Over the last four decades, vulnerability discourse has become ubiquitous—not only in everyday language, but also in fields as diverse as philosophy, medicine, humanitarian aid, environmental studies, and economics. We commonly say that humans are, by our very own nature, vulnerable beings, that infants are vulnerable research subjects, that orangutans are vulnerable to extinction, that certain ecosystems are vulnerable to destruction, that certain economic systems are vulnerable to financial crises, and so on. But what is vulnerability, and why does it matter from a moral point of view?

This chapter opens with three cases of vulnerabilities among humans and animals. These cases serve as an illustration to show that while we all probably have some intuitive understanding of what vulnerability amounts to, it is not easy to identify the common feature in virtue of which the individuals in these scenarios are deemed vulnerable. In the second section, I distinguish between different uses of vulnerability ascriptions. I explain that vulnerability is a dispositional concept. From this follows that we should distinguish the reasons why an entity is vulnerable from its conditions of manifestation and from its different forms of manifestation. Furthermore, I introduce the two most important conceptions of vulnerability in the literature: vulnerability as a universal property of all human and non-human animals (called universal or ontological vulnerability), and vulnerability as a property of only some groups or individuals in some contexts (situational or circumstantial vulnerability). Lastly, I outline the conditions that a satisfactory account of vulnerability should fulfill, before proceeding to my own definition of vulnerability in Chap. 3.

2.1 Illustrations of Vulnerability

In 1932, the U.S. Public Health Service gave the green light to a study that became one of the biggest scandals in biomedical research. At that time, those suffering from syphilis could not yet count on any effective treatment. The usual cures—such as mercury—were often more harmful than beneficial, leading researchers to wonder whether leaving syphilis untreated might be more helpful to patients than taking any medication at all. This was the starting point of the now-infamous “Tuskegee Study of Untreated Syphilis in the Negro Male.” In this context, researchers investigated the natural development of untreated syphilis, in the hopes of better understanding the disease. In total 399 black sharecroppers suffering from syphilis, from Macon County, Alabama (a rural, predominantly black, and rather poor area) were enrolled in the study, while 201 men with a similar societal background but without syphilis were selected as a control group. The participants were promised treatment for what researchers called “bad blood,” and were provided with free meals, medical care, and stipends for their burials once they died. These conditions seemed attractive to many of the participants.

In 1947, penicillin became widely available as the recommended treatment for syphilis. However, the researchers running the Tuskegee study refrained from administering these life-saving antibiotics to their study subjects; instead, they stuck with their original plan to investigate the full natural course of untreated syphilis as a medical condition—with disastrous effects for the study participants.

In 1965, a young social worker and epidemiologist called Peter Buxtun, who was working for the U.S. Public Health Service, learned about the study, which was still ongoing. Outraged by what he heard, he twice filed a complaint—to no avail. In 1972, he decided to leak information about the study to Jean Heller, an investigative journalist. Heller published articles on this story in July 1972 in the Washington Star and the New York Times. Her articles caused a public outcry, leading Federal Agencies to appoint an ad hoc advisory panel (with nine members from different scientific backgrounds) to review the study’s design. This panel’s investigation showed that the men participated freely in the study, but that they were not aware of their medical condition, the exact nature of the study, nor its aim. Furthermore, penicillin was withheld from them, and they were not informed that this effective treatment existed. The board concluded that the study was ethically unjustified, and recommended stopping it. By the end of the study, numerous men had died and an untold number of wives, children, and other people close to the infected men had inadvertently been infected with syphilis. A class-action lawsuit was filed, and a settlement was reached for ten million dollars. Furthermore, the survivors of the study, along with some of their family members, received free medical care for the rest of their lives.

Evidently, reparation alone was insufficient: actions were needed to avoid such exploitative practices in the future. Shortly thereafter, the Belmont Report—a code with principles and rules governing medical research on humans—was established, and was partly influenced by the misconduct of the Tuskegee syphilis study. While its main ethical principles are respect for persons, beneficence, and justice, the report also contains a significant section on vulnerable research subjects:

One special instance of injustice results from the involvement of vulnerable subjects. Certain groups, such as racial minorities, the economically disadvantaged, the very sick, and the institutionalized may continually be sought as research subjects, owing to their ready availability in settings where research is conducted. Given their dependent status and their frequently compromised capacity for free consent, they should be protected against the danger of being involved in research solely for administrative convenience, or because they are easy to manipulate as a result of their illness or socioeconomic condition. (The National Commission for the Protection of Human Subjects 1979)

This was one of the ways in which the terms “vulnerable subjects,” “vulnerable groups,” and “vulnerable populations” found their way into guidelines for research with human subjects.

In the 1990s, another scandal broke—this time involving not humans, but animals. Cows are herbivorous animals that normally feed on grass; in industrialized farming, however, cheap commercial feeds are frequently used as an alternative. For years, British farmers fed cattle (destined for meat and milk production) a protein supplement called “meat-and-bone meal.” This meal contained the remains of other animals, such as cattle or sheep. Unfortunately, it did not always come from healthy animals. Due to this feeding practice, many cattle were infected with bovine spongiform encephalopathy (BSE), also called “mad cow disease.” BSE is a neurodegenerative disease that results in abnormal behavior like aggression, severe moving issues, weight loss, and eventually death.

Scientists first assumed that this disease could not be transmitted to humans and other animals. However, they rapidly had to change their view: cats, mice, and zoo animals fed with contaminated meat also fell ill. And soon, it was discovered that this disease also affects humans who eat contaminated meat: they develop variant Creutzfeldt-Jakob disease, which is fatal.

The epidemic reached a peak among cattle in 1992 and 1993. Other European countries besides the United Kingdom, such as Germany and France, were also affected. Over 100,000 cases of cows suffering from BSE were confirmed. In an attempt to stop the disease, 4.4 million cattle were slaughtered and their carcasses thrown away.

What is wrong with this? The cattle and other affected animals were sentient beings. While they could not formulate it in words, they had an interest in the avoidance of unnecessary suffering, an interest in pursuing species-typical behavior, an interest in being fed species-typical food, and an interest in continued existence, that is, to be able to continue living pleasant moments. Cattle commonly exist for the sake of humans, insofar as it is humans who bring them into existence for their meat, their hides, and their milk. While farmers were responsible for these animals, they nonetheless made them vulnerable by relying on a cheap feeding supplement that was inadequate for herbivores. Inadvertently, they directly harmed the cattle under their care by their actions, ultimately also putting many human lives at risk.

The Tuskegee study and the emergence of BSE are two cases that involve intuitively clear cases of vulnerability. And there are many others. Another, more recent, example involving vulnerabilities is the separation of over 2000 children from their parents at the United States’ southern border in 2018. Consequent to U.S. President Donald J. Trump’s “zero tolerance” migration policy, illegal immigrants from Central and South America were stopped at the U.S. border, where parents were frequently separated from their children. The parents were prosecuted for illegal entry into the United States. But since their children could not be charged with a crime and thus not be detained in jail, these minors were brought into detention camps, which, as the media and lawyers frequently reported, offered deplorable conditions: cold and over-crowded rooms, insufficient or unwholesome food, and not enough sanitary facilities. Due to the ensuing public outcry, on June 20, 2018, President Donald Trump signed an executive order restricting family separation. Nonetheless, subsequently reuniting the families often proved difficult in practice, as the situation was chaotic and officials were missing a central database with the details of all the family members and their locations. Today, it remains an open question whether all these children have been reunited with their family, not to mention how they will cope with their traumatic experiences in the future.

Taken together, these cases illustrate that many different groups can be deemed vulnerable, and that vulnerability is context-sensitive—that is, the groups that can be identified as vulnerable vary depending on the context. The examples also show that we all probably have some intuitive understanding of what vulnerability involves. In practice, however, it seems hard to pinpoint what actually renders the humans and animals in the examples above vulnerable. Is it the fact that they were exploited or instrumentalized? Is it the fact that they were not duly respected? Is it that they were in a situation of dependency? Or is it that their lives or livelihoods were at risk? Again, the answer seems to depend on the scenario.

In addition, one may object to the view that only some groups or individuals are considered vulnerable in certain contexts. Are we not all vulnerable beings, by our very nature? Indeed, we are all mortal, embodied beings who may be harmed and wronged. And if we are all vulnerable by nature, is it not problematic to refer only to certain groups and populations as vulnerable, as is currently done in many fields, such as medical research, healthcare, and humanitarian aid?

Furthermore—one could add—if so many different groups are deemed vulnerable, does not the concept of vulnerability become meaningless? Do vulnerability ascriptions add anything useful to moral and political discourse? What purpose does it serve to describe some groups of humans and animals as vulnerable, if we are all vulnerable by our very nature? And can animals really be described as vulnerable, after all?

The following sections address these questions: I will analyze the concept of vulnerability along with its different possible uses, and introduce the notions of universal and situational vulnerability.

2.2 Vulnerability: The Conceptual Landscape

Vulnerability ascriptions are common in everyday language. Consider the following examples:

  1. (1)

    Infants are vulnerable research subjects.

  2. (2)

    Sahid is vulnerable to poverty.

  3. (3)

    Aadhya’s computer is vulnerable to viruses.

  4. (4)

    The ecosystem is vulnerable to irreversible damage.

  5. (5)

    Blue whales are vulnerable to extinction.

  6. (6)

    Iran is vulnerable to political sanctions.

As I will demonstrate below, these examples will help to identify the structure of vulnerability ascriptions:

  1. (i)

    vulnerability ascriptions commonly have the form “X is vulnerable to Y”;

  2. (ii)

    vulnerability is a dispositional property;

  3. (iii)

    the manifestations of vulnerability are often perceived negatively;

  4. (iv)

    the manifestations of vulnerability are externally caused; and

  5. (v)

    vulnerability is not an essential property of humans and animals, contrary to what one may intuitively think.

The examples (1)–(6) above reveal the common structure of vulnerability ascriptions: vulnerability can be ascribed both to living entities (such as humans, animals, and the ecosystem) and to non-living entities (the computer and Iran). Furthermore, most entities X are vulnerable to a particular Y. This means that if we want to meaningfully ascribe vulnerability to some entity X, we have to specify the Y to which X is vulnerable: we must say that an ecosystem is vulnerable to destruction, an economic system is vulnerable to financial crisis, or a state is vulnerable to attack.

Vulnerability ascriptions refer to some possible state of being, rather than an actual one. If we say that blue whales are vulnerable to extinction, we are claiming not that blue whales are extinct, but that they may likely become extinct. If a newscaster informs us that some countries are vulnerable to political sanctions, she is not saying that sanctions are already in place, but that they may likely be instated. That is, as the above examples show, vulnerability ascriptions express that some Y may happen to X. Vulnerability thus refers to some potential state that has a certain likelihood, under certain circumstances, of manifesting itself. This means that vulnerability is a modal property—more precisely, a dispositional property.

What are dispositional properties? There is no agreement to date on a generally accepted definition of dispositions in philosophy. Broadly speaking, a dispositional property is the property of an object that is liable to manifest a certain state if certain conditions obtain. Thus, dispositions refer to some possible state of affairs for a certain object. Insofar as they express the potentiality or possibility, under some circumstances, that an object will manifest a certain state, dispositional properties entail a probabilistic component. Accordingly, it can be said that dispositions depend on subjunctive conditionals and entail the possible behavior or manifestation of a certain object. By way of illustration, the opposite of dispositional properties are categorical properties, such as shape, (molecular) structure, massiveness, or triangularity: a stone is massive (or not), and a triangle is triangular (or else it is not a triangle). They do not have any special relation to modality and conditionals, contrary to dispositional properties. Further examples of dispositional properties are solubility and fragility. They manifest if certain conditions are fulfilled: a fragile glass shatters if it falls from a table. Sugar dissolves if it is put in liquid. Therefore, we can say that dispositions have a characteristic form or state of manifestation. There can be one manifestation for a certain object (such as the breaking of a fragile vase), or several, in which case one speaks of multiple realizability. Vulnerability seems to be multiply realizable: it does not have only one possible and specific form of manifestation; rather, several are conceivable.

In general, three different points can thus be distinguished for dispositions: (i) the reason why some X is disposed to Y (e.g., a fragile glass is disposed to break because of its atomic structure); (ii) the conditions under which Y manifests (e.g., a glass breaks if it is dropped); and (iii) the manifestation itself (e.g., the shattering and breaking of the glass).

Significantly, the same analysis applies to vulnerability, a convincing account of which should be able to distinguish and explain the following three points:

  1. (i)

    the reasons why an object is vulnerable;

  2. (ii)

    the conditions under which vulnerability is manifested; and

  3. (iii)

    the manifestations of vulnerability.

The reasons explain why an object is vulnerable; the conditions represent the causes of manifestation; and the manifestation can be regarded as the resulting state. A convincing account of vulnerability thus needs to explain the reasons for vulnerability, its conditions of manifestation, and what its different forms of manifestation are. I will spell out these points in the next chapters.

Moreover, the manifestation of vulnerability is often seen as something negative, which should be avoided. The examples (1)–(6) above state that some entities, such as countries, the ecosystem, animals, and humans are at risk of manifesting some undesirable state Y, and imply that we care about this risk. Indeed, we usually regard Y as negative and would prefer that Y did not occur. For example, by saying that “an ecosystem is vulnerable to destruction,” we mean that the ecosystem is at risk of being destroyed—a resulting state that we regard as adverse and undesirable, as we prefer that the ecosystem remains in its original state, or at least that it does not change in this way. Vulnerability ascriptions in the case of non-sentient entities signify that the entity in question is valued in its current state, yet is disposed to be adversely affected.

In the case of the vulnerability of sentient beings, preventing the manifestation of vulnerability is in the interest of the beings themselves, while in the case of non-living entities, their disposition to manifest a certain state has no value for the entity in question, as they do not have any experiential well-being of their own. Thus, it is in the interest of sentient vulnerable beings to avoid the manifestation of vulnerability for their own sake—contrary to vulnerable objects such as ecosystems or countries. Conversely, invulnerability means that the entity in question cannot be adversely affected. In what follows, I will concentrate on the vulnerability of humans and animals, as living, sentient beings.

The manifestations of vulnerability are externally caused or brought about. Vulnerability is thus an extrinsic property, rather than an intrinsic one (McKitrick 2003; Armstrong 2017). If we analyze the vulnerability ascriptions above, we notice that vulnerability does not necessarily inhere within things or individuals. Rather, entities are caused to manifest a certain state by external circumstances or actions. For example, it is not something in the nature of blue whales that makes them prone to going extinct. Rather, it is human behavior—by whaling, polluting oceans, and overfishing—that renders whales vulnerable to extinction. Likewise, it is surely not the case that some individuals possess some intrinsic properties that make them more likely to endure poverty, while others lack these properties altogether. There is nothing intrinsic to some persons which makes them more prone to suffer from poverty than others. Rather, poverty has to do with external circumstances, such as being born into a place with limited access to resources—be it because these resources are depleted by others, or because there are just not enough resources available for the individuals living there.Footnote 1 Finally, if my computer is vulnerable to viruses, this does not necessarily have anything to do with my computer; rather, it has to do with my computer’s interactions with the world. Only by surfing the World Wide Web and clicking on suspicious websites, opening links from unknown sources, downloading problematic content and not renewing anti-virus programs, do I put my computer at risk and thereby render it vulnerable to catching viruses. Vulnerability is therefore an extrinsic property of objects, not an intrinsic one.

Consequently, the conditions of manifestations of vulnerability are relational: a manifestation of vulnerability usually occurs due to some interaction between the vulnerable individual and the world. Causes of the manifestation of vulnerability may thus be the individuals themselves when they interact with the world, the circumstances, or other living beings. One may object on the grounds that if individuals manifest vulnerability themselves, for instance by harming themselves when taking part in risky sports, the reason for manifestation is intrinsic. But this interpretation is shortsighted, as it overlooks the fact that some interaction with the world was still involved, even if the individual was the main source or cause.

Are all humans and animals vulnerable by their very nature? That is, are they essentially vulnerable, or could they lack vulnerability as a property? Essential properties of an object are properties that this object must necessarily have, and therefore could not lack without ceasing to exist. Accordingly, if vulnerability were an essential property of humans and animals, they would not conceivable without it. In order to determine whether this is the case, a thought experiment can help: can we imagine non-vulnerable humans and animals and would we still consider them to be humans or animals, respectively? This seems to be the case. Consider, for example, embryos in a very early developmental stage, say a few days after conception in the blastocyst stadium: they may be destroyed, but since they do not have—until a later stage of their development—any actual interests, they cannot count as actually vulnerable, but only potentially so. Yet they are human (or animal) nonetheless. The same applies to dead humans and animals, or those in an irreversible vegetative state: they are non-vulnerable to the extent that they can no longer be adversely affected, yet they remain human beings and animals. Therefore, vulnerability is an essential feature neither of humans, nor of other animals.

2.3 Universal and Situational Vulnerability

Two different understandings of vulnerability dominate the literature: universal or ontological vulnerability on the one hand, and situational or circumstantial vulnerability on the other. Those who construe vulnerability as universal start from the assumption that vulnerability is a shared property of all humans and sometimes animals as well—that is, that we are all vulnerable by our very nature. Those who understand vulnerability as situational assume that some individuals are rendered vulnerable in certain situations or contexts—that is, they become a vulnerable population that should be afforded special protection and additional attention. Furthermore, some accounts combine these two understandings of vulnerability in their writings. I will discuss these two conceptions of vulnerability in turn, beginning with situational vulnerability.

Vulnerability often seems to be situational: we frequently talk about vulnerable refugee groups, vulnerable minority groups, vulnerable medical patients, vulnerable research subjects, and so on in specific contexts. When we refer to vulnerable groups, we imply that they should be afforded special protection, attention, or assistance. For example, if we say that children are a vulnerable refugee group, we are implying that we need to pay special attention to their needs in asylum accommodation and refugee camps, as these may often be overlooked.

We often assume that it is intuitively clear who the vulnerable individuals or groups are. As a consequence, for a long time, many authors and guidelines (e.g., in medical research) did not concern themselves with defining vulnerability. Typically, they listed socially salient groups, including the homeless, the poor, children, racial minorities, prisoners, refugees, the marginalized, the elderly, and the like (see, e.g., the National Commission for the Protection of Human Subjects 1979; ICH Steering Committee 1996).

There are several problems with list-based approaches toward vulnerability, however. First, such lists lack an organizing criterion or characteristic that is shared by all the individuals mentioned, and it is not clear why precisely these people should be considered vulnerable. Second, these lists are often vast, such that nearly everyone appears to be vulnerable, which may result in criticism of the use and application of the term “vulnerability”. Third, describing so many different populations as vulnerable may lead to the stigmatization of some of these groups: if women, racial minorities, or the elderly are considered vulnerable, this characterization may foster discriminatory attitudes toward these groups along with opposition to the claim that they deserve special protection. Fourth, listing vulnerable populations misses the fact that among the group members, there are always certain individuals who are not particularly vulnerable. A failure to recognize this discrepancy can result in overly protective measures, which may be inappropriate for some individuals. Thus, overly comprehensive lists of vulnerable populations that fail to include exact criteria for vulnerability ascriptions may not only be misleading and stigmatizing, but also counterproductive, since they may result in opposition to the idea of special protection for those who actually need it. Conversely, the lists may not be comprehensive, and failing to recognize some vulnerable populations can potentially lead to negative consequences.

Some organizations have recognized these problems and changed their strategy. For example, the Council for International Organizations of Medical Sciences (CIOMS) defined vulnerability in its 2002 guidelines as follows: “Vulnerable persons are those who are relatively (or absolutely) incapable of protecting their own interests. More formally, they may have insufficient power, intelligence, education, resources, strength, or other needed attributes to protect their own interests” (Council for International Organizations of Medical Sciences 2002: Guideline 13). In the subsequent paragraphs, the Council mentions various groups as vulnerable, including—to list just some—medical and nursing students, subordinate hospital and laboratory personnel, employees of pharmaceutical companies, members of the armed forces or police, residents of nursing homes, people receiving welfare benefits or social assistance, the unemployed, patients in emergency rooms, patients with incurable disease, individuals who are politically powerless, members of communities unfamiliar with modern medical concepts, and persons who have serious diseases (potentially disabling or life-threatening). In 2016, the CIOMS revised its guidelines; tellingly, it no longer presents a long list of vulnerable individuals. It states that “a traditional approach to vulnerability in research has been to label entire classes of individuals as vulnerable. The account of vulnerability in this Guideline seeks to avoid considering members of entire classes of individuals as vulnerable” (Council for International Organizations of Medical Sciences 2016: 57). Instead, the Council now lists alleged properties of the vulnerable, such as diminished capacity to give informed consent, being in a hierarchical setting that may pressure people to partake in research, and being institutionalized.

Justifying what exactly makes these groups vulnerable is no simple task, though. Other accounts ascribe vulnerability to groups of individuals who are more likely to be exploited (Macklin 2003), who are unable to protect or safeguard their own interests (Nickel 2006), who lack basic rights (Zion et al. 2000), who are susceptible to additional harm (Kottow 2003), or who are at risk of having unequal opportunities to achieve maximum possible health and quality of life (Danis and Patrick 2002). It is suggested that, due to their status as vulnerable individuals, they should be afforded special protection and additional attention—for example, when they enroll in medical research or when they happen to be hospitalized.

All these definitions present difficulties. For instance, some persons are able to give free and informed consent, or may have access to the health care system, but can be considered vulnerable nonetheless. One example is a hospitalized yet conscious person suffering from a rare and severe disease. Since research in the field of rare diseases is not lucrative, there may be no adequate medication or treatment available, making it likely that the person will receive either no medication at all or an unsuitable treatment. And so we may want to consider this person particularly vulnerable within the healthcare sector (but not necessarily in other contexts). Another example could be women in countries with repressive gender attitudes. If women are considered inferior to men, their needs are likely to be ignored in many domains—regardless of their ability to give informed consent. Therefore, while the lack of some ability (such as decisional capacity) can certainly render a person vulnerable, it is insufficient as a characteristic of vulnerability on its own.

Could exploitation provide the criterion? That is, are the vulnerable those who are more likely to be exploited? The Tuskegee study described above would seem to confirm this suggestion. However, the notion of exploitation is notoriously vague. It is not obvious what it fully includes, since this depends on the definition and interpretation of the concept. If exploitation is defined as taking unfair advantage of someone and benefitting from it, there may be situations in which a person is not exploited yet still vulnerable. This will be the case for those who are more likely to be victims of indiscretions and breaches of secrecy in healthcare, or those suffering from rare or terminal diseases, who are more likely to take inappropriately high risks in medical research in the hopes of being cured. A definition of vulnerability that is based on exploitation as its main criterion therefore fails, not only because the concept of exploitation is unclear, but also because it is not a sufficient criterion for defining vulnerability.

This lack of a clear criterion for ascribing vulnerability hinders the identification of vulnerable individuals in different domains, preventing these individuals’ being afforded the protective measures they are due. It may also lead to the neglect of some individuals who are not regarded as vulnerable—although they ought to be. Other negative consequences may involve unwarranted paternalism, undue overprotection of some individuals, and stigmatizing attitudes toward the vulnerable.

But there is a further, more important, challenge. The view that only some individuals and groups in some domains are vulnerable is sometimes criticized on the basis that describing only some as vulnerable disregards the fact that everyone is vulnerable by their very nature. Some philosophers and bioethicists argue that vulnerability is universal. On their view, it would be divisive to deny that vulnerability is a shared human property, as failing to recognize this basic fact may lead to stigmatization. We should not think of vulnerability as something that only concerns other groups and populations (i.e., not us), as we are all vulnerable (Danis and Patrick 2002). Thus, recognizing only some people’s vulnerability ignores the reality that we are all, to some extent, vulnerable by our very nature.

This view, which defends what can be called a universal understanding of vulnerability, is intuitively appealing: all humans may experience various unpleasant events over their lifetime. Among other things, we may seriously hurt ourselves, fall sick temporarily or suffer from life-threatening diseases, become injured in natural catastrophes, be attacked by other humans or animals, become dependent on others at various points of our lives, be treated disrespectfully or have our autonomy disregarded, and finally, face the risk of dying at any moment of our lives. All these incidents seem to instantiate what we would intuitively call “vulnerability.” Living beings are vulnerable to various forms of harm, such as diseases, accidents, attacks, hunger, fire, and, ultimately, death. Universal vulnerability is mostly seen as a property of all humans due to our very nature, as forming part of the human condition, or as a consequence of our finitude. Vulnerability is therefore mostly (but not exclusively) linked to dependency, to having a body, to the possibility of encountering harm, and to being mortal (Callahan 2000; Kemp 2000; Rendtorff and Kemp 2000; Thomasma 2000; Rendtorff 2002; Kottow 2003; Gert 2004; Hoffmaster 2006; Harrosh 2012).

What are the consequences of universal vulnerability? Its implications vary in the literature. Some philosophers regard universal vulnerability as purely descriptive, considering any qualification of it as normatively relevant, a naturalistic fallacy. That is, they claim that vulnerability is an essential part of being human which cannot have any ethical dimension (Kottow 2004: 284). Other authors derive individual obligations from a universally shared vulnerability: we have obligations toward others because they are vulnerable beings. The vulnerability of other individuals is then, according to this view, one reason for us to have moral obligations toward them (Goodin 1985: 42). Still others derive institutional duties from vulnerability: care ethicists, for example, often claim that vulnerability should be taken into consideration not only by moral theories, but also by political and social systems. Institutions, they argue, ought to acknowledge and account for the general vulnerability and dependency of human beings by recognizing them symbolically and materially (Dodds 2007; Engster 2019). Other theorists claim that universal vulnerability is the foundation of equality, even of human rights. Given that we are all equally vulnerable—according an argument by Martha Fineman (2008)—political theories should be responsive to our shared vulnerability. That is, societal institutions should be created in ways that take vulnerability into account, and they should ensure that no group is disadvantaged or given undue benefits: societal assets should be distributed equally. Another author, Bryan Turner (2006), regards vulnerability as the foundation of human rights: vulnerability, he claims, is closely associated with fundamental rights, such as the right to life. According to his account, vulnerability is the foundation for respecting others. In order to reduce humans’ various vulnerabilities and dependencies, we create political institutions that ensure collective security; at the same time, these institutions may be imperfect and increase our vulnerability. Thus, human rights are rights that we need as protection from our vulnerable and dependent nature.

There is a tension between the situational and universal conceptions of vulnerability. Indeed, these two disparate interpretations seem contradictory and thus irreconcilable: the idea that vulnerability encompasses everyone conflicts with the view that it is a property restricted only to some. Due to this conflict, it remains undecided whether vulnerability has some normative pull, or whether it is instead a descriptive concept. Thus, it is an open question whether vulnerability entails some action-guiding component and requires that something be done about it, or whether it merely describes the human condition without any normative implications. Moreover, not resolving this conflict may result in opposition to the idea that the vulnerable should be afforded special protection: if all human beings are vulnerable by their very nature, how can special protection only for some be justified?

What is needed, then, is an account of vulnerability which encompasses universal vulnerability, but which can, at the same time, account for the increased situational vulnerability of some individuals or groups in some contexts, who should consequently receive special attention. Some authors have identified this problem and proposed comprehensive definitions of vulnerability. In her article “Elucidating the Concept of Vulnerability: Layers, Not Labels,” Florencia Luna (2009) argues in favor of the view that vulnerability is a relational concept, determined by the relationship of the person in question to a certain context or set of circumstances. She argues against regarding vulnerability as a label or category that can be applied to certain persons, and so she rejects lists of vulnerable (sub-) populations. Instead, she proposes a layered approach: individuals can have one or several (sometimes overlapping) layers of vulnerability, which arise in certain situations or contexts. That is, vulnerability is not a permanent condition or label that one has. Rather, one can have more or fewer layers of vulnerability. Such an approach has the advantage of flexibility, and it can account for the idea that everyone is vulnerable, to some extent. At the same time, it allows for the identification of some individuals as particularly vulnerable and in need of special protection and attention. However, this account does not clearly specify the circumstances under which someone acquires an additional layer of vulnerability. In another article, Luna suggests with her co-author a number of examples: a lack of autonomy, problems with voluntary informed consent, cognitive impairments, stigmatization, the disrespect of one’s rights, and the risk of exploitation (Luna and Vanderpoel 2013). The different layers a person accumulates should be identified; afterward, appropriate measures must be found in order to reduce the effects of these layers. But this incomplete list does not allay the worry that it may prove difficult in practice to identify all the potential layers of vulnerability without overlooking some. These layers must also be distinguished from the layers of vulnerability that cannot be reduced or prevented.

One very influential taxonomy of the sources of vulnerability was formulated by Wendy Rogers, Catriona Mackenzie, and Susan Dodds (Rogers et al. 2012). They argue that vulnerability is an under-theorized concept that needs to be further investigated and defined, given its importance in contemporary bioethics. In their work, they distinguish among three overlapping types of vulnerability, according to the source from which the vulnerability stems: inherent, situational, and pathogenic vulnerability. All humans share the same inherent vulnerability, due to our having a body, our needs and our dependency on others. Situational vulnerability results from the individual, social, political, economic, or environmental situation of an individual or group. Finally, pathogenic vulnerability (a subtype of situational vulnerability) stems from actions that aggravate existing vulnerabilities or cause further ones. Each source of vulnerability can be dispositional or occurrent. Occurrent vulnerabilities should be diminished, and dispositional ones prevented from becoming occurrent.

The account of Rogers et al. has the advantage of resolving the conflict between the two diverging understandings of vulnerability currently found in the literature. While inherent vulnerability describes vulnerability as a human condition, pathogenic and situational vulnerability can account for the increased vulnerability of certain individuals. However, this conception encounters two challenges. First, the account’s grounding of moral obligations directly in vulnerability—that is, we have obligations toward others precisely because they are vulnerable—runs into a circularity problem: the vulnerable do not have a claim to protection because they are vulnerable. Rather, if certain types of vulnerability give rise to moral obligations, they do so because we accept the principle that those affected or concerned by these vulnerabilities have a claim to be protected. This claim cannot be grounded in vulnerability itself—as this would be question-begging—but only in other claims, which we accept on other grounds (e.g., that we should prevent unnecessary harm and wrongs from occurring). Second, it is unclear whether Rogers et al.’s three types of vulnerability refer to the very same concept (Armstrong 2017: 186). Their account encompasses many phenomena, but it is unclear whether they all fall under what we commonly perceive as vulnerability.

Thus, no satisfying definition of vulnerability can be found in the literature, so far. This lack of a precise understanding of vulnerability—due to differing definitions and the unclear scope of the concept—has far-reaching negative consequences. On the one hand, it may lead to the neglect of some individuals who are not duly recognized as vulnerable and who are consequently denied the special protection they need. On the other hand, it may also result in the overprotection and stigmatization of those groups inappropriately designated as vulnerable. Furthermore, there may be opposition to the idea that some individuals in particular domains are in need of special protection and additional attention: if all human beings are vulnerable by their very nature, how does one justify special protection only for some? Does vulnerability not have any normative pull, after all?

And these are not the only problems. Recently, authors working on vulnerability were confronted with a new challenge: some authors went a step further and started ascribing vulnerability to animals. At first sight, this reasoning seems intuitive, for two reasons. First, many animals share basic features and properties with humans. After all, they can be harmed, they have a body and needs, and they are mortal beings. Thus, it seems evident that many animals partake in a universally shared vulnerability. The question then becomes how much this vulnerability counts from a moral point of view: does it matter morally, and if so, does it count as much as or less than human vulnerability? Second, animals can also be described as situationally vulnerable. Many animals share the properties often deemed relevant for increased vulnerability: for example, animals are unable to give informed consent. Some animal groups, such as the animals used in research laboratories, are unable to protect their own interests and are highly dependent on their care-givers. Due to these vulnerabilities—so it is argued by different authors—the animals concerned should benefit from protective measures similar to those afforded to humans in particular contexts, notably special protection and additional attention. In the case of animal research, this would mean that recommendations for vulnerable groups should be applied to research animals (Beauchamp et al. 2012; Ferdowsian and Choe 2013). This would imply, for instance, that research with these animals would only be permissible if there were minimal or no risk to the animals and if the research were in the interest of the animal. In other cases, it could also mean that some research protocols involving animals are problematic from a moral point of view and should therefore not be conducted. The implications of situational animal vulnerability for other contexts, such as zoos and circuses, remain unclear, as they have not yet been fleshed out in the philosophical literature.

All these open questions and problems have led to criticisms and concerns regarding the concept of vulnerability and its use. Some theorists argue that vulnerability does not add much to moral discourse (Levine et al. 2004), while others recommend abandoning the term of vulnerability altogether in some domains, as other—clearer—notions could be used to fulfill its function (Vladeck 2007; Wrigley 2015).

Nonetheless, vulnerability remains an established concept in political, medical, and ethical discourse. Despite its current vagueness, the term “vulnerability” will unlikely disappear from our language, nor can we erase it entirely from our vocabulary. Indeed, vulnerability discourse performs a meaningful function: to point out those individuals who may need special protection and additional attention. To abandon the concept of vulnerability without sacrificing this function, we would have to establish a new concept with equal normative force—but this new concept would probably suffer from similar shortcomings. Therefore, it would be best to come up with a more convincing and nuanced account of vulnerability.

2.4 Conditions for Defining Vulnerability

Why can’t we find a convincing definition or account of vulnerability in the literature? Besides the problem outlined above regarding the scope of vulnerability, there are other reasons. First, we all probably have some intuitive grasp of what vulnerability is and amounts to. And since many philosophers presuppose that we all have a shared understanding of vulnerability, they do not feel the need to define it properly. However, it is by no means clear that we all understand the same thing when we speak of vulnerability. Vulnerability can be described in various ways, because it is a dispositional concept. When defining vulnerability, one may focus on why a being is vulnerable; one can focus on the conditions under which vulnerability manifests; or one can focus on its different forms of manifestation. That is, different accounts of vulnerability often focus on its different aspects, without specifying which one. For instance, one could focus solely on the reasons why a being is vulnerable and thus explain vulnerability by referring to our embodied nature or to the fact that we are mortal beings with needs and interests. Another definition could focus on the different manifestations of vulnerability, such as being harmed or exploited. However, a satisfying account of vulnerability should not solely focus on one aspect, but rather encompass the whole picture. Second, the groups we deem vulnerable may change over time. For instance, the Belmont Report was a reaction to the Tuskegee syphilis study, described above, in which a poor, rural racialized population was used in research to study the development of untreated syphilis. Many definitions of vulnerability are in reaction to cases that previously posed ethical problems. The interpretation of situational vulnerability therefore evolves over time, and different features are emphasized depending on the context.

So how should we proceed? What should a convincing account of vulnerability look like? What conditions or desiderata should it fulfill? A convincing account of vulnerability should resolve the conflict between the universal and situational understanding; it should be able to handle the different case scenarios of what we would intuitively consider instances of vulnerability; it should also be able to explain why and how vulnerability matters from a moral point of view; and finally, it should be able to clarify both the connections and the differences between vulnerability and other important concepts, such as justice, dependency, autonomy, sentience, and the like. In sum, to satisfactorily respond to these challenges, convincing accounts of vulnerability should fulfill the following three conditions:

  1. (i)

    be able to account for both the universal and situational conceptions of vulnerability;

  2. (ii)

    be easily applicable to different cases and scenarios—that is, be general in scope; and

  3. (iii)

    be able to explain why some vulnerabilities are morally relevant and require action.

But what type of definition is the best candidate to fulfill these conditions? There are different types of definitions, and thus different possible ways to proceed. Yet many classical ways of defining vulnerability turn out to be inadequate.

If one wished to explain vulnerability by relying on its etymology and the history of the term, it would be defined as follows according to the Oxford English Dictionary: a vulnerable entity (from Latin vulnerabilis, wounding) is an entity that may be wounded, is susceptible of receiving wounds or physical injury, is open to attack or injury of a non-physical nature, and the like. According to this dictionary definition, vulnerability amounts to being exposed to the possibility of physical or mental harm. Being “vulnerable” in this sense does not mean that someone is actually harmed, but that they are susceptible to harm, such as pain and suffering. Vulnerable entities may be harmed under some circumstances. Accordingly, this definition emphasizes that vulnerability is a dispositional concept, since it makes reference to the potential state of an object, as outlined earlier. However, this approach to defining the concept is clearly insufficient: by focusing solely on the different forms of manifestation, it leaves out the reason for vulnerability as well as its ascription conditions. Consequently, such an approach cannot tell us much about how the concept should be used in contemporary bioethics or in other fields, such as humanitarian aid.

A second possibility for defining vulnerability is to consider its use in everyday language and conversations. Such an approach, frequently used by ordinary-language philosophers, is called a lexical definition. In ordinary language, “vulnerable” means that one may be attacked or injured in some way, or that one is susceptible to being emotionally damaged or offended. Definitions focusing exclusively on the usage of a term are often criticized, however, since they do not necessarily rely on univocal criteria for the concept to be defined. Furthermore, such definitions are often too broad, as they do not address the scope of the definiendum (the expression that is being defined). This complicates the determination of who the vulnerable individuals are and what properties they commonly share. In addition, such an approach usually just lists different uses; that is, it refrains from arguing in favor or against a certain use, and it cannot mediate between, or reconcile, different diverging understandings. We cannot, therefore, properly deduce what vulnerability means in a certain domain from the common-language use.

A third option for defining vulnerability is to present a stipulative definition, by which one determines the meaning of a term oneself, for the purpose of one’s project. In practice, this means that one establishes or fixes a meaning and continues working with it, regardless of the concept’s former uses. Such a definition might be useful if one is looking for a working definition by which a certain problem or specific case can be resolved, as when international organizations sometimes define minority groups as vulnerable populations; insofar as their aim is to protect these specific groups, it is not necessary to thoroughly investigate what vulnerability means in general. However, this is not a viable option for the present study, which aims to determine what can be understood as vulnerability in bioethics. The definition I am seeking here should, as outlined before, be formal in scope, resolve the conflict regarding the scope of vulnerability, and cover different intuitive case examples of increased vulnerability. A stipulative definition is diametrically opposed to this aim—it would simply posit what I wish to investigate.

Given the various conditions under which vulnerability manifests itself, one cannot straightforwardly indicate the necessary and sufficient conditions of vulnerability ascriptions. Vulnerability seems too wide and too diverse a concept to be defined in this way. One way out of this problem might be to define vulnerability in terms of overlapping circles of family resemblances. The concept of “family resemblance” goes back to Ludwig Wittgenstein’s Philosophical Investigations. For the (later) Wittgenstein, “the meaning of a word is its use in the language” (Wittgenstein 2009: § 43)—but there is often no uniform use of a certain term, and there may be several different meanings. According to Wittgenstein, one can discern, among the various uses of a term, a complicated network of overlapping and crisscrossing similarities. This is what he calls family resemblances (Familienähnlichkeiten in German). They do not spell out any necessary and sufficient conditions but rather consist of similarities. Wittgenstein presents the concept of “games” as an example of family resemblances: different sorts of games are all linked by similarities, in virtue of which it is intuitively clear to us what a “game” is, even though we would have trouble defining it explicitly (Wittgenstein 2009: § 67).

To define vulnerability using a family resemblance approach, one must examine different cases of vulnerability ascriptions and outline the different uses of said concept. Their constituents may differ from case to case, but, according to Wittgenstein’s theory, there will be similarities or resemblances between them. As pointed out earlier, however, it is not clear what vulnerability actually is. Given its different uses, one cannot simply point to something and describe it as vulnerable—the concept is too complex for that. Unlike games, there probably aren’t different types of vulnerability that can be subsumed under one umbrella term—only different types of manifestation. In his family resemblance account, Wittgenstein seems mostly concerned with the definition of umbrella terms that can be instantiated in different forms, which does not apply to vulnerability.

A final option for defining vulnerability is to present an explicative definition. Such a definition respects some central uses of ordinary language, but is stipulative about others, in order to guarantee an applicable definition for a certain domain. That is, such definitions are, to a certain extent, ameliorative—they aim to improve the concept at stake. The theory of explication goes back to Rudolf Carnap, who defines it as follows: “By the explication of a familiar but vague concept we mean its replacement by a new exact concept” (Carnap 1947: 7). More precisely: “By the procedure of explication we mean the transformation of an inexact, prescientific concept, the explicandum, into a new exact concept, the explicatum. Although the explicandum cannot be given in exact terms, it should be made as clear as possible by informal explanations and examples” (Carnap 1962: 3). One chooses an explicative definition for concepts that are inexact and unsystematic. Thus, the aim of an explication is to offer an improved definition of a problematic everyday-language concept that needs to be defined for a certain use and domain. Accordingly, the new definition need not be extremely close to the meaning of the concept in ordinary language, nor need it capture all aspects of the concept. It should refer to its essential aspects yet remain close enough to the former concept that it can be used in its stead. Explications are thus never completely exact. One presents an explicative definition for a certain purpose, in order to work with it in a certain domain. Its appropriateness depends on whether it is a satisfactory definition or not, and whether it resolves the problems of former definitions. The correctness of an explication thus depends on the context for which the definition was made.

According to Carnap, to explicate a concept, one has to begin by looking at the everyday-language use of some fuzzy concept in a certain domain in which it is used. That is, one analyses exactly what one wishes to define, and specifies for which context. Context-dependency is important, since understanding the use of a concept in a certain context can already provide some clarification as to what precisely is meant by the explicandum in said context. Furthermore, it helps by including essential parts of the concept while excluding those that are not important for a particular domain.

Carnap outlines four conditions that an explication must fulfill: (1) it should be similar to the concept to be defined; (2) it should be exact; (3) it should be fruitful; (4) it should be simple (Carnap 1962: 5). Regarding the first condition, the explication should feature some similarity to the original concept concerning its content. However, this similarity need not be as close as possible: an overlap is sufficient. Explications do not correct the rules of language, but rather suggest a change to the rules of language for a certain domain. To explicate a term does not imply that the explicatum always replaces the explicandum. The use of the explicandum in some contexts and for some purposes is still appropriate. In the case of vulnerability, this means that we can still use the concept in our everyday language. However, when we refer to a more “technical” use of it, designating certain groups or referring to universal vulnerability and its importance, we would refer to the new term. By “fruitful,” Carnap meant that one could formulate universal statements and that the explicatum can be compared and linked with other similar (scientific) concepts (Carnap 1962: 6). And by “exactness,” he meant that we can insert the concept into a system of scientific concepts.

In the case of vulnerability, this means that the definition should be formal and broad in scope, so that it can be easily applied to different scenarios and situations within one domain. One should be able to infer who the vulnerable are from an explicative definition. Additionally, one should see the relationship of the explicatum of vulnerability to other concepts more clearly. In the case of vulnerability, these other concepts would be dependency, fragility, sentience, and the like. The last condition Carnap outlines is the simplicity of the definition. According to Carnap, simplicity only becomes important if one has a choice between several useful concepts, and one must choose the best one; in such a scenario, he recommends choosing the simplest of the concepts.

Carnap recommends explicative definitions for scientific, empirical, logical, or mathematical objects; that is, he restricts himself to concepts that belong to science (his example is “pescis” for fish). One could now object that it is not an appropriate definition type for vulnerability, as vulnerability does not fall into any of these categories. However, nothing rules out the possibility of explicative definitions for other, less scientific concepts. A definition of vulnerability should fit into, and be consistent with, a general system of similar concepts and assumptions in one domain—but this system does not necessarily need to be based on logico-mathematical or empirical concepts. Hence, the fact that vulnerability cannot be defined in scientific terms or based on empirical evidence does not constitute an insurmountable objection to seeking an explicative definition for it. One can still present an exact definition that outlines specific rules of ascription and helps to resolve the question of what vulnerability means in some domains. Thus, although Carnap’s explication was originally intended for scientific terms, one can nevertheless present a fruitful explication of vulnerability. Having made these clarifications, I shall now proceed to my own definition of vulnerability.