Just Caring for Caregivers: What Society and the State Owe to Those Who Render Care

Traditional considerations of justice for those who require caregiving have centered on what is due to the dependent person. However, considerations of justice also bear strongly on what is due to the caregiver. I focus on unpaid dependency work, too long treated as a private matter rather than a public concern. More is owed to those who render care: the division of labor is unjust, the nature of dependency work creates vulnerabilities for caregivers, and unpaid caregivers are disadvantaged in the world of paid work. Obligations to mitigate these facts are ultimately based on the truth that all members of society at some point in their lives benefit from caregiving and that noncaregivers benefit unfairly from the heavy distribution of dependency work to a small number of certain kinds of individuals. It is necessary to ask which agents of justice are responsible for remedying this state of affairs, and how. I propose a distributed scheme of obligation in which members of society and the state, as arbiter of social responsibility, share responsibility for the remedy. It is incumbent upon us as a society to refrain from making vulnerable the most essential among us, to reap benefits without sowing unjust burdens.


Introduction
Considerations of justice with respect to dependent persons in need of care often center on what is due to the recipient of care. Dependent persons may be patients who require intense caregiving outside of medical settings, children, elderly persons who can no longer live completely on their own, or disabled persons at any stage of life. Such a focus makes sense: dependent persons, after all, have often experienced great difficulty in receiving adequate care. As important as such 1 I am grateful to Anna Gotlib and FPQ's rigorous reviewers for critical feedback without which this paper would not have its current form. I am also grateful to my writing group at Southern Illinois University-Edwardsville, which supported me through multiple iterations of this paper. analysis is, considerations of justice also bear strongly on what is due to the caregiver who performs such dependency work, and different social policy recommendations result from considerations of the caregiver than result from simply considering the recipient of care. While I have argued elsewhere that justice for paid dependency workers is inseparable from justice for the recipient of care (Reiheld 2013), I focus here on what society owes to those who render care.
In the United States, for example, millions of people care for elderly parents or other dependent friends and family members in informal caregiving arrangements. Dementia care, only one kind of such caregiving, illustrates this nicely: 75% of persons with dementia are cared for by friends and family; in 2007, 10 million Americans were providing such care (Brodaty and Donkin 2009, 217). Millions of other caregivers are caring for strangers in exchange for pay in arrangements that vary from regulated formal pay to informal compensation for assistance. Care is rendered by a combination of paid and unpaid labor in service to people who are temporarily or permanently dependent. Caregivers often find their own health damaged through injury or stress with no respite in sight (Schulz and Beach 1999). Paid dependency workers, citizens, and especially immigrants face exploitation, low wages, and poor labor protections in both Canada and the United States (Reiheld 2013;Boyd 2006;Weir 2005). Both paid and unpaid dependency workers are more likely to be female than male. Unpaid dependency workers, both male and female, often have other formal paid employment unrelated to their caregiving functions. These formal workplaces have criteria of success and systems for promotion that penalize those who do not put paid work first. And the number of persons who require medical caregiving and elder care is increasing due to changing medical technologies and demographics, both in the U.S. and globally (World Health Organization 2002).
Such care may involve providing assistance for those who can no longer perform Activities of Daily Living (ADLs) such as dressing or bathing oneself, but also medical caregiving such as wound care, physical therapy, and proper medication administration. Such work is often referred to as "direct care," care which attends directly to a wide variety of daily needs. The caregivers who perform this sort of dependency work can be nonprofessionals and thus, unpaid. Eldercare in particular exemplifies this: nearly one-fourth of eldercare providers on any given day are unpaid-39.6 million people-spending an average of 3.2 hours each daily providing care (U.S. Department of Labor 2013). The great range of those who require caregiving and the care required demonstrate a key fact of caregiving: that all members of society, from birth to death and in between, at some point in their lives benefit from it.
In what follows, I focus on unpaid direct care, a form of dependency work that has too long been treated as a private matter rather than a public concern. I include unpaid medical caregiving as well as the more general considerations of caregiving for dependent children and elderly or disabled persons. Persons who do this work are generally picked out by their relationship with the dependent individual, e.g. family and close friends, rather than by the professional descriptions that accompany paid work. I elaborate on the lives of unpaid caregivers and briefly show why their lots are unjust rather than merely unfortunate, for at least three reasons: the division of labor is unjust, the nature of dependency work creates vulnerabilities for caregivers, and unpaid caregivers are disadvantaged in the world of paid work. I argue for obligations to mitigate these injustices, drawing on feminist philosophers' arguments that all members of society at some point in their lives benefit from caregiving, and noncaregivers benefit unfairly from the heavy distribution of dependency work to a small number of certain kinds of individuals. Yet we must go further than arguing for the moral basis of society's obligations to those who render care. We must also ask which agents of justice are responsible for remedying this state of affairs, and how.
I propose a distributed scheme of obligation in which members of society, and the state as an arbiter of social responsibility, share responsibility for the remedy. I consider an obvious objection to this scheme, that such obligations to care for caregivers may be just but are too expensive to be fulfilled, and offer several replies. Let us begin with further consideration of why society owes anything at all to those who render care, and why this generates both general obligations and specific obligations for particular agents.

I. The Origin of Obligations of Justice to Caregivers
There are many ways of parsing obligations when considering justice. One might ground them in responsiveness to rights claims. Compatibly and equally plausibly, one might ground obligations in a duty to avoid the perpetuation of injustice. Either way, the obligation is a way of satisfying a kind of moral debt that is owed; to say that one is not obligated is to say that nothing is owed. And yet much is indeed owed to caregivers.
Many discussions of justice begin with basic human rights, which are presumed to be universal. As Onora O'Neill (2001) notes, documents such as the 1948 Universal Declaration of Human Rights approach justice by "proclaiming rights": they proclaim "what is to be received, what entitlements everyone is to have," but say "very little about which agents and agencies must do what if these rights are to be secured" (183). The "allocation of obligations matters, and some means of designing and enforcing effective allocations is required if any ascription of rights is to have practical import" (184). It is obligations, after all, that are "the active aspects of justice" (185). Any universal claim, whether about general human rights or the sort of claims we are about to encounter regarding what society owes to caregivers with respect to rights claims and avoiding injustice, is hobbled without the ability to ascribe specific obligations to specific agents. My concern in this paper is to lay out general obligations based in an expanded notion of reciprocity and fairness aimed at reducing caregiver vulnerability, and subsequently to offer a scheme for assigning particular obligations to particular agents. Caregiving has traditionally been seen as a dyadic relation between the caregiver and the dependent person. In the paid-professional context of medical caregiving, the dyad is illustrated by the predominance of the terms "nurse-patient relationship" and "physician-patient relationship" as well as the more general "provider-patient relationship." This same notion extends to the relationship between unpaid caregivers and dependent persons. While there is great moral power in these obligations of care that individuals have to one another, this emphasis on caregiving as a dyadic relationship between individuals is problematic and belies the true nature of dependency work as a necessary function in a healthy society. However, it should come as no surprise. As Hilde Lindemann (2006) points out in her consideration of classic philosophical concepts, concepts that have become embedded in American culture, the traditions of reasoning about our obligations to one another are riddled with flawed assumptions that often go in sets, which she refers to as "pictures" (72-81). Two especially relevant pictures, for the purposes of this essay, are those of the independent person and the public-private split in society.
Though we are quick to judge those who we believe ought to be performing caregiving duties for dependent persons, we see dependency as a thing to be avoided in part because of the flawed picture of the person. Lindemann argues that this flawed picture privileges the notion of persons as fundamentally independent of others, as "detached from other people" and "self-sufficient" with "just as much social power as everyone else" (73). The dependent person cannot fit into this picture which, as Margaret Urban Walker has said, is "none of us at all times and many of us at no times" (1998,(21)(22). One of the critics of this picture is Eva Feder Kittay (1998;, who considers it in the context of America's welfare system. She observes that both the Right and the Left, liberal and progressive politicians, assume that a person "can only be incorporated as a full citizen by fulfilling the role of 'independent' wage earner" (1998,124). So powerful is this picture of the person that it is shared between widely disparate political value systems which share the claim that the goal of welfare reform should be paid work.
We see this ideal in its opposite: the "'icon of dependency' in America is the single black mother on welfare, who combines the characteristically dependent attributes of women, blacks and the poor" (Weir 2005, 309). Ronald Reagan famously caricatured such women as "welfare queens," doing no real work at all, a "controlling image" of black women (Collins 2008, 78) that shapes our views of women rendered vulnerable by caregiving. Paradoxically, in American culture our willingness to judge those we think ought to be doing unpaid caregiving for those to whom they are attached is paired with the ideal of being independent: detached, with equal social power to anyone else, and self-sufficient. And yet we are attached, do not have equal social status, and we are not self-sufficient-all of us have been dependent and will be dependent again at some point due to injury or age, whether temporarily or permanently. We see caregiving as damaging our independence, we see the dependent person as worth less, and all of this makes unpaid caregiving in particular appear to be without value. The flawed picture of the person as selfsufficient and equal in social status complicates our thinking about dependency and about caregiving for dependent persons.
The second flawed picture Lindemann identifies also complicates our thinking about dependency and caregiving. This picture depicts society as consisting primarily of two spheres, the public and the private. Though there is overlap in reality, we conceptualize them as distinct. The public and private spheres are distinguished not only by the spaces in which they take place-literally public spaces where one can be seen by others, and private spaces where one has privacy such as the home or a bathroom-but also by what sorts of considerations are seen to belong to each: The public sphere is the one in which people's freedom is secured by rights. . . . These rights . . . can fairly be summed up as the right not to be interfered with. . . . The public sphere is also the place for impartiality. If the public sphere is the sphere of rights, the private sphere is the sphere of the good. It's here that people pursue their various "thick" visions of the best way to live. . . . The private sphere is the one in which relationships and the responsibilities that arise from them are frequently unchosen. It's the place of favoritism, because it's the sphere of friendship, love, and families-relationships in which another person is singled out for special consideration rather than treated like everybody else. (Lindemann,(76)(77) Allison Weir (2005) suggests that such patterns are instantiated in "global institutions of public work and private family that render these two spheres inherently conflictual" (308). Because of this picture, it is an uphill battle to argue that events that occur in the private sphere can be the subject of rights or the purview of justice. For example, marriage is a conceptualized as private, and it was not until the mid-1970s that American states began to consider that it might be possible for a husband to rape his wife; the last states (Oklahoma and North Carolina) to withdraw marital exceptions to rape as a crime did so as recently as 1993 (Bergen 1999).
It is this distinction between private and public that makes caregiving seem to be a matter of one-to-one relations, of an individual's obligations to another dependent individual. As Weir says in an observation combining both the picture of society and the picture of the person, "the valuation of care is romanticized, feminized, and privatized, so that it is inevitable conflicts with the masculine model of work-with the valuation of independence, of wage-earning and individual career success in the public sphere" (2005,314). It is this flawed picture of society, combined with the flawed picture of the person, that makes it seem so obvious that caregiving is no concern of justice at all. This is, however, patently not the case. Caregiving is a fine illustration of a set of conditions in which the private-public distinction is detrimental to right action. It is precisely because of reliance upon this flatly false division between the public and the private that traditional theories of justice are very poorly suited to deal with caregiving, which is relegated to the realm of the "private." As Kittay has said, "Theories of the just state . . . fail to include among the social goods those that bear on the needs of dependency workers, dependents, and the relations of dependency" (1998,131).
How so? Caregiving is a relationship not only between caregivers and dependent persons but also between caregivers and society, and thus spans the alleged public-private divide. As Martha Fineman argues, "without aggregate caretaking" for all persons but especially for children, "there could be no society, so we might say that it is caretaking labor that produces and reproduces society. . . . The uncompensated labor of caretakers is an unrecognized subsidy, not only to individuals who directly receive it, but more significantly, to the entire society" (2004,48). This is not simply a matter of individual members of society treating individuals fairly. Rather, caregivers are owed, at a minimum, decent treatment by those whom they serve-society as a whole-as a matter of reciprocal justice. According to Susan Moller Okin (1989), some injustices result from social structures that create vulnerabilities that would not otherwise exist. Consider the social structures embedded in marriage, often a basis for caregiving relationships. If institutions such as work and government determine wages, benefits, and social services based on assumptions about the institution of marriage and the financial security and caregiving it often entails, then individuals are made vulnerable by this system in the eventuality that marriage is undesirable, inaccessible, or insufficient for individuals.
Since the vulnerable individual is vulnerable because of society, society has an obligation both to attend to the vulnerability and to alter the structures that cause it (Okin 1989, 138-139). Many employment and benefit systems make caregivers vulnerable because they assume there is access to free dependency work from another person. Meetings and work hours, demands for unscheduled labor to complete projects, required travel for work, and requirements that welfare recipients seek paid work all assume that the worker is not the caregiver and that someone else is. All the while, society denigrates and fails to support such caregivers because of the high value placed on paid work. As Eva Feder Kittay and Ellen Feder (2002) have pointed out, these caregivers and their families generally bear the costs of caring for dependent individuals, "yet the social benefits of such care are distributed throughout society" (3). Benefitting from socially necessary labor while not giving back commensurately and actively denigrating those who perform such labor violates reciprocity and is thus deeply unjust. Society as a whole, and component members and institutions of society that do not support caregivingsuch as noncaregiving individuals, the state, and corporations-are free riders on the unpaid labor of caregivers (Folbre 2001;Fraser 1994, 613;Parks 2003, 489) and contribute to the vulnerability of caregivers.
Such vulnerabilities accrue in what Florencia Luna (2009), writing on research ethics and medical consent, has called "layers of vulnerability" (121). Rather than seeing certain persons as intrinsically vulnerable to coercion or exploitation, Luna argues that we should see them as "rendered vulnerable" by different layers of vulnerability (128). While the reader may be forgiven for thinking this is just another way of conceptualizing intersectionality, it is not, for these layers can be removed by changing "a particular situation that makes or renders someone vulnerable" (129). One need not remove all layers of vulnerability in order to reduce the vulnerability to which a person finds herself subject. Removing even some can constitute a significant improvement for the vulnerable person. Thus, on grounds of reciprocal obligations, society owes caregivers decent treatment and reformation of relevant social structures. At the end of this paper, I will propose just such specific measures to remove the resultant layers of vulnerability.
I have characterized the ground of reciprocal obligation here as one between caregivers and society rather than, more classically, between caregiver and patient; caregivers provide a service that is essential to society and thus society owes caregivers relevant compensation and improved treatment that reduces vulnerability. However, one need not embrace a traditional one-to-one reciprocal obligation in order to find a ground for society having obligations with respect to caregivers, dependent persons, and the caregiving relation. There is no doubt that caregivers are made vulnerable by the caregiving relation. In Kittay's 1998 article "Welfare, Dependency, and a Public Ethics of Care" and her 1999 book Love's Labor, she seeks a system in which unpaid caregivers-overwhelmingly women, though increasingly men 2 -will not be made vulnerable by their work. This is precisely the sort of theory of justice that I am after, as well. One of Kittay's signature contributions to social and political philosophy is her principle of doulia, "a concept of interdependence that recognizes what is not precisely a relation of reciprocity, but a relation [characterized] as 'nested dependencies.' These link those who need help to those who help, and link the helpers to a set of supports" (1998, 133; my emphasis). Claims to such a set of supports are rooted in the caregiver's "entitlement" to expect that "one can give care without the caregiving becoming a liability to one's own being" (1999,66). This is less a matter of one-to-one reciprocation, as I had argued for earlier with the caregiver-society obligation, and more a concept of "social cooperation" (Kittay 1998, 133) and "reciprocity-inconnection" (Kittay 1999, 67) rather than the classic "exchange reciprocity"-you scratch my back, I'll scratch yours. Kittay herself describes the nested dependencies of doulia as a matter of "what goes around comes around," adding, "A connectionbased equality calls upon those within the nested set of social relations to support the dependency worker sufficiently, so that she is not made unduly vulnerable as she answers to the vulnerability of another" (1999,170). She extends this to the context of welfare, arguing that welfare should be expanded beyond the poor to focus less on poverty and more on vulnerability through caregiving. In other words, we refocus our welfare obligations from a ground of charity to a ground of reciprocity qua social cooperation. This would mean that anyone who might be made vulnerable through caregiving should be given access to welfare, without focusing on means testing as the criterion for eligibility.
This claim is a universal claim of general obligation. But recall O'Neill's comment regarding universal human rights. We cannot adjudicate universal rights claims without some way of assigning particular obligations to particular agents. As I shall argue, particular obligations cannot fall-must not fall-only on government as the agent of public policy.
Before I take on that task, it should now be clear that on grounds of both one-to-one reciprocal obligations and doulia's more nuanced nested reciprocities of social cooperation, society owes caregivers decent treatment and reformation of relevant social structures. To see what specific obligations follow from this general social obligation to create a just system of caregiving, we must further examine the specific facts of caregiving.

II. The Labor of Dependency Work Carries Unique Burdens Which Are Unjust
Although caregivers perform work that must be done so that all of society may benefit, caregivers not only experience a substantial drain on time and energy but can also experience illness, injury, or death as a result of their dependency work. A 1999 study of elderly spousal caregiving (Schulz and Beach 1999) found that caregivers who reported strain had a 63% increase in mortality during the four-year study period relative to other caregivers and noncaregiving controls. The authors conclude that "being a caregiver who is experiencing mental or emotional strain is an independent risk factor for mortality among elderly spousal caregivers" (1999,2215). The authors also note that previous work has shown increased morbidity among those caregivers of all ages who live with the care recipient (2219). Such burdens add layers to the vulnerability of caregivers, even moreso to a surprising demographic: in 3.2% of households with children, at least one child performs dependency work for a family member, usually for an adult but sometimes for a sibling. Children who provide help with ADLs and additional duties of care report abnormally high levels of belief that expressing their emotions is not valuable, that too much is asked of them, and that they cannot rely on anyone else. They also report lower levels of attendance and participation at school, and inability to complete homework (Hunt et al.,(10)(11)(12), further layering vulnerability.
Hiring paid dependency workers to relieve these burdens and remove this vulnerability from unpaid caregivers may seem like a solution to the problem of imposing unjust burdens. But it is not. All too often, unjust burdens are simply transferred to a class of worker with low protections, low salary, and long hours (Reiheld 2013;Kittay 1998;Weir 2005). Many of these workers are women of color and/or migrants who enter the U.S. illegally or on restrictive work visas like the H-2B Temporary Non-Agricultural Worker visa. In Canada, the Live-In Caregiver Program (LCP) allows immigration for the explicit purpose of providing care for children, seniors, or people with disabilities, without supervision-and thus also without recourse or appeal in the event of mistreatment-in a private household; workers who enter Canada under the LCP are almost all women (Boyd 2005, 5). As Weir has noted, such women are compelled to leave their children behind in their country of origin or, if they bring them along, are unable to provide care for their own children while providing care for others at wages insufficient to hire safe caregiving of their own. Wages for an au pair performing childcare work in the U.S. can be under $140 a week with no real oversight or way of restricting work hours (Weir 2005, 191-193). One of the eight lowest-paying jobs in the United States is personal and home care aides; when paid, these workers receive an average of only $9.75 an hour (Tahmincioglu 2014). The reality of such exploitation indicates that displacing the burden moves the target of obligations from unpaid caregivers to poorly paid and poorly protected caregivers, who already suffer additional layers of vulnerability imposed by immigration. This is no solution to problems of injustice.
If we owe anything to those who render care, then we owe dependency workers a more reasonable level of burdens than they typically have. That some caregivers shoulder the burden of their own premature deaths or increased morbidity, learn that they cannot rely on others, or experience educational impairment is surely too much. In this case, there is an obligation to reduce the mental and emotional strain experienced by unpaid caregivers by peeling away vulnerabilities where possible, and to do so without simply displacing them onto vulnerable paid caregivers. After establishing additional obligations that society owes to unpaid caregivers, I will show how they can be assigned to particular agents.

III. Division of Labor Is Unjust
The division of labor within caregiving also demonstrates classic injustice. Noncaregivers benefit unfairly from the heavy distribution of dependency work to a small number of certain kinds of individuals. The relevant line of distribution of unpaid caregiving is gender. 3 Women perform roughly two-thirds of all unpaid caregiving (Family Caregiver Alliance 2012). One cause of this gender discrepancy is what Jennifer Parks (2003) describes as socially pervasive "representations of caretaking as unskilled, voluntary, private 'women's work'" that has "resulted in the ghettoizing of care work, turning it into unpaid, altruistic familial labor or poorly paid labor provided by a transient, exploited, impoverished workforce" (493).
Gendering of such work as "women's work" isn't the only reason for its low social valuation, however. Marilyn Waring argued in Counting for Nothing: What Men Value and What Women Are Worth , originally published in 1988, that the gendered labor done primarily by women is actually both unpaid and largely invisible to global means measuring economic productivity precisely because it is private. In a follow-up essay, Waring (2003) points out that, historically, work that is "consumed" inside the home is considered private household labor and thus is invisible to economic measures; only work "produced" to be consumed by others is measured by economic measures such as the United Nations System of National Accounts. Thus, paid caregiving done outside of private relations has economic value, whereas work done inside the home by unpaid caregivers who are generally family or friends is significantly devalued by society's primary economic and other social measurements. As economist Nancy Folbre (2001) has said, "Conventional economic theory concedes that human capital . . . is an important factor of production. . . . If human capital is so important, maybe we should pay more attention to how and where it is actually produced" rather than treating people as "exogenously given" to the economic system by forces outside of it (71).
Perhaps because of this socioeconomic devaluing of dependency work, it is performed by men far less often than it is performed by women. One direct indicator of this disparity in unpaid caregiving is the varying degree to which men and women leave work, or take leave from work, to attend to familial responsibilities. More and more men are doing so but the disparity still exists: in caregiving overall, two-thirds of caregiving is performed by women (Family Caregiving Alliance 2012).
Use of leave under the Family and Medical Leave Act (FMLA) mirrors these numbers: according to a 2012 nonprofit report based on data from the U.S. Department of Labor (DOL), 56% of those who took leave in 2011 were women (NPWF 2013, 1), meaning that 44% of those who took leave in 2011 were men. The report notes that "the rate of leave taking among men has increased in small but steady increments in the 20 years since the FMLA's enactment" (NPWF 2013, 1). Shifting the metric from the percentage of leave-takers who are a certain gender to the percentage of men taking leave in the previous 18-month period, the report documents the trend: in 1995, 13% of men reported taking leave; in 2000, 14 percent; in 2005, 16%. Of all those taking leave, 19% reported taking leave to "care for a parent's, spouse's or child's health condition" (NPWF 2013, 1).
This may appear promising in terms of the degree to which men are increasingly performing women's work, and it is. But the numbers are not as straightforward an indicator as they might first appear. After all, the FMLA simply requires that employers retain the employee and allow them to return to work after the leave period has ended. It does not require that they be paid during the time of the leave. This explains an ongoing gender disparity: the nonprofit report also points out that "a growing share of workers need leave but do not take it-most commonly because unpaid leave is unaffordable-and women bear a disproportionate burden" (NPWF 2013, 2). In fact, women made up 64% of those who needed leave but did not take it.
Gender disparity in caregiving, and the burdens thereof, become even clearer when we look away from those who take leave from paid work to do caregiving and focus on those who leave paid work entirely. In October of 2014, of those who left work because of "family responsibilities," 29% were male (71,000 men) and 71% female (176,000 women). A year later in October of 2015, the number of men remained 71,000 but the number of women leaving work because of "family responsibilities" had decreased to 135,000. So while there was no increase in the number of men leaving work because of "family responsibilities," they now constitute 34% of those doing so (BLS 2015). From year to year, the percentages fluctuate but one thing is clear: this labor is still heavily gendered, and those who leave paid work to do unpaid caregiving are still overwhelmingly women.
The costs of leaving work are significant and often unmitigated. In rare cases, persons who quit work to care for a seriously ill family member are still eligible to collect unemployment; eligibility depends on each state's requirements for severity of illness and which family members are covered. Folks who leave the workforcewhether they be men or women-to care for ill family members suffer degradation of their subsequent earning capacity in addition to loss of income for the duration of caregiving. The burdens of continuing to do paid work while also taking on the second unpaid job of caring for dependent persons, and the burdens of leaving the workforce to care for dependent persons, fall predominantly on women.
Of course, if indeed women were somehow better suited to perform "women's work," this distribution of labor might not be unfair. 4 After all, not all inequalities are injustices. However, the skills required to perform activities of daily living and administer medications and simple medical treatments are clearly skills that men also possess. Indeed, the average man is stronger and taller than the average woman and may be better suited to much of the physical labor involved in caring for a dependent adult: bathing, lifting, dressing, and so forth. And yet it is women who do most of the caregiving and on whom the burdens of caregiving fall most heavily, women who are least likely to take leave from paid work and most likely to leave the labor market. That certain kinds of work are assigned according to attributes irrelevant to the performance of that work is a hallmark of injustice, and we see it here.
Remedy will require a change in the basic structure of family-within marriages, sibling relationships, and other aspects of family life. Kittay has acknowledged this, noting that de-gendering dependency work may require "public programs of educating for dependency work-especially for young boys and men" (1998,137). This is part of a change in the sociocultural perception and valuation of any such thing as "women's work." We should no longer devalue unpaid work such as caregiving simply because it is unpaid, simply because it cannot easily be measured in economic terms (Folbre 2001;Waring 2003Waring , 2004, simply because it is done by women and, even when paid, receives a lower wage.
Thus, there is an obligation to alter social perceptions, attitudes, and basic structures that result in gendered division of, and devaluation of, the labor of caregiving. I shall argue for one more such general social obligation to caregivers before we address which agents must fulfill such duties.

IV. Treatment of Unpaid Dependency Workers in the Context of Paid Work Is Unjust
Economic pressures often drive businesses to concentrate only on what will serve their bottom line, 5 yet businesses are also part of free-riding society. As Susan Moller Okin (1989) says: . . . many full-time employers assume, in innumerable ways, that "someone" is at home at least part-time during the day to assume primary responsibility for children. . . . It is assumed by the high degree of geographical mobility required by many high-level management positions. It is also implicit in the structure of the professions, in which the greatest demands are placed on workers at the very peak of their child-bearing years. (156) Work is equally as insensitive to those who care for any sort of dependent person. Many of these aspects became apparent in the discussion of leave from work, and leaving work, in the previous section. And yet, as Okin notes, nothing in the nature of work makes it impossible to adjust to the fact that people are caregivers as well as workers (1989,5). She argues that "any just and fair solution to the urgent problem of women's and children's vulnerability must encourage and facilitate the equal sharing by men and women of paid and unpaid work, of productive and reproductive labor" (171).
It might be argued that the FMLA does just this. Indeed, when Okin's Justice, Gender, & the Family was published in 1989, the FMLA was not yet law. Were a good solution to the problem Okin describes so well already on the books, we would need only concern ourselves with implementation. However, the FMLA is deeply unsatisfying. For one thing, it requires only unpaid leave, forcing paid workers into exactly the kind of choice discussed in the last section: to take leave and lose pay, or to continue working and take on the second unpaid job and unpredictable hours of caregiving. For another, it only covers employees who work for a "covered employer" at a worksite within 75 miles of which that employer employs at least 50 people, for whom they have worked at least 12 months cumulative and 1,250 hours during the previous 12 months (U.S. DOL 2015). Approximately 40% of the actual work force is not even eligible due to these and other restrictions, including what Ellen Bravo calls a "very narrow definition of family," which does not include samesex partners or one's own grandparents (Ludden 2013a). In addition, the FMLA only allows leave from 12 workweeks during any 12-month period (U.S. DOL 2015). This is insufficient for most caregiving, much of which ranges well above the covered 3 months. According to the Bureau of Labor and Statistics, duration of eldercare was less than 1 year for only 22.9% of caregivers, while 26.9% of caregivers provided care for 1 to 2 years, and on the high end, 15.4% had provided care for 10 years or more (BLS 2015). Despite its flaws, FMLA is a boon for those individuals who qualify and who can afford to take advantage of it.
Far superior for the purposes of my argument are various paid family leave pilot programs being implemented in several states that have "paid family leave insurance programs" (Ludden 2013a). These seem similar to the workers' compensation insurance programs in which employers pay into the program and the program pays out to employees who qualify. The issue of who qualifies, however, and for how long still holds. How long will benefits last? Will parents of any gender-fathers as well as mothers-receive them for caring for their children, and for the same duration? Will married or unmarried children of either gender-sons as well as daughters-receive them for caring for their parents? What of grandchildren who care for their grandparents, who are currently not eligible for FMLA coverage? What of close friends, nieces, or nephews of the dependent person who perform unpaid dependency work? Will they also be eligible? We must ask ourselves whether what we want is to support certain kinds or degrees of family-based dependency work, or whether we want to support dependency work per se because of its value to society. It is the latter, a kind of universal support for caregivers, that Kittay (1999) argues for when she calls for a "universalistic model" (128) of entitlement to a "socially supported situation in which one can give care without the caregiving becoming a liability to one's own well-being." (66) Without changing the basic structures of paid work, including programs that allow paid leave as well as unpaid leave, and for a wider variety of those who often perform unpaid dependency work, caregivers will continue to be vulnerable precisely because they provide care. Thus, there remains unfulfilled an obligation to reform the basic structure of paid work so that it does not inherently disadvantage those who perform unpaid work, and a concomitant obligation to alter social perceptions of who should engage in caregiving. Now that we have seen the general social obligations that society owes to unpaid caregivers, we can determine which agents are responsible for fulfilling them.

V. The Remedy: A Distributed Scheme of Obligations, Its Policy Ramifications, and Required Changes in the Basic Structures of Society
I have established that the obligations society has to caregivers are grounded in the fact that society demands dependency work and is thus responsible to dependency workers for a just distribution of burdens. The obligations are to reduce emotional and mental strain on unpaid caregivers. to alter social perceptions, attitudes, and basic structures that result in gendered division of the labor of caregiving. to reform the basic structure of paid work so that it does not inherently disadvantage those who perform unpaid dependency work. to alter social perceptions of who should engage in caregiving.
These are, as Kittay would have it, universal obligations. Addressing each to whatever degree possible will remove layers of vulnerability, reducing injustice. For instance, the first obligation, to reduce emotional and mental strain on unpaid caregivers, is indeed a general claim, even a universal one. However, the obligation is strongest to those who suffer the most, such as children who perform dependency work, as discussed earlier in this paper. By removing layers of vulnerability from the most vulnerable, we best satisfy this obligation. We satisfy it by removing layers of vulnerability from any caregiver made vulnerable by emotional and mental strain. What does society owe to those who render care? What is just caring for caregivers? To fulfill these obligations. But as I have argued, we are only partially done once we understand obligations of justice in relation to caregivers. The final step is to be able to assign responsibilities for implementing these obligations to particular agents.
Classically, the distinction between public and private has been used to assign these obligations, deriving unfairly limited public policies and relegating such work to those held responsible for the private sphere. With those who have critiqued the public-private distinction before me, I am calling for a principled erasure of the flawed picture of society on the grounds that it does not suit the pursuit of justice in this case (or, it could be argued, in most others). This opens up a whole new range of agents and concerns that had previously been exempt from dealing with dependency work, especially in its unpaid form. How can we specify those agents and their obligations?
Seeking to provide a "robust and realistic account of agents of justice who are to carry the counterpart obligations," Onora O'Neill (2001) focuses on the assignment of obligations to what she calls "agents of justice" (186). O'Neill describes two major kinds of agents of justice (hereafter referred to as AoJs): primary AoJs and secondary AoJs. These differ in their capabilities. This matters, because agents of justice must have "specific, effectively resourced capacities which they can deploy in actual circumstance. . . . It is the specific capabilities of agents or agencies in specific situations . . . that are relevant to determining which obligations of justice they can hold and discharge. . . .
[This] provides a seriously realistic starting point for normative reasoning" (189). Primary AoJs have the capacity to determine how principles of justice are to be institutionalized within a certain domain and typically have some means of coercion. They also have the ability to construct other agents or agencies with specific competencies regarding that institutionalization. Primary AoJs could be a body like a legislature or one without a formal structure like a loosely organized town meeting, or an individual like a monarch or elected leader-in sum, the state (181). Secondary AoJs, on the other hand, are typically thought of as contributing to justice by meeting the demands of primary agents and of justice more generally. Secondary AoJs can be nonprofit organizations like the Robert Wood Johnson Foundation, local churches, the Alzheimer's Association, the American Nursing Association, or even individuals within the community whose actions can have effects on justice relations within the community (181).
As O'Neill sees it, if agents have the capability to fulfill the remedy, then they have the obligation to fulfill the remedy (184-185). One does not become an agent of justice because one has served justice, but rather because it is possible for one to do so. According to O'Neill, we can determine the obligations any particular AoJ has by assessing that agent's capabilities in light of the overall task of achieving justice and eliminating injustice. In O'Neill's eyes, all of the AoJs she discussed (primary or otherwise) have capabilities that can contribute to greater justice. That they may fail to do so does not make them any less agents of justice, though it does make them less praiseworthy or more blameworthy depending on the nature of that failure. Responsible AoJs work together in a distributed scheme of specific obligations in order to fulfill the duties of justice: "Justice has to be built by a diversity of agents and agencies that possess and lack varying ranges of capabilities, and that can contribute to justice-or to injustice-in more diverse ways than is generally acknowledged" (O'Neill 2001, 194). However, the primary AoJ must be convinced that certain principles of justice apply, for without the primary AoJ, pervasive change is unlikely.
O'Neill concerns herself with the role of secondary AoJs in a weak state, a state that does not use its power to pursue injustice-a rogue state-but rather one that is incapable of pursuing justice or compelling it (190). She contends that when states fail as agents of justice, it is because they "lack a specific range of capabilities that are needed for the delivery of justice" (190). In these cases, secondary AoJs step in to take over these functions to the best of their capabilities. While they may have some of the necessary capabilities, they will lack the enforcement ability that is the legitimate authority of government.
However, I contend that secondary AoJs have a role to play even in strong states beyond just "badgering" the primary AoJ to pursue justice (think of the American civil rights movement and women's movements). Even in strong states, secondary AoJs can actually implement cultural and practical changes that fulfill obligations within their capabilities. Based on O'Neill's conception of secondary AoJs, in the case of meeting obligations to those who render care, we should indeed draw on both strong primary agents of justice and secondary ones. We ought to expand the notion of agents of justice well beyond the kinds of organizations O'Neill envisions to include quite small community groups and even individuals.
Nonetheless, primary AoJs are important, for only they have the enforcement capabilities and large-scale implementation capabilities necessary for sweeping change. They are uniquely capable of removing certain layers of vulnerability and fulfilling particular aspects of obligations with respect to caregivers. These include family leave policies and labor conditions that make caregiving less burdensome, and remove layers of vulnerability, for caregivers who perform paid work outside the home. It may be tempting in the United States, where many people have a strong resistance to government involvement in business, to suggest that policies should at most provide guidance to those who wish to voluntarily comply. Alas, as Nancy Folbre (2001) has pointed out, "bad capitalisms" too often outcompete good ones: "Employers who assume their fair share of social costs . . . will operate at a disadvantage" (186) if others are permitted to refuse to do so. Policies that require employers to give reasonable paid leave, to not penalize employees for caregiving, and so forth, must involve mandatory compliance. Mandatory compliance requires enforcement mechanisms, and only the teeth of a primary AoJ can effect such change. While both secondary and primary AoJs can remove layers of vulnerability and do their part to fulfill society's obligations of justice to caregivers, the most effective endeavors in pursuit of justice will indeed, as O'Neill said, require varying ranges of capabilities and thus a diversity of agents and agencies.
Within this diversity, many secondary agents of justice have a substantive role in realizing obligations to caregivers. There are some aspects of supporting caregivers and removing layers of vulnerability that government alone lacks the capability to perform, not because it is weak but because of what is required. While doulia and one-to-one reciprocity are the grounds of social obligations to caregivers, the state alone-as the primary AoJ-will not be sufficient for fulfilling those social obligations. In order to effectively fulfill social obligations to caregivers, a concerted scheme of distributed obligations and interlocking policies will need to be put in place. Consider the following proposal.
Unpaid care work cannot continue to be performed primarily by individuals, relegated to the private sphere. The burdens of intense caregiving are too great for both children and adults performing dependency work, as I have shown. To reduce emotional and mental strain on unpaid caregivers, we will need the state and other agents of justice to offer actual assistance to caregivers in the performance of their duties. I concur with Nancy Fraser when she suggests that some socially "supported care work would be located outside of households altogether-in civil society" (1994,613). Examples of concrete ways this has been done in the past include respite care centers where dependent persons can be brought one or more days a week, where their needs will be met by others. I have already warned against displacing the burdens of caregiving onto those already burdened in exploitative ways. I recommend against that. However, there are ways to set up respite care centers and in-home care assistance such that those who staff these endeavors are individual agents of justice from within the community who volunteer to help, or who are decently paid with workplace protections. 6 Existing iterations or respite centers have been local and not consistently available, due to the lack of involvement by a widely effective primary AoJ, and prone to exploitation of paid staff. Such centers can be supported by a combination of government funding and volunteerism, by the involvement of civil society. However, volunteers of both genders will not be forthcoming unless the state or strong, respected secondary agents of justice can broker a national debate on caregiving in order to indicate a symbolic endorsement of the reconsideration of the value of caregiving and who should perform it, thereby also raising the issue of the basic structure of the family. The need for such public policies is reinforced by Fraser, who notes that "Women today often combine breadwinning and caregiving, albeit with great difficulty and strain. . . . [The state must] ensure that men do the same, while redesigning institutions so as to eliminate difficulty and strain" (1994,611). However, the state cannot accomplish this simply by regulation.
Members of society who have so far been unwilling to take on caregiving must change their willingness to take on such duties and become agents of justice, themselves. Distributing caregiving more fairly and widely will lessen its burdens on all caregivers. This will also require a change in society's culture of work. At the very least, employers would not seek to skirt inadequate but relevant laws such as the Family and Medical Leave Act. Acting rightly, government would enforce such policies with substantive disincentives. But there will also have to be a significant change in the underlying assumptions of employers. Per Okin, they could no longer assume that all paid workers have someone at home to take care of the family. Doing so makes caregivers who are doing paid work particularly vulnerable. In the same vein, Fraser proposes that, in a work culture which supports caregiving on grounds of justice rather than charity, "all jobs would assume workers who are caregivers, too . . . and all would have employment-enabling services" (1994,612) such as flexible scheduling, several different promotional tracks, and leave-paid, and not overly restrictive-to care for dependent family members. Employers are capable of contributing to justice as secondary agents of justice, and they are obligated to do so. A strong primary agent of justice dedicated to fulfilling these obligations would ensure it.
One of the most compelling objections to increased involvement by the state in supporting members of society has always been that it costs too much. This objection carries even more weight as the global economy, and the U.S. economy with it, hobbles slowly out of the recession that began late in the first decade of this century. This is compelling not because money is the goal of all human endeavor, but because all government-run social programs require funding; adding increased funding to directly support caregivers, and an infrastructure to coordinate support for caregivers, is no small thing. Taxes will go up or the state will have to borrow funding on top of the already significant national debt. Smaller government and fiscal responsibility are run over roughshod by such a proposal, folks might say, as they have done in response to the Affordable Care Act of 2010 (a.k.a. Obamacare). There are at least three ways to respond to the cost-based objection.
First, one might reject the consequentialist perspective wholesale and say that money is no object when doing the right thing, especially something so fundamental to the lives of humans as support for our nested dependency upon one another. However, this seems to ignore that doing the right thing requires resources. Without expanding the pool of resources in some way, doing this right thing will require undoing some other (perhaps also right) thing.
This leads me to a second possible reply: I do not reject consequentialism outright, but I reject the cost-based objection's focus on the negative consequences, namely the size of government, the increase in taxes, and fiscal responsibility. I am entirely comfortable with expanding the pool of resources in order to do more right things. I believe that we must mitigate the consequences that fall upon the tens of millions of unpaid caregivers in the United States who perform this essential labor, contributing to the well-being of not just the dependent person but society as a whole. To focus on plausible duties to restrict the growth of government or keep taxes and debt stable for this and future generations is too petty and narrow a consequence, when faced with how critical it is to maintain and support caregiving relationships. Justice, as I have argued in this paper, has so strong an argument in favor of support for caregiving that the concomitant obligations to caregivers far outweigh any plausible duties to restrict the growth of government or keep taxes and debt stable. Now I come to the third and final reply to the cost-based objection. It may well be that we can all accept a modified consequentialism based on the importance of support for caregiving relationships, and that we can also embrace a principled pursuit of justice in support of caregiving relationships. Doing so will not be cheap, but the distributed schema of obligations keeps the costs as low as possible while holding agents of justice throughout society accountable for exerting their capabilities to achieve a more just society. After all, it is not only the primary AoJgovernment-that is responsible. Employers can change their family leave policies and work-schedule expectations, offering flex time, supporting telecommuting, or adjusting expectations of availability for late meetings so that family members can reliably be home in time to provide promised care. Non-governmental organizations such as the Alzheimer's Association or the National Alliance on Mental Illness, social welfare organizations such as Rotary, the Knights of Columbus, or the YMCAs of America, and local communities may be able to set up respite care programs or socalled "adult day care." Friends and neighbors might offer to stay with and care for the dependent person to give respite and support. Relatives whose gender or positioning have allowed us to assume caregiving was not their job can either take on more of the caregiving or give better support to the caregiver.
By distributing obligations and calling on all sectors and members of society to improve support for caregiving relationships, we can accomplish justice in a way that is not only just-asking those who benefit from, and are capable of supporting, caregiving relationships to do both-but also cost-effective if we are to do it well.

VI. Conclusion
Recent years have seen an increasing number of primary agents of justice playing a key role in support for caregiving. States and municipalities have been experimenting with innovative solutions to the flaws in the reach and scope of the FMLA. Vicki Shabo, director of the National Partnership for Women and Families, has been tracking such solutions: [Ten] cities and states have passed paid sick days or family leave or expansions to family leave laws. New York City put in place a paid sick days law which covers every worker in the city, guarantees more than one million access to paid sick time and the rest access to unpaid sick time. That means you can't lose your job if you're sick or a family member is sick or needs medical care. Portland and Jersey City did the same things. Then Rhode Island, most significantly, became the third state in the country to have a family leave insurance program in place. Which means that workers there would be able to take time when a new child joins the family or when a family member is seriously ill. (Ludden 2013b) All good news, to be sure, and these options provide more reinforcement from primary agents of justice that this is indeed an obligation that a just society has toward its members. However, if I am correct, laws alone will not do the trick. American culture will have to change, a change helped and not hindered by the work of primary agents of justice, and multiple primary and secondary agents of justice will need to become involved. Laws like the FMLA are promising but constitute a patchwork, and cannot do the job alone. If employers fight compliance or employees remain afraid to take leave because of the consequences of doing so, no amount of family leave in the world will suffice. Nonetheless, these are promising changes, like the changes before them, for they indicate that even in the throes of recession, as employers progressively strip benefits from employees (Ludden 2013b), there is nonetheless increasing acknowledgement that the current structure of American society ill-serves unpaid caregivers. Piece by piece, we can strip away the layers of vulnerability imposed on caregivers by the way we (fail to) value such work. To embrace that as a society calls for each agent to embrace their share of fulfilling obligations to unpaid caregivers.
I contend that the capabilities for fulfilling society's duties towards caregivers determine the specific obligations of agents of justice. They also imply necessary involvement of noncaregiving men and women, employers, social welfare organizations, non-governmental organizations, and the state at all levels. Without this distributed involvement, the burdens of caregivers may increase as caregiving needs increase and will, at best, remain at the current unjust status quo. With it, to whatever degree possible, we relieve these burdens incrementally, layer by layer. A full deployment of responsible agents of justice may well eliminate them.
The personal actions of those who serve the public good must not remain a matter for private redress. Surely it is incumbent upon us as a society to refrain from making vulnerable the most essential among us, to reap benefits without sowing unjust burdens.