Introduction: Meta-Ethics and Normative Ethics
I undertake two main tasks in this chapter. First, I aim to provide a brief overview of the
chapters in this volume and to show how they are related to one another. Second, I aim to
introduce the issues in moral philosophy that are addressed in the volume, and to do so in a way
that is accessible to general readers with little background in philosophy. Because of my second
aim, I discuss the chapters in the order that seems best pedagogically. My choice of which
chapters to emphasize also reflects my pedagogical goal.
As we go about our lives, we face many decisions. Some of the decisions seem to concern
only ourselves and people with whom we are intimate, such as decisions about behavior within
the family. Other decisions concern our responsibilities in our job. Some concern our relationship
to the state or the law, such as decisions about whether to abide by the tax code or whether to join
the armed forces. People who have governmental roles sometimes make decisions about
controversial social issues, such as the morality of capital punishment or the justice of the tax
system. All of us who live in democratic societies need to make decisions about such issues if we
intend to vote responsibly. Moral philosophy addresses the many abstract ethical and
philosophical issues that arise when we attempt to make such decisions in a reflective and
Of course, some decisions have little moral import, but moral considerations have a
bearing on a great many of our decisions. A person’s decision-making can also be shaped,
however, by considerations of self-interest, law, etiquette, custom, and tradition, and people in
professional roles who are subject to codes of “ethics” may take such codes into account in their
decisions. The question therefore arises, What distinguishes moral considerations from other
kinds of consideration? What does morality require? Does morality determine what we ought to
do all things considered? These questions are addressed in various chapters in the volume.
For present purposes, we can take a person’s moral beliefs to be the beliefs she has about
how to live her life when she takes into account in a sympathetic way the impact of her life and
decisions on others. This statement is more vague than I would like, and it prejudges certain
questions, but it is a place to begin. It is worth saying at the outset, moreover, that in this volume,
“morality” and “ethics” are used interchangeably.
This volume focuses on theoretical questions that can arise in thinking about any practical
issue as well as general moral questions of theoretical importance. Applied ethics is an area of
moral philosophy that focuses on concrete moral issues, including such matters as abortion,
capital punishment, civil disobedience, drug use, family responsibilities, and professional ethics.
Can war be just? Is euthanasia ever justifiable? This volume focuses, however, on questions that
are more abstract than these. For example, what kinds of actions are right or wrong? These
questions may seem far removed from concrete issues of everyday importance, but anyone who
tries to think her way through a practical problem, such as the question whether euthanasia can
ever be permitted, can eventually be led to the kinds of questions addressed in this volume. The
chapter by Gerald Dworkin is motivated by this point; Dworkin examines various philosophical
moral theories in an effort to see how well they are suited to help us with practical questions. All
of the chapters, however, deal with the abstract issues I am pointing to.
These issues can usefully be divided into two categories. First are general moral issues.
What kinds of actions are right or wrong? What kind of person should one be? What are the
moral virtues? What, in general, has moral value? What kinds of things make a person’s life go
well? What does justice require? Most generally, how should we live our lives? In answering any
of these questions, one would be making a moral claim or a claim with moral implications.
Normative moral theory aims to provide answers to the general moral questions that fall into this
category. Theories of this kind are sometimes called “first-order” in contrast with the “second-
order” theories that deal with questions in the second category.
The second category includes issues or questions about morality and moral judgment. Are
there moral truths? Do we simply have a variety of feelings and attitudes about moral issues, with
there being nothing in virtue of which one side of a disagreement is correct and the other
incorrect? Are there moral “properties”? For example, is there a property or characteristic that a
kind of action can have of being wrong in the way that there is a property a kind of action can
have of being unpopular? If so, is wrongness analogous to unpopularity, in that it is a relation
between an action and the attitudes of a group of persons? Or is wrongness a more “objective”
property? When a person makes a moral claim, is she expressing a belief or is she merely
expressing a feeling or an attitude, such as approval or disapproval? Is it possible to have moral
knowledge? What is the relation between morality and rationality? Would it be rational to
commit oneself to morality? Answering such questions does not require making a moral claim.
It requires making a claim about moral claims or about morality. This explains why the issues in
this category are called “second-order” or “meta-ethical”.
The chapters in this volume defend a variety of positions in both normative moral theory
and meta-ethics. The first part of the volume contains the chapters on meta-ethical issues and the
second contains the chapters on normative issues. Issues in these two areas are much more
closely connected than might seem to be the case, given what I say in this introduction. But it
will be easier to introduce the material if I discuss the two areas separately.
A philosophical study of morality is very different from a sociological or anthropological
study, or a study from the perspective of biology or psychology. One important difference is that
in moral philosophy we do not distance ourselves from our own moral views in the way we
would if we were engaged in a study of one of these other kinds. We do not take the fact that
people, including ourselves, have moral views as merely a datum to be explained. Our goal is not
merely to explain data of this kind, whether it be the distribution of moral beliefs and attitudes, or
the occurrence of selfish or altruistic actions. Rather, in moral philosophy, the correctness or
cogency or defensibility of moral claims, convictions and attitudes, and the probity of various
behaviors, are among the things at issue. Normative ethics makes moral claims in its own right.
Meta-ethics does not do this, yet, despite this, it is morally engaged. For among its central
questions are the questions whether any moral claims are true, and whether it is rational to
commit oneself to acting morally. One cannot answer such questions without taking a position on
the correctness or cogency of people’s moral convictions.
Moral realism takes an optimistic view on the issue whether moral convictions can be
correct or cogent. In the opening chapter in the volume, Geoffrey Sayre-McCord characterizes
moral realism as the position that (1) there are moral facts, (2) people’s moral judgments are
made true or false by the moral facts, and (3) the mere fact that we have the moral beliefs we
have is not what makes the moral facts be as they are. This is a highly abstract view that may be
difficult to grasp. For this reason, I am going to begin with an example.
Many people find it plausible that the requirements of morality are determined by God’s
commands. This idea is a useful starting place because most people understand it immediately,
and because it points the way to the divine command theory, which is a kind of moral realism.
Philip Quinn defends a divine command theory in his chapter. The idea is, for example, that lying
is morally wrong (if it is wrong) due simply and exactly to the fact that God has commanded that
we not lie. More generally, Quinn holds that a kind of action is morally obligatory just in case
God has commanded that actions of that kind be performed, and, he also holds, God’s
commanding that an action be performed is what makes it obligatory. So he holds that actions
can have the properties of being obligatory, permissible, or forbidden–these are standardly called
the “deontic” properties–and he holds that such properties depend on God’s commands. God’s
commands bring it about that the wrong actions are wrong and the required actions are required.
Views of this kind have been discussed by philosophers for centuries, and indeed the
standard objection to them is derived from a discussion in Plato’s dialogue, Euthyphro. The
objection takes the form of a dilemma. Either actions are commanded by God because they are
obligatory or they are obligatory because they are commanded by God. The first alternative is
incompatible with Quinn’s divine command theory, since the theory holds that what makes an
action obligatory is God’s commanding that it be performed. On this view, actions are not
obligatory independently of God’s commands, so God could not take an action’s being obligatory
as a reason to command it. But the second alternative seems unacceptable. For it seems to allow
the possibility of God’s commanding something arbitrary or horrible, and in that case, according
to the theory, the action would be obligatory. Quinn discusses the story in Genesis (22: 1-2) in
which God orders Abraham to sacrifice his son, Isaac. The divine command theory seems to
imply that in this case it was obligatory for Abraham to sacrifice Isaac, and indeed that whatever
God commanded Abraham to do would be obligatory, no matter how arbitrary or horrible.
Quinn’s answer to the challenge is that God’s goodness ensures that His commands are
not arbitrary. To make this reply work, however, Quinn cannot say that goodness depends on
God’s will in the way that the obligatoriness of an action depends on God’s commands, for if he
said this, the Euthyphro objection would come back to haunt him. (Is what God wills good
because He wills it, or does He will it because it is good?) What Quinn says instead is that
something is good just in case it resembles God in a relevant way. God is the standard of
Goodness. Since God resembles Himself, He is good. Deontic or duty-related properties depend
on God’s commands, but axiological or evaluative properties, such as goodness, do not.
The difficulty with Quinn’s approach is that the fact that God is good does not seem to
guarantee that His commands will not be horrible if His being good is simply a matter of His
resembling Himself. It is trivial that God resembles Himself, but if God is perfectly good, this is
a substantive and important moral fact. It would be different if there were an independent
standard of goodness and if God qualified as perfectly good when measured against this standard.
But if we added an independent standard of goodness to the theory, we would be leaving behind
Quinn’s idea that all moral statuses depend on God.
The chief problem with the divine command theory can be seen if we consider people
who do not believe that there is a God. An atheist could accept that actions are obligatory just in
case they are commanded by God, but since an atheist holds that there is no God, she would be
committed to denying that any actions are obligatory. She would be committed to denying that
any actions whatsoever are right or wrong. On Quinn’s view about goodness, she would also be
committed to denying that anything whatsoever is good or bad. Even a theist would be
committed to holding that if God does not exist, then nothing is right or wrong, good or bad.1
This implication of the divine command view is surely implausible. Even if there is no God,
there are cases of harming others, coercing them, torturing them, and so on, and it is difficult to
believe that such actions are not wrong, and that there is nothing bad about them, although this is
implied by the divine command theory if God does not exist. Surely one would not accept this
implication of the theory if one thought there were an alternative. And there are alternatives, as
we shall see, including other kinds of moral realism.
For my purposes in exploring the kinds of moral realism and anti-realism, it will be useful
to define realism somewhat differently from the way Sayre-McCord defines it. I shall take moral
realism to combine the following five doctrines.
(1) There are moral properties (and relations).2 There is, for example, such a thing as
wrongness. The divine command theory implies that actions have the property of being wrong
when God has commanded that they not be performed. It implies that if God exists and has
commanded that we not perform certain actions, those actions are wrong. Hence, on these
assumptions, it also implies the second doctrine of moral realism: (2) Some moral properties are
instantiated. For example, some actions are wrong. Moral realism also includes two doctrines
about moral thought and language: (3) Moral predicates are used to ascribe moral properties. And
(4) moral assertions express moral beliefs. When we call an action “wrong” we are ascribing to it
the property wrongness and we are expressing the belief that the action is wrong. Finally, moral
realism includes a doctrine designed to clarify its first thesis: (5) The moral properties, in that
they are properties, have the metaphysical status that any other property has, whatever that status
is.3 This doctrine belongs in the list because some philosophers who reject moral realism think
that we can call wrongness a “property” without misusing English even though wrongness is not
a property that would be recognized in an adequate metaphysics. An adequate metaphysics must
give some account of the status of properties such as redness and deciduous-ness. These are not
moral properties, of course, and they differ in a variety of ways from any moral property.
Nevertheless, a moral realist insists that wrongness is like these properties in that it is also a
property, and that, in this respect, it has the same metaphysical status as all other properties.
Moral realists disagree about various things, but they disagree chiefly about the nature of
the moral properties. We can think of a realist theory as proposing a “model” that explains the
nature of these properties. The divine command view sees wrongness as analogous to the
property of being unlawful. It sees morality as, in effect, a divine legal system. Other versions of
realism propose other models. There are both naturalistic and non-naturalistic versions of
realism, where naturalism treats moral properties as “natural” properties. Quinn’s divine
command theory is a kind of non-naturalism, or it certainly appears to be. For Quinn holds that
the goodness of something is a matter of its resembling God; God is the standard of goodness. As
usually understood, however, God is not part of the natural world.
Naturalistic moral realism is defended in the chapter by Nicholas Sturgeon. Sturgeon
holds that the moral properties are ordinary properties, akin to a variety of ordinary garden-
variety properties, such as the property of being a quarter dollar or the property of being
deciduous. He does not attempt to give an account in non-moral terms of what rightness or
wrongness are. He thinks that there is no adequate reason to suppose that moral properties are
any more problematic or puzzling than are the properties that are theorized about in biology or in
psychology, such as being deciduous or being in pain. The latter properties supervene on the
basic physical natures of things in the sense that, roughly, any biological or psychological change
in a thing depends on some underlying change in the physical nature of the thing. Similarly,
Sturgeon holds, moral properties supervene on the basic nature of things. But just as we do not
expect to be able to characterize the biology of a tree in non-biological terms, we should not
expect to be able to characterize the moral nature of an action or an institution or a person in non-
moral terms. We should not expect, that is, to be able to specify in non-moral terms exactly
which natural properties are the moral properties. On this point, Sturgeon disagrees with most
philosophers who have thought about ethical naturalism. Most have thought that the viability of
naturalism depends on there being, for each moral property, a true reductive identity statement
that identifies that property with a natural property picked out in non-moral terms. As Sturgeon
says, they have thought that “ethical naturalism must be, in this sense, reductive.” Sturgeon
denies that this is so. He thinks, moreover, that to understand the moral properties, there is no
substitute for normative theorizing. To understand what justice is, we need to think about what
makes for just institutions. Meta-ethics, then, is continuous with normative moral theory.
Moral naturalism is attacked vigorously in the chapter by Jonathan Dancy. Dancy is a
realist, but he thinks that naturalism is indefensible because it is unable to make sense of the
normativity of moral judgment. There are, unfortunately, a variety of ways to understand
normativity. The basic idea is that when all goes well, a person’s moral judgments guide her
actions. Suppose, for example, that a person thinks that she ought to help people in countries
suffering from famine, and suppose that she receives a letter from Care asking for a donation to
help people suffering from famine. In this case, if all goes well, she will be motivated to make a
donation (Smith, 1994, p. 7). Moral judgment, especially judgment about what one ought to do,
has a kind of characteristic direct relevance to action or choice. This idea is unfortunately vague,
and in a paper on the topic, I distinguish three “grades” of normativity and argue that moral
naturalism can accommodate all three (Copp, 2004).
Dancy disagrees. He thinks that, to understand the normativity of moral judgment, we
must take the moral properties to be intrinsically normative. The problem for naturalism is, he
thinks, that no natural property is intrinsically normative. We can express his argument in terms
of the idea of a moral fact–a fact consisting of something’s having a moral property. Naturalists
claim that moral facts are natural facts. But Dancy argues that moral facts are normative and that
no natural fact is normative. Why not? He holds that natural facts are not directly and
immediately relevant to a decision about what to do in the way that normative facts are.
One might turn Dancy’s argument into an argument against moral realism. J.L. Mackie
argued for a position called the error theory, according to which there are no moral facts (Mackie,
1977, chapter 1; see also Joyce, 2001). The error theory says, in effect, that moral beliefs have the
status of superstitious beliefs, such as beliefs in hobgoblins. Mackie offered several arguments
for his view, including an argument something like Dancy’s. Mackie held that the moral
properties, if there were any, would be intrinsically normative. Rightness would have “to-be-
doneness” built into it. He thought that such a property would be “queer,” and unlike “anything
else in the universe.” He therefore concluded that there are no such properties. Accordingly, he
held, all basic moral claims are false.4 In effect, Mackie took Dancy’s line of reasoning, added
the premise that all properties are natural, and concluded that there are no moral properties. In so
doing, he rejected one of the central doctrines of moral realism.
Mackie’s error theory is a highly controversial. It implies that nothing is morally wrong.
This is as hard to accept as the implication of divine command theory that if God does not exist,
nothing is wrong. There are cases of harming others, coercing them, torturing them, and so on. It
is difficult to believe that such actions are not wrong, although this is implied by the error theory.
Three premises are on the table: first, that moral judgment is normative, second, that no
natural property is normative, and third, that there are no non-natural properties. In arguing for
non-naturalism, Dancy accepted the first two of these premises but rejected the third. Assuming
the truth of moral realism, he argued from the first premise to the conclusion that the moral
properties are normative, and so he thought that, given the second premise, the moral properties
must be non-natural. Mackie was not prepared to assume the truth of moral realism. He accepted
all three premises and was led to the error theory. But it is possible to accept all three premises
without accepting the error theory. One can be led, instead, to non-cognitivism, which is another
form of moral anti-realism. Like the error theory, it denies that there are moral properties, but it
proposes to explain the normativity of moral judgment in another way.
The core idea of non-cognitivism is the thesis that the state of mind of a person who
accepts a (basic) moral claim is not a belief or any other kind of cognitive state, but is, instead, a
conative state or a motivational state, akin to a desire. Any fully developed version of non-
cognitivism would need to say exactly what kind of state is involved, but we can neglect such
details here. The view could be that the relevant state of mind is an “attitude.” In his chapter,
Simon Blackburn speaks of “stances.” The root idea is that, for example, a person who accepts
that capital punishment is wrong is in a state of mind that could most accurately be described as
an attitude of disapproval of capital punishment or a stance of disapproval. Non-cognitivists hold
that moral assertions express such conative stances rather than beliefs. (Because it takes a thesis
of this kind to explain the meaning of moral assertions, non-cognitivism is often described as
“expressivism.”) What would lead one to accept this view?
Blackburn begins with the idea that cognitive states such as beliefs, and conative states
such as desires, have different “directions of fit.” A belief represents the world as being a certain
way and it tends to go out of existence, or should tend to go out of existence, when we have
evidence that the world is not that way. Conative states are different. A desire need not go out of
existence when we have evidence that the world is not the way we desire it to be. If my car fails
to start one morning, my belief that it is reliable should tend to go out of existence, but I might
still desire that it be reliable. If I do have this desire, I will be motivated to have the car repaired.
In this sense, conative states such as desires have a different direction of fit from beliefs. They do
not represent the world as being one way rather than another. Their function is to motivate action
rather than to represent the world. Blackburn holds that moral states of mind have the direction
of fit of desires and other conative states. They are “directive” rather than “representational.” If a
person holds that she ought to help the victims of famine, and if she receives a letter from Care
asking for a donation, then, if all goes well, she will be motivated to make a donation. For,
according to the non-cognitivist, to hold that one ought to help is to have a stance that supports
helping. It is, inter alia, to have an inclination or desire to help.
Philip Kitcher argues, in his chapter, that the best biological explanation of the existence
of altruistic behavior supports non-cognitivism. In his view, evolutionary biology supports the
idea that the function of moral attitudes is to create motivation for the kinds of altruistic behavior
that improve social cohesion. We accept a system of moral rules, but its content is not shaped by
antecedently existing moral truths. As he says, “The criterion of success [of a system of moral
rules] is not accurate representation, but the improvement of social cohesion in ways that
promote the transmission of the system itself.” One might combine Kitcher’s view, according to
which moral codes have the function of improving social cohesion, with the view that moral
truths are “grounded in” the tendency of a system of moral rules to improve social cohesion. The
result would be a cognitivist moral functionalism.5 Kitcher holds, however, that there is no need
to postulate the existence of moral truths in order to explain altruistic behavior.
A non-cognitivist clearly would have difficulty accepting any of the doctrines that
constitute moral realism. She denies that moral assertions express moral beliefs, for she holds
that there are no moral beliefs to express. She will also want to deny that there are moral
properties. For if there are moral properties, then surely it is possible to believe that something
has a moral property, and presumably such a belief would qualify as a moral belief. For instance,
there is a state of mind that we could express by saying “Torture is wrong,” and if there are moral
properties, including the property wrongness, it would be difficult to deny that this state of mind
qualifies as a belief that ascribes wrongness to torture. So the non-cognitivist will be led to deny
that there are moral properties. Of course, if there are no moral properties, then there are no
moral properties to be instantiated or to have any kind of metaphysical status, so she will deny
two more realist doctrines. And, finally, she will deny that moral predicates are used to ascribe
moral properties. For it would be odd to hold that moral predicates are used to ascribe moral
properties while denying that an assertion, say, of the sentence,“Torture is wrong,”expresses the
belief that torture has the property thereby ascribed. Accordingly, non-cognitivism gives one
reason to deny all five of the doctrines that constitute moral realism.
The problem is that moral thought and discourse at least appear to be cognitive in nature.
As Blackburn says, everyday moralizing has a “realist surface”. We speak of people as having
moral beliefs. We speak of moral beliefs as true or false. A person who holds, say, that capital
punishment is wrong would have difficulty denying that capital punishment has the characteristic
or “property” of wrongness or of being wrong. She would have difficulty denying that the term
“wrong” is used to talk about wrongness and to express beliefs about things that are wrong.
Accordingly, everyday moralizing seems to commit us to four of the realist doctrines. The
missing doctrine is the récherché thesis that wrongness has the same metaphysical status as other
properties. In everyday moralizing, we do not worry about metaphysical issues. Perhaps, then, the
difference between realism and a plausible anti-realism would boil down to this fifth doctrine.
Blackburn aims to develop a position that accepts and explains the realist surface of
everyday moral discourse without abandoning the underlying, anti-realist doctrine of non-
cognitivism. He calls his view “quasi-realism”. In his view, there are merely moral stances, such
as moral approval and disapproval, but we have come to speak as if such stances are beliefs and
as if there are properties such as wrongness. He sometimes calls his position “projectivism,”
drawing an analogy with the way a slide projector can make it seem as if there is, say, a tree in
front of us when there is, in reality, only the play of light on a wall. His idea is that in doing
metaphysics, we see that there are no moral properties, but, in ordinary moralizing, we speak as if
there were, thereby projecting our moral stances out into the world. The trouble is that a quasi-
realist will be tempted by ‘minimalism’ about our use of the term “property” – a view that allows
us to say a ‘property’ is ‘expressed’ by every predicate in the language, including moral
predicates, but that denies this has any metaphysical significance. If Blackburn accepts such a
minimalism, he would be forced to agree that so-called moral properties have the same
metaphysical status as other so-called properties.
Where are we then? An anti-realist denies at least one of the five realist doctrines, but a
quasi-realist may find it difficult to deny any of them, given the realist surface of moral discourse
and the availability of minimalism. Yet Blackburn would deny that he is a realist. In the end, he
distinguishes his position from realism on the ground that, as he says, whatever we call them,
moral states of mind have the “directional” direction of fit rather than the “representational.”
That is, in effect, he denies that there are moral beliefs. Strictly speaking there are only stances.
Recent work by Blackburn and others has made it difficult to draw a clear and bright line
between moral realism and anti-realism. In his chapter, Sayre-McCord attempts to clarify
matters. Blackburn and other non-cognitivists and quasi-realists need to be clear about what they
reject in moral realism. In some ways, moral realists face a more difficult burden, however. As
Sayre-McCord explains, they need to explain the nature of the moral facts, how we can have
knowledge of them, and why these facts give us reason to act in one way rather than another.
It is highly plausible that a person who has a moral conviction is in a relevant conative
state of some kind, such as a state of approval or disapproval. A person with the conviction that
capital punishment is wrong is naturally said to disapprove of capital punishment, and in saying
that capital punishment is wrong, she would naturally be taken to express disapproval of it. This
idea is fully compatible with moral realism, however (Copp, 2001). Many predicates in our
language are “colored”, to use Frege’s term (Frege, 1984, pp. 161, 185, 357). For example, there
are impolite terms for various ethnic groups that are used both to predicate membership in the
group and to express an attitude of contempt. Moral predicates could be colored in a similar way.
They could be used to predicate a moral property, such as wrongness, and also to express a
corresponding attitude, such as disapproval. This idea is not a problem for moral realism.
Indeed, it is compatible with moral realism to go beyond this and treat moral judgment as
concerned at root with the appropriateness of moral attitudes, such as approval and disapproval,
disgust and shame. Blackburn’s projectivism holds that moral judgment involves a potentially
misleading projection of such attitudes onto a morally neutral reality. We might instead see the
moral attitudes as responses to features of the world that make them appropriate. A moral
property might then be understood as a “response-dependent” property, much as color properties
are often taken to be properties the nature of which is that they tend to cause certain associated
visual experiences.6 A number of research programs are exploring this idea. Justin D’Arms and
Daniel Jacobson lay out the geography of the territory in their chapter. They distinguish
projectivism from “perceptivism”, which holds that the moral sentiments are responses to, or
perceptions of, morally relevant features of the world. They distinguish a purely dispositional
variety of perceptivism from the “sensibility theory” that has been proposed by John McDowell
(1985). Ultimately they argue that the projectivist and perceptual metaphors are both misleading.
What they find plausible is an idea that both views share–the sentimentalist idea that, as they say,
“evaluation is to be understood by way of human emotional response.”
Michael Slote explores a related idea. He sees moral sentimentalism as contrasting with
rationalism, by which he means the view that reason rather than sentiment is the source of moral
judgment and moral motivation. He sees sentimentalism as a position that straddles both
normative and meta-ethical issues, since he thinks it goes hand in hand with a virtue theoretic
approach in normative ethics and with a plausible account of the nature of moral properties. The
chief moral sentiment, in his view, is empathic concern. He holds, for example, that moral
goodness consists in empathic concern for others.
One might worry that sentimentalism supports a kind of relativism since the empathic
concern of different people might be engaged by different things. Slote thinks he can avoid this
worry since, on his account, the reference of our moral terms is fixed by our actual empathic
reactions, not by reactions we might have in merely possible circumstances. But it is not clear
what rationale can be given for taking our actual empathic reactions to fix what counts as right
and wrong. Perhaps our actual reactions can be improved morally. Moreover, it is possible that
different people or cultures actually have very different empathic reactions to things. Given this,
it is not clear how best to understand Slote’s theory. Suppose that my empathic concern is
engaged by thoughts of capital punishment but yours is not. In this case, Slote’s account could be
taken to imply that it is indeterminate whether capital punishment is wrong. Or it could be taken
to imply that capital punishment is wrong-relative-to-me but is not wrong-relative-to-you. It is
not clear, then, that Slote’s sentimentalism can avoid a troubling relativism.
Notice that there is a kind of “normative relativism” that is highly plausible. For instance,
it is plausible that whether telling a lie would be wrong depends on the circumstances. It might
not be wrong to lie to Alice if telling her the truth would distract her while she is doing neuro-
surgery. The underlying idea could be expressed crudely by saying that any plausible moral
evaluation depends on, or is “relative to,” the circumstances. This thesis is surely highly
plausible, but I want to focus on a kind of meta-ethical relativism that is much more interesting
James Dreier advocates a relativism of this kind in his chapter. In his view, the moral
“properties”, such as rightness, are not monadic properties, but are actually relations to the moral
standards of relevant person(s) or groups. For example, there may be rightness-relative-to-Alice
as distinct from rightness-relative-to-Bill, and an action that is right-relative-to-Alice might not
be right-relative-to-Bill. Here is an analogy. Weight is a relation between an object’s mass and
the local gravitational field. This is why an object has a different weight on the moon than it has
on the earth. The relevant gravitational field must be specified or assumed before we can fully
understand an assertion to the effect that something has a given weight. Similarly, in Dreier’s
view, a system of moral rules must be specified or assumed in order for us to understand what
proposition is expressed by an assertion to the effect that something is right or wrong. In contexts
in which different moral systems are at issue, assertions to the effect that something is “right”
will express different propositions and different rightness-relations. Dreier proposes a “speaker-
relativism,” according to which the moral system of the speaker is the relevant one. If Alice says
“Capital punishment is right,” she expresses the proposition that capital punishment is permitted
in her moral system, whereas if Bill says this, he expresses the different proposition that capital
punishment is permitted in his moral system. Of course, it is possible that Alice and Bill accept
different systems so that what Alice says is true but what Bill says is false.
Dreier thinks that this view is supported by the widely-accepted thesis that there is an
“internal connection” between moral judgment and moral motivation, the thesis that, necessarily,
a person who believes she morally ought to do something is thereby motivated to some degree to
do it. Stephen Darwall has called this thesis “judgment internalism” (Darwall, 1983, pp. 54-55).
Judgment internalism figures in many arguments in meta-ethics. Blackburn invokes it in arguing
for quasi-realism, and Mackie invoked it in arguing for the error theory. Dreier thinks that
speaker relativism can explain judgment internalism without the counter-intuitive implications of
these theories. If speaker-relativism is true, a person who asserts sincerely that she ought not to
do something is thereby expressing a belief, and she presumably must also have motivations that
incline her not to do the thing. For, in Dreier’s view, accepting a moral system is a matter of
having certain relevant attitudes and motivations. So it appears that speaker relativism explains
the connection between moral belief and motivation that is postulated by judgment internalism.
Judgment internalism is controversial, however, as Sayre-McCord explains. Certain
forms of moral realism conflict with it. The divine command theory does not ensure that there is
a connection between moral belief and motivation. Sturgeon’s moral naturalism and Dancy’s
non-naturalism both reject it and are in this sense “externalist.” Sturgeon argues that externalism
is actually more plausible than internalism. And there are familiar objections to internalism. It
seems possible, for example, for a depressed person to lose all motivation to do the right thing.
Her beliefs about what is right could remain unchanged while her motivations waste away.
One serious objection to Dreier’s relativism is the “disagreement argument.” Speaker
relativism seems to imply that if Alice says, “Capital punishment is right,” and Bill says, “It is
not the case that capital punishment is right,” they have not disagreed. Alice has expressed the
proposition that capital punishment is right-relative-to-Alice, and Bill has not denied this. He has
expressed the proposition that it is not the case that capital punishment is right-relative-to-Bill.
But this seems implausible. Surely Alice and Bill have disagreed in the imagined situation.
Dreier would respond that there is a pragmatic disagreement between them; they would be
expected, say, to vote in different ways in a referendum on capital punishment. But Dreier’s view
has odd implications. On his view, for example, if Alice says, “Capital punishment is right,” Bill
could reply, coherently and truly, by saying, “I agree with you, and, in addition, it is not the case
that capital punishment is right.” This would be a very puzzling conversation! Intuitively, what
Alice says contradicts what Bill says in saying “It is not the case that capital punishment is right.”
A position that has counter-intuitive implications is difficult to defend, but we should not
conclude that it is impossible to defend. I myself have attempted to support a kind of meta-ethical
relativism against the disagreement argument (Copp, 1995, pp. 218-223).
Several of the authors we have discussed agree that morality is in some fundamental way
the province of the sentiments. Blackburn, Kitcher, Slote, Dreier, and D’Arms and Jacobson
agree about this, although they disagree about the details. An alternative view is that morality is
fundamentally the province of practical reason. To understand this, we need to look at details.
Serious complications arise immediately, for there are theoretical issues about practical
reason that are similar to the issues we have been discussing about morality. There are first-order,
normative issues: What are the basic factors that determine which actions are rational and which
are not? And there are second-order meta-theoretical issues: Are there truths about rational
behavior? Is there a property that an action can have of being rational? Do claims about the
rationality of actions express beliefs or do they merely express non-cognitive attitudes such as
approval or disapproval? We will set aside most of these questions.
The essay by Peter Railton explores Humean and neo-Humean theories of practical
reason and their relation to morality. A neo-Humean theory holds that rationality is basically a
matter of efficiency in serving one’s intrinsic ends or goals, where a person’s intrinsic goals are
taken as given–or as they would be if the person had more accurate information. On the standard
neo-Humean view, it is a contingent matter whether a person has a good practical reason to be
moral, for people’s goals vary widely. A person who had no goal that would be well-served by
morally appropriate behavior would have no practical reason to act morally. True, most people
have the goals of avoiding punishment and the disapproval of others, and it may be that these
goals typically give them good practical reason to act morally. But this would be a purely
instrumental reason to act morally, and the existence of such a reason would be a contingent
As against this position, some philosophers hold that an adequate account of morality
must show it to be a necessary truth that every person who is subject to morality has good
practical reason to be moral. If we accept this claim, there are at least four ways to proceed. One
is to concede that there may be rational agents who are not subject to morality because they lack
good practical reason to be moral. This approach seems to embrace a skepticism about morality.
A second strategy is to adopt the view that a person’s goals, which, on the standard neo-Humean
view, determine what she has reason to do, also determine what morality requires of her. This
position is a version of ethical egoism, which we will discuss briefly when we turn to issues in
normative ethics. A third strategy involves amending the neo-Humean account of practical
reason in an attempt to avoid the skeptical result. The difficulty is to motivate such an
amendment without giving up the basic idea that rationality is instrumental to serving one’s
intrinsic goals. A fourth strategy would involve abandoning the neo-Humean view by arguing
that compliance with morality is partly constitutive of rationality. Aristotelian and Kantian
theories take this approach.
Some theories of the latter kind have been called “constructivist” (Rawls, 1980). They
can be seen as constructing ethics out of a theory of practical reason or as “reducing” morality to
practical reason. Versions of the second and third strategies can also be seen this way.
David Gauthier took the third of these strategies in arguing for a contractarian moral
theory (Gauthier, 1986). In much of life, we need to cooperate and coordinate our actions with
other people. As Railton explains, however, even if each person acts rationally, according to a
standard neo-Humean account, so that everyone serves her goals as well as possible given what
everyone else is doing, it may be that everyone would have done better at serving her goals if
everyone had acted differently. A situation that illustrates this possibility is a so-called prisoners’
dilemma. In a prisoners’ dilemma, no-one can do better for herself, given what everyone else is
doing, but everyone could do better for themselves if everyone were to act otherwise. To achieve
the situation that is better for everyone, however, each must forego attempting to achieve what
would be best for herself. Gauthier concludes from this that it is not always genuinely rational to
attempt to maximize one’s own advantage. He went on to argue, in effect, that morality exists to
solve problems of cooperation and coordination that are modeled by the prisoners’ dilemma.
Intuitively, rational persons ought to be able to cooperate. Gauthier thinks that a plausible
account of rationality would dictate complying with agreements to cooperate, provided that the
other parties to the agreement were also likely to comply. So, he concludes, rational persons
would not be disposed to maximize their own advantage in general and without restriction.
Instead, rational persons would be “constrained maximizers.” They would be disposed to comply
with systems of norms, mutual compliance with which would be mutually advantageous, in
situations in which it is reasonable to believe that those with whom they are interacting are
similarly disposed. This means that a rational person would comply with morality provided that
doing so promised to be mutually advantageous and provided that enough others were likely
enough also to comply. This, in brief, is Gauthier’s contractarianism.
There are two main objections. First, even in Gauthier’s view, it is a contingent matter
whether a given person has good practical reason to be moral. Whether she does will depend on
whether enough others are likely to comply and on whether morality promises to benefit her in
the circumstances, given her abilities and goals. It might seem that an adequate account would
show morality to have a stronger and more internal connection to rationality than this. Second,
Gauthier’s view treats morality as merely of instrumental value. It might seem that it is
intrinsically important to treat people fairly and that it is a mistake to view fairness as worthy of
respect only to the degree that it serves our goals to adopt a disposition to be fair.
Kantian approaches are intended to show morality to have the intrinsic value and tight
internal connection to rationality that, so-far, has seemed elusive. Kantian moral theory is a
fertile area of contemporary research that is especially interesting because of the way it seeks to
link meta-ethical issues with issues in normative ethics. The volume includes two chapters on
Kantian theory. In one, Stephen Darwall develops and defends a Kantian connection between
morality and rationality. In the other, Thomas E. Hill, Jr., explicates Kantian approaches to
normative ethical theory. The volume also includes a chapter by Julia Annas that, among other
things, outlines an Aristotelian account of the connection between morality and rationality.
The basic Kantian doctrine is that moral obligations are categorical. There are, however,
different views about how best to spell out this idea. For Darwall, the idea is that it is necessarily
the case that if an action is morally wrong, there is a reason not to do it; moreover, crucially, this
reason has “genuine normative weight”, such that anyone who is deliberating rationally will take
it into account as appropriate and assign it conclusive weight. Darwall accepts this thesis but he
sees it as difficult to support. In summary, he argues that neo-Humean theories cannot
accommodate it and that typical forms of moral realism also cannot accommodate it. He argues
as well that Christine Korsgaard’s recent neo-Kantian attempts to support it are unsuccessful
(Korsgaard, 1996). Indeed Darwall thinks that Kant’s arguments need to be supplemented.
Darwall’s own argument begins from an idea of moral responsibility. A moral agent is
responsible for complying with the demands of morality, and responsibility implies the capacity
to respond to the moral demands placed on oneself. Moral agents view each other as responsible,
moreover, in that they hold each other liable to respond to the demands placed on them.7 Darwall
holds that an assumption of “reciprocal accountability” of this kind is essential to the practice of
holding people to be subject to moral obligations, and he argues that reciprocal accountability
presupposes that other people can see the reasons for acting the way we say they are obligated to
act. It also presupposes that the reasons in question are independent of the variable ends or goals
these people might have, for we put forward claims of moral obligation to people merely as
moral agents, not as people with special ends or goals. Moreover, in putting forward a demand,
we assume the person addressed is capable of complying. Hence, in putting forward such
demands we presuppose that people can act on reasons that are independent of their variable ends
or goals. We presuppose that, in this sense, people are autonomous and capable of acting on
As we saw, Darwall begins with a conception of moral responsibility. In his chapter, John
Fischer explores a variety of conceptions of moral responsibility and their connection to the idea
of free will. His main focus is on the challenge of causal determinism. We typically take it that
we have the freedom to choose what to do from a menu of alternatives, each of which is open to
us. But the thesis of causal determinism says that everything we do is caused deterministically by
events that happened in the past. It seems to follow from this that we do not have the freedom to
choose. For it seems to follow that the “choice” we make from the “menu of alternatives”
available to us at a given time was determined by events that happened prior to the choice. If so,
it seems, we lacked the power to choose or to do otherwise than we did.
The thesis of causal determinism challenges moral theory in a variety of places. It seems
to imply that we are not free to determine how we will act. It may even imply that we have no
obligation to do anything other than what we actually do. For it is standardly assumed that we
have an obligation to do something only if we can do it–“ought” implies “can”–and causal
determinism seems to imply that we have no power to act differently than we actually act.
Finally, the thesis of causal determinism appears to imply that we lack moral responsibility for
our actions. For it is often assumed that we are morally responsible for doing something only if
we could have done otherwise, and causal determinism seems to imply that we have no power to
do other than we actually do. Fischer explores all of these worries.
In turning from meta-ethics to normative ethics, we turn from issues about ethics to issues
in ethics. We turn to questions such as, What kinds of actions are right or wrong? What kind of
person should one be? There are many theories about these issues. In thinking about the
differences among them, it is helpful to consider the answers they give to two closely related
questions. What is the basic matter of moral concern? And what are the fundamental or basic
moral truths? The disputes posed by these questions are central to normative ethics.
First, what is the basic or fundamental matter of moral concern? Is it the kind of life we
should live? Is it the kind of person we should be? Is it the actions we perform? Is it the kind of
character we have? Is it our motivations or intentions? Is it goodness or value–either the
goodness in a person’s own life, or the overall goodness of the state of the world and the
condition of people in the world? Second, what are the fundamental or basic moral truths? Are
they propositions about the kind of life we should live? Are they propositions about the kind of
person we should be? Are they about the kinds of actions we are required to perform, or about
the kind of character we ought to have, or about our motivations or intentions? Or are they
propositions about goodness or value? Typically, a theory that proposes or argues that certain
moral truths are basic to ethics then attempts to support other moral propositions by deriving
them in one way or another from the basic truths. But theories can differ in how they attempt to
do this, and they can also differ in their views about the exact status of the truths they take to be
basic. Of course, a theory could instead reject the idea that certain moral truths are basic in any
interesting sense. And a theory could take it that all or several of the matters of concern are
It is useful to categorize moral theories on the basis, inter alia, of the positions they take
on these disputes. As we will see in what follows, there is a tendency for a theory to take the
same position on both disputes. That is, there is a tendency to hold that the basic moral truths are
propositions about the basic matter of concern. In classic virtue theories, for example, the basic
concern is with the kind of person we should be, or with the kind of character we should have,
and, in these theories, propositions about what kind of person to be or about what kind of
character to have are treated as fundamental to morality. In the ethics of care, the basic concern is
with relationships motivated by care, and the basic moral truths are about such relationships. In
standard deontology, the basic concern is with right action or moral duty and the basic moral
truths are propositions about our duties. In Kantian theory, the basic concern is with rational
agency. The fundamental moral truths are judgments about rational agency, such as judgments
about the maxims that a rational agent could will to be universal laws or judgments about the
respect owed to rational agency. In rights-based theories, the basic concern is with rights, and the
fundamental moral truths are propositions about the rights we have. Consequentialism presents a
more complex situation, however. In consequentialism, the basic truths are or include
propositions about intrinsic value or goodness. In different kinds of consequentialism, however,
different things are taken to be matters of basic concern. In act consequentialism, the basic
concern seems to be with right action, and the rightness of an action is a matter of the value of its
consequences. In virtue consequentialism, the basic concern is with our character, and the best
traits of character are those, the having of which tends to lead to the best consequences. In all
forms of consequentialism, however, the basic truths are or include propositions about goodness.8
The two disputes I have been discussing may seem intractable, but they are in the
the background of a debate that has dominated normative moral theory, a debate about the theory
of right action. The moral assessment of actions is a central concern in our moral life. In any
situation, we can wonder what would be the right thing to do. A theory of right action attempts to
answer the question, What are the basic factors that determine which actions are right and which
are wrong? Or, what are the right-making properties of actions? A theory of right action is
shaped by a conception of what is fundamental to morality. Theories that disagree about the
content of the basic moral truths, or about the basic matter of moral concern, can be expected to
disagree as well about right action. They will differ about the basic right-making properties.
To be sure, some normative theories do not aim to provide a theory of right action. In
their chapters, Julia Annas proposes a kind of virtue ethics and Virginia Held defends an ethic of
care, neither of which provides a theory of right action. They would deny that moral philosophy
needs to provide such a theory, or perhaps that it can provide one. They would argue that disputes
over right action have distracted moral theory from more central concerns.
Among approaches that do aim to provide a theory of right action, the central divide is
between consequentialist and non-consequentialist theories. Consequentialist theories share the
basic idea that the rightness of an action depends in some way on the promotion of the good.
Hence, consequentialism grounds the theory of right action in a theory of intrinsic good, or a
theory of value. It is in this way that consequentialism takes propositions about the good to be
basic or fundamental. It is difficult, however, to draw the distinction between consequentialism
and non-consequentialism in a precise way, and the distinction has sometimes been contested.
The problem is that different kinds of consequentialism specify the right-making property in
different ways, even if all specify that it is a function of the promotion of goodness. We can say
that a consequentialist theory of right action proposes a criterion that takes the rightness of an
action to be a function of the promotion of intrinsic goodness. But different theories propose
different functions, and consequentialists also disagree about what things are intrinsically good.
Non-consequentialist theories of right action include deontological theories, rights-based
theories, and Kantian theories. The term “deontological” is often used to describe any such
theory. But as I use the term, deontological theories are those that take the basic matter of moral
concern and the fundamental moral truths to be about the rightness of actions or about our duties.
Understood in this way, Kantian theories and rights-based theories are not best viewed as kinds
of deontology. They are non-consequentialist, but they share with consequentialism the idea that
judgments about the rightness of action are derivative. In consequentialism, they are derivative
from judgments about value or goodness. In Kantian theories, they are derivative from judgments
about rational agency. In rights-based theories, they are derivative from judgments about rights.9
It is convenient to begin with consequentialism because the best known consequentialist
theories have a relatively simple structure and because other kinds of normative theory typically
situate themselves in relation to consequentialism. I therefore turn to the chapter on value theory
by Thomas Hurka. Value theory is important in its own right, which is sufficient reason to
consider it, but consequentialism lacks content unless it is combined with a theory of value.
It is important to distinguish the idea of an intrinsic good from the idea of an instrumental
or extrinsic good. Instrumental goods are good or valuable only because of something else they
bring about–something which is good in itself–whereas intrinsic goods are good in themselves. It
is plausible, for instance, that enjoyment and understanding are intrinsic goods whereas money is
good only instrumentally–because of the intrinsic goods it can perhaps buy. The distinction
between the intrinsic and the instrumental can be drawn in different ways, as Hurka explains. The
main point, however, is that our concern should be with intrinsic goods. The first step is to come
to an understanding of what things are intrinsically good.
Hurka holds, very plausibly, that there is a great variety of intrinsic goods. He argues
against hedonism, which is the view that only pleasure is intrinsically good, and against desire
theories, which hold that the good in a person’s life is her getting what she desires
intrinsically–or what she would so desire if she were rational and informed. He favors a kind of
perfectionism; that is, he favors a view according to which the good is not determined by desire
but rather should guide desire, and according to which pleasure is not the only intrinsic good.
Perfectionist theories set standards for our improvement or betterment, both with respect to what
we desire and with respect to what gives us pleasure. Most perfectionist theories are pluralistic,
listing a variety of goods including such things as knowledge, friendship, creativity, and moral
virtue. Hurka discusses strategies a perfectionist theory might follow to explain the unity in the
set of intrinsic goods and to explain how various kinds of goods can be compared.
The most simple kind of consequentialist theory is “act consequentialism,” according to
which the morally required action in a situation is the action that, among the agent’s options,
produces, or would produce, the most good. But there is an enormous variety of consequentialist
theories, and debates about their plausibility and formulation are astonishingly complex.
To begin with, consequentialists disagree about the theory of intrinsic good. Some are
hedonists; some accept a desire theory; and some are perfectionists. A hedonist who accepted a
simple act consequentialism would be committed to saying, for example, that a person is morally
required to visit a friend in hospital just in case this is the option that would produce the most
pleasure overall. Indeed, she would be committed to saying that a person is morally permitted to
visit a friend in hospital only if there is no alternative that would produce more pleasure. But a
perfectionist might hold that friendship is intrinsically good and, moreover, that the direct
expression of friendship itself has great intrinsic value. Because of this, she might hold that there
is always moral good to be gained by expressing friendship through such acts as visiting a friend
in hospital. Hence, unless a person with a friend in hospital could do more good in some other
way, she is permitted and indeed required to visit her friend.
Consequentialists disagree about other matters as well. Most important, they disagree
about how to formulate the criterion of right action–about the precise relation between goodness
and rightness. A modest amendment of act consequentialism would take into account the fact
that the consequences of an action can be uncertain or unfixed. It would say that the rightness of
an action depends on the expected value of its consequences rather than the actual value of its
consequences–the expected value of an action is a measure constructed by taking the value of its
consequences in different possible scenarios, weighing these values by the probability of the
scenarios, and aggregating the weighted values into a measure of the overall value of the action.
Some consequentialists favor a simple and direct criterion, such as the act
consequentialist criterion, but there are alternatives.10 Some favor a much more indirect criterion.
An example is “rule consequentialism” according to which an act is morally required if and only
if it is required by the code of rules the currency of which in society would have the best
consequences. There are varieties of rule consequentialism, depending on how precisely we
understand such things as the currency of a code of rules. In principle, a rule consequentialist
might think that a person is morally required to visit a friend in the hospital because treating
friends this way is a generally beneficial practice–even if the consequences on a particular
occasion are less good than the consequences would be of not visiting the friend.
So far we have been comparing direct and indirect criteria of right action. But consider
the question of how such a criterion is to be used. The question is whether people ought to think
about what to do by applying the criterion. There is a debate about this both among act
consequentialists and between them and their critics. This debate is also sometimes described as
concerned with a kind of indirection. The so-called direct view says to apply the criterion in
moral decision making. An act consequentialist who took this view would recommend that we
decide what to do by considering which of our actions would have the best consequences. She
would recommend, in effect, that we pursue the good directly. She would treat the act
consequentialist principle as a decision procedure as well as a criterion of rightness. McNaughton
and Rawling discuss some of the problems with this view. The so-called indirect view treats the
principle simply as a criterion of rightness and rejects the idea that it is to be used in general as a
decision procedure. It says that the question of how to decide is itself one that is to be determined
by the criterion (Bales, 1971). On this view, a consequentialist theory recommends that we
decide what to do in the way that the criterion implies is the right way. For act consequentialism,
this is the way of deciding such that deciding in that way would have the best consequences. The
right way to decide might not involve the direct pursuit of good consequences, for it might be
best to decide what to do by following traditional moral rules without giving any thought to
consequences. Perhaps, for example, it would be best to be moved directly by friendship, in
visiting our friend in the hospital, rather than to worry about costs and benefits. The calculating
attitude that weights costs and benefits could have negative consequences for our friendships and
for other intrinsic goods. In light of problems with the direct view, act-consequentialists tend to
favor this indirect view. McNaughton and Rawling and other critics argue that the indirect view
is also problematic.
There are, then, many forms of consequentialism. Anyone defending consequentialism
must choose his poison. Anyone attacking it as a general style of theorizing must attack every
variety. She must find some underlying mistake or problem that is common to all kinds of
consequentialism. In doing so, she must bear in mind the variety of theories of value as well as
the variety of forms of direction and indirection.
The complexity among alternatives to consequentialism is at least equally striking as the
complexity among forms of consequentialism. The Ten Commandments offer a familiar
deontological view. But even here we must remember that rule consequentialism might
recommend the Ten Commandments as the best set of rules for our society.
In their chapter, David McNaughton and Piers Rawling aim to defend a “Rossian”
deontology of the kind that was first articulated by David Ross (1930). Rossian deontology
postulates a plurality of basic moral principles, such as the principle not to harm people and the
principle of promise keeping. The duties postulated by these principles are prima facie in that
they can conflict with one another, and when they do, the relative importance of the conflicting
duties must be weighed in order to determine what to do all things considered.
A Rossian principle may seem to imply that a relevant corresponding property of actions
is always right-making or wrong-making. For example, the principle that we ought not to lie may
seem to imply that lying always at least tends to be wrong. Some “particularists” would argue,
however, that no property of actions is always right-making or wrong-making in a way that
would support the truth of a Rossian principle. In their chapter, Mark Lance and Margaret Little
aim to clarify what is at issue in debates about particularism. On their account, particularism is
the denial that there are true moral principles with all of the classical characteristics of being
exceptionless, explanatory, and epistemically useful. On this showing, Rossian deontology may
be a kind of particularism because it allows that there are exceptions to its basic moral principles.
Traditional deontology recognizes three significant moral statuses. First are constraints,
such as the duty not to kill innocent people. These duties constrain us even when a prohibited
action has good consequences. For example, to take a far-fetched example, the duty not to torture
prohibits torturing Allan even if by doing so we could prevent someone else from torturing Bill
and Carol. Second are duties of special relationship, such as duties of friendship and duties of
family. And third are options. We normally think that there is a limit to how much good we are
morally required to bring about. Traditional deontology agrees that there is a limit and gives us
options to pursue our own projects even in circumstances where we could otherwise do more
good. McNaughton and Rawling object to consequentialism mainly on the basis that it cannot
account for options and the duties of special relationship. They think, for example, that duties of
friendship are morally basic in a way that consequentialism misses, since it sees everything of
moral significance as boiling down to issues about the impersonal good. Moreover our concern
for our own lives and personal projects is basic. Rule consequentialism may make room for
options, but only if the currency of a system of rules with options works best overall. This makes
room for options but without giving a fundamental significance to our own personal concerns.
The most surprising aspect of McNaughton and Rawling’s view is that they reject
constraints. Deontology has been bedeviled for thirty years by a line of argument according to
which deontological constraints are paradoxical. The idea is basically as follows: If it is
forbidden to torture Allan, then it must be a bad thing if Allan is tortured. But suppose that
someone else will torture Bill and Carol unless I torture Allan. If it is forbidden to torture, it must
be worse (other things being equal) if two people are tortured than if only one is tortured. So it is
better (other things being equal) if I torture Allan, thereby ensuring that Bill and Carol are not
tortured, than if I do not torture Allan, thereby ensuring that Bill and Carol are both tortured.
Given this, it seems, it would be paradoxical if there were a constraint against torture that
prohibits torturing one person to prevent the torturing of two. Yet the idea that my torturing one
can be justified by the fact that I would be saving two from torture is a consequentialist idea. It
appears, then, that instead of imposing constraints against torturing, a plausible view would treat
torturing as a bad to be avoided. It would be a form of consequentialism.11
McNaughton and Rawling do not draw the consequentialist conclusion, but they find the
argument against deontological constraints to be successful. They therefore adopt a deontology
that rejects constraints of a traditional kind, such as constraints against lying and torture. They do
hold, however, that there are a variety of proscriptions that are not constraints. For example, they
hold that there is an absolute prohibition against killing someone when one’s only motivation is
personal gain and when there are no (other) reasons to kill. What they deny is that there are
“proscriptions that admit the possibility of, and forbid, their own violation to good effect.”
The defensibility of this overall position needs to be investigated. Part of the problem is
that McNaughton and Rawling accept duties of special relationship even though such duties are a
kind of constraint. I have a duty to care for my children even if, by neglecting them, I would set
an example that would lead to an overall improvement in parents’ caring for their children. It is
not clear why duties of this kind survive the critique of constraints when the duty not to torture
does not. Moreover, intuitively, there is a constraint against torture. Intuitively, it would be
morally wrong (other things being equal) to torture one person even if this is the only way to
prevent two other people from being tortured.
McNaughton and Rawling hold that the Rossian principles are the most basic and
fundamental normative moral truths. Accordingly they reject a variety of attempts to derive or to
ground deontology. It may be possible, however, to provide deontology with a kind of extra-
moral grounding even if McNaughton and Rawling are correct that the Rossian rules are the most
fundamental moral truths. Such a grounding could perhaps embrace constraints. There is a view
according to which, roughly, a system of moral rules is justified or authoritative just in case its
currency in society would improve social cohesion and otherwise enable a society to meet its
needs (Copp, 1995). It might be argued that a deontological moral code that includes constraints
would be best suited to filling this role, and that this is sufficient to ground a deontology with
constraints. This kind of grounding is structurally similar to the grounding that might be offered
in rule consequentialism, but it is not consequentialist. Rule consequentialism depends on a view
about the value of states of affairs and uses this view in justifying the code of rules the currency
of which would have the best consequences. But the strategy at issue here does not depend on the
idea that social cohesion is valuable. It proposes an extra-moral grounding of morality rather than
a moral justification of a code of rules.12
The concept of a deontic constraint is closely related to the concept of a moral right. A
right of the kind at issue–a claim-right–entails a constraint on others, an obligation not to treat
the right-holder in a specified way. So understood, there is a right against being tortured only if
there is a constraint against torture. Some philosophers have aspired to build a rights-based moral
theory in which propositions about rights are taken to be basic (Mackie, 1978). But it is more
natural to see rights as one element in a pluralistic deontology.
One of the central goals in the chapter by Hillel Steiner is to show what would be lost in a
moral theory that failed to recognize claim-rights. Robert Nozick pointed out that it is possible to
treat the minimization of rights-violations as an end to be achieved in a “utilitarianism of rights,”
a kind of consequentialism that treats the minimization of rights-violation as the central moral
good. However, such a view does not treat rights as entailing the existence of constraints. So
even if it recognizes a “right against torture,” it does not recognize a claim-right against torture.
Consequently, Steiner wants to argue, it fails to establish a proper moral status for persons.
Nozick argued that claim-rights “reflect the underlying Kantian principle that individuals are
ends and not merely means” (Nozick, 1974, pp. 28-32). Steiner agrees.
One might think that Nozick’s Kantian approach answers the paradox of deontology, but
this is unclear. For it can seem paradoxical that Allan’s status as an end could preclude me from
treating him as a mere means even if my doing so is the only way I can prevent Bill and Carol
from being treated as mere means. The difficulty may only have been moved to a new level.
David Brink aims to defend a kind of perfectionist consequentialism in his chapter. He
agrees in broad terms with McNaughton and Rawling that there are no traditional deontological
constraints and that an adequate moral theory must give a plausible account of options and of
duties of special relationship. He thinks, however, that an agent-relative consequentialism can do
the trick. An agent-relative consequentialism can give special significance to the concerns and
projects of the agent, thereby making room for options, and it can give special significance to the
consequences of actions for those to whom the agent stands in special relationships, thereby
making room for the duties of special relationship. Brink sees the “self-referential altruism” that
was proposed by C.D. Broad as a model of such a theory (Broad, 1971).
Traditional consequentialism is agent-neutral. It takes the consequences of an action–or
of the currency of a rule, et cetera–for the good of anyone to matter, and to matter to the same
degree (to the rightness of the action), provided that the degree of good effect is the same. There
is also, however, an agent-relative form of consequentialism, called ethical egoism, according to
which the right action is the action that would have the best consequences for the agent. Brink’s
view is a close relative of ethical egoism if not a kind of egoism. Brink holds that the
consequences of an action that determine whether it is right are consequences for the agent’s
good. Good consequences matter only if and to the extent that they are good for the agent.
Brink holds, however, that the good of an agent is not restricted to goods that “fall
within” the agent’s life. If you and I are related in certain ways, he thinks, it can be intrinsically
good for me that a good falls in your life. For example, if you are a friend, an enjoyment
experienced by you might be intrinsically good for me as well as for you.
Here Brink takes up a line of argument about personal identity that is found in the work
of Derek Parfit. As time goes by, I pass through a variety of psychological states. Many of these
are continuous with other states or are connected to others in the way that a memory is connected
to the event of which it is a memory, or in the way that an early childhood plan to become a
firefighter can be continuous with one’s later career as a teacher by means of a chain of
decisions. Parfit proposed that a stream of psychological states over time constitutes a person just
when, roughly, the events in the stream have a characteristic kind and degree of connectedness
and continuity. He proposed to reduce personal identity to psychological connectedness and
continuity (Parfit, 1984). Brink suggests that, since you and I can have interlocking plans and
lives, there might be the same kind of psychological connectedness and continuity between our
psychologies as there is within each of our psychologies. The difference is perhaps only a matter
of degree. But if so, then perhaps the difference between distinct persons is no more significant
morally than the difference between distinct stages in the life of one person. If personal identity
boils down to psychological connectedness and continuity, then its moral significance boils down
to the significance of psychological connectedness and continuity.
This line of reasoning suggests that the moral value to me of a good received by someone
depends on the degree to which the person is psychologically connected to and continuous with
me. An agent-relative consequentialism can hold that consequences of an action that determine
its moral status are consequences for the good of those who are psychologically connected to and
continuous with the agent, and that the degree to which such consequences affect the action’s
moral status depends, other things being equal, on the degree to which those affected are
psychologically connected to and continuous with the agent.
The Parfitian theory of personal identity raises issues in metaphysics that lie outside the
realm of moral philosophy. But let me briefly raise a worry about Brink’s use of it. Suppose that
you and I are friends and both of us have headaches. On Brink’s view, your headache has the
same kind of significance to my good as my own headache has, even though, since your
headache is not as closely connected as my headache is to the psychological stream that is
identical to me, your headache has a lesser degree of significance for my good than my headache
does. This seems incorrect. Intuitively, my headache diminishes my good directly, by its very
nature, while if yours diminishes my good, it does so only instrumentally or indirectly, because I
care about you. On Brink’s view, if I have one dose of painkiller, then my duty with respect to
the use of it is determined by the effect it would have on each of the headaches weighted by the
degree to which the person whose headache would be helped is psychologically related to me. So
if our headaches are roughly equally bad and would benefit from the painkiller to roughly the
same degree, I would be wrong to give it to you. This also seems incorrect. It illustrates the
affinity between Brink’s view and ethical egoism, for a standard kind of ethical egoism would
have the same implication.
It may seem at this point that in order to make progress in the debate between deontology
and consequentialism, we need to seek to ground in some way the approach we take to normative
issues. It is time, then, to turn to Kantian moral theory. Kantian theory seeks to ground moral
judgments in a meta-ethical doctrine about the relation between morality and rationality.
Thomas E. Hill, Jr., explores the variety of ways in which Kant, and contemporary
philosophers who are applying and extending Kant’s views, deal with normative issues. Kant
holds that the fundamental principle of morality is the Categorical Imperative, but he offers
several different formulations of it. Hill examines problems in the interpretation and application
of each of these formulations. He begins with the formula of universal law: “Act only on that
maxim by which you can at the same time will that it should become a universal law.” (Kant,
2002, p. 222 [4: 402]).13 His fundamental worry about this formulation is that it does not seem to
explain what is wrong with wrongful actions, such as failing to help others. As he says, the
wrongness of slavery does not seem to be explained by pointing out that it is impossible for
everyone to act on the maxim of the slave-owner. Kant’s second formulation of the Categorical
Imperative is the so-called formula of humanity: “Act in such a way that you treat humanity,
whether in your own person or in any other person, always at the same time as an end, never
merely as a means” (Kant, 2002, pp. 229-30 [4: 429]). Alan Donagan (1977) has interpreted this
formula as requiring respect for persons. Hill thinks that the idea of respecting persons is too
vague to guide action. He suggests viewing Kant’s formulations of the Categorical Imperative as
different attempts to describe a point of view that can shape discussion and deliberation, rather
than as attempts to state a precise criterion of right action or a precise decision procedure. On this
basis, he proposes the idea of a Kantian legislative perspective, a perspective from which we can
deliberate about proposed moral rules. In this, Hill is building on Kant’s idea of a “kingdom of
ends” (Kant, 2002, p. 234 [4: 433]).
One might view Thomas Scanlon’s recent “contractualist” proposal as likewise proposing
a perspective for moral deliberation rather than a precise criterion or decision procedure. Moral
deliberation, says Scanlon, is fundamentally a matter of “thinking about what could be justified
to others on grounds that they, if appropriately motivated, could not reasonably reject” (Scanlon,
1998, p. 5). An action is wrong, he says, if it “would be disallowed by any principle” that “could
not reasonably be rejected, by people who were moved to find principles for the general
regulation of behavior that others, similarly motivated, could not reasonably reject” (Scanlon,
1998, p. 4). McNaughton and Rawling discuss Scanlon’s approach.
Moral theory has been dominated by the debate about right action that we have been
discussing, and many philosophers regret this. Virtue theory holds that the most fundamental
matter of moral concern is the character of a virtuous person. The ethics of care holds that the
most fundamental matter of moral concern is caring relationships. Both approaches aspire to turn
normative theory away from a preoccupation with right action, and toward an assessment of the
broader issues of how to live and what kind of person to be. It is not that these approaches hold
that issues about right action are unimportant. The idea is that they are secondary issues and that
they cannot properly be understood until we have an adequate theory of moral virtue or of caring.
Any complete moral theory would have to make room for the idea of virtuous, caring
agency. Nothing prevents an account of the virtues from being incorporated into a pluralistic
moral theory alongside an account of moral duty. It can also be incorporated into a
consequentialist framework.14 But some philosophers, inspired in many cases by their reading of
Aristotle’s moral philosophy, believe that a theory that is adequate to the subtle experience of a
mature moral agent must take moral character to be the basic moral concern. Virtue ethics, so
understood, is widely seen to have great promise, and in recent years, a number of new
approaches to virtue have appeared in the philosophical literature (see Copp and Sobel, 2004).
Julia Annas advocates an ambitious program of virtue ethics. In her chapter, she lays out
the structure of virtue theory as it was developed in what she calls the classical version of the
virtue ethics tradition. Such a theory was first articulated in a clear way by Aristotle, but Annas
holds that the basic features of Aristotelian virtue ethics are common to all ancient ethical theory.
Some contemporary versions of virtue ethics reject certain aspects of the classical theory and can
best be understood by comparison with the classical version of virtue theory. Annas holds,
however, that the classical version is the most attractive and defensible.
Theories in the classical tradition claim that moral virtue is necessary if one is to flourish.
Annas insists that this does not mean that these theories ground virtue in self-interest, for in the
classical view, flourishing is explained as consisting in part in being virtuous. A virtuous person
must be fair, kind, generous, and so on, and her virtues lead her to be wholehearted in doing
things for others. Virtue is a “disposition to do the right thing for the right reason in the
appropriate way–honestly, courageously, and so on.” It therefore involves acting both with an
appropriate affect–with sympathy, for example–and with an appropriate understanding of the
reasons for so acting.
In view of the latter point, one might think that virtue ethics cannot avoid problems in the
theory of right action. Annas explains, however, that in the classical tradition, ethical
understanding is viewed as involving the acquisition of something like a skill rather than learning
a criterion of right action. A virtuous person has the skill to determine the right way to act. The
rest of us may need to use principles and rules. But, as Annas explains it, virtue ethics denies that
there is a criterion of right action. In virtue theory, it is true that, roughly, the right action is the
action that a virtuous person would perform. But this is not intended to specify a right-making
property. It is not meant to serve as a principle, or a criterion, or a decision rule.
The ethics of care sees a disposition to care appropriately for others as the chief
characteristic of a morally desirable psychology. Such a disposition can be viewed as a virtue. In
her chapter, however, Virginia Held rejects the idea that the ethics of care is a kind of virtue
theory on the ground that its focus is on caring relations between people rather than on caring
dispositions. The ethics of care clearly is not a virtue theory in the classical tradition discussed by
Annas, for it rejects the idea that the proper exercise of practical reason is needed to enable one
to determine how to act. It holds that the moral emotions, such as empathy and sensitivity, guide
us to act properly. Beyond this, the ethics of care stresses the moral importance of meeting
people’s needs, especially the needs of people to whom we are related either intimately or in a
relation that brings special responsibility, such as the relation to an infant. Society includes
persons in various degrees of dependency. Caring is the glue that holds this together.
One could perhaps view the ethics of care as supplementing more traditional theories by
stressing the importance of the moral emotions and situations of dependency. Yet it is intended
as a new approach, on a par with deontology, consequentialism, and virtue theory. The ethics of
care developed out of philosophical reflection on the implications of feminist insights for moral
theory. Carol Gilligan’s work in moral psychology was highly influential. Gilligan (1982) found
that while boys tended to interpret certain stories as raising issues of justice, girls tended to
interpret the stories as raising issues of care. Some philosophers found this suggestive of a new
approach to ethics, and argued for the superiority of the perspective of care. Held cautions us,
however, that issues of justice arise within caring relationships, so that a complete theory cannot
Both virtue theory and the ethics of care deny that moral understanding depends on a
knowledge of principles of right action. These approaches to normative theory therefore tend to
be sympathetic to particularism, which is discussed in the chapter by Lance and Little.
In order to evaluate the various theories we have been discussing, philosophers construct
imaginary examples and then compare what the theories say about the examples with their
“moral intuitions.” I followed this strategy in objecting to some of the theories we have
discussed. Philosophers pursue a similar strategy in evaluating meta-ethical theories, for a meta-
ethical theory can be tested to see whether it conflicts with pre-theoretical beliefs about morality.
It might be objected that our moral intuitions may merely reflect our own parochial culture and
that our pre-theoretical intuitions may rest on naivete and inadequate thought. In his chapter,
Michael DePaul examines in detail the methodology of seeking a “wide-reflective equilibrium”
between theory and intuition. He argues that there is no sensible alternative since the method
basically consists in reflecting thoroughly and then trusting the conclusions we reach.
Moral philosophy can have an immediate significance for our lives that many other
abstract areas of philosophy do not have. Normative theories have implications for how we are to
live. And while meta-ethical theories may not have such implications, they can have implications
for how we are to understand the implications of normative theory, so they can affect our
understanding of claims about how we are to live. It is appropriate, therefore, to inquire into the
relation between the theories we have examined and moral practice.
This is the topic of Gerald Dworkin’s chapter. Dworkin argues that we need to make use
of moral principles in order to satisfy a normative requirement on responsible moral inquiry and
discourse–the requirement of “consistency,” or systematic coherence. This is the requirement to
conduct moral inquiry and discourse in such a way that our decisions about how to live are not
“arbitrary,” but are “principled”, in a familiar intuitive sense. He therefore argues, by implication,
that an adequate moral theory must articulate and defend moral principles.
This has been an introduction to moral theory wrapped around an introduction to the
chapters in this volume. The volume will have served us well if it helps to raise the level of
debate in moral philosophy and to foster a heightened level of responsiveness and reasonableness
in moral discourse.
Introduction: Meta-Ethics and Normative Ethics
* Several people provided helpful comments on earlier drafts of this essay and I am
grateful to all of them. I am especially grateful to Daniel Boisvert, Jamie Dreier, Tom
Hurka, Kirk Ludwig, David McNaughton, Marina Oshana, Piers Rawling, Geoff Sayre-
McCord, Jon Tresan, Anton Tupa, and Crystal Thorpe.
1. Theists often hold that it is a necessary truth that God exists. On this view, the
conditional that if God does not exist, there are no obligations, has a necessarily false
antecedent. There is controversy about the evaluation of conditionals with necessarily
false antecedents, but a discussion of the controversy would be beyond the scope of this
chapter. It seems to me that an adequate account would treat the above conditional as
following from the divine command theory, for its consequent follows from the
conjunction of its antecedent with the theory. That is, there is a valid argument from the
conjunction of the proposition that God does not exist and the divine command theory to
the proposition that there are no obligations. (I am grateful to Kirk Ludwig for helpful
discussion of this issue.) If one takes the first horn of the Euthyphro dilemma instead of
the horn chosen by Quinn, one can avoid this difficulty. For on this view, God commands
that we do our duty, but our duties are obligatory independently of His commands.
Hence, God’s non-existence does not, or would not, mean we have no obligations. But
this view is incompatible with the divine command theory.
2. Moral realism is compatible with any theory about the nature of properties, including
nominalism. See below, note 3. In what follows, I treat relations, such as the relation of
being morally better than, as a kind of property. On some theories, rightness and
wrongness themselves are best understood as relations. See Copp, 1995, pp. 218-223. See
the chapter by James Dreier.
3. That is, the first realist doctrine is to be interpreted such that the term “property,” as it
occurs there, ascribes the same metaphysical status to moral properties, such as
wrongness, as it ascribes to a non-moral property such as redness when it is predicated of
such a property. A moral realist can be a nominalist, for although she says there are moral
properties, she says they have the metaphysical status that any other property has,
whatever that is. Some philosophers would deny that there are any properties at all. But I
take it that they do not mean to deny that red things have the “characteristic” of being
red. They mean to reject the standard philosophical theories about the nature of such
characteristics. They would agree that sentences such as, “There are properties such as
redness,” can be used to express truths, but they reject standard philosophical theories of
their truth conditions. If so, they may be in a position to accept moral realism.
4. A moral claim is “basic” in the sense at issue just in case it is (or could be expressed in
English by a sentence) of the form, ‘A is M,’ where ‘M’ is replaced by a moral predicate
and ‘A’ is replaced by a term that refers to or picks out a person, action or action-type,
character-trait, social institution, or the like. A moral realist would say that a basic moral
claim ascribes a moral property. An example of a basic claim is the claim that capital
punishment is wrong. The proposition that nothing is morally wrong is not basic.
5. For an example of a cognitivist functionalism that is roughly of this kind, see Foot, 2001.
For a critique, see FitzPatrick, 2000.
6. For the idea of a “response-dependent” property, see Wedgwood, 1998.
7. A similar view is proposed in Oshana, 1997. Various other conceptions of moral
responsibility are discussed by John Fischer in his chapter.
8. My thinking about the two central disputes, and especially about the idea of a basic
matter of moral concern, has been influenced by Shelly Kagan’s discussion of
foundational normative theories, and especially by his idea that consequentialist theories
can have different “evaluative focal points.” See Kagan, 1998, pp. 202-204. I have
benefitted from the helpful comments of Daniel Boisvert, David McNaughton, Piers
Rawling, and Jon Tresan.
9. To be sure, Kantian theory takes rational agency to be valuable, and rights-based theories
take rights to be valuable. But they take judgments about rational agency, or about rights,
respectively, to be basic or fundamental, not judgments about value. They do not qualify
as consequentialist merely because they would agree that what, for them, is the basic
matter of moral concern is also valuable.
10. According to act consequentialism, as I formulated it, the right action is the available
action that would maximize the good. One might instead think that any alternative is
permitted provide it is above a threshold. Brink discusses a variety of possible views.
11. This is a crude presentation of an argument that first appeared in Nozick, 1974, pp. 29-
31, and that was then elaborated in detail in Scheffler, 1982, chapter 4.
12. A position of this kind is developed in Copp, 1995. The idea is that a moral code that is
“justified” thereby has a truth-grounding status, a status such that relevantly
corresponding moral propositions are true. For example, if a justified code includes a
constraint against torture, then it is true that torture is wrong. Braybrooke (2003) argues
that the position proposed in Copp, 1995, falls within the natural law tradition, broadly
conceived. As he puts it, “Natural law theory founds moral judgments on what, given the
nature of human beings and ever-present circumstances, enables people to live together
in thriving communities” (p. 125). See Copp, 1995, pp. 201-209.
13. Numbers in brackets refer to the volume and page number in the standard Prussian
14. This can be done in different ways, as illustrated in Hurka, 2000 and Driver, 2001.