The Pragmatics of Moral Motivation

Abstract

One of the most prevalent and influential assumptions in metaethics is that our conception of the relation between moral language and motivation provides strong support to internalism about moral judgments. In the present paper, I argue that this supposition is unfounded. Our responses to the type of thought experiments that internalists employ do not lend confirmation to this view to the extent they are assumed to do. In particular, they are as readily explained by an externalist view according to which there is a pragmatic and standardized connection between moral utterances and motivation. The pragmatic account I propose states that a person’s utterance of a sentence according to which she ought to ϕ conveys two things: the sentence expresses, in virtue of its conventional meaning, the belief that she ought to ϕ, and her utterance carries a generalized conversational implicature to the effect that she is motivated to ϕ. This view also makes it possible to defend cognitivism against a well-known internalist argument.

Introduction

In metaethics, we are frequently asked to imagine people who employ moral language but lack the corresponding moral motivation. It is generally presumed that our responses to such thought experiments provide strong support for internalism, the view that there is a necessary and nontrivial connection between moral judgments and motivation. Correspondingly, it is generally presumed that they constitute a problem for externalism, the view that no such connection holds. It is consequently taken for granted that internalism has an important advantage over externalism on the ground that it offers a compelling explanation of our conception of the relation between moral language and motivation, whereas externalism merely can appeal to the impression that amoralists appear conceivable. In the present paper, I argue that this presumption is unfounded. On closer inspection, our conception of the relation between moral language and motivation does not supply the foundation to internalism it is usually thought to do. In particular, I argue that it provides at least as strong support to an externalist account in terms of a pragmatic and standardized connection between moral utterances and motivation.

In the next section, I examine five dimensions of internalist claims that yield different versions of internalism. An important upshot of this survey is that it allows us to demarcate the version of internalism needed to function as a premise in an influential internalist argument against cognitivism. In Sect. 3, I examine the various types of thought experiments that internalists adduce in support of their view. Most importantly, I argue that none of our assumed responses to these cases provide convincing evidence for any of the various versions of internalism, and that all of these responses are equally explainable by an externalist and pragmatic view. The rest of the paper is devoted to defending such a view. In Sects. 4 through 5, I lay out the Dual Aspect Account with regard to moral motivation: A person’s utterance of a sentence according to which she ought to ϕ conveys both the belief that she ought to ϕ and that she is motivated to ϕ. Whereas the first is conveyed by means of the conventional meaning of the sentence, the latter is conveyed by the fact that an utterance of such a sentence carries a generalized conversational to this effect.¹ In Sect. 6, I argue that this pragmatic account can explain our responses to the various thought experiments we considered earlier. Therefore, I recommend that externalists adopt this account so as to explain our conception of the relation between moral language and motivation. The resulting view bestows cognitivists with an effective tool to uphold their claim against the internalist argument, since it makes it possible for them to adopt externalism at the same time as they can explain our conception of the relation between moral language and motivation. The general line of argument in the paper suggests that, in consideration of the fact that it has been quite hard to formulate a viable version of internalism and that this kind of claim has a number of difficulties, there is reason to think that the proposed externalist and pragmatist view is preferable to internalism.

Dimensions of Internalism

The following type of claim can be considered as a generic formulation of Moral Judgment Internalism (MJI):

Necessarily, if a person judges that she morally ought to ϕ, then she is, at least to some extent, motivated to ϕ.

We shall first see that this generic formulation can be specified and modified along five dimensions with the consequence that there are different versions of MJI.² We shall then see that the most metaethically significant version of MJI needs to be understood so as to fulfill five conditions, answering to the first claim in each dimension.³

(i) Conceptual Necessity vs. Non-Conceptual Necessity This dimension concerns how “necessary” in MJI should be understood. According to a conceptual version of MJI, “necessary” is understood as “conceptually necessary,” and, hence, as being a matter of the meaning of moral terms and sentences. According to a non-conceptual version of this view, it is understood as some other type of necessity, e.g., metaphysical necessity. In metaethics, claims about meaning are commonly understood in terms of what mental states sentences express in virtue of their conventional meaning. Think of a sentence of the type “I ought to ϕ.” Assume that a person who fully understands the meaning of the sentence asserts or accepts it. It is then plausible to assume that she has to be in a certain mental state. More exactly, it is reasonable to assume there is a minimal mental state that she needs to be in, order to it to be consistent with the meaning of the sentence for her to accept or assert it. This might be put by saying that the sentence expresses such a mental state: a “moral judgment.”⁴ Accordingly, a conceptual version of MJI would maintain that the sentence, in virtue of its conventional meaning, expresses a moral judgment such that if a person holds it, she is motivated to ϕ.⁵

(ii) Generalizable vs. Non-Generalizable This dimension concerns whether MJI can be generalized to other types of moral judgments that involve a certain moral concept (e.g., ought) than those referred to in a generic formulation of this claim. As is clear from this formulation, MJI is typically characterized in terms of a particular type of moral judgment. First, it concerns judgments to the effect that ϕing has a certain moral characteristic, e.g., that ϕing ought to be performed. Second, of these judgments it concerns self-addressed moral judgments, e.g., a person’s judgment that she ought to ϕ. Third, of these judgments it concerns those that relate to a person’s present or future actions. According to a generalizable version of MJI, this claim is generalizable to other kinds of moral judgments than those referred to in a generic formulation of this view.⁶ According to a non-generalizable version of MJI, it is only applicable to the kind of moral judgments referred to in this formulation.

(iii) Intrinsic vs. Extrinsic This dimension concerns whether motivation is constitutive of the nature of moral judgments. According to an intrinsic version of MJI, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ, and this judgment is by itself sufficient for her to be motivated to ϕ.⁷ According to an extrinsic version of MJI, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ,⁸ but the judgment is not by itself sufficient for her to be thus motivated.

We can furthermore distinguish between different types of extrinsic MJI by considering to what extent motivation is extrinsic to moral judgments. According to a weak extrinsic version of MJI, it is the nature of a person’s moral judgment in conjunction with some further fact that explains her motivation to ϕ. According to a strong extrinsic version of MJI, the nature of a person’s moral judgment does not play any part at all in an explanation as to why she is motivated to ϕ. For instance, it might be suggested that we classify a person’s judgment as a moral judgment only if she is motivated to ϕ, but that the nature of the judgment does not play any part in the explanation of her motivation.⁹

We will next consider two further dimensions of MJI. In each of these dimensions, the first version of MJI exists in both intrinsic and extrinsic variants, whereas the second version of MJI only exists in extrinsic variants.

(iv) Unconditional vs. Conditional This dimension concerns whether the necessary connection between moral judgments holds for all persons or only for those who satisfy some condition relating to their psychological state. According to an unconditional version of MJI, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ, and this holds for any person irrespective of her psychological state.¹⁰ According to a conditional version of MJI, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ if she fulfills a certain condition C with regard to her psychological state, for example, that she is rational or “normal.”¹¹, ¹² Advocates of conditional MJI think that the connection between moral judgments and motivation might collapse if a person who holds a moral judgment suffers from a mental condition such as depression, apathy or emotional disturbances. They therefore advocate conditional MJI where the fact that a person does not satisfy C can explain such cases.¹³

(v) Individual vs. Communal This dimension concerns whether the necessary connection between moral judgments and motivation holds on an individual or a communal level. According to an individual version of MJI, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ.¹⁴ According to a communal version of MJI, it necessary that if a person judges that she ought to ϕ, then she is part of, or suitable connected to, a community in which a significant number of people are motivated in accordance with their moral judgments.¹⁵ Advocates of the latter view maintain that a person cannot hold a moral judgment unless she is part of a moral community containing moral institutions and practices that sustain the connection between people’s moral judgments and motivation in general. This can, in turn, be accounted for by maintaining that morality is an essentially social phenomenon that requires the existence of such institutions and practices for its existence.¹⁶

Moral Judgment Externalism (MJE), as will understand this view, rejects MJI irrespective of how it is interpreted along the various dimensions delineated above. Thus, MJE denies that there is any nontrivial and necessary connection between moral judgments and motivation.

It is generally assumed that MJI is of utmost significance on the ground that this view in conjunction with the Humean theory of motivation entails that cognitivism is false and lends support to expressivism. Cognitivism is standardly understood as the semantic claim that moral sentences express moral judgments that consist in beliefs, and expressivism as the semantic claim that moral sentences express moral judgments that consist in desire-like states. The internalist argument can be construed as follows. According to a generic formulation of MJI, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ. However, according to the Humean theory of motivation, a belief is not by itself sufficient for motivation; to be motivated, it is necessary to have a distinct mental state in the form of a desire.¹⁷ It is concluded that moral judgments do not consist in beliefs, which means that cognitivism is false, and that they involve desires, which is an argument for expressivism.

We can now see that MJI needs to be understood according to the first version of this claim in each of the various dimensions (i)–(v) in order to function as a premise in this argument. (i) Assume a non-conceptual version of MJI: such a claim is not a semantic claim about what kind of moral judgments that moral sentences express. In that case MJI in conjunction with the Humean theory of motivation does not provide any support to expressivism as standardly understood, since moral sentences would not express moral judgments with the relevant connection to motivation or desire.¹⁸ (ii) Assume a non-generalizable version of MJI: such a claim does not generalize to other moral judgments than those mentioned in a generic formulation of such a view. In that case there is reason to doubt the plausibility of the general conclusions of the argument: that cognitivism is false and that there is reason to adopt expressivism. (iii) Assume an extrinsic version of MJI: a moral judgment is not by itself sufficient for motivation. Assume further the Humean theory of motivation. The conjunction of these views is compatible with moral judgments consisting in beliefs. For example, it might be a person’s moral belief in conjunction with a certain further fact that explains why she is motivated to ϕ, where her motivation is constituted by a desire.¹⁹ (iv), (v) Assume a conditional or a communal version of MJI. As these are instances of extrinsic MJI, similar considerations hold for these views.²⁰

Corresponding to the various versions of MJI, there are several distinct ways for cognitivists to counter the internalist argument. However, I will argue that cognitivists should respond to this argument by denying MJI altogether and adopt MJE. However, they need to complement MJE with a pragmatic account so as to explain our conception of the connection between moral language and motivation.

The Alleged Support of Internalism

It is generally presumed that MJI has a significant advantage over MJE on the ground that it is able to explain our conception of moral language and motivation. In the present section, I will examine this supposition by considering the various types of thought experiments that are assumed to show that MJI is explanatorily superior to MJE in this regard. The discussion follows my description above of the different dimensions of MJI, (i)–(v).

Individual Cases

The overwhelmingly most frequent argument for MJI invokes cases that concern the connection between a person’s moral utterance and her motivation.²¹ The following scenario is representative:

C1. Anne and Bill are watching a TV programme about people suffering from famine. They start to talk about charity. Anne says to Bill “Actually, I ought to give some money to those who are starving.” There is no indication that her utterance is insincere, e.g., that she is lying or is ironic. At the same time, Anne has no motivation whatsoever to donate any money to the people who are starving.²²

We respond presumably to such cases by finding them puzzling and we might even want to question whether the person in question actually thinks (i.e., judges) that she ought to ϕ, even if we do not have any evidence that she is insincere.

(i) It should first be pointed out that our responses to the type of thought experiments that internalist make use of only can provide support to a conceptual version of MJI, since these cases clearly are assumed to invoke our linguistic intuitions. The internalist explanation of our responses is thus that we know, granted that we are competent language users, that a sentence of the type “I ought to ϕ” expresses a moral judgment such that if a person holds it, then she is motivated to ϕ (given that certain requirements are fulfilled).

However, contrary to what usually is assumed to be the case, our response to cases such as C1 provides very little support to a conceptual version of MJI. We are told in these cases that a person utters a sentence to the effect that she ought to ϕ without being motivated to ϕ. However, MJI is not a thesis about the connection between utterances of such a sentence and motivation, but about the connection between the moral judgment such a sentence, in virtue of its conventional meaning, expresses and motivation. Consequently, our response is compatible with another explanation than the one proposed by internalists: what explains that we find a case such as C1 puzzling is that a person’s utterance of a sentence to the effect that she ought to ϕ standardly conveys that she is motivated to ϕ, not that the sentence she utters expresses a moral judgment that is necessarily accompanied by motivation. Our response thus leaves open to the possibility that even if it is anomalous for a person to utter such a sentence without being accordingly motivated, it is possible for a person to hold such a judgment without being accordingly motivated. Later on we will see that this view also can explain why we might come to doubt that a person who makes such a moral utterance, without being motivated to ϕ, thinks that she ought to ϕ.²³

This argument is similar to a well-known objection against inferences from use to meaning. When internalists appeal to cases such as C1, they argue in effect that we feel that it is odd to use a certain type of moral sentence in the absence of motivation, and they take this to indicate that the sentence expresses a moral judgment that is necessarily accompanied by motivation. A number of philosophers have pointed out that from the fact that it would be odd to use a given expression in certain circumstances we cannot draw any definitive conclusion as regards its meaning.²⁴ In the same vein, I argue that the fact that it is odd to use a certain type of moral sentence unless one is accordingly motivated should not make us infer MJI. These philosophers often maintain that we instead should adopt a pragmatic explanation of why it is odd to use an expression in certain circumstances. Consider:

C2. Anne and Bill are professors at the philosophy department. They are discussing which students that will be admitted to the master programme in ethics. Anne says to Bill “David or Saul will be admitted to the master programme.” There is no indication that her utterance is insincere, e.g., that she is lying or is ironic. At the same time, Anne thinks that both David and Saul will be admitted to the programme.

It is plausible to assume that we respond to C2 by finding it puzzling since a person’s utterance to the effect that David or Saul will be admitted to the master programme strongly indicates that she does not think that both of them will. We might even hesitate to ascribe to her the belief that David or Saul will be thus admitted in that case. However, according to the received view, the sentence “David or Saul will be admitted to the master programme” does not entail that not both of them will be admitted to the programme. To formulate the point in the vocabulary utilized above: The sentence does not, in virtue of its conventional meaning, express a belief entailing that not both David and Saul will be admitted to the programme. Rather, it is a person’s utterance of a sentence of the type “X or Y” which standardly conveys that not both X and Y are the case. Thus, our response to C2 should not be explained by the meaning of the sentence in question. It should rather be explained by there being a standardized pragmatic connection between utterances of this type of sentence and the mentioned piece of information.²⁵ Importantly, this suggests that the explanation of why we find C1 puzzling might be of the same kind as the explanation of why we find C2 puzzling: there is a standardized pragmatic connection between utterances of a certain type of sentence and a certain piece of information.

In reply, internalists might object that there is another type of case that provides stronger support to their view:²⁶

C3. Anne is alone at home watching a TV programme about people suffering from famine. Anne starts to think about charity. She thinks to herself that she ought to give some money to those who are starving. At the same time, Anne has no motivation whatsoever to donate any money to the people who are starving.

Internalists might argue that we respond to cases such as C3 by finding them puzzling; indeed, they might even be inclined to declare that such cases are impossible. The idea is then that this response would give stronger support to MJI than our response to C1, since C3 concerns moral judgments rather than moral utterances.

It is first worth noticing that internalists appeal to cases such as C3 much less frequently than they appeal to cases such as C1. This is in itself remarkable since, as just noticed, our response to them would provide stronger evidence for MJI. Indeed, it might even be suggested that the fact that internalists appeal to cases such as C1, rather than C3, suggests a pragmatic explanation on the lines indicated above.

More importantly, the fact that we find cases such as C3 puzzling does not help the sake of MJI. The reason is that we respond in the same manner to cases where our response cannot be explained in terms of meaning but needs to be explained pragmatically. Consider:

C4. Anne, who is a professor at the philosophy department, is alone working. She is thinking about which students that will be admitted to the master programme in ethics. Anne thinks to herself that David or Saul will be admitted to the master programme. At the same time, Anne thinks to herself that both David and Saul will be admitted to the programme.

It is plausible to assume that we respond to C4 by finding it puzzling because describing a person as thinking that David or Saul will be admitted to the master programme strongly indicates that she does not think that both of them will. However, as our discussion of C2 makes clear, this should not be explained by the conventional meaning of the sentence “David or Saul will be admitted to the master programme.” In view of our considerations in relation to C2, it is plausible to assume that it should be explained by there being a standardized pragmatic connection between utterances of a sentence of the type “X or Y” and the contention that not both X and Y are the case. More precisely, it is plausible to assume that we have, so to speak, internalized this connection in such a way that it has become part of our notion of the connection between “X or Y” and “not X and Y.” Importantly, this means that our response to C3 does not give any advantage to MJI over an externalist and pragmatic account. In other words, just as our response to C4 should be explained by means of pragmatic considerations, it might be that our response to C3 should be explained in this manner as well.

(ii) The type of cases that internalists appeal to in defence of their view, such as C1 and C3, concern a particular type of moral judgments, corresponding to the type of judgments referred to in a generic formulation of MJI. This means that our responses to such cases do not provide support to a generalized version of MJI according to which such a claim can be generalized to other kinds of moral judgments than those referred to in a generic formulation. It is plausible to doubt, however, that MJI can be thus generalizable in view of the numerous types of contexts a certain moral concept can be involved in. Importantly, if MJI is not generalizable it is reasonable to question whether it can constitute a conceptually necessary claim. The reason is that such a claim reasonably has to be generalizable to all judgments involving the same moral concept irrespective of how they diverge with regard to the various aspects of moral judgments mentioned earlier. Unfortunately, I cannot pursue this argument in the present paper as it brings up a number of complex topics.

(iii) It can next be observed that our response to cases such as C1 or C3 provides support neither to an intrinsic nor an extrinsic version of MJI.

According to intrinsic version of MJI, a person’s judgment that she ought to ϕ is by itself sufficient for her to be motivated to ϕ. Our responses to cases such as C1 and C3 do not provide evidence to this version of MJI. They merely show that if we are told that a person utters a certain type of moral sentence, or holds a certain kind of moral judgment, without being accordingly motivated, we are puzzled and perhaps even inclined to protest that she cannot really be holding the moral judgment in question.²⁷ There is nothing in this response which shows that the judgment itself, or even a part of it, needs to be involved in an explanation of the motivation in question. It should particularly be observed that our responses to cases such as C1 or C3 do not provide support to an intrinsic version of MJI over an extrinsic variant of this view. The explanation of our response might be that we are told that a person judges that she ought to ϕ without being motivated to ϕ, and that we tacitly presume some further relevant consideration. For instance, it might be that we classify a person’s judgment as a moral judgment only if it is accompanied by motivation, although the judgment need not be involved in the explanation of the motivation.²⁸ As we presume that the person figuring in these cases holds a moral judgment, we are puzzled when we are told that she is not accordingly motivated.

According to an extrinsic version of MJI, a person’s judgment that she ought to ϕ is not sufficient by itself for her to be motivated to ϕ. Our responses to cases such as C1 and C3 do not lend support to this view for the simple reason that they do not make explicit reference to any additional consideration, such as the one just mentioned, which would explain why we respond to these cases by finding them puzzling. (But see the discussion on conditional MJI below.)

(iv) We might next consider the dimension pertaining to unconditional and conditional MJI.

According to a unconditional version of MJI, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ, and this holds for all persons irrespective of their mental condition. First, this view has the same type of difficulty as an intrinsic version of MJI. Plausibly, we generally presume that people do not suffer from, say, depression or apathy in the absence of information to this effect. As a consequence, the explanation of why we respond to cases as C1 or C3 as we do might be that we presume that the person in question does not suffer from any mental condition of the relevant kind. Second, consider:

C5. Identical to C1, except for the additional information: There are indications that Anne is deeply depressed.

It seems plausible to assume that we do not find cases such as C5 particularly puzzling, since it makes reference to a possible explanation of the person’s lack of motivation. Thus, there is reason to think that a person who judges that she ought to ϕ need not be motivated to ϕ quite irrespective of her mental condition.

According to a conditional version of MJI, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ if she fulfills condition C. What I said above suggests that this view gets support from a comparison between C5 and another type of case. Consider:

C6. Identical to C1, except for the additional information: There is no indication that Anne suffers from a mental condition such as depression, apathy, emotional disturbances, or the like.

It is plausible to assume that we find cases such as C6 puzzling. Moreover, it might be argued that since we find C6 more puzzling than C5, a conditional version of MJI is preferable to an unconditional version of this view.

However, the fact that we respond differently to cases such as C5 and C6 give limited support to a conditional version of MJI, since there is an alternative pragmatic explanation of our reactions. These responses amount to the following: we are more puzzled by cases (such as C6) in which a person utters a sentence to the effect that she ought to ϕ, is not motivated to ϕ, and she is reported not to suffer from a mental condition like depression, than cases (such as C5) in which she is reported to suffer from such a inadequacy. According to the relevant conditional version of MJI, the explanation of our responses is that the sentence the person utters expresses a moral judgment such that if a person holds it, she is motivated to ϕ, provided that she fulfills a certain condition C with regard to her mental condition. However, there is a pragmatic explanation of our responses: A person’s utterance of a sentence to the effect that she ought to ϕ standardly conveys that she is motivated to ϕ, granted the assumption that she does not suffer from a mental condition of the relevant kind. I will return to this explanation in Sect. 6.

Communal Cases

A number of internalists have appealed to cases that concern the connection between the moral language in an entire community and motivation.²⁹ Consider:

C7. There is an alien community which in many respects is like ours. People in this community utter sentences to the effect that certain actions ought to be performed. For example, when discussing charity, some of them might say things like “Actually, I ought to give some money to those who are starving.” There is no indication that their utterances are generally insincere; e.g., there are no indications that they generally are lying or are ironic. No one in this community who makes such utterances has, or has ever had, any motivation whatsoever to perform any of these actions.³⁰

We respond presumably to such cases by finding them puzzling and we might even come to question whether people in this community have a moral language.

Let us start with observing that our response to cases such as C7 does not give any support to an individual version of MJI. According to this view, it is necessary that if a person judges that she ought to ϕ, then she is motivated to ϕ.³¹ However, what we find puzzling in C7 is not that there is a single person who utters, or judges, that she ought to ϕ and that she fails to be motivated to ϕ, but that there is a whole community in which there are persons who utter that they ought to ϕ and that each of them fails to be motivated to ϕ. Our response leaves consequently open the possibility that a person might hold a moral judgment even if she is not accordingly motivated.³²

(v) Advocates of communal versions of MJI argue that although our response to cases such as C7 does not provide support to individual version of MJI, it provides support to communal MJI. In particular, they take our response to suggest that in order for a person to judge that she ought to ϕ, she must be part of a community in which a significant number of people are motivated in accordance with their moral judgments, even if she need not be motivated to ϕ herself.

However, our response to cases such as C7 provides little support even to communal MJI. The most important reason from the present perspective is that these cases concern the connection between utterances of moral sentences and motivation, not the connection between moral judgments and motivation. They leave consequently open the possibility that there is no necessary connection between moral judgments and motivation even at a communal level. In particular, they are compatible with a pragmatic account according to which there is a standardized connection between moral utterances and motivation. Consider:

C8. There is an alien community which in many respects is like ours. People in this community utter sentences of the type “X or Y.” For example, when discussing which students that will be admitted to the master programme in ethics, philosophy professors might say things like “David or Saul will be admitted to the master programme.” There is no indication that their utterances generally are insincere; e.g., there are no indications that they generally are lying or ironic. No one in this community who makes such utterances thinks, or has ever thought, that not both X and Y is the case. For example, none of the mentioned philosophy professors thinks, or has ever thought, that not both David and Saul will be admitted to the master programme.

We are presumably liable to find such cases puzzling. According to the received view, the explanation of our response is not that a sentence of the type “X or Y,” in virtue of its conventional meaning, expresses a belief which entails that not both X and Y are the case, even granted that the person in question is part of a community in which a significant number of people who holds the first belief hold the second. As before, the explanation is rather that there is a standardized pragmatic connection between utterances of a sentence of the type “X or Y” and the mentioned piece of information.³³

It might be objected that we find communal cases, such as C7, more puzzling than individual cases, such as C1. It is this difference proponents of communal MJI employ so as to make plausible the claim that even if it is possible for a person to judge that she ought to ϕ without being motivated to ϕ, it is not possible for a person to judge that she ought to ϕ unless she is part of a community in which there is a significant correlation between moral judgments and motivation.³⁴ It might then be maintained that a pragmatic account cannot explain this difference between the individual and communal level. But this difference is not surprising on the suggestion that there is a standardized pragmatic connection between moral utterances and motivation. In C7 no one who utters that she ought to ϕ is motivated to ϕ, whereas this is the case only as regards one person in C1. Similarly, corresponding to our responses to C1 and C7, we presumably find C8 more puzzling than C2. Furthermore, it is reasonable to argue that the pragmatic connection between moral utterances and motivation is a matter of the function moral language has in an entire community.

Let us summarize. First, we have seen that none of the responses that we are presumed that have in relation to the thought experiments internalists employ provides convincing support to any of the various versions of MJI. Importantly, it follows that these responses do not provide evidence to a version of MJI that can function as a premise in the internalist argument against cognitivism. It should also be stressed that in reaching this conclusion, I have presumed that internalists are correct about our responses. In other words, externalists need not contest that we have these responses in order to challenge MJI. Second, I have indicated that our responses to these cases can be explained by an externalist account that exploits a standardized pragmatic connection between moral utterances and motivation. Third, as the metaethical literature bears witness, internalist claims have a number of well-documented difficulties.³⁵ In view of these considerations, there is reason to further investigate the plausibility of an externalist and pragmatic alternative. However, in order to be plausible, this view needs to account for the fact that utterances of moral sentences standardly convey motivation in a way which has been taken to suggest that it should be explained in terms of the conventional meaning of these sentences. In what follows, I will argue that externalists can provide such an account in terms of Paul Grice’s notion of general conversational implicature.³⁶

Two Purposes of Moral Discourse

A governing idea in Grice’s theory of conversational implicatures is that we contribute to the conversations in which we are engaged in the ways that are required to fulfill the mutually accepted purposes of these conversations. He formulates a number of maxims which make it possible for us to fulfill this goal. The two I will make use is the maxim of relation, which states that one should make one’s contributions so as to be relevant in consideration of the purposes of the conversation in question, and the second maxim of quantity, which states that one should not make one’s contributions more informative than is required in consideration of these purposes.³⁷ In Grice’s view, the fact that we are understood to conform to these maxims can make our utterances convey information that is not part of the conventional meaning of the sentences we employ. He refers to this information as “conversational implicatures.”

To illustrate, assume that Anne and Bill work together with trying to assemble a bookcase and that Anne says to Bill “The screwdriver is over there.” It is reasonable to assume that, given certain features of the context, Bill understands Anne’s utterance to convey that she wants him to give her the screwdriver. This cannot be explained by the conventional meaning of the sentence she utters. Simply put, it should rather be explained by the fact that a mutually recognized purpose of their conversation is that they should put together the bookcase and that they are assumed to make utterances that are relevant given this purpose. On these assumptions, the most reasonable understanding of Anne’s utterance is that she wants Bill to give her the screwdriver.³⁸

Grice distinguishes between two kinds of conversational implicatures. To see the difference, assume that an utterance of a sentence s conversationally implicates q.

It is a particularized conversational implicature (PCI) in case an utterance of s does not standardly conversationally implicate q. The utterance carries the implicature in question only on the assumption that the context in which it is uttered has certain specific features justifying this understanding of it. (The case above is an example of PCI.)

It is a generalized conversational implicature (GCI) in case an utterance of s standardly conversationally implicates q. The utterance carries the implicature in question even if the contexts in which it is uttered do not have any specific features that justify this understanding of it. Hence, a GCI does not need contextual features of this kind to arise but arises unless there are any specific circumstances that defeat it.³⁹

An important idea in Grice’s theory is that utterances carry conversational implicatures because there are certain mutually recognized purposes of the conversations in question. In what follows, I will maintain that externalists should claim that there are two mutually accepted purposes that generally are present in moral conversations.

(1) The first general purpose of moral conversation is to communicate what the participants of the conversation believe about moral matters. For example, in moral conversations people communicate their beliefs that certain actions ought to be performed. This exemplifies the kind of purpose Grice describes as “giving and receiving information.”⁴⁰

This purpose corresponds to what constitutes the conventional meaning of moral sentences according to cognitivism. A main objective of the present paper is, as pointed out in the introduction, to defend MJE in order to save cognitivism from the internalist argument. Consequently, in what follows I will investigate the prospects of combining cognitivism with an externalist and pragmatic view. It should be emphasized that the view I am defending does not entail a particular version of cognitivism but is compatible with any relevant version of this view. However, it presupposes one important but uncontroversial requirement: A sentence to the effect that a person morally ought to ϕ entails that she has a moral reason to ϕ.⁴¹ A reason to ϕ consists, according to a widely accepted view, in a fact which “counts in favour” of ϕing.⁴² In other words, a sentence to the effect that a person morally ought to ϕ entails that ϕing has some feature which, from the point of view of morality, counts in favour of that person ϕing. This requirement does not entail any particular conception of moral reasons. In particular, it does not entail that moral reasons constitute normative reasons where such reasons are understood in terms of rationality. According to such a notion, if a person thinks that she has a normative reason to ϕ, then she is motivated to ϕ if she is fully rational.⁴³ My account is thus not committed to this notion, even if it is compatible with it.

As the first purpose corresponds to the conventional meaning of moral sentences according to cognitivism, it should be uncontroversial to assume that moral conversations generally have that purpose on this view. In contrast, the second purpose is not in this way connected to what constitutes the conventional meaning of moral sentences on cognitivism. It is clear from Grice’s discussion, however, that what purposes conversations have need not be directly connected to the meaning of the sentences uttered in them.

(2) The second general purpose of moral conversations is to influence action and, in particular, to make people perform, or avoid to perform, certain actions.⁴⁴ This exemplifies the purpose Grice describes as “influencing or directing the actions of others.”⁴⁵

It is widely agreed in metaethics that moral language is essentially practical. Generally, if a person employs a sentence to the effect that ϕing has a certain moral characteristic, then we presume that she has a certain action-guiding attitude towards ϕing. Indeed, our responses to the cases we considered above confirm this view. A natural explanation of why moral language is practical in this way is, briefly put, that moral language thereby fulfills an important function: it makes it possible for us to regulate one another’s behaviour in directions that we find desirable. For instance, assume that, generally, if a person utters a sentence to the effect that ϕing ought to be performed, then we presume that she wants that ϕing is performed. It is plausible to suppose that this, all things equal, will make us more inclined to ϕ than if she had not made the utterance. Similarly, it is likely to make us more inclined to assist other people in ϕing. Further, assume that, generally, if a person’s utters a sentence to the effect that she ought to ϕ, then we presume that she is motivated to ϕ. This is also, all things equal, likely to make us more inclined to ϕ than if she had not made the utterance, since it indicates that she wants that ϕing is carried out. Likewise, it is likely to make us more inclined to assist her and others in ϕing. The explanation of this feature of moral language is presumably very fundamental: in order to live together in a way which promotes our own welfare and the welfare of our family and mates, we need to avoid conflicts and to cooperate with one another. As a result, what actions we carry out will depend on what attitudes we think people have towards them.

Granted the assumption that moral language generally functions so as to regulate behaviour, it is plausible to assume that moral conversations generally have as a purpose to influence action. If the purpose of moral conversations to influence action is successfully achieved, this will have as an overall result that we regulate one another’s behaviour by means of moral language, and, hence, that moral language fulfills this function. As it is plausible to think that moral language has this function, it is plausible to think that our moral conversations generally have the purpose to influence action. On the other hand, if it were not a general purpose of moral conversations to influence action, it is difficult to see how moral language would be able to fulfill this function, since moral language in that case would need to work in a way that does not match the purpose we have in employing it. It is also important to observe that, in Grice’s view, we do not need to be consciously aware what the purposes are of the conversations we are engaged in. Grice’s contention is particularly plausible on the view that the purpose to influence action is a general purpose of moral conversations, since this view helps to explain that the purpose has become habitual and, hence, stays in the background of our moral considerations. We should rather think of it as a tacit assumption that guides our understanding of moral utterances without us necessarily being consciously aware of it.

Metaethicists who reject cognitivism and accept expressivism should agree that moral conversations have the general purpose of influencing action. To see this, suppose that it were not a purpose of moral conversations to influence action. In that case, a person’s utterance of a sentence to the effect that ϕing ought to be performed would not be able to influence action, even if the sentence expresses a certain desire-like state. This is so since we would not identify the utterance as saying anything that is relevant in the conversation in question. Accordingly, expressivists repeatedly stress that the purpose of moral language is to influence action.⁴⁶

The Dual Aspect Account

In what follows, I will argue that advocates of MJE should adopt the Dual Aspect Account (DAA) as regards moral motivation. This account states that a person’s utterance of a sentence to the effect that she ought to ϕ conveys two things, corresponding to the two purposes of moral discourse identified above. First, such a sentence expresses, in virtue of its conventional meaning, the belief that the person in question ought to ϕ. Second, an utterance of this type of sentence carries a GCI, a standardized conversational implicature, to the effect that she is motivated to ϕ. As the first aspect merely is a statement of cognitivism, I will focus on the second aspect.

Let us start with considering an utterance of a sentence to the effect that a certain person ought to ϕ, but where the person in question need not be the speaker herself. We observed in particular three things above. First, a sentence to the effect that a person ought to ϕ entails that she has a moral reason to ϕ, which in turn entails that ϕing has some feature which, from the point of view of morality, counts in favour of her ϕing. Second, given the objectives of the present paper, it is reasonable to assume that moral conversations have two general purposes: to communicate beliefs about moral issues and to influence action. It was also observed that we need not be consciously aware of the purposes of the conversations we are engaged in. Third, a very basic way to influence other people’s actions is to inform them about our attitudes. We observed in particular that if our moral utterances convey that we want that certain actions are performed, or that we are motivated to perform certain actions, this will have the function to influence other people’s actions in ways that we find desirable.

Now, in view of these considerations, it is reasonable to understand someone who utters a sentence to the effect that a person ought to ϕ as wanting that the person in question ϕs. That is, if someone utters a sentence which entails that, from the point of view of morality, there is a reason for a person to ϕ, and moral conversations have as a general purpose to influence action, and a basic way of doing so is to inform one another about our attitudes towards these actions, we understand the speaker as wanting that the person in question ϕs. The reason is that, in consideration of the purpose of moral conversations to influence action, her utterance would lack in relevance unless she wants that the person ϕs. Put differently, it is difficult to see what the point would be for someone to utter a sentence which entails that there is a moral reason for a person to ϕ, in a conversation that has as a purpose to influence action, unless she wants that the person in question ϕs.

Let us next consider a person’s utterance of sentence to the effect that she ought to ϕ. We have seen that an utterance of a sentence to the effect that a person ought to ϕ conveys that the speaker wants that the person in question ϕs. In case a person utters a sentence according to which she ought to ϕ, this means that her utterance conveys that she wants herself to ϕ; in other words, that she is motivated to ϕ.

This can be accounted for in terms of Grice’s notion of the maxim of relation, which declares “Be relevant.”⁴⁷ In Grice’s view, we are justified to assume that a person who is engaged in a conversation contributes to it in the ways that are required to fulfill its mutually accepted purposes. An essential means of doing so is to follow the maxim of relation and, hence, only make utterances that are relevant in view of the purposes of the conversation in question. As we have seen, in order to explain that a person who utters a sentence to the effect that she ought to ϕ contributes to a moral conversation in a way which is relevant given the purpose of such conversations to regulate behaviour, we should understand her to be motivated to ϕ. Consequently, the person’s utterance conversationally implicates that she is motivated to ϕ.⁴⁸, ⁴⁹

Moreover, this is a GCI rather than PCI. We noted above that, in order to regulate people’s behaviour, moral conversations generally have as a purpose to influence action. This means that moral conversations have this purpose throughout the various contexts in which people are involved in such conversations. In turn, this means that a person’s utterance to the effect that she ought to ϕ conversationally implicates that she is motivated to ϕ throughout such contexts. In other words, such an utterance standardly carries this implicature. In particular, given that moral conversations generally have the purpose to influence action, they have this purpose also in the absence of specific contextual features. That is, they have this purpose even if the context lacks any special features indicating that this is the purpose of the conversation in question. This purpose thus works as a kind of tacit assumption that governs our understanding of moral utterances. As a result, a person’s utterance to the effect that she ought to ϕ conversationally implicates that she is motivated to ϕ even if the context in which it is uttered lacks special features of the mentioned sort; such an utterance carries the implicature unless there are any special circumstances that defeat it.

It might be objected that there are cases in which moral conversations do not have the purpose to influence action but a person’s utterance to the effect that she ought to ϕ nevertheless conveys that she is motivated to ϕ. This might be the case when the subject matter of the conversation is not practically significant; for example, when it concerns matters people cannot influence or when it concerns abstract moral issues.

To respond to this objection, let us start by observing that there are various commonly accepted views that constitute a set of background presumptions about moral utterances and beliefs. For instance, it is generally the case that someone who utters a sentence to the effect that she ought to ϕ is motivated to ϕ. It is not difficult to explain this presumption on the view defended above, as it maintains that such utterances carry a GCI to the mentioned effect. According to another background presumption, it is generally the case that a person who thinks that she ought to ϕ is motivated to ϕ. There are presumably a number of psychological and social explanations of this assumption, such as moral upbringing, socialization, and certain fundamental human features and conditions.⁵⁰ Another explanation concerns normative reasons. As observed above, the view defended in this paper is not committed to the claim that a sentence according to which a person ought to ϕ entails that she has a normative reason to ϕ, where normative reasons are understood in terms of rationality. However, it is not farfetched to assume that we regularly believe that we have strong normative reasons to do what we morally ought to do. Moreover, it is plausible to assume that we are motivated to do what we think we have strong normative reasons to do in so far as we are rational. On the assumption that we generally are rational, it can consequently be expected that we regularly are motivated to do what we think we morally ought to do.

Return now to a person who utters a sentence to the effect that she ought to ϕ in a conversation which does not have as a purpose to influence action. In consideration of the kind of background assumptions I just mentioned, it is nevertheless plausible to assume that we understand the person’s utterance as conveying that she is motivated to ϕ.

This can be accounted for in terms of Grice’s second maxim of quantity, which declares “Do not make your contribution more informative than is required.”⁵¹ In order to understand the person above as complying with this maxim and, hence, as not providing more information than is called for, we need to understand her utterance in such a way that she has no reason to submit any information that modifies it. In particular, we need to assume that she has no reason to inform us that what ordinarily is taken for granted to be the case as a matter of fact is not the case in the particular situation at hand. As a result, it seems that the best explanation of her utterance is that there is no exception to the background presumptions I mentioned above. Consequently, the person’s utterance conversationally implicates that she is motivated to ϕ.⁵²

Once again it is a matter of a GCI rather than a PCI. As I suggested above, it plausible to think that these background presumptions are quite commonly taken for granted in the various contexts in which people are engaged in moral conversations. This means that a person’s utterance to the effect that she ought to ϕ conversationally implicates that she is motivated to ϕ throughout such contexts. The fact that people who are engaged in a moral conversation share certain background presumptions about moral utterances and beliefs is part of its context, but since these presumptions standardly are taken for granted, they are not special contextual features. They are, in other words, not features that merely belong to individual contexts, but rather standing contextual features of moral conversations. As a result, a person’s utterance to the effect that she ought to ϕ conversationally implicates that she is motivated to ϕ even in the absence of special contextual features; such an utterance carries the implicature granted that there are no special circumstances that defeat it.

We have seen that even if we assume that a moral conversation lacks the second purpose, a person’s utterance according to which she ought to ϕ might carry a GCI to the effect that she is motivated to ϕ. Hence, an utterance of this type might carry such a GCI even if the purpose of that particular conversation is not to influence action. The reason is, as we have seen, that there are certain commonly shared background presumptions that generally are present in moral conversations.

Explaining our Conception of Moral Language and Motivation

We are now in the position to see that DAA can explain our conception of the connection between moral language and motivation as it manifests itself in our responses to the various types of thought experiments we considered in Sect. 4.

(i) As far as DAA is concerned, there is no conceptually necessary connection between moral judgments and motivation, contrary to what a conceptual version of MJI maintains. That is, DAA does not state that a sentence of the type “I ought to ϕ,” merely in virtue of its conventional meaning, expresses a judgment such that if a person holds it, she is motivated to ϕ. In spite of this, DAA can explain our response to individual cases such as C1. The reason is that it states that a person’s utterance of a sentence according to which she ought to ϕ carries a GCI to the effect that she is motivated to ϕ, which means that utterances of such a sentence standardly convey this piece of information. As there is such a standardized connection between utterances of this type of sentence and motivation, we are puzzled if a person makes such an utterance without being motivated to ϕ.

What are we to say about someone who makes such an utterance without being accordingly motivated? Given that a sentence of the type “I ought to ϕ” does not convey motivation to ϕ merely by means of its conventional meaning, it is not strictly accurate to claim that her use of the sentence is semantically incorrect. But we are surely justified to object that her use of the sentence is extremely misleading in view of the fact that an utterance of such a sentence carries the mentioned GCI.⁵³ It is also understandable if we experience her use of it as plainly incorrect. As there is a standardized correlation between uses of this type of sentence and motivation, it is easy to get the impression that moral motivation is a matter of the meaning of the sentence, as a conceptual version of MJI maintains.⁵⁴

Assume that the person’s utterance is made in a moral conversation that has both the purposes we identified earlier. In that case her utterance violates the maxim of relation, since it is not relevant in view of the purpose of influencing action. Furthermore, her utterance violates the second maxim of quantity, since it conveys information that is uncalled for: the utterance conversationally implicates that she is motivated to ϕ although she is not. If the moral conversation she is involved in does not have the second purpose, her utterance only violates the latter maxim, but we will still find it puzzling.

Moreover, DAA can explain why we might come to doubt that a person who utters that she ought to ϕ, without being motivated to ϕ, even thinks (i.e., judges) that she ought to ϕ. First, a person’s utterance of a truth-evaluable sentence standardly conveys that she asserts it and, hence, that she entertains the belief expressed by the sentence.⁵⁵ Second, according to DAA, a person’s utterance of a sentence according to which she ought to ϕ carries a GCI to the effect that she is motivated to ϕ, which means that such an utterance standardly conveys this information. As a consequence, it is plausible to suggest that, if a person utters a sentence of this type without being motivated to ϕ, we might come to doubt that she asserts it and, hence, that she believes that she ought to ϕ. Compare: According to an influential view, a person’s utterance of a sentence of the type “X or Y” standardly conveys that not both X and Y are the case, and this can be accounted for in terms of a certain GCI. If a person utters “David or Saul will be admitted to the master programme” at the same time as we are told that she believes that David and Saul will be admitted, we might come to doubt that she asserts the sentence and, hence, that she believes that David or Saul will be admitted to the programme.

It can further be argued that DAA is able to explain why we find cases such as C3 puzzling in which a person is reported to think (i.e., judge) that she ought to ϕ at the same time as she lacks motivation to ϕ. Let us first observe that given the background assumptions considered in the last section, we presume that a person who thinks that she ought to ϕ is motivated to ϕ because of factors such as upbringing, socialization, and the connection between morality and normative reasons. This means that DAA already accommodates the notion that there is a close connection between moral thought and motivation.

As I have already indicated, DAA is able to provide a further explanation of our response to cases such as C3. To see this, let us start by considering our response to the corresponding non-moral case in the form of C4. We saw that our response to C4 should not be explained by the conventional meaning of a sentence of the type “X or Y,” but that it has to be explained pragmatically. It is widely assumed that an utterance of a sentence of the type “X or Y” standardly conveys that not both X and Y are the case, and that this can be accounted for in terms of a certain GCI. Our response to C4 thus makes it reasonable to assume that this has formed our conception of the relation between “X or Y” and “not X and Y.” As a consequence of this fact, we find it misleading to describe a person as thinking that David or Saul will be admitted to the master programme at the same time as she is described as thinking that both of them will. For the same reason, the very thought of a person who combines these beliefs might seem paradoxical. Now, DAA provides an analogous explanation of our response to cases such as C3. That is, it might be suggested that the standardized pragmatic connection between moral utterances and motivation is so strong that it has come to affect our conception of the connection between moral beliefs and motivation. As a consequence, we find it misleading to describe a person as thinking that she ought to ϕ at the same time as she is described as not being motivated to ϕ. Moreover, we might find the very thought of such a person paradoxical.

It is further worth observing that on the standard amoralist objection against MJI, we find it possible to conceive of a person who thinks that she ought to ϕ without being motivated to ϕ, even if we may find such a person peculiar. This impression is not difficult to explain on DAA. An amoralist is conceivable on this view because there is no conceptually necessary connection between moral judgments and motivation. Yet, we find her utterly peculiar since our conception of the connection between moral beliefs and motivation is influenced by the fact that there is a standardized connection between certain uses of moral language and motivation.

(ii) As far as DAA is concerned, there is no necessary connection between moral judgments and motivation. That is, there is no necessary connection of this kind that needs to apply to other moral judgments in the manner a generalizable version of MJI maintains. As a result, DAA can account for the fact that our responses to cases such as C1 and C3 do not provide any evidence that MJI is generalizable.

It is also important to observe that, according to DAA, the standardized pragmatic connection that holds between utterances of a certain type of moral sentence and motivation need not be generalizable to utterances of other types of moral sentences. More precisely, it is only a person’s utterance of a particular type of sentence that needs to carry a GCI to the effect that she is motivated to ϕ: a type of sentence which entails that she has a moral reason to ϕ. An essential part of the explanation as to why a person’s utterance of a sentence according to which she ought to ϕ carries a GCI to the effect that she is motivated to ϕ consists in the fact that such a sentence entails that she has a moral reason to ϕ. It is particularly this fact which explains why it is inconsistent with the maxim of relation to make such an utterance without being motivated to ϕ in a moral conversation which has as a purpose to influence action.⁵⁶ This means that utterances of moral sentences which do not entail the existence of moral reasons of this kind need not convey that the person in question is accordingly motivated.⁵⁷ It should be stressed, though, that DAA is compatible with the view that utterances of other types of moral sentences convey other kinds of action-guiding states by pragmatic means.

(iii) As far as DAA is concerned, motivation is not constitutive of moral judgments, contrary to what an intrinsic version of MJI maintains. That is, DAA does not state that a person’s judgment that she ought to ϕ is by itself sufficient for her to be motivated to ϕ. This view is consequently able to explain why our responses to cases such as C1 and C3 do not provide support to the view that there is something in the nature of the moral judgment that explains her motivation to ϕ. As already stressed, I suggest that DAA should be combined with MJE. This means that in order for a person to be motivated to ϕ, she needs to be in a motivational state that is entirely distinct from her moral judgment.⁵⁸

From the perspective of DAA, the underlying reason as to why a person’s utterance according to which she ought to ϕ carries a GCI to the effect that she is motivated to ϕ is that it thereby fulfills the purpose of moral conversations to influence action. What is most important from this perspective is consequently that a person who makes such an utterance is motivated to ϕ, not what motivates her to ϕ. At the same time it should be recalled that she utters a sentence which entails that she has a moral reason to ϕ. This means that when a person’s moral utterance carries a GCI to the effect that she is motivated to ϕ, her motivation to ϕ typically can be explained by her considerations regarding her moral reason to ϕ.⁵⁹ As a consequence, DAA can explain why we might be puzzled if a person who makes such an utterance is motivated to ϕ but is not motivated by any moral considerations in relation to ϕing.

(iv) It should next be noticed that DAA is able to explain why we are more puzzled by cases (such as C6) in which a person utters that she ought to ϕ, is not motivated to ϕ, and is reported not to suffer from any mental condition like depression or apathy, than by cases (such as C5) in which she is reported to suffer from such a deficiency. This also means that DAA can explain why a conditional version of MJI might seem more plausible than an unconditional variant of this view.

First, it should be observed that even if, as externalists maintain, a person’s judgment that she ought to ϕ and her motivation to ϕ are entirely distinct mental states, her mental condition can influence to what extent she is motivated in accordance with her moral judgment.⁶⁰ Second, it is plausible to assume that it is part of folk psychology that a person’s motivation to do what she thinks she ought to do might be negatively influenced by her mental condition. It is therefore plausible to suppose that this aspect of folk psychology is part of the kind of background presumptions regarding the connection between moral judgments and motivation I mentioned in the last section.

Assume that a person utters that she ought to ϕ, but that she is not motivated to ϕ. Moreover, assume that her utterance is made in a context in which there are no indications that she suffers from a mental condition such as depression, apathy or emotional disturbances. In consideration of the mentioned background presumption regarding the connection between moral judgments and motivation, it is very likely that they find such a case more puzzling than a case in which a person makes such an utterance without being motivated to ϕ, but where there are indications that she suffers from such a mental condition. In the second case, there is a plausible explanation, in the form of her mental condition, as to why a person says that she ought to ϕ, and thinks this to be the case, at the same time as she fails to be motivated to ϕ. Importantly, in such a case the person’s utterance need not carry a GCI to the effect that she is motivated to ϕ, as a result of the contextual evidence regarding her mental condition.⁶¹ Thus, it is plausible to understand DAA as saying that a person’s utterance according to which she ought to ϕ carries a GCI to the effect that she is motivated to ϕ, provided that she is not understood to suffer from a mental condition such that it can influence her motivation to ϕ negatively.

(v) Let us finally observe that DAA can explain why we find communal cases, such as C7, puzzling. We have seen that, on this view, the fact that a person’s utterance according to which she ought to ϕ carries a GCI to the effect that she is motivated to ϕ is a matter of the function of moral language. This view assigns two major functions to moral language: first, to inform one another about our beliefs about moral matters and, second, to influence one another’s actions. The latter means that moral language is regarded as an essentially social phenomenon since it has as a function to regulate the behaviour of people who live together in a community in such a way that their actions are adjusted to one other. Now, in cases such as C7 the moral language in an entire community seem to lack the second function in view of the fact that no one is motivated in accordance with his or her moral utterances. As a result, we find such cases quite bewildering. Indeed, on DAA it might even be claimed that people in this community do not have a moral language on the ground that their manner of interacting with one another lacks a function which constitutes a moral discourse.⁶²

We saw earlier that considerations of communal cases have made some internalists adopt a communal version of MJI. It is worth observing that DAA can be combined with a picture of moral language which is quite similar to this view. A communal version of MJI allows that a person might hold a moral judgment without being accordingly motivated, granted that she is part of a community in which a significant number of people are motivated in accordance with their moral judgments. As for DAA, it is possible to maintain that a person might make a moral utterance without being accordingly motivated, granted that her utterance is made in a community where moral language has as a function to influence action. That is, unless the utterance is made in a community where moral language has this function, it cannot influence action in which case it is doubtful whether it can be regarded as a genuine moral utterance.

Conclusion

It is frequently claimed in metaethics that we have “internalist intuitions.” In this paper, I have in effect argued that this manner of speaking is misleading. We do not have internalist intuitions in the sense of having intuitions that support internalism. What we do have are intuitions that indicate that there is a very close connection between moral language and motivation, but these intuitions can be as readily explained by externalism. Moreover, I have suggested that externalists should explain our conception of moral language and motivation by adopting the Dual Aspect Account: A person’s utterance of a sentence according to which she ought to ϕ conveys two things: the sentence expresses, in virtue of its conventional meaning, the belief that she ought to ϕ, and her utterance of the sentence carries a generalized conversational implicature to the effect that she is motivated to ϕ. This view, I argued, is available to any cognitivist who embraces externalism since it does not rest on any particular version of cognitivism. Moreover, it enables cognitivists to defend their view against the internalist argument, since it makes it possible for them to embrace externalism at the same as they supply an account of our conception of the connection between moral language and motivation.