Shaming Encounters: Reflections on Contemporary Understandings of Social Inequality and Health

Introduction

The publication in 2009 of The Spirit Level (referred to from here on as TSL) continues to provoke debate about how social inequality shapes population health in those countries which have passed through the epidemiological transition. The authors, Wilkinson and Pickett (from here on W&P), assign a key role to income inequality in their explanation for the well-known and long-standing social gradient observed in health and seen across a wide range of social policy concerns (Lynch et al., 2004; Marmot, 2010; Wilkinson and Pickett, 2007, 2009). Those interested in this area will know that the strength of the relationship between income inequality and health remains a subject of dispute amongst epidemiologists, despite a recent review by Rowlingson which found strong support for a correlation and ‘some rigorous studies’ (2011: 5), indicating a causative relationship. The intention here is not to address this debate, but to provide a sociologically oriented critique and re-formulation of one aspect of W&P’s psychosocial explanation for the observed relationship between health and (income) inequality – that is, the role and function of invidious or shaming comparisons. Others have commented on the philosophical, theoretical and methodological questions raised by attempting to make the leap from epidemiological and survey data to theorising about mechanisms and have highlighted ‘the dangers of using such data to develop complex social explanations for health inequalities’ (Forbes and Wainwright, 2001). The bodies of work which we discuss in this article, and which we argue provide nuanced ways of understanding what might be happening in unequal societies, are often located in a very different epistemological space to positivist social epidemiology. The intention here is not to pit these bodies of knowledge against each other but to explore, looking at one particular aspect of psychosocial theorising rooted in social epidemiology (that concerning shame), how ideas drawn from culturally oriented class analyses might serve to extend contemporary understandings of the causes of health inequalities. But, like Forbes and Wainwright (2001), we fully acknowledge the enormous contribution to understandings of the causes of health inequality made by Wilkinson, Pickett and others. Without their work these critiques would not be possible.

The literature we address here allows factors such as exploitation, ‘othering’ and how people resist to be considered in the context of shame and serves to put flesh on the bones of the social epidemiology in W&P’s work. In this article, therefore, we seek to advance three proposals. First, we argue that the conceptualisation of shame in TSL can be extended by drawing on scholarship from the sociology of health and illness, facilitating a more dynamic interpretation of the social epidemiology presented in TSL. Second, both agency and resistance are largely absent from TSL and we propose two resources which people can deploy (with varying degrees of success) to make sense of their social position (and which may also be protective of health). These are destigmatisation strategies and collective imaginaries; integrating these with psychosocial theorising, we argue, provides a more nuanced theoretical framework for understanding how shaming social comparisons may operate to shape health. In this regard, our argument is consistent with those who have lamented the epistemological ‘thinness’ of understanding at the core of much contemporary social epidemiology (Forbes and Wainwright, 2001; Lamont, 2009; Popay et al., 1998; Scambler, 2007, 2012) and we echo calls for greater incorporation of theory as a response to this ‘thinness’ (Williams, 2003).

Finally, our third argument is that using shame in this sociologically informed manner casts a light on some of the specifics of how shame may operate in unequal societies and its inter-relationship with wider political and economic forces such as the growth of neoliberalism (Bambra, 2011; Coburn, 2010; De Vogli, 2011; Scambler, 2012). We begin by briefly describing W&P’s arguments, before outlining our attempt to extend their framework.

The Social Epidemiology of Inequalities in Health

The key issues at stake regarding the causes of health inequalities within social epidemiology can be swiftly stated. The inverse relationship between income inequality and health (the main thrust of TSL) applies only in those countries which have passed through the ‘epidemiological transition’ where the epidemic and largely infectious diseases of poverty cease to be the major causes of mortality, and are replaced by chronic and degenerative conditions, familiar in the developed world. As national income per capita increases, there is a rapid rise in life expectancy, but above a certain threshold, further increases in national income per capita yield little or no increases (Wilkinson, 1994). At this point, W&P (2009: 29) contend that ‘National standards of health and other important outcomes are substantially determined by the amount of inequality in a society’, and it is the extent of (income) inequality which underpins the social gradient seen in morbidity and mortality. Developed countries have a mortality pattern where the excess is amongst those of working age who are exposed to the ‘pressures of working lives’ (Hall and Taylor, 2009: 84) and it is this pattern of mortality and morbidity with its strong connection to the role of (psycho) social stressors in mediating ill health and mortality which provides the basis for the explanation for inequalities in health posited in TSL.

Within social epidemiology, there has been a fierce debate about the respective contributions of what are referred to as material, neo-material and psychosocial factors in accounting for the health gradient (Lynch et al., 2004; Scambler, 2012). ¹ Again, our intention is not to review this debate but to adopt a similar perspective to that of Hertzman and Siddiqui that, ‘each hypothesis has been presented as mutually exclusive of (and, in fact in competition with) the others … however, it is our belief that these pathways operate together in different combinations and permutations in different contexts’ (2009: 43). Psychosocial approaches can, and usually do, integrate material factors impacting on health and then explain the excess morbidity and mortality remaining after material factors are taken into account (Marmot and Wilkinson, 2006). We now turn to the psychosocial explanation and the place of shame, advanced in TSL.

Shame and Social Comparison in TSL

W&P argue in TSL that the social gradient in morbidity and mortality can be explained with reference to inequality via three key psychosocial pathways: stress in early life, lack of friends or social engagement and – the focus of this article – the invidious or shaming comparisons which aggravate the negative consequences of greater social inequality. These comparisons get inside the body and impact on health via psycho-neuro-immunological pathways, the biology of which is becoming increasingly well understood (McEwen, 2005; Sapolsky, 2005). In highlighting invidious comparisons and the shaming experiences these generate, W&P draw on the work of sociologist Thomas Scheff, for whom shame is seen as ‘The social emotion’ (1990: 79, emphasis in the original), and which for humans, as intersubjective, evaluative beings, means exposure to the potential threat of exclusion from the bonds and connections which are central to our lives. W&P argue:

Greater inequality seems to heighten people’s social evaluation anxieties by increasing the importance of social status. Instead of accepting each other as equals on the basis of our common humanity … getting the measure of each other becomes more important as status differences widen … If inequalities are bigger, so that some people seem to count for almost everything and others for practically nothing, where each one of us is placed becomes more important. (2009: 43–44)

To further illustrate their argument, they point to an increase in anxiety, narcissism and rates of depression in recent decades, arguing that these reflect ‘the extent to which we do or do not feel at ease and confident with each other’ and that this, in turn, reflects important characteristics about the social spaces people inhabit. Certain kinds of stressors, what they refer to as ‘social evaluative threats … threats to self-esteem or social status where others … negatively judge performance’ (2009: 38), are those that are most salient for health, and, where these threats occur in situations which are uncontrollable, they result in the greatest stress and the most negative health outcomes. ² The consequences of invidious comparisons are social as well as biological, with more unequal societies tending to be more aggressive, less trusting and have poorer overall health.

This then, is the context for the debate around the relationship between shame, social comparison and health in TSL. In the next section, we propose ways in which sociological critiques of the limitations of social epidemiology can be deployed to extend W&P’s formulation of shame and how shame might operate in social spaces; in part by looking at the known connection between humiliation and entrapment (shame-like phenomena) and depression and by locating contemporary shame in the context of the growth of discourses associated with neoliberalism.

Extending the Understanding of Shame in The Spirit Level

In TSL, shame is described as ‘the range of emotions to do with feeling foolish, stupid, ridiculous, inadequate, defective, incompetent, awkward, exposed, vulnerable and insecure’ (2009: 41) resulting in us internalising how we imagine we are seen by others. Consequentially, shame functions to shape behaviour in ways that ‘provide the basis for conformity throughout adult life’ and, ‘the reason why Scheff calls shame the social emotion is because he sees it as the psychological force underpinning both conformity and obedience to authority’ (2009: 41). But Scheff developed what he meant by shame being ‘the social emotion’ beyond viewing it as a simple social mechanism for producing conformity in a way which raises interesting questions about the contemporary nature of shame and the relationship between shame and inequality. For Scheff, shame is a social emotion because it relates to threats to a social bond, with the emphasis being on the latter, rather than on the production of conformity, as W&P suggest:

If … shame is the result of threat to a social bond, shame would be the most social of the basic emotions. Fear is a signal of danger to the body, anger a signal of frustration and so on … [but these] are not uniquely social. Grief also has a social origin, since it signals loss of a bond. Shame, since it involves even a slight threat to the bond, is pervasive in virtually all social interaction … all human beings are extremely sensitive to the exact amount of deference they are accorded. Even slight discrepancies can generate shame and embarrassment. Equally important is … a sense of shame. That is, shame figures in most social interaction because although members may only occasionally feel shame, they are constantly anticipating it. (Scheff, 2000: 97, first emphasis in the original, second emphasis added)

What this means is that the threat of shame is always present (if only latently and as a possibility) in our everyday encounters with others and, importantly, is present in all modern societies so as to be, arguably, a universal human experience. Scheff located his understanding of shame in earlier sociological thinking and proposed that future work should attempt to explore; ‘the key hypotheses on collective shame … that shame is increasing in modern societies but at the same time awareness of shame is decreasing’ (2000: 98) and that it is the least affluent who are most likely to be shamed by their status (acknowledging that this raises important methodological issues). Thus, shame can be prompted by social comparison and anxieties around status as argued in TSL; however, as shame is always present, understanding the experience of shame in relation to contemporary inequality has to go beyond establishing its presence or extent. Whilst it is arguable that greater inequality may simply serve to produce more shame, the question for social epidemiologists seeking to understand shame in unequal societies may not be, is shame present, how much of it is there and does it function to enforce social conformity? (it is present, even if not easily observable). But rather, do unequal societies do something which makes shame more pervasive and more corrosive to health? Related questions are, do the dynamics of contemporary inequality in the context of the growth of neoliberal discourses mean that protections from and resistance to shame are more problematic, or have these declined or disappeared? Is shame more problematic in unequal societies because it combines with other features of inequality, for example, the ‘re-commodification’ ³ of social goods which is a feature of contemporary neoliberalism (Coburn, 2004, 2010; Esping-Andersen, 1990).

To seek some purchase on the importance of shame to health and its potentially health damaging consequences we can turn to the well-known literature exploring the social origins of depression (Brown et al., 1995). At the heart of this literature is the experience of humiliation and shame; the fear of exclusion, of not feeling worthy of being held in mind, respected or attended to by others (Farmer and McGuffin, 2003). Space considerations preclude a detailed review, but what have been found to be central to depression are losses, and, specifically, losses which entail experiences which are humiliating or entrapping. These result in a doubling of depression onsets, with even higher rates among the most disadvantaged (Brown et al., 1995; Kendler et al., 2003). Such experiences have been demonstrated across a wide variety of countries and cultures (Broadhead and Abas, 1998), indicating that there may be something intrinsic to human experience which requires recognition and respect and which is damaged by shame and humiliation or entrapment.

The most harmful shame in the context of depression is chronic rather than acute. This may be similar to what Sayer (2005: 153) has described as ‘low level shame’, characteristic of unequal societies and far more difficult to access than acute shame, forming not just a backdrop to life but a sense of being woven into everyday experiences, and experienced as part of the habitus. For the least advantaged this can mean repeated exposure to numerous minor and major incidences of disrespect, mis-recognition or symbolic violence, starting in childhood and running throughout the life-course (Bourdieu, 1991; Sennett and Cobb, 1972). Shame as formulated in TSL highlights shaming events, resulting from explicit and invidious social comparisons; as Sayer puts it, ‘an intense, sometimes burning shame that follows from specific actions’ (2005: 153), as illustrated by the work of Gilligan (1996) with the emphasis on the loss and saving of face following shaming encounters (and the violence which may ensue from such efforts). These kinds of events are clearly important and likely to have health damaging consequences, but our point is that the more corrosive, day-to-day experiences of shame may not always follow this pattern.

Furthermore, as Sayer argues, class ideology means that structured into society is the expectation that working-class people must compete on the same terms as other classes, but without the same resources and advantages and are thus more likely to fail, to be seen to have failed, and experience themselves as failing – a process he describes as ‘structural humiliation’ (2005: 161). This experience of low level shame and structural humiliation, we would argue, may be as salient to the internalisation of inequality as the acute shaming events, or invidious shaming comparisons which are emphasised in TSL.

Understanding the place of shame in relation to population health is to see it not as straightforwardly connected to income inequality in quite the way described in TSL, but as shaped by the social, historical and political contexts in which inequality occurs, as others have argued (Coburn, 2004, 2010; Scambler, 2012). This in turn means exploring how working-class people, those who fare the least well in unequal societies, are seen by themselves and others (since shame involves both internal and external appraisals of the self) and the discourses that can be drawn upon to protect the self. We need to consider what resources might be available to people in responding to or protecting themselves from shaming events and the comparisons which might follow.

Shame, Status and Inequality and Social Class

There is of course a long history of the working classes being the objects of derision and either the targets of middle- or upper-class attempts to ‘improve’ them or the recipients of sentiments that can best be described as class hatred or contempt (Jones, 2011; Lawler, 2008). The contemporary form in the UK that this derision takes is the discourse around ‘chavs’ and the incomprehension and vitriol heaped upon ‘their’ supposedly flawed consumption, lifestyle choices and willingness to work (Jones, 2011). It also takes the form of judgements about the body (a discourse which is often gendered), and about how money is spent, particularly in the purchasing of goods to which status attaches, which is seen as shameful evidence of distorted priorities and an indicator that there is no real poverty any more. Looking at previous discourses, Wise describes how, in the slums of Victorian England, the overwhelming majority of children were insured by their parents, ‘a pauper grave being considered more shameful and tragic than a pauper life’, and how ‘many social explorers could not comprehend why parents would jeopardise their outlay on food and shelter in order to meet the weekly insurance policy payment’ (2009: 124). Not only was there a lack of comprehension of the need to avoid such shame, but also claims that the insured children were being murdered by their parents to rid themselves of unwanted lives and secure the insurance monies. However, investigations revealed that insured children had a lower mortality than their actuarial tables would predict – pointing towards the poignant reality that the insured children were the most loved and valued. To a contemporary ear these sorts of judgements appear both appalling and heartbreaking, but today’s attempts by parents to avoid shame or protect their children from shame by buying designer clothes and electrical goods are similarly mocked, or at best disapproved of.

It is not just the middle class or the more affluent who ‘other’ in this way as a way of shoring up anxieties about the self (and to protect against shame). Shildrick et al., in a study in the northeast of England examining the experience of adults in insecure work, found that even those on the very lowest incomes routinely talked about an often imagined ‘other’ poor in ways that were ‘demeaning and disrespectful’ (2010: 30), with the authors commenting that:

… we would explain this … as a local discourse that borrows from a powerful and widespread stereotype that demonises those who are unemployed or living in poverty and … bolsters a personal sense of self-respect and pride in managing to get by in hard conditions. (2010: 38)

Jones’ book, Chavs: The Demonization of the Working Class (2011), focuses on related themes, arguing that the changing structures of work and, correspondingly, the working class, combine with neoliberal ideology to provide numerous opportunities to shame and demean those who are workless or in low-status jobs. Central to this argument is the idea that neoliberal ideologies of personal blame and responsibility not only legitimise the devastating consequences of economic restructuring and the loss of manufacturing (and other) industries, but contribute to the damage by vicious stereotyping and othering of the least advantaged as feckless, dangerous and risible. Jones quotes the Labour MP Stephen Pound:

I genuinely think that there are people out there in the middle classes, in the church and the judiciary and politics and the media, who actually fear, physically fear the idea of this great, gold bling-dripping, lumpenproletariat that might one day kick their front door in and eat their au-pair. (2011: 131)

One element in the contemporary transformations which we have seen in recent decades has been the growth of neoliberalism. Neoliberalism has ideological as well as structural dimensions, with ideology providing a rationale and a legitimation for the social and economic priorities of the overall project: re-commodification of social goods, shrinking of the welfare state, erosion of trades unions and workplace rights and reduction in wages, combined with an attack on the idea that dependence can ever be a legitimate state (De Vogli, 2011; Navarro, 2008). Our point is that both the ideological and structural consequences of neoliberalism may shape the contemporary experience of shame. This, of course, has been the theme of other authors working outside epidemiology. For example, Sennett (1998) argues that:

The social bond arises most elementally from a sense of mutual dependence. All the shibboleths of the new order treat dependence as a shameful condition: the attack on rigid bureaucratic hierarchy is meant to free people structurally from dependence; risk taking is meant to stimulate self-assertion rather than submission to what is given … None of these repudiations of dependence as shameful, however, promotes strong bonds of sharing. (1998: 139)

Arguably, neoliberalism has both opened up additional spaces where the working class can be shamed and has undermined what might have been sources of resistance, which might protect health. However, we know that people are not simply passive; they endeavour to ameliorate and to resist harm or shame, drawing on the protections which are available to them. We now consider two such protections (albeit imperfect): these are destigmatisation strategies and collective imaginaries and we suggest that these offer possibilities for extending the arguments of W&P.

Destigmatisation Strategies, Collective Imaginaries and Resistance

We have argued that shame in TSL is conceptualised in a relatively narrow way with a focus on shame as engendering social conformity: ‘Pride is the pleasure and shame the pain through which we are socialised, so that we learn from early childhood onwards, to behave in socially acceptable ways’ (2009: 41). The shame process is seen as unidirectional, as bearing down on the individual as a result of their position in a hierarchy. But whether a person experiences or is able to resist shaming comparisons relates both to position in a hierarchy and to the resources – personal, social, cultural and political – which can be drawn upon to protect or to resist.

Lamont’s (2000) comparative study of French and American workers explored how the men she interviewed struggled to position themselves in ways that acknowledged dominant political discourses (‘American Dream’ ideas of failure and success, for example) but also afforded protection from invidious comparisons. Whilst both groups of workers (French and American) deployed some similar techniques, the French workers appeared to have greater protection from the shame or invidious social comparisons associated with their class position, because they drew on a more politically sophisticated understanding of class and exploitation and thus shamed themselves less as individuals. The American workers, she argues, were less successful in constructing discourses which might be protective of their identities, utilising a strategy of placing a high value on morality and religion as criteria for worth and attempting to avoid the primarily economic evaluations which threatened to position them negatively in the eyes of others. But the key point for our discussion here is that both groups of workers attempted to resist the shame-inducing comparisons and part of this capacity to resist was shaped by the wider repertoires and narratives available to them.

That the working class are not passive recipients of shame, that they attempt to protect themselves and resist, is of course well documented in the sociological literature on health and inequality (Charlesworth et al., 2004; Popay et al., 2003). Our point is that this is seldom addressed in contemporary social epidemiology and that understandings of health inequalities would be strengthened by greater dialogue across epidemiology and the sociology of health and illness.

Lamont goes on to comment that the health inequalities literature pays little attention to the ways in which resistance to stigma or shame may be important in mediating the health consequences of inequality, and describes the ‘destigmatisation’ strategies which groups use to challenge or resist discriminatory stereotypes or behaviours leading to their exclusion:

Considering destigmatisation strategies can broaden our understanding of the effect of racism and discrimination on health. Research has clearly shown the impact of inequality and discrimination on physical and mental health. However, social epidemiologists rarely consider how responses to inequality and discrimination can modulate this impact, and those who consider them tend to have a thin understanding of the role of meaning and the cultural environment in shaping these responses. (2009: 152, emphasis in original)

A significant dimension of destigmatisation strategies refers to what repertoires or cultural scripts are available and whether these are likely to facilitate or undermine health. Lamont (2009) proposes that when people are faced with threats to the self or the community they feel part of they may draw group boundaries determining who is ‘in’ and who is ‘out’, and construct narratives of who is worthy of being ‘in’ and what the meaning of the group is to the self. It is apparent that these processes can result in protection (for example, the ‘Hispanic Paradox’) ⁴ or the reverse in the sorts of ‘othering’ and exclusion described above by Shildrick et al. (2010). What follows from this is, first, that incorporation of such insights allows for a more complex understanding of shame and social comparison than that seen in much contemporary epidemiology; and, second, the importance of incorporating agency into these understandings. Lamont (2009) explicitly refers to the ‘thin understanding’ which characterises much current social epidemiology (Forbes and Wainwright, 2001) and this seems particularly apposite to a consideration of W&P’s proposals regarding the links between shaming comparisons and health. Indeed, Lamont’s (2009) arguments seem particularly salient during a social and economic period when many traditional working-class organisations and sources of solidarity and myth-making have declined, with consequences for the health of both individuals and communities. Lamont argues that in societies where no ‘alternative value systems’ are available, there is a likelihood that stigmatised groups will be both passive and more likely to internalise negative stereotypes prevailing as explanations for social and economic failure. We now turn to consider the notion of collective imaginaries.

Collective Imaginaries

This process of destigmatisation (and the potentially shaming social comparisons arising in unequal societies) can be further augmented by integrating the idea of ‘social’ or ‘collective imaginaries’ (Bouchard, 2009; Sewell, 2009). These refer to:

… sets of representations composed of symbols, myths, and narratives that people use to portray their community or nation and their own relationship as well as that of others to it. By virtue of their contributions to collective identity … by presenting a community’s past in a particular way, collective narratives influence the expectations of its members about the future, suggesting paths of collective development available to the community and strategies of action feasible for individuals within it. The moral valence of such representations lends them influence, but they have cognitive and emotional impact as well, conjuring up templates for action from the past. (Hall and Lamont, 2009: 12)

Such ‘collective imaginaries’ may provide part of a repertoire which explains, defends and rationalises our place in the world to ourselves and others and, crucially, may also offer a means of protecting the self against the symbolic violence that inequality can engender. It is not just actual qualities or resources (social, cultural and symbolic capital, income, education and contacts, for example) available to individuals and communities, but the extent and nature of the collective imaginaries available to be drawn upon to construct accounts and to reconfigure identities, which may protect or engender pride, self-esteem and efficacy, or indeed shame (Lamont, 2009; Lawler, 2009). This may be illustrated in Reynolds and Brady’s (2012) study of the health benefits of trade union membership, which found that households with a union member were healthier than those without. The relationship was not fully explained by the increased income, which could be assumed to be a consequence of membership – raising the possibility that this is explained in part by access to collective imaginaries.

However, these collective imaginaries are of course shaped by the broader political economy, where it is arguable that the growth of neoliberalism weakens the repertoires or collective imaginaries available to many (Frost and Hoggett, 2008; Gibson, 2007; Sewell, 2009). Whilst the UK (compared to the USA) has retained historically higher levels of both collective and state organisation, the collective imaginaries available to working-class communities based around collective provision, solidarity and trades unionism have declined, with more marginalised sections of the population being characterised in shaming terms, such as underclass or chavs (Jones, 2011; Skeggs, 2004). Framing historical changes using the concept of collective imaginaries may provide a way of reconfiguring our understandings of social comparison and shame, and the link between these and health.

Collective imaginaries can provide a sense of what it means to belong to a particular community and who is or is not included, be that community geographical, social or political. This may facilitate the integration or rejection, in whole or in part, of what are considered to be the dominant values of a society and a sense of what an individual might be expected to do when faced with particular challenges. In addition, collective imaginaries can shape the capacity to seek help either individually or mobilise with others to accomplish change.

As well as individual agency, collective imaginaries can also shape ‘second order agency’ (Hoggett, 2001), which is ‘agency which brings about a change of pattern in the life of an individual or group’. With the weakening of the trade unions and the growth of neoliberalism, ‘the kind of collective agency that was possible to those lacking economic and cultural capital even twenty years ago – rent strikes, consumer boycotts, industrial action – now seems a thing of the past’ (Frost and Hoggett, 2008: 441). Help-seeking too can flow from ideas about what ‘a person like me’ could be morally legitimated to do, being rooted in the political traditions or myths that form part of the imaginary. Eliciting the cooperation of others is vital to health and well-being in societies which make many demands of individuals and especially so where there is uneven provision of public services such as childcare or housing. As others have commented (Hall and Lamont, 2009; Marmot, 2005), eliciting cooperation is often linked with status, as status is ‘an all purpose social lubricant conditioning the co-operation one receives from others’ (Hall and Taylor, 2009: 93) and collective imaginaries can shape the likelihood of a person feeling it legitimate to seek help and from whom.

Collective imaginaries can enhance or erode feelings of inclusion or social isolation and, importantly, are located in a historical context shaping people’s ideas both about their futures and about the agency that they can bring to bear to their life circumstances. Whilst these imaginaries can constrain and have the potential in certain sets of circumstances to ghettoise, they are an important source of aspects of identity forming part of a repertoire which can protect against inequality or can explain how inequality has ‘got under the skin’ in circumstances of social or historical change. Conversely, the absence or weakness of such imaginaries may result in highly individualised experiences of social and economic change with few resources to draw upon to counter dominant ideas which may be shaming. The power of such imaginaries is further documented in Walkerdine and Jiminez’s (2010) study of a South Wales ex-steel town where almost agonising levels of shame were found amongst older men (ex-steel workers) in relation to their sons’ employment in the low-paid service sector jobs which were often the only work available. Men were described as refusing to speak to their sons if wearing pizza delivery uniforms and the young men themselves struggled with the stigma of jobs that were seen as feminine and shameful, also reporting feeling that such attacks on them were legitimate. The imaginaries allowed for partial protection of the self and a particular construction of masculinity, but at a high price as the outcome for many was that being unemployed was the only way to retain some vestiges of pride and masculinity.

Collective imaginaries, therefore, can provide a means of conceptualising how shame or invidious social comparisons may be resisted, but this is unlikely to be an entirely successful endeavour. What ideas and discourses dominate in descriptions of stigmatised groups, what claims are made for how they became stigmatised, and the ideas or ideologies which form part of the accounts that can be drawn upon to resist or deny stigmatised identities will all shape the extent to which they result in negative health consequences. What emerges from these ways of understanding is a rather different construction of the psychosocial perspective of TSL. Shame and the social gradient in health cannot be fully accounted for without including resistance and the resources which make resistance (and resilience) possible and this, in turn, needs to be anchored in a broad historical, social and political context (Williams, 2003).

Conclusion

In this article, we set out to critically evaluate one dimension of W&P’s (2009) explanation for how income inequality impacts on population health. In so doing, we sought to link their central arguments about the role of shaming comparisons and health with perspectives from the sociology of health and illness around the nature of shame and the role of agency and resistance. Lamont’s (2009) work provides a reminder of the connections that can be drawn between the growth of a neoliberal economic and political agenda and the corresponding decline in working-class organisations in recent years, including what this means for the social or collective imaginaries constructed through a class-based view of history (and, concomitantly, what this might mean for health). Notions of solidarity and class resistance now have far less force and salience than they once did, reducing their potential to offer (protective) meaning in lives that may be harsh, precarious or difficult (Pearlin et al., 2005). With declining trades union membership in the UK and continental Europe (Grainger, 2008), there has been a loss of many of the structures within communities (labour organisations, labour clubs, bands) which both conveyed and sustained identity and pride and which could generate ideas to protect against unwarranted shaming and symbolic violence (Hall and Taylor, 2009). As these movements and ideas have become marginalised, everyday explanations for inequality have shifted and become increasingly individualised with the space for shame becoming wider. Arguably, this may also mean that the health inequalities divide will widen, not just because of the practical, or neo-material consequences of political decisions, but because of the loss of collective imaginaries (Reynolds and Braby, 2012).

Returning to TSL and its key arguments, increasing income inequality alone seems unlikely to be the sole determinant of inequalities in health. Unequal societies are increasingly characterised by both a widening income gap and greater and more unequal difficulties and daily pressures for the poorest (Pearlin et al., 2005; Smith, 2005; Turner and Avison, 2003) – assumed in much political discourse to be the responsibility of individuals – coupled with a decline in the resources, both practical and ideological, which might protect against these. Therefore, our argument in this article is that many of the most disadvantaged have a life experience of repeated, low level shame (Sayer, 2005), alongside a diminution of the collective imaginaries which might allow protection or some form of resistance.

This is a different way of understanding how shame and social comparison might operate to that presented by W&P, where social forces are seen to shape and bear down on people. By integrating ‘everyday’ resistance, it is possible to see that, ‘Humans can only be shaped if they have structures and powers which resist some influences and yield to others. Without capacities for resistance then, like air, we could not be shaped’ (Sayer, 2005: 32), and that, ‘our relationship to the world is not simply one of accommodation … but at least in some ways, one of wanting to be different and wanting the world to be different’ (2005: 35).

Reconceptualising shame in these ways allows us to see how people in their everyday lives draw on practical and discursive resources and resist, both habitually and consciously. Much social epidemiology does not take this into account and this fails to acknowledge one aspect of our capabilities as humans.

Thus, exploring the role of shame may be accomplished by locating it within a wider social, political and historical context, seen as shaped by the growth of neoliberal ideology; that is, shame and its impacts are not best understood simply as the by-products of the extent of the income divide (De Vogli, 2011; Scambler, 2012; Sewell, 2009). Indeed, Frost and Hoggett link the growth of neoliberal ideology to the more overarching idea of loss and its impact on class and identity, arguing:

… the loss of one’s own history (as history is largely not written by the powerless), the loss of a sense of the achievements of one’s group or class, the loss of valued role models, icons and heroes, present and past, … the individualisation of such experiences of hurt and loss can leave individuals feeling very much alone; shame in particular can lead to a withdrawal from intimacy, networks, connectedness. (2008: 448)

Exploring the role of shame in relation to health inequality means embracing and looking beyond explanations rooted in epidemiology and at the strategies employed by individuals to defend against, rationalise or grapple with their circumstances. It also entails locating these struggles in historical and political, as well as social, contexts. In the context of recent calls to highlight the health consequences of injustice (Popay et al., 2010), such an approach may offer much to those interested in taking forward the ideas of Wilkinson and Pickett (2009) in relation to the socially corrosive consequences of inequality.