Does having deep personal relationships constitute an element of well-being?

Deep personal relationships involve deep mutual understanding and strong mutual affection. This paper focuses on whether having deep personal relationships is one of the elements of well-being. Roger Crisp put forward thought experiments which might be taken to suggest that having deep personal relationships has only instrumental value as a means to other elements of well-being. The different conclusion this paper draws is that having deep personal relationships is an element of well-being if, but only if, the other people involved have qualities that merit affection for these people.

I am using the term 'well-being' as a synonym of 'welfare', 'flourishing', and 'personal good' (for a survey of theories of well-being, see Fletcher 2016). Beyond question, having deep personal relationships can bring people pleasure, emotional support, knowledge, a sense of achievement, and many other benefits. However, such relationships can also be causes of disappointment, frustration, danger, and more generally, personal disaster. This paper is not about the benefits or harms to which deep personal relationships can be instrumental. The paper is about whether deep personal relationships constitute elements of well-being.
By asking whether having deep personal relationships is an element of well-being, I am not asking whether someone can have a life that is positive in terms of well-being without deep personal relationships. I accept that a life that misses out on deep personal relationships might have enough other good elements in it to come out net positive in terms of the well-being of the person who has that life. The question I am addressing is whether having a deep personal relationship is constitutively beneficial to the parties to the relationship, even if having a deep personal relationship is not a necessary condition for having a decently good life.

I
The Elements of Deep Personal Relationships. Some personal relationships are very shallow. These might be brought about by only a few things, such as proximity plus a shared interest in an activity or object. Such a relationship might last only a short time because the proximity or the shared interests are lost. Or the cause of the relationship's brevity might be that one or other participant in the relationship learns something about the other one that kills the relationship. I am not here concerned with shallow personal relationships. Nor am I concerned with personal relationships that are intermediate between shallow and deep. My focus is exclusively on deep personal relationships.
What are the essential features of deep personal relationships? One essential feature of deep personal relationships is a multi-faceted understanding of each other's personality and character. I comment below on the vagueness in that proposition. But let me acknowledge immediately that this multi-faceted understanding need not be complete and need not be articulated.
Two people's multi-faceted understanding of each other's personality and character is very typically gained through a shared history of activities together. At the limit, such shared activities might be merely conversations between the parties. But, of course, a more secure evidential base for beliefs about someone's personality and character would include very many observations of behaviour, since these might bear on the truth of what the person has asserted. Indeed, precisely because people can learn so much about one another from observing behaviour, a shared history of activities that generates a multi-faceted understanding of each other's personality and character might involve vanishingly little conversation, none of which is about especially meaningful topics. People can get to know one another well enough for a deep personal relationship without ever discussing with one another their parents, life goals, or the meaning of life.
Developing a deep and multi-faceted understanding of something as complex as a person's personality and character takes time. Such an understanding will need to range over the person's thinking and behaviour in a variety of different contexts and in the face of different kinds of challenges. Moreover, personality and character change over time, and at least some of a person's personality and character changes need to be understood in order for an understanding of her to be deep enough for a deep personal relationship with her. Only in fantasy can a deep and multi-faceted understanding of something as complex as the personality and character of a person be achieved in a moment. Such points seem to me to be reflected in the stress some writers put on shared history.
But must the multi-faceted understanding of each other's personality and character be the product of a shared history of joint activities? Most definitely that route is the normal one. But, with a little imagination, we can describe a case where two people develop multi-faceted understanding of each other's personality and character without having a shared history of joint activities.
Imagine that Sarah and Anne read very many reports and books about one another. Imagine that their reading up on one another is not a joint activity, even if their studies of reports and books about one another happen to be simultaneous. Suppose that, from the reports and books, they learn a lot of the most important facts about each other's personality and character. Reading all this material, making sense of it, and embedding it in memory takes time. But by the end of this time, I think Sarah and Anne might have a degree of understanding of one another that meets what we might call the knowledge threshold for having for a deep personal relationship. And yet they did not acquire this understanding through a shared history of joint activities.
That example may not seem persuasive. Some people might think that for two people to have a deep personal relationship with one another, they must have a multi-faceted understanding of each other's personality and character that comes from interaction with one another. If interaction is absolutely necessary, then Sarah and Anne's having read lots of books and reports about one another is just not good enough. I do not myself believe that, but I recognize that others might.
Nevertheless, undeniably, in at least a very high percentage of cases in which two people acquire a multi-faceted understanding of each other's personality and character, they in fact acquire it from a shared history of joint activities. The presumption that a multi-faceted understanding of each other's personality and character comes from a shared history of joint activities is so strong that I suspect many people employ shared history as a proxy for a multi-faceted understanding of each other's personality and character.
The second essential feature of deep personal relationships is that the parties involved must have strong affection for one another (Aristotle 1985. Admittedly, there is a kind of deep personal relationship between people who are long-standing enemies. There can be instrumental benefits-as well as, much more obviously, instrumental costs-from having enemies. Whether or not two people's being enemies counts as their having a deep personal relationship, the idea that being enemies with someone constitutes an element of well-being is very implausible. Thus, henceforth in this paper my discussion of deep personal relationships will be restricted to ones involving strong affection. The strong affection I have in mind necessarily involves strong concern for the other's welfare. Strong concern for the other's welfare has attitudinal, motivational and dispositional components. The attitudinal component is such that I don't have strong affection for you unless I am pleased when things go well for you and displeased when things go badly for you (Darwall 2002, ch. 3). The motivational component is such that to have strong affection for you I must be motivated to benefit you and to avoid harming you (Parfit 2011, pp. 146-7). One part of the dispositional component of strong affection is that, as Troy Jollimore (2011, ch. 3) contends, I must be disposed to interpret what you do and say in ways favourable to you. The other part of the dispositional component is that I must be disposed to behave at least sometimes in ways I think beneficial to you (Pettit 2015, ch. 1).
These motivations and behavioural dispositions will sometimes lose out in conflicts with other motivations and dispositions. In order to have strong affection for you, it is not necessary that I always do what I think is beneficial to you. And yet, if I never do what I think is beneficial for you despite my having opportunities, then whether I really do have strong affection for you is thrown into question.
Whether strong affection necessarily involves partiality is controversial. Many writers have claimed that the concept of a deep relationship is such that we don't have a deep relationship with one another unless we give one another priority over other people (for example, Raz 1995, p. 41;Scanlon 1998, p. 218;Scheffler 2001, pp. 100-1, 121-2;Kolodny 2003, pp. 150-1;Jollimore 2011, pp. 29, 112-13, 124, 147, 160, 167, 171). But imagine two people stranded together on a remote island away from the rest of the world.
Suppose they come to know one another extremely well and care intensely about one another. Part of their concern for one another would presumably be instrumental, since both would have needs for companionship and help. But let us imagine that each person's concern for the other is not wholly instrumental. To make this point especially clear, we can imagine that these two grow so fond of one another that each would be willing to sacrifice his own life to save the other's life. With deep knowledge about and intense concern for one another, these two people undeniably have a deep personal relationship.
Now let us add the extra stipulation that, as far as these two people are aware, there are no human beings anywhere else. Hence they cannot be prioritizing one another over other people, since they are not aware that there are any other people. This example thus illustrates the possibility of two people's having strong affection for one another without actually giving priority to one another over other people. Moreover, we can think of examples that do not involve remote two-person islands. Imagine a small tribe living far away from everyone else and unaware of the existence of anyone else. Again, unaware of anyone else outside their group, these people have strong affection for one another without actually prioritizing one another over outsiders.
Admittedly, the idea that strong affection necessarily involves partiality might be formulated in some sort of counterfactual or dispositional form. The idea might be that the two people stranded together who have strong affection for one another and know of no one else in the world could be thought of as disposed to give priority to the objects of their affection over any others, were there any others. Likewise, the mutually affectionate members of the small tribe who are unaware of the existence of anyone outside the tribe could be thought of as disposed to give priority to the objects of their affection over any others, were there any others.
The example of the two people on the remote island is imaginary, though hardly novel. The example of the remote tribe is not imaginary but there are, at most, few tribes left who do not know that the world contains people outside the tribe. Indeed, nearly all of the world's current adult population are aware that the world contains a vast number of people. For people who know that the world contains a vast number of people, there is no realistic prospect that the strong concern had for friends can be mirrored by an equally strong concern for each of a vast number of others. I take it to be an empirical fact about human psychology that human beings are capable of sustaining strong affection for only a limited number of other individuals. If that is an empirical fact, then whatever strong affection people have must be focused on only some others, rather than spread evenly across all of very many others.
People's strong affection for some others will, at least in many contexts, lead them to prioritize the people for whom they have strong affection over anyone else for whom they don't have strong affection. The qualification 'in many contexts' is needed because some contexts forbid prioritizing on the basis of either affection or any other personal connection. In a well-designed state, for example, if you have the job of deciding whom to hire as the hospital's ambulance driver, you are restricted to making the decision solely on the basis of criteria such as driving skills, reliability, composure under pressure, and so on. If you are guided solely by the application of such criteria, you will completely ignore whatever differences in affection you have for the various applicants. However, when you are deciding to whom to allocate your own time, attention, and other resources, you are always allowed and sometimes required to give some degree of priority to those with whom you have deep relationships (Cottingham 1998, p. 11).
In this section, I have contended that one necessary element of deep personal relationships is that the parties to it have a multi-faceted understanding of each other's personality and character. In nearly all cases, such an understanding is the product of a shared history of joint activities. Another necessary element of deep personal relationships is that the parties to it have to have a strong affection for one another. I outlined how this affection has attitudinal, motivational and dispositional components.
I did not add that strong mutual affection must lead to mutual commitment. Mutual commitment is not sufficient for a deep personal relationship. Admittedly, the explicit mutual commitment in a marriage ceremony creates a very important relation between the parties in the eyes of social custom and the law. Furthermore, two people's making uncoerced promises to one another gives each moral rights against the other, and thus the promises create a morally important relation between the parties. But, remembering that our focus here is on the question of whether having deep personal relationships is an element of well-being, we have to admit that a marriage's legal significance and the moral significance of an exchange of promises are insufficient severally or jointly to create the kind of relationship that could plausibly be considered an element of well-being. If the two parties either don't know one another well or don't have strong affection for one another, their marriage does not necessarily create a deep personal relationship between them, though some marriages are lucky enough to lead to mutual understanding and affection.
Not only is explicit mutual commitment not a sufficient condition for having a deep personal relationship; it is not a necessary condition either. Though explicit mutual commitment might be typical in deep personal relationships, I assume it is possible to have a deep personal relationship without explicit commitment. What really matters is whether the parties know one another very well and have strong affection for one another, not whether they have explicitly pledged commitment.
It is possible to have a deep personal relationship with someone without even implicit mutual commitment. Sometimes people have strong affection for someone without being aware of this fact. When people are not even aware of their own strong affection for someone else, they might not make even an implicit commitment to a deep personal relationship with that person. However, if, for a significant duration, two people share a deep understanding of one another and strong affection for one another, they can have a deep personal relationship even if they have not made implicit or explicit commitments to one another.
I have already acknowledged that developing a deep and multifaceted understanding of something as complex as a person's personality and character takes time. Strong mutual affection, however, is sometimes ignited quickly. While developing a deep understanding of one another takes significant time and igniting affection does not, both the understanding and the affection have to persist for a significant duration in order for there to be a deep personal relationship. A personal relationship that lasted for only a brief episode was not a deep personal relationship. about well-being hold about the status of deep personal relationships. I then consider a familiar line of argument for concluding that a so-called objective list theory of well-being should hold that having deep personal relationships is an element of well-being.
Hedonists think that how much a person's life benefits her or him is solely a matter of how much net pleasure she or he gets over the course of the life. Hedonists thus hold that having a deep personal relationship is not an element of well-being, but at best a means to obtaining net pleasure. If instead we believe that having a deep personal relationship is an element of well-being, we apparently are not hedonists.
In reply, hedonists might contend that the pleasures of friendship are distinct in kind, higher in quality, and irreplaceable by any other pleasure. Even if that is correct, the hedonist still attributes non-instrumental value to the pleasures of friendship, rather than the friendship itself. And we are now very familiar with thought experiments in which someone is given the illusion of having friends, and thus is given what he firmly believes are the experiences of having friends, without really having friends (Nozick 1974, pp. 42-5;Pettit 2015, pp. 117-19). If hedonism concentrates on the properties of a subject's experiences as perceived from the subject's perspective at the time of the experiences, then the experiences produced by the illusion of a deep relationship could provide more pleasure than the experiences of actually having a deep relationship. Thus an implication of any form of hedonism that concentrates on the properties of a subject's experiences as perceived from the subject's perspective at the time of the experiences is that having a deep personal relationship is not an element of well-being.
Desire-fulfilment theories of well-being maintain that what constitutes our well-being is the fulfilment of our desires. According to desire-fulfilment theories, our having deep personal relationships benefits us if and only if our desires are fulfilled to a greater extent than they would be if our lives lacked such relationships, whether or not our net pleasure increases. Note the contrast with hedonism, which holds that our having deep personal relationships benefits us if and only if we get more net pleasure as a result, whether or not we desire to have such relationships.
A different theory of well-being holds that how much a person's life benefits her or him depends on the extent to which that life contains multiple elements. Since this theory of well-being holds that there are multiple elements, the theory is sometimes called pluralism about well-being. Since the theory holds that what the elements of well-being are is not determined by subjective desire, the theory is often called objective list theory (Parfit 1984, pp. 499-502).
Different varieties of objective list theory pick out different elements. Nearly all varieties of the objective list theory admit that how much a person's life benefits her or him is partly a matter of how much net pleasure she or he gets. The main debate is not about whether net pleasure is on the objective list, but rather about what else is. Perhaps the leading variety of the objective list theory is one that follows Derek Parfit (1984) and James Griffin (1986Griffin ( , 1996 in taking the elements of well-being to be pleasure (or enjoyment), important knowledge, significant achievement, autonomy, and deep personal relationships.
How can we ascertain whether something is on the list of elements that together determine how much a person's life benefits her or him? We can run the following kind of thought experiment (Hooker 2015). We imagine any two possible lives an agent might have, on the assumption that these possible lives are as much alike as possible except that one of these lives contains the item in question and the other does not. We then think about whether the life containing that item would be more beneficial to the agent than the other life. If the correct answer is that the life containing the item would not be more beneficial to the agent than the other life, then that item is not on the list of elements of well-being.
For illustration, imagine an agent whose life could have either of two possible futures. These alternative futures are as much alike as possible in terms of amounts of net pleasure, important knowledge, significant achievement, and autonomy. However, in one of these possible futures the agent's life contains deep personal relationships and in the other the agent's life doesn't contain deep personal relationships. If two possible futures contain the same amounts of net pleasure, important knowledge, significant achievement, and autonomy, then the life with deep personal relationships seems more beneficial to the agent.
If the life with deep personal relationships seems more beneficial to the agent, what is the best explanation of this? Well, an explanation that suggests itself is that what seems to be so is so, that is, the life with deep personal relationships is more beneficial to the agent. And the natural explanation of why the life with deep personal relationships is more beneficial to the person living that life is that having deep personal relationships is on the list of elements that constitute additions to well-being.
However, a rival possible explanation is as follows. By hypothesis, the two possible futures being compared are as much alike as possible with the proviso that one includes deep personal relationships and the other doesn't. Nevertheless, the fact that one of these possible futures contains deep personal relationships cannot help but generate differences in the levels of other goods, and these other differences are what account for the superiority of one possible future to the other. Our thought experiment was supposed to isolate one variable. However, the rival possible explanation claims that we have not managed to isolate one variable, because other variables remain in play and are making the pivotal difference. The contention might be that we cannot help assuming that the possible future containing deep personal relationships would contain more pleasure, significant achievement or important knowledge than the possible future without deep personal relationships.
There are two points to be set against that contention. First, the possible future containing deep personal relationships might well contain some associated emotional pain that the possible future without deep personal relationships does not contain. Having a deep personal relationship makes one vulnerable to sympathetic distress when the other person in the relationship is harmed. There is also vulnerability to the emotional pain that would come from the deterioration or ending of the deep relationship. Hence the extra pleasure from having deep personal relationships might be counterbalanced by extra emotional pain.
Second, to the extent that someone is devoting her time and attention to developing personal relationships, she is passing up opportunities for kinds of significant achievements and important knowledge and pleasure that do not come from deep personal relationships. The life without deep relationships might get enough pleasure, knowledge and achievement from other sources to counterbalance exactly the pleasure, knowledge and achievement that the other life gets from the deep relationships it has.
Our thought experiment needs to compare two possible lives that really are equal in terms of quantities of pleasure, knowledge, achievement and autonomy. Because of the two points made in the previous two paragraphs, there is the possibility that one of two possible lives-which really are equal in terms of quantities of pleasure, knowledge, achievement and autonomy-contains deep personal relationships while the other of these two lives does not. In this comparison of possible lives, we have isolated the one variable of whether a life contains deep personal relationships. If, in this comparison, we think that the life with deep personal relationships contains more well-being than the life without deep personal relationships, the explanation of that thought must be that we think that deep personal relationships do constitute an element of well-being.

III
Crisp's Counter-argument. Recently, a line of attack on that conclusion has been suggested by a thought experiment posed by Roger Crisp: Titania wakes. The first person she sees is Bottom. The second person she sees is Top. Because of the magic juice Oberon has sprinkled in Titania's eyes, she has developed a deep affection for Bottom. . . . Titania's affection for Bottom is a highly contingent psychological fact about her. Had she seen Top first, she would have felt the same affection for him instead. (Crisp 2018, p. 8) 1 Crisp poses this thought experiment in a discussion of whether there are moral reasons that come from partiality. My focus is on whether having deep personal relationships is an element of well-being. Although the focus of Crisp's discussion is different from mine, Crisp's arguments against the idea that there are moral reasons that come from partiality can be recast as an argument against the idea that having deep personal relationships is an element of well-being. This recast argument is that, although affection is very often instrumentally valuable, the Titania case dramatizes that affection is too contingent and arbitrary to be valuable in itself. One person's affection for another is not enough to constitute a deep personal relationship. However, Crisp (2018, p. 9) adds, 'Nor does it matter if the affection in question is reciprocated (imagine that Oberon had caused Bottom to feel for Titania what she feels for him).' In other 1 Many of the examples and arguments in Crisp (2018) are repeated in Crisp (2021). I focus on the 2018 paper because it is in print as I type these words and the 2021 paper is not quite. words, even if the affection is mutual, this mutual affection might be too contingent and arbitrary to be valuable in itself.
Very many people hold that affection is utterly subjective in the sense that affection always depends completely upon likes and impulses for which there are causal explanations but not normative reasons. However, what if someone told us that he has strong affection for a saucer of mud (cf. Anscombe 1957, pp. 70-1)? Suppose that when we ask about the aesthetic properties of the saucer of mud, he denies there is anything beautiful about the colour, texture or shape of the mud. We then ask how this saucer of mud is connected to other things, in expectation that, for him, the saucer of mud has symbolic value (Kagan 1998;Keller 2013, pp. 66, 68). Suppose he asserts that this saucer of mud has no special connections to other things and thus no symbolic value. If he admits that the saucer of mud has no aesthetic or symbolic value, his statement that he has strong affection for this saucer of mud is bewildering. Affection is not a fitting attitude to have towards a saucer of mud that has no aesthetic or symbolic value.
'In general, an attachment must have a worthy object to be valuable', according to Raz (2001, p. 19). This view seems to me correct. Obviously, affection is the kind of attachment relevant here. Affection is a fitting attitude to have towards a person only if either that person has qualities that warrant the affection (Hooker 1999;Keller 2000Keller , 2013; Abramson and Leite 2011; Jollimore 2011; Howard 2019) or one has a duty of affection towards that person. 2 The qualities that warrant affection for a person are undefeated reasons for having affection for the person. Many people have qualities sufficiently endearing to merit affection though these people also have bad qualities. I won't try to say anything informative here about how much good is enough to defeat a particular quantity of bad.
Limited beings that we are, we are incapable of knowing about, much less having affection for, every single person with qualities warranting affection. The principle that 'ought' implies 'can' seems to apply here. If this principle does apply, then we should reject the idea that we ought to have affection for every single person who has qualities warranting affection (cf. Jollimore 2011, p. 16). Reasons for having affection for a person that come from the entity's good qualities must therefore be reasons that make affection sensible and permit it but do not require it.
I do not suppose that someone who has affection for someone must be able to articulate the qualities of the person that warrant the affection (Keller 2000, pp. 164-5). Why a particular person warrants affection might be a long story, involving lots of fine distinctions and weighing up of good qualities versus bad. And why a person warrants affection might involve facts about her of which other people can easily be unaware. The person herself might be unaware of many of them. Thus people might have strong affection for someone that is warranted by qualities of her that these people cannot specify and that she herself doesn't know she has. Now, one of the most common causes of affection is beauty. However, does beauty make affection fitting? There are at least four worries about the idea that beauty makes affection fitting.
One is that beauty is a surface phenomenon. Shouldn't appropriate affection be based on something deeper than what is only skin deep? Consider as an example someone who on the one hand is captivatingly beautiful but on the other hand is completely selfish, dominating, sadistic and dishonest. Admiring the beauty of this person would be one thing; having affection for such a person would be another thing entirely.
We often hear people say that affection for someone is justified because, whatever her surface features, she is 'beautiful inside'. However, the judgement that someone is 'beautiful inside' is not an aesthetic judgement about the innards of the person. Rather, the judgement is normally that she has a strikingly good character. Skindeep beauty seems not to merit deep affection; being 'beautiful inside' does merit deep affection, but the endearing feature here is the strikingly good character, not really beauty.
The second worry about the idea that beauty makes affection fitting is that beauty is something for which people are not responsible and so cannot deserve credit. Crisp (2018, p. 9) intimates that people should be given credit for their good qualities only if these qualities are 'the result of free and voluntary actions they have taken'. Well, how people look is obviously not entirely a product of their choices, though the looks of people who are beyond a certain age partly depend upon the choices they have made.
I think we need not engage here in a discussion about whether people can be given at least a little credit for how they look. For we can justifiably like people for qualities they have even if they cannot take any credit for those qualities. Imagine someone born with a friendly disposition and then trained from a young age to be virtuous and skilled, especially in activities that we find particularly interesting. These qualities of the other person make our affection for her fitting even if she cannot take credit for these qualities. Equally, we can justifiably dislike people for bad qualities that were implanted in them (Raz 2001, p. 24 n. 9;Sher 2007Sher , 2009.
The third worry about taking beauty to be a good justification for affection is that the beauty of living things is normally transitory. If affection for someone is based on that person's beauty, will affection fade as the beauty fades? Obviously, the worry about fading also applies to qualities other than beauty, for example, generosity of spirit, memory, analytical power, creativity, strength, and so on. Should affection based on these qualities fade as the qualities fade?
The fourth worry is that very many people are very beautiful. Likewise, very many people are extremely knowledgeable, very many tremendously kind, very many wonderfully witty, and so on. If we have affection for some people simply because of such good qualities, shouldn't we have affection for everyone else who possesses the same good qualities to the same degree? Crisp (2018, p. 9) writes, 'I love my wife because she is beautiful, kind, and witty. But in that case why should I be partial to my wife, and not to all those who are beautiful, kind, and witty?' IV Shared History. Crisp himself points to an answer to the problem about fading qualities and to the problem that there are too many objects with good qualities for us to be able to love them all. In §i, we saw that deep personal relationships have as essential components not only mutual affection but also a multi-faceted understanding of each other's personality and character. Although gaining a multi-faceted understanding of each other's personality and character by means other than a shared history of joint activities is possible, the nearly universal means of gaining a multi-faceted understanding of each other's personality and character is through a shared history of joint activities (Keller 2000, p. 171;Raz 2001, p. 30).
Shared history expands with time, and time is the very thing that often corrodes good qualities such as beauty, energy, memory, and quickness of mind. As a shared history builds up between people, they might find that their understanding of and affection for one another grows. Furthermore, each might find endearing the other person's past abilities, achievements and generosity, even if the other person's abilities have faded and this person now has little left to give (cf. Keller 2000, p. 165). As Jollimore (2011, p. 140) observes, the lover 'will be especially inclined to see the image of the past in the face of the present'. In many ways, therefore, the accumulation of shared history might well correlate with a deepening of a relationship (Hurka 2006;2011, ch. 7;2017;Kolodny 2010, esp. pp. 42-54).
Clearly, shared history does not, on its own, create a valuable relationship. The shared history might not have shown the two participants anything endearing about the other (Jollimore 2011, p. 185 n. 18). Even more, the shared history can be one of making each other miserable. Indeed, extended contact with someone might reveal his terrible character. In that case, others would be right to back (or run) away from him.
Turn now to cases where a shared history leaves people opaque to one another. A shared history of mutual incomprehension does not lead to a deep personal relationship. Indeed, the role of a shared history in sustaining deep personal relationships is based largely on the presumption that this shared history will have given each party important knowledge about the other, including knowledge of endearing qualities. Thus, where the shared history does not yield knowledge of one another's endearing qualities, a deep relationship cannot be grounded in the shared history. Now consider cases where a shared history leads two people to have various false beliefs about one another's endearing qualities. Suppose Ivor's shared history with Natasha convinces Ivor that she consistently and strongly wants his life to go well for his sake, as opposed to just when and to the extent that her welfare somehow depends on what happens to him. Suppose the belief Ivor has about Natasha is false: the truth is that she cares about him only in so far as his welfare is instrumental to other concerns she has. So, given the important false belief that Ivor has about Natasha, do they really have a deep relationship at all? The answer is no. Although she appeared to be his friend and he thought she was, in reality she was not.
I am not suggesting that false beliefs about someone, even false beliefs about someone that arise from a shared history with that person, always preclude having a deep personal relationship with that person. Suppose that a shared history between you and me has led you to think I am both self-aware and courageous; however, actually I am self-deluded and cowardly. If there is considerable shared history between us on the basis of which we have many important true beliefs about one another and deep affection, we might have a deep relationship even if one of us has important flaws of which the other is unaware (thank goodness). While in order to have a deep relationship with each other we have to know a considerable number of important truths about one another, we don't have to know all important truths about one another.
Return now to the example in which Ivor's and Natasha's shared history leads Ivor to believe that Natasha stably and strongly wants his life to go well for his sake, but in fact Ivor's belief about Natasha is false, since really her concern for him is only instrumental. In this example, Ivor and Natasha do not have a deep relationship. But what prevents them from having a deep relationship is not that he has false beliefs about her. It is rather that her concern for him is only instrumental. She doesn't have the kind of affection for him that is necessary for a deep relationship.
As remarked earlier, the concept of a deep personal relationship has grey boundaries. There is a scale running from 'no personal relationship between this person and that person' to 'very shallow personal relationship between this person and that' to 'moderately substantial personal relationship between this person and that' to 'very deep personal relationship between this person and that'. Of course, there are lots of points along the scale between the ones I have mentioned. And there are no sharp lines between the categories on this scale. We should therefore acknowledge the imprecision in the condition that a considerable degree of mutual affection and mutual knowledge is necessary for a deep personal relationship.
However, even this imprecise condition provides us with an escape from the point that there are an overwhelmingly large number of people who have good qualities and thus for whom affection would be fitting. Earlier, I contended that we are psychologically unable to have strong affection for every one of a vast number of individuals. Even if that were not true, we should consider the implications of the conjunction of the following two facts. First, the main route to a deep and multi-faceted understanding of someone's personality and character is via a history of shared activities. Second, mortal beings that we are, we have a limited amount of time to get to know others deeply. Therefore, there are limits to the number of people we can get to know well enough in order for us to have deep personal relationships with them. Immortal beings might have enough time to form innumerable deep personal relationships. We mortal beings don't.

V
Crisp's Objection to Taking Shared History Plus Affection to Be Enough. Crisp poses the following thought experiment (2018, pp. 10-11). Suppose Titania, Top and Bottom do have a shared history of joint activity. And suppose this shared history of joint activity has not caused them to have special affection for one another. However, then Oberon sprinkles magic juice, and this juice induces intense affection between Titania and Bottom. Given the shared history between Titania and Bottom, is the injection of intense affection in the imagined scenario enough to give them a deep personal relationship? I suggested earlier that referring to shared history is best interpreted as a proxy for reference to a deep and multi-faceted understanding of each other's personality and character. Let us interpret Crisp's example as stipulating that the shared histories between Titania and Bottom and Titania and Top have given each a deep and multi-faceted understanding of the other. But these deep and multifaceted understandings have not sparked deep affection. Then Oberon administers his magic juice, which causes strong affection between Titania and Bottom (or Top).
If the magic juice wears off too quickly and the affection fades with it, then no deep personal relationship between the two is created. For, as I wrote earlier, a significant period of mutual understanding and affection is necessary for there to be a deep personal relationship. If there was a significant period before the magic juice wears off and the affection fades, then there was a deep personal relationship, but the deep personal relationship ended when the affection ended.
Crisp points to the massive contingency in who has affection for whom. Had Oberon sprinkled the magic juice differently, the affection would have been between Titania and Top. Again, if this affection lasted long enough, the deep personal relationship would have been between Titania and Top instead of between Titania and Bottom. Such contingency is unsettling, and yet we cannot plausibly deny that whom any one person ends up forming a relationship with is highly contingent.
Suppose that administering magic juice clarified Titania's and Bottom's vision so that now each sees clearly some of the other's qualities for the first time. (A very similar thing can happen without magic: one person can know another well but fail to perceive or appreciate some of his qualities until a chance remark or even a nonverbal revelation exposes these qualities so openly that strong affection for him ignites immediately.) If the magic juice works by causing an epistemic improvement that ignites affection between Titania and Bottom, then, once the strong affection between them has lasted for a considerable period, a deep personal relationship between Titania and Bottom exits. Now suppose that the magic juice somehow works without revealing to Titania or Bottom previously unknown qualities of the other. I think that, even in this case, once the strong affection between them has lasted for a considerable period, Titania and Bottom have a deep personal relationship. (Whether this deep personal relationship is valuable is a matter I return to below.) However, some people have different reactions to (a) the case where Titania and Bottom have a multi-faceted understanding of one another's personality and then magic juice is needed to help them see the other's good qualities, and (b) the case where the magic juice works without revealing to Titania or Bottom previously unknown qualities of the other. Some people might say that while in (a) a deep personal relationship does result, in (b) what results is not a deep personal relationship. The idea here might be that the affection needed for a deep personal relationship cannot result from something as indiscriminate as the application of magic juice.
Beyond question, if Titania and Bottom fall in love with one another because of magic juice, the cause of their love would not provide good reason for the love. The same would be true if they fell in love with one another because of each other's beauty, ancestry, money or popularity. Think how often what initially attracts people to one another is something that does not actually warrant the affection they have for one another. (For example, two people might have been attracted to one another by the fact that they shared an enemy. Or Jack might have fallen for a Jill because she looked like his favourite movie character.) Very often, although what initially attracts people to one another are qualities or connections that do not justify strong affection, the relationships nevertheless last. And some of these relationships achieve such an impressive degree of depth that we would have to classify them as deep relationships.
Thus there are cases in which people were initialled attracted by qualities of each other that do not actually merit love, and yet, after a significant duration of multi-faceted knowledge about one another and strong mutual affection, these people do share a deep personal relationship. Once we remember that the affection in what ends up being deep personal relationships is in many actual cases sparked by qualities that do not warrant the affection, I cannot see why we would deny the conceptual possibility that, in the fanciful case in which the affection was caused by magic juice, the relationship could end up being a deep one.
Whether the relationship two people have with one another is a deep one is one thing; whether having a deep personal relationship is valuable enough to constitute an element of someone's well-being-I am now persuaded by Crisp's examples-is another. His examples show, I think, we need to distinguish between deep personal relationships that are valuable enough to constitute an element of wellbeing and deep personal relationships that are not valuable enough to constitute an element of well-being. As I will argue, deep personal relationships that are valuable enough to constitute an element of well-being are ones between people who do have qualities that merit the other's affection. And deep personal relationships that are not valuable enough to constitute an element of well-being are ones between people who do not have such good qualities.
I posed the thought experiment in which we compare two possible lives that really are equal in terms of quantities of pleasure, knowledge, achievement and autonomy, but where one of these lives contains deep personal relationships and the other does not. The conclusion from that thought experiment I hoped at least most people would draw was that any possible life with deep personal relationships would have a higher level of well-being than any other life that is equal in terms of quantities of pleasure, knowledge, achievement and autonomy but devoid of deep personal relationships. And I hoped people would infer from that conclusion that having deep personal relationships constitutes an element of well-being. But Crisp's example of Titania and Bottom mandates a qualification to that conclusion: deep personal relationships constitute an element of well-being only if the relationships are between people who do have qualities that merit the other's affection.
To consider a case in which the people do not have the qualities that merit the other's affection, let us imagine that Titania finds herself with strong affection for Bottom, and that nothing about Bottom warrants her strong affection for him. We might try imagining that the magic juice sparks her affection for Bottom without at all affecting her negative judgements about him. Suppose, for instance, that both before and after the magic juice is administered, she judges that he is a selfish and cruel hypocrite, with no admirable qualities. Nevertheless, she wakes up one morning to find herself with strong affection for him. If Titania feels strong affection for Bottom but sincerely judges him to be completely awful, her attitudes towards him are incoherent and should strike her as so.
In some cases, there are people for whom one is morally required by one's relation to that person to have affection for that person, if possible. I will come back to such cases. But let us suppose the case we are considering now is not such a case: Titania stands in no relation to Bottom that imposes on her a requirement to feel affection for him. Where one's affection for another is not required by a special relation to that person, one's affection for that person is under pressure not to exceed one's positive evaluations of that person. Now let us imagine instead that Titania makes no evaluative judgements about Bottom both before the magic juice was administered and after. In this case, after the magic juice has been administered, she finds herself with strong affection for Bottom but without positive evaluative judgements about him. Again, wouldn't Titania think her affection for Bottom mystifying? How could she make sense of her intense affection for him unless she thought he has at least some good qualities? I am not proposing that Titania must be able to articulate what qualities of Bottom provide undefeated reasons for her strong affection for him. I am proposing that having long-lasting affection for someone does not cohere with being unable to make at least some positive evaluative judgements about him.
Another possibility is that the magic juice causes Titania not only to have strong affection for Bottom but also to judge-incorrectlythat he has wonderful qualities. In this scenario, Titania is not guilty of incoherence between her attitudes toward Bottom and her judgements about him. But she either has an empirically mistaken view of his character and personality or is evaluating his character and personality in terms of the wrong criteria. 3 Either way, she is badly mistaken to judge him favourably.
Let us turn to a point that holds no matter whether Titania makes negative judgements about Bottom's bad qualities, no judgements about his bad qualities, or mistaken positive judgements about his bad qualities. If Titania has strong affection for Bottom over a sustained period of time although he does not actually have qualities that merit the affection, then the scenario is terrible for Titania, whether or not she ever realizes it. To the extent that she devotes her affection, attention and time to him, she wastes limited resources on an undeserving person. Admittedly, there might be some instrumental benefits of Titania's having a deep personal relationship with Bottom. However, we are interested here not in instrumental benefits but in constitutive benefits. If she has a deep personal relationship with someone lacking qualities that merit her affection, this relationship does not constitute a benefit to Titania. On the contrary, this scenario must be a nightmare for anyone with Titania's well-being at heart. As Raz (2001, pp. 39-40) wrote, 'love demeans the lover if bestowed on an unworthy object, admiration is ridiculous if directed to the commonplace, respect is perverted when those respected do not merit it'.
Let us now turn from voluntary relationships to involuntary ones. There seems to be something morally wrong with people who do not have affection for their children, siblings and parents-unless these children, siblings or parents have behaved in such a hostile way that claims to affection are forfeited. Affection for someone can not only be justified by the other person's good qualities but also be morally required because of a special connection between these two people. For illustration, suppose two siblings have a deep personal relationship even though one of them does not have good qualities and both of them know this. A deep personal relationship can obtain in such a case without there being either incoherence or delusion about each other's qualities. Imagine that the bad sibling's affection for the good sibling is caused by recognition of the good sibling's good qualities, and the good sibling's affection for the bad sibling is caused by recognition of duty.
In such a case, the deep personal relationship can be an element of the bad sibling's well-being, since this sibling's affection for the other does reflect the other's goodness. But the deep personal relationship would not be an element of the good sibling's well-being (though, other things being equal, the good sibling's having strong affection for the bad sibling makes the good sibling's life morally better than it would otherwise be).

VI
Conclusion. This paper has argued that having deep personal relationships is an element of your well-being if but only if the other people in these relationships have qualities that merit your affection. Having deep personal relationships with people who don't merit affection may bring you some instrumental benefits or may be morally admirable, but it does not constitute a benefit to you. I will close by making an observation rather than offering an argument. Objective list theorists, who take well-being to be constituted by various elements, might start off saying the list contains pleasure, knowledge, achievement, deep personal relationships, and autonomy. In the face of counterexamples, these theorists often narrow the specification of these elements. Pleasure is narrowed to innocent pleasure. Knowledge is narrowed to important knowledge. Achievement is narrowed to achievement of worthwhile goals (Griffin 1986, pp. 64-6;1996, pp. 19-20;cf. Bradford 2015). Similarly, the lesson to be drawn from Crisp's Titania examples is that having deep personal relationships is an element of well-being