Balancing words, balancing lives: framing vulnerability in times of crisis

Words matter, especially in times of crisis. This article analyses the complexities of political discourse on vulnerability by considering the case of the Dutch COVID-19 response. Our study finds that the framing of vulnerability as a predetermined and naturalised condition, linked to old age and pre-existing medical conditions, draws attention away from aspects of precarisation tied to economic position, social class, cultural background and living conditions. This rhetorical strategy can be understood as a practice of de-responsibilisation through which attention is rhetorically diverted from the way(s) in which political authorities are implicated in producing, exacerbating or failing to mitigate vulnerabilities.


Introduction
Words matter, especially in times of crisis. The words we use to describe people, events and phenomena focus our attention on particular aspects, while drawing attention away from others. When it comes to vulnerability, the effects of the terms we choose to describe diverse forms and degrees of susceptibility to harm, loss and suffering can be especially damaging. While some expressions call for care, attention and action, others omit or conceal realities of precarity, minimising the moral and political responsibilities these entail.
This article examines the complexities of political discourse on vulnerability by considering the case of the COVID-19 response in the Netherlands. Our study highlights the reductive way in which even well-meaning governments may be inclined to define vulnerability and identifies the harmful consequences that can follow from an overly narrow understanding of this notion. Scrutinising how vulnerability is framed in the political domain is an important task because language shapes not only our views of the world but also our awareness of social issues and our willingness to respond with ameliorative action.
As Carol Bacchi (2009: 40) argues, political discourse represents social problems in particular ways. These limit what can be thought or said about a given issue, attribute persons a certain role in relation to the problem (which, in turn, may influence the way they are perceived by others, as well as themselves), and have a material impact on the lives of those affected. Critical engagement with the underlying assumptions that shape political discourse can call attention to alternative ways of perceiving the problem and thereby create opportunities for different policy decisions. It is thus important to consider how vulnerability is presented in political discourse and what alternative understandings are conceivable.
Our analysis finds that vulnerability is predominantly framed in the Dutch policy discourse on the COVID-19 response as a predetermined and naturalised condition, linked to old age and pre-existing medical conditions, drawing attention away from aspects of precarisation related to economic position, social class, cultural background and living conditions. This framing avoids sensitive questions about who holds responsibility for the precarious position of certain groups in society that places their members at high(er) risk of severe detrimental consequences for their health and general well-being due to the pandemic and subsequent crisis response. The focus on vulnerability as a physical condition of frailty limits the government's framing of this notion in critical ways and leads to tensions between its rhetoric and practice. Importantly, the discursive emphasis on a naturalised and medicalised notion of vulnerability can be understood as a practice of de-responsibilisation through which attention is rhetorically diverted from the way(s) in which political authorities are implicated in producing, exacerbating or failing to mitigate vulnerabilities.
Our main argument is that this reductive notion of vulnerability as a condition determined by factors that lie beyond personal or political responsibility should be replaced by a multifaceted conception of vulnerability that acknowledges the multitude of sources from which human vulnerability springs, including those in which political powers are implicated. While the rhetoric used does not adequately reflect this complexity, it shines through in policy decisions that tacitly acknowledge various facets of vulnerability in seeking to balance the detrimental effects of the crisis in terms of the life, health and well-being of (most of) those affected.
The first part of the article introduces the notion of vulnerability and identifies some of the challenges of dealing with vulnerability in times of crisis. The second part explains the methods and approach of the study. The third part presents our findings, highlighting the tensions between an explicit, narrow understanding of vulnerability as set out in the policy discourse and an implicit, broader sense of vulnerability that underpins (certain) policy decisions. The fourth part discusses our findings with a critical eye for what is omitted in this political framing of vulnerability. The fifth part offers guidelines for finding a solution for the identified problems, calling for a multifaceted notion of vulnerability.

Vulnerability in times of crisis
Vulnerability is a complex notion, as it takes different shapes and affects people (as well as other beings) in various ways. Scholars have noted that there are two broad ways in which vulnerability may be understood (Mackenzie et al, 2014: 4-7). The first refers to a susceptibility to harm that is shared by all living beings (Mackenzie et al, 2014: 4-5; on this notion, see also, for example, Butler, 2004;Nussbaum, 2006;Ricoeur, 2007;Fineman, 2008). Our embodiment leaves us exposed to suffering, as our bodies may be wounded and harmed. As social beings, people are furthermore vulnerable to social injuries as well, which may result from processes of humiliation, rejection or exclusion. This vulnerability is inherent and universal. Although we are not equally vulnerable to all forms of harm and suffering, nobody can completely discard the risk of being hurt by others or avoid the need for support from others.
A second understanding of vulnerability highlights the differences that exist between individuals in how likely they are to be affected by certain harms and threats (Mackenzie et al, 2014: 4-5; on this notion, see also, for example, Goodin, 1985;Luna, 2009;Schroeder and Gefenas, 2009). While all human beings share a universal vulnerability, people are not equally exposed to risks. Judith Butler (2016: 21) notes, for example, how 'the degree to which any particular body is vulnerable at any place and time is inextricably bound up with social institutions and cultural norms'. Not only do these institutions and norms help determine how likely people are to be affected by particular threats and risks to their well-being, such as losing their source of income, developing (severe) health issues or being subjected to discrimination, but they also influence the resources that individuals and groups have available to deal with adversity, as well as the prospects for remedial action to be taken on their behalf. When policies focus on vulnerable groups in society, this differential notion of vulnerability is at stake.
In times of crisis, vulnerabilities are likely to multiply and deepen as people are exposed to new susceptibilities to suffering or inabilities to meet their needs, and already-existing proneness to these factors heightens. In the COVID-19 pandemic, for instance, all people became more vulnerable, as each of us was exposed to a greater likelihood of suffering harms, be it from getting ill or dying ourselves, losing loved ones, not receiving timely or adequate healthcare, being deprived of opportunities for education, work or other forms of meaningful activities and development, or being unable to meet our needs in terms of interpersonal contact and the enjoyment of personal freedoms, to name just a few examples. Responding to this multiplicity of forms of vulnerability poses a formidable political challenge.
It is therefore not surprising that vulnerability is a salient theme in scholarship on the pandemic. Vulnerability (and social vulnerability) has been examined, for instance, as an important risk factor in COVID-19 deaths (Kim and Bostwick, 2020), as contributing to the impact of the crisis and ensuing counter-measures on particular groups in society (De Vries et al, 2022), and as a prominent political feature of human life highlighted in times of crisis (Ghandeharian and FitzGerald, 2022). By examining how vulnerability is framed in the Dutch pandemic response, this article identifies the detrimental consequences of applying a reductive understanding of this notion and illustrates the need to develop a multifaceted perspective of vulnerability.

Tracing vulnerability in policy discourse
Our study is based on a discourse analysis of government letters sent from the Dutch Ministry of Health, Welfare and Sport (VWS) to the Dutch Parliament (Tweede Kamer), which justify the policy decisions made by the cabinet (for details, see Appendix 1). The main corpus consists of 26 letters, which were sent in the period from 27 February to 27 September 2020. The analysis is limited to letters from the first wave, as this period sets the baseline for the crisis response. While changes (in emphasis) have occurred since this period, the focus on the protection of the vulnerable remains central throughout the Dutch COVID-19 response.
The main corpus includes all letters that followed on advice given by the Outbreak Management Team (OMT) (an advisory board of [bio]medical experts) and meetings of the Administrative Adjustment Consultation (BAO) (a board of public officials who consulted on the implementation of the policy plans). All letters that present general accounts of the measures taken to address the spread of COVID-19 in the Netherlands were included. Additional letters on specific measures, including the closure of nursing homes for visitors, were consulted but were not included in the main corpus. All letters are publicly available through the website of the Dutch Parliament.
The letters were analysed in three steps. First, we used the software R to check whether 'vulnerability' was indeed a central term in the government discourse. We did so by uploading all the letters of the main corpus into R and creating a list of the most commonly used terms in the letters. We identified all morally loaded terms in the list (ranging from terms like 'good' and 'together' to 'vulnerability' and 'responsibility') and then checked how often particular semantically related word groups, associated with the identified morally loaded terms, occurred in the corpus. This meant that to determine how often vulnerability was mentioned in the letters, we did not just look for 'vulnerability' ('kwetsbaarheid'), but also for various declensions of the Dutch word 'vulnerable' ('kwetsbaar', 'kwetsbare' and 'kwetsbaren').
Second, we conducted a close reading of all the letters, with particular attention to passages containing references to vulnerability, using discourse-analysis methods proposed by Bacchi (2009) and Sevenhuijsen (2004) (see later). Third, references to vulnerability were grouped and analysed to determine the meaning(s) attributed to this notion and the principal uses made of the term. To aid this step, we used R to identify all references to 'vulnerability' and 'vulnerable' in the main corpus, producing a list of the sentences in which these terms occurred. This allowed us to identify who was described as 'vulnerable' and what meaning(s) were attributed to 'vulnerability' (for an overview, see Figure 1).
Our discourse analysis is inspired by Bacchi's (2009) work on the representation of social problems in discourse and Selma Sevenhuijsen's (2004) work on tracing normative content in policy documents. In our analysis, we draw particularly from Bacchi (2009: 40) in examining the effects of the framing of vulnerability on policy decisions and those affected. We consider how the way in which vulnerability is represented in policy discourse limits what can be thought and said about vulnerability as a social issue, who (if anyone) is held to be responsible, how this affects the lives of those involved, and what alternative understandings are possible.

Vulnerability as frailty
Vulnerability plays a central role in the Dutch government's crisis policy discourse. On 12 March 2020, the Dutch Minister of Health, Welfare and Sport explained that 'the core of the government policy is that we want to maximally protect the health of people that are vulnerable ' (25 295, no. 124: 3). Protecting the health of vulnerable people is one of the three stated pillars of the policy, together with maintaining the capacity of hospitals and maximally controlling the spread of the virus (25 295, no. 179: 1).
Vulnerability lies at the basis of the perceived necessity to takes measures that have a severe impact on public life. In letters from March 2020, a shift is visible from a focus on the negative socio-economic effects of restrictive measures towards a heightened sense of the urgency of curbing the spread of the virus in order to protect critical healthcare facilities and vulnerable people. A concern for proportionality is expressed by the Minister of Health in his letter of 12 March 2020: We should definitely not do too little, but also not too much. This entails that we should think well about potential negative consequences, we should not underestimate the economic and social effects. This is why we need to continue to look for proportional measures that also leave room for individual considerations and common sense. (25 295, no. 124: 1) Just a few days later, on 17 March, the minister claimed that the government was forced to take radical decisions to limit the spread of the virus given the inability of the healthcare system to deal with the sharp rise in contagions (25 295, no. 176). The decision to check the spread of the virus, he announced, aimed 'to ensure that there is always sufficient capacity to help people that are the most vulnerable' (25 295, no. 176: 1).
The choice to place vulnerability at the core of the political discourse demonstrates concern for people's well-being and a commitment to avoid or alleviate suffering and harm. It shows that the government took seriously the detrimental effects that the crisis was bound to have on the lives and well-being of those affected. Yet, vulnerability tended to be narrowly defined.
In the government letters, vulnerability is usually linked to health. The statement from the minister of health that the policy aims to 'maximally protect the health of people that are vulnerable' (25 295, no. 124: 3), as well as the coupling of the need to protect vulnerable people with the need to safeguard critical healthcare facilities, as mentioned earlier, suggest a strong tie between vulnerability and health. The association of vulnerability to health is also reflected in the semantic linkages in the letters, which show who or what is identified as vulnerable (see Figure 1). 'People' are most often called 'vulnerable' (76 out of 258 mentions), followed by 'elderly' (30 out of 258 mentions), 'health' (24 out of 258 mentions), 'patients' (24 out of 258 mentions) and 'persons' (22 out of 258 mentions). When people or persons are called 'vulnerable', this is often followed by the specification that they are vulnerable due to old age or underlying medical conditions. In his letter of 12 March 2020, the minister of health calls, for example, on everyone to restrict visits to these groups: 'People are urgently requested to limit visits to vulnerable people, that is to say the elderly and people with weakened immune systems' (25 295, no. 124: 2).
Other categories that are called vulnerable in the letters include 'children' (ten mentions), 'families' (eight mentions), 'people receiving mental healthcare' (six mentions), 'students' (three mentions), 'homeless people' (one mention) and 'people with disability' (one mention). 2 While the government does therefore recognise that other sources of vulnerability exist, as expressed in concerns over the predicaments of children exposed to domestic violence or other concerns about safety in their homes (see, for example, 25 295, no. 200: 21), or susceptibility to harm and suffering related to mental health (see, for example, 25 295, no. 199: 13), homelessness (25 295, no. 249: 33) or disability (25 295, no. 428: 25), the frequency of references to these groups is significantly lower than those made to the elderly, patients or people with medical conditions. This suggests that the government places concerns over vulnerability in a medical sense as central to its discourse.
The view of vulnerability expressed in the letters is thus predominantly focused on susceptibility to suffering severe health risks in the case of an infection with the SARS-CoV-2 virus due to a weakened physical state caused by old age and/or poor health. This speaks to one particular sense of the Dutch word 'kwetsbaar', which captures an array of meanings that would be expressed in English as frail, vulnerable or precarious. The meaning in the letters is most akin to frailty, which stands for the unstable relationship between health and disease (Thomas, 2008). Van der Meide et al (2015: 1) note that 'frailty is associated with, but not determined by, aging' because the notion is used to describe a physical condition of lacking in strength and prone to illness and infirmity. This understanding of vulnerability as a form of frailty is connected to the biological and medical sciences (Thomas, 2008). The prominent place it takes in the policy discourse on the COVID-19 response in the Netherlands may therefore be unsurprising given that this policy is strongly influenced by the recommendations of the OMT, an advisory board predominantly made up of (bio) medical experts.
Vulnerability, on the other hand, refers to a form of susceptibility to harm related to our being in the world as embodied and social beings who can be hurt by others or through natural and social conditions. While all people are vulnerable, they are not necessarily all frail (or at least not at all times). Frailty has to do with the phase of one's life (at the end of the human life span) or health conditions that put one in a delicate state. Vulnerability instead refers to the sense in which all human beings, even the most able-bodied, can be reduced to suffering through the infliction of physical or psychological harms, or the inability to meet their needs.
The Dutch notion of 'kwetsbaarheid' can also be used to refer to precariousness. Precarity refers to uncertainty about meeting one's needs in life, which affects some people more than others. Social and political conditions are important here because this uncertainty derives in large part from the absence of adequate protections for, and safeguards of, the means required to guarantee one's well-being. Those living in a state of precarity are balancing on a thin line that holds them within the sphere of persons who are seen as valuable contributors to society but risk losing this position through even a small change in their circumstances (Thomas, 2008;Lorey, 2015). Precarity thus expresses itself as a condition of social and economic instability and unsafety that puts people at risk of failing to meet even their fundamental needs.
The medically focused framing of vulnerability in the government letters highlights aspects of frailty and generally leaves out other dimensions of vulnerability, particularly those related to precarity. While vulnerability is mostly used as a label attributed to specific groups given their frail bodily condition, it is important to acknowledge that the letters do recognise in a limited sense that other sources of vulnerability exist. Yet, the number of mentions of other vulnerable groups, such as children, people with mental health issues, students, people with a disability or homeless people, is significantly lower, as noted earlier. Our analysis thus shows that the government places concerns over vulnerability in a medical sense centrally within its discourse. While this is an understandable response to a public health emergency, the deep impact and prolonged nature of the crisis require a more careful weighing of various facets of vulnerability and a critical appraisal of how the predominant focus on medical aspects related to physical frailty influences policy decisions.

The political and societal impact of the narrow view of vulnerability
The narrow understanding of vulnerability, as expressed in government discourse, impacted on policy decisions and on broader views in society about who was to perform what role in the pandemic. One important effect of this reductive view of vulnerability was that it fed into a dichotomous view of society, which rhetorically divided the population into a frail minority of elderly people and persons in weak health who should allegedly be protected by the healthy and resilient majority.
This divisive logic was reflected, for example, in the decision of the government to close all nursing homes for visits on 19 March 2020. This closure followed the logic of 'shielding' vulnerable populations by 'fencing them off '. 3 While nursing homes were, sometimes literally, fenced off from the outside world (with the notable exception of care workers, who continued to commute between the care institution and their homes), society at large was subjected to a partial lockdown. Schools, day-care centres, restaurants, bars, sport clubs and coffee shops were closed, and public events were cancelled, though shops remained open. People were urged to work from home and not to visit older people or people with underlying health conditions.
The justifications used in policy discourse presented these restrictive measures as necessary to protect vulnerable members of society. Scholars note that this argument was used as a 'mantra' because it 'entices an intuitive reflex to act and tries to persuade people to show solidarity' (Maeckelberghe, 2021: iv50). While the majority of the Dutch population supported the imposed restrictions at the time (RIVM, 2020), they also met with resistance. Some commentators held that the impact of the measures on the economy, healthcare sector and culture were disproportionate (van Dongen and van Mersbergen, 2020). Less diplomatically, one columnist expressed the view that 'the corona crisis screws up the prosperity of our children due to the inability to accept that old people have to die sometime' (Zwagerman, 2020). This example illustrates how the logic of distinction could lead to polarisation.
The response of an elderly man gave voice to the indignation felt at the crude opposition between generations that spoke from this representation of affairs: 'We all suffer from the corona crisis. Young and old. And then you're supposed to say: "Well, the older person has to choose death so the young can continue with their lives". This is not how it works!' (BNNVARA, 2021). While the appeal to the general population to protect older people and those who are frail was meant, presumably, to elicit solidarity, this example shows that the sharp juxtaposition between an allegedly vulnerable minority and a purportedly healthy and resilient majority could also backfire.
The polarising dimension of the dichotomous view derives, at least in part, from its essentialist tendency, as groups, rather than individuals, are attributed vulnerability based on a particular aspect of their identity, namely, being elderly and/or having a weakened immune system. This perspective does not take into account individual differences but, rather, considers particular characteristics of people's (social) identity to be leading in determining how to treat them. This may cause resistance, both among members of the allegedly able-bodied majority who are unable or unwilling to shoulder the burdens required to keep those considered vulnerable safe, and among members of the purportedly vulnerable minority who do not actually experience the kind of frailty that underpins the image of vulnerability expressed in the policy discourse.

Reflecting the complexity of vulnerability in policy decisions
Despite the discursive emphasis on a narrow sense of vulnerability as frailty, a broader understanding of vulnerability implicitly seemed to inform certain policy decisions and the development of the crisis response over time. The Dutch government's choice of a partial, rather than a full, lockdown may be understood, for example, as a response to the multitude of vulnerabilities that were exacerbated, exposed and produced by the COVID-19 crisis. It is important to note here that despite its rhetorical claims to 'maximally control' the virus and to protect the vulnerable 'as well as possible', the Dutch government did not declare a full lockdown, which (allegedly) would have been the best strategy for reaching these stated aims. By limiting the scope of the lockdown, the authorities responded to the vulnerabilities of Dutch citizens in terms of loss of income and meaningful activities, and exposure to stress and isolation, which, in all likelihood, would have been more keenly felt in the case of a full lockdown.
This suggests that while it may be natural to revert to a reductive, medicalised understanding of vulnerability in response to an unprecedented public health crisis, it proves difficult to ignore the complex, multifaceted nature of vulnerability. A certain awareness of this reality also shines through in the fact that the government tacitly acknowledged alternative grounds for vulnerability because it made funds available to secure the livelihood of people whose income would be rendered uncertain by the crisis (35 420, no. 2) and eased regulations allowing young people to apply more easily for social welfare (25 295, no. 249: 33). In the government letters, attention was paid to the need to ensure support for children in unsafe domestic settings (25 295, no. 219: 27) and guarantee access to shelter for homeless people (25 295, no. 200: 20). The government also established that no people were to be evicted from their homes except in cases of criminal activity or severe nuisance (25 295, no. 249: 33).
The complex nature of vulnerability was further reflected in changes in policy regarding the closure of nursing homes. The decision to fence off residents of nursing homes at the start of the crisis was heavily criticised for ignoring the impact this measure had on their quality of life (Pen, 2020). In May 2020, the decision was taken to reopen all nursing homes without infections, and in early June, the minister of health even stated that from 15 June, all nursing homes were required to implement eased regulations (25 295, no. 386: 23-4).
The government thus re-evaluated its stance on what matters most in the case of vulnerable people living in nursing homes. Whereas a national closure of nursing homes was deemed unavoidable and necessary in March 2020, it was seen as an unacceptable restriction just a few months later, favouring instead a more localised approach in which care professionals could balance safety against wellbeing. Evidently, lack of resources in terms of protective gear and test capacity played a role in setting the conditions in which a full closure was considered needed at the outset of the pandemic, but the reorientation of the policy at least implicitly acknowledged the complexity of vulnerability as a relational and situational concept.
While this discussion of the policy implications is far from exhaustive, it highlights the tension between a narrow view of vulnerability, as emphasised in policy discourse, and a richer understanding, at times reflected in actual policy practice, which corresponds more adequately to the diverse needs, demands and concerns of differently situated individuals and groups in society. The government led (some of) these calls in developing policies to protect people from major setbacks and in reorienting its policy towards nursing homes over time. This practical expansion of the notion of vulnerability did not suffice, however, to fully capture the scope of vulnerabilities to which people were exposed. Crucial dimensions of vulnerability were overlooked, as were the vulnerabilities of particular groups in society.
Our analysis suggests that while the Dutch government stressed a reductive, medicalised understanding of vulnerability, the complexity of vulnerability proved hard to overlook, even in a public health crisis. Implicit recognition of the broader scope of vulnerability was acknowledged in the practical expansion reflected in actual policy decisions. Yet, more is needed to guarantee truly inclusive policies: a reconceptualisation of the meaning of vulnerability in the policy domain is required to ensure that vulnerabilities are recognised based on the (potential) harm and suffering they entail, rather than the ease with which they can be addressed in politics.

Taking responsibility for vulnerability
The understanding of vulnerability presented by the government in the letters analysed is troubled by significant limitations. Our findings show that the government discourse on the COVID-19 response tended to reduce vulnerability to frailty, contributing to a dichotomous view that labels certain groups as vulnerable due to their alleged weak health compared with a strong and resilient majority. Not only does this understanding lead to a potentially divisive view of society, but it also proved problematic in the face of opposition by older people who did not recognise themselves as being in the frail state attributed to them, as well as members of the allegedly resilient majority who felt vulnerable on other accounts. Even if frailty is acknowledged as a disposition that, generally, marks the lives of older people, the idea that the right response to this form of vulnerability is to fence them off in times of a pandemic proved problematic, as this policy deepened other forms of vulnerability, including those related to loss of social connection and quality of life.

The vulnerability of invisibility
While these tensions are real and serious, it is important to note that they do receive attention in the government letters and are thus recognised as social issues that mark the lives of people who are seen to matter. A deeper problem with the framing of vulnerability as found in the government letters is that the vulnerabilities of certain people are not considered at all.
Certain vulnerabilities are more easily recognised than others. This follows from the fact that it is politically more difficult and (potentially) inopportune to acknowledge vulnerabilities that are produced, exacerbated or not adequately addressed by authorities. Children exposed to domestic violence or other forms of harm may well have been rendered more vulnerable, for example, due to continued budget cuts for youth care and assistance, as implemented by the previous governments led by the prime minister. While vulnerable children are on the radar of the government, other groups are not even mentioned in the government letters. Precarious groups like prisoners and drug addicts are not brought up. Undocumented migrants are referred to only once in the letters and only with regard to the situation in the overseas territory of St. Maarten. This is remarkable because their number in the Netherlands is estimated to be between 23,000 and 58,000 (van der Heijden et al, 2020) and this group is susceptible to various forms of harm and loss, even more so in the crisis, due to the precarious nature of their working and living conditions.
Pharos -a Dutch expertise centre that explores inequalities in health and access to healthcare -notes how a comparison between expected mortality rates and actual mortality rates for the first weeks of the pandemic in the Netherlands revealed that this number was 10 per cent higher for people with a migration background (Heijdenrijk et al, 2020). Patricia Heijdenrijk and her colleagues (2020) argue that the risk of contagion is influenced by living conditions, which are frequently less favourable for people with a migration background: (Labour) migrants often live under bad conditions, with several people in small houses on holiday resorts, or with several generations in too small a house. Poverty and social inequality often coincide with work that cannot be carried out from home and circumstances that do not allow for extra protective measures.
The authors also draw attention to the fact that groups with a low social and economic status, such as people living in poverty, homeless people, prisoners and drug addicts, experienced many issues due to the crisis measures. Despite the profound vulnerabilities these groups face, scarce notice was given to their predicaments. As Heijdenrijk and her colleagues (2020) note: 'in the midst of all the billion euros costing measures and big gestures, these people receive little attention and they are pushed closer to the fringes of our society'.
The situation of refugees is discussed in government letters on the impact of the COVID-19 crisis on the asylum procedure, but the discourse frames the pandemic not as an emergency situation that raises concerns about the well-being of asylum seekers but, rather, as a logistical issue to be resolved. In their letter of 23 April 2020, the minister and state secretary of justice and safety and the minister of legal protection announced that the asylum procedure would be suspended from 15 March and that new arrivals in the Netherlands would be brought to an emergency shelter location after having been registered and medically tested. Once there, asylum seekers would be 'restricted in their freedom' (35 300-VI, no. 126: 4) to limit health risks and reduce concerns from employees and people living in the vicinity regarding contamination. In response to criticism by non-governmental organisations, the ministers and state secretary stated that 'a more open form of shelter would lead to too much unrest for municipal boards and neighbours no. 126: 4). This letter illustrates that concerns among municipal board members, employees and people living in the vicinity of emergency shelters were taken more seriously than the right of asylum seekers not to be unduly restricted in their freedom while awaiting recommencement of the asylum procedure.

The politics of de-responsibilisation
Our analysis suggests that focusing on the vulnerability of some members of society is more politically convenient than engaging with the vulnerabilities of others whose predicaments may be more complex and less appealing to the general public. By putting forward the vulnerability of frail older people and people with weak health in its discourse, the government could justify the need for restrictive measures without engaging with complicated questions about personal and political responsibility for the way in which such vulnerabilities arise in the first place. By focusing attention on groups of people who supposedly simply happen to be susceptible to severe physical harms in case of infection with the SARS-CoV-2 virus, the government could appeal to solidarity in recognition of the needs of persons with a lot to lose in terms of their health, without going down the slippery slope of acknowledging political responsibility for its involvement in setting the conditions that render some people more prone to be severely affected by the pandemic and crisis measures than others. The naturalisation of vulnerability in the case of older people and people with underlying medical conditions thus allows the government to sidestep critical questions of responsibility.
This sidestepping may be understood as a discursive strategy of 'de-reponsibilisation', through which the government brushes off its political responsibility, both for setting the conditions that render certain people more susceptible to suffering severe harms due to the crisis and for how it allocates importance to the different vulnerabilities of various groups. We propose the term 'de-responsibilisation' to designate: a distinct form of responsibility shirking, which consists in rhetorically drawing attention away from the way(s) in which a particular agent is implicated in producing, exacerbating or failing to mitigate vulnerabilities of persons (and/or other living beings). 4 De-responsibilisation is not without its costs, however. Besides the risk of moral failure and loss of legitimacy implicated in the avoidance of political responsibility, this move also led to tensions and inconsistencies between rhetoric and practice, as discussed earlier. The government did seek to respond to the vulnerabilities of other groups and individuals in society, even if these vulnerabilities were not always adequately reflected in the accompanying discourse. While it is a step in the right direction for the government to acknowledge and respond to a wider set of vulnerabilities in action, more should be done to ensure that critical vulnerabilities are not overlooked due to the social, political or legal status of the persons or groups affected. Vulnerabilities should be addressed based on the severity of the harm, loss or suffering that is likely to follow, and the likelihood of this coming about, not the ease with which vulnerabilities can be used as a rhetorical tool to rally society behind one's cause.

Towards a multifaceted understanding of vulnerability
To provide a guideline for a more reflective and responsible engagement with the notion of vulnerability in the context of the crisis and beyond, we propose an alternative approach that can help guide government decisions and bridge the gap between rhetoric and practice. For this approach, we draw inspiration from the work of Florencia Luna, who argues for a layered notion of vulnerability: Vulnerability should not be understood as a permanent and categorical condition, a label that is attached to someone given certain conditions (such as a lack of power or incapability) that persists throughout its existence. It is not a black or white concept, that is, a fixed label that includes or excludes a particular group. Rather it should be seen as layered and inessential. (Luna, 2009: 129) Building on Luna's insights, we propose that various facets of vulnerability in the COVID-19 crisis can be charted more effectively by considering vulnerability along three axes that correct the reductive, essentialist and static nature of the dichotomous and naturalised view of vulnerability by considering the dimensions of vulnerability, experiences of vulnerability and conditions of vulnerability. While these axes are not exhaustive, they can help bring into sharper focus the aspects that are lost from sight in the narrow understanding of vulnerability linked to bodily frailty and focus attention on aspects of precarisation that dispel the framing of vulnerability as a naturalised condition.
First, vulnerability has various dimensions. Perceiving vulnerability as bodily frailty is too limited, as it omits other dimensions of vulnerability related to the embodied, social, moral and political aspects of the human condition. Vulnerability in the crisis can thus not simply be reduced to susceptibility to suffering severe health risks as the consequence of an infection with the SARS-CoV-2 virus. Even the category of bodily aspects of vulnerability is far richer than this narrow view suggests. Not only does the pressure on healthcare facilities caused by the pandemic entail vulnerabilities for people in need of medical care whose treatments are postponed or whose illnesses are diagnosed late, but loneliness and stress can also negatively impact on people's physical health, as can changes in eating and exercising habits, or the inability to understand health guidelines due to difficulties in understanding the language in which these guidelines are presented.
Besides embodied aspects of vulnerability, it is also important to consider other dimensions of vulnerability. Mackenzie et al (2014: 1) point to interpersonal dimensions of vulnerability, as they note how 'as social and affective beings we are emotionally and psychologically vulnerable to others in myriad ways: to loss and grief; to neglect, abuse, and lack of care; to rejection, ostracism, and humiliation'. In the COVID-19 crisis, a key aspect of these social and affective dimensions of vulnerability involved the isolation and distancing required to curb the spread of the virus, which led to loneliness and a longing for closer contact. Related to these social and affective dimensions is moral vulnerability, which exposes people to distress and tension in situations where moral values and principles conflict. The crisis led to tragic choices, as when care workers had to decide whether to let relatives visit dying patients when this could potentially jeopardise the safety of others under their charge (Robert et al, 2020).
The crisis also exposed the economic dimensions of vulnerability, as discussed earlier, where economic vulnerability refers to susceptibility to loss of one's standard of living. This may lead to milder forms of suffering and harm in cases where social welfare can guarantee people's basic needs, or have a detrimental impact when people's ability to meet even their fundamental needs is undermined. Political vulnerabilities, as related to undue restrictions of one's freedom, were a focal point of resistance to restrictive measures taken in the crisis response. Developmental vulnerability, which involves impediments that keep people, and particularly children, from developing their capacities or lead to disparities in their ability to do so, also drew attention because the closure of schools led to inequalities between children due to differences in the domestic settings and abilities of parents to support teaching at home. This overview, though far from complete, shows that the dimensions of vulnerability are manifold and that a reductive view does not suffice to capture the array of vulnerabilities to which people are exposed in the crisis.
Second, experiences of vulnerability need to be considered to avoid essentialist and generalising uses of the term. Van der Meide et al (2015: 8) note that the attribution of vulnerability can be alienating for (older) people, as they tend not to recognise themselves in the frail status ascribed to them. Furthermore, vulnerability may be experienced in different ways, and the meaning(s) attached to it by others may differ from the meaning(s) attributed by the people affected. The government's response aimed to save the lives of vulnerable persons, which is an understandable response to a public health crisis and takes up the political responsibility for its citizens in a conscientious manner. At the same time, this response, particularly as shaped into a strategy of 'fencing off ' the frail, did not necessarily respond well to the experiences of (all) those the policy sought to protect.
In the Netherlands, priority is usually given to ensuring people's quality of life, rather than necessarily to prolonging their life. Questions about quality of life seemed to be put aside, leading to a sharp contrast with the usual commitment to allow individuals a say in choices that severely affect their lives and to consider which measures improve their well-being. 5 This issue also shows that the various dimensions of vulnerability are not neatly separated but frequently linked. Susceptibility to loss of quality of life is, in itself, a vulnerability that, in turn, impacts people's vulnerability in terms of physical health, as mental well-being constitutes a factor that affects bodily health. In fact, care workers and relatives have argued that the prohibition of visits led to premature deaths in nursing homes due to the detrimental effects of isolation and loneliness (De Goede, 2020).
Third, conditions of vulnerability need to be considered in order to advance a richer understanding of vulnerability during the COVID-19 crisis. Conditions refer to situational aspects that influence people's susceptibility to harm, loss and suffering. As Luna (2009: 129) notes, it is conducive to approaching the question of vulnerability 'not by thinking that someone is vulnerable, but by considering a particular situation that makes or renders someone vulnerable'. In the crisis, older people living in nursing homes were exposed to greater risk of contagion, for instance, due to the fact that they were residing in facilities with many other persons and with regular contact between care workers who moved from one resident to another (and from the care institution to their homes), which may be marked as spaces of vulnerability. Similarly, as mentioned earlier, Heijdenrijk and her colleagues pointed to the increased risk of contagion of (labour) migrants due to their working and living conditions, which renders them more vulnerable given these circumstances.
Through consideration of the various dimensions, experiences and conditions of vulnerability, the use of this term as a label that tends to be reductive, essentialist and static can be corrected. In particular, this approach helps bring into focus aspects of precarisation that are not fully acknowledged in the government's framing. It thereby aims to counter the tendency towards de-responsibilisation and to contribute to more open and transparent reflection of the factors that need to be weighed in policy decisions.

Conclusion
This article has analysed the complexities of political discourse on vulnerability by examining the case of the COVID-19 response in the Netherlands. Our findings indicate that the Dutch government tended to define vulnerability in a reductive way, closely linked to a medical sense of frailty. While this framing may be understandable as an immediate response to an unprecedented public health crisis, it leaves out many important aspects of vulnerability. To remedy these ills, a reconceptualisation is needed that reflects various dimensions, experiences and conditions of vulnerability.
It is true that a richer understanding, which considers these different facets, may prove difficult to handle in political rhetoric, as it complicates matters, whereas the aim of the justification of restrictive measures is often to rally people behind a cause. However, as shown, not only can the stark opposition of vulnerable older people and people in weak health with resilient others not necessarily lead to solidarity, but it can also be divisive. A reductive focus on vulnerability as bodily frailty furthermore does not do away with other forms of vulnerability or the need to consider these in policy decisions. It represents choices as far simpler than they actually are and yet does not, in itself, make these decisions any more straightforward. The observed tensions between the policy discourse and policy decisions seem to derive, at least in part, from the fact that the discourse does not adequately reflect the complexity of the issue at hand, which cannot be overlooked in policy decisions without making unbalanced choices.
At the same time, it is important to acknowledge that it is practically impossible to respond to all forms of human vulnerability in an adequate manner. As Daniel Engster (2019: 112) rightly notes: 'since resources will always be limited and human vulnerability is almost limitless, societies will further need to make difficult choices about which vulnerabilities they wish to protect against most strongly'. These choices should be a matter of political debate, however, and should not be foreclosed by a narrow depiction of the dilemma at hand. The COVID-19 crisis presents a unique occasion for the Dutch government to engage with these questions and with the underlying issues that render some persons more vulnerable than others. Crises constitute opportune moments to correct wrongs of the past, as they put into perspective the vulnerabilities of all and allow for a revaluation of previous choices. It is the task of a responsive and responsible government to respond with care and attention to these vulnerabilities, and to consider which are most dire and require to be addressed most urgently. 6 Notes 1 The list of terms in Figure 1 is not exhaustive. The letters also include 21 references to 'vulnerable groups'. This term was not included because it referred to different categories. In most cases, 'vulnerable groups' referred to the elderly and people with underlying health issues, though this term was also occasionally used to refer to socially vulnerable groups. The letters also include references to 'vulnerable countries' (five mentions), 'vulnerable residents' (four mentions), 'vulnerable situations' (three mentions), and 'vulnerable target groups' (two mentions), as well as some other miscellaneous entities. These were excluded to enhance the clarity and readability of the figure. 2 This overview is based on an analysis with R, which considered the five terms preceding and following the word 'vulnerable' or 'vulnerability'. It does not offer a comprehensive overview of the government's framing of vulnerability. The government did call attention, for example, to patients receiving mental healthcare and persons with intellectual disabilities residing in care institutions as vulnerable people whose needs require adjustments to the general visiting regulations for care homes (see, for example, 25 295, no. 199: 11). This was not reflected in the number of times people receiving mental healthcare or disabilities were called 'vulnerable' in cases where these words did not occur within close distance of each other (that is, within the five words preceding and following the search term). 3 In a television speech on 16 March 2020, Prime Minister Rutte announced that people who run a higher risk should be shielded from the virus (Rijksoverheid, 2020). The choice of the Dutch term 'afschermen' hints at a notion of separation, with the vulnerable held behind protective screens, similar to the English notion of 'fencing off '. 4 The term 'responsibilisation' is used in governmentality literature to denote a process through which responsibilities are shifted from authorities to individuals (Wakefield and Fleming, 2009: 277-8). 'De-responsibilisation' could therefore be understood as the reverse process, through which state agents reassume responsibilities previously attributed to others. Our use of the term differs from this understanding in focusing on its meaning as a process opposed to the assumption of responsibility. In our view, no other word better captures the signification of a process through which responsibility is avoided through rhetorical diversion. 5 It is telling to note that the Dutch term for joint decision making ('inspraak') was mentioned only once in the analysed government letters. 6 The limits of this article did not allow for a discussion of the international dimension of this question, though vulnerabilities are evidently not limited to country borders. For a discussion of the international aspect, see Maeckelberghe (2021).

Funding
This work was supported by the Netherlands Organisation for Health Research and Development (ZonMw) under Grant 10430032010005.