Kennedy School of Government, Harvard University, 79 JFK Street, Taubman 370, Cambridge MA 02138, USA. E-mail: robert_putnam{at}harvard.edu
That social capital has been, at least, faddish in international policy and social scientific circles during the last decade is beyond doubt. After several decades of intellectual and political hegemony on the part of an individualistic philosophy that claimed that there is no such thing as society,1 advocates of the social capital perspective argued for renewed emphasis on the importance of social networks and norms in many spheres of our livesfrom job placement to democratic governance to health. To be sure, in a number of specific sub-disciplines, such as the sociology of labour markets or criminology or social epidemiology, researchers were already well aware of the importance of social networks.* The new social capital agenda capitalized on this prior work, and by framing the underlying issues in terms of a broader intellectual agenda, gave added prominence to it. Researchers in many different fields (not just those in which social networks were already on the agenda) have contributed to an exponential growth in research on social capital. (Figure 1, drawn from a forthcoming major new treatise on social capital by David Halpern.) At the same time political leaders (mostly from the centre-left, but some from the centre-right) were attracted to the new concept and to its emphasis on community, social networks, and civic renewal.
|
On both the theoretical and empirical fronts the last decade has witnessed a lively and fundamentally productive debate about social capital. Critics have expressed doubt about the methodological soundness of some empirical claims about the supposedly powerful effects of social capital, and researchers have responded with innovative and more powerful research designs. In fields as diverse as development economics, criminology, immigration, health, and even evolutionary biology, this debate is well underway, following conventional canons of scientific inquiry, precisely as one would hope for any proposed scientific innovation. Probably the final verdict will be mixedwith social capital turning out to be very important in some areas and less so in othersbut only the gradual dialectic of careful research will tell.
On the theoretical side, progress has been made in the last decade to clarify and sharpen certain distinctions that were insufficiently clear in the earliest work on social capital. Some initial definitions of the concept of social capital were extremely broad, encompassing not merely social networks and norms of reciprocity, but also many other aspects of social life, such as culture, institutions (such as the rule of law), and sometimes even the supposed beneficial consequences of the networks, norms, and institutions. For example, James S Coleman, justifiably regarded as the intellectual progenitor of the field, defined social capital functionally as features of social organization, such as trust, norms, and networks, that facilitate action.2 However, more than a decade ago it was widely recognized, by both critics and advocates of the concept, that to include consequences (such as improved social efficiency) within the definition of social capital would render claims about its causal importance tautological. It is now generally understood that like physical and human capital, social capital can be used for pro-social or anti-social purposes.
But even if those alleged consequences are excluded from the definition, a very broad definition always risks intellectual sloppiness, so in recent years many figures in the field have converged toward a lean and mean definition of social capital as social networks and norms of reciprocity.** If social capital is to be useful as a scientific (and practical) concept, advocates need to abjure slippery definitions that encompass features of society that do not rest on social networks, which everyone agrees are at the core of the concept.
Even in this restricted sense, however, social capital is, fundamentally, not a single thing (of which one might have a lot or a little), but a conceptual category that includes many different types of networks and norms that might be used for many different purposes. Like physical and human capital, social capital comes in many forms, not all fungible. A dentist's drill and an oil-rigger's drill are both bits of physical capital, but they are not interchangeable. Development economists often speak generically of human capital as a source of productivity, even though a degree in cosmetology and a degree in cosmology are not strictly interchangeable. One of the most important frontiers in social capital theory, in fact, is precisely the development of a comprehensive, rigorous, and empirically relevant typology of different forms of social capital (or social networks). How do social networks differ from one another in ways that are relevant to their consequences? What types of networks are most important for job placement, or for neighbourhood safety, or for mental health, or for pluralist democracy?
Important advances have been made in this arena. Well before the recent interest in social capital, for example, Mark Granovetter showed that strong ties (that is, networks with a high degree of closure, so that people in the network mostly know one another) are less effective for job-seekers than weak ties (that is, less dense, more open-ended networks).3 Others have distinguished between episodic, single-stranded ties (loose connections) and more durable, multi-stranded (or multiplex) ties, and it seems likely that the latter are more valuable for social support and health than the former.4 Still others have argued that vertical networks (linking patrons and clients, for example) are less conducive to democracy than horizontal networks.5
Probably the most widely discussed typological distinction has been between bridging social capital (that is, inclusive social networks that cut across various lines of social cleavage, linking people of different races, ages, classes, and so on) and bonding social capital (that is, exclusive social networks that are bounded within a given social category). It is plausible to conjecture, for example, using terms introduced by Xavier de Souza Briggs, that while bonding social capital is crucial for getting by, bridging social capital is especially important for getting ahead.6 In deeply cleaved societies, like Northern Ireland or Bosnia, the main social capital deficit involves not insufficient bonding networks, but insufficient bridging networks.7 The crucial test of this distinction, however, (like the other putative typologies I have mentioned) will only come when these conjectures have been tested with rigorous empirical research: Can bonding and bridging networks be reliably distinguished, and do they have measurably distinctive consequences?
The important new paper by Simon Szreter and Michael Woolcock, Health by association: social capital, social theory and the political economy of public health,8 makes several valuable contributions to this theoretical frontier in the field of social capital. I wish here to comment on several themes in this paper.
First, Szreter and Woolcock suggest that to bridging and bonding should be added a third category of linking social capital. More precisely, they suggest that linking networks are an especially important sub-type of bridging networks, namely, networks that connect actors of different degrees of institutional power. I am myself agnostic about whether this distinction will prove to be valid, reliable, and substantively important. Certainly, networks that link the powerless to the powerful might in principle merit singling out for special attention. In principle, norms of respect and networks of trusting relations between people who are interacting across explicit, formal or institutionalized power or authority gradients sounds like a promising, perhaps even essential, ingredient in any recipe for the sort of civil society that could underpin policies and practices to foster social justice and public health.
As Szreter and Woolcock concede, however, not all vertical networks have pro-social consequences, especially if the disparities in power within the network are substantial and mainly exercised by the powerful to control the powerless, as is true in most patronclient ties. In Italy, for example, the frequency of contacts between citizens and local officials is strongly negatively correlated with social trust and political responsiveness and economic growth, since in that setting such personalistic linking networks are symptomatic of patronclient exploitation (ref. 9, pp. 99101). So presumably we would need to distinguish between what we might call responsive linking and unresponsive or exploitative linking. Working out how to make that distinction in a non-tautological waynot an easy taskmust presumably be at the very top of the agenda that Szreter and Woolcock are proposing.
Indeed, in its current incarnation, I am not convinced that Szreter and Woolcock have entirely avoided the same sort of tautological trap from which social capital theory has only recently escaped. As their essay moves from theory to empirics in the concluding section on 19th century British public health, the primary evidence offered for the claim that British public health benefited at various times from linking social capital is precisely the responsiveness of the British state to public needs. But if linking social capital is defined operationally as good government, then the claim that linking social capital conduces to good government is not an interesting empirical discovery, but a mere tautology.
The relatively generous [pre-1834] Poor Law is said to be the evidence for pre-Victorian linking social capital, but in fact the Poor Law represented state action, which is precisely what is supposed to be explained by linking social capital, so that (at least in that case) the authors' substantive claim represents not a novel discovery, but a tautology.8 Paternalistic attitudes and the founding of subscription hospitals, used here as operational indicators of linking social capital, seem quite distinct conceptually from social networks, so that it seems misleading to count them as social capital at all.8 Similarly, Joseph Chamberlain appears to have been a remarkable political leader who presided over epochal reforms in municipal services and public health, but his actions cannot themselves be the evidence for the existence of linking social capital that might in turn explain the health outcomes. The authors' evidence for effective linking social capital in the Swedish case appears to be a record of adequate material assistance to the poor, thus again making the connection between linking social capital and responsive public policy not causal, but tautological.8 Government responsiveness cannot be both the explicans and the explicandum.
As a distinguished sociological methodologist patiently instructed me many years ago, social scientists should seek brittle hypotheses that visibly fracture when they encounter discomfiting evidence. We need to be able to establish in advance what facts about late Victorian Birmingham would (if they proved to be facts) disconfirm the thesis that linking social capital (rather than something else, like efficient technocratic government or enlightened political leadership) was a key to improved public health.
This is not an insoluble dilemma, assuming that researchers can develop reliable and valid measures of linking social capital, just as they are now developing such measures for bridging and bonding social capital. In fact, in recent work (still preliminary) in the US my students and I seem to have uncovered suggestive evidence that interpersonal ties between grassroots activists and municipal leaders may have an independent effect on community revitalization. So my caution on this first issue is not rooted in theoretical first principles, but in the ever-present need for empirical rigor. Without that rigor, the social capital research programme would justifiably be dismissed.
A larger issue addressed by Szreter and Woolcock concerns the connections among three big variables that seem to be implicated in much current discussion about the sociological origins of healthequality, social capital, and the state. Let me begin with the latter two factors. Unfortunately, Szreter and Woolcock endorse the common canard that students of social capital ignore the role of the state in public and social life. They say that one danger in the social capital literature has been an over-emphasis on voluntary associations, alone, as the key to healthy social capital and a tendency to cast the state only in the negative terms of an impersonal and monolithic big brother figure. Since they offer no citation at all for that strong criticism, I can only respond autobiographicallyin my own work, at least, nothing could be further from the truth.
It is, to be sure, important to distinguish between social capital and the state analytically, not because the state is unimportant in social life, but because it is very important. One crucial issue is how the state and social capital are related, and that issue cannot even be posed if the two are conceptually conflated. At this gross-level analysis it certainly must be the case that the two are reciprocally related. State actions powerfully affect patterns of social capital, and patterns of social capital powerfully affect the state. My first empirical work in this field, which Szreter and Woolcock generously mention, explored precisely how state actions (in particular, the policy performance of Italian regional governments) were conditioned by social capital.9 Conversely, my first essay on social capital theory argued explicitly (and with emphasis in the original) that:
Social capital is not a substitute for effective public policy, but rather a prerequisite for it and, in part, a consequence of it. Social capital, as our Italian study suggests, works through and with states and markets, not in place of them. The social capital approach is neither an argument for cultural determinism nor an excuse to blame the victim.... Wise policy can encourage social capital formation, and social capital itself enhances the effectiveness of government action.... [then speaking to an American public audience, I concluded]: The social capital approach promises to uncover new ways of combining private social infrastructure with public policies that work, and of using wise public policies to revitalize America's stocks of social capital.
In short, I agree with Szreter and Woolcock that the state (or public policy) must be embraced in any understanding of how social capital influences well-being, including health, while being slightly bemused by the claim that this view is novel.
Finally, Szreter and Woolcock usefully and appropriately describe an even larger theoretical framework that encompasses the role of social inequality in health, as elaborated in the important debates among Richard Wilkinson, Michael Marmot, John Lynch, and others. In order to explain my own perspective on this broader set of questions, it is convenient to lay out a crude causal diagram. (In the interests of parsimony, I ignore here such factors as genes, medicine, nutrition, and education, all of which presumably also affect health.) Figure 2 is intended, not to answer questions, but to allow us to frame them clearly.
|
Arrow (3) is implicated in the work of Wilkinson and others. As I have argued elsewhere, the general positive correlation between measures of equality (for example, income distribution) and measures of social capital is quite robust. There is an extremely interesting argument to be had, however, about which way the causal arrow operates here. Wilkinson and many other observers who have noticed this correlation seem to have assumed without much investigation that the arrow must flow from the hard reality of income distribution to the softer realities of social capital (or social solidarity). That is certainly possible, but we should not dismiss too quickly the logical possibility that the arrow runs in the other direction, that is, that economic equality or inequality in any society is to some extent a consequence of social capital or social solidarity. The issue is (to take a single example) whether levels of social capital are so high in Sweden (as they are, by almost any measure) because the gap between rich and poor is relatively narrow, or whether instead Swedes have chosen policies (public and private) that encourage equality precisely because they have high levels of social connectedness and mutual solidarity. Perhaps under the influence of economists, we tend to assume that economic facts (like income distribution) are exogenous, that is, an uncaused first cause, but recent reflection by economists like Paul Krugman and Tony Atkinson have raised the issue of whether social norms of solidarity (or, by contrast, excessive material individualism) might have a powerful independent effect both on wage determination and on redistributive public policies. Resolving the two-headed Arrow (3) in any rigorous empirical way will be dauntingly difficult, but I mean here to argue that we should not resolve it simply by a priori assumption.
Arrow (4) represents the direct effects of inequality on health. The exact nature and strength of this linkage is, of course, a matter of controversy between Wilkinson, Lynch, and others. On balance, like Szreter and Woolcock, I am inclined to believe that the adverse effects of inequality on health flow through multiple channels, both material and psychophysiological, but I have little to contribute substantively to that debate.
Finally, Arrow (5) refers, generically, to all the ways in which social capital (by which, it will be recalled, I mean social networks and the associated norms of reciprocity) may affect health. As I have noted earlier, and as is summarized briefly in Figure 3, some of these pathways have long been explored by social epidemiologists under the rubric of social support. Other possible mechanisms include the potential effects of social networks on health-related communication (as in some accounts of AIDS), of social networks on social identity and risky behaviour (as in some accounts of adolescent smoking), of social networks on access to material resources (as in many accounts of health problems in areas of rural poverty), of social capital on the resolution of dilemmas of collective action (as in many accounts of environmental threats to health), and of social connectedness on physiological processes (as in accounts of how living in a dog-eat-dog society could raise basic stress levels, as well as some animal studies of the physiological effects of isolation). I see no reason to assume that this list of possible mechanisms through which social networks influence health is exhaustive. On the contrary, one of the reasons that I find this field so exciting (as an observer, to be sure, rather than as an active contributor) is precisely the sense that the current debates embodied ever so schematically in Figures 2 and 3 presage important advances in our understanding of how social connections affect health and thus promise important opportunities for improving public health, especially among the more marginalized of our fellow citizens. I thus concur entirely with the concluding sentence of Szreter and Woolcock: the social capital lens is important for public health not because it promises a cheap or simple prescription, but because social networks and the associated norms of reciprocity appear to have multiple, powerful effects on health.
|
![]() |
Notes |
---|
** A lively debate continues among researchers about whether trust should be seen as part of the definition of social capital or as merely a close correlate and proxy. In my view the essential element is trustworthiness, that is, a readiness to act in accord with the obligations of reciprocity. Trust (that is, the perception by others that one is trustworthy) is typically a correlate of trustworthiness and thus of social capital, but unwarranted trust (that is, trust divorced from trustworthiness) is merely gullibility, which is not part of the conception of social capital. For an elaboration of this point, see: Putnam RD. Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster, 2000, pp. 13437.
So far as I can tell, credit for coining these labels belongs to Ross Gittell and Avis Vidal. In: Gittell R, Vidal A. Community Organizing: Building Social Capital as a Development Strategy. Thousand Oaks, CA: Sage, 1998, p. 8.
The prosperous community: social capital and public life. In: The American Prospect, no. 13 (Spring 1993), pp. 3542, quotation at p. 42. This passage was prefaced by a polemic against American conservatives that made my views even more explicit: During the [first] Bush Administration community self-reliancea thousand points of lighttoo often served as an ideological fig leaf for an administration that used the thinness of our public wallet as an alibi for a lack of political will. Conservatives are right to emphasize the value of intermediary associations, but they misunderstand the potential synergy between private organization and the government.
For useful recent overviews of the myriad possible pathways between social connection and health, see Berkman LF, Glass T, Social integration, social networks, social support, and health, (pp. 137173), and Kawachi I, Berkman L, Social cohesion, social capital, and health, (pp. 174190), both in: Kawachi I, Berkman L (eds). Social Epidemiology New York: Oxford University Press, 2000.
![]() |
References |
---|
![]() ![]() |
---|
2 Coleman JS. Foundations of Social Theory. Cambridge, MA: Harvard University Press, 1990, ch. 12, esp. p. 302.
3 Granovetter M. The strength of weak ties. Am J Sociol 1973;78: 136080.[CrossRef][ISI]
4 Wuthnow R. Loose Connections: Joining Together in America's Fragmented Communities. Cambridge, MA: Harvard University Press, 1998.
5 Putnam RD. Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press, 1993.
6 de Souza Briggs X. Doing democracy up-close: culture, power, and communication in community building. J Plann Educ Res 1998; 18:113.[ISI]
7 Nelson BJ, Kaboolian L, Carver KA. The Concord Handbook: How to Build Social Capital Across Communities (2003). http://concord.sppsr.ucla.edu/concord.pdf.
8 Szreter S, Woolcock M. Health by association? Social capital, social theory and the political economy of public health. Int J Epidemiol 2004;33:
9 Putnam RD, Leonardi R, Nanetti RY. Making Democracy Work. Princeton, NJ: Princeton University Press, 1994.