Download An Internalist Dilemma - University of Colorado Boulder

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Document related concepts

Consequentialism wikipedia, lookup

Speciesism wikipedia, lookup

Ethics wikipedia, lookup

Morality and religion wikipedia, lookup

Neeti Sastra wikipedia, lookup

Lawrence Kohlberg wikipedia, lookup

Pikuach nefesh wikipedia, lookup

John McDowell wikipedia, lookup

Morality throughout the Life Span wikipedia, lookup

Alasdair MacIntyre wikipedia, lookup

Moral disengagement wikipedia, lookup

Moral development wikipedia, lookup

Lawrence Kohlberg's stages of moral development wikipedia, lookup

The Theory of Moral Sentiments wikipedia, lookup

Thomas Nagel wikipedia, lookup

Emotivism wikipedia, lookup

Ethical intuitionism wikipedia, lookup

Secular morality wikipedia, lookup

Morality wikipedia, lookup

Thomas Hill Green wikipedia, lookup

Moral relativism wikipedia, lookup

Moral responsibility wikipedia, lookup

Internalism and externalism wikipedia, lookup

Transcript
An Internalist Dilemma—and an Externalist Solution
Caj Strandberg
ABSTRACT. In this paper, I argue that internalism about moral judgments and motivation faces a dilemma.
On the one hand, a strong version of internalism is able to explain our linguistic intuitions as regards the
connection between moral language and motivation, but fails to account for the fact that people who suffer
from certain mental conditions need not be accordingly motivated. On the other hand, the most plausible
version of a weaker form of internalism, rationalist internalism, avoids this difficulty, but fails to explain our
linguistic intuitions. Moreover, I argue that externalism can account for the first notion by employing Grice’s
concept of generalized conversational implicature and that this view is able to account for the second notion
as well. Consequently, there is reason to think that externalism is preferable to internalism.
1. Introduction
There is broad agreement in metaethics that there is a strong connection between moral
language and motivation, but significant disagreement as to how it should be understood.
According to internalism, there is a necessary relation between moral judgments and
motivation whereas externalism denies this. It is widely taken for granted that internalism
has initial credibility since it seems able to explain our conception of this connection
whereas it is less clear whether externalism is able to do so. At the same time, there is
disagreement within the internalist camp about how strong the necessary relation between
moral judgments and motivation should be understood to be. On strong internalism, it
holds for all persons whereas it on weak internalism only holds for those who satisfy a
certain condition.
In Sections 2 through 6, I argue that internalism faces a dilemma: Strong internalism
can explain our linguistic intuitions as regards the connection between moral language and
motivation, but fails to account for the fact that some persons need not be motivated in
accordance with their moral judgments. The most plausible version of weak internalism,
rationalist internalism, can account for the latter notion, but not the first. In Sections 7 and
8, I suggest that externalism can explain the first notion in terms of Paul Grice’s concept of
generalized conversational implicature and that it can account for the second notion as
well. As internalism faces the dilemma whereas externalism avoids it, there is reason to
prefer the latter view to the former.
2. Strong Internalism
The following formulation of internalism about moral judgments and motivation is
standard in metaethics: It is necessary that if a person judges that it is morally right that she
s, then she is motivated to . However, there are different views that are labelled
1
‘internalism’ in the literature and I will not be concerned with all of them.1 The kind of
internalist claims I am interested in have four basic claims in common. First, I understand
internalism as a conceptual claim. It can be understood to state that a sentence to the effect
that it is right to  expresses, in virtue of its conventional meaning, a judgment such that if
a person holds it, she is motivated to . As I use ‘judgment’, it is neutral between
cognitivism and non-cognitivism: it might consist in a belief or a non-cognitive state.2
Second, as a natural corollary, motivation is internal to such moral judgments. More
precisely, such a judgment is sufficient for motivation by consisting in, entailing (without
consisting in), or by causing a motivational state.3 Third, a person’s moral judgment
suffices for her to be motivated to some extent to ; ing need not be what she is most
motivated to do. Fourth, it is noticeable that a mental state can be understood as a
dispositional state which might be activated in the form of corresponding occurrent state or
as such an occurrent state.4 This means that internalism entails that it is conceptually
necessary that if a person judges that it is right that she s—where the moral judgment can
be understood as a dispositional state or an occurrent state—she is motivated to  in the
sense of being disposed to ; it does not entail that she has an occurrent motivation to .
Corresponding claims hold for what I call ‘normative internalism’ below. Lastly,
internalist claims vary in strength depending on whether they apply to all persons or
merely persons who fulfil a certain condition.
Externalism entails the denial of internalism. On this view, there is no necessary
connection of this kind between a person’s moral judgment and her motivation to act.
Consider:
Strong internalism: It is conceptually necessary that, for any action  and person S, if
S judges that it is morally right that she s, then she is motivated to .5
1
In particular, I will not discuss views according to which there is a necessary connection between the
practice of making moral judgments in a community and motivation (see e.g. Tresan (2004), pp. 143–165,
and Bedke (2009), pp. 189–209), and my solution to the internalism dilemma is compatible with this view.
2
In case a person utters a sentence to the effect that it is right to , we presume that she thinks that it is right
that she s unless she or the context indicates otherwise. I will adhere to the simpler formulation in what
follows. For a useful discussion of how to understand internalism, see Miller (2008), pp. 234–236.
3
For a possible example of a causal reading, see Smith (1994), e.g. p. 179.
4
See e.g. Mele (2003), pp. 30–33, and Persson (2005), p. 55. Internalism as I formulate it is neutral as to
whether motivation consists in a desire or some other mental state.
5
Some authors can be understood to support strong internalism; see e.g. Blackburn (1998), pp. 59–68;
Lenman (1999), pp. 441–457; Joyce (2001), pp. 17–29, and Gibbard (2003), pp. 152–154.
2
One of the main arguments for internalism is that it is able to explain our linguistic
intuitions as regards the connection between moral language and motivation. In fact,
virtually all authors who are attracted by internalism appeal explicitly or implicitly to this
consideration and it might be regarded as the prime basis for internalism. Importantly, this
fundamental argument for internalism is employed by strong as well as weak internalists,
in spite of the fact that their claims differ considerably in strength, thereby suggesting that
it provides support to internalism in general. In the typical argument, we are asked to
imagine a person who utters a sentence to the effect that it is right to  but who has no
motivation whatsoever to .6 Thus, try to imagine a person who says ‘Actually, it’s right to
give some money to those who need it more than I do’, but who is not motivated at all to
actually give any money to the needy. We doubtless respond to such thought experiments
by finding them puzzling. After all, in case a person utters such a sentence, we strongly
presume that she is motivated in accordance with it, and in case she is not, we want an
explanation as to why she makes the utterance. Strong internalism has a ready explanation
of our response. On this view, a sentence saying that it is right to  expresses, in virtue of
its meaning, a motivating moral judgment. Moreover, this holds for all persons. The
explanation of why we find the mentioned cases puzzling is thus that we know, in so far as
we are competent language users and know the meaning of the sentence, that any person
who holds such a judgment is motivated to act in accordance with it.
3. Rationalist Internalism
Most internalists have come however to the conclusion that strong internalism is mistaken
because there might be cases where a person thinks that it is right that she s, but is not
motivated to , because she suffers from some mental condition, such as apathy,
depression, emotional disturbance, addiction, or compulsion.7 This has made them suggest
a weaker type of internalism:
6
For just a few examples, see Stevenson (1944), pp. 16–17; Frankena (1973 (1958)), p. 100; Korsgaard
(1986), p. 9; Dreier (1991), pp. 13–14; Dancy (1993), p. 4; Smith (1994), p. 60; Blackburn (1998), pp. 48,
52–53; Lenman (1999), pp. 443–446; Joyce (2001), pp. 26–27; van Roojen (2002), p. 32; Finlay (2004), pp.
206, 212; Shafer-Landau (2003) p. 156; Ridge (2007), p. 51, and Bedke (2009), pp. 189–209. (Not all of
these authors ultimately defend internalism although they find it attractive.) Many commentators take for
granted that internalism is plausible without providing any arguments for this view; however, it seems
plausible to assume that they are led to this view by considerations of our linguistic intuitions in the way
indicated in the text. I argue against this kind of arguments in REF.
7
See e.g. Stocker (1979), pp. 738–753, and Mele (2003), Ch. 5.
3
Weak internalism: It is conceptually necessary that, for any action  and person S
who satisfies condition C, if S judges that it is morally right that she s, then she is
motivated to .
Weak internalism constitutes a type of internalism where condition C can be specified in
different ways. The resulting claims have in common the idea that what explains that a
person is not motivated  is that she does not satisfy C.8
In order for weak internalism to be a proper version of internalism C has to be
understood in a way which does not trivialize the internalist claim. Weak internalism
should be understood to maintain that the content of a person’s moral judgment is such that
it provides a genuine explanation as to why a person who satisfies C is motivated in
accordance it. I take this restriction to mean that C cannot be understood in such a way that
it becomes trivially true that a person who satisfies C is motivated in accordance with her
moral judgment. Thus, C cannot be understood to simply state that a person who satisfies it
is motivated in accordance with her moral judgments. Similarly, it cannot be understood as
a mere ad hoc condition, such as a negation of a disjunction of mental conditions which is
thought to result in the fact that a person is not accordingly motivated.
It has been shown extremely difficult for internalists to spell out C in a way which is
compatible with this restriction. According to one suggestion, C is understood as ‘normal
conditions’. According to another, C consists in the absence of certain general motivational
disorders.9 It is difficult to see that these understandings of C can be elucidated in a way
which is not susceptible to the indicated difficulty.
The version of weak internalism which seems most likely to avoid this difficulty is
one which specifies C in terms of rationality:
8
It might be thought that weak internalism is not an instance of internalism at all on the ground that
motivation is not internal to moral judgments on this view. As a person might hold a moral judgment without
being motivated in accordance with it, such a judgment cannot be sufficient by itself for motivation.
However, even if it might be the case that moral judgments are not sufficient for motivation for all persons,
they can be thus sufficient for some persons, namely those who satisfy condition C.
9
For the first suggestion, see Dreier (1990), pp. 13–14; for the second, see Dancy (1993), p. 25, and
Svavarsdóttir (1999), p. 165. Of course, these authors are aware of the difficulties of formulating weak
internalism; see esp. Dreier. For criticism, see e.g. Roskies (2003), pp. 53–55.
4
Rationalist internalism: It is conceptually necessary that, for any action  and
rational person S, if S judges that it is morally right that she s, then she is
motivated to .
On rationalist internalism, it has to be something about a moral judgment which explains
that it is only if a person is rational that she needs to be motivated in accordance with it.
The most plausible explanation is that moral judgments consist in judgments about what
there are normative reasons to do. It seems evident that there is a connection between a
person’s normative judgments, her rationality, and her motivation to act. In case a person’s
motivation to act are not in accordance with her own views about what she has reason to
do, this is grounds to think that she is irrational.
This is the intuitive line of thought which leads up to rationalist internalism, but we
need to formulate it more exactly:10
(1) Rationalism: It is conceptually necessary that, for any action  and person S, if S
judges that it is morally right that she s, then she judges that she has a normative
reason to .
Rationalism captures the notion that moral judgments consist in normative judgments.
(2) Normative internalism: It is conceptually necessary that, for any action  and
rational person S, if S judges that she has a normative reason to , then she is
motivated to .
Normative internalism captures the notion that a person is irrational unless she is motivated
in accordance with her normative judgments. In order to function as a premise in this line
of thought, it should be understood as a conceptual claim and motivation should be internal
to the normative judgment.11
10
This kind of reasoning is familiar in metaethics but in recent times it is most associated with the writings of
Michael Smith and Christine Korsgaard; see e.g. Smith (1994), esp. Ch. 3 and 5, and Korsgaard (1986), pp.
5–25. For views, see e.g. Audi (1997), pp. 137–138; Wallace (2006), pp. 182–196, and Wedgwood (2007),
esp. Ch. 1.
11
Normative internalism as formulated in (2) is, as far as I know, seldom explicitly defended although it
might be presumed in various arguments, i.e. in arguments for rationalist internalism. For clear formulations
5
We may now infer:
(3) Rationalist internalism: It is conceptually necessary that, for any action  and
rational person S, if S judges that it is morally right that she s, then she is motivated
to .
What I said above implies that rationality has to be understood in such a way that it does
not trivialize rationalist internalism and it might be argued that it is difficult to find a
characterization of rationality which satisfies this constraint.12 However, having identified
condition C in terms of rationality is a significant improvement and I will not be concerned
with this problem in what follows.
We saw earlier that strong internalism seems able to explain our linguistic intuitions
as regards moral language and motivation, but that it is too strong since it cannot account
for cases where there connection between moral judgments and motivation is defeated.
Now, it might be suspected that weak internalism faces the reverse problem. Weak
internalism seems able to explain the latter fact. However, it might be suspected that in
case a person utters that an action is right, but is not accordingly motivated, we would find
this puzzling even if we do not have any view as to whether she satisfies C and even if we
know that she does not satisfy C. As a consequence, weak internalism would be too weak
to explain our conception of the connection between moral language and motivation. In the
two subsequent sections I argue that the most promising version of weak internalism—
rationalist internalism—is vulnerable to this difficulty. Rationalism in (1) has been
questioned by a number of metaethicists, but I will argue by focusing on normative
internalism in (2).13
4. The Problem Child: Normative Internalism
In metaethical discussions about the connection between normative judgments, rationality,
and motivation, a person’s various normative judgments are often considered in isolation
from each other. Normative internalism as formulated in (2) might then seem plausible.
of it, see e.g. Darwall (1990), p. 258; Smith (1994), p. 148; Wedgwood (2002), p. 347, and Setiya (2004), p.
268. (However, of these authors only Smith explicitly endorses it.)
12
For some critics of rationalism, see Sayre-McCord (1997), pp. 64–65; Svavarsdóttir (1999), pp. 164–165;
Miller (2003), p. 221, and Roskies (2003), p. 53.
13
For critical discussions of rationalism, see e.g. Foot (1978), pp. 157–173, and Brink (1989), Ch. 4.
6
However, when we consider a person’s various normative judgments we can see that this
view is much less appealing.
One type of cases which provide difficulties for (2) are the following. Assume that a
person thinks that what she has absolutely strongest reason to do in a certain situation is to
perform an action , but that she also thinks that she has an extremely much weaker reason
to perform another action ψ. Assume further that she thinks that  and ψ are practically
incompatible in the sense that she cannot do both. In such a case it seems reasonable to
maintain that she has to be motivated to  in so far as she is rational.14 Accordingly, the
following claim seems plausible:
It is conceptually necessary that, for any action  and rational person S, if S judges
that she has strongest normative reason to , then she is motivated to .
However, normative internalism in (2) is a much stronger claim. It entails that it
conceptually necessary that for each of the countless actions she thinks she has a reason to
perform, she is motivated to do that particular action however weak she thinks that reason
is. According to (2), it is conceptually necessary that the person above is motivated to ψ as
well in so far as she is rational. It follows that she has to be motivated to ψ in order to be
entirely rational and that she is irrational to the extent she is not motivated to ψ. However,
it appears that she might be completely rational even if she is not motivated to ψ.15
To see this, consider an example. Assume that a person who is seriously ill knows
that undergoing a certain medical treatment is the only way to save her life. She
consequently thinks that what she has absolutely strongest reason to do is to accept the
treatment. However, imagine that her doctor informs her that if she goes through the
treatment, she will not be able to, say, drink coffee for one minute. Assume that she
therefore thinks that she has some, but extremely much weaker, reason to decline the
treatment. We might accept that she might be motivated to decline the treatment even if she
is rational. However, (2) entails that it conceptually necessary that she has to be motivated
to decline the treatment in order to be entirely rational; that it follows that she is to a
Moreover, she needs presumably to be most motivated to . For some defenders of this view, see Scanlon
(1998), pp. 25–30; Parfit (2001), pp. 15–40; Broome (2007), pp. 349–374¸ and Wedgwood (2007), esp. Ch.
1. However, there are philosophers who deny it; see Gert (2004).
15
See Sinnott-Armstrong (1992), p. 302; Copp (1997), pp. 45–46; Wallace (2006), pp. 187–188, and Gert
(2008), pp. 16–17.
14
7
certain extent irrational unless she is motivated in that way. I think that we accept that she
might be entirely rational even if she is not motivated to decline the treatment.
Consider a more dramatic example. Assume that a person knows that the only way to
save ten people’s lives is to call the ambulance. He consequently thinks that what he has
strongest reason is to do so. But assume that he also realizes that in case he does, he run the
risk of getting a stain on his new suit in the telephone booth. He therefore thinks that he
has some, but extremely much weaker, reason not to call the ambulance. Again, it seems
mistaken to claim that, as a matter of conceptually necessity, he has to be motivated in that
way in order to be entirely rational and that he is irrational to the extent he is not.
It might be objected that these arguments do not take into consideration that the fact
that a person is motivated to perform an action merely entails that she is disposed to do it,
not that she has an occurrent motivation to do it. However, I do not think that rationalist
internalists can save (2) merely by making the reference to dispositions explicit. First, I
think we respond in the same way if we formulate the case under consideration in terms of
dispositions. Return to the person who thinks that she has an extremely strong reason to
accept a medical treatment and an extremely weak reason to decline it because she thinks
that going through the treatment would mean that she cannot drink coffee for one minute.
It does not seem conceptually necessary that she has to be disposed to decline the
treatment in order to be entirely rational. And it seems incorrect to claim that it is
conceptually necessary that she is irrational to the extent she is not disposed in that
manner. Second, we should be careful not to draw the conclusion that a person has a
certain disposition on the wrong grounds. According to a simple conditional analysis of
dispositions, an object is disposed to M when conditions C are present in so far as the
object has an intrinsic property B such that the object would M in C. It should first be
stressed that what is relevant to the present issue is not simply whether it is plausible to
think that a person has a certain disposition. The relevant issue is much more limited,
namely whether it conceptually necessary that if a person thinks that she has absolutely
strongest reason to , and a much weaker reason to ψ, then she has to be disposed to ψ in
order to be entirely rational. We should thus not be confused by the simple fact that we
easily can come to think of various circumstances in which the person in our example is
disposed to decline the treatment as this does not show that the mentioned view is correct.
Moreover, it has shown to be extremely difficult to specify conditions C even for simple
examples of dispositions, such as fragility, and there is no reason to think that dispositions
8
to act constitute any exception. As is made clear in the literature, it is not correct to claim
that an object is fragile merely because we can imagine some conditions that would make
an object crackle.16 A similar point holds for dispositions for action. These two
considerations suggest that we should not be misled to think that the person in our example
has to be disposed to decline the treatment merely because we can come to think of certain
conditions in which she would do so.17 Moreover, the fact that it is difficult to specify in
what conditions a person has a disposition to perform an action indicates that even if (2) is
true, it is not an apparent conceptual truth. I will return to a similar point below.
Thus far, we have been concerned with one type of cases which provide a difficulty
for (2). However, there are also other kinds of cases considerations of which put doubts on
this view.18
We might first consider a variant of the type of cases we considered earlier. Assume
again that a person thinks what she has strongest reason to do is to accept a certain medical
treatment because undergoing the treatment would save her life, but that she also thinks
that she has a much weaker reason to decline the treatment because it would make her
unable to drink coffee for one minute. However, assume now that she also thinks that if she
is motivated to decline the treatment, this will make it more difficult for her to actually get
it. It might for example be the case that she thinks that if she is motivated to decline the
treatment, this will start a psychological process in her which ultimately will make her give
in and decline the treatment. In that case it seems even more incorrect to maintain that she
needs to be irrational unless she is motivated to decline the treatment.19
16
For discussion of this point, see e.g. Johnson (1992), pp. 221–263; Lewis (1999), pp. 133–151, and Choi
(2006), pp. 369–379.
17
It might be argued that a person who thinks that she has a strong reason to accept a medical treatment, but
a weak reason to decline it, has to be less disposed to accept the treatment than she otherwise would have
been, in order to be entirely rational. In this sense, it might be objected, she is disposed to decline treatment.
It is a difficult issue what it means to say that a disposition is weakened (cf. Mele (2003), Ch. 7), but two
comments are worth making. First, it does not seem conceptually necessary that a person who thinks she has
an extremely strong reason to accept a medical treatment has to be less disposed to accept it merely because
she thinks she has an extremely weak reason to decline the treatment on the ground that she will not be able
to drink coffee for one minute, even if she is rational. It seems reasonable to think that a rational person can
be unaffected by her various normative judgments that concern such petty considerations. Second, it can be
argued that the kind of phenomena appealed to in this objection is consistent with my argument. The fact that
a person is less disposed to accept the treatment than she otherwise would have been does not seem to entail
that she is disposed to decline the treatment. Indeed, the latter seems to be a disposition to perform a quite
different action.
18
This kind of cases is presumably controversial due to the distinction between object-given and state-given
reasons. Some authors deny that there are such reasons (e.g. Parfit (2001), pp. 21–23), whereas other accept
their existence (e.g. Rabinowicz and Rønnow-Rasmussen (2004), pp. 411–414).
19
If anything, it seems that she is rational in case she is not motivated in that way; however, the latter claim
is not needed for the argument. It might be the case that a rational person would not in fact give in to her fear
9
There is also an entirely distinct kind of cases which provide a difficulty for (2).
Assume that a person thinks that she has a reason to  and a reason to ψ. Assume further
that she thinks that her reason to  is exactly as strong as her reason to ψ. Lastly, assume
that she thinks that  and ψ are practically incompatible. According to (2) it is conceptually
necessary that she is motivated to perform both  and ψ, in order to be entirely rational.
However, it is not obvious that this is correct, especially not if the actions in question are
inconsequential. Consider a person who thinks about what to eat for dinner. She thinks that
her reason to have pasta is a strong as her reason to have a pizza. We might also assume
that she thinks that which alternative she chooses is of very little importance, but that she
cannot have both. Now, we can admit that she might be motivated to have both pizza and
pasta even if she is entirely rational, but it is questionable whether it is conceptually
necessary that she has to be thus motivated in order to be entirely rational and that she
needs to be irrational to the extent she fails to be motivated in that manner.20 In such cases,
it seems that, from the perspective of rationality, the person is free to lack one of the
motivations and merely be led by her pure inclinations.
Thus, in the different types of cases we have considered it seems plausible to say that
it is rationally permissible for the person in question to be motivated to perform the action
in question, but it does not seem that she needs to be rationally required to be thus
motivated and that she has to be irrational to the extent she is not motivated in that way.21
Moreover, this conclusion finds support in our concept of rationality which also internalists
appeal to. The term ‘irrational’ is used to categorize various forms of failure in mental
functioning which is shown by the various cases standardly used to exemplify it, such as
apathy, depression, emotional disturbance, addiction, and compulsion. It is such cases that
advocates of rationalist internalism appeal to when they introduce the notion of rationality
to save internalism. Moreover, it is such cases that justify our criticism against people
when we consider them as being mentally malfunctioning in the relevant respect. In
ordinary talk we often use other terms to signify behaviour and motivation which is due to
such failure, e.g. ‘confused’, ‘doesn’t make sense’, ‘strange’, ‘incomprehensible’, or
of medical treatments or even have such a fear. However, it does not seem to be part of the concept of being
rational that one knows that such situations do not occur.
20
This view also suggests that in such a case the rational person has to be equally motivated to  and ψ;
however, this might cause problems when she is about to form an intention to either  or ψ.
21
Cf. Wedgwood (2002), p. 349.
10
‘weird’.22 Now, our responses to the examples above provide evidence that we do not think
that it is conceptually necessary that the people in the kind of cases we have considered are
mentally malfunctioning in the relevant way. As a consequence, we do not seem think that
they need to be irrational. Accordingly, we would not maintain that a person who denies
that the people in these cases are irrational is not linguistically competent as regards what it
means to say that a person thinks that she has reason to perform an action.
Thus far, I have argued that there are some reasons to think that normative internalism
in (2) is false in which case there is reason to doubt rationalist internalism in (3).
Advocates of (2) might want to object that whether a claim is conceptually necessary need
not be obvious and that this only can be determined by a careful philosophical analysis
which requires much reflection.23 They can then argue that it might seem possible to think
that a person is entirely rational even if she does not fulfil the requirement in (2) although
this thesis actually is true.
In reply, it should first be pointed out that even if considerations of such examples do
not demonstrate that (2) does not hold, they provide some evidence that it does not, since
they indicate that competent language users need not adhere to it. It seems fair to claim
that, in view of these responses, the burden of proof is on defenders of (2).
However, what is most important for the present inquiry is that even if (2) is true, it is
not an apparent conceptual truth which in turn means that rationalist internalism in (3) is
unable to explain our relevant linguistic intuitions concerning the connection between
moral language and motivation. To see this, assume for the sake of the argument that (2) is
correct. Nevertheless, our responses to the kind of cases we have considered above show
that it is not apparent to us as competent language users that the connection between
normative judgments, rationality, and motivation stated in (2) is the case; on the contrary,
our responses suggest that it is quite reasonable for us to doubt this view. However, in
order for (2) to be a part of an explanation of our straightforward responses to various
cases, it would have to be in the foreground of our conception of this connection, at least to
the extent that we should not outrightly question it. This means that (3), which is logically
dependent on (2), cannot explain why we find it so puzzling that a person who utters that
an action is right is not motivated to do it. In other words: If it is not apparent to us that a
person who thinks that she has a reason to perform an action needs to lack in rationality
22
23
Cf. Gert (2004), p. 143.
See e.g. Smith (1994), pp. 35–39.
11
unless she is motivated to do it, it cannot be apparent to us that a person who thinks that it
is right to perform an action, where this is assumed to entail that she thinks that she has a
reason to do it, needs to lack in rationality unless she is not motivated in that way.
Likewise, if it reasonable for us to doubt the former, it is reasonable for us to doubt the
latter. As a consequence, (3) cannot explain our straightforward linguistic intuition that the
cases internalists appeal to are so puzzling.
In view of these arguments, the best advocates of rationalist internalism can do is
presumably to adopt a weaker and more plausible version of normative internalism than
the one in (2). The strongest version of this view which is consistent with our responses to
the different types of cases we considered above is the following:
(2*) Normative internalism: It is conceptually necessary that, for any actions  and ψ
and any rational person S, if S judges that she has a normative reason to , then she is
motivated to , if she does not judge that she has a normative reason to ψ which is
stronger than, or at least as strong as, her normative reason to , where  and ψ are
practically incompatible.
It should be clear that (2*) corresponds to the plausible view mentioned above, namely that
it is conceptually necessary that if a person thinks that she has strongest reason to , then
she is motivated to  in so far as she is rational. It should also be clear that (2*) is not
vulnerable to the difficulties which trouble (2). It does not entail that in case a person
thinks that she has strongest reason to , but a much weaker reason to ψ, where  and ψ are
practically incompatible, she needs to be motivated to ψ in order to be entirely rational.
And it does not entail that in case a person thinks that she has a reason to  that is as strong
as her reason to ψ, where  and ψ are practically incompatible, she needs to be motivated
to both  and ψ in order to be entirely rational.24
Now, rationalism in (1) together with the new version of normative internalism in
(2*) entails the following version of rationalist internalism:
24
I do not want to commit myself to any of the claims above. In particular, I think Joshua Gert has provided
convincing arguments for a distinction between justificatory and requiring reasons which constitute problems
for any general view according to which there is a necessary connection between normative judgements,
rationality, and motivation; see Gert (2004), esp. Ch. 8.
12
(3*) Rationalist internalism: It is conceptually necessary that for any actions  and ψ
any rational person S, if S judges that it is morally right that she s, then she is
motivated to , if she does not judge that she has a normative reason to ψ which is
stronger than, or at least as strong as, her normative reason to , where  and ψ are
practically incompatible.
Thus, the advantages with adopting (2*) come at a certain prize: a version of rationalist
internalism which is considerably weaker than (3).25
We might end this section by considering an argument against my reasoning. David
Copp has directed a similar objection as the first one above against Michael Smith’s
version of normative internalism.26 Smith admits that, on his view, a rational person needs
to have a desire corresponding to each of her normative judgments and he does not deny
that this view has counterintuitive consequences. However, he argues that on this view it is
possible to explain why the strength of a rational person’s desires to perform actions
accords with the strength of her normative beliefs about these actions, i.e. her beliefs as
regards how strong reasons she has to perform these actions.27
In reply, it might first be wondered why one should want to explain the strength of a
rational person’s normative judgments in terms of the strength of her desires in this way in
case one is not already attracted to normative internalism. After all, there are alternative
explanations.28 In addition, it seems possible to explain the strength of a rational person’s
normative judgments in terms of the strength of her desires even if each of her normative
judgments does not correspond to an existing desire. Assume that each of a rational
25
It might be objected that rationalist internalists can avoid the argument above by modifying rationalism in
(1) instead of modifying normative internalism in (2). The best way to spell out this suggestion seems to be
in terms ought judgments. This line of thought can be formulated as follows: (O1) Rationalism: It is
conceptually necessary that, for any action  and person S, if S judges that she morally ought to , then she
judges that she has strongest normative reason to . (O2) Normative Internalism: It is conceptually necessary
that, for any action  and rational person S, if S judges that she has strongest normative reason to , then she
is motivated to . (O3) Rationalist Internalism: It is conceptually necessary that, for any action  and rational
person S, if S judges that she morally ought to , then she is motivated to . Admittedly, (O2) is more
plausible than (2). On the other hand, (O1) is much more implausible than (1). After all, it seems evident that
a person may think that she morally ought to  but at the same time think that she has a stronger reason to ψ.
Moreover, it might be argued that (2) is more plausible than (O1) in which case (3) might seem more
appealing than (O3) even if the former view is not ultimately correct either.
26
Copp (1997), p. 46. Cf. Gert (2008), pp. 16–17.
27
Smith (1997), p. 92.
28
Copp (1997), p. 46. For instance, Copp argues that a rational person can deliberate about what to do, reach
a conclusion as regards what she has strongest reason to do and then form a desire to perform that action.
Gert argues for a version of normative externalism which provides another explanation: Gert (2004), Ch. 8.
13
person’s judgment to the effect that she has a reason to perform a certain action is
associated with a desire of a particular strength. In that case the strength of normative
judgments accords with the strength of the desires with which they are associated and we
can provide the explanation Smith requires. However, this view does not mean that it is
conceptually necessary that a rational person actually has the desire with which her reason
is associated. It is compatible with (2*) according to which it is conceptually necessary that
she is motivated to perform the action if she does not also judge that she has a stronger
reason to do another action. In the same vein, R. Jay Wallace observes that in some
circumstances ‘the most rationality might require of me would be a kind of conditional
disposition to action’.29 Hence, the arguments above seem compatible with the existence of
the explanation Smith requires. Furthermore, it is possible to explain the strength of a
rational person’s normative judgments in terms of the strength of her existing desires in a
way which does not entail (2). Assume that there is a correspondence of the kind Smith
maintains: in a rational person each normative judgment has a counterpart in an existing
desire. However, it might be the case that such correspondence holds, and even that it is
necessary, but that it is not conceptually necessary. This suggests that the strength of a
rational person’s normative judgments can be explained in terms of the strength of her
existing desires even if (2) does not hold. Finally, the basic point of my argument is that
(2) cannot be part of an explanation of our linguistic intuitions as regards the relevant
cases, not that this claim does not hold. As a consequence, in so far as the plausibility of
rationalist internalism in (3) rests on (2), it cannot explain our relevant linguistic intuitions.
5. An Internalist Dilemma
In the last section, I argued that there are reasons to think that normative internalism in (2)
is false and that even if it is not, it is explanatory impotent in a way which implies that
rationalist internalism in (3) is unable to explain our linguistic intuitions as regards moral
language and motivation. Therefore, rationalists internalists are led to adopt (2*) which
means that they have to adopt (3*). However, in this section we will see that (3*) is
Wallace (2006), p. 188. This is not only compatible with the possibility that the person’s desire to  might
be overridden in case she thinks that she has a stronger reason—and hence a stronger desire—to ψ. It is also
compatible with the possibility that the desire to  might be annihilated in such a case. It might also be
argued that there are situations where a person is more rational in case she lacks motivation to do what she
thinks she has reasons to do; see Arpaly (2000), pp. 488–513.
29
14
incapable of explaining our relevant linguistic intuitions and that, as a consequence,
internalism faces a dilemma. There are two basic arguments for this view.
First, (3*) is not sufficient to explain our linguistic intuitions as regards the
connection between moral language and motivation. As we have seen, it is generally
acknowledged that if a person utters a sentence to the effect that it is right to , e.g.
‘Actually, it’s right to give some money to those who need it more than I do’, without
being motivated to , we find this puzzling. According to (3*), it is conceptually necessary
that if a person thinks that it is morally right that she s, she is motivated to  on two
conditions: first, that she is rational and, second, that she does not think that she has a
conflicting reason to ψ which is stronger than, or at least as strong as, her reason to . This
means that (3*) is able to explain why we find it puzzling if a person were to utter a
sentence to the effect that it is right to , without being motivated to , only in case we are
in the position to assume that these two conditions are fulfilled. However, we would
clearly find the mentioned case perplexing even if we do not have any views as to whether
the second condition is satisfied. That is: in case a person utters a sentence to the effect that
it is right to , but is not motivated to , we would find this puzzling even if we do not
have any opinion at all as to whether she thinks that she has a stronger, or at least as strong
reason, to do another action ψ. Moreover, on the view under consideration we cannot
presume that a person who utters a sentence to the effect that it is right to  does not think
that she has a stronger, or at least as strong reason, to ψ. As a result, (3*) is unable to
explain our response.
Second, (3*) is not necessary to explain our linguistic intuitions as regards the
connection between moral language and motivation. Assume again that a person utters a
sentence to the effect that it is right to , such as ‘Actually, it’s right to give some money to
those who need it more than I do’, but has no motivation to . However, assume now that
we also know that she does think that she has reason to perform another action ψ which is
stronger than her reason to . We may now ask if we would find this case puzzling. I think
we would. We would find it puzzling that a person makes such an utterance even if we
know that the case has a feature which, according to the view under consideration, would
account for the fact that she is not motivated in accordance with it. We would presumably
wonder why she says that it is right to  when she is not motivated to ; similarly, we
15
would feel that her utterance lacks any appropriate point in that case.30 The version of
rationalist internalism in (3*) seems incapable to explain this. According to this claim, the
case in question has a feature—the person thinking that she has stronger reasons to ψ than
to —which would account for the absence of her motivation to . As we know that the
case has this feature, this fact would justify the response that the case under consideration
is not particularly strange. However, since we still find it puzzling that she utters a
sentence to the effect that it is right to , (3*) is unable to account for our response.
We can now summarize the internalist dilemma. On the one hand, if advocates of
internalism about moral judgments and motivation adopt strong internalism, they can
explain our linguistic intuitions in relation to moral language and motivation, but fail to
explain why a person might hold a moral judgment without being motivated. On the other
hand, if they adopt rationalist internalism, their view is susceptible to another difficulty. In
fact, they face yet a dilemma. If they adopt normative internalism in (2) so as to defend
rationalist internalism in (3), they accept a version of normative rationalism which there
are reasons to doubt and which cannot be part of an explanation of our responses to the
kind of cases internalists appeal to. As a result, the original version of rationalist
internalism, (3), cannot explain our conception of moral language and motivation.
However, if they adopt a more plausible version of normative internalism, (2*), they have
to accept a weaker version of rationalist internalism, (3*), which is incapable of explaining
this conception.
It is worth stressing that the arguments above do not entail that rationalist internalism
in (3*) is false. What they show, if correct, is that this view is incapable to explain our
conception of moral language and motivation. However, this conclusion has significant
implications for our reasons to accept internalism about moral judgments and motivation.
As I observed above, a main reason to adopt this kind of internalism is that it is able to
provide such an explanation; indeed, it is commonly presumed to be the reason for this
view. Importantly, other views, such as rationalism or normative internalism, are often
defended in order to establish rationalist internalism whose basic virtue is that it provides
the mentioned explanation. We have seen, however, that there are grounds to think that
other versions of internalism than rationalist internalism in (3*) are vulnerable to various
difficulties. We have also seen that (3*) is the only version of internalism which both is
30
At least, this is the case unless she explains why she makes the utterance in spite of not being accordingly
motivated.
16
able to account for the fact that people might not be motivated in accordance with their
moral views and able to avoid the arguments I put forward in the last section. However, in
this section I have argued that this version of internalism is unable to explain our
conception of moral language and motivation. It follows that we have grounds to be
sceptical about internalism as regards moral judgments and motivation in general.
6. Reconsidering Strong Internalism
Is there any way in which internalists can avoid the dilemma? One way would be to
maintain another version of weak internalism than rationalist internalism. This possibility
cannot be completely excluded, but it is difficult to see that condition C can be specified in
another way which does not run the risk of trivializing the internalist claim.
Another way for internalists to avoid the dilemma would be to uphold strong
internalism but account for cases where we think that a person is not motivated to act in
accordance with her moral judgement in another way than weak internalism. Strong
internalists might try to do so by stressing the distinction between motivation understood
as a disposition and an occurrent motivation. On this suggestion, a person who judges that
it is right that she s, but who does not appear to be motivated to , is actually motivated to
do the action in virtue of having a disposition to . As a consequence of the mental
condition she finds herself in, her disposition to  is not activated and she does therefore
not have any occurrent motivation to  which manifests itself in her behaviour.
This suggestion deserves thorough discussion, but two considerations should be
mentioned.31 First, it is doubtful whether it is able to account for all relevant cases. Strong
internalists agree, as we have seen, that there might be cases where a person judges that it
is right that she s, but does not have any occurrent motivation to , because of the mental
condition she finds herself in. They also have to agree that a person’s mental condition
may influence her dispositions to act, not only her occurrent motivation. This also holds for
moral judgments. Consider a person who thinks that it would be right of her to help her son
stop taking drugs but is severely depressed. Clearly, her mental condition might have the
effect that she becomes less disposed to help her son, not only that her disposition fails to
get activated into an occurrent motivation. Now, given these assumptions, advocates of the
present suggestion owe us an explanation of why we should rule out the possibility that
31
For related suggestions, see e.g. Gibbard (2003), p. 154. Cf. Blackburn (1998), p. 61–63. I argue against
this defence of internalism more fully in REF.
17
there are cases where a person judges that it is right that she s, but where her mental
condition makes her disposition to  go out of existence. Indeed, if (i) there are cases
where a person judges that it is right that she s but does not have any occurrent motivation
to  because of her mental condition, and (ii) people’s dispositions to act in accordance
with their moral judgments can be severely undermined by their mental conditions, a
plausible explanation of (i) seems to be that the person in question does judge that it is
right that she s but is not disposed to  because of her mental condition.
Second, the suggestion under consideration runs the risk of drawing an arbitrary line
between, on the one hand, cases where a person judges that it is right that she s and is
motivated to  and, on the other hand, cases where a person does not judge that it is right
that she s because she is not motivated to . Consider someone who holds judges that it is
right that she s but who is severely depressed. As we have seen, the proponents of the
suggestion at issue have to accept that her mental condition may weaken her disposition to
. Having admitted that much, they also have to admit that her mental condition might
have the effect that her disposition to  becomes extremely weak. However, they are
committed to the view that as soon as her depression gets so severe that it extinguishes her
disposition to  altogether, she does not judge that is right that she s any longer. But why
think that she stops holding the moral judgement precisely when her disposition has gone
out of existence? Why at this particular point—why not before or after? In lack of an
explanation, it seems arbitrary to draw the line there and not somewhere else. It is hard to
se that proponents of strong internalism can provide any explanation that does not run the
risk of becoming question begging.
7. An Externalist Solution
Lastly I would like to briefly argue that externalism is able to avoid the dilemma which
internalism faces. Externalism entails the denial of internalism. In particular, on this view it
is not the case that a sentence to the effect that it is morally right to  expresses, merely in
virtue of its meaning, a judgment such that if a person holds it, she is motivated to , not
even if she satisfies a non-trivial condition such as rationality.
It is generally thought that externalism has difficulties to explain our conception of
the connection between moral language and motivation. However, I think it can be argued
that externalists are able to do so by employing Paul Grice’s concept of generalized
18
conversational implicature. More exactly, I think externalists should adopt the following
view:
A person’s utterance of a sentence according to which it is morally right to  carries a
generalized conversational implicature to the effect that she is motivated to .
Let us very briefly recall Grice’s notion of generalized conversational implicature.32 That
an utterance carries an implicature is mainly a result of the fact that we are presumed to
make utterances which adhere to ‘the principle of cooperation’. According to this principle,
we are to make utterances which contribute to the mutually accepted purpose of the
conversation in which we are involved. We adhere to this principle by following certain
maxims which state, for instance, that we only are to make utterances that are relevant in
relation to the purpose of the conversation. As a result of our adherence to the cooperative
principle and the maxims, an utterance of a sentence can carry information that is not part
of the conventional meaning of the sentence: a conversational implicature.33 As regards
particularized implicatures, an utterance carries such an implicature in virtue of the fact
that the context in which it is uttered has special features which are distinctive of it. As
regards generalized implicatures, an utterance carries the implicature even if the context in
question does not have any special features; especially, it does so unless there are any
special contextual features which defeat the implicature. As a result, in this case the
utterance standardly carries the implicature and it constitutes a default understanding of the
utterance. Grice points out that it might mistakenly be thought that a generalized
implicature is part of the meaning of a sentence since it is standardized and consequently
systematically accompanies utterances of the sentence throughout various contexts.
The idea that a moral utterance carries a generalized implicature to the effect that the
person who utters it is motivated is, I think, interesting as it stands. It is common in
philosophy and linguistics to appeal to the notion of generalized implicature in cases where
it is contested whether something is part of the meaning of a certain linguistic expression
or not; surprisingly, however, it has not yet been applied to the debate between internalism
and externalism. Moreover, I defend this view in more detail elsewhere and in the present
32
Grice (1989 (1975)), pp. 22–40. For alternative pragmatic accounts, see Copp (2001), pp. 1–43; Copp
(2009), pp. 167–202, and Finlay (2004), pp. 205–223. See also Boisvert (2008), 169–203. I argue that my
account is preferable to these views in REF.
33
From now on, I omit ‘conversational’.
19
paper I am chiefly interested in how it solves the internalist dilemma.34 For these reasons, I
will merely indicate the bare outline of it here.
To simplify, the suggestion is based on three simple assumptions. First, it is obvious
that a sentence to the effect that it is morally right to  entails that there is a moral
reason—a reason according to morality—to .35 Second, it is plausible to assume that it is
a general purpose of moral conversations to influence actions; moral conversations
generally have as a purpose to contribute to people performing, or not performing, certain
actions. Moreover, it is a basic fact about human psychology that one general way to affect
people’s behaviour is to let them know about our attitudes towards it.
Consider a person who is involved in a moral conversation and utters a sentence to
the effect that it is right to , e.g. ‘Actually, it’s right to give some money to those who
need it more than I do’. In consideration of the purpose of such conversations, it is
plausible to assume that she wants that  is performed. The basic reason is that it does not
seem relevant to utter a sentence which entails that there is a moral reason to  in a moral
conversation where a general purpose is to influence actions unless one wants that  is
performed. In other words, given that a general purpose of moral conversations is to
influence actions and a person utters a sentence which entails that there is moral reason to
, we understand her as wanting that  is performed, since if she does not want this, her
utterance lacks in relevance. Such an utterance would simply lack point in consideration of
the purpose of the discourse. Hence, the person’s utterance implicates that she wants that 
is performed. Moreover, this is a matter of a generalized implicature. As I mentioned, it is
a general purpose of moral conversations to influence actions; that is, it is a purpose which
pervades throughout various contexts independently of their special features. As a
consequence, an utterance of sentence to the effect that it is right to  standardly carries
this implicature and it qualifies therefore as a generalized implicature.
As there is a standardized connection between utterances of a sentence saying that it
is right to  and the mentioned attitude, we naturally associate such a sentence with this
attitude and need not to work out the implicature in order to recognize that utterances of
34
REF.
Note that this does not mean that such a sentence entails that there is a normative reason, in the sense
considered above, to . Hence, this view does not entail rationalism in (1).
35
20
the sentence convey the attitude.36 In other words, when we encounter an utterance of such
a sentence we can immediately infer that the person has the attitude in question without
reflecting on the particular circumstances of the utterance. As a result, we might come to
think that there is a conceptually necessary connection between a sentence to the effect that
it is right to  and a motivating attitude in the manner internalists maintain.37
It might be objected that a moral utterance can convey that a person has a motivating
attitude towards an action even in a context where it is not a purpose of the conversation to
influence action.38 The present view is however able to explain this. As we have seen, a
person’s utterance of a sentence according to which it is right to  carries a generalized
implicature to the effect that she wants  to be performed. As a consequence, there is a
strong background presumption that someone who uses such a sentence does want  to be
performed. This means that a person’s utterance of such a sentence carries this implicature
unless she cancels it by making an additional utterance to the effect that the background
assumption does no hold in that particular case or the context cancels the implicature by
containing information to this effect.
When a person utters a sentence to the effect that it is right to , we presume that she
thinks that it is right that she s unless another utterance of hers she or the context indicates
otherwise; we assume she does not make an exception for herself unless something
indicates so.39 As a result, a person’s utterance of such a sentence carries a generalized
implicature to the effect that she wants that she s and, hence, is motivated to , unless she
makes an additional utterance, or the context contains information, which cancels the
implicature.40
36
This is a phenomenon commonly observed in linguistics; see e.g. Morgan (1998 (1978)), pp. 651–652. In
REF I demonstrate that the generalized implicature under consideration can be worked out and that it is
cancellable.
37
Cf. Grice (1989 (1975)), pp. 37–38.
38
For example, in certain circumstances a moral conversation might be entirely theoretical.
39
This can be explained by the fact that an utterance of such a sentence carries a generalized implicature to
the effect that it is right that she, the speaker, s.
40
Another objection against my proposal is that it merely explains that we find it puzzling that a person
utters a sentence to the effect that it is right to  without being motivated to , not that we find it puzzling
that a person thinks (judges) that it is right that she s without being accordingly motivated. This is a
complicated issue that I cannot do justice to here, but I would like to three points. First, internalists
standardly appeal to the first kind of cases, but very seldom to the second kind of cases, which indicate that
the former are most in want of explanation. Second, there are various plausible explanations as to why we
find the second kind of cases puzzling which is consistent with externalism as I understand it. For example, it
can be partially explained in terms of moral upbringing and fundamental human characteristics, such as
sympathy and need to cooperate with others. (Cf. Svavarsdóttir (1999), pp. 183–187, and Shafer-Landau
(2003), pp. 159–160.) It should also be observed that the present suggestion is compatible with a version of
21
We are now in the position to see how this suggestion makes it possible for
externalism to avoid the dilemma which troubles internalism.
First, this view is able to explain our linguistic intuitions as regards moral language
and motivation which is considered to be the main argument for internalism. In case a
person utters a sentence to the effect that it is right to , but lacks motivation to , we
would find this puzzling. Internalism explains this by claiming that such a sentence
expresses, in virtue of its meaning, an intrinsically motivating moral judgment. On the
view suggested here, externalism explains our response by claiming that it is the utterance
of such a sentence which carries a generalized implicature to the effect that the person in
question is motivated to . As it is a matter of a generalized implicature, and such an
utterance standardly carries the implicature, there is a strong presumption that a person
who makes such an utterance is motivated to . Consequently, in case a person is not
motivated to  in spite of making such an utterance, we find this puzzling. We can justify
this response further by referring to the fact that she has not adhered to the cooperative
principle. In particular, she has violated this principle since she has made an utterance
which is not relevant in consideration of the general purpose of moral conversations. Thus,
this view avoids the part of the internalist dilemma which troubles weak internalism.
Second, this view is able to account for cases where a person holds a moral judgment
but is not motivated to act in accordance with it because of her mental condition.
According to externalism, it is not conceptually necessary that if a person judges that it is
right that she s, she is motivated to . Consequently, a person who suffers from a certain
mental condition, such as severe depression, can judge that it is right that she s without
being motivated to .41 Thus, this view avoids the part of the internalist dilemma which
troubles strong internalism.
internalism according to which there is a necessary connection between the practice of making moral
judgments in a community and motivation, but not between each individual moral judgment and motivation.
There are reasons to think that this view is able to explain the second type of cases. (See Tresan (2004), pp.
143–165, and Bedke (2009), pp. 189–209.) Third, it does not seem farfetched to assume that our moral
thinking is influenced by the fact that our moral utterances carry the generalized implicature I identified
above. Indeed, this seems to be common phenomenon as regards other instances of generalized implicatures.
(Cf. Williamson (2009), pp. 1–37.) In particular, in view of the fact that this connection between uses of
moral language and motivation is standardized, it might have made us think in accordance with it.
41
On externalism, it is not the case that a person’s judgement to the effect that it is right that she s is
sufficient by itself for her to be motivated to . What motivates her is a motivational state which is external to
the judgement. If she fails to be motivated to  because of a certain mental condition, such as severe
depression, the explanation might be that her condition has defeated this external motivational state.
22
8. The Externalist Solution and Rational Internalism
It should also be observed that the externalist account I propose is not susceptible to my
objections against rationalist internalism in Section 5. This solution clearly does not have
the difficulties afflicting rationalist internalism in (3) since it is not committed to normative
internalism in (2). It is especially important to see that the present account does not have
the difficulties afflicting rationalist internalism in (3*) as this it is the most plausible
version of this claim.
Let us start by noting that it does not have the first difficulty of (3*). According to
(3*), it is conceptually necessary that if a person thinks that it is right that she s, she is
motivated to  on condition she is rational and does not think that she has a conflicting
reason to ψ which is stronger than, or at least as strong as, her reason to . However, on the
view suggested here there is no conceptually necessary connection between moral
judgments and motivation of this kind. This means that our conception of the connection
between moral utterances and motivation is not constrained by the person in question
fulfilling the mentioned conditions. It is a person’s utterance of a sentence to the effect that
it is right to  which carries a generalized implicature which conveys that she is motivated
to . This means in turn that in case a person makes such an utterance, but is not motivated
to , we might find this puzzling even if we do not have any views at all as to whether she
thinks she has a stronger, or at least as strong, reason to perform to another action ψ. The
explanation is that such an utterance carries the mentioned implicature irrespective of
whether she satisfies the mentioned conditions or not.
Let us next note that it does not have the second difficulty of (3*). In case a person
utters a sentence to the effect that it is right to , but is not motivated to , and we know
that she does think she has stronger reasons to ψ, we might still find it puzzling that she
makes such an utterance. The reason is that her utterance carries this implicature even if
she does not satisfy the mentioned conditions. As I suggested above, we would presumably
wonder why she makes the utterance in spite of not being motivated to  and feel that her
utterance lacks point in such a situation. The present account readily explains this by
referring to the fact that her utterance does not conform to a general purpose of moral
conversations.
The fact that generalized implicatures are cancellable is relevant to what I said above.
There are basically two ways in which an implicature can be cancelled: the person who
23
makes the original utterance can annul the implicature by making an additional utterance
or the context of the utterance is such that it annuls the implicature.42 We have seen that
there might be cases where a person who thinks that it is right that she s fails to be
motivated to  as a result of her mental condition. It is therefore plausible to think that
there might be cases where a person utters a sentence to the effect that it is right to  but
where we do not assume that she is motivated to  because we believe that she finds
herself in a certain mental condition. The fact that generalized implicatures are cancellable
is able to account for this fact. Thus, the generalized implicature which conveys that the
speaker is accordingly motivated can be cancelled in case the speaker, or the context of her
utterance, indicates that she finds herself in a mental condition which makes it the case that
is she is not motivated to . Suppose that a person utters a sentence saying that it is right to
 but then adds an additional utterance to the effect that she feels, say, very depressed.
Alternatively, suppose that a person makes such a moral utterance in a context in which it
is generally known that she is, say, severely depressed. Moreover, suppose that in either
case it is commonly assumed that her mental condition relates to  in such a way that it is
reasonable to believe that it can defeat her motivation to . In either case, the generalized
implicature in question can be cancelled, in which case we do not assume that she is
motivated to .
9. Conclusion
In this paper, I have argued that internalism about moral judgments and motivation faces a
dilemma. Strong internalism is able to explain our linguistic intuitions as regards the
connection between moral language and motivation, but fails to account for cases where
moral judgment and motivation take separate paths. Conversely, the most plausible version
of weak internalism, rationalist internalism, is able to account for such cases, but fails to
explain our conception of moral language and motivation. I also argued that externalism
should maintain that people’s moral utterances carry a generalized conversational
implicature to the effect that they are accordingly motivated. This view, I argued, is able to
explain both the notions that internalism fails to account for. As a consequence, there is
reason to think that externalism is to be preferred to internalism.
42
Grice (1989 (1975)), p. 39.
24
References
Arpaly, Nomy (2000), ‘On Acting Rationally against One’s Best Judgment’, Ethics, Vol. 110, pp. 488–513.
Audi, Robert (1997), ‘Moral Judgment and Reasons for Action’, Ethics and Practical Reason, ed. by Garrett
Cullity and Berys Gaut, Oxford: Clarendon Press, pp. 125–160.
Bedke, M. S. (2009), ‘Moral Judgment Purposivism: Saving Internalism from Amoralism’, Philosophical
Studies, vol. 144, pp. 189–209.
Blackburn, Simon (1998), Ruling Passions, Oxford: Clarendon Press.
Boisvert, Daniel (2008), ‘Expressive-Assertivism’, Pacific Philosophical Quarterly, Vol. 89, pp. 169–203.
Brink, David (1989), Moral Realism and the Foundations of Ethics, Cambridge: Cambridge University Press.
Broome, John (2007), ‘Does Rationality Consist in Responding Correctly to Reasons?’, Journal of Moral
Philosophy, Vol. 4, pp. 349–374.
Choi, Sungho (2006), ‘The Simple vs. Reformed Conditional Analysis of Dispostions’, Synthese, Vol. 148,
pp. 369–379.
Copp, David (1997), ‘Belief, Reason, and Motivation: Michael Smith’s The Moral Problem’, Ethics, Vol.
108, pp. 33–54.
Copp, David (2001), ‘Realist-Expressivism: A Neglected Option for Moral Realism’, Social Philosophy and
Policy, Vol. 18, pp. 1–43.
Copp, David (2009), ‘Realist-expressivism and Conventional Implicature’, Oxford Studies in Metaethics, ed.
by R. Shafer-Landau, Vol. 4, Oxford: Oxford University Press (forthcoming).
Dancy, Jonathan (1993), Moral Reasons, Oxford: Blackwell.
Darwall, Stephen (1990), ‘Autonomist Internalism and the Justification of Morals’, Noûs, Vol. 24, pp. 257–
267.
Dreier, James, (1990), ‘Internalism and Speaker Relativism’, Ethics, Vol. 101, pp. 6–25.
Finlay, Stephen (2004), ‘The Conversational Practicality of Value Judgment’, The Journal of Ethics, Vol. 8,
pp. 205–223.
Foot, Philippa (1978), ‘Morality as a System of Hypothetical Imperatives’, Virtues and Vices, Oxford:
Blackwell, pp. 157–173.
Frankena, William (1976 (1958)), ‘Obligation and Motivation in Recent Moral Philosophy’, Perspectives on
Morality. Essays by William K. Frankena, ed. By K. E. Goodpaster, Notre Dame: University of Notre
Dame Press, pp. 12–23.
Gert, Joshua (2004), Brute Rationality, Cambridge: Cambridge University Press.
Gert, Joshua (2008), ‘Michael Smith and the Rationality of Immoral Action’, The Journal of Ethics, Vol. 12,
pp. 1–23.
Gibbard, Allan (2003), Thinking How to Live, Cambridge, Mass.: Harvard University Press.
Grice, Paul (1989 (1975)), ‘Logic and Conversation’, Studies in the Ways of Words, Cambridge, Mass.:
Harvard University Press, pp. 22–40.
Johnson, Mark (1992), ‘How to Speak of Colors’, Philosophical Studies, Vol. 68, pp. 221–263.
Joyce, Richard (2003), The Myth of Morality, Cambridge: Cambridge University Press.
Korsgaard, Christine (1986), ‘Skepticism About Practical Reason’, The Journal of Philosophy, Vol.
LXXXIII, pp. 311–334.
Lenman, James (1999), ‘The Externalist and the Amoralist’, Philosophia, Vol. 27, pp. 441–457.
Lewis, David (1999), ‘Finkish Dispositions’, Papers in Metaphysics and Epistemology, Cambridge:
Cambridge University Press, pp. 133–151.
Mele, Alfred. (2003), Motivation and Agency, New York: Oxford University Press.
Miller, Alexander (2003), An Introduction to Contemporary Metaethics, Cambridge: Polity Press.
Miller, B. C. (2008), ‘Motivational Internalism’, Philosophical Studies, Vol. 139, pp. 233–255.
Morgan, J. L. (1998 (1978)), ‘Two Types of Convention in Indirect Speech Acts’, Pragmatics, ed. by A.
Kasher, Vol. IV, London: Routledge, pp. 639–657.
Parfit, Derek (2001), ‘Rationality and Reasons’, Exploring Practical Philosophy. From Action to Values, ed.
by Jonas Josefsson, Björn Petersson and Toni Rønnow-Rasmussen, Hampshire: Ashgate, pp. 17–39.
Persson, Ingmar (2005), The Retreat of Reason, Oxford: Oxford University Press.
Rabinowicz, Wlodek and Toni Ronnøw-Rasmussen (2004), ‘The Strike of the Demon: On Fitting Proattitudes and Value’, Ethics, Vol. 114, pp. 391–423.
Ridge, Michael (2007), ‘Ecumenical Expressivism: The Best of Both Worlds?’, Oxford Studies in
Metaethics, ed. by Russ Shafer-Landau, Vol. 2, pp. 51–76.
Roojen, Mark van (2002), ‘Humean and Anti-Humean Internalism About Moral Judgments’, Philosophy
and Phenomenological Research, Vol. 65, pp. 26–49.
Roskies, Adina (2003), ‘Are Ethical Judgments Intrinsically Motivational? Lessons from “Acquired
Sociopathy”’, Philosophical Psychology, Vol. 16, pp. 51–66.
25
Sayre-McCord, Geoffrey (1997), ‘The Metaethical Problem’, Ethics, Vol. 108, pp. 55–83.
Scanlon, T. M. (1998), What We Owe to Each Other, Cambridge, Mass.: Harvard University Press.
Setiya, Kieran (2004), ‘Against Internalism’, Noûs, Vol. 38, pp. 266–298
Shafer-Landau, Russ (2003), Moral Realism. A Defence, Oxford: Clarendon Press.
Sinnott-Armstrong, Walter (1993), ‘Some Problems for Gibbard’s Norm-Expressivism’, Philosophical
Studies, Vol. 69, pp. 297–313.
Smith, Michael (1994), The Moral Problem, Oxford: Blackwell.
Smith, Michael (1997), ‘In Defense of The Moral Problem’, Ethics, Vol. 108, pp. 84–117.
Stevenson, Charles (1944), Ethics and Language, New Haven: Yale University Press.
Stocker, Michael (1979), ‘Desiring the Bad: An Essay in Moral Psychology’, The Journal of Philosophy,
Vol. LXXVI, pp. 738–753.
Svavarsdóttir, Sigrún (1999), ‘Moral Cognitivism and Motivation’, The Philosophical Review, Vol. 108, pp.
161–219.
Tresan, Jon, ‘De Dicto Internalist Cognitivism’, Noûs, Vol. 40, pp. 143–165.
Wallace, R. Jay (2006), ‘Moral Motivation’, Contemporary Debates in Moral Theory, ed. by James Dreier,
Oxford: Blackwell, pp. 182–196.
Wedgwood, Ralph (2002), ‘Practical Reason and Desire’, Australasian Journal of Philosophy, Vol. 80, pp.
345–358.
Wedgwood, Ralph (2007), The Nature of Normativity, Oxford: Oxford University Press.
Williamson, Timothy (2009), ‘Reference, Inference and the Semantics of Pejoratives’, The Philosophy of
David Kaplan, ed. by J. Almog and P. Leonardi, Oxford: Oxford University Press (forthcoming).
26