Reasons and ‘because’*

ABSTRACT

The paper argues that action explanations of the form ‘because p’ do not indicate that reasons are non-psychological facts or propositions. ‘Because p’ has two different uses: In the explanatory use, ‘because’ operates on the alleged fact that p. In the reason-giving use, however, ‘because’ operates not on p, but on the agent’s belief that p: she does not describe but express her reason. I conclude that a proper analysis of reason-giving ‘because’-utterances suggests that reasons are mental states.

KEYWORDS:

• Reasons
• ‘because’
• expressivism
• anti-psychologism
• psychologism

1. Introduction

Anti-psychologism, the view that motivating reasons are non-psychological facts or propositions, has often been driven by considerations such as the following: in giving a reason for her action, the agent is concerned, not with her own mental states, but with some alleged fact about reality. For example, with ‘because John is in London’, Mary gives a reason for her going to the British capital and in doing so considers only John’s whereabouts; neither is her own mental state mentioned in the utterance, nor is it the intentional object in the reason-giving act. From this, many anti-psychologists conclude that Mary’s motivating reason is not a state of mind but rather some state of affairs: the fact or proposition that John is in London.

The anti-psychologistic argument starts from what I consider to be an important insight. In giving a reason, Mary thinks about John, not about her own mental states. Yet, I claim, anti-psychologists wrongly infer that the fact or proposition that John is in London is Mary’s reason. That Mary thinks only about John’s abode patently does not, at least not by itself, entail or even suggest that reasons are non-psychological facts or propositions. Assumptions to the contrary are based, I argue, on a mistaken view of reason-giving. Once the view stands corrected, the same insight suggests that psychologism is correct after all.

The paper unfolds as follows. I will show in Section 2 that the anti-psychologistic conclusion plausibly derives from the assertion view of reason-giving, that is, the position that Mary, with ‘because John is there’, reports on the explanation, justification, or motivation of her action. Section 3 disputes the assertion view, thus depriving the anti-psychologist argument of a crucial premise. I show that Mary does not assert that John’s being in London explains, justifies, or motivates her going to London. Instead, I claim in Section 4, Mary’s reason-giving act is a case of rational grounding: Mary displays that she grounds her intention in a belief which she expresses by giving a reason. I conclude, in Section 5, that the observation that Mary considers only non-psychological facts suggests strongly (and ironically) that her motivating reason is the belief that John is in London. Once we properly reflect on the nature of reason-giving, it emerges that the anti-psychologistic argument backfires: that Mary is concerned with John supports the traditional psychologistic position.¹

In the discussion, I will establish a distinction between first-person present-tense reason-giving and (third-person) action explanation proper. While Sections 2–5 focus on the first-person case (‘I go to London because John is there’), I will discuss the third-person variant (‘Mary goes to London because John is there’) in the final Section 6 and demonstrate that anti-psychologists cannot draw support from such cases either. I will end with the suggestion that anti-psychologism draws a lot of its motivation from the erroneous assumption that first- and third-person cases are subject to essentially the same analysis.

Some remarks on the limitations of the present paper are in order. The main aim, at least in Sections 2–5, is to clarify what I take to be a central misunderstanding in the debate on the nature of motivating reasons, which shows itself in the argument from first-person reason-giving. I will therefore be exclusively concerned with this argument and the conclusions that can be derived from it. My main thesis is that the argument from first-person reason-giving does not support anti-psychologism. In addition, I suggest that reason-giving is best understood in terms of rational grounding. As a positive theory, I offer an expressivist analysis of rational grounding and conclude that the argument from first-person reason-giving actually supports psychologism. There are other arguments for anti-psychologism, but except for the argument from third-person action explanation (in Section 6) they will not be discussed here.² Also, the argument from first-person reason-giving has sometimes been taken to demonstrate that ‘epistemic’³ and ‘normative’⁴ reasons are non-psychological facts or propositions, too. My critique could be transferred to these arguments, but for the sake of straightforward presentation I concentrate here on the arguments concerning ‘motivating’ reasons alone. This brings us to terminology. I take it that, properly speaking, motivating reasons rationalise actions by rationalising the corresponding intentions.⁵ If Mary gives a reason for going to London, she gives a reason for her intention do so. When I here speak of the agent’s action, this should, in the appropriate circumstances, be understood as shorthand for her corresponding intention. Finally, as indicated above, some anti-psychologists claim that reasons are facts or states of affairs, and others that they are propositions, true or false.⁶ While the precise type of anti-psychologism is relevant when it comes to other difficulties, e.g. with respect to the ‘error problem’,⁷ it is immaterial for the purposes of the present paper. To keep things simple, I will henceforth often not differentiate between propositional and factual versions of anti-psychologism and then indiscriminately use the word ‘proposition’ for true or false propositions, states of affairs, facts, etc.

2. The argument from reason-giving

To the question ‘Why do you go to London?’, Mary may respond with the words,

1. I go to London because John is there.⁸

In the ‘because’-clause she does not, at least not verbally, refer to herself, but only to John’s whereabouts. Jonathan Dancy takes this piece of linguistic evidence at face value and concludes (if we transfer his ideas to our case) that Mary’s reason is not a belief but rather the state of affairs that John is in London: ‘Some reason-givings offer … as a reason … something that seems not to be a state of the agent at all, but a state of affairs’ (2000, 15; cf. 99). Maria Alvarez, using an example of her own and switching from linguistic to phenomenological considerations, arrives at a similar conclusion:

[W]hat motivates me to give [my cousin] the money is that he needs it: it is that that seems to me to make the action of giving him money right or appropriate and not my believing that he needs it. For, if my reason had been my believing that he needs the money, then, when deciding whether to give him the money, my concern would be with how things are with me, in particular, with my own state of mind, rather than with how things are with my cousin, in particular, with his financial situation. (Alvarez 2010, 131; italics in the original)⁹

In considering the reasons for an action, the agent does not necessarily think about her mental states; her thoughts may well be about nothing but worldly states of affairs (that the cousin needs money; that John is in London). Anti-psychologists conclude that motivating reasons are non-psychological propositions. Mary’s reason is not the belief, but rather the fact or proposition that John is in London.

This type of argument, call it the argument from (first-person present-tense) reason-giving, is rarely made fully explicit and comes in various forms and formulations. The following may be an acceptable reconstruction with respect to Mary’s case:

(P1)	Mary is concerned with the proposition that John is in London.
(P2)	The reason Mary gives is (identical to) the proposition that Mary is concerned with.
(P3)	Mary’s motivating reason is (identical to) the reason Mary gives.
(C)	Mary’s motivating reason is (identical to) the proposition that John is in London.

The argument is valid, so we must examine its premises. Premise (P1) is meant to capture the central phenomenological insight from first-person reason-giving. When Mary gives a motivating reason for her action, the reason ‘in the light of which [she] acts’ (Dancy 2000, 106), she is concerned not with herself, but with John’s abode.¹⁰ As mentioned above, I think that premise (P1) reflects an important insight, and hence, without further ado, I assume its truth.

The basic idea of premise (P3) is that, in determining the agent’s motivating reason, I must take her point of view seriously: ‘The aim of the explanation of action is to give, so far as possible, the agent’s own perspective on things, so as to reveal the light in which the action was done’ (Dancy 2000, 108).¹¹ That Mary speaks about John’s whereabouts in ‘explaining’ her action must hence be taken as a strong indicator of what constitutes the true motivating reason. Of course, we might object to premise (P3) on the grounds that Mary may be dishonest or insincere: her true reason for going to London may have nothing to do with John, the reference to him being but an excuse to meet up with Joan. Yet the issue is not whether Mary’s true reason pertains to John or Joan, but whether the reason is some worldly state of affairs or whether it is a belief of hers. To avoid unnecessary complications, I will therefore assume that the agents are sincere when they put forth their reasons and, given this proviso, accept that premise (P3) is correct as well. The argument from reason-giving thus stands and falls with premise (P2).

Anti-psychologists usually do not justify this premise, at least not explicitly, but the most natural construal makes it a consequence of the view that to give a reason is to assert some relation between the action in question and what is mentioned in the ‘because’-clause. In this spirit, Arthur Collins claims that in a ‘reason giving explanation’, the agent ‘make[s] intelligible to himself what he might do, and then what he does’ (1997, 121; cf. 112): according to Collins, the agent reflects on the action (or the corresponding intention) and then, after some deliberation, reports what she considers its explanation to be. As he describes a situation in which the agent takes the ferry ‘because the bridge is closed’: ‘[The agent] thinks of his action as something he is doing because he is faced with a closed-bridge situation’ (Collins 1997, 122).¹² When the agent says ‘I take the ferry because the bridge is closed’, Collins suggests, the agent articulates the hypothesis that he takes the ferry because the bridge is closed. Alvarez similarly argues that in reason-giving acts, the agent states what makes the action ‘seem right, or appropriate, or desirable to the agent’ (2010, 131). And Tim Henning holds that, in reason-giving, the agent provides an ‘account’ (2018, 4) of her action in which it is specified what it is that she takes ‘to weigh in favor of’ the action (2018, 163, 170; similarly 169). The exact formulations vary from author to author, and there may be considerable differences between reporting what explains, what motivates, and what counts in favour of an action. Yet, common to all proponents of the argument from reason-giving is the view that, in giving a reason, the agent asserts that some alleged state of affairs explains, motivates, or speaks in favour of her action. For lack of a better term, I will speak of this kind of conception as the assertion view of (first-person present-tense) reason-giving: according to the assertion view, with the utterance of (1), Mary asserts that there is a reason-relation (variously fleshed out as some explanatory, motivational, or justificatory relation) between the alleged fact that John is in London and her intention or action of going there.

The assertion view straightforwardly yields premise (P2): if Mary articulates the thesis that her going to London is explained, justified, or motivated by John’s being there, ‘because’ is used to operate on the proposition that John is in London, which is therefore the ‘explanans’ given. Assuming, furthermore, that the ‘explanans’ Mary provides for her action is the reason she gives for that action, the reason Mary gives is whatever she is concerned with in giving a reason. This yields premise (P2). Together with premises (P1) and (P3), we arrive at the conclusion that Mary’s motivating reason is the proposition that John is in London: reasons are those things the reason-giving agent speaks and thinks about. This, I take it, captures the essence of the anti-psychologistic rationale in the argument from reason-giving.

One more remark on anti-psychologism to forestall a possible misunderstanding: To say that reasons are non-psychological propositions is not, of course, to say that mental states do not have a part in reason-giving acts. Anti-psychologists usually hold that without the belief that p the agent cannot sincerely give a reason in referring to p:¹³ in order to sincerely maintain that John’s being in London is a reason for her action, Mary must believe that John is in London. Yet, anti-psychologists insist, that belief is required for sincere reason-giving should not distract us from the fact that, while belief would be indispensable for reason-giving thus construed, this does not mean that Mary’s believing would be (part of) the reason given. Mary’s reason then is still the proposition that John is in London alone.

3. Updating on reasons

If the assertion view of reason-giving is correct, then anti-psychologists rightly hold that motivating reasons are non-psychological propositions. But is the assertion view correct? In the literature, there is little by way of explicit argument that would support the claim that first-person reason-giving acts are reports of reason-relations. Perhaps the assertion view is taken to be the natural position, not in need of corroboration or even proper articulation. And indeed, what else should reason-giving be, if not a form of asserting a thesis on the explanation, justification, or motivation of the action? After all, Mary’s utterance of (1) looks like the report of some reason-relation between a proposition and an action.¹⁴

Compare the case in which Mary gives a reason with a case in which she undisputedly explains her action and thus asserts some relation between a certain fact and her action. Suppose Mary has attended a lecture on, say, neurophysiological action explanation and now uses her newly acquired knowledge when she accounts for her going to London with the words,

(2)	I go to London because my XYZ-neurons have fired.

Mary cites the firing of her XYZ-neurons, a non-psychological state of affairs, as the cause of her action. The proclaimed explanans is that Mary’s XYZ-neurons fired, not that she believes so. Of course she also believes that her XYZ-neurons fired. But this belief is necessary only for entertaining (and sincerely asserting) the explanatory thesis; it does not figure as part of the explanation provided.

As Mary’s reason-giving (1) has the same grammatical form as the physiological explanation (2), it is tempting to construe it along the same lines and thus to adopt the view that Mary, in uttering (1), provides a thesis on some relation too: accordingly, Mary asserts that John’s being in London explains, motivates, or speaks in favour of her action, just as she asserts that the firing of her XYZ-neurons is the cause of her going to London. That (1) and (2) have the same grammatical form seems to support the assertion view of reason-giving and hence the view that (1) is the articulation of a hypothesis relating Mary’s action to John’s being in London.

The analogy soon breaks down, however. While Mary may qualify her physiological explanation with, say, ‘supposedly’, ‘probably’, or ‘certainly’ (as in ‘Probably I go to London because my XYZ-neurons fired’), she cannot so modify her reason-giving utterance. Modification with sentential adverbs presupposes theoretical distance: Mary would have to be conceived of as somewhat alienated from the action or decision: as if she first realises that she performs some action or has some intention (‘Oh, interesting, I am about to go to London’), and then, after pondering various possible explanations (‘Why am I doing this?’), sets out to provide some positive account (‘Now I’ve got it: probably I go to London because John is there’). Yet this would mean to drastically and dramatically misconstrue the agent’s attitude towards her own action. Mary has a characteristic first-person agent perspective: she gives a reason for her intending, she does not explain some intention she finds and contemplates.¹⁵

Let me hasten to add that there are indeed situations in which an agent reflects on her action or intention and wonders about its explanation. An agent might find herself on the way to London, dressed for a journey, purse and ticket in hand, and begin to wonder why she is about to go there. And she might, after some deliberation, hypothesise that it’s ‘because John is there’. In such a case, the affirmation of (1) is not an act of reason-giving but, like (2) in the situation discussed above, an assertion about an explanatory relation. Yet such a situation, possible as it is, is not the one that we currently investigate (and which we usually have in mind). In the situation we examine, Mary does not adopt a theoretical stance from which she would be able to explore, and then articulate, with more or less confidence, a hypothesis on the explanation of her action or intention: she simply gives a reason for her action. And this is a very different thing.

The difference between reason-giving and explaining manifests itself in the patterns of rational update behaviour. Consider explanations first. An agent can only (rationally) believe that the explanandum is explained by the explanans if she believes that both are actually the case. Hence, if the agent gives up her belief in the explanans, she gives up her belief in the explanation. Suppose, for example, that Mary learns that the professor was wrong and that her XYZ-neurons have actually not fired. She then no longer believes that her action is explained by the firing of her XYZ-neurons.¹⁶ Her belief in the explanandum remains unaffected, however: upon realising that this particular physiological account is unavailable, Mary does not abandon (or even weaken) her conviction that she intends to go to London. On the contrary, it’s because Mary continues to think that she is about to go to London that she is prepared to look for an alternative explanation once she abandons the original one. If Mary comes to believe that her XYZ-neurons have not fired, she gives up the explanatory hypothesis (2) and looks for a new explanation.

If the assertion view were correct, Mary’s reason-giving by reference to John’s whereabouts would have to display the very same update pattern: upon giving up the idea that John is there, she would have to abandon a hypothesis: that her going to London is explained (or justified, or motivated) by the fact that John is there. Mary would then search for a new hypothesis regarding the explanation (justification, motivation) of her intention of going to London, and would hence still believe that she intends to go to London. (If she didn’t believe that she intends to go to London there would, after all, be no reason to search for an explanation.) What she would not do, however, is abandon this intention. Her reaction to the information that John is in London would not be to change her action plans, but to revise the hypothesis on the explanation (justification, motivation) of these plans. Yet this is not the rational update pattern if we understand Mary’s utterance as an act of reason-giving.

Suppose Mary’s belief that John is in London is based on Julie’s testimony (‘I saw John on Piccadilly’), but Mary later finds out that Julie mistook Jim for John, and therefore abandons the belief that John is in London.¹⁷ When Mary stops believing that John is in London, it would be quite odd if she continued with her preparations and revised her views on the explanation or motivation of her action. We surely don’t expect Mary to go on packing and then, taking her bag, purse, and keys, make her way to the station, all the time wondering, ‘Why am I doing this? It’s not because of John. So is it perhaps because of Joan?’ Rather, we would expect Mary, as soon as she finds out about Julie’s mistake, to unpack her clothes, sit down and, perhaps with a sigh, prepare a cup of tea: Mary gives up her intention to go to London.

Upon abandoning her belief that John is in London, Mary will not look for a different explanation for her intention, but rationally give up that intention.¹⁸ If this is correct, believing that John is in London is not a condition that would enable Mary to explain some previously ascertained intention; it is a condition allowing her to rationally form or sustain the intention to go to London. This would also account for the fact that, if she gives up that belief, she gives up the intention, too. Importantly, while the new information on Julie’s mistake affects Mary’s plans to go to London, it does not affect the connection between her intention and her belief. On the contrary, Mary’s intention to go to London is based on the belief that John is there. And this is precisely why her intention to go to London vanishes with the belief that John is there. That this connection is preserved can be seen from the fact that, if Mary regains confidence in the proposition that John is in London – e.g. because of some other friend who can tell John and Jim apart – Mary would start packing again.

My analysis is confirmed if we consider not only change of categorical belief but change in degrees of belief. So let’s rewind a bit and suppose that Mary does not yet know that Julie mistook Jim for John, but is merely informed about the fact that John has a doppelganger, Jim, and that it is hence an open question whether Julie’s report (‘I saw John at Piccadilly’) can be trusted. Mary does not yet give up the belief that John is back, but her credence is significantly lowered. As her credence in John’s being in London decreases, Mary’s willingness to undertake the journey will likewise diminish. She might still want to go, but the more doubts she has about John’s actually being in London, the less sacrifices will she be prepared to make. We can also describe this in decision-theoretic terms: Mary’s expected utility of going to London is determined by her desire to see John and her credence in John’s being in London. If her credence in John’s being there decreases, so does the expected utility (and hence the firmness of her intention) of going to London.¹⁹

To sum up: The assertion view is incorrect. In an act of reason-giving, the agent does not assert that some (alleged) fact (‘John is there’) explains, justifies, or motivates a certain action (‘I go to London’). She does not report some reason-relation. Premise (P2) is therefore unsupported and the argument from reason-giving deprived of a crucial premise. Yet we can draw a positive conclusion, too: reason-giving is to be construed as providing a rational ground for an intention. In Sections 4 and 5, I will explore the notion of motivating reasons favoured by such a grounding-view of reason-giving.

4. An expressivist analysis of the reason-giving ‘because’

A theory of reason-giving as rational grounding must account for the following: First, the action, or rather the intention, is not necessarily an object of theoretical reflection. The agent does not, or at least need not, think about her intention. A fortiori she does not provide a thesis on its explanation, justification, or motivation. Second, the agent presents her intention to go to London as grounded in her belief that John is there; the intention depends not on some non-psychological state of affairs but on some doxastic state of mind. I will now propose a (tentative) analysis of rational grounding based on the theory of expressive acts. I begin with some central ideas.²⁰

In seriously and competently affirming a sentence ‘p’, the agent asserts the semantic content of that sentence, i.e. she asserts that p. The semantic content of the uttered sentence is the proposition expressed by the sentence. In addition, she performs an expressive act: she expresses the belief that p. The expressive content of the act is the mental state expressed by the speaker.²¹ Both contents are systematically related: the semantic content is the content of the mental state expressed, i.e. the content embedded in the expressive content. With an assertive utterance ‘John is in London’, Mary asserts that John is in London and expresses her corresponding belief. This duality of contents is particularly easy to portray for assertions, but it is a general feature of language. To prepare for my discussions below, consider another example: by uttering ‘I’ll go to London’, Mary not only refers to the object of her intention – her going to London – but also expresses her intention to carry out that action.²²

Let me apply the theory of expressive acts first to the case in which the agent asserts an explanatory thesis. When Mary states (2), she proposes a physiological explanation of her action and thereby expresses a belief with the complex content that she goes to London because her XYZ-neurons fired. The word ‘because’ is used to conjoin two sentences, ‘I go to London’ and ‘My XYZ-neurons have fired’, in order to combine their semantic contents into a complex whole. That is, the connective ‘because’ is used to form a causal hypothesis which then figures as the content of the doxastic attitude expressed by the subject in performing the explanatory act. Mary expresses the belief that her going to London is explained by the firing of her XYZ-neurons. Due to the factivity of ‘because’ the belief in the explanation presupposes that Mary believes the explanans as well. As soon as Mary stops believing that her XYZ-neurons have fired, she as a rational agent gives up the belief that the firing of her XYZ-neurons explains her going to London. Thus, my analysis fully accounts for the update behaviour associated with the articulation of explanatory hypotheses.

What happens when Mary gives a reason, i.e. presents her ground for the intention to go to London? If Mary, with (1), were to report some reason-relation, she would express a complex belief. She would express a single doxastic state with the content that she goes to London because John is there. And the more she came to doubt that John is in London, the less she would be convinced of this explanatory hypothesis: if she finally realised that Julie was wrong and came to believe that John is not actually in London, she would abandon the belief in this hypothesis altogether and search for a new hypothesis regarding the explanation (or justification or motivation) of her intention.

Yet, as Mary grounds, not explains, her intention, the assertive reading cannot be the correct take on (1). I conclude that in this case the word ‘because’ does not function as a sentential connective which conjoins two semantic contents, that Mary goes to London and that John is there, into a complex whole. Rather, I suggest, ‘because’ has a paratactic use: it concatenates two sentences which must be considered independently.²³ I therefore propose an expressivist analysis of rational grounding. The function of ‘because’ is to pragmatically relate, not the semantic, but the expressive contents of the two adjacent utterances: with ‘I’ll go to London’, Mary expresses her intention to go to London; with ‘John is there’, she expresses the belief that John is in London; and with ‘because’, finally, Mary expresses that this belief is her ground for the intention. That is, in reason-giving, the agent does not express a single propositional attitude with a complex proposition as its content; rather, she expresses two propositional attitudes, in this case, an intention and a belief, and uses ‘because’ to express the former’s being rationally grounded in the latter: with (1), Mary expresses that her intention to go to London is based on the belief that John is there.²⁴

It is crucial to observe that Mary does not assert that her belief grounds her intention: Mary does not speak about the rational dependence of her intention on her belief. Nor does she express a belief in such a rational dependence. The relation between her belief and her intention is not the object of Mary’s concern; the objects of her concern are not even her belief and her intention themselves. Mary does not necessarily believe that her belief that John is in London grounds her intention of going there. And if she were to entertain this complex belief, this would surely not be what she expresses when she gives a reason with an utterance of (1).²⁵

My analysis explains why Mary begins to waver in her decision to go to London as soon as she discovers that John has a doppelganger: the basis of Mary’s intention is the belief that John is in London, and as this belief wanes, so does the intention, based on this belief, of going to London. This also explains why she cancels her trip altogether once she is informed about Julie’s mistake: Mary then gives up the intention-grounding doxastic state. Precisely because Mary’s intention of going to London is based on the belief that John is there, she is required to abandon the intention if she learns that Julie has mistaken Jim for John. And for the very same reason, Mary would pack again if she learned from some other source that John is in London after all.

The expressivist analysis of rational grounding fully vindicates Mary’s rational reaction to new information relating to John’s whereabouts. I will therefore presuppose this analysis in my discussion. It should be noted, however, that I do so in the present context only to offer some positive theory of rational grounding: my account of reasons in the next section is independent of the expressivist analysis of rational grounding; it requires only that the assertion view is wrong, i.e. that reason-giving is not the articulating of an explanatory hypothesis, and that the agent’s intention is dependent on her doxastic state in the manner discussed in the last two sections.

5. The reason in ‘because’

Let us return to our original question. What is the reason Mary gives for her trip to London? On the face of it, the grounding-conception of Mary’s reason-giving allows for two different answers. According to the first, the reason Mary gives is identical with the object of her concern, that is, the semantic content of the sentence following ‘because’: the reason is the proposition that John is in London, i.e. whatever Mary speaks and thinks about in a reason-giving act. This would yield premise (P2) and hence the thesis of anti-psychologism. The alternative, psychologistic proposal would identify the reason that Mary gives with the expressive content of the utterance following ‘because’, i.e. her belief that John is in London.²⁶ This alternative replaces (P2) with

(P2′)

The reason Mary gives is (identical to) her belief in the proposition she is concerned with.

Which of the two candidates is the right one? It is no use to insist on behalf of the anti-psychologist that Mary thinks only of John and not of herself. Proponents of (P2′), like the advocates of (P2), can (and should!) fully accept premise (P1). When Alvarez remarks that she is interested only in her cousin’s needs, and Henning insists (referring to an example of Parfit’s) that ‘[i]f you believe that there is a fire in your hotel, the last thing that will be of significance to you is the presence of that mental state’ (Henning 2018, 3; cf. 85), that is surely correct,²⁷ but it does nothing to decide between (P2) and (P2′). The question is not what Mary is concerned with, but what reason she gives for her action when, in giving the reason, she is concerned with John’s whereabouts. The choice between (P2) and (P2′) must be made on independent grounds.

At first it may appear that there is a stand-off: that psychologism and anti-psychologism are on an equal footing since, in reason-giving, the agent expresses a doxastic attitude which in a sense involves both the believing and the proposition believed; it may therefore appear that the choice between (P2) and (P2′) has to be made with recourse to considerations independent of (first-person present-tense) reason-giving, or even that it is a matter of mere stipulation.²⁸ Once we review the matter, however, it turns out that the grounding-conception is not in any way neutral on this point.

When Mary, by uttering (1), makes it clear that her intention is based on her believing that John is in London, the word ‘because’ does not function as a sentential connective whose sense relates the semantic contents of the concatenated sentences; it is used to indicate that the expressive contents of the two adjacent utterances, an intention and a belief, stand in some grounding relation. ‘Because’ is used to operate on expressive, not semantic, content, which shows that Mary bases her intention on a belief, not on some non-psychological proposition. If we stick to the natural assumption that the reason given is whatever ‘because’ is used to operate on, the reason Mary gives is the doxastic state she expresses with ‘John is in London’.²⁹

Once we replace the assertion view of reason-giving with a position according to which ‘because’ is used to operate on propositional attitudes, the reason Mary gives turns out to be the belief, not the proposition, that John is in London. We must therefore replace (P2) with (P2′). From this, together with (P1) and (P3), we conclude that Mary’s motivating reason is her belief that John is in London.³⁰ The agent’s reason is what she expresses, not what she is concerned with. The observation that Mary, in deciding what to do, considers only where John might be does not call for a non-psychologistic view of reasons, but only for the realisation that the assertion view of reason-giving must give way to the grounding-view.

Anti-psychologists worry that, in saying that Mary intends to go to London because ‘she believes that John is in London’, we – the philosophical theorists – fail to do justice to her motivation, because when she gives a reason for her action or contemplates what to do, she does not think or speak about her own belief.³¹ The charge is that psychologists misconstrue the agent’s reasons by referring to doxastic states in their account, while the agent herself refers to non-psychological propositions. Yet, a little reflection shows that this charge is unwarranted, rooted in the failure to acknowledge that there are different linguistic forms associated with assertion and expression, respectively: Mary and we use different linguistic devices to indicate the very same reason.

Nobody but Mary is able to ground her intention in her belief. And nobody but Mary is in a position to express her belief, which she does by asserting its content. Mary expresses her reason, the belief that John is in London, by reporting the (alleged) fact that John is there. Rational grounding and expressing a mental state are both reserved to the agent. (That’s why I have restricted myself to the first-person present-tense case.) We, on the other hand, cannot express another person’s reasons; we can only report them. And if we do so, we must refer to her beliefs. The very same reason that Mary expresses with (1) is reported by us with the words ‘because Mary believes that John is London’. Hence, in describing Mary’s reason by reference to her belief, we do not fail with respect to the task of giving ‘the agent’s own perspective on things, so as to reveal the light in which the action was done’ (Dancy 2000, 108): there is nothing left out or, for that matter, illegitimately added, if we refer to Mary’s belief in explaining her action. Likewise, when we say of the generous donor that she gives money ‘because she thinks her cousin needs the money’, or of Parfit’s agent that she jumps out of the window ‘because she believes that there is a fire’, we represent the agent’s motivation by referring to her mental state and claiming that it is this mental state which motivates her intention or action.

To emphasise: When we talk about an agent’s reasons and hence about her doxastic states, we do not suggest or imply that, in her own act of reason-giving, the agent would have to think or speak about her own believing. On the contrary! When we explain Mary’s action with the words ‘because she believes that John is in London’, we articulate the view that Mary, in giving her reason, is concerned specifically with John’s whereabouts. It is quite ironic, but it is only because anti-psychologists are right with respect to this part of the phenomenology of reason-giving, in particular right in their insistence on (P1), that classical psychologism can be the correct theory of motivating reasons in the first place.

6. Non-psychological explanations

Up to this point, I have focused on first-person (present-tense) acts of reason-giving and found that their full linguistic phenomenology, including the fact that the agent thinks about some worldly state of affairs and not about her own state of mind, actually supports psychologism. Yet accounting for actions is not limited to the first person. So let me briefly consider whether third-person action explanation holds an anti-psychologistic promise.

Marty might explain Mary’s action with the words,

(3)	Mary goes to London because John is there,32

which is grammatically nothing but the third-person variant of (1). The ‘because’-phrase in (3) is even identical to that in (1). Like Mary, Marty speaks only about John’s abode, not about Mary’s mindset. It might therefore be tempting to derive support for anti-psychologism from (3). In a nutshell, the argument would go as follows. Marty, like Mary, resorts to the non-psychological proposition that John is in London and thus asserts that this proposition is Mary’s reason. Mary’s alleged reason is hence John’s being in London, not the belief that John is in London.

Of course I cannot reject this line of reasoning on the same grounds I used to undermine the argument from first-person reason-giving. Marty’s utterance of (3) is clearly an assertion of an explanatory hypothesis. It indeed reports an explanatory relation between Mary’s action of going to London and the fact that John is there. This is confirmed by the update argument: if Marty were to give up the belief that John is in London, he would also give up the belief that (3) is true. Utterance (3), like (2) and unlike (1), articulates an explanatory hypothesis. The connective ‘because’ is used here to operate on the semantic contents of the adjacent sentences and hence also on the proposition that John is in London.

If there is a problem with the argument from third-person action explanation, it must be the assumption that (3) describes Mary’s reason. Many philosophers have made this assumption.³³ I will argue that this assumption is mistaken.

When Marty affirms (3), he cites the fact that John is in London as an explanatory antecedent of Mary’s action: John’s being in London is claimed to be at the beginning of some train of events the end of which is Mary’s action. Yet the truth of this claim is perfectly consistent with very different developments in between:

(S1)	John is in London; Julie sees John on Piccadilly and tells Mary so. As Mary wants to see John, she packs for London. Upon being asked why she intends to go there, Mary replies with (1): ‘I go to London because John is in London.’
(S2)	John is in London; Julie sees John on Piccadilly and tells Mary so. Yet Mary does not care about John, she only wants to see Joan. Knowing this, and to get her to London anyway, John persuades Julie to also convince Mary of Joan’s presence in the capital. Upon being asked why she intends to go there, Mary replies: ‘I go to London because Joan is in London.’
(S3)	John is in London, and to lure Mary there, John persuades Julie to hide the fact that he is there and instead to convince Mary of Joan’s presence. Upon being asked why she intends to go to London, Mary again replies: ‘I go to London because Joan is in London.’

Marty’s explanation (3) is correct in all three situations, in particular also in situations (S2) and (S3): it is John’s being in London which ultimately generates in Mary the belief that Joan is in London, which in turn is responsible for her intention to go there. But this shows that, despite the grammatical analogy, there is a striking asymmetry between Marty’s utterance of (3) and Mary’s utterance of (1):³⁴ while (3) is true in all of these situations, (S3) precludes Mary’s sincere utterance of (1), as she lacks the belief that John is in London. And in situation (S2) this belief, while present, does not ground her intention of going to London: the alleged fact that John is in London is irrelevant, her intention being based on the belief that Joan is in London. Mary can sincerely utter (1) only in situation (S1), in which her intention of going to London is based on the belief that John is there. Yet even in this situation, Mary’s sentence (1) and Marty’s sentence (3) are used in utterly different ways. Mary uses these words to give a reason, Marty employs them to explain an action. This is why, if John were not in London, (3) would be false, while Mary’s reason-giving (1) would still stand: even if John happens not to be in London after all, her action is still grounded in her belief that John is there. And this is what she expresses with (1).

For (3) to be true, Mary need not believe that John is in London. John’s being in London may explain Mary’s going there in the way described by (S3). And in cases in which she entertains the belief that John is in London, (3) can be true without making use of this belief: the belief that John is in London need have nothing to do with her motivating reason (see (S2)). Marty’s explanation of Mary’s action by reference to John’s whereabouts does hence not, as such, refer to Mary’s John-related reasons. In fact, it does not, in itself, suggest that Mary’s action is based on any reason at all, as a fourth scenario shows:

(S4)

John is a scientific genius. Upon his return to London, he is finally able to control Mary’s XYZ-neurons and make them fire in such a way that Mary forms the intention of going where he is. Upon being asked why she intends to go to London, Mary is unable to provide a reason. She can only speculate on some possible causal explanation with the words: ‘I am not sure. Perhaps because my XYZ-neurons fired.’

In situation (S4), Mary has no motivating reason for her action at all; there is only a cause for her action: Mary goes to London because her XYZ-neurons fired. And these XYZ-neurons fired because John is in London. So, Marty’s claim (3) is true also in this case, even though Mary does not have any reason for her action. It follows that Marty’s explanation (3) cannot be understood as referring to Mary’s reason, let alone some possible John-related reason. The anti-psychologistic argument from third-person action explanation fails. Marty, while thinking and speaking about John’s whereabouts in explaining Mary’s action, does not explain that action by reference to her reason.

This is not to say that sentence (3) might not be used to indirectly point to Mary’s reason. If, e.g. the conversation is concerned with Mary’s motivation for going to London, and it is a salient possibility that Mary wants to see John and that his presence in London would not go unnoticed by Mary, Marty, by uttering (3), might well (be taken to) implicate scenario (S1). And perhaps it is situations such as these that philosophers have in mind when they claim that explanation (3) refers to Mary’s reasons. Note two things, however. Firstly, Marty’s reference to Mary’s reason is then a matter of Gricean pragmatics, not of semantics. As Marty’s implicature is highly context-dependent, we cannot infer anything about Mary’s reasons from the explanation (3) alone: If, e.g. it were contextually salient that Mary is interested in Joan, not in John, and that John, once in London, would do anything to get Mary to London, deceptions included, (3) might be taken to implicate either (S2) or (S3) instead. And if it were contextually salient that John is an evil genius of the sort described in (S4), we might take Marty’s (3) to implicate scenario (S4) and hence that Mary does not have any reason for her action at all. Secondly, even if we focus on cases in which (3) generates an implicature about some John-related reason in the spirit of (S1), this does nothing to support anti-psychologism, as the implicature need not be some non-psychological proposition. In my view, of course, if the implicature is in the spirit of case (S1), Marty implicates that Mary believes that John is in London. If Marty were then to avoid all pragmatic vagaries and to flat-out report Mary’s John-related reason, he would have to say, ‘Mary goes to London because she believes that John is there.’

7. Conclusion

To the question ‘Why do you/does Mary go to London?’, Mary and Marty may both respond with ‘Because John is there’. Correctly observing that Mary gives a reason for her action and that Marty explains her action, anti-psychologists conclude that there are cases in which the reason addressed is the non-psychological fact or proposition that John is in London. They construe Mary’s reason-giving act along the lines of an explanation, and thus feel confident in the cogency of the argument from first-person reason-giving. And they interpret Marty, in his explanation, as referring to Mary’s reason, and therefore wield the argument from third-person action explanation.

I have argued that these considerations are unsound, based on a conflation of reason-giving and explaining proper, a conflation no doubt furthered, if not actually brought about, by the fact that Mary and Marty use the very same ‘because’-phrase. Once the two cases are disentangled, it is obvious that case (3) does not support the view that reasons are non-psychological facts or propositions. Marty uses ‘because’ to operate on the semantic content and, hence, on a non-psychological proposition. Yet his explanation has, considered in itself, nothing to do with Mary’s reasons. And when Mary uses the ‘because’-phrase to give a reason, ‘because’ operates on expressive content and thus on mental states. There need not be a reason in ‘because John is in London’. But if there is, it appears to be the agent’s belief that John is there.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This work was supported by the Fritz Thyssen Stiftung.

Notes

1 For proponents of psychologism – the view that reasons are mental states – see, for instance, Davidson (1963); Smith (1987; 1992); Wallace (2003). For psychologism about theoretical reasons specifically, see in particular Turri (2009) and Grajner (2016).

2 I won’t, for example, discuss Dancy’s ‘normative constraint’ (2000, 103) and the reasons associated with it. For discussions of the relation between motivating and normative reasons, see Wallace (2003) (from a psychologistic perspective) and Mantel (2017) (from an anti-psychologistic perspective).

3 For anti-psychologistic accounts of epistemic reasons, see, e.g. Stampe (1987); Skorupski (1997); Collins (1997); Hyman (1999); Dancy (2000); Williamson (2000; 2017); Comesaña and McGrath (2014); Glüer and Wikforss (2018); Littlejohn (2018). A very early advocate seems to be Brown (1955, 353).

4 Mantel (2016, 89), for example, gives an argument of this sort in favour of anti-psychologism with respect to normative reasons.

5 I realise that this assumption may be controversial. I make it mainly for purposes of simple and straightforward presentation and trust that the main point holds up even if this assumption were rejected.

6 Variants of factualism can be found in Collins (1997) and Dancy (2000; 2005). For variants of propositionalism, see, e.g. Alvarez (2010; 2017); Comesaña and McGrath (2014); Williamson (2017); Mantel (2017); Henning (2018); Hawthorne and Magidor (2018).

7 The ‘error problem’ is the main problem for the factualist version of anti-psychologism: Mary appears to have a reason for going to London even if John is not in fact there. For a discussion, see, e.g., Collins (1997); Dancy (2000); Alvarez (2010); Comesaña and McGrath (2014); Henning (2018). A particularly pressing problem for propositionalism is that the reason-operator, like the ‘because’-operator, seems factive, which would preclude the possibility of describing false propositions as reasons. For a discussion of the problem, and a sophisticated defense of propositionalism in this respect, see Henning (2018). (I dispute one of the central assumptions of Henning’s defense in Freitag 2024.)

8 Of course, Mary may use alternative linguistic forms to give a reason. In response to the ‘why’-question, she may simply say: ‘John is there,’ thus pragmatically indicating that with the utterance of this sentence she gives a reason. Note that my expressivist analysis – see Section 4 below – essentially applies to this case as well. Observe also that reason-giving acts in the form ‘My reason is that John is there’ do arguably not indicate that reasons are non-psychological facts or propositions, either. Pryor (2007) plausibly claims that the ‘is’ in ‘my reason is’ does not express identity, but specifies the propositional content of the reason, thus the content of the belief expressed.

9 See also the Othello example in Alvarez (2017), Sect. 3.1. Considerations of this sort are frequently found in the literature, e.g. in Collins (1997); Wiland (2018, 203); Henning (2018, 163); and, with respect to normative reasons, in Mantel (2016).

10 Sometimes, of course, the reason-giving agent may be concerned with her own state of mind as well. If a person sees a shrink to be cured of her paranoia, she might give a reason for her action with the words ‘because I think I am being followed.’ The reason then is a second-order belief, the belief that she thinks that she is being followed; the reason is the agent’s belief that she has a mental disorder that needs a cure. Interestingly, an utterance of ‘I think that p’ is expressively ambiguous. As I have argued in a number of papers, it can also be used to express a first-order belief. Thus ‘because I think that I am being followed’ can also be used as giving a reason for going to the police. I cannot here pursue this interesting complication, but see Freitag (2018); Freitag and Kraus (2022); Freitag and Yolcu (2021); and Freitag and Bräuer (2022) for an explanation and defense of the expressive ambiguity of ‘I think.’

11 Cf. Collins (1997, 120–121); and Henning (2018), 3–4.

12 Compare also Dancy (2000, 15 and 108–109).

13 This might be the idea behind the anti-psychologistic treatment of Mary’s belief as a mere ‘enabling condition’ for explanations (for a discussion, see Dancy 1995, 427; 2000, 127–128; compare also Alvarez 2010, 26; 2017, Sect. 3.2; and Henning 2018, 172–173).

14 Anti-psychologists treat the first-person present-tense case in analogy to third-person cases (as in ‘Mary goes to London because John is there’) and first-person past-tense cases (‘I went to London because John was there’). As these latter cases are indeed to be analysed as reports of explanatory relations (see Section 6 below), this yields additional support for the attribution of the assertion view to anti-psychologists.

15 For the distinction between first- and third-person perspectives with respect to reason explanations, see also Wallace (2003, 431–432).

16 The same result is obtained if the speaker, though maintaining the explanans, gives up the belief that the explanans causes the explanandum, e.g. when she comes to realise that the (alleged) explanans occurs only after the explanandum.

17 Note that Mary need not disbelieve that John is in London. It suffices that she gives up the belief that John is there.

18 I here assume that Mary does not have further sufficient reasons to go to London. Similarly Dancy (2000, 127): ‘in the absence of the believing … the action would not then have been done at all, or … if it had, it would have been done for another reason.’ With respect to the epistemic case, Williamson (2000, 62) makes an analogous observation.

19 Needless to say, what is true for practical reasons holds also in the case of theoretical reasons. If Mary’s reason for believing that John has returned from the States is given by (1), her belief that he is back in Europe weakens once she realises that John has a doppelganger. And as soon as Mary abandons altogether her belief that John is in London, she also abandons her belief that John is back from the States.

20 In the present paper, I assume that ‘express’ (in the case of speaker expression) and its cognates refer to a non-factive relation; a mental state may be expressed without itself being present (e.g., when the speaker is insincere). My main argument is, however, independent of this assumption. For other non-factive conceptions of ‘express,’ see Austin (1962, 14 ff.); Searle (1969, 65); Harnish (1976); Kemmerling (2002); and Davis (2003). For advocates of a factive understanding of (self-)expression, see, e.g., Green (2007); Williams (2013). I here provide only a sketch of the expressivist theory, which has been presented and elaborated in Freitag (2018); Freitag and Kraus (2022); Freitag and Yolcu (2021); Freitag and Bräuer (2022).

21 I will here describe expressivism in its simple, statist form: the expressive content of an utterance is a mental state. As I will later declare reasons to be expressive contents in reason-giving acts, this yields a statist view of reasons. According to an alternative conception, expressive contents are propositions about mental states, e.g. the proposition that Mary believes that John is in London, which would yield that reasons are psychological propositions. (Propositional expressivism is discussed in Freitag and Yolcu 2021.) Since my main aim is to save psychologism from the argument from reason-giving, it is presently immaterial whether expressive content is conceived as a psychological state or as a proposition describing such a psychological state.

22 I here assume that ‘I’ll go to London’ is indeed the expression of an intention and not a prediction.

23 For the suggestion of a paratactic understanding of ‘because’ in its reason-giving use, see Breul (1997); Couper-Kuhlen (1996); Catasso (2015). Note that I do not propose a semantic ambiguity of ‘because.’ For the purposes of the present paper, I remain neutral in this regard. I claim that ‘because’ has two uses, which results in two different possible interpretations of ‘because’-utterances.

24 In this last sentence, I transcend the limitations of a statist conception of expressive content (see fn. 21 above). For reasons of simplicity, I will not here enter the intricacies of propositional expressivism.

25 I do not even exclude the possibility that persons can give reasons even though they do not think about their own first-order mental states at all. They may be able to give reasons, much as persons who don’t know the laws of statics are able to build a supporting wall.

26 ‘Because’ would still be ‘factive,’ but ‘factivity’ would now concern the propositional attitudes expressed by the adjacent sentences.

27 This is correct at least in the sense that the agent should not or need not pay attention to her own mental state.

28 Cunningham (2022) suggests that the dispute between psychologism and anti-psychologism might be a merely verbal dispute. Mantel (2016, 85) seems to take the disagreement as being a matter of stipulation and distinguishes between a ‘motivating consideration’ and a ‘motivating reason.’ For the suggestion of a terminological differentiation, see also Wallace (2003). With respect to normative reasons, Sylvan and Sosa (2018, 559) also assume that it’s a matter of linguistic convention.

29 That, in reason-giving, the speaker expresses her reason (her belief), is also suggested, in passing, in Turri (2009, 507).

30 This result ties in with the observation that reasons, like beliefs, come in degrees of strength. Before Mary gives up the belief that John is in London, her reason is weakened by the realisation that John has a twin. Mary may still have a motivating reason for going to London, but the reason is diminished by this information. Note that it is hard to reconcile anti-psychologism with the idea that reasons might come in different strengths.

31 Henning, for example, says: ‘[I]f we cite other facts … that hold independently of the agent’s considerations [such as the fact that the agent believes that p; W.F.], we fail to present the considerations in the light of which she acted’ (2018, 169; his emphasis). Collins even charges the psychologist with ‘withdrawing the explanation’ the agent herself gives (1997, 108).

32 I here discuss only third-person explanations. The same type of analysis would apply to first-person past-tense explanations (as in ‘I intended to go to London because John was there’) and to first-person present-tense cases used as explanations.

33 The assumption is made, e.g., in Collins (1997, 108); Dancy (2000, 119; 2005, 28); Wallace (2003, 433n1); Alvarez (2010, 18 and 29); Henning (2018, 163).

34 Again, I restrict myself to the reason-giving use of (1). If (1) is understood as an explanatory claim, its utterance by Mary is true if and only if (3) is true.