Philosophy reading homework questions





The article is difficult to read,so play close attention to the questions on the homework and focus on the readings.

MLA format with size 12 font Time roman single spaced

You can skip sections 2 and 6. All answers must be in your own words.

1. McGrath is not arguing that we do not, or can not, know


moral truths, just that we do not know


of them.

a. What sorts of moral truths does she think we cannot know? (1-2 sentences)

b. Give an example (not from the reading). (1-3 sentences)

2. Why, according to McGrath, are we unable to recognize moral experts? (1-3 sentences)

3. How is McGrath’s argument relevant to Raz’s view? (1-3 sentences)

4. Assume McGrath is correct about the answers to 1a and 2; given this, in your opinion, what should judges do when they believe the law is substantially morally mistaken (when they think a particular law is fundamentally unjust)?

5. Give the best argument you can that, in some cases, we can either recognize that someone is a moral expert, or at least recognize that someone is NOT a moral expert.

Key words


: Key terms: “authority,” “content independent reason.” Be able to explain what each means. Be able to give plausible examples of authorities outside the context of the law (you don’t have to agree that these are authorities, you just need to have some examples that others will find plausible). If I give examples, be able to explain whether or not these terms apply to them.

First and second order reasons:

Key terms: “first order reason,” “second order reason,” “exclusionary reason” (this last one is a concept from Raz and is explained in the Hurd reading). Be able to explain what each of these are. Be able to give plausible examples of each (you may not believe that, e.g., exclusionary reasons exist, but you should be able to give an example that would be plausible to someone who does believe in these types of reasons). Be able to give examples in which a particular fact is (or gives) both first and second order reasons. If I give you examples, be able to explain whether they are first or second order reasons (or both or neither).

Causal, explanatory, and justificatory reasons

: Key terms: “causal reason,” “explanatory reason,” “justificatory reason.” Be able to give an example of causal reasons and explanatory reasons that are not moral justificatory reasons. Be able to give examples of moral justificatory reasons that do not cause any event or explain any event. Be able to give examples of moral reasons that also do cause or explain events.

Moral Disagreement and Moral Expertise
Sarah McGrath
The phenomenon of persistent ethical disagreement is often cited in
connection with the question of whether there is any ‘‘absolute’’ morality,
or whether, instead, morality is in some sense merely ‘‘a matter of personal
opinion’’. Citing disagreement, many people who hold strong views about
controversial issues such as the permissibility of abortion, eating meat, or
the death penalty deny that these views are anything more than ‘‘personal
beliefs’’. But while there might be inconsistencies lurking in this position,
it is not obviously at fault for according the facts about disagreement some
epistemic weight.
This paper addresses the question of whether and to what extent moral
disagreement undermines moral knowledge. The most familiar arguments
from disagreement in the literature purport to establish conclusions about
the metaphysics of morality: that there are no moral facts, or that there are
no moral properties, or that the moral facts are relative rather than absolute.
Of course, the conclusions of some such metaphysical arguments might be
perfectly consistent with the existence of considerable moral knowledge. For
example, even if there is some successful argument from disagreement to
the conclusion that moral facts are relative rather than absolute, this might
very well be consistent with our having just as much moral knowledge as we
Earlier versions of this paper were presented at the 2005 Dubrovnik Metaphysics and
Epistemology Conference and at the Second Annual Metaethics Workshop at the
University of Wisconsin, Madison. I am grateful to the members of those audiences for
their questions and feedback. Thanks also to Frank Arntzenius, Alex Byrne, Elisabeth
Camp, John Collins, Adam Elga, Delia Graff Fara, Elizabeth Harman, Tom Kelly,
Sarah-Jane Leslie, Ted Sider, and anonymous referees for Oxford Studies in Metaethics
for helpful comments on earlier drafts.
Sarah McGrath
ordinarily take ourselves to have. (Although of course, such an argument
might alter our conception of what it is that we know.) On the other hand,
a metaphysical argument from disagreement which successfully showed that
there are no moral facts would presumably rule out the possibility of moral
By contrast, epistemological arguments from disagreement purport to
undermine moral knowledge by showing that, regardless of the metaphysics
of the moral facts, we are not in a position to have anything like the
amount of moral knowledge that we ordinarily take ourselves to have.
For reasons that I explore below, there are various respects in which
epistemological arguments from disagreement present a more formidable
skeptical challenge than metaphysical ones. My main goal in this paper
is to develop an epistemological argument that creates a difficulty for our
controversial moral beliefs and to explore the extent to which it succeeds.
As a representative metaphysical argument, consider J. L. Mackie’s wellknown ‘‘argument from relativity’’ (1977: 36–8). According to Mackie,
‘‘radical differences between first order moral judgments’’ provide a compelling reason to doubt ‘‘the objectivity of values’’. While it is not entirely
clear what Mackie means when he denies the objectivity of values, he does
seem to mean, minimally, that all claims to the effect that something has a
certain moral property are false. If Mackie is right about this, then we have
very little moral knowledge—far less than we thought we had. Perhaps one
could know that nothing is morally wrong, but one could not know of any
particular action that it is morally wrong—for all claims to the effect that
a particular action is right or wrong are false. ( Just as, having learned that
there aren’t any witches, one could know that Marilyn Manson is not a
witch. What one can’t know is that anybody is a witch.)
Significantly, Mackie does not think that scientific disagreement supports
an analogous conclusion about science. He argues that the skeptical inference
is compelling in the moral but not in the scientific case because moral
and scientific disagreements have different explanations. While scientific
disagreement is best explained by the fact that scientists draw different
conclusions from inadequate evidence, disagreement about moral codes is
better explained by ‘‘people’s adherence to and participation in different
ways of life’’ (p. 36). In the moral case, ‘‘the causal connection seems
to be mainly that way round: it is that people approve of monogamy
because they participate in a monogamous way of life rather than that
they participate in a monogamous way of life because they approve of
Moral Disagreement and Moral Expertise
monogamy’’.¹ The hypothesis that moral codes are mere reflections of
ways of life better explains the pattern of moral variation than does the
hypothesis that different people have different ‘‘seriously inadequate and
badly distorted’’ perceptions of objective values (p. 37). Thus, there are no
objective values.
One immediate concern about Mackie’s argument is that it seems to
prove too much: it is not true that, in general, where differences in belief covary with differences in ways of life, we ought to draw similar conclusions.
For example, within the United States, beliefs about evolutionary theory
seem to satisfy the relevant criteria. According to a Harris Poll conducted
in the summer of 2005, only one-fifth of Americans believe that human
beings evolved from other species; only half think that other plants or
animals did; 64 percent believe that ‘‘human beings were created directly by
God’’. The poll shows that variation in these beliefs reflects differences in
the ways of life of the individuals who hold them, in the sense of reflecting
the religious, political, and cultural features of the communities to which
they belong: individuals who embrace creationism are more likely to be
from the South, to be Republicans, to be religious, and to lack college
educations. By contrast, Democrats, those from the Northeast and West,
and those with college educations are more likely to believe in evolutionary
theory. But while it does seem that people’s beliefs about evolutionary
theory reflect their ways of life in this sense, this does not support any
surprising metaphysical conclusions about the facts at issue. In particular, it
does not support the conclusion that there are no truths about the origins of
the human species, or that all claims about human origins are false. Perhaps
Mackie is correct in holding that the pattern of difference in moral beliefs
corresponds to a pattern of difference in the cultural norms prevailing in
the communities in which individuals were raised. But even if that is true,
it does not show that there are no moral facts.
Of course, more could be said on behalf of Mackie’s argument. In
particular, one might argue that moral controversy and the controversy
about human origins are disanalogous in ways that ultimately prove crucial.
I will not explore arguments to that effect here, since I do not claim that
the present difficulty is decisive. My purpose in raising this prima facie
¹ This quote leaves out some details: on Mackie’s view, people’s moral beliefs typically
reflect ‘‘idealizations’’ of the ways that they actually live rather than simply reflecting
those ways of living. So, for example, ‘‘the monogamy in which people participate may
be less complete, less rigid, than that of which it leads them to approve’’ (p. 36). And
some revisions in moral beliefs are explained not by changing idealizations but by the fact
that people aim for consistency: thus, someone might change her belief about whether
same-sex marriage is wrong because it conflicts with her other beliefs about what features
of a relationship are relevant to whether people ought to marry.
Sarah McGrath
difficulty for Mackie’s argument is to highlight a quite general challenge
for those who would have us draw conclusions about the metaphysics
of morality from the existence of moral disagreement: such arguments
naturally invite the charge that they prove too much. In order to successfully
respond to this charge, proponents of such arguments must explain why we
should not draw the same surprising metaphysical conclusions wherever we
find apparently similar phenomena. Why, for example, doesn’t widespread
religious disagreement show that there is no fact of the matter about whether
any gods exist, or that such facts are relative? Notoriously, it is difficult
to explain why moral disagreement cries out for the metaethicist’s favored
metaphysical conclusion while similar disagreement in other domains does
not. ( Just as we would not want to conclude that all beliefs about human
origins are false, so also we would not want to conclude that such beliefs
are all relative, or by nature are knowable only by some special faculty of
intuition.) Of course, this is not to say that the relevant explanation cannot
be provided; only that the task of providing it cannot be avoided, and is far
from trivial.
A second potential vulnerability for metaphysical arguments from disagreement is that in general such arguments have the form of inference to
the best explanation arguments, according to which the best explanation
of the kind of disagreement that we find in the moral domain is the
preferred conclusion of the proponent of the argument: that there are no
objective values, or, alternatively, that moral facts are relative facts, or that
what look like moral claims are really just expressions of emotion, and so
on. Because metaphysical arguments are inference to the best explanation
arguments, one who offers such an argument must show that her favored
conclusion better explains the data than any alternative hypothesis does.
One competing hypothesis is the perfectly mundane one that the questions
with respect to which we disagree are difficult ones, and at least some of
us are getting them wrong; the others include the wide range of surprising
candidate metaphysical hypotheses familiar from the metaethics literature.
Again, there is no guarantee that such a case cannot be made on behalf
of some preferred explanation. The point is just that it is not enough to
point to a hypothesis that would adequately explain the relevant features
of moral disagreement if it were true: one must show that the hypothesis
better explains those features than would any competing hypothesis if it
were true.
Thus, any metaphysical argument for the skeptical conclusion that we
have little or no moral knowledge immediately inherits two potential
vulnerabilities. First, to the extent that parallel reasoning applied to other
domains would lead to conclusions that we are unwilling to accept,
it is potentially vulnerable to the charge that it proves too much or
Moral Disagreement and Moral Expertise
overgeneralizes. Second, inasmuch as such an argument is an inference
to the best explanation argument, it is vulnerable to the provision of
formidable competing explanations of moral disagreement. In the next
section, I consider a line of epistemological argument which possesses
neither of these vulnerabilities.
Consider the following passage from Henry Sidgwick’s The Methods of
[I]f I find any of my judgments, intuitive or inferential, in direct conflict with
a judgment of some other mind, there must be error somewhere: and if I have
no more reason to suspect error in the other mind than in my own, reflective
comparison between the two judgments necessarily reduces me temporarily to a
state of neutrality. (p. 342)
Moreover, according to Sidgwick, ‘‘the absence of such disagreement
must remain an indispensable negative condition of the certainty of our
beliefs’’ (p. 342).
Let us call a belief CONTROVERSIAL just in case it satisfies the
condition to which Sidgwick draws our attention. Thus your belief that p
is  if and only if it is denied by another person of whom
it is true that: you have no more reason to think that he or she is in error
than you are. Of course, a belief might be controversial without being
. This is the case, for example, when some view that you
hold is disputed, but you have reason to think that those who dispute it are
more likely to be in error than you are.
As we have noted, Sidgwick holds that no belief that is 
can be certain. But a parallel claim about knowledge also seems attractive.
That is, it seems plausible that
If one’s belief that p is , then one does not know that p.
Suppose that you and your friend Alice intend to take the train together
but discover that you have different views about what time it is scheduled
to depart: you think that the train departs at a quarter past the hour, while
she thinks that it departs at half past. Perhaps you have some good reason to
think that Alice is the one who has made a mistake. For example, perhaps
you know that she arrived at her view by consulting a train schedule that is
out of date, while you arrived at yours by consulting the current schedule.
Or perhaps you know that Alice is prone to carelessness with respect to
such matters, as she has a past history of having made similar mistakes.
Sarah McGrath
But suppose instead that you have no such reason to think that it is Alice
who has made the mistake: as far you know, it is just as likely that you are
mistaken as that she is. In that case, it seems that your belief about what
time the train leaves does not amount to knowledge.²
Of course, it’s clear enough that your belief does not amount to knowledge
if you are in fact the one in error, i.e., if your belief about what time the
train leaves is false. But even if your belief is true, and Alice is the one
who has misread the schedule, it seems that your belief does not amount
to knowledge provided that you have no good reason to think that she is
the one who has made the mistake. Even if your belief would amount to
knowledge in the absence of Alice’s holding a contrary belief, the fact that
she believes as she does can preclude your knowing in the circumstances. For
plausibly, this would be a case in which misleading evidence undermines
This suggests the following epistemological argument for a certain kind
of moral skepticism:
P1 Our controversial moral beliefs are .
P2  beliefs do not amount to knowledge.
C Therefore, our controversial moral beliefs do not amount to knowledge.
The first premise and the conclusion of the argument refer to ‘‘our
controversial moral beliefs’’. By this, I mean our beliefs about the correct
answers to the kinds of questions that tend to be hotly contested in the
applied ethics literature as well as in the broader culture: questions about
² Cases broadly similar to this one have recently been discussed in the epistemology
literature devoted to the question of how we should respond to ‘‘peer disagreement’’.
This literature has not directly addressed the question of how disagreement affects
knowledge, which is our primary concern here. Significantly, however, a number of
contributors to this literature (notably Feldman 2006, Christensen 2007, and Elga 2007)
either endorse or express considerable sympathy for the view that peer disagreement
should lead the peers to suspend judgment about the disputed question. Presumably,
if one ought to suspend judgment as to whether p, then one does not know that
P. Kelly (forthcoming) explicitly argues against the view that one is rationally required to
suspend judgment in the face of peer disagreement but holds that one should nonetheless
become less confident of one’s original opinion, and that, all else being equal, as the
number of peers on both sides of the issue increases, the push towards agnosticism grows
³ Does the fact that your true belief is denied by the relevant kind of person suffice to
undermine its status as knowledge, or must one also be aware that it is denied by such
a person? This is a special case of a substantive and disputed question in epistemology,
the question of whether (or in what circumstances) the existence of misleading evidence
undermines knowledge when it is not possessed by the would be knower. For discussion,
see Harman 1973, Lycan 1977, and Ginet 1980. In what follows, I sidestep this issue by
focusing on cases in which one is aware of the disagreement.
Moral Disagreement and Moral Expertise
the circumstances (if any) in which it is morally permissible to administer
the death penalty, or to have an abortion, or to eat meat, or about how
much money we are morally obligated to donate to those in dire need,
and so on. It is clear that our beliefs about the answers to such questions
are controversial ones. It is of course much less clear that they are also
, i.e., that P1 is true. A good part of what follows is devoted
to scrutinizing this claim. I begin, however, with a few preliminary remarks
about the argument.
First, one who endorses the argument might remain studiously agnostic
about the metaphysics of morality, and in particular, about whether there
are any moral facts. That is, one who endorses the argument need not take a
stand on whether such facts exist, or even on what, if anything, the relevant
kind of disagreement suggests about their existence. The contention of one
who endorses the argument is rather that the kind of disagreement that we
find with respect to controversial moral questions precludes our knowing
the correct answers to these questions, regardless of whether such questions
have correct answers.
Second, the conclusion of the argument is that our beliefs about controversial moral matters do not amount to knowledge. The conclusion is
not that it is unreasonable to hold those beliefs in the face of disagreement, or that we are rationally required to suspend judgment with respect
to controversial moral matters. However, if the argument is successful,
then the skeptic would seem to have made significant headway towards
establishing these apparently stronger claims. For it has been argued, with
considerable plausibility, that if one is not in a position to know whether
p, then the reasonable course is to suspend judgment about whether p
until further evidence becomes available; that is, one should not believe
when one is in no position to know ⁴. Thus, if the above argument is
sound, then this would at the very least seem to put considerable pressure on the idea that it is rational for us to maintain our controversial
moral views.
Third, in the previous section, we noted that metaphysical arguments
from disagreement generally take the form of inference to the best explanation arguments, and that this fact presents a potential line of resistance
to such arguments. Notice that the epistemological argument presented
here is not best reconstructed as an inference to the best explanation argument. The suggestion is not that the best explanation of the
disagreement is that no one knows; rather, the suggestion is that the circumstances of the disagreement are inconsistent with one’s knowing. Thus,
⁴ This conclusion will be especially attractive to those who take knowledge to be the
aim of belief; for defense of this claim, see esp. Williamson 2000.
Sarah McGrath
the argument does not share at least one of the two potential vulnerabilities
characteristic of metaphysical arguments.
It is less clear, however, that the argument avoids the second potential
vulnerability of metaphysical arguments: that of susceptibility to the charge
of overgeneralizing, or proving too much. This issue is the focus of the next
In Section 2, we noted that metaphysical arguments run the risk of
overgeneralizing. On the face of it, Mackie’s argument that there are no
objective values would, if it succeeded in showing claims about value
to be false, show the same for claims about human origins. Since we can
confidently assume that it would be a mistake to draw that conclusion about
human origins, it seems that we can conclude that Mackie’s argument falls
short of showing that there are no objective values. Does the epistemological
argument from disagreement similarly overgeneralize?
One might suspect that the epistemological argument does prove too
much. Thus, in responding to a similar line of argument to the one under
consideration here, Russ Shafer-Landau poses the following dilemma:
Either intractable disagreement among consistent intelligent parties forces them to
suspend judgment about their contested views, or it doesn’t. If it does, then we must
suspend judgment about all of our philosophical views, as well as our belief that
there is an external world, that I am an embodied being, that the earth is older than
a second, etc. All of these have been challenged by brilliant, consistent, informed
skeptics over the millennia.
Alternatively, if we are warranted in any of our beliefs, despite the presence
of such skepticism, then justified belief is possible, even in the face of persistent
disagreement. And so we could retain our moral beliefs, especially those we have
carefully thought through, despite an inability to convince all of our intelligent
opponents. (2004: 108–9)
Here, the suggestion is that our controversial moral beliefs are in the same
epistemic boat as our beliefs that there is an external world and that the earth
has existed for more than one second. If this were the case, then we could
safely conclude that the argument from disagreement does prove too much,
since (I assume) we do know that there is an external world and that the
earth has existed for more than one second.
However, the idea that beliefs of this kind and our controversial moral
beliefs are equally jeopardized by disagreement seems dubious. After all, my
belief that the earth is older than one second faces much less opposition than
my belief that the death penalty is morally impermissible. Even if it is true that
Moral Disagreement and Moral Expertise
brilliant skeptics have disputed the former⁵, they are vastly outnumbered
by reasonable people who disagree. By contrast, with respect to, say, the
moral permissibility or impermissibilty of the death penalty, the division of
opinion is not that of lone geniuses vastly outnumbered by the opposition.
As Shafer-Landau interprets the skeptical challenge, it is the absence of
unanimity among the relevant class of people which suffices to generate the
skeptical conclusion. Even the existence of a single formidable dissenter who
cannot be won over would suffice to undermine whatever justification one’s
belief originally enjoyed. This interpretation allows him to plausibly suggest
that such a requirement, if consistently applied, would yield a sweeping and
global skepticism. However, this is not the most charitable interpretation
of the skeptical challenge. On a more charitable construal of that challenge,
it is the fact that there is a substantial division of opinion with respect to
controversial moral questions that undermines the possibility of knowing
the answers to those questions.
In short, the beliefs that the earth is older than one second and that there
is an external world are not . Even if these beliefs have on
occasion been denied by some, including some of formidable intelligence
(etc.), it does not follow that one has no more reason to suspect error in such
minds than in one’s own. Plausibly, one does have such reasons, reasons
provided by facts about the distribution of opinion among the relevant class
of people. If you and Alice have conflicting beliefs about what time the
train is scheduled to depart, then it might be that both of your beliefs are
. However, if you and Alice subsequently discover that ten
other people have independently arrived at your belief while none shares
hers, your belief is no longer rendered  by the fact that Alice
denies it. For now you do have reason to think that she is the one who has
made the mistake. On the other hand, her belief—supposing she maintains
it—is : she lacks any parallel reason.
Of course, it is no objection to the skeptical challenge under consideration
that it fails to single out our controversial moral beliefs. Parallel arguments
might be constructed to show that one lacks knowledge with respect to a
significant number of topics—for example, philosophy, public policy, and
⁵ One might quibble with Shafer-Landau’s choice of examples. Skeptics about the
external world or the past are best understood, I think, not as disputing first order
propositions about the world such as there is an external world or the world is older than
one second, but rather epistemic propositions such as we know that there is an external
world or we know that the world is older than one second. It is clear that some have
held views inconsistent with the latter propositions; it is less clear that anyone has held
views inconsistent with the former. Still, if it could be shown that our controversial
moral beliefs are no worse off than the relevant epistemic beliefs, I would take this to
constitute an adequate vindication of the ‘companions in guilt’ strategy for resisting
disagreement-inspired moral skepticism.
Sarah McGrath
religion.⁶ But this does not show that the argument over-generalizes. For
it is far from clear that the answers to much disputed questions in such
domains are known; in any case, that some of us have such knowledge
is not a datum to which one might appeal in attempting to discredit the
The previous section defended the argument against the charge of overgeneralization. This section addresses the question of why one should think
that one’s beliefs about disputed moral questions are  in
the first place. We have emphasized that, even if a belief is controversial, it
might not be . That is, even if the truth of a given belief
is contested, a person who holds that belief might have good reason to
think that anyone who thinks otherwise is more likely to be wrong than
she herself is. Indeed, some beliefs might be extremely controversial without
being . Consider again our earlier example of evolutionary
theory. The proposition that human beings evolved from other species is
vigorously denied by many, but it would be a mistake to conclude that it is
therefore not known by any of those who believe it. Indeed, the fact that it
is denied by many does not even preclude its being known by some who are
relatively unfamiliar with the scientific evidence in its favor. Crucially, the
proposition in question is not controversial among those who are known to
possess the relevant expertise. Certain scientific questions might be highly
controversial among the population as a whole, but when a consensus or
near consensus exists among those with the relevant expertise, one need
remain in a state of agnosticism only for as long as it takes to discover the
content of that consensus. Thus, despite the large number of people who
deny that human beings have evolved from other species, awareness of the
expert consensus on the opposite side of the issue provides good reason to
think that those who deny it are in error.
It might be thought that there is a parallel defense of one’s controversial moral beliefs. That is, it might seem plausible that, although many
dispute these beliefs, they are not , because they are not
controversial among ‘‘the moral experts’’. This raises two questions. First,
are there genuine moral experts? And second, if there are, how can they be
recognized—either by themselves or by others—as such? Let us set aside
the first question, and concede for the sake of argument that individuals
with genuine moral expertise exist. How might they be identified?
⁶ Cf. van Inwagen 1996.
Moral Disagreement and Moral Expertise
The task of identifying those with genuine expertise will be a much
less straightforward matter in some domains than in others. For the most
part, the epistemology literature devoted to the topic of disagreement has
focused on the idealized case, in which facts about relative expertise and
‘‘epistemic peerhood’’ are treated as given; the question that has dominated
that literature concerns how we should respond to disagreement with our
epistemic peers or equals.⁷ But in actual, real-life cases, others do not
typically wear their relative levels of competence on their sleeves. Of course,
on occasion they do: most of us have good reason to think that the person
whose shirt reads ‘‘Expert Plumbers’’ is someone to whose judgment we
should defer with respect to whatever plumbing questions might arise. But
in other cases, facts about relative levels of expertise and competence are far
from transparent.
In general, identifying those with genuine expertise in some domain will
be most straightforward when we have some kind of independent check, one
not itself subject to significant controversy, by which we can tell who is
(and who is not) getting things right. In certain domains, it is relatively easy
for us to acquire evidence which bears straightforwardly on questions about
relative expertise. Consider, for example, weather forecasting. Two weather
forecasters might offer what seem to be equally compelling cases for their
conflicting predictions about what tomorrow’s weather will be like. But
once tomorrow’s weather rolls in, we will have an answer to the question
of which of today’s two conflicting predictions was more accurate. Thus, in
the weather forecasting case, inductive track record evidence about who is
more reliable is relatively easy to acquire. Moreover, crucially, such evidence
can be readily assessed and assimilated by the layperson: one need not be
an expert weather forecaster in order to reliably identify those who possess
genuine expertise with respect to weather forecasting.
But significantly, we possess no similar independent check for moral
expertise. If moral expertise stands to morality as weather forecasting
expertise stands to weather, then a moral expert would be someone who
consistently arrives at the correct answers to non-trivial moral questions (or
at least, someone whose reliability with respect to such questions significantly
exceeds that possessed by the average person, when the average person does
not form his moral opinions by deferring to a moral expert). Given
such a straightforward understanding of moral expertise, there is nothing
particularly problematic about the idea that some individuals possess such
expertise. The difficulty lies in arriving at compelling grounds for attributing
such expertise, either to oneself or to others. A natural suggestion is that
the possession of certain academic credentials, or professional concern with
⁷ See esp. the work referred to in n. 2 above.
Sarah McGrath
ethics, is good evidence that one possesses reliable moral judgment. I am
acquainted with the ethics literature in a way that my plumber is not;
moreover, I have taught ethics classes to college students and attended
conferences devoted to the subject. He can claim no similar experiences.
But in the absence of an independent check on my relative ability to
therefore get the answers right, such facts would seem to constitute a
relatively meager basis on which to conclude that I am his superior with
respect to the reliability of my moral judgment. Again, contrast a case in
which we know that one of two weather forecasters is more reliable than the
other on the basis of his superior past track record. It would be a mistake, I
think, to suppose that in these circumstances I have anything like the kind
of evidence for the superiority of my moral judgment that is available in the
weather-forecasting case.
Simply put, there is no obvious way to locate oneself in the space of
moral expertise relative to others. It is true that professional philosophers
who work in applied ethics have thought about the arguments longer
than the average person has. Here, as elsewhere in philosophy, this has
not resulted in a convergence of opinion. Yet even if these professionals
were to converge on the view that, say, killing is no worse than letting
die, on the grounds that no adequate metaphysical basis for imputing
moral significance to this distinction could be found, it is not clear that
ordinary people of the opposite conviction need treat this as conclusive.
For it is less clear in the moral case than in various other cases that
reliable judgment with respect to the relevant domain is the typical upshot
of formal training. Here again the lack of an independent check seems
crucial. If a moral expert is someone who tends to get the hard questions
right, then good moral training is presumably whatever confers the relevant
capacity. That studying structural engineering at MIT is good training
for solving the kinds of problem that confront structural engineers can
be more or less readily checked by, for example, examining the stability
of bridges built by MIT-trained engineers. But in the moral case, since
it is unclear how to check who is getting things right, it is unclear how
to check whether MIT is a good place for moral training. Thus, while
one might think that good moral training would consist in taking a
series of ethics courses devoted to the critical examination of arguments
on both sides of divisive issues, an equally plausible answer might be
that good moral training consists in being raised by virtuous people who
devote relatively little time to scrutinizing arguments for and against their
views. Similarly, one might think that the best training for appreciating
the permissibility or impermissibility of causing animal suffering would
involve, among other things, witnessing such suffering. But we could just
as easily imagine that the judgment of those best acquainted with the
Moral Disagreement and Moral Expertise
slaughterhouse tends to become artificially deadened to the thought that
animals matter.
If the population is substantially divided about, say, the moral permissibility of abortion in certain circumstances, then, assuming that there is
some non-relative fact of the matter, a large number of us are wrong.
Unfortunately, we possess no analogue to an eye exam, by which we might
determine whose moral vision is askew and whose is in good working order.
Thus, the truth about where one stands in the space of moral expertise
might prove elusive, even for intelligent, thoughtful people.
The upshot of these considerations is that it is quite unclear how one
might argue, in a way that is not transparently question-begging or circular,
that one’s controversial moral beliefs are uncontroversial among the moral
experts. But if, for all one knows, there is no consensus among the moral
experts in favor of one’s controversial moral beliefs, then one cannot appeal
to the existence of such a consensus in order to show that those beliefs are
not .
In some ways, moral disagreement seems to parallel the diversity of opinion
as to which shade of green is unique green. Unique green is that shade of
green that is neither bluish nor yellowish. When asked to select the shade
which is unique green, different subjects with normal color vision will select
different shades.⁸ As in the case of our controversial moral views, opinion
about which shade is unique green not only fails to be unanimous, but is
substantially divided. Perhaps if there were relatively widespread agreement
as to which shade is unique green, then the dissenting judgments of a few
who possessed otherwise normal color vision could be dismissed. But the
fact that the actual division of opinion is substantial suggests that human
beings are not reliable detectors of the relevant property. That relevantly
similar creatures—creatures with the same type of visual system—arrive at
different verdicts when similarly situated seems to show that that kind of
creature is simply not well equipped to detect the presence or absence of
the property in question. That human beings are not, as a species, reliable
detectors of unique green seems to tell against crediting any individual with
⁸ See Hardin 1988 for detailed information about the phenomenon. Hardin himself
draws a conclusion analogous to Mackie’s, viz. that nothing is colored. Cohen 2003
draws a relativist conclusion similar to Harman’s: that objects have colors only relative
to perceivers and circumstances of viewing. Byrne and Hilbert (2003) conclude that we
should suspend judgment about which particular things are unique green.
Sarah McGrath
knowledge that a certain shade is unique green, particularly if the individual
knows of this general lack of reliability and has no good reason to think
that he is exceptional in this respect.
Note that although questions about which shade of green is unique
green are hard questions for human beings, such questions do not present
themselves to us as difficult ones. In fact, most subjects are quite confident
of their initial judgments; each person’s view strikes her as obviously correct.
This seems parallel to the moral case: in the moral case too, many find that
their own views about controversial moral questions strike them as obviously
correct. Cases in which we are quite confident in our original judgments,
only to discover there is a substantial division of opinion among those
with relevantly similar cognitive capacities, highlight the fact that there is
more than one way to discover the relative difficulty of a given intellectual
task. While one might learn that a given problem is hard by attempting to
think it through and finding oneself struggling or unable to come up with
an answer, one might, alternatively, learn that a given problem is hard by
discovering that beings with the same cognitive capacities have arrived at
wildly different answers. One might learn that a particular philosophical
problem—say, about what makes it the case that I am the same person
over time—is difficult by attempting to answer it, and finding oneself at
a loss. But alternatively, one could learn that it is hard by discovering that
there is a great deal of controversy about it. In the case of unique green,
learning about the disagreement is the crucial way of finding out that the
question is hard: each individual finds a shade that seems straightforwardly
neither bluish nor yellowish to her; it is only upon discovering the extensive
variation in judgment among those with similar cognitive capacities that
the intellectual task is revealed to be difficult. The fact that with respect
to various controversial moral questions, many of those on both sides of
the issue experience their own view as obviously true suggests that here, as
in the case of unique green, the better route to appreciating the relative
difficulty of the problem is the more indirect of the two.
Judgments to the effect that a given shade is unique green are controversial; are such judgments also ? Suppose that, in fact, some
among us are reliable detectors of unique green: the initial judgments of
members of this subpopulation are quite accurate, and non-accidentally
so. Relative to the general population then, the members of this group
are ‘‘experts’’ in the straightforward sense employed above. Consider their
position, once they come to learn that others—whose color vision otherwise
resembles their own—make contrary judgments. Do the experts have more
reason to think that it is the others who are in error than that they are? It
seems that they do not. For although their judgments are, ex hypothesi, more
accurate, they have no reason to think so. Thus, even true, reliably formed
Moral Disagreement and Moral Expertise
beliefs about which shade of green is unique green are . On
the assumption that  beliefs are not knowledge, neither are
these beliefs.⁹
But while the case against attributing knowledge to even reliable detectors
of unique green is quite strong, one might accept this conclusion while
denying that the kind of disagreement that surrounds our controversial
moral beliefs plays a similarly undermining role. The challenge, then,
would be to point to some compelling difference between the moral and
color case. Of course, there are some potentially relevant disanalogies. Here
I will mention two, and argue that neither is sufficient for a successful
defense of moral knowledge.
First, in the case of unique green, the subjects arrive at their judgments
completely independently of one another. You select a particular shade as
unique green, and those who select some other shade as unique green are
neither influenced by one another, nor by some common influence. This
stands in sharp contrast to the moral case, in which individuals do not
arrive at their views about controversial moral issues in isolation. One might
claim, then, that disagreement in the moral case creates significantly less
skeptical pressure than it does in the unique green case, inasmuch as those
on the other side of a particular moral controversy did not independently
converge on their view.¹⁰
Now it is certainly true that, in some cases, learning that those who think
otherwise did not arrive at their view independently can substantially reduce
the skeptical pressure on one’s own view. Consider an extreme case: the
population is more or less evenly divided with respect to some question, with
large numbers of people on both sides. Suppose you learn that all or almost
all of those on the other side of the issue believe as they do because they
unquestioningly defer to the judgment of a single charismatic individual
whom they regard as a guru. If most of those on your side of the issue arrived
at their view independently of one another, then it seems that you might
reasonably conclude that your belief is not  despite being
controversial. After all, this would be a case in which there is substantial
convergence on your view among individuals who made up their minds independently of one another. (Even if the guru himself arrived at the opposite
opinion on his own, he is greatly outnumbered by those on the other side.)
⁹ Notice that this conclusion can be accepted even by those drawn to reliabilist
accounts of knowledge. On sophisticated versions of such views (e.g. Goldman 1986),
true, reliably produced belief does not amount to knowledge when the subject is in
possession of undermining evidence. (See Goldman’s discussion of this point, 1986:
¹⁰ On the importance of independence, see esp. Goldman 2001 and Kelly (forthcoming).
Sarah McGrath
The difficulty with this response is just that actual moral controversies do
not seem to exhibit the kind of asymmetry that might make the disagreement
less threatening for either of the two sides. Granted, the many people who
contest some controversial moral opinion of yours did not converge on
their contrary view independently: the correct explanation of why that view
is held by a substantial number of people will undoubtedly attribute a great
deal to mutual influence, influence of common sources, and the like. But it
is not the case that those who share your view have independently converged
upon it. Thus, although it is true that people arrive at their judgments
about unique green in relative isolation as compared to the moral case, it
is far from clear that this disanalogy helps to defuse the skeptical challenge
facing our moral beliefs.
A second potentially relevant disanalogy between moral controversy and
the unique green case is the following. Judgments to the effect that a
particular shade is unique green are non-inferential judgments. But for
many of us, that is not how things are with our controversial moral beliefs.
Many of us can provide reasons or arguments in favor of our controversial
moral beliefs and against those of our opponents. Suppose, for example,
that I disagree with Alice about whether abortion is morally permissible.
She says that the fetus has a right to life, and that it follows that abortion
is impermissible. But I have read Judith Jarvis Thomson’s ‘‘In Defense of
Abortion’’, and so I can supply an argument that even if the fetus does
have a right to life it does not follow that abortion is impermissible. I
take myself to have rebutted her argument and thus to have more reason
to think that she is in error than that I am; I thus conclude that the fact
that Alice disagrees with me about abortion does not render my belief
More generally, one might think that non-inferential beliefs face skeptical
pressure from disagreement that is not faced by beliefs that are based
on arguments or discursive considerations.¹¹ Those who find this line
of thought plausible will think that the key disanalogy between moral
disagreement and the diversity of judgments about unique green lies here:
many of us take ourselves to have compelling arguments for our controversial
moral convictions while judgments about unique green are non-inferential,
brute judgments.
Of course, the mere having of reasons cannot be sufficient to defuse the
skeptical challenge. Presumably, if one can only offer bad reasons for one’s
¹¹ For example, Walter Sinnott-Armstrong argues that moral intutionism leads to
moral skepticism inasmuch as the moral intuitionist maintains that moral beliefs are
non-inferentially justified. According to Sinnott-Armstrong, ‘‘disagreement creates a
need for inferential justification’’ (2002: 312).
Moral Disagreement and Moral Expertise
view, that is not sufficient to break the symmetry between oneself and one’s
opponent. And neither is the offering of genuine reasons on behalf of one’s
controversial beliefs, when those genuine reasons can be matched by similar
reasons on the other side. After all, in the case of unique green, each of
us can at least cite as a reason that this shade appears unique green to me.
As each person’s reason seems equally compelling, the symmetry remains,
along with the skeptical pressure.
But suppose that one correctly recognizes that the argument on which
the other person bases her belief is fallacious. One is then in a position to
conclude that one’s own belief is not rendered  by the fact
that this person holds a contrary view. If one could do this more generally,
then one could establish that one’s belief was not . But
often those who engage in moral debate are dialectically skilled proponents
of the rival views. In such cases, there will be non-fallacious arguments on
both sides. The disagreement will then effectively reduce to one about the
relative plausibility of the fundamental premises from which the arguments
proceed. However, once the disagreement has been reduced to the question
of whose premises are more compelling, the gap between the case of moral
disagreement and the case of unique green seems to close. Of course, the
premises of my argument seem more compelling to me than the premises of
Alice’s argument; but by the same token, the premises of Alice’s argument
seem more compelling to her than the premises of my argument.
Alice might show me pictures that motivate a premise of her argument for
the conclusion that abortion is impermissible. But the pictures might not
move me to agree that that the premise is true. Can I break the symmetry,
then, by assuring myself that the reasons that I have are more compelling
than hers? This seems no better than simply privileging my judgment about
a given shade of green over Alice’s contrary judgment.
Once again, it seems that I need some principled line of reasoning by
which to privilege my judgment over that of those with whom I disagree.
The final section of this paper examines a recent account of such reasoning
due to Adam Elga.
In his recent paper ‘‘Reflection and Disagreement’’, Elga defends a view
known in the epistemology literature as the ‘‘the equal weight view’’.
According to the equal weight view, one is required to give equal weight to
the judgment of an epistemic peer as to one’s own judgment. You consider
someone your epistemic peer with respect to a given question just in case:
in advance of either of you reasoning about the issue, you would have
Sarah McGrath
predicted that the person in question was just as likely as you to arrive at the
correct answer. For example, if I would have predicted that you and I would
be equally likely to arrive at the correct solution to some mathematical
problem in advance of our actually performing the calculation, then I
consider you my epistemic peer with respect to that problem. According to
the equal weight view, if you and I arrive at different answers, I am required
to suspend judgment.
On the face of it, the equal weight view seems to have far-reaching
skeptical consequences, requiring us to suspend judgment with respect to
countless controversial questions. Elga, however, argues that this is not the
case. His general strategy is to show that one’s circle of epistemic peers
includes only those with whom one is in substantial agreement on issues
closely related to the one under dispute. Thus:
Consider Ann and Beth, two friends who stand at opposite ends of the political
spectrum. Consider the claim that abortion is morally permissible. Does Ann
consider Beth a peer with respect to this claim? That is: setting aside her own
reasoning about the abortion claim (and Beth’s contrary view about it), does Ann
think Beth would be just as likely as her to get things right?
The answer is ‘‘no’’. For (let us suppose) Ann and Beth have discussed claims
closely linked to the abortion claim. They have discussed, for example, whether
human beings have souls, whether it is permissible to withhold treatment from
certain terminally ill infants, and whether rights figure prominently in a correct
ethical theory. By Ann’s lights, Beth has reached wrong conclusions about most
of these closely related questions. As a result, even setting aside her own reasoning
about the abortion claim, Ann thinks it unlikely that Beth would be right in case the
two of them disagree about abortion … The upshot is that Ann does not consider
Beth an epistemic peer with respect to the abortion claim. (pp. 492–3).
It is clear enough how this general line of thought might be adapted so as
to apply to the argument with which we are concerned. Again, a belief of
yours is  if and only if it is denied by another person of
whom it is true that: you have no more reason to think that he or she is in
error than that you are. But in what circumstances do you have no more
reason to think that the other person is in error than that you are? Perhaps:
exactly when, in advance of either of you reasoning about the case at hand,
you would have predicted that the other person was just as likely as you to
arrive at the correct answer. That is, we might take a  belief
to be one that is disputed by someone who is an epistemic peer in Elga’s
sense. In that case, one might appeal to the line of reasoning Elga provides
and hold that one has significantly fewer  beliefs than one
might have thought, since those with whom one frequently disagrees over
controversial moral questions are outside one’s circle of peers. Following
Elga’s lead, one might say: even though Ann knows that Beth disagrees
Moral Disagreement and Moral Expertise
with her about abortion, this has no tendency to make Ann’s view about
abortion .
Elga anticipates a natural objection that runs as follows: Ann cannot
legitimately take her own views on the surrounding issues for granted and
use them as a basis for concluding that Beth is more likely to get things
wrong with respect to abortion. Rather, Ann should think of the entire
cluster of related issues as a single compound issue, and take into account
Beth’s disagreement about this single compound issue. Once she does
this, Ann will no longer be in a position to penalize Beth for having, by
Ann’s lights, false views about the surrounding issues. Hence, the case for
skepticism is restored.
In response, Elga offers the following:
Consider the cluster of issues linked to abortion. Contrary to what the objection
supposes, Ann does not consider Beth a peer about that cluster … That is because
there is no fact of the matter about Ann’s opinion of Beth, once so many of Ann’s
considerations have been set aside … To set aside Ann’s reasoning about all of these
issues is to set aside a large and central chunk of her ethical and political outlook.
Once so much has been set aside, there is no determinate fact about what opinion
of Beth remains. (pp. 495–6).
He motivates this claim with an example: plausibly, there is no determinate
answer to the question of what your opinion of Jennifer Lopez is, setting
aside your views that humans have bodies and that the Earth exists (p. 25).
However, while it seems right to say that there is no fact of the matter
about your opinion of Lopez setting aside your beliefs about human
embodiment and the existence of the Earth, the same maneuver seems
less plausible when applied to the case of Ann and Beth. Recall that Elga
characterizes Ann and Beth as ‘‘two friends at opposite ends of the political
spectrum’’. We might then think of Ann as a conservative Republican
who takes abortion to be morally abhorrent in most circumstances, and
Beth as a liberal Democrat who thinks that it is morally permissible in
most circumstances. No doubt, we would expect Ann and Beth to disagree
about a wide range of moral issues. Significantly, however, this is perfectly
consistent with a very substantial amount of moral agreement between the
two. Indeed, we would expect Ann and Beth to agree about the answers
to any number of moral questions. We would expect them to agree, for
example, that slavery is morally abhorrent, that it is wrong to cause others
pain for the sake of one’s own amusement, that lying is prima facie wrong,
and about countless other issues. Moreover, notice that many of the issues
on which they are likely to agree are highly non-trivial, at least when judged
by world-historical standards. (Consider, for example, their shared belief
that slavery is morally abhorrent.) With respect to moral sensibility, we
Sarah McGrath
would expect Ann and Beth to resemble one another far more than either
resembles, for example, a committed Nazi, or an ancient Hittite lord for
whom the thought that slavery is abhorrent was simply not on the moral
map. In short, Ann and Beth’s disagreements about abortion and related
matters, although substantial, almost surely take place against a relatively
wide background of shared moral beliefs. It seems wrong, then, to say
that there is no fact of the matter about Ann’s opinion of Beth’s moral
judgment, setting aside abortion and the many related issues about which
they disagree. Indeed, once these disagreements are bracketed, the relatively
wide background of agreement seems to tell in favor of Ann’s taking it that
Beth is more or less equally likely to get the hard questions right.
Once again, a comparison with the case of unique green seems apt.
Subjects with normal color vision make contrary judgments as to which
shade is unique green. Thus, those contrary judgments take place against
a relatively substantial background of shared color judgments. Plausibly,
this fact tends to increase the skeptical pressure: a shared background of
agreement strengthens the case for counting the conflicting judgments
as . The same would seem to be true in the moral case.
Elga’s proposal suggests that I can simply rank myself above those who
disagree with me about controversial moral issues on the grounds that our
disagreement is substantial. But this seems like a dubious procedure for
locating myself relative to others in the space of moral expertise—all the
more so when those who disagree do so against a wide background of
agreement. After all, according to me, they usually get it right.
Byrne, A., and Hilbert, D. (2003) ‘‘Color Realism and Color Science’’ Behavioral
and Brain Sciences 26: 3–21.
Christensen, D. (2007) ‘‘Epistemology of Disagreement: the Good News’’ Philosophical Review 116: 187–217.
Cohen, J. (2003). ‘‘Perceptual Variation, Realism, and Relativization, or: How I
Learned To Stop Worrying And Love Variations In Color Vision (Open peer
commentary on Byrne and Hilbert, Color Realism and Color Science)’’ Behavioral
and Brain Sciences 26: 25–6.
Elga, A. (2007) ‘‘Reflection and Disagreement’’ Noûs, LXI(3): 478–502.
Feldman, R. (2003) Epistemology (Upper Saddle River, NJ: Prentice-Hall).
(2006) ‘‘Epistemological Puzzles about Disagreement’’ in Epistemology Futures,
ed. S. Hetherington (Oxford: Oxford University Press).
Ginet, C. (1980) ‘‘Knowing Less by Knowing More’’ Midwest Studies in Philosophy.
5: 151–62.
Goldman, A. (1986). Epistemology and Cognition. (Cambridge, MA: Harvard
University Press).
Moral Disagreement and Moral Expertise
(2001) ‘‘Experts: Which Ones Should You Trust?’’ Philosophy and Phenomenological Research 63.
Hardin, C. L. (1988) Color for Philosophers (Indianapolis: Hackett).
Harman, G. (1973) Thought (Princeton: Princeton University Press).
and Thomson, J. J. (1996) Moral Relativism and Moral Objectivity (Cambridge,
MA: Blackwell).
Harris Poll (2005) ‘‘Nearly Two-thirds of U.S. Adults Believe Human Beings
Were Created by God.’’ Harris Interactive, Inc. http://www.harrisinteractive.
com/harris poll/index.asp?PID=581
Kelly, T. (2005). ‘‘The Epistemic Significance of Disagreement’’ in Oxford Studies
in Epistemology, ed. J. Hawthorne and T. Gendler Szabo (Oxford: Oxford
University Press), 167–96
(forthcoming) ‘‘Peer Disagreement and Higher Order Evidence’’ in Disagreement, ed. R. Feldman and T. Warfield (Oxford: Oxford University Press).
Lycan, W. (1977) ‘‘Evidence one does not possess’’ Australasian Journal of Philosophy
55: 114–26.
Mackie, J. L. (1977) Ethics: Inventing Right and Wrong (New York: Penguin).
Rosen, G. (2001) ‘‘Nominalism, Naturalism, Epistemic Relativism’’ Philosophical
Perspectives 15: 69–91.
Shafer-Landau, R. (2004) Whatever Happened to Good and Evil? (Oxford: Oxford
University Press).
Sher, George. (2001) ‘‘But I Could Be Wrong’’ Social Philosophy and Policy 18:
Sidgwick, H. (1907/1981) The Methods of Ethics (Indianapolis: Hackett).
Sinnott-Armstrong, W. (2002) ‘‘Moral Relativity and Intuitionism’’ Philosophical
Issues 12: 305–28.
van Inwagen, P. (1996). ‘‘It is Wrong, Always, Everywhere, and for Anyone, to
Believe Anything, Upon Insufficient Evidence’’ in Faith, Freedom and Rationality,
es. J. Jordan and D. Howard-Snyder (Lanham, MD: Rowman & Littlefield).
Williamson, T. (2000). Knowledge and its Limits (Oxford: Oxford University Press).
University of Pennsylvania Law School
Penn Law: Legal Scholarship Repository
Faculty Scholarship
Second-Order Reasons, Uncertainty and Legal
Stephen R. Perry
University of Pennsylvania Law School,
Follow this and additional works at:
Part of the Judges Commons, Jurisprudence Commons, Legal History, Theory and Process
Commons, Philosophy of Mind Commons, and the Political Theory Commons
Recommended Citation
Perry, Stephen R., “Second-Order Reasons, Uncertainty and Legal Theory” (1989). Faculty Scholarship. Paper 1354.
This Article is brought to you for free and open access by Penn Law: Legal Scholarship Repository. It has been accepted for inclusion in Faculty
Scholarship by an authorized administrator of Penn Law: Legal Scholarship Repository. For more information, please contact
. 1)
It is an obvious and banal fact that reasons for action often conflict,
which is simply to say that they often call for different and incompatible
courses of action. Ordinarily we resolve such conflicts by assessing the
relative weight or strength of all the relevant reasons and then deciding
in favor of that action which has the greatest overall support. T his process, which Professor Joseph Raz calls determining what ought to be
done on the balance of reasons, 1 is clearly a fundamental and commonly
employed mode of practical reasoning. T he reasons which we take into
account when relying on this mode are, in Raz’s terminology, first- order
reasons; they are reasons for action that have been drawn directly from
considerations of interest, desire or morality. 2
It is one of Raz’s most important philosophical insights to have
noticed that determining what ought to be done on the balance of firstorder reasons is not the only mode of practical reasoning upon which
people rely. Sometimes we decide what to do on the basis of what R az
calls second- order reasons, which he defines as “reason[s] to act on or
refrain from acting on a reason. ” 3 T he most important category of second- order reasons recognized by Raz is that of exclusionary or peremptory reasons. T hese are reasons to refrain from acting on a reason.
Exclusionary reasons give rise to the possibility of another ty pe of conflict between reasons in addition to first- order conflicts, namely, confl icts
between a firs t- order reason and an exclusionary reason for not acting on
the first- order reason. When such a conflict occurs, according to Raz,

Assista nt Professo r of Law, McGi ll U ni versit y. I a m indebted to Ken Kress and Joseph
R az for comme nts on ear lier drafts. I would also lik e to thank Ronald Dworkin and J eremy
Waldron fo r he lpful di sc ussions on topics whi c h a re dea lt with in thi s paper.
2. !d . at 34.
3. Jd . a t 39; see a/so J. RA Z, THE AUTHORITY OF LAW 16-17 (1979).
[Vol. 62:913
the exclusionary reason always prevails just by virtue of being a reason of
a higher order. 4 The exclusionary reason does not override or outweigh
the first- order reason but simply excludes it from consideration by the
Raz’s distinction between first- and second-order reasons is a major
contribution to the philosophy of practical reason and the found ation
upon which much of his work in legal and political as well as in practical
philosophy is built. He has used it, for example, to analyze the concept
of a ru le in a more precise and fruitful manner than had previously been
possible, namely as an exclusionary reason of general application which
is also a first- order reason for action. 5 This analysis has permitted him to
show how rules are able to provide an intermediate level of practical reasoning which is capable of mediating between concrete decisions and ultimate reasons or values. 6 The idea of rules as mediating devices in
practical reasoning figures in turn as an important element in Raz’s analysis of practical and political authority, an analysis which he then draws
upon in his formulation and defense of legal positivism. The distinction
between first- and second- order reasons is thus the heart of a comprehensive and powerful system of thought in which legal and political philosophy are shown to be deeply rooted in the soil of a subtle and carefully
worked-out practical philosophy.
As Raz has correctly observed, “[p]hilosophers have tended too
often to avoid facing the complexities of practical reasoning with its
multi-level assessments …. Many … pessimistic conclusions are based
upon a confusion between the epistemological difficulties in establishing
the validity of ultimate values and the logical difficulties in explaining the
considerations often found in practical reasoning. ” 7 This Paper will
attempt to show that Raz himself has underestimated the complexities of
multi-level assessments in practical reasoning because he circumscribes
the possible categories of second-order reasons too narrowly. Raz’s own
definition and theoretical utilization of second-order reasons emphasize
the possibility of isolating a level of practical reasoning from the considerations and values which ultimately justify the decisions being taken.
The notion of a second-order reason is in fact far richer than Raz allmvs,
4. J. RAZ, supra note I, at 46.
5. !d. at 73; see also Raz, Promises and Obligations, in LAw, MORALITY
221-22 (P.M.S. Hacker & J. Raz eds. 1977).
6. See, e.g., J. RAZ, THE MORALITY OF FREEDOM 58 (1986).
7. J. RAZ, supra note I, at 94-95.
91 5
and there exist second-order reasons for action upon which people commonly rely, mainly in institutional contexts, that do not have the isolating effect he describes. Reliance on these reasons requires at least some
fam iliarity, and often a great deal, with the ultimate values and other
justifying reasons which figure in first- order practical reasoning. The
most important context in which second- order reasons of this non-isolating ty pe are to be found is legal reasoning. T heir prominence in common
Jaw reasoning in particular casts doubt upon at least certain aspects of
Raz’s posi tivism and lends support to what is essentially a Dworkinian
theory of law.
The Paper is di vided into two main Parts. The first is concerned
with certain aspects of practica l philosophy generally, while the second
focuses on the philosophy of law. The first Part begins with a brief overview of Raz’s account of how exclusionary reasons figure in practical
reasoning. This is followed by a discussion of the general nature of reasons and of the forms which practical reasoning can take when agents are
uncertain about what right reason requires. While this may seem at first
to constitute something of a detour away from the main concerns of the
Paper, it will lead to the clarification of some basic notions and the formulation of certain distinctions which will facilitate the development of
the argument to follow. A critical examination of Raz’s account of the
concept of authority is then undertaken. After it has first been shown
that Raz’s notion of an exclusionary reason is ambiguous, it is argued
that reliance on the type of exclusionary reason that figures in his analysis of authority, which will be referred to as a subjective exclusionary
reason, represents a rational strategy that agents can or should adopt in
order to deal with uncertainty about what action right reason demands.
Once this is understood, it becomes possible to define two new types of
subjective second- order reasons that allow agents to formulate strategies
for dealing with uncertainty which, unlike the exclusionary approach,
enable them to continue to take the underlying first- order reasons into
T he second Part of the Paper, which deals with the philosophy of
law, begins by outlining how Raz makes use of the idea of a (subjective)
exclusionary reason in order to defend a particularly powerful and probably definitive version of positivism . It is then shown that the two newlydefined categories of second- order reasons can be utilized to formulate
an interpretation of the common law process of decisionmaking which is
superior to Raz’s own positivist interpretation. This offers strong support for the conclusion that positivism cannot provide a complete
[Vol. 62:913
account of the foundations of law, and that at the very least it has to be
supplemented by a theory which in its most fundamental respects resembles that of R onald Dworkin.
It has already been mentioned that Raz defines a second- order reason as a reason to act for a reason or to refrain from acting for a reason.
H e pays very little attention, however, to positive second- order reasons,
which are reasons to act for a reason. It is negative second- order reasons,
which are reasons to refrain from acting for a reason , that figure most
prominently in his theoretical work. 8 He also calls negative secondorder reasons exclusionary or peremptory reasons. 9 R az defends the thesis that determining what ought to be done on the basis of an exclusionary reason is a distinct mode of practical reasoning by pointing to the
ways in which people actually do deliberate about what they ought to do.
F or example, a soldier who has been given an order to do an action
which he thinks cannot be justified on the balance of reasons will ordinarily regard the order as a reason for him not to act on his view of the
merits of the case rather than as simply another reason which is to be
added to the balance of reasons. As Raz says, “we would be disregarding
[the soldier’s] own conception of the situation if we were to say that he
regards the order as an overriding first-order reason.” 10 The soldier conceives of the order as an exclusionary reason which is also a first-order
reason to do the action that he was ordered to perform. Suppose that the
soldier nonetheless disregards the order and acts on his own, correct
assessment of the balance of reasons. H is superior offi cer will now, says
Raz, ” be torn between conflicting feelings,” since he is fa ced with conduct which he concedes was right on the merits but which he nonetheless
thinks was wrong in disregarding an exclusionary reason. 1 1 The conflict
felt by the superior officer is an indication that he is aware that the soldier’s action can be assessed in two different and incompatible ways. It is
an indication, that is to say, that he recognizes that there are two distinct
modes of practical reasoning, one of which demands that a person act on
Raz discusses negative and positive second-order reasons in
RAZ, supra note 3, at 17.
9. For criticism of Raz’s notion of an exclusionary reason see Gans, lWandatory Rules and
Exclusionary Reasons, 15 PHILOSOPHIA 373 ( 1986).
J. RAZ, supra note 1, at 42; cf id. at 74-75; J. RAZ, supra note 3, at 22-23.
J. RAz, supra note i, at 43.

91 7
his own assessment of the balance of reasons and the other of which
demands that he disregard that assessment.
Raz argues that such practical phenomena as promising, com plying
with an order, acting on advice, taking a decision and follo wing a rule
must all be a nalyzed in terms of the acceptance of an exclusionary reason
for actio n. 12 But his concern is not just with the phenomenology of practical reaso ning . H e wishes to establish mo re than that peopl e do, in fact ,
sometimes determ ine what they ough t to do by reasoning in an exclu sionary rr1an n.:r. H e also maintains that there can be valid exclusionary
reasons, i.e., there can be excl usionary reasons upon which people are
justified in ac ting. It has already been mentioned that Raz analyzes the
notion of a ru le as a general exclusionary reason which is also a firstorder reason for some particular course of action. As he points out, it is
quite plausible to suppose that indi viduals will sometimes be better off in
their everyday lives if they follov rules of thumb which have this logical
structure than they would be if they were to decide what to do in each
relevant situation by assessing the balance of reasons on a case by case
basis. Following a predetermined course of action could well reduce
overall error when one has to act in circumstances of impaired rationality, for example. 13 I t is also obvious that following a rule which one has
adopted in advance in order to deal with a particular type of situation
that is expected to recur could in principle be justified by the time and
effort which would be saved in not having to reassess the balance of reasons on every relevant occasion. 14
R az further argues tha t it can be rational not simply to follow an
exclusionary reason which one has formulated for oneself, at a tim e
before the rel evan t situation or series of situations has arisen, but also to
treat the utterances of another person as exclusionary reasons for action.
To treat another person (or an institution) as a practical authority, for
example, means that one accepts the directives of that person (or institution) as exclusionary reasons applicable to oneself. 15 When one person
has authority over another the former possesses a type of normative
power which enables him or her to change the latter’s protected reasons,
these being exclusionary reasons which are also first-order reasons . 16
12. See id. at 49-84; see also J. RAZ, supra note 3, at 21-25, 30. On promises see Raz, supra
no te 5, at 210-2 8.
13. J. RAZ, supra note I, a t 37-38 , 59-60.
14. !d. at 59- 60; see also R az, supra note 5, a t 224.
15 . J. RA Z, supra note 3, at 26.
16. !d. at 21; cf 1. RAZ, supra note 6, at 24. Raz defines protected reason s in J. RAZ , supra
note 3, at 18.
[Vol. 62 :91 3
Raz defends what he calls the ” service conception” of authority, which
regards authorities as “mediating between people and the right reasons
which apply to them, so that the autho rity judges and pronounces what
they ought to do acco rding to right reason.” 17 T he normal and primary
way of justifying the legitimacy of an authori ty is to show that the person
over ·whom authority is claimed “is likely better to com ply with reasons
which apply to him (other than the alleged d irectives) if he accepts t he
directives of t he alleged authority as aut horitati vely bind ing and tries to
follow them, rather than by t rying to follovv the reasons which appl y to
him directl y.” 16 Raz refers to this claim as the normal justification thesis.
According to the service conception, aut horities should base their directives on reasons which apply to their subjects. (Raz calls both reasons
which apply to the subjects and reasons wh ich are meant to reflect such
reasons dependent reasons. 19 ) But a legitimate authority cannot carry
out its mediating role unless its subjects treat its directives as reasons for
action which replace some of the reaso ns which would otherwise be relevant in assessing what they ought to do and not simply as reasons that
are to be ad ded to those original reasons. 20 Subjects should, in other
words, treat the authority’s directives as exclusionary reasons for action.
Raz’s analysis of the concept of legitimate authority effectively
rebuts the anarchist claim that the concept is necessarily incompatible
with rationality, since he shows that the directives of an authority could
in principle be valid exclusionary reasons which it would be rational for a
person to follow .2 1 He also outlines a number of ways in which the legitimacy of a practical authority, including a political authority such as the
government of a state, could in fact be established. 22 One way is to show
the authority to be wiser than the individual in determining what ought
to be done in a particular type of situati on. Another requires a demonstration that the authority is in a better position than the individual to
achieve what the latter has reason to ac hieve but cannot, such as soiving
coo rdination problems or changing the structure of prisoner’s dilemmatype situations. 2 3 Raz is of the opinion that such considerations are in
fact capable of justifying political authority, but only on a piecemeal
basis and only up to a point. The extent of a government’s authority is
17. Raz, A u1hori1y. Law and
note 6, at 55-56.
18. J. RAZ, supra note 6, at
!9. J. RA Z, supra note 6, at
20. !d. at 57-59 .
21. J. RA Z, supra note 3, at
22. J. RAZ, supra not e 6, at
23. J. RAZ , supra no te 6, a t
Mo ra!iry, 68 THE MONiST 295 , 299 ( 1985); see also J. RAZ, supra
53; see also Raz, supra note 17 , at 299.
27 ; J. RAZ, supra note 6, a t 57 , 68.
75; cf J. RAZ, supra no te I, at 63-64.
9 19
likely to vary from individual to individual, and will in the case of most
persons be more limited than the unrestricted authority which governments actually claim for themselves.24
There are three other points about Raz’s conception of an exclusionary second- order reason which are worth mentioning here. The fir st is
that exclusionary reasons can differ in scope. This simply means that an
excl usionary reason may exclude only som e of the reasons which would
otherw ise be applicable in a gi ven situation, and that the range of
excl uded reasons can vary. 2 5 T hus an authority, for example, can be limited not just by the kinds of acts which it is justified !n regulating but also
by the kinds of reasons upon which it may rely and which its decisions
will preempt. 2 6 The second point is that Raz uses the notion of an exclusionary reason to explain the concept of an obligation or duty. This is
accomplished in the following way: “An action is obligatory if it is
required by a categorical rule, i.e. [an exclusionary] rule which applies to
its subjects not merely because adherence to it facilitates achievement of
their goals.’m , T he third point is that being subject to the authority of
another and therefore, according to Raz, possessing a duty to treat that
other’s directives as exclusionary reasons, does not entail that one is not
permitted to form a judgment of what ought to be done on the balance of
reasons. Authority does not require a “surrender of judgment” in this
strong sense; the only thing which is necessarily excluded is action on
one’s judgment of what ought to be done. 28
Section C undertakes a critical inquiry into the nature of exclusionary reasons as these figure in R az’s analysis of the concept of authority.
This will enable us to see that there are in fact more types of secondorder reasons than Raz acknowledges. First, however, it will be helpful
to say something about the general nature of reasons for action, paying
particular attention to the fact that practical reasoning must often take
place under conditions of uncertainty. This will clarify certain matters
24. !d. at 80.
25 . J. RA Z, sup ra no te I, at 40; J. RAZ , sup ra note 3, a t 22.
26. J. RA Z, supra note 6, at 46- 4 7.
27. Raz, supra note 5, a t 223; cf J. RAZ, supra note 3, at 234-35. S ee also J. RAZ, supra note
6, at 186, 195; J. RAZ, supra note I, at 76 (moral duties are ex c lusionary in nature). Donald Regan
argues pe rsuasively th a t R a z’s account o f a uthorit y, while broad ly co rrec t, cannot sus tain th e conclusion that th e subjects of even a legitima te auth o rit y have a duty to obey the auth o rity’s direc tives.
See Regan, Au1hori1y and Value: Refieaing on R az ‘s 1l,foralily of Freedom, 62 S. CAL. L. REV. 995
( 19 89) . T his is not an issu e which will be discu ssed in this Paper.
28. J. R AZ, supra note 6, at 29; J. R AZ, supra note 3, a t 26 n.25 .
[Vol. 62:913
and yield certain di sti nctions which will be of use in the disc ussion to
Raz on the Na ture of Reasons
In Practica l R eason and Norms an d other rel ati vely early works Raz
dravvs a distinction between two different ·;;.;ays that we common ly talk
” actJOn
· .-‘9 H
, . sometimes
1 01.c reasons
a b out reasons ror
:· e notes tna{
we speaK
as facts, while at other times we speak of them as beliefs. (The concept of
a “fact” is to be understood here as incl udin g values and moral principles
as well as events and processes. 30 ) Thus we refer both to the fact that it
wi ll rain and to a belief that ii will as reasons for carrying an
umbrella. R az acknowledges that one could distinguish in this way
between two distinct notions of reason, but he goes on t o say this :
Only reasons understood as facts are normatively significant; only they
determine what ought to be done. To decide what we should do we
must find what the world is like, and not what our thoughts are like.
The other notion of reasons is relevant excl usively for explanatory purposes and not at all for guiding pu rposes. 3 1
In thus maintaining that only reasons understood as fac ts are normatively significant it would appear that Raz is not simply stating the
obvious truth that it is only facts about the physical world, morality, and
one’s desires and interests that ultimately determine what ought to be
done. H e seems to be saying in addition that practical reasoning can be
adequately characterized without reference to the beliefs which agents
hold about their reasons, so that a theory of practical reasoning need
only refer to the idea of reasons as facts . A gents are to be regarded as
deliberating about what to do on the assumption that certain facts which
could serve as reasons obtain (or that they do not obtain). Sometimes
they will discover after they have acted tha t this assumption did not
hold, and it is in this situation t hat the ;dea of reasons understood as
beliefs serves its explanatory fun ction: “It is mostly when we come to
believe that the reason on which we relied does not obtain that we cite
our belief in it as a reason.” 32 But for ex ante guiding purposes, as
opposed to ex post explanatory purposes, the assump ti on is to be
29. J. RAZ, supra no te 1, a t 16-19: see also Raz. Inlroduction, in P RACTICAL REASONING 1, 24 (J. Raz ed. 1978).
30. J. RAZ, sup ra note 1, at 17-18.
31. !d. at 18.
32 . !d. According to Raz, ta lk of reasons as beliefs could be replaced even in explan a tory
con texts by a m od e o f speak in g wh ic h co nform ed wit h hi s own analysis if we were to sa y aft er the
fact that we did no t have a reaso n but that we had a reaso n fo r thinking we had a rea so n.
regarded as sound; practical reasoning is essentially just a matter of
m anipulating statements about states of affairs w hich, if they obtained,
would be both facts and reasons, where the agent’s beli ef in the tru th of
t hose statements is taken as given. T hus Raz defines an atomic complete
reason as “the fact(s) stated by the premises of a sound practical inferr:nce ·with no red undant premises. ” 33 (Of the different reasons t hat can
comprise a complete reason, those which state valid goals are operative;
o. 1l ~; fners are refe rred to as auxilimy. 3 -+) Raz then defines practi ca l reasoning as ”the transition (not necessarily conscious) from belief in the
pcerrdses to acceptance of the putative conclusion of a practical infer enr: t::.” 35 (The conclusion of a practical inference is always, according to
36 )
,.)…..,-:.. :: “~J r1 ,o.o,…1l t;c
i ‘ .’::
.•. ,_, s t”‘~em

We offer the bestcustom writing paper services. We have done this question before, we can also do it for you.

Why Choose Us

  • 100% non-plagiarized Papers
  • 24/7 /365 Service Available
  • Affordable Prices
  • Any Paper, Urgency, and Subject
  • Will complete your papers in 6 hours
  • On-time Delivery
  • Money-back and Privacy guarantees
  • Unlimited Amendments upon request
  • Satisfaction guarantee

How it Works

  • Click on the “Place Order” tab at the top menu or “Order Now” icon at the bottom and a new page will appear with an order form to be filled.
  • Fill in your paper’s requirements in the "PAPER DETAILS" section.
  • Fill in your paper’s academic level, deadline, and the required number of pages from the drop-down menus.
  • Click “CREATE ACCOUNT & SIGN IN” to enter your registration details and get an account with us for record-keeping and then, click on “PROCEED TO CHECKOUT” at the bottom of the page.
  • From there, the payment sections will show, follow the guided payment process and your order will be available for our writing team to work on it.