1 . Preliminary Clarifications ²

Let us start with some background. In what follows we will use the word ‘love’ to refer to romantic love, as in the example of Romeo and Juliet. Although there is no universally accepted definition of romantic love, it is widely assumed that romantic love is an attitude that a person, the lover, has towards another person, the beloved, and that exhibits distinctive emotional and behavioural patterns that do not typically occur in friendship, family relations, or other kinds of attachment. This assumption will suffice for our purposes. As it will turn out, the point we make about justification is compatible with different definitions of romantic love.

Since we understand love as an attitude, we take it to be distinct from the loving relationship that can obtain between the lover and the beloved. Romeo’s love for Juliet does not reduce to his romantic entanglement with Juliet. Love may cause or sustain a loving relationship, typically in the form of a socially regulated practice such as dating or marriage, but what matters for us is the attitude rather than the relationship. One may certainly wonder whether the initiation, continuation, or termination of a loving relationship is justified. But the issue here is whether love itself can be justified.¹

The justification of love essentially concerns the reasons for love, where the latter are understood normatively, as distinct from causes or motivations. To say that one’s love for a certain person is justified is to say one has reasons to love that person. Similarly, one’s love for a certain person is unjustified when one does not have reasons to love that person, or one has reasons not to love that person. This does not necessarily mean that, when one’s love for a certain person is justified, or unjustified, something follows about what one should or should not do. Normativity in the sense that matters here does not necessarily have prescriptive implications. What it requires is simply that love can be judged as right or wrong in terms of some suitable appropriateness condition.

But where do reasons for love stem from? Why should Romeo have reasons to love Juliet, rather than, say, Juliet’s cousin? The most natural answer to this question is that one’s reasons to love a given person are grounded in the properties of that person. Romeo would say that he loves Juliet in virtue of the way she is. Juliet’s cousin is not like Juliet in some respects that matter to Romeo. The underlying idea is that the justification of love essentially involves the properties of the beloved, given the preferences and inclinations of the lover. The property view, as we will call it, may be regarded as an instance of a more general principle that seems well-grounded, namely, that the justification of an attitude towards an object is based on the properties of the object.

At least three remarks will help to clarify the property view as we understand it. First, it is not part of the view that the justification of love depends on objectively valuable or generally appreciated properties of the beloved. Although it may actually happen that some properties of the beloved are objectively valuable or generally appreciated, nothing in our discussion of reasons for love will hinge on this fact. In particular, the properties involved in the justification of love need not be moral qualities. As far as the property view is concerned, features such as sex appeal, musicality, or physical prowess can be as relevant for the justification of love as honesty, generosity, or intelligence.²

Second, the qualification ‘given the preferences and inclinations of the lover’ is intended to suggest that the criteria in terms of which the properties of the beloved are evaluated ultimately depend on the features of the lover. To judge a property of the beloved as good or bad is to judge it as good or bad for the lover. This explains, among other things, the previous remark about moral qualities. Even though generosity is generally appreciated as a moral quality, nothing prevents one from valuing selfishness instead, and thus appropriately loving a selfish person. Of course, one’s reasons for loving a given person can also be assessed from a third-person perspective, rather than from a first-person perspective. But this does not mean that the properties of the beloved are to be judged according to the third person’s standards.

Third, the property view does not entail that the justification of love is transparent to the lover, that is, it does not entail that one is always fully aware of the reasons for one’s love. In this respect, there is no guarantee that first-person judgments are more accurate than third-person judgments. Lovers may easily be wrong about which of the properties of their beloveds justify their love. So we differ from those who assume that an attitude is justified by a normative reason only if the subject of the attitude is aware of that reason.³

The structure of the paper is as follows. Section 2 outlines the fact about irreplaceability that we take for granted as a starting point of our discussion. Section 3 considers three alternatives to the property view and explains why we do not regard them as viable options. Sections 4 and 5 draw two key distinctions: one concerns the very notion of justification, the other concerns the claim that beloveds are irreplaceable. Section 6 examines different readings of the latter claim and suggests that there is a plausible sense in which the property view does not entail that beloveds are replaceable, although there is also a sense in which the entailment holds. Section 7 provides some concluding remarks about reasons and properties.

2 . The Irreplaceability Intuition

Any attempt to provide an account of reasons for love faces a crucial problem. Love seems hardly justifiable on the basis of general truths of the kind that make actions or emotions justifiable. Consider fear, for example. When one is justified in fearing something, one’s justification is based on some clearly identifiable properties of the object of the fear. If you see a shark while you are swimming in the ocean, you are justified to fear that shark. Your justification is based on some properties of that shark -being a predator, having strong jaws, and so on- that generally make sharks dangerous for humans. This is why, if you were facing another shark, it would be equally appropriate for you to fear it. Love differs from fear in this respect. It is hard to identify some set of properties of Juliet such that Romeo would love any maiden with those properties. So, it is not clear how Romeo’s love for Juliet can be justified by appealing to such a set of properties.

The problem may be phrased in terms of fungibility. If x’s love for y is justified by certain properties of y, it seems to follow that any other person with those properties would be equally lovable for x. But it is very unlikely that x is willing to regard y as fungible in this way. A similar formulation of the problem can be given in terms of trading up. If x’s love for y is justified by certain properties of y, and z instantiates those properties to a greater degree, it seems to follow that y has a reason to switch y with z. But again, it is very unlikely that x finds it appropriate to do that. We will not distinguish between these two formulations, and talk generically about replaceability. Lovers tend to believe that their beloveds are irreplaceable, and there seems to be a grain of rationality in this belief.⁴

How strong and how reliable is the belief that beloveds are irreplaceable? Pretheoretical judgments do not always converge in this regard, so there is plenty of room for disagreement. But at least it can be taken for granted that replaceability does not hold in general; that is, it is not in general true that if one is justified in loving a person with certain properties, then one is thereby justified in loving another person with those properties. This assumption -which we will call the irreplaceability intuition- seems to state a relatively uncontentious fact on which philosophers and non-philosophers tend to agree.

Note that the irreplaceability intuition as we understand it does not say that replaceability never holds, or that it should not hold. Some philosophers, for example Goldie, have made claims along these lines:

But then let’s assume that James, who loves Mary, meets another woman, Rose, who is more elegant, more charming, more intelligent, and a better cook than Mary. Is this reason to change his affections to Rose, or at least to love her more than Mary? Surely not. And we can go further than that: surely there would be something wrong with his love for Mary if he were to change his affections and come to find Rose more lovable just for these reasons.⁵

However, we believe that such claims are too strong to be taken for granted on intuitive grounds. As far as we can see, there is no clear intuition to the effect that it is never reasonable, or that it is always wrong, to change the object of one’s love. Goldie’s example is potentially misleading, for it seems to imply that it is morally bad to stop loving Mary, or to break up with her, so it diverts attention from the main point, which concerns the reasons for loving Rose. If there is a moral issue about James and Mary -which concerns James’ constancy, dedication, or loyalty to Mary- it is different from the question whether James has reasons for loving Rose. As explained in section 1, we take the justification of love to be distinct from the justification of loving relationships, so any moral consideration about the latter need not be relevant for the former.

The fact that nothing stronger than the irreplaceability intuition can be assumed as uncontentious emerges clearly when one thinks about imaginary cases in which beloveds are replaced by duplicates that are indistinguishable from them. Imagine a replica of Juliet, call her Juliet*, a person that Romeo is unable to distinguish from Juliet. Romeo would certainly love Juliet* if he were unaware that she is a replica -say, if the Capulets secretly switched Juliet with her during the night- and it is equally plausible that his attitude could easily change once he realized that she is a replica. But does Romeo have reasons to love Juliet*? As far as we can see, there is no clear intuitive answer to this question. Until the question itself is made more precise, different answers seem equally plausible.⁶

3 . Some Alternatives to the Property View

Although it is undeniable that the irreplaceability intuition poses a serious challenge to any account of reasons for love, it would be wrong to regard it as a compelling argument against the property view. At most, it shows that a naive version of the property view is untenable. In this section we will briefly discuss three alternative views, none of which seem very promising to us.

The first option is simply to deny that love can be justified, endorsing the considerations against the rationality of love advanced by Frankfurt, Zangwill, and others. On this view, the properties of Juliet do not justify Romeo’s love for her, because nothing can provide such justification: there are no such thing as reasons for love.⁷

We will make no attempt to argue against this view, which questions from the very start the idea that love is subject to normative considerations. If one thinks that love is entirely irrational or a-rational, one will hardly be moved by considerations about the initial plausibility of this idea. Conversely, in order for our discussion of the irreplaceability intuition to be minimally interesting, it has to be granted at least as a working hypothesis that there are reasons for love. What we will say here about the property view may be regarded as conditional on that hypothesis.

The other two options are less radical, in that they reject the property view -or a naive version of it- without denying that love can be justified. One of them is to claim, following Kolodny, that reasons for love stem from the relationship between the lover and the beloved, rather than from the properties of the beloved. On this view, what Romeo really values is his relationship with Juliet, rather than Juliet’s properties. The loving relationship is the source of the normativity of love.⁸

We doubt that this view can provide a satisfactory account of reasons for love. As explained in section 1, the justification of love must not be confused with the justification of the loving relationship that can obtain between the lover and the beloved. On the face of it, is not even necessary to be involved in a loving relationship in order to have reasons for love. As Protasi has argued, unrequited love can be justified. One can have reasons to love a certain person even when there is no way of getting involved in a loving relationship with that person.⁹

A more specific way to deal with the irreplaceability intuition within the framework of the relationship approach is to appeal to the historical connection between the lover and the beloved. Although other persons may instantiate some properties of the beloved -say, elegance- they do not have the same “historical” properties, that is, they are not tied to the lover historically. If Alf and Betty share some experiences that are important for Alf -for example, they painted their house together- the value that Betty has for Alf may depend at least in part on those experiences. Therefore, Betty and her replica Betty* are not equal from his point of view.¹⁰

This account is plausible to some extent. The fact that Betty* has no historical connection with Alf may be taken to explain why Alf does not regard Betty and Betty* as interchangeable. However, the appeal to historical properties has its own problems. On the one hand, it would make little sense to assume that all the historical properties of Betty matter to Alf, because some historical properties of Betty are clearly irrelevant, and also because trivially no other person can have exactly the same historical properties as Betty, unless some form of haecceitism is assumed. On the other hand, if one takes any restricted subset of historical properties of Betty, one runs into exactly the same kind of worry that can be raised in connection with ordinary properties. Just as it is not obvious that Alf would love any other woman who is as elegant as Betty, it is not obvious that he would love any other woman who painted a house with him.¹¹

The third option is to claim that the justification of love crucially differs from the justification of emotions such as fear, because it stems from the individuality of the beloved. Romeo loves Juliet in virtue of her being Juliet, rather than in virtue of the way she is. As Kraut, Grau, and Goldie would put it, the justification of love is de re, as it is directed towards that very person.¹²

This view, however, does not seem to provide a viable alternative to the property view. One way to interpret the claim that love is de re is to say that the properties of the beloved do not matter. But this seems wrong. Suppose that Alf and Betty fall in love with each other, decide to live together, and after a year Alf starts spending every night at home playing videogames with his friends. If Betty loved Alf qua individual, that is, independently of his properties, she would have no reason to change her attitude towards Alf: Alf is still Alf, the same person she fell in love with. But many people would agree that if Betty stops loving Alf, she does it for a reason.¹³

Another way to interpret the claim that love is de re is to say that the individual value of the beloved ultimately depends on the beloved’s being loved by the lover. In this case the idea would be that Juliet is individually valuable for Romeo because Romeo loves her, and not the other way round.¹⁴

This interpretation, however, seems unable to answer the question of justification in the way we expect. How can Romeo’s love for Juliet be justified by the individuality of Juliet, as opposed to her properties, if the only intelligible sense in which the individuality of Juliet has a value for Romeo is that Romeo loves her? If someone claimed that your fear of the shark is justified by some negative value of the shark which ultimately depends on its being feared by you, that claim would definitely not provide an account of justification in the sense that we have in mind. The disanalogy between love and fear cannot be this, or so we believe.¹⁵

4 . Two Ways of Having Reasons

In order to elucidate the irreplaceability intuition and its implications for the property view, we will draw some distinctions that we take to be crucial. The first distinction concerns the very idea of justification. There are basically two ways to understand reasons for actions or emotions: in one case the point of view from which reasons are individuated is internal to the epistemic perspective of subject, in the other it is external. Therefore, the justification of an action or emotion can be understood either internally or externally.¹⁶

Imagine that Alf needs to go to the supermarket to buy some food. Alf has plenty of evidence for thinking that the supermarket is open -it is Friday, it is 5 pm, and so on- and no evidence for thinking that it is closed. So he goes to the supermarket and buys what he needs. In this case it is clear that Alf’s action is justified. But there are two ways to explain why. One is to say that Alf has a reason to go to the supermarket because he is justified in believing that the supermarket is open. The other is to say that Alf has a reason to go to the supermarket because the supermarket is actually open. According to the first explanation, the reason is internal to Alf’s epistemic perspective, in that it depends on the evidence available to Alf. According to the second, the reason is external to Alf’s epistemic perspective, in that it depends on facts about the supermarket.

To see that these two ways to individuate reasons are independent of each other it suffices to consider variations of the case just described in which one of the two conditions is not satisfied. Imagine that the supermarket unexpectedly closes at 4:59 pm because a dangerous snake escaped from the zoo has been seen hiding under a shelf. In this case Alf’s belief is justified but false. So, even though there is a sense in which he has a reason to go to the supermarket, the internal sense, there is another sense in which he has a reason not to go to the supermarket, the external sense. A different variation of Alf’s case may produce the opposite result. If Alf believes without justification that the supermarket is open -or does not believe it at all- but the supermarket is actually open, he has a reason to go to the supermarket only in the external sense.

A similar example can be provided about the justification of an emotion. Imagine that Betty is swimming in the ocean and she notices something that looks like the dorsal fin of a shark. Since she knows much about sharks, and she is aware that sharks often cross that area, she has sufficient evidence for thinking that there is a shark in front of her. As a matter of fact there really is a shark in front of her. Betty is frightened and swims away as fast as she can. In this case it is clear that Betty’s fear is justified. But there are two ways to explain why. One is to say that Betty has a reason to be afraid because she is justified in believing that there is a shark in front of her. The other is to say that Betty has a reason to be afraid because there actually is a shark in front of her.

As in the previous case, to see that these two ways to individuate reasons are independent of each other it suffices to consider variations of the case just described in which one of the two conditions is not satisfied. Imagine that what Betty is seeing is an innocuous robot controlled by a group of scientists through radio waves. In this case Betty’s belief is justified but false. So, even though there is a sense in which she has a reason to be afraid, the internal sense, there is another sense in which she has a reason not to be afraid, the external sense. A different variation of Betty’s case may produce the opposite result. If Betty believes without justification that there is a shark in front of her -or does not believe it at all- but there actually is a shark in front of her, she has a reason to be afraid only in the external sense.

As these examples show, the internal justification of an action or emotion requires that the subject has justified beliefs about some external conditions that are relevant for the action or emotion, or at least has evidence for such beliefs. On the other hand, the external justification of an action or emotion purely concerns external conditions that are relevant for the action or emotion, independently of what the subject is justified in believing. As we will suggest, the same distinction can be drawn in the case of love, although so far the debate on reasons for love has mainly focused on internal justification.

Perhaps a hint at this distinction surfaces in Brogaard’s account of rational love. According to Brogaard, rational love satisfies two constraints. The first -call it C1- rests on the notion of “proper fit”:

For romantic love to be rational, there must be a good fit between the beloved’s qualities and the loving feeling. [ . . . ] Many different sets of qualities can guarantee this type of proper fit. The set of qualities that I have and the set of qualities that you have may very well make both of us worthy of love.¹⁷

The second constraint -call it C2- concerns the accuracy of the perception that the lover has of the beloved:

For your love to be rational, you must accurately perceive the qualities of the beloved that form the causal basis for the love in question. [ . . . ] If, on the other hand, you are in love with your own fantastical creation of your beloved instead of your beloved as she really is, your love is irrational.¹⁸

Brogaard seems to take C1 and C2 as both necessary and jointly sufficient. She explains that there are cases of irrational love in which both constraints are violated, and cases in which only one of them is violated. Suppose that a wife believes that her husband is a generous and hard-working man, which explains at least in part why she loves him, but that in reality he is slowly poisoning her with arsenic so he can inherit her money. In this case both C1 and C2 are violated: there is no good fit between the husband’s qualities and the wife’s love for him, and the wife misperceives qualities of the husband that partially form a basis for her love. A case in which only C2 is violated, instead, is that in which the lover idealizes the beloved even if the beloved is perfectly lovable. Finally, a case in which only C1 is violated is that in which one’s love is unrequited, even if one does not misperceive the qualities of the beloved.¹⁹

First of all it must be noted that ‘accurately perceive the qualities of the beloved’ can mean either that the lover has justified beliefs about the beloved, or that the lover has true beliefs about the beloved. Accordingly, a violation of C2 may be understood either as a case in which the lover has some unjustified belief about the beloved, or as one in which the lover has some false belief about the beloved. We think that the most plausible reading is the first. There is a clear sense in which justified belief suffices for rationality, even in absence of truth. Suppose that the wife’s belief that her husband is a good man is justified because he is deceiving her so ably that she has no evidence to think otherwise. In this case we would say that the wife is rational, in spite of the falsity of her belief. On the reading just suggested, Brogaard’s two examples of violation of C2 are compelling. If the wife is unjustified in believing that her husband is a good man, she is irrational in some sense, independently of whether he is poisoning her.

The first example -that in which the husband is actually poisoning her- also provides a clear illustration of how C1 can be violated. The third example, on the other hand, is less convincing. As explained in section 3, we believe that unrequited love is not intrinsically irrational. If there is a clear case of violation of C1 without violation of C2, it is rather the one considered above in which the wife is justified in believing that the husband is a good man because he ably deceives her.

More generally, it seems that the plausible cases of violation of C1 are cases in which some beliefs of the lover about the beloved are false, independently of whether those beliefs are justified. Similarly, it seems that the plausible cases of violation of C2 are cases in which the lover has unjustified beliefs about the beloved, independently of whether those beliefs are true.

In the light of the distinction between internal and external justification, C1 and C2 may be regarded as criteria pertaining to two different but equally legitimate ways of understanding reasons. According to the internal understanding, if the wife is justified in believing that her husband is a good man, she has a reason to love him. According to the external understanding, if her husband is poisoning her, she has a reason not to love him. Brogaard seems to talk about these two senses of justification as if they were different components of a single sense. Of course, this might be just a different way of framing the same point, without any substantial disagreement. But in any case we think that our distinction helps to spell out the difference between C1 and C2.

5 . Interpretations of Irreplaceability

A second distinction that must be drawn concerns the claim that the beloved is irreplaceable. Suppose that Romeo is justified in loving Juliet, and that the Capulets secretly switch Juliet with Juliet* during the night. In section 2 we observed that, even though Romeo might love Juliet* as long as he is unaware of the switch, his attitude could easily change once he realizes that she is not Juliet. This observation may be understood in at least two ways. One is to take it as a statement about what Romeo is likely to believe about his reasons to love or not to love Juliet*, the other is to take it as evidence about those very reasons, independently of what Romeo has in mind.

Let x, y, and z stand respectively for the lover, the beloved, and the replica. The first reading of the claim that y is irreplaceable for x takes one or the other of the following forms:

(I1) and (I2) differ in that (I2) is stronger than (I1). Both claims are interesting, but they are psychological rather than normative. So they are not directly relevant to the question of justification. What matters to that question are x’s reason to love y, rather than x’s beliefs about such reasons. After all, x’s beliefs may be false.²⁰

The second reading of the claim that y is irreplaceable for x concerns x’s reasons to love y, independently of what x has in mind:

As in the previous case, (I4) is stronger than (I3). (I3) is definitely plausible, and we take it to be an adequate formulation of the irreplaceability intuition. Some philosophers may be willing to endorse (I4) as well. However, as noted in section 2, (I4) can hardly be taken for granted. So we will not deal with it. The irreplaceability intuition as we understand it says that it is not always right to switch one’s love to another person, it does not say that it is always wrong.

A further clarification concerns the term ‘replica’ which occurs in (I1)-(I4). Sometimes, those who appeal to the irreplaceability intuition seem to take for granted that z lacks some properties that ground x’s love for y. For example, in the case of Alf and Betty*, the point may be that Betty* does not have all the properties that ground Alf’s love for Betty, for Alf’s love for Betty is grounded at least in part on the fact that they painted their house together, and Betty* did not share that experience with him. However, in the present discussion it cannot be taken for granted that z lacks some properties that ground x’s love for y, for the question to be addressed is precisely how x’s love for y can be grounded. The alternative option to be considered is that z is a perfect replica of y in the sense that it shares with y all the properties that ground x’s love for y. Betty* might be a perfect replica of Betty in this sense, in spite of lacking some historical property of Betty. Although it is hard to accept that a replica has all the properties of the original, we must leave open the possibility that z has all the properties of y that matter to x’s love for y.

6 . The Intuition Dissected

Now let us focus on (I3), the claim that it is not the case that, if x is justified in loving y, and z is a replica of y, then x is thereby justified in loving z. Once it is granted that ‘replica’ can be read in the two ways just explained, and that justification can be understood either internally or externally, the irreplaceability intuition can be dissected by reasoning on the following four cases.

Case 1. Suppose that z is not a perfect replica of y and that x is internally justified in loving y. Then, (I3) is definitely true. Either x does not know that z is a replica, in which case x is internally justified, but not thereby justified, in loving z, or x is aware that z lacks some relevant properties, in which case x is not internally justified in loving z.

Case 2. Suppose that z is not a perfect replica of y and that x is externally justified in loving y. Then, (I3) is definitely true. In this case x is not thereby externally justified in loving z, given that z is not a perfect replica of y.

Case 3. Suppose that z is a perfect replica of y and that x is internally justified in loving y. Then, (I3) is definitely true. In this case it may happen that x is unaware that z is a replica, as in case 1, and so that x is internally justified, but not thereby justified, in loving z.

Case 4. Suppose that z is a perfect replica of y and that x is externally justified in loving y. In this case we are inclined to say that (I3) is false. It seems that x does have reasons to love z, namely, the same reasons that justify x’s love for y.

In other words, (I3) is true as long as justification is understood internally, no matter whether z is a perfect replica of y, and also if justification is understood externally and z is not a perfect replica of y. But it is not guaranteed to be true when justification is understood externally and z is a perfect replica of y.

Of course, this analysis might be questioned. One might insist that what makes love unique is that it essentially involves irreplaceability, so it is simply wrong to think that y and z can be equally valuable for x. We do not have an argument against such an objection. As observed in section 2, pretheoretical judgments about irreplaceability do not always converge, and we have tried to explain this lack of convergence by laying out different readings of (I3). But if you see irreplaceability as unrenounceable, and you think that reasons for love must be understood accordingly, then you have no need of any such explanation. We simply disagree on the explanandum.

Another possible move is to question the relevance of the notion of external justification. One might contend that, although the distinction between internal and external justification applies to many kinds of actions or emotions, it does not apply to love, because the only sense in which love can be justified is the internal sense. As in the previous case, we do not have an argument against such an objection. So far we have reasoned under the hypothesis that, as far as justification is concerned, love is analogous to many actions or emotions. But if you reject this hypothesis, then you simply deny that cases 3 and 4 make any sense. So, again, we have reached a standoff.

To leave aside such radical objections is to refrain from thinking that the justification of love is utterly sui generis. We are convinced that it is necessary to do so in order to fully understand reasons for love. Our discussion of (I3) rests on this conviction. Assuming that (I3) captures the irreplaceability intuition, there is a clear sense in which the property view does not entail that beloveds are replaceable: internal justification does not necessarily transfer to replicas. As we have seen, however, this does not mean that the irreplaceability intuition rules out replaceability tout court. There is also a sense in which beloveds are replaceable: external justification transfers to replicas, provided that they are perfect.

7 . Properties As Reasons

The line of thought developed in the foregoing sections suggests that the properties of the beloved play an essential role in the justification of love. If Romeo’s love for Juliet is justified, it is justified in virtue of some properties of Juliet. However, we have not offered a positive account of reasons for love. So the question remains of how the properties of a person can provide reasons for loving that person, compatibly with the irreplaceability intuition.

In the last few years, some attempts have been made to address this question. At least three examples deserve attention. Naar has argued that the justification of love depends on the qualities of the beloved as they are manifested in the context of a relationship with the lover. His view aims to explain, among other things, how treating someone as irreplaceable can be fitting.²¹

According to Protasi, we love a person not only in virtue of properties shaped and experienced in a loving relationship, but also in virtue of perspectival properties whose value can be properly assessed outside such a relationship. Beauty is a paradigmatic example of perspectival property: when I say I love a person in virtue of her beauty, I am saying that I love her in virtue of the fact that she looks beautiful to me.²²

According to Jollimore, when one loves a person, one actively attends to the person’s valuable properties in a way that coherently provides one with reasons to treat the person preferentially, so there is a distinctive epistemic attitude that the lover has towards the beloved’s properties. The lover typically is in a position to notice qualities of the beloved that others might not notice.²³

Naar, Protasi, and Jollimore describe in different ways how the interaction between the lover and the beloved enables the lover to gain some sort of privileged access to the qualities of the beloved. When x loves y, y’s properties are given to x in a way in which z’s properties are not given to x, even if z is a perfect replica of y. This leads x to believe that y could not be replaced by z, and in some sense x may be justified in believing so. However, independently of what x believes, x could be in a relationship with z instead of being in a relationship with y. And if x did love z instead of y, then x would experience z’s properties in the same way in which x is now experiencing y’s properties.

In other words, property-based accounts such as these essentially focus on internal justification. It is surely reasonable to say that, if x actively attends to y’s properties in a way that provides x with reasons to love y, and z is a perfect replica of y, then x is internally justified in treating y as irreplaceable. Or at least, x is internally justified in loving y in some way in which x is not internally justified in loving z. However, as far as we can see, this does not prevent x from being externally justified in loving z. So, even if there is nothing incorrect in the accounts discussed, we find them incomplete in some way.

Note that recognizing the role of external reasons is not quite the same thing as acknowledging that the justification of love may not be transparent. The non-transparency assumption stated in section 1 is neutral with respect to the distinction between internal and external justification, because both kinds of justification may fail to be transparent. One may be internally justified in loving a person -because one’s evidence provides the relevant reasons- even though one is not fully aware of one’s internal justification. External reasons are also non-transparent, of course, but they transcend one’s evidence. Unless one wants to maintain that the only sense in which love can be justified is the internal sense, as in the objection discussed above, the appeal to “manifested”, “perspectival”, or “attentively selected” properties does not provide the whole story about the justification of love.

Then what does provide the whole story? Perhaps there is no such thing as the whole story. We have no theory to offer about the properties that constitute reasons for love, and we doubt that such a theory can be found. The fact is that different lovers value different properties, and this variation is largely independent of the standards of beauty or morality that hold in their social environment. Even though there may be actual convergence among lovers on some generally appreciated qualities -as noted in section 1- there is no principled way to characterize a definite kind of property as the kind that matters to the justification to love.

This is not quite the same thing as to deny that there are general truths that ground the justification of love. Presumably there are such truths. But they have a relatively low degree of generality, that is, they are inherently local. To say that Romeo’s love for Juliet is justified is to say that for one like Romeo it is appropriate to love one like Juliet. So there is a regularity that concerns different pairs of persons, including Romeo and Juliet. But this regularity leaves out many pairs of persons, such as Romeo and Juliet’s cousin. That is, it is not appropriate for one like Romeo to love one like Juliet’s cousin, because one like Juliet’s cousin lacks some properties that matter to Romeo (and to those like him). This is why Romeo is not justified in loving Juliet’s cousin.

In more general terms, it may be the case that x loves y, and there is a class of persons such that, for any member z of that class, it is appropriate for x to love z. Since this class is comparatively small, many people around x do not belong to it. So x is not justified in loving any of them. The disanalogy between love and fear observed in section 2 can be explained along these lines. If you are swimming in the ocean and see a shark, there is a very large class of sharks such that, for any member of that class, it is appropriate for you to fear that shark. This is to say that the truth that grounds your justification to fear the shark in front of you has a much higher degree of generality.

The thought that truths about lovers are inherently local accords with our analysis of the irreplaceability intuition. In particular, it explains why replaceability does hold in the external sense for perfect replicas: to say that x is externally justified in loving z is to say that z belongs to the class of persons who are appropriate objects of love for x. A perfect replica of Juliet, unlike Juliet’s cousin or other maidens in Verona, would belong to the class of persons who are appropriate objects of love for Romeo. So, Romeo would be externally justified in loving her.²⁴