is the
issue of how it connects with other aspects of our psychology, in particular with the
science of human psychology. In lecture II of Political Liberalism (PL) Rawls gives the
final section (§8) the puzzling title ‘Moral Psychology: Philosophical not Psychological’
(PL p. 86). How can a ‘psychology’ not be ‘psychological’? one wants to ask. He then
begins the section as follows:
1. This completes our sketch of the moral psychology of the person. I stress that it
is a moral psychology drawn from the political conception of justice as fairness. It is
not a psychology originating in the science of human nature but rather a scheme of
concepts and principles for expressing a certain political conception of the person and
an ideal of citizenship. (PL pp. 86-7)
A first thought here might be that Rawls is now using the expression ‘moral psychology’
in a different way from that in which he uses it in TJ, as a way of capturing ‘a certain
political conception of the person and an ideal of citizenship.’ When one looks at the
detail, however, it turns out that Rawls has much the same extension in mind; indeed he
alludes here specifically to his discussion of moral psychology in TJ (cf. PL p. 86 fn. 35).
There is a difference of emphasis here, to the effect that our sense of justice includes a
commitment to an ‘ideal of citizenship’ which goes along with his general reorientation
towards a ‘political conception of justice as fairness’, but the substance of the position is
much the same, and it is worth emphasizing that he still uses the phrase ‘moral
psychology’ – it has not become ‘political psychology’.
So, if ‘moral psychology’ here is still much what it was, what does Rawls mean by
Thomas Baldwin: Rawls and moral psychology
saying that it is ‘Philosophical not Psychological’? I take it that the contrast he seeks to
indicate is between ‘a psychology originating in the science of human nature’ and a
psychology which includes reference to principles of justice. So, one might say, the
contrast would have been better expressed as ‘Moral Psychology: Political not Scientific’.
But even that, on reflection, wrongly suggests that the disjunction – political or scientific –
is to be understood in an exclusive way. For this cannot be Rawls’s view: if, as he
maintains, moral sentiments such as guilt are inseparable from natural attitudes such as
sociability (cf TJ p. 428 above), a ‘scientific’ psychology of these natural attitudes should
make reference to a moral psychology which deals with the moral sentiments and the
moral/political conditions under which these sentiments develop and are exercised. So
there cannot be a full separation between scientific and moral psychology. Rawls’s
treatment of this relationship is, I think, a bit uncertain. In TJ, in §74, which is called ‘The
Connection between Moral and Natural Attitudes’, Rawls distinguishes a thesis he
endorses, that the absence of feelings of guilt implies the absence of attachments such as
friendship (TJ p. 425) from the converse thesis, which he does not endorse:
while feelings of indignation and guilt, say, can often be taken as evidence for such
affections, there may be other explanations. (TJ pp. 425-6)
So here the position is clear: appropriate natural attitudes are sufficient, but not necessary,
for the related moral sentiment. In PL, however, the position seems to be reversed:
Human nature and its natural psychology are permissive: they may limit the viable
conceptions of persons and ideals of citizenships and the moral psychologies that may
support them, but do not dictate the ones we must adopt.
That is the answer to the objection that our account is unscientific. (PL p. 87)
If natural attitudes are only ‘permissive’, then they are clearly not sufficient for particular
moral sentiments; but one way of interpreting the thesis that they ‘limit’ moral sentiments
would be that they are necessary for them.
This is not by itself a big point and there are ways of bridging the apparent
difference between these positions. My preference is for the later approach, since it seems
to me sensible to treat the natural attitudes as broadly necessary but not sufficient for
moral sentiments. But the important point remains that there is a substantive connection
here (whatever it is), and thus that a full psychology of human nature needs to engage with
the moral aspects of human psychology, i.e. with moral psychology. If a ‘science of
human nature’ or a ‘natural psychology’ are, by stipulation, value-free investigations, then
Thomas Baldwin: Rawls and moral psychology
of course moral psychology will not be included within them; and I take it that Rawls is
using these phrases in this way in PL because he here takes it that ‘science’ connotes a
kind of truth-directed empirical inquiry which is unavailable in moral and political theory.
But this practice is questionable: given, to use his own phrase, ‘The Connection between
Moral and Natural Attitudes’ that is set out in TJ, it might have been better to argue that
moral psychology shows how a proper understanding of human life, a true ‘science of
human nature’, has to make room for values insofar as ‘some view of justice enters into
the explanation of the corresponding sentiment’ (TJ p. 430), i.e. the explanation of the
moral sentiments.
II Rawls and metaethics
Rawls’s approach to this issue connects with his Quinean metaethical perspective which is
the main thesis of his 1975 paper ‘The Independence of Moral Theory’
. Rawls here
develops the brief discussions in TJ in which he rejects the application of the
analytic/synthetic distinction within moral philosophy which would imply giving priority
to questions of definition over substantive issues of principle (TJ pp. 44-5, 506-7). If
anything, Rawls urges, the priority runs in the other direction. Just as the advances in
logical theory due to the work of Frege, Russell and others have profoundly transformed
the philosophy of logic and of language, he suggests,
Once the substantive content of moral conceptions is better understood, a similar
transformation may occur. It is possible that convincing answers to questions of the
meaning and justification of moral judgments can be found in no other way. (TJ p. 45)
Thus in the 1975 paper Rawls argues that
A relation of methodological priority does not hold, I believe, between the theory of
meaning, epistemology, and the philosophy of mind on the one hand and moral
philosophy on the other. To the contrary: a central part of moral philosophy is what I
have called moral theory; it consists in the comparative study of moral conceptions,
which is, in large part, independent. (‘Independence’ CP p. 301)
So insofar as Rawls has a metaethical perspective, it is a ‘bottom-up’ rather than a
‘top-down’ approach that he favours. An important element of this bottom-up approach is
a willingness to engage with psychology: for psychology lies at the heart of moral theory
This paper is reprinted in Rawls’s Collected Papers (CP), and all my references to it and
to others papers by Rawls are to this edition.
Thomas Baldwin: Rawls and moral psychology
since such a theory aims to provide
a deeper understanding of the structure of the moral conceptions and of their
connections with human sensibility; ….. We must not turn away from this task because
much of it may appear to belong to psychology or social theory and not to philosophy.
For the fact is that others are not prompted by philosophical inclination to pursue
moral theory; yet this motivation is essential, for without it the inquiry has the wrong
focus. (‘Independence’ CP p. 302)
Thus moral philosophy has to include the psychological aspect of moral theory (‘without it
the inquiry has the wrong focus’); and so far from moral psychology being dependent
upon philosophy of mind, the dependence runs, if anything, in the other direction.
Although by and large abstract philosophical debates about mind and body do not intersect
with moral theory, where there are connections, as between natural attitudes and moral
sentiments, philosophy of mind has to accommodate itself to moral psychology, to our
having a psychology which is not wholly value-free.
An important application of this ‘independence thesis’ occurs in Rawls’s
exposition of his doctrine of ‘Kantian constructivism’ in moral theory. This is the doctrine
that substantive moral principles which define the requirements of justice and the virtues
are to be ‘viewed as specified by a procedure of construction (...), the form and structure
of which mirrors both of our two powers of practical reason as well as our status as free
and equal moral persons’ (Lectures on the History of Moral Philosophy p. 237). Rawls
insists that this doctrine is not an application of some more general metaphysical or
epistemological doctrine concerning truth:
Furthermore, it is important to notice here that no assumptions have been made about a
theory of truth. A constructivist view does not require an idealist or a verificationist, as
opposed to a realist, account of truth. Whatever the nature of truth in the case of
general beliefs about human nature and how society works, a constructivist moral
doctrine requires a distinct procedure of construction to identify the first principles of
justice. (‘Kantian Constructivism’ CP pp. 351-2)
So Rawls’s constructivism is grounded firmly in his moral theory, in his Kantian faith in
the absolute priority of our autonomy with respect to the demands of morality.
Throughout his writings Rawls is mainly concerned to contrast his constructivist
doctrine with the position of the ‘Rational Intuitionist’ who holds that fundamental moral
principles, ‘first principles’, are not constructed but discovered through a capacity for
Thomas Baldwin: Rawls and moral psychology
intuitive insight. Historical figures such as Clarke are taken to exemplify this position, as
are 20th century philosophers such as Moore and Ross. The familiar objections to this
position, such as Mackie’s ‘Argument from Queerness’, appeal to general metaphysical
and epistemological considerations; in the light of his independence thesis it is not
surprising that is not the way in which Rawls argues against the Rational Intuitionist.
Instead his argument is rooted within his moral theory, in the priority he gives to the value
of autonomy:
Yet it suffices for heteronomy that these <first> principles obtain in virtue of relations
among objects the nature of which is not affected or determined by the conception of
the person. Kant’s idea of autonomy requires that there exist no such order of given
objects determining the first principles of right and justice among free and equal moral
persons. Heteronomy obtains not only when first principles are fixed by the special
psychological constitution of human nature, as in Hume, but also when they are fixed
by an order of universals or concepts grasped by rational intuition, as in Plato’s realm
of forms .. (‘Kantian Constructivism’ CP p. 345)
I shall return in a moment to Rawls’s comment here about Hume, but, I want first to
consider this critical discussion of rational intuitionism. For the rational intuitionist, first
principles are truths which express fundamental moral facts which both obtain altogether
independent of our grasp of them and place moral demands upon us that have no ground in
our capacities. Rawls, following Kant, takes it that this position is essentially a secular
version of a Divine Command theory and that moral demands cannot be in this way
altogether independent of us; morality can secure its authority over us only by answering
to our nature as free rational beings. This difference then brings with it two further
differences. First, although the constructivist agrees with the intuitionist that morality is
objective, he rejects any implication that moral principles are true (‘KC’ CP p. 340) since
he takes it that truth and facts are inseparable and he denies that there are any moral facts
because ‘the idea of constructing facts seems incoherent’ (PL 122). Second, they differ
significantly with respect to moral psychology: for the intuitionist, moral theory implies
only that our psychology include a capacity for intuitive moral insight; but because the
constructivist starts from a conception of our nature as free rational persons it assumes a
richer moral psychology (‘KC’ CP p. 346).
This second difference brings me back to the issue postponed earlier, namely
Rawls’s attitude to Hume. This is, I think, an important issue, since Rawls says all too

Thomas Baldwin: Rawls and moral psychology
little directly about the relationship between his Kantian constructivism and the ethical
non-cognitivism which has been such a prominent feature of 20th century ethical theory.
The main reason for this, I take it, is that insofar as the roots of 20th century ethical non-
cognitivism lie in general metaphysical and epistemological anti-realist doctrines such as
logical empiricism Rawls held that it is not sensible to attempt to deal with the position:
his independence thesis implies that it is a mistake to make the discussion of moral theory
in this way dependent on extraneous doctrines. One might well object that actually non-
cognitivist positions are best grounded in accounts of human motivation, in the ‘Humean
psychology’ which treats reason as essentially passive. The objection is, I think, correct:
but it precisely implies that it is in the context of his discussion of Hume that one might
hope to find Rawls’s attitude to non-cognitivism; and this implication is born out by the
fact that he tentatively endorses a ‘projectivist’ interpretation of Hume’s position
(Lectures p. 96).
So what then was Rawls’s objection to Hume? It is that his ‘psychological
naturalism’ is a kind of heteronomy (‘KC’ CP p. 345). In elaborating this objection in
‘Kantian Constructivism’ Rawls interprets Hume’s position as a kind of reductive
naturalism. What Rawls seems to have in mind here is that Hume treats our moral
sentiments as just special cases of more general natural sentiments whose function and
application can be understood without reference to any moral concepts. Rawls sums up his
judgment of Hume as follows:
What is distinctive of his view is that it seems to be purely psychological and to lack
altogether what some writers think of as the ideas of practical reason and its authority.
(Lectures p. 50)
I shall not pursue the question of the justice of this verdict; what interests me here is the
way in which Rawls damns Hume’s moral theory for being too dependent on psychology
while equally insisting, as against the Rational Intuitionist, on the importance of moral
psychology within his Kantian moral theory. The key issue here seems to be one of
reduction. For Rawls what is fundamental to moral theory is a certain conception of us as
‘moral persons’ with two fundamental moral powers, a capacity for ‘a conception of the
good’ and ‘a sense of justice’. Moral theory is then supposed to show how putative moral
principles are justified by viewing them as the normative framework hypothetically
adopted by a community of moral persons who find themselves living together; and within
this theory moral psychology is the part which explains the development and application
Thomas Baldwin: Rawls and moral psychology
of the fundamental moral capacities. Since the notion of a moral person with these moral
capacities is here taken to be fundamental, our moral psychology is not here taken to be
derivable from more general psychological capacities and dispositions in the way that, for
Rawls, Hume seeks to demonstrate. Nonetheless, as we have seen earlier, Rawls himself
acknowledges that our moral psychology is connected to our ‘natural attitudes’, as indeed
any sensible psychology must allow. So there is a delicate issue as to how this relationship
is understood – how there can be a relation of non-reductive dependence of moral
psychology upon ‘natural’ psychology. I shall come back to this; but I want first to tidy up
two loose ends.
First, following Rawls, I have identified the issue of psychological reduction as
fundamental to the Hume/Kant argument. Suppose a contemporary ethical non-cognitivist
were to agree in rejecting the reductive thesis, would that commit such a person to being a
Rawlsian Kantian? Clearly not! For such a non-cognitivist, while agreeing with Rawls that
moral judgments express our fundamental and irreducible moral psychology, would deny
that there is any objectivity to be constructed on this basis. By contrast Rawls holds that by
viewing moral principles as the collective expression of the psychology of moral persons
who are both rational and reasonable, unforced and systematically adaptable agreement
concerning moral principles is possible. Whether Rawls’s optimism is warranted is an
issue I shall not pursue here.
Second, when discussing Rawls’s critical treatment of Hume above I cited Hume’s
‘psychological naturalism’ as Rawls’s main complaint. And indeed when elaborating on
this Rawls says that, because ‘it formulates definitions of the basic moral concepts in
terms of nonmoral concepts .... naturalism is a form of heteronomy’ (‘KC’ CP p. 345 fn.
8). Rawls’s uncomplaining use here of the term ‘naturalism’ is in sharp contrast with his
position in TJ, where he writes:
A second approach (called naturalism by an abuse of language) is to introduce
definitions of moral concepts in terms of presumptively non-moral ones ... (TJ p. 506)
I draw attention to this earlier complaint because I think it is legitimate; what is at issue
between the reductionist and a non-reductionist such as Rawls is not well described as a
disagreement as to whether or not morality is grounded in human nature. It is, rather, a
disagreement concerning human nature, namely as to whether or not moral concepts have
a fundamental role in characterising our psychology. It is because Rawls holds that they do
have such a role that he summarises his conclusion in TJ in the striking phrase:
Thomas Baldwin: Rawls and moral psychology
In this sense we might say that humankind has a moral nature. (TJ p. 508)
It is in this sense, as I said earlier, he should hold that moral psychology belongs to the
science of human nature.
III Moral development
I turn now back to the beginning and more of the detail of Rawls’s account of moral
psychology, in particular to his account of the development of a sense of justice. In chapter
8 of TJ Rawls argues that our sense of justice is our disposition to act in accordance with
the principles of justice for their own sake, primarily as they concern our behaviour
towards fellow citizens within a well-ordered society. Rawls argues that this disposition is
a natural development of a deep tendency to reciprocity which is exemplified in our
psychological development: within the family the child develops a capacity for love as it
grows up in an environment within which it is itself loved - its capacity for love
reciprocates the love it has received; outside the family, in ‘associations’ which it joins the
growing child develops a capacity for friendship, trust and responsibility as it is itself
treated in these ways by other members of the associations to which it belongs; and finally
as a young adult it internalises the requirements of justice in virtue of the fact that through
being treated with respect and fairness by other citizens it recognises that self-respect is
dependent upon treating others in ways which show respect for them. This approach
implies that justice itself is fundamentally a kind of reciprocity, that one owes it to others
to accord them the status of someone who has the right to live a fulfilling life of their own
that one wants them to accord to oneself. For it is by thinking of justice in this way that
one can see how the limited reciprocity of the first two stages becomes a reciprocal
disposition which is a ‘sense of justice’ when it is extended to apply to relationships with
just anyone – typically fellow citizens, but in principle strangers as well.
The first issue to be addressed concerning this three-stage development is Rawls’s
observation, mentioned earlier, that the account ‘refers to an institutional setting as being
just’, so that ‘some view of justice enters into the explanation of the corresponding
sentiment’ – the sense of justice (TJ p. 430). This claim obviously has to be understood in
the context of his constructivism, so that the explanatory role here of justice is not that of a
distinctive fact which structures the context in which this development takes place. For,
according to Rawls, there are no such moral facts. Instead Rawls’ position is that this
moral development is accomplished by ties of personal affection and loyalty through
Thomas Baldwin: Rawls and moral psychology
which we come to see the point of fair practices which underpin the relationships from
which we benefit, such as growing up within a family, sharing a house with a group of
friends, or living as a citizen in a well-ordered society. Hence for Rawls it is by thinking of
moral principles as constructed in a way which expresses the equal freedom of moral
persons that we can understand the role of social practices and institutions which
incorporate these principles as enabling the reciprocal psychological development which
issues in a sense of justice.
In TJ Rawls suggests that we should think of this development as one which
involves successive transformations in the kinds of desire that we recognise
The three laws describe how our system of desires comes to have new final ends as we
acquire affective ties. These changes are to be distinguished from our forming
derivative desires ...... <Instead the laws> characterise transformations of our patterns
of final ends that arise from our recognising the manner in which institutions and the
actions of others affect our good. (TJ 432)
In his later writings Rawls writes of our capacity for ‘principle-dependent’ and
‘conception-dependent’ desires, as opposed to ordinary ‘object-dependent’ desires (PL pp.
82ff.), and this provides a slightly different way of thinking about this account of moral
development. Object-dependent desires are desires whose objects are personal goods
whose characterisation as such relies on no moral or other normative principle; by contrast
specifying the objects of principle-dependent desires such as fidelity involves moral
principles, and similarly specifying the objects of conception-dependent desires involves
moral ideals such as citizenship. Can one match these three types of desire with the three
stages? The match is fairly easy to see at stages two and three: the growing child who
becomes trustworthy through shared activities with friends can be thought of as someone
who begins to develop principle-dependent desires; and for Rawls the final development
of a sense of justice is accomplished as one identifies oneself as a member of a potentially
well-ordered society in which can aspire to the ideal of citizenship. The first stage,
however, is not to be thought of as that in which one becomes susceptible to object-
dependent desires, since these are all too prevalent anyway. Instead what is important at
this first stage is, I think, the development of a capacity to care about others, to make their
good one’s own good, since this is a prerequisite of the capacity for friendship which
enters into Rawls’s second stage. Initially, of course, the others in question are those who
care about one themselves, most notably the members of one’s family. So, at least for the
Thomas Baldwin: Rawls and moral psychology
purpose of completing the match between Rawls’s three stage account of moral
development and his three-way hierarchy of desires, one might say that Rawls’s hierarchy
of desires needs to be augmented by inserting a category of ‘relationship-dependent
desires’ between the object-dependent and the principle-dependent desires. Object-
dependent desires come for free, and do not mark a significant stage of moral
development; and the first stage of development requires, not principle-dependent desires,
but those which involve care for others to whom the subject, typically a child, is connected
by an affectionate relationship, such as the love between members of a family.
In setting things out in this way I have been trying to explore the way in which a
Rawlsian moral psychology might be thought to work. The sketch of moral development
above does, I think, meet the twin requirements of neither tacitly drawing on ‘Rational
Intuitionism’, the intuitive appreciation of moral truths as such, nor turning out to be a
form of reductive naturalism which derives moral sentiments from non-moral natural
attitudes. For although the account of moral development involves a hierarchy of desires,
starting from non-moral object-dependent desires, the progression is achieved though
transformations in which the subject’s motivational set is thought of as enhanced as a
result of including principles and ideals which the subject recognises as informing
practices and institutions that are essential to his own good as he grasps a broader sense of
his own identity as a member of groups in which his own good is dependent upon that of
others and vice-versa. Thus the rational basis of these principles and ideals is indeed such
that they can be viewed as requirements of Rawls’s Kantian constructivism.
IV The Problem of Stability
So far, then, so good for this sketch of the way a Rawlsian moral psychology might work.
But an important further consideration comes from a key role that this psychology is
intended to fulfil in Rawls’s theory of justice, namely that of providing a solution to the
‘problem of stability’. This is the problem of showing that a state whose political
institutions are just will be reasonably ‘stable’ in the sense that such a state need not rely
primarily on coercion to ensure its citizens obey the law and support the state’s institutions
because, for the most part, these institutions and laws enjoy the support of the citizens
anyway. The contribution of moral psychology to the solution of this problem was
supposed to be that it would show how the requirements of justice are in fact broadly
‘congruent’ with the interests of individual citizens, such that, as Rawls puts it, ‘being a
Thomas Baldwin: Rawls and moral psychology
good person is a good thing for that person’ (TJ p. 349).
Rawls’ first thought, in his 1963 paper ‘The Sense of Justice’, was that in showing
that the development of a sense of justice is the natural psychological outcome of life
within a society whose institutions are broadly just one shows that such a society is stable
(‘SJ’ CP p. 105). For there is a virtuous circle whereby the institutions and moral
sentiments reinforce each other. In TJ Rawls starts by repeating this line of thought in
chapter 8, arguing that the moral psychology of his conception of justice as fairness is
more conducive to stability than, say, the psychology that one would associate with a
utilitarian conception of justice (TJ §76 ‘The Problem of Relative Stability’ pp. 434-41).
But he does not treat this comparison as the end of the matter, since only a few pages later,
at the start of chapter 9, he says that only now is he in a position to deal properly with the
task of showing that justice as fairness and goodness as rationality are congruent (TJ p.
450). As he later acknowledged it is not clear what is going on here:
Throughout Part III too many connections are left for the reader to make, so that one
may be left in doubt as to the point of much of chs. 8 and 9 (‘Justice as Fairness:
Political not Metaphysical’, 1985, CP p. 414)
My own hypothesis is that Rawls may have felt that he could not simply rely on his
moral psychology to vindicate the congruence thesis for the reason that his account of the
development of a sense of justice was primarily causal and not normative: it showed how
one would expect a sense of justice to be inculcated among those growing up in a just
society; but it does not thereby show that it was good for them to have this motivation.
And without a demonstration of this, the ‘problem of stability’ was not fully resolved,
since if it remained an open question whether acting justly was good for one, one could
not reliably expect people to obey the laws of a just state despite their sense of justice.
I myself think there is something right about this but Rawls’s own favoured
account of the congruence of individual good and morality, namely that ‘the desire to act
justly and the desire to express our nature as free moral persons turn out to specify what is
practically speaking the same desire’ (TJ p. 501) seems to me to assume too much of his
own Kantian moral psychology. Of course, if the account of our individual good is
specified by reference to the full Kantian account of our psychology, it is not surprising
that justice and goodness are ‘practically speaking the same desire’. But now the practical
doubt which motivates the search for congruence has to be redirected at the question as to
whether our individual good is captured by the Kantian moral psychology. So, like most
Thomas Baldwin: Rawls and moral psychology
other readers of TJ, I do not find Rawls’s account of congruence, and thus his ultimate
solution to the ‘problem of stability’ persuasive.
Notoriously, there is a sense in which Rawls himself came to agree about this. For
in the 1992 ‘Introduction’ (esp. pp. xvi-xvii) to Political Liberalism Rawls explains that he
himself came to see that his Kantian argument for the congruence of goodness and justice
was unsatisfactory as a general solution to the problem of stability since his Kantian moral
theory was but one of several reasonable comprehensive moral theories (‘comprehensive
conceptions of the good’). This ‘fact of reasonable pluralism’, as he calls it, implies that
no one moral theory can be employed in a political philosophy which aims to provide
arguments that will be persuasive for all reasonable citizens; hence, he infers, it cannot be
right to rely on the Kantian theory to solve the problem of stability. This last move is not
unchallengeable: one might think that because the problem of stability is essentially
theoretical, insofar as it is not solved by explaining how a just society nurtures a sense of
justice among its citizens, one cannot expect to avoid drawing on one’s moral theory to
resolve it – even while recognising that there are other ‘reasonable moral theories’ which
will promote different solutions. But Rawls thinks that a different line of thought is
available: if one can show that there is a conception of justice which expresses the
reasonable political aspirations of adherents of different moral theories who acknowledge
the fact of reasonable pluralism, one will thereby be in a position to show that a state
which realises this conception should be stable, since ‘it can win its support by addressing
each citizen’s reason, as explained within its own framework’ (PL p. 143).
To follow this line of thought would take me well away from moral psychology
onto the topic of whether Rawls has a non-circular account of what it is for a conception of
the good to be ‘reasonable’. Instead I shall briefly address an issue that it may well be felt
that I should addressed earlier, namely that in my discussion I have failed to take proper
account of the shift in Rawls’s thought marked by his 1985 paper ‘Justice as Fairness:
Political not Metaphysical’, a shift which brings with it an increasing emphasis on
specifically political questions and correspondingly less attention to questions of moral
theory. One way to think of the shift is as a shift of concern from moral theory to political
philosophy: his 1975 thesis that moral theory is independent of metaphysics and
epistemology is replaced by the 1985 thesis that ‘the aim of justice as fairness as a
political conception is practical, and not metaphysical or epistemological’ (‘JF: Political
not Metaphysical’ CP p. 394); equally he writes that the title of his 1980 lectures ‘Kantian
Thomas Baldwin: Rawls and moral psychology
Constructivism in Moral Theory’ is misleading and that a better title would have been
‘Kantian Constructivism in Political Philosophy’ (‘JF: Political not Metaphysical’ CP p.
389). In my judgment, however, he does not abandon moral theory. Rawls continues to
suggest that justice as fairness is ‘a natural moral conception that can stand on its own
feet’ (‘JF: Political not Metaphysical’ CP p. 411); and in Political Liberalism he observes
right at the start (PL p. 11) that a political conception of justice is a ‘moral conception’,
one which is ‘worked out for a special kind of subject, political, social and economic
institutions’. So moral theory is necessarily still very much in play in Rawls’s later
writings. What is different, however, is that the role of theory in working out a moral
conception appropriate to political, social and economic institutions has to accommodate
‘the fact of reasonable pluralism’. So insofar as this moral conception involves the public
legitimation of political, social and economic institutions it can draw only on those aspects
of moral theory that are not in dispute between different reasonable theories – a project
which leads Rawls to rely increasingly on the ‘idea of public reason’ as the moral basis for
his political philosophy.
One result of all this is that moral psychology remains important to Rawls’ project,
even though this is conceived as primarily directed at political philosophy. This is
primarily because he remains committed to ‘political constructivism’ as a moral thesis,
and thus to the thought that the content of a legitimate political conception of justice can
be represented as ‘constructed’ by persons with the two basic ‘moral powers’, ‘a capacity
for a sense of justice and for a conception of the good’ (PL p. 93). But, having indicated
briefly why moral psychology, and in particular a sense of justice, remain central to his
concerns, I want to return to the problem of stability as he left it in TJ. For in his
discussion there of stability, and in particular of the congruence of justice and individual
good, he introduced some suggestive ideas which remain largely unexploited by him but
which, I think, can be used not only to provide a better solution to the problem of stability
than the Kantian approach he favoured but also to enrich his account of moral psychology
generally by inducing a transformation of it into social psychology or, perhaps better,
moral sociology.
V From moral psychology to moral sociology
The line of thought I have in mind is that which Rawls introduces in chapter 9, starting
with the conception of a ‘social union’ which he takes from Humboldt (TJ p. 459 esp. fn.
Thomas Baldwin: Rawls and moral psychology
4). A social union is a collective institution whose members cooperate in a type of joint
activity in order to achieve valuable ends which they cannot bring about without such
cooperation. Rawls gives the example of an orchestra as a social union of this kind: for it
is only within an orchestra which brings together musicians of many different kinds that
the individual musicians can take part in performing great orchestral works. In cases of
this kind, he writes, ‘persons need one another since it is only in active cooperation with
others that one’s powers reach fruition. Only in a social union is the individual complete’
(TJ p. 460). The existence of social unions shows us something important about the way in
which individuals with different abilities need to collaborate with each other in order to
achieve valuable ends; for Rawls the significance of this is that a just state can be regarded
as a social union, a ‘social union of social unions’ (TJ p. 462). This involves more than
just the familiar thesis that political cooperation is essential for the achievement of
individual goods: instead, if the social union model is to be applicable there has to be some
collective good (like the performance of a symphony) which is not available without the
collective participation of the citizens who are members of this social union, the just state,
and which is a way of fulfilling the goals of each individual citizen (‘Only in a social
union is the individual complete’). Rawls suggests that the collective good is just ‘the
public realisation of justice’ which meets the requirement of providing a distinctive form
of self-fulfilment because ‘the collective activity of justice is the pre-eminent form of
human flourishing’ (TJ p. 463).
This line of thought, though suggestive, is not obviously persuasive. And Rawls in
a way acknowledges this when, at the very end of the book, he sets himself to say why, in
his view, justice and individual good are congruent. For having alluded to the conception
of a just state as a social union, and introduced the ‘Aristotelian principle’ which tells us
that we will find much greater fulfilment in the complex activities which require
collaboration with others than in simple pleasures, all that he can say to connect justice to
this greater fulfilment is that ‘to share fully in this life we must acknowledge the principles
of its regulative conception, and this means that we must affirm our sentiment of justice’
(TJ p. 500); and having said this, he then moves on to his favoured Kantian defence of
congruence. What is going wrong here is that the social union model for the state does not
really work: the Rawlsian liberal state is not comparable to an orchestra, or a sports team
(another of his examples), an institution whose members rely on each other’s
complementary activities to achieve a valuable end which cannot be achieved in any other
Thomas Baldwin: Rawls and moral psychology
way. Rawls’s suggestion that ‘the public realisation of justice’ counts as such an end, for
example, is unpersuasive. It is not clear how it requires complementary activities by the
citizens as opposed to the public authorities whose responsibility it is. Perhaps if Rawls
were to agree with the communitarians that the state is a collective association with some
dominant political goal that supposedly meets the requirement of providing self-fulfilment
for all citizens, such as the establishment of a classless society, he could use the social
union model for the purposes of his congruence thesis. But, of course, that is exactly not
the way in which Rawls conceives of his liberal state.
Yet one should not for this reason dismiss altogether all of the themes that enter
into Rawls’s discussion of the idea of a social union, in particular the suggestion that
the members of a community participate in one another’s nature: we appreciate what
others do as things which we might have done which they do for us, and what we do is
similarly done for them. Since the self is realized in the activities of many selves,
relations of justice that would be assented to by all are best fitted to express the nature
of each. (TJ p. 495).
Rawls’s line of thought here is reminiscent of the kind of the kind of reciprocity that came
up earlier in connection with his account of our moral development, the three-stage
development of a sense of justice via the place of love in ‘the morality of authority’ and
that of trust in ‘the morality of association’. Rawls never connects this conception of
reciprocity that is central to his moral psychology with his later discussion of congruence;
but I want to propose that there are connections to be made here which enable one to fill
out both his moral theory and his moral psychology. A good place to start is with the good
which is for Rawls of primary importance: self-respect. He writes
It is clear then why self-respect is a primary good. Without it nothing may seem worth
doing, or if some things have value for us, we lack the will to strive for them. All
desire and activity becomes empty and vain, and we sink into apathy and cynicism.
Therefore the parties in the original position would wish to avoid at almost any cost
the social conditions that undermine self-respect. (TJ p. 386)
Suppose we now apply to self-respect an extension of the developmental approach
involving reciprocity from Rawls’s moral psychology, so that self-respect is held to be
dependent upon respect by others whom one respects oneself. It now becomes almost
embarrassingly easy to argue for the congruence of justice and individual good. The
argument runs as follows.
Thomas Baldwin: Rawls and moral psychology
Start from Rawls’ thesis about the primary value of self-respect:
(i) Any reasonable conception of the good will acknowledge that self-respect is a
primary good.
Add my proposal about the dependence of self-respect on respect for others:
(ii) The achievement of self-respect is dependent upon reciprocal relationships of
mutual respect;
Now add a Rawlsian thesis about justice as reciprocity -
(iii) The conception of justice as reciprocity is the conception of principles whose
institutional realisation would affirm the mutual respect of citizens for each other
It does now follow that
(iv) Any rational conception of the good will bring with it a ‘conception-based desire’
to living in accordance with justice, at least in a well-ordered society.
The point to attend to here is the proposal that self-respect is dependent upon
reciprocal relationships of mutual respect. I take this to be a truth of moral psychology, or
rather, moral sociology, since it essentially concerns relationships between people. One
way to think about this proposal is to view it as comparable to Hegel’s famous thesis
concerning self-consciousness, that it is dependent upon the consciousness of one by
others of whom one is oneself conscious:
Self-consciousness exists in and for itself when, and by the fact that, it so exists for
another; that is, it exists only in being acknowledged. (Phenomenology of Spirit §178)
This ‘acknowledgment’ takes the form of mutual ‘recognition:
They recognize themselves as mutually recognizing each other (Phenomenology of
Spirit §184)
Just what Hegel’s thesis here amounts to is notoriously obscure and disputed, and I shall
not attempt to elucidate it. But I mention by way of preparing the ground for the work of
another German philosopher whose work I do want to discuss briefly in this connection –
Axel Honneth. For what Honneth does in his book The Struggle for Recognition is to use
Hegel’s early discussion (in the Jena period which precedes his Phenomenology of Spirit)
of the way in which self-consciousness is dependent upon recognition by others as a way
of introducing a three stage theory of moral ‘recognition’ which is comparable to Rawls’s
three stage account of the development of moral psychology. And my optimistic
hypothesis is that if one were to insert something like Honneth’s theory into Rawls’s
account, thereby transforming Rawls’s moral psychology into a moral sociology, some at
Thomas Baldwin: Rawls and moral psychology
least of the problems Rawls encountered in bringing his theory of justice to a satisfactory
conclusion would be alleviated. This may sound outrageous: but there are in fact some
existing areas of overlap – Honneth draws extensively on G. H. Mead’s account of the self
and the generalised other in Mind, Self and Society – to which Rawls also alludes when
presenting his three stage theory of moral development (TJ p. 410 fn. 11).
One can get an introductory sense of Honneth’s theory from the following
passages from The Struggle for Recognition:
Hegel’s model starts from the speculative thesis that the formation of the practical self
presupposes mutual recognition between subjects. Not until both individuals see
themselves confirmed by the other as independent can they mutually reach an
understanding of themselves as autonomously acting, individuated selves. For Hegel,
this thesis has to be the point of departure, since it discloses, as it were, the basic
structural feature of the societal subject-matter with which he is concerned in his
theory of ethical life. (SR p. 68)
The second thesis (and the first that is constitutive for Hegel’s conceptual model)
asserts, on the basis of intersubjectivist premises, the existence of various forms of
reciprocal recognition, which are to be distinguished according to the level of
autonomy they make possible for an agent. Both the System of Ethical Life and the
Realphilosophie contained at least a tendency to assume – with regard to ‘love’, ‘law’
(Recht), and ethical life – a sequence of recognition relations, in the context of which
individuals reciprocally confirm each other to an increasing degree as autonomous and
individuated persons. (SR p. 69)
Honneth them summarises his own position, arrived at by a critical discussion of the
positions of Hegel and Mead, with the following table (SR p. 129, slightly modified):
Thomas Baldwin: Rawls and moral psychology
Patterns of love law ethical life
Mode of emotional cognitive social esteem
Recognition support respect
Dimension of needs, emotions moral traits, abilities
Personality responsibility
Forms of primary legal relations community of value
Recognition relationships (rights) (solidarity)
(love, friendship)
Practical basic self-respect self-esteem
Relation-to-self self-confidence
Forms of abuse, rape denial of rights, denigration,
disrespect exclusion insult
Threatened physical social integrity ‘honour’,
component integrity dignity
of personality
There is too much here to take in all the details, and much that one might want to
question. For example, although it is clearly correct to distinguish self-respect from self-
esteem (as Rawls does not), it seems to me that self-esteem is much more closely tied to a
sense of the value of one’s achievements than to one’s sense of ‘honour’ or dignity, which
connect rather with self-respect; for to possess dignity is to be worthy of respect whereas
we esteem those who have achieved something worthwhile. So one may want to alter
some of the entries in Honneth’s table. The question for me, however, is whether it is
appropriate at all to suppose that this three stage schema can be fitted ‘on top of’ Rawls’s
three stage account of moral development.
Thomas Baldwin: Rawls and moral psychology
There is, I think, no great conflict concerning the first stage, Honneth’s level of
‘love’ and Rawls’s ‘morality of authority’ which depends on mutual love. There is
however some tension concerning the second stage, between Honneth’s level of ‘law’ and
Rawls’s morality of association since Rawls thinks of his associations as essentially
informal, and thus lacking the status of ‘law’. But it is easy to see that Rawls’s account
might be extended to include, along with a sense of one’s duty to one’s friends, a broader
sense of ‘moral responsibility’, as Honneth puts it. The two schemata do appear to diverge
more decisely at the third stage, however, where Rawls is concerned with a sense of
justice and not with ‘social esteem’, as Honneth puts it. Yet it seems to me that the
differences here should not be exaggerated, since Rawls’s sense of justice is to be thought
of as a motivational disposition informed by a ‘conception-dependent desire’, a desire
dependent on an ideal conception of citizenship, which is not too far from Honneth’s
invocation of ‘community of value’ and ‘solidarity’ as the forms of mutual recognition
characteristic of this third pattern of relationship. The source of tension here is that
Rawls’s liberal theory of justice does not aspire to be a full theory of ‘ethical life’ of the
kind Honneth sketches, insofar as ethical life concerns social actions and relationships
that, for the liberal, are not political since they are essentially private. Yet as soon as one
thinks about supposedly private matters, such as the family or sexuality, it is apparent that
they have important political dimensions, even for the liberal. So if ‘ethical life’ is
construed as conceived as essentially public, then it does after all fall within the scope of
Rawls’s political conception. Hence there is less disparity between Rawls and Honneth
than may appear at first. Still, it is not my aim to show that their positions match up in all
respects. All I need there to be is sufficient similarity in their aims for it to be sensible to
propose Honneth’s emphasis on mutual recognition as a valuable enhancement to Rawls’s
moral psychology since it connects directly with his own emphasis on reciprocity.
The thesis I want to take from this comparison is, therefore, that mutual
recognition plays an essential role in the constitution of oneself as an autonomous moral
subject, responsible for oneself and happy to play one’s part in furthering the welfare of
others who are different. Very briefly, putting Rawls and Honneth together yields a story
of the following kind. Initially as a child one experiences oneself as a member of a family,
with needs and feelings which receive attention just because one’s parents and others
demonstrate that they love one for one’s own sake. This emotional security gives one a
sense of one’s bodily integrity and makes it possible for one to form affectionate ties with
Thomas Baldwin: Rawls and moral psychology
others who enjoy the same kind of emotional security. As these friendships broaden and
contact with others extends to participation in games and other collective activities, issues
of trust and respect arise. Here the key claim is that it is through finding oneself trusted by
others who are more experienced and whom one trusts oneself that one gains self-
confidence and self-respect. One comes to recognise oneself as trustworthy, and thus as a
morally responsible subject who is motivated to do what is required because of one
identifies oneself as someone who can be trusted to do this. This, then, takes one up to the
level of Rawls’s ‘morality of association’ and there is, I think, no sharp boundary between
this and Honneth’s sphere of law or ‘right’ (Recht). As I have already indicated, Rawls and
Honneth diverge a bit in their accounts of third level, but this divergence is a difference
that need not imply disagreement. Rawls’s emphasis on the development of a sense of
justice comes within the context of his overall conception of justice as fairness which
draws on the mutual recognition among citizens that all are entitled to the opportunity to
lead a fulfilling life so that the only legitimate differences between them in power and
wealth arise within institutions that aim to do the best for the worst off. So to have a
Rawlsian sense of justice is not simply to be motivated to obey the law because it is the
law: it is to be moved to work to uphold egalitarian institutions which promote the ideal of
justice as fairness; and this disposition draws on a fundamental mutual recognition among
citizens of their collective commitment to the ideal of justice as fairness, in particular to
the ideal that the lives of the worst off are to be worthwhile however difficult it may be to
achieve this. What Honneth’s emphasis on ‘self-esteem’ and solidarity add to this is a
commitment to valuing others for their achievements and abilities in a context in which
they similarly value one’s own different achievements and abilities and thereby develop
one’s capacity for self-esteem. Plainly this does not conflict with Rawls’s ideal, and one
might even connect them by taking it that justice as fairness aims to give everyone the
opportunity for self-esteem.
Two things now follow from this, one congenial to Rawls’s project, the other a
challenge to his underlying moral theory. The first point arises from the fact that by
extending his moral psychology into a moral sociology according to which the attainment
of moral autonomy is dependent upon recognition by others, the issue of the congruence
between justice and individual good is resolved. For my own good includes my status as a
moral subject, and this now emerges as dependent on my recognition as such by others
whom I equally recognise as moral subjects and therefore treat with justice, as persons
Thomas Baldwin: Rawls and moral psychology
who deserve respect and the opportunity to lead a fulfilling life. The line of thought here is
similar to that which Rawls essays when he writes about our membership of social unions,
and says that ‘the self is realised in the activities of many selves’ (TJ p. 495). But the state
is not a social union, and interdependent patterns of mutual recognition achieve in this
argument by honest toil what Rawls could only gesture at.
The second point is an implication of the first. If moral autonomy is not just a
‘given’, a fundamental moral capacity inherent in our moral psychology, in the way that it
is treated in Rawls’s moral theory, but is understood as constituted through the
internalisation of interdependent moral relations of mutual recognition of moral worth,
then moral theory will have to take as its starting point, not the autonomous individual
moral person who figures so prominently in Rawls’s writings, but moral relationships
between people in which they expect, require or hope for things from each other. Moral
psychology will have to give way to moral sociology in moral theory. That change, I
think, is a serious challenge to Rawls, especially to the project of ‘Kantian constructivism’
in both moral theory and political philosophy. If Honneth is right, we need instead a
programme that one might call ‘Hegelian recognition’. But I shall not speculate further on
its implications.

Sign up to vote on this title
UsefulNot useful