1 The obvious exceptions are spousal relationships and parental relationships to children (but not vice versa). The former are easily accounted for by reference to the marriage contract or some similar explicit exchange of vows. The latter cannot always be seen as clear examples of the voluntary entering of a relationship, especially when abortion and/or contraceptive devices are not readily available. My primary concern is with those familial relationships that do not appear, in any cases, to be voluntarily entered.

2 I will briefly consider this option in section II below.

3 Simmons, A. John Moral Principles and Political Obligations (Princeton, NJ: Princeton University Press 1979), 13Google Scholar

4 Also, I am not using the term ‘obligation’ to refer to a moral requirement with greater weight than a duty. In fact, some obligations, such as certain obligations of gratitude (if we have such obligations), may be fairly weak, while some duties, such as certain duties of mutual aid, are extremely stringent.

5 Simmons, in Moral Principles and Political obligation, takes an obligation to be ‘a moral requirement generated by the performance of some voluntary act (or omission)’ (14). Thus my terminology is not in any sense standard, but, I suggest, it is useful for the discussion of obligations to family members. However, in my discussion of Belliotti in section IV below, I will deviate from my own terminology, which is specifically rejected by Belliotti (see note 33 below).

6 See also Simmons, Moral Principles and Political obligation, 15.

7 Of course it might be said that your nature and the nature of the act of throwing a life preserver to you are not sufficient to explain why I have the duty to throw that life preserver to you: we also need to state that I happened to be standing on the bank of the river. But notice that once I am in a position to throw that life preserver to you, your nature and the nature of the action provide sufficient explanation of my duty. On the other hand, even when I am in a position to give you ten dollars (I have the money and you are standing in front of me with your hand out), appeal to your nature and the nature of the action are not sufficient to explain or ground an obligation, unless we can appeal to a promise that I made to you. It might be responded that if we redescribe the action as an act of promise-keeping, then appeal to the nature of the act is sufficient to ground an obligation. However, this is not a normatively neutral redescription of the act: to say that it is an act of promise-keeping is to say that it is obligatory in so far as genuine promises generate obligations. However, there are difficulties here that I cannot address, and that need not be resolved for my purposes.

8 I have characterized natural duties as grounded by the nature of the persons to whom the duties are owed, while special obligations are not so grounded. Promise keeping and, as I will argue, obligations to intimates are grounded by the character of the relationship between obligor and obligee. I have left it open that other types of special obligations may be grounded on some fact other than the relationship between obligor and obligee, although all special obligations will be grounded by something other than or more than the intrinsic nature of the obligee and the intrinsic character of the required action.

9 This fact about special obligations to intimates raises difficulties for my account that I will discuss in section III, 2 below.

10 For a discussion of the voluntarist objection to associative or special obligations more generally, see Samuel Scheffler, ‘Families, Nations, and Strangers,’ The Lindley Lecture at the University oi Kansas (October 17, 1994). For my response to Scheffler, see my ‘Associative Obligations, Voluntarism, and Equality,’ Pacific Philosophical Quarterly 77 (1996) 289–309Google Scholar.

11 Of course, children may be required to care for older parents for any number of reasons, including utilitarian reasons. But such utilitarian reasons do not constitute special obligations: we have a general duty to promote or maximize well-being, and caring for parents may be a good or even the best way to fulfill such a duty. (If one is willing to view utilitarian duties as owed to other persons, then utilitarian duties are natural duties.) As I indicate in the next section, however, I am concerned to find the source of genuinely special obligations between family members.

12 See, for example, the discussion of obligations of fair play in Nozick, Robert Anarchy, State, and Utopia (New York: Basic Books 1974), 90ffGoogle Scholar. See also Simmons, Moral Principles and Political Obligations, chap. 5. Obligations of gratitude are often appealed to in order to explain what adult children owe to their parents. See, for example, Jecker, Nancy S. ‘Are Filial Duties Unfounded?’ American Philosophical Quarterly 26 (1989) 73–80Google Scholar; and McConnell, Terrance Gratitude (Philadelphia: Temple University Press 1993), chap. 7Google Scholar.

13 Simmons, A. John ‘External Justifications and Institutional Roles,’ The Journal of Philosophy 93 (1996), 30CrossRefGoogle Scholar

14 If one is unhappy with my earlier attempt to distinguish special obligations from natural duties (where obligations of promise-keeping are included under the former heading), one can, for the purposes of my discussion, simply take as a starting point the claim that we have certain duties and obligations, including duties of mutual aid and obligations to keep the promises that we make. We can then see the deflationary account as holding that any obligations that we have to family members must be reducible (in a straightforward way) to some type or types of obligation or duty that were included in our initial set. If we understand the deflationary view in this way, then my view presented in section III will still not be deflationary, because I assimilate familial obligations to obligations of promise-keeping but I do not reduce the former to the latter.

15 For a more complete discussion of special obligations and utilitarianism, see my ‘Relatives and Relativism,’ with Fumerton, Richard Philosophical Studies 87 (1997) 143–57Google Scholar.

16 ‘Filial Morality,’ The Journal of Philosophy 83 (1986), 446

17 In her well-known paper ‘What Do Grown Children Owe Their Parents?’ (in Having Children: Philosophical and Legal Reflections on Parenthood, O'Neill, Onora and Ruddick, William eds. [New York: Oxford University Press 1979] 351–6)Google Scholar, Jane English argues that ‘the duties of grown children are those of friends and result from love between them and their parents’ (351). My account of familial obligations accepts, in some form, that claim of English's; however, my account is an attempt to explain why the features shared by friendships and certain familial relationships do in fact generate obligations, a question upon which English's paper is silent. By offering such an account, we will be in a far better position to respond to the views that I discuss in the next section.

18 I am grateful to Thomas Hurka for pointing out to me the need to differentiate between forced intimacy and genuine intimacy.

19 I am not, of course, talking about a project of having friends, or of making friends. I am speaking of a friendship with a certain person as a project. The latter but not the former essentially involves a specific other.

20 I do not mean my language of mutual projects and our projects to be taken as a rejection of individualism or of what has come to be known as ‘atomism’: I am not postulating some metaphysical ‘us’ distinct from the individual parties to the friendship. Rather, I am pointing out a fact about the character of the project of friendship and of the sorts of attitudes of the parties to a friendship. For a discussion of such projects that does reject individualism, see Taylor, Charles ‘Cross-Purposes: The Liberal Communitarian Debate,’ Liberalism and the Moral Life, Rosenblum, N. ed. (Cambridge, MA: Harvard University Press 1989) 159–82Google Scholar.

21 The having of such a project can be more or less explicitly articulated in the minds of the parties to the friendship; often, such articulation only occurs in times of crisis.

22 So, unlike with respect to most promises, there is an open-endedness about the content of duties to friends: they are required to care about one another and to maintain the friendship (of course, these are prima facie requirements that can be overridden). If Tracy is my friend, then, there are no specific acts, perhaps, that I am required to perform, unless certain specific acts are necessary conditions of my caring for Tracy and for sustaining our friendship. But, as with promising, what I owe is correlated with what I have chosen: in one case, I have chosen the content of my promise, in the other I have chosen the character of the shared project. I will return to this point below.

23 With respect to promising, it is the voluntary character of a promise that renders the resultant obligations in conformity with the voluntarist requirement. An account of promising would need to answer the second question: why do promises create such demands?

24 An analogous confusion sometimes arises in discussions of emotions. It is often said that we cannot choose to feel a particular emotion at a particular time. That is true, if we mean that we cannot choose to be, for example, angry, in the way that we can choose to say ‘I promise.’ But, over the course of time, we can make choices and perform actions that will influence the types of persons we become, such as whether we are people who get angry in certain types of circumstances. So the feeling of a particular emotion may not be voluntary in any straightforward way, but that does not mean that emotions are beyond our control.

25 See also Kupfer, Joseph ‘Can Parents and Children Be Friends?’ American Philosophical Quarterly 27 (1990) 15–26Google Scholar, for an interesting discussion of the relationships that can be realized between parents and children, although I disagree with his claim that parents and adult children cannot be friends.

26 When I use the term ‘adult,’ I do not have any specific age in mind. So I do not mean to exclude teenagers. The ages at which persons can begin to develop friendships with parents can vary a great deal.

27 See note 12 above for references to works that develop accounts based on gratitude. One way to develop a more general account based on the realization of goods is to take an Aristotelian approach to special obligations; for example, see Hurka, Thomas Perfectionism (New York: Oxford University Press 1993), 134–6Google Scholar.

28 For an interesting critique of the various sociobiological enterprises in the context of E.O. Wilson's work, see Kitcher, Philip ‘The Hypothalamic Imperative,’ from Vaulting Ambition (Cambridge, MA: The MIT Press 1985) 417–34Google Scholar; reprinted in Issues in Evolutionary Ethics, Thompson, Paul ed. (Albany: State University of New York Press 1995) 203–23Google Scholar.

29 ‘Evolutionary Ethics: A Phoenix Arisen,’ in Issues in Evolutionary Ethics 225-47.

30 Liberalism and the Limits of justice (Cambridge, UK: Cambridge University Press 1982), 179

31 ‘Honor Thy Father and Thy Mother and to Thine Own Self Be True,’ The Southern Journal of Philosophy 24 (1986), 152

32 Actually, Belliotti's view on this point is unclear. He seems to think that it does not matter whether we hold that children owe abusive parents prima facie obligations or we hold that such abusive parents ‘fall outside the realm of consideration’ (157). But, in order to be consistent, he (or anyone who thinks that biology generates moral requirements) has to hold the former: if genetic contribution generates moral requirements, then, even if such requirements can be overridden by other moral considerations, they still exist in abusive contexts. At least, such is the case, if one accepts what has come to be known as generalism, as opposed to particularism. For a defense of particularism, see Dancy, Jonathan Moral Reasons (Oxford: Blackwell 1993), chaps. 4–7Google Scholar.

33 Belliotti uses the term ‘moral requirement’ in order to avoid the debate about the distinctions between obligations, duties, and ‘ought’ judgments, so, in my discussion of his view, I will adhere to his terminology. See Belliotti, n. 10 (161).

34 Belliotti actually suggests another argument: just as we are permitted to pursue our own interests to a greater extent than we pursue those of persons generally, so we are required to pursue those of persons who are ‘metaphysically closer’ to us to a greater extent (154). But the move from permissions to requirements in this argument is completely unwarranted. I think that the worry about a reliable source of obligations to parents is the primary motivation for the appeal to biology and so I focus on such an argument in the text.

35 The Journal of Philosophy 91 (1994) 333-63

36 See Walzer, Michael Spheres of Justice (New York: Basic Books 1983)Google Scholar.

37 The voluntarist worry is not, of course, the only worry about such a view. How can we determine which cultural understandings, if any, are shared? What is the relevant culture to which to appeal to understand the role of daughter or sister? I cannot here discuss all of the difficulties that face such a view.

38 These cases also reveal why Hardimon's appeal to birth as being different from impressment is useful, if at all, only in the case of biological familial relations. The more important worry about his contrast is that we can imagine many roles altered so that persons are born into them. Suppose that persons were born into the role of sailor in the navy as a result of having a parent in the navy and being born on board a naval vessel. And it might be ‘meaningful, rational, or good’ that persons play the role of sailor in the navy. See also Simmons, ‘External Justifications and Institutional Roles,’ 34-5.

39 I would like to thank Richard Fumerton, Thomas Hurka, and two anonymous referees for the Canadian Journal of Philosophy for helpful comments on earlier versions of this paper.