Key Words: agency, diversity, interpretive sociology, social structure.
We are excited and encouraged by the enthusiasm that our work on ambivalence has generated. Considering the comments in confidential reviews and, now, the public responses that you have just read has been intellectually stimulating. We thank all of the reviewers for the time and effort that they spent reviewing and responding to our work, especially Sara Curran for her innovative and challenging feedback. Like our colleagues, we commend Alexis Walker for proposing this opportunity for open debate and exchange. The willingness to publish theoretical work and to foster discussion supports the goal of articles such as ours to move the understanding of family relations forward by building ideas.
Clearly, we are not the first to suggest ambivalence as a useful concept in the study of family relations. However, ours is among very few attempts to develop ambivalence as a distinctly sociological concept connected to strong theoretical traditions in our discipline. Our response addresses four key issues. First, we hope to put to rest the notion that ambivalence can somehow be subsumed within the solidarity model. Then we discuss three interlocking issues related to our thinking about family relationships: assumptions about "science," the distinction between theory and ideology, and views about what constitutes the "stuff" of sociology. We use this approach to highlight fundamental differences (and sometimes, parallels) between our views of theory, society, and doing sociology and those of our commentators. This thematic approach also minimizes repetition and maintains our focus on major points. Before exploring these four issues, we make a summary comment or two about each of the responses.
Vern Bengtson, Roseann Giarrusso, Beth Mabry, and Merril Silverstein have launched a defence of the solidarity model more than a critique of our conception of ambivalence. Nonetheless, their discussion indicates serious misunderstandings about ambivalence and a quite different perspective from ours about family relations, science, and theory. Kurt Luscher is in fundamental agreement with us about the value of ambivalence as a concept for studying intergenerational relations, but he takes exception to some of our interpretations of his work with Pillemer (Lubscher & Pillemer, 1998), espouses some assumptions about science and theory that we do not share, and presents his own typology on ambivalence. His closing outlook suggests the promise and growing relevance of ambivalence to studying ever more diverse family relations. Finally, Sara Curran provides a good example of how attempts to think in new ways, as we have tried to do in presenting a sociological conception of ambivalence, can prompt others to take off with an idea in both expected and unexpected directions. Despite some points of divergence, she shares with us the value of theorizing before empirical evaluation, provides a sound summary of our article, and discusses motivation and accountability as ways of extending our conceptualization of ambivalence.
AMBIVALENCE AND SOLIDARITY: COMPATIBILITY? YES. CO-OPTATION? No.
Rather than rehash the history and shortcomings of solidarity as a concept, we focus on the assertion by Bengtson and his colleagues that ambivalence can somehow be subsumed within the solidarity model. Counter to their claim (Bengston, Giarrusson, Mabry, & Silverstein, 2002, p. 570), we do not suggest that ambivalence is incompatible with solidarity. This is largely because we do not consider them to be on the same plane. We develop a theoretical context for the sensitizing concept of ambivalence as a motivator of action (negotiating relationships). In contrast, solidarity is a set of operational definitions that reduces to six variables that may or may not vary together (in other words, may covary or operate independently).
Theoretical Concepts Versus Measures
The longitudinal, multigenerational data produced by the research projects at the University of Southern California, Los Angeles, are among the best in the world. Yet, the initial goal of the solidarity framework to provide "a means of registering" the effects of widespread societal changes on "the family" (Bengtson et al., 2002, p. 572) helps explain a focus on measuring effects or outcomes that continues today. Consequently, the key contribution of work in the solidarity school has been to help capture the conditions of relationships at different points in time, not to provide an understanding of how these conditions came to be through a theoretical conceptualization of solidarity. We can as readily employ and interpret these measures by calling them what they are (how much family members see each other, help each other, agree with one another, care about each other, and how far apart they live) as by calling them some form of solidarity (associational, functional, consensual, effectual, normative, structural). The compatibility of ambivalence and solidarity rests in the fact that the various dimensions of the solidarity classification scheme are useful for assessing current states or (transient) outcomes of negotiating ambivalence.
The elusiveness of solidarity as a concept and Bengtson and his colleagues' confusion about ambivalence are indicated in their assertion about how the solidarity variables can be used to measure ambivalence: "a disjuncture between any of the six dimensions of solidarity could be indicafive of ambivalence" (2002, p. 571). So, too, could complete accord between and among dimensions (e.g., a mother and daughter who love each other dearly and see each other very often, in part so that the daughter can deliver extensive aid). These measures enhance understanding only when they are related to concepts with theoretical underpinnings; they cannot speak for themselves, even through sophisticated quantitative analysis.
Dialectic Versus Opposites
Claims of solidarity's dialectical properties and concern with structural conflict are simply not supported in early or more recent formulations. However, one reason for Bengtson and his coauthors' belief in the possibility of incorporating ambivalence within the solidarity model may be their misconstrued conception of dialectic. As a counter-claim to our view that the solidarity perspective cannot address the dialectic of love and hate, Bengtson et al. state that "each of the multiple dimensions of solidarity is distinct (orthogonal), and each represents a dialectic: intimacy and distance ... agreement and dissent ... dependence and autonomy . . . integration and isolation . . . opportunities and barriers ... familism and individualism ... (2002, pp. 570-571). Creating antonyms to describe the ends of each continuum (e.g., rather than saying more or less geographic proximity, one says opportunity vs. barrier) completely misses the coexistence of contradictory conditions as core to understanding family relationships. How can one describe living near to someone and living far from someone as a dialectic? One can only be in one place at a time. Clearly, we have a different understanding of dialectic. Like most sociologists, we believe that dialectic emphasizes contradictions and the resulting structural strains that often provoke change (see Pfohl, 1994). Complementary treatments of dialectic have been applied to cognitive functioning by psychologists (see Kenyon, 1988) and to critical gerontology by humanists (Moody, 1988).
The general appeal of ambivalence as a concept is levied as an indictment. Yet, surely, it is solidarity's appeal that rests on being all things to all people. Bengtson and his colleagues argue as a strength the fact that "the dimensions of solidarity represent the core social elements of family life. These dimensions can be organized in an almost infinite number of arrangements that can suit many theoretical orientations, including functionalist, conflict, and critical theories" (2002, p. 572). In our view, this is precisely its weakness. The equal utility of solidarity's dimensions to all perspectives reveals the concept of solidarity as a theoretical charlatan.
Thus, asking whether the centrality of ambivalence to family life means that the solidarity paradigm is facing a legitimacy crisis (Bengtson et al., 2002, p. 573) misses the point that, despite 30 years of interesting work using solidarity measures, there simply never has been a clear theoretical basis for pursuing solidarity as a concept or paradigm. Even after this exchange one is still left asking, What is solidarity? At the same time, the repeated interpretation of the solidarity model as implicitly functionalist and normative by a variety of writers demonstrates the impossibility of avoiding theoretical positions and debates. Our conceptualization of ambivalence specifies underlying features of social structure and social relationships, providing a theoretical framework for interpreting observations about family ties. Although the solidarity measures may be useful for analyzing family relationships conceptualized as ambivalent, the concept of ambivalence cannot be subsumed within these measures.
The responses to our article raise fundamental issues about science, theory, and the focus of sociology-sometimes explicitly, often implicitly. References to hypothesis testing (Curran, 2002; Uscher, 2002), quantitative versus qualitative data (Bengtson et al., 2002), operationalization (Bengtson et al., Uscher), and empirical measurement and validation (Bengtson et al., Curran, Luscher), reflect assumptions of traditional scientific method in sociology that we do not share.
Interpretive Approach Versus Positivism
Like many sociologists today, we favor an interpretive perspective. To us, approaching research by asking questions rather than stating hypotheses is more appealing. Why? Because hypotheses bring with them the weight of old assumptions of science that tend to view the researcher as expert, to frame the enquiry before it has begun, to specify variables rather than ideas, and to predetermine the relationships that matter. Asking research questions means we are more open to learning from the subject (respondent) and to discovery in the field, enhancing the possible emergence of unanticipated ideas and the relationships between them. Such an approach can be applied to either quantitative or qualitative data. Thus, we believe that the real rift alluded to by Bengtson and colleagues in their tangential simplification of quantitative versus qualitative approaches is that between the positivism of old-school scientific method and the interpretive approach. Although typically aligned with qualitative method, an interpretive perspective need not be restricted to it.
Diversity: Inclusion Versus Exclusion
The epistemological difference between our views is reflected in how Bengtson and his colleagues treat diversity. Their claim to value and to examine diversity by virtue of constructing configurations of families based on solidarity measures is problematic (2002, p. 571). There is no doubt that one can build many family configurations based on how often people see each other, help each other, and so on, but how does this address diversity based on gender, class, age, ethnicity, and race? This avoidance of diversity is compounded in the subsequent description of "the natural world of families" as coherent and of less common forms as less coherent, less stable, and inherently short-lived (p. 572). How does this respect and explore diversity? Where do lone-parent families, gay or lesbian couples, long-distance commuting partners, or cohabiting partners fit into this picture as other than inherently inferior family forms? As Luscher puts it well, "referring to an understanding of a given social institution, especially the family, as a given natural phenomenon is often adopted in order to support normative judgments regarding its form or structure" (2002, p. 587).
Diversity is also slighted by the restrictive view of quantitative analysis held by Bengtson and his coauthors when they assert, "Marginal cases that do not fit the more common patterns are considered `deviant,' and like any outliers are prime subjects for in-depth qualitative study" (2002, p. 572). How does this respect real-life diversity? Why must outliers be studied differently? Why are qualitative methods relevant only to deviant cases? Why assume that processes of family life are somehow different for the "marginal" cases than for the "common patterns"? Gaining insight into the variations of social life can take us a long way toward explaining commonalities.
In our conceptual applications of ambivalence to the family ties of gay and lesbian adults (Connidis, 2001a) and of divorced individuals (Connidis & McMullin, 2001), we find that examining diversity helps us to learn more about all family relationships.
Our approach to developing the concept of ambivalence reflects our general view that we cannot measure something before we know what that something is. Even the proponents of traditional scientific method view construct validity as an essential feature of good empirical measurement. Yet, the focus in research has tended toward reliability, often at the expense of validity. We suggest that there are two key reasons for this: on one hand, the relative ease of assessing reliability quantitatively and the compatibility of this focus with empiricism, on the other, the relative difficulty of assessing construct validity because this requires a clear concept against which to assess the measure.
Conceptual Development Before Empirical Measurement
We have taken as our focus the conceptual development of sociological ambivalence. Frankly, we do not see the value in rushing out to measure it as a solution to clarifying the concept. As we have noted, one of the fundamental flaws of solidarity is its underdevelopment as a concept. Like Curran (2002, p. 580), we are concerned about the prospects of "fundamental blind spots in research design" when "inadequate theoretical attention" precedes empirical evaluation and have, therefore, focused our efforts on conceptual development.
Conflict in Structured Social Relations
What does it mean to engage in conceptual development? To us, conceptual development requires specifying the theoretical foundations of a concept. Thus, we devoted attention to discussing the theoretical ideas from the critical and symbolic interactionist perspectives that inform our sociological conception of ambivalence. One of these ideas is that conflict is a feature of social life. This is not to say, as both Luscher (2002) and Bengtson et al. (2002) suggest, that "conflict is an inevitable feature of interpersonal relationships" (Bengtson et al., p. 569) or that ambivalence is "equated with conflict" (Luscher, p. 587). Rather, our view is that conflict is an inevitable feature of the structured social relations of class, gender, age, race, and ethnicity, not of interpersonal relations or ambivalence. Indeed, if we had argued that conflict is an inevitable feature of interpersonal relations, then we could not have proposed that solidarity is one possible outcome of efforts to resolve ambivalence (see p. 562).
Theory and Ideology
Loscher (2002, p. 587) goes one step further, arguing that, because we acknowledge our guiding theoretical assumptions and state that "society is more accurately characterized as based on conflicting interests than on consensus," we are taking an ideological rather than an analytical position. Of course, the logic of this argument is problematic because it assumes that "good" theory must not be ideological; that it both must and can be value free. We believe that societies are themselves built upon ideological views and that our theories must expose the ideas that are the basis for structured social relations. In order to make claims about how social actors relate to one another, we must specify a position on how the social world works. Consequently, we find no solace in claims that a concept is equally compatible with all theoretical positions (see earlier discussion on solidarity). Critical sociology and symbolic interactionism have both moved theoretical thinking forward by making us examine what we take for granted-from our daily interactions with one another to the patterned relationships of social structure. We do not believe that the "truth" about the ideological bases of social arrangements, including those related to family, will be revealed through empiricism.
Structural Versus Role Conflict
Bengtson et al. (2002) criticize our conception of structured ambivalence for not differing much from symbolic interactionist approaches to role theory. However, their misunderstanding of our argument is evident in their parenthetical equation of ambivalence with agency (p. 570). But it is our treatment of ambivalence as a feature of structured social relations that is key. Our consequent emphasis on relative resources and power means that our conception of ambivalence cannot be reconciled with earlier versions of role theory.
Early formulations of symbolic interactionism paid limited attention to the significance of social structure and power and tended to view conflict as embedded in roles and, hence, at the individual level (mother and paid worker). We view conflict as the outcome of structured social relations made manifest in social domains at the institutional level (such as family and paid work; see Marshall, Matthews, & Rosenthal, 1993). Thus, in our formulation, the mother-worker role set is not inherently paradoxical and contradictory nor is it necessarily characterized by conflict. Rather, such role sets are experienced as ambivalent only when structured social relations are embedded in conflicting social domains. To say that our conception of ambivalence "sensitize[s] us to the important issue of contextualizing family relationships within multiple social statuses and roles" misses the point. We contextualize a concept that emphasizes the coexistence of harmony and conflict in a theoretical model that addresses social structure and individual agency.
Ambivalence and Social Change
In sum, combining central tenets from the critical and interactionist perspectives has been the basis for exposing a social world that is far more complex, conflicted, and dynamic than is typical of the view from other perspectives. Perhaps even more important, it is a view that also differs from our experience as relatively privileged social actors. The concept of ambivalence applies this awareness to our understanding of relationships among family members.
Our views on theory relate to two other points: the application of theoretical concepts to social change and policy and the view of ambivalence as negative. Bengtson and his colleagues (2002) congratulate us for highlighting the relevance of structured ambivalence to social policy and change (p. 569). The key and perhaps only way that our research can make such a contribution is by specifying concepts in relation to our understanding about the larger social world. If family ties are studied within a theoretical view of society as working in the interests of all its members, then attempts to apply research results will inevitably target families rather than social structural arrangements. This leads to a focus on individuals and their private troubles rather than on society and public issues (Mills, 1959; see Connidis, 2001b on family). The fact that Bengtson and his colleagues view solidarity as a conceptual chameleon, equally at home with all theoretical approaches, does not clear them of this concern. Failure to stake a theoretical claim simply means that the application of their findings on solidarity is equally amenable to different perspectives, including those that favor maintaining the status quo.
Ambivalence and "Natural" Order
Both Luscher (2002) and Bengtson et al. (2002) suggest that we treat ambivalence as an inherently negative state. Is ambivalence negative? And is it causally prior to solidarity? In some ways, both of these questions converge on the extent to which any condition is seen as a natural state for social actors. A unique feature of being human is that we can alter social conditions that are mistakenly considered natural and essential. We bristle at such treatments of families and relationships because arrangements that are viewed as natural or essential are more likely to be evaluated positively and less likely to be the object of change.
Our initial response to Luscher's implication (2002, p. 587) that we treat ambivalence as negative was to respond, "No we don't. Ambivalence just is. It is a given; it is a condition of relationships." But, then we paused when we realized the implicit assumption of essentialism in such a statement. Our sociological conception of ambivalence embeds the ambivalence of relationships in social structural arrangements. To the extent that these are negative (e.g., oppressive, especially to some groups), then it is fair to say that some of the ambivalence that results from such arrangements is also negative. It is, of course, this assumption that leads to proposals about how to make a better world for those who experience more ambivalence because of structurally created conditions. Indeed, a key contribution of a sociological understanding of ambivalence is to connect individual experience to relationships, to social institutions, and to social structure and, hence, to recommendations for social change. So, we must conclude that there is a component of ambivalence that is negative.
Beyond this, however, we also argue that there is an inevitability about ambivalence as a feature of social relationships. Our desires and objectives as social actors are not always compatible but must still coexist with one another, with those of other social actors, and with the patterned relationships that our interactions have created over time. This is a feature of social life that is neither positive nor negative and is why ambivalence cannot be reduced to negative feelings as Bengtson et al. suggest (2002, p. 574). It also points to a causal order in which ambivalence comes first, as a condition that prompts social action, the outcome of which at any given point in time may be relationships characterized primarily by ambivalence, conflict, or solidarity (meaning relative harmony). Here, we are in agreement with Lischer (2002, p. 591).
The depiction of the bonds between parent and child or of the courting couple as evidence that solidarity comes first (Bengtson et al., 2002, p. 574) instead describes an idealized view of both relationships (Gillis, 1996). Wanting to have an intense and committed relationship with another coexists with wanting to do so at least partly on one's own terms, a particularly ambivalent situation. This makes the first dispute over a bill or dirty diaper unlikely as the first confrontation with having to negotiate ambivalence, although optimism about the chances of a positive outcome may be higher at earlier stages of many relationships.
THE STUFF OF SOCIOLOGY
Finally, what is the stuff of sociology? What is it that we as sociologists are uniquely equipped to add to our understanding of family ties? Fundamentally, it is the linking of individuals as social actors to their relationships with family members, to various institutions (including the family), and to social structure. We have proposed a sociological conception of ambivalence as a vehicle for making these connections.
Multidisciplinary, Multilevel Analyses of Ambivalence
Differences in disciplinary perspective and progress are exemplified in the development of ambivalence as a concept. As Luscher (2002) illustrates, ambivalence has a relatively long history in psychiatry, psychoanalysis, psychotherapy, and psychology. Curran's (2002) suggestion to incorporate motivation and accountability highlights a unique contribution made by an approach informed by economics. Both point to the value of a multidisciplinary approach to family relationships.
Consider, for example, the issue of motivation raised by Curran (2002). At the sociological level, the motivation to enter, maintain, and dissolve relationships (see Curran, p. 580) relates to questions about what leads individuals to engage in the relatively patterned behavior of family relations: getting married or entering long-term partnerships; having children; and maintaining (or dissolving) ties with partners, siblings, parents, and children. Institutional arrangements related to family ties are based on socially structured relations that encourage particular commitments by particular groups and distribute resources differentially in all institutional arrangements, including families. For many, contradictory conditions and limited agency combine to create ambivalent family relations.
However, Curran (2002) seems also to be referring to a more personal or individual level of motivation that concerns why someone chooses to start, maintain or dissolve a relationship with a particular person (caring for a particular family member or marrying a particular individual) and invokes individual motivations at the psychological level (e.g., love for that particular individual). We believe that questions of this type are more adequately addressed by psychologists than sociologists. Thus, the ideas of motivation and dialectic (see earlier discussion) reinforce the need for multiple levels of analysis involving different disciplines. Directing multidisciplinary efforts at developing the concept of ambivalence will assist this effort by informing our understanding of family ties at the levels of the individual, the relationship, institutional arrangements, and structured social relations. Unlike Luscher (2002, p. 585), and as Curran observes (p. 579), our sociological conception of ambivalence applies to relationships as well as to social structure and this constitutes one of the contributions of our perspective.
At the moment there are variations among disciplines in the level of conceptual development of ambivalence. The longer history and firmer foundation of the psychological conception makes it better equipped for developing operational definitions and empirical measures. Given significant disciplinary differences in paradigmatic assumptions (see Dannefer, 1984), we cannot simply transfer empirical measures from psychology to address sociological questions. Instead, we must first develop sound conceptualizations grounded in our discipline so that, in the longer run, we can contribute the unique view of sociology to our understanding of family ties. From a sociological perspective, developing ambivalence as a concept that links social structure and individual agency is crucial.
Social Structure and Agency
A multilevel analysis of family relations is facilitated by a sociological approach that links structured social relations to core social institutions and to the assumed obligations and privileges of individuals in their daily interactions. Unlike Luscher (2002), who conflates institutions and social structure, we specify social structure as sets of social relations based on class, age, gender, race, and ethnicity that produce lasting patterns of inequality in society. In turn, these structured social relations shape the arrangements embedded in institutions, including families. In the case of family ties, the differentiated distribution of resources and responsibilities among family members reflects the norms, rules, and beliefs that are built into structured social relations (see Giddens, 1984).
Curran's suggestion to include accountability in our model (2002, p. 580) parallels our discussion of legitimate excuses (Finch, 1989) as a concept that captures the influence of both rules and resources in our negotiations of family ties. Regardless of whether family relationships are chosen, once we are involved, we are embedded in institutional arrangements that include a view of socially structured obligations among family members. We are "implicated" (Sprey, 1991). Consequently, negotiations to avoid responsibility focus on formulating excuses that establish what we can (and cannot) do, not what we will (and will not) do (Finch). The grounds for legitimate excuses vary by gender and reflect greater accountability among women to provide caring work (see, e.g., Aronson, 1992).
Curran's discussion of interior versus exterior sanctions (2002, p. 582) highlights further variability in structural constraints based on the number of social locations in which one is embedded and the relative degree of entanglement in a particular social location (domain). Curran's equation of managing ambivalence with accruing accountability benefits and avoiding accountability costs (p. 582) is a complementary extension of our focus on ambivalence in relationships. A larger issue it minimizes, however, is how to alter the distribution of social domains or locations to which various groups are accountable, as opposed to how to minimize the ambivalence experienced when one is embedded in multiple social domains (the employed mother who is seen as the likely catalyst of change in Curran's example). One threat of focusing too much attention on the worthy effort of recognizing care work as a public good is to miss the point that it is a public good that can be performed by a broader spectrum of citizens and state-supported initiatives than is currently the case.
Curran (2002) addresses agency by raising the question of motivation and indicates the need for a clearer definition of agency. In our article, we defined agency simply as the attempt of individuals to exert control over their lives (p. 563). To clarify, we make the ontological assumption along with many sociologists that individuals act with agency, that is, they have the capacity to act with intention (Giddens, 1984). This definition should clarify that agency cannot be equated with 11 whether to provide or receive care," as Curran (p. 579) parenthetically suggests.
In relation to a family member's need for support, exercising agency may mean making a decision to care, or not to care, or not to make a decision, or to ignore the situation, or to wait and hope that someone else will provide care, or to redefine the situation as one where support is not "really" needed, and so on. Thus, to care or not to care are only two of many possible decisions. Because Curran's discussion of motives for agency is, in fact, a discussion of motives for providing care, we are drawn back to accountability and ambivalence. Thus, the answer to Curran's closing question (p. 583), "Why do some individuals continue to provide care work?" lies not "with explaining the motives for individual agency" but with exploring ambivalence and accountability.
The current use of ambivalence in studies throughout the world (the United States, Canada, Europe, the United Kingdom and Scandinavia) and the debate and interest that this concept has generated at conferences (including the Gerontological Society of America and the American Sociological Association), attest to a clear readiness to explore new avenues to understanding family relations. In proposing a theoretical context for ambivalence as a sociological concept, we believe that we have enhanced its potential utility for exploring how family relationships are negotiated and how they connect with individual lives, other institutional arrangements, and social structure.
Does our conception of ambivalence represent a paradigm shift? No. It is one more step in a broader and ongoing paradigm shift away from an idealization of social structure as serving everyone's interests toward a critical view of society. This general shift moves us from the quantitative empiricism of traditional scientific methods and notions of objective "truth" toward an interpretive approach that appreciates a subjective view and approaches quantitative and qualitative data critically; from the simplification required by empirical measurement toward the complexity allowed by conceptual thinking; from unilevel to multilevel approaches; from middle-class ideals to not merely noting difference but exploring diversity; and from dualisms to dialectics.
Does ambivalence have all of the answers? No. But, in questioning dominant ways of thinking about family relations, it does lead us to ask better questions and to consider how strong theoretical traditions in sociology can be applied to our thinking and research on family relations. In the long run, ambivalence may become an unnecessary concept. We will be content if, in the interim, the net effect of exchanges like this one is to move us beyond complacently accepting the periodic lament about our limited theorizing while applying the same measures in our research on family relations. Our hope in connecting family life to individual agency and social structure through the concept of ambivalence is that future explorations will address the action and processes of family life rather than its steady states. We agree with Curran's conclusion (2002, p. 584) that our "theory requires several more steps of elaboration before it can be empirically tested." If our conception of ambivalence has helped to encourage the application of critical sociological ideas about social structure, social actors, and conflict to family ties, then we have succeeded.
Aronson, J. (1992). Women's sense of responsibility for the care of old people: "But who else is going to do it?" Gender and Society, 6, 8-29.
Bengtson, V., Giarrusso, R., Mabry, J. B., & Silverstein, M. (2002). Solidarity, Conflict, and Ambivalence: Complementary or Competing Perspectives On Inter
generational Relationships? Journal of Marriage and Family, 64, 568-576.
Connidis, I. A. (2001 a). Family ties and aging. Thousand Oaks, CA: Sage.
Connidis, I. A. (2001b, November). The intergenerational ties of gay and lesbian adults and step kin: A conceptual discussion. Paper presented at the 5411, meeting of the Gerontological Society of America, Chicago, IL.
Connidis, 1. A., & McMullin, J. A. (2001, July). Negotiating family ties over three generations: The impact of divorce. Paper presented at the International Association on Gerontology 171" World Congress. Vancouver, B.C.
Curran, S. (2002). Agency, Accountability, and Embedded Relations: What's Love Got to Do With It? Journal of Marriage and Family, 64, 577-584.
Dannefer, D. (1984). Adult development and social theory: A paradigmatic reappraisal. American Sociological Review, 49, 100-116.
Finch, J. (1989). Family obligations and social change. Cambridge, UK: Polity Press.
Giddens, A. (1984). The constitution of society: Outline of a theory of structuration. Cambridge, UK: Polity Press.
Gillis, J. R. (1996). A world of their own making: Myth, ritual and the quest for family values. Cambridge, MA: Harvard University Press.
Kenyon, G. M. (1988). Basic assumptions in theories of human aging. In J. E. Birren & V. L. Bengtson (Eds.), Emergent theories of aging (pp. 3-18). New York: Springer.
Luscher, K. (2002). Intergenerational Ambivalence: Further Steps in Theory and Research. Journal of Marriage and Family 64, 585-593.
Luscher, K., & Pillemer, K. (1998). Intergenerational ambivalence: A new approach to the study of parentchild relations in later life. Journal of Marriage and the Family, 60, 413-425.
Marshall, V. W., Matthews, S. H., & Rosenthal, C. J. (1993). Elusiveness of family life: A challenge for the sociology of aging. In G. L. Maddox, & M. P Lawton (Eds.), Annual review of gerontology and geriatrics, 13 (pp. 39-72). New York: Springer.
Mills, C. W. (1959). The sociological imagination. New York: Oxford University Press.
Moody, H. R. (1988). Toward a critical gerontology: The contribution of the humanities to theories of aging. In J. E. Birren & V. L. Bengtson (Eds.), Emergent theories of aging (pp. 19-40). New York: S rineer.
Pfohl, S. (1994). Images of deviance and social control: A sociological history. New York: McGraw-Hill.
Sprey, J. (1991). Studying adult children and their parents. Marriage and Family Review, 16, 221-235.
INGRID ARNET CONNIDIS AND JULIE ANN MCMULLIN
University of Western Ontario
Department of Sociology, University of Western Ontario, London, Ontario, Canada N6A 5C2 (firstname.lastname@example.org).