Bioethics is widely seen to protect a vital principle: autonomy ( Steffen, 2016). Ensuring the defence of patients’ rights, their free exercise of choice, and respecting integrity is categorised under a critical principle, autonomy, closely related to rights and freedoms. Ensuring respect for individual autonomy has helped advance individuals’ interests, especially given that, throughout the history of bioethics, this principle has often been disputed, ignored, or violated ( Bazzano et al., 2021). Such a guiding principle has recently come under intense criticism in the face of emerging political tensions, such as end-of-life issues and anti-vaccine movements, igniting ‘toxic individualism’ and ‘community excesses’ ( Bourret et al., 2015). At the same time, whether it is abortion, surrogacy, vaccination, or medically assisted suicide, never before has the principle of autonomy been called upon in such an unprecedented way. As such, it seems to be the driving force of all the great affirmations of human will in bioethics and in other disciplinary spheres ( Le Coz, 2018). The heritage of Descartes, philosopher of modernity, and that of Hegel, for whom “the right to subjective freedom constitutes the critical and central point which marks the difference between modern times and antiquity” ( Hegel, 1975, p. 233) and more generally the liberal dynamics of the spirit of the Enlightenment have all emphasised this form of emancipation. Such a general will for emancipation disrupts the existing narrative of paternalism, including in care, empowering patients and ensuring the respect of their choices and consent ( Rendtorff, 2002). Stemming from a Kantian approach in philosophy, then further codified within different rules, laws, and legislation, autonomy is a triumphant ethical principle whose application is deemed to ease moral dilemmas ( O’Neill, 2022).

First formulated as a principle in the Belmont Report ( National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979), then brought into widespread application in medical ethics through Beauchamp and Childress’ principlism theory, autonomy encompasses a wide range of definitions, applications, and assumptions ( Beauchamp & Childress, 1979). As such, the current state of autonomy, and its future as a principle, has ignited large-scale discussion and some dispute. However, it can be argued that the underlying root of most tensions could, at least in part, be a confusion, misunderstanding, or disagreement on what autonomy is and involves. For example, in abortion, a conflict seems to arise between a fetus’s prospective autonomy and the woman’s autonomy in not bringing it to term. In medically assisted suicide, autonomy’s legitimacy as a guiding principle is significantly questioned. In anti-vaccination groups, autonomy is used to justify refusing vaccination, endangering collective well-being.

A large part of the academic literature is concerned with politicising or sanctifying this principle, fearing a “new hegemony of autonomy” ( McLean, 2010, p.34), granting autonomy a rigid superiority over other valuable principles, such as beneficence, non-maleficence, and justice ( Gillon, 2003). Here, a scoping review of the literature on autonomy will be conducted to observe, sort, and summarise those claims and the literature’s general disagreements on autonomy. We focus on the following questions: is criticism of the principle of autonomy legitimate, and how can we make sense of these accusations while distinguishing the genuine limitations of such a central principle? Can we ethically decide to violate this principle to guarantee others? Or, better still, can we address such criticism by refining the principle of autonomy through different approaches and practical applications? Finally, by introducing discussions that stem from the literature surrounding autonomy, such as narrative and relational autonomy, or by harmonising the principles together, can we find a consensus that addresses these issues while consolidating autonomy?

Autonomy has benefitted from a vast literature ( Steckermeier, 2021). Discussing a principle as extensive as autonomy thus requires recalling a foundational literature on autonomy, particularly from authors such as Kant, Mill, and Berlin. However, autonomy as an interdisciplinary term is not solely confined to philosophy, and perhaps the most practically important piece of work on autonomy that must be discussed stems from medical ethics, notably from principlism’s cofounders, Beauchamp and Childress.

First, for Immanuel Kant, autonomy is both the foundation of our moral obligations and the core motive for respecting persons as intrinsically worthy of respect. Kant advocated that a biologically determined or socially conditioned act cannot be autonomous: genuine autonomy stems solely from a self-given law or maxim ( Kant, 1785). In this sense, we all have a duty to act autonomously. To act out of duty consists of following a moral law that we give ourselves, hence the idea that we are the author of the moral law ( Kant, 1785). We are, in this sense, autonomous, not because we are subject to such law, but because, as authors of the law, we freely submit to it. Therefore, Kant considered that acting out of duty and acting autonomously are one and the same ( Kant, 1785).

However, if we are considered autonomous as we give ourselves our own moral law, then that begs the question, what is such law, and from where does it stem? Kant considered that this moral law stems from reason: it is through reason that we act autonomously, as reason reveals which ends are worth pursuing ( Kant, 1785). To determine which ends are worth pursuing, one must call upon the categorical imperative. The categorical imperative is a universal and independent requirement: “it has to do not with the matter of the action and what is to result from it, but with the form and the principle from which it results; and what is essentially good about it consists in the disposition, whatever the result may be” ( Kant, 1785).

There are two formulations of the categorical imperative, and both are essential to properly understand the Kantian perspective on autonomy. The first is universalisation: “Act only in accordance with that maxim through which you can at the same time will that it become a universal law” ( Kant, 1785). As such, all contradictory actions, or those which present a conflict, cannot be universalised. For example, when universalised, abstaining from vaccination results in a logical contradiction: “if autonomous individuals are allowed to violate autonomy in self and others by inflicting harm, we are imagining a world in which everyone is autonomous but also a world in which there is no genuine autonomy” ( Gandjour, 2022, p.12). We therefore all have a duty to get vaccinated.

The second aspect of the categorical imperative is non-instrumentalisation: “Act in such a way that you treat humanity, whether in your own person or in the person of any other, always at the same time as an end, never merely as a means” ( Kant, 1785). Kant values respect for oneself and for others as one and the same principle. The duty of respect is inherently tied to the rationality and autonomy of persons, two traits inherent to humanity. Indeed, for Kant, the ability to act autonomously gives human life its dignity ( Kant, 1785). Autonomy grants the difference between people and things: persons can exercise autonomy, but objects and (non-human) animals cannot ( Kant, 1785).

Respecting autonomy means treating people as ends in themselves: hence the idea that it is immoral to use people in the name of general well-being, as stated by the categorical imperative. The categorical imperative stems from autonomy, as, since persons are deemed autonomous, it is therefore illegitimate to ‘objectify’ or instrumentalise them. For example, while we do have a duty to vaccinate ourselves, enforcing such duty may be considered immoral: “to deny an adult the right to make their own decisions, however mistaken they are from some standpoint, is to treat them as simply means to their own good, rather than as ends in themselves” ( Gandjour, 2022, p. 12).

In this sense, the notion of autonomy fits directly into the deontological perspective and therefore has a strong disciplinary history within this normative model. However, “it would be a mistake to consider that autonomy counts only from this perspective” ( Mauron, 1996, p.1). Thus, for example, classical utilitarian theorists insist on the characteristic of autonomy as the foundation of personal freedoms. For John Stuart Mill, autonomy is the capacity of individuals to make rational and considered decisions according to their interests and values. Mill advocated that everyone should be free to pursue their own happiness as long as it did not harm others. Autonomy is, as such, an essential requirement for the foundation of a free and fair society, as autonomy encourages individuals to thrive and contribute meaningfully to society ( Mill, 1859). Mill defended that governments and other institutions should respect individual autonomy and should only intervene to protect people from harm caused by the actions of others. Often quoted when discussing minimalism in ethics , Mill defended, in his work On Liberty, that the only legitimate reason a community can use force against one of its members, against its own will, is to prevent harm from being done to others ( Mill, 1859). Compelling someone for their own good, whether physical or moral, does not provide sufficient justification. One cannot oblige someone to either act or abstain from a decision under the pretext that it would be better for them or would make them happier, even if, in the opinion of others, it would be wise or even right to do so ( Mill, 1859). While these may be good reasons for reasoning, persuading, or begging them, it is not a legitimate enough reason to not coerce or punish them should they act that way ultimately ( Mill, 1859). When the decision concerns the person themselves, and solely them, their independence and autonomy are absolute, states Mill. As such, the individual is sovereign over himself, his own body and his own mind ( Mill, 1859). Autonomy is, therefore, a fundamental principle of individual freedom and personal responsibility, which should be protected and encouraged in all spheres of human life ( Mill, 1859).

While autonomy is a principle of individual freedom, freedom itself should be investigated further. Isaiah Berlin often addressed the issue of freedom, particularly in Two Concepts of Liberty. Berlin considered autonomy as the enactment of a person to control their own life and make free and informed decisions. It is a form of positive freedom, which consists of the ability to realise oneself as an autonomous individual ( Berlin, 2002). Berlin also stresses that autonomy presupposes the possibility of choosing among different, numerous options rather than being limited to a single possibility imposed by any form of authority ( Berlin, 2002). Autonomy, in this sense, resembles pluralism, which serves both as a universal moral principle and sustains freedom of choice, no matter what the person’s cultural, historical, social or political context ( Haidt, 2012).

Such foundational literature therefore informs the view that autonomy serves as the basis for human rights ( Mill, 1859), inherently encompassed within human nature ( Kant, 1785). In this sense, autonomy is a distinctive characteristic of the human person and the manifestation of human dignity ( Kant, 1785).

2.2.1 An examination of autonomy’s history

Autonomy gradually left philosophy to touch on other disciplinary fields – passing from political philosophy, to law, then profoundly overhauled in bioethics, still an emerging discipline ( Watkin and Davis, 2022). These changes have led to the qualification of autonomy as a principle: a general theorem or law that has numerous applications across disciplines, and that serves as the foundation for a system of belief. These changes were: •

In the social and political context. In North America, particularly in the 1960s and 1970s, individual autonomy entered and was the centre of most political claims. Movements for civil rights, women’s rights, and patients’ rights all emphasised the importance of individual autonomy in decision-making and the general exercise of freedom ( Bazzano et al., 2021).

The Belmont Report, written by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research after Tuskegee, was the first occasion on which autonomy was defined as a bioethics principle, alongside the principles of beneficence and justice ( Kristinsson, 2009). It defined autonomy as the right of individuals to make informed and free decisions about their own lives, particularly their health. The report stresses the importance of respecting the autonomy of individuals to promote their well-being and dignity as human beings. Concretely, the Belmont Report specifies the characteristics of autonomy, including that individuals must be informed of the possible risks and benefits of a medical or research intervention, as well as all available alternatives. The Belmont report also emphasises that respect for autonomy is not an absolute right for patients to do what they want. Healthcare professionals and researchers must also adhere, to balance autonomy, to the principles of beneficence and justice, as well as established laws and ethical standards.

One of the Belmont Report’s authors, Tom Beauchamp, further developed these principles ( Nagai et al., 2022, p. 159). In Principles of Biomedical Ethics, first published in 1979 and co-written with James Childress, such a principle-based approach, or principlism, aimed at providing a systematic structure for analysing and resolving ethical issues in medicine. As a result, the four principles of principlism – autonomy, non-maleficence, beneficence and justice – were developed to respond to the ethical challenges posed by technological medical advances ( Beauchamp and Childress, 1979).

While the principle of autonomy was developed to respond to the need to respect the right of patients to make informed and free decisions concerning their health, the principle of non-maleficence (which dates from ancient times: as the primum non nocere principle) was developed to ensure that healthcare professionals do not harm patients. Likewise, the principle of beneficence was developed to ensure that healthcare professionals do what is in the interest of patients’ health, while the principle of justice, broadly, attempts to ensure that healthcare resources are distributed fairly ( Beauchamp and Childress, 1979).

Principlism, qualified as “the most popular work” in bioethics, “played a pivotal role in establishing and shaping” autonomy and its construction in bioethics ( Shea, 2020, p. 442).

2.2.2 Contemporary understandings of autonomy

Generally, and beyond principlism, the sense of autonomy that the bioethics community has embraced resembles Berlin’s concept of positive liberty rather than negative liberty; autonomy is the freedom to act (positive liberty) rather than freedom from interference (negative liberty). While Berlin’s account of positive liberty is seen as shaped by Kant, autonomy is presently considered in bioethics as quite distinct from the Kantian perspective ( Jennings, 2016). Indeed, while Kant’s conception focused on logical formalism and severe rationalism and, most importantly, the rejection of all forms of biased motivations, autonomy is now synonymous with living according to one’s values and principles and protecting informed consent ( Jennings, 2016). Autonomy presently is solely to be self-sovereign and does not account for moral ramifications such as the categorical imperative that Kant considered fundamental ( Jennings, 2016).

The most notable difference between Kantian autonomy and the ‘default conception’ of autonomy currently embedded in bioethics, sometimes called the ‘empirical conception’ ( Takala, 2001), is that Kantian autonomy is overly imbued with rationalism ( Warburton, 1999). Rationalism divides a person into a higher and lower self, in which the higher is associated with reason and the lower with passions, desires, and a-rational aspects. The ‘default’ conception in bioethics stresses that autonomy mainly refers to acting freely, whether based on rational or emotional considerations ( Takala, 2001).

Although this conception of autonomy shares particular points in common with Mill’s, it also differs in several aspects ( Stirrat and Gill, 2005). First, Mill was primarily concerned that autonomy served as the right of each individual to pursue their own happiness, whereas the principle of autonomy in bioethics is primarily centred on individual health-based decisions. Furthermore, the principle of autonomy in bioethics is often limited by other ethical principles, such as the principles of non-maleficence and beneficence, which require health professionals to act in the patient’s health interest and avoid causing harm. Thus, while Mill was primarily concerned with individual liberty and happiness, the principle of autonomy in bioethics is typically balanced with other ethical considerations.

Whether autonomy should be balanced is the core of the disagreement in the literature, with some advocating that autonomy is paramount, others seeing it as inter pares. The below proposed methodology needs to consider the diversity of viewpoints in the literature.

This review adopts a scoping review methodology, as defined by Arksey and O’Malley (2005) and expanded by Levac et al. (2010), which is appropriate for mapping broad, conceptual questions and identifying gaps in heterogeneous literature. The primary objective is to examine how the concept of autonomy is defined, debated, and potentially reconfigured across disciplines, particularly within bioethics. The review also explores whether scholarly consensus exists or whether a fragmentation of views signals an epistemic or normative gap.

No formal review protocol was registered prior to the initiation of the review. Consequently, no pre-registered protocol is publicly available. However, a PRISMA-compliant flow diagram summarising the number of records identified, screened, excluded, and included is provided in the online repository referenced in the Data Availability section.

The search was conducted using primarily Google Scholar, as it aggregates results from various databases and disciplines, making it particularly suited for interdisciplinary inquiries. Additional searches were performed via Elsevier, Taylor & Francis, and ResearchGate. The final search was executed in April 2023.

Sources considered included peer-reviewed journal articles and book chapters from philosophy, ethics, bioethics, law, psychology, and behavioural economics. The inclusion of multiple disciplines aimed to ensure conceptual saturation and epistemic diversity.

Inclusion criteria were as follows: •

Publications discussing the concept, critique, or application of autonomy, with a focus on normative and philosophical aspects.

Exclusion criteria included: •

Publications that mention autonomy only in passing, without conceptual analysis (n = 12).

Articles meeting the inclusion criteria but with limited focus on autonomy were retained separately as background literature but excluded from detailed charting.

The primary search strings were combinations of: •

“principle of autonomy” with or without “bioethics”

Searches began with general terms and were iteratively refined as additional key terms emerged inductively from initial results. For example, “relational autonomy” and “narrative autonomy” were integrated after appearing frequently in early articles.

Titles and abstracts were manually screened for relevance, and the initial keyword searches retrieved 110 documents. Full texts were reviewed when titles or abstracts indicated possible inclusion. This process eliminated 33 records based on the exclusion criteria given above. The final number of sources included was 72, and can be consulted in the Data Availability section below.

The selection process was conducted manually. Initial inclusion was based on title and abstract screening, followed by full-text assessment. A visual representation of the process, consistent with the PRISMA-ScR format, is included in the flow diagram.

No calibrated or pre-tested forms were used. Instead, a manual charting strategy was adopted, aligned with the research goals. For each article, the main arguments, relevant definitions, and notable passages were documented and written down. The process was conducted independently by the authors, without duplicate screening. Recurrent themes and arguments were noted to identify patterns and epistemic tensions.

The following data items were extracted: •

Definitions of autonomy (e.g., relational, narrative, procedural)

As this is a scoping review aimed at mapping the conceptual landscape rather than evaluating empirical effects, no formal critical appraisal of the quality of individual sources was conducted. The inclusion of diverse theoretical positions was intentional to reflect the complexity of the topic.

The review produced four thematic clusters: 1.

General theoretical discussions of autonomy, particularly critiques of the principle.

Thematic saturation indicated recurrent tensions between autonomy as a universal moral principle and as a context-sensitive ethical construct. Relational and narrative conceptions often provided critiques of traditional liberal notions of autonomy, proposing alternatives that better account for vulnerability, social embeddedness, and cultural specificity.

It is important to recall that autonomy, in its original formulation and in the way it is discussed in the existing literature, is a principle. A principle is a moral rule, or general norm that is intended to guide moral conduct ( Ravitsky, 2022). A principle, therefore, aims to ease and guide the justifications of ethical choices: if one acts this way rather than another in such a situation, on what grounds could someone say that such a decision is morally right? Justifying moral actions often requires calling on a principle, a guiding maxim that justifies moral conduct. Principles, however, are not absolute, nor were they formulated to be absolute ( Mauron, 1996; Ravitsky, 2022). They were formulated as approximate rules on how to act, and in no way as “mechanically applicable geometric axioms” ( Ravitsky, 2022).

This distinction is essential, as it seems to be missed by most of the literature that criticises the principle of autonomy: while sceptical scholars do define autonomy as a principle, they tend to neglect the reality that it should be operationalised and considered as one. Childress (1990) notably argues that the criticisms stem from misunderstandings and misinterpretations, often resulting from flawed formulations from autonomy’s defenders. As a result, critics who attack the concept of autonomy inevitably end up criticising particular conceptions of autonomy, “often the least defensible ones” ( Childress, 1990). Another mistake is to fail to appreciate that ‘autonomy’ is a shorthand expression for the original enunciation; “the principle of respect for autonomy”. As such, the literature tends to forget, first, that autonomy is a principle, and, secondly, that what is at issue is about ‘respecting’ autonomy, not necessarily “guaranteeing”, “ensuring”, or “protecting” it. “Respect for autonomy” implies a lighter responsibility and is a less restrictive principle than the literature on autonomy may suggest, as it is far less absolute in scope ( Childress, 1990). It is possible to respect autonomy, in the sense of listening and respectfully considering a person’s articulated will, while still emphasising the need for medical care to which a person may not consent. In the existing literature, many articles (e.g., Gillon, 2003; Stirrat and Gill, 2005) assume that respect for autonomy is a synonym for prioritising autonomy. Autonomy is not, and should not, be granted absolute authority during the care process or over the other principles.

Principles, since they are approximate, can, therefore, conflict with one another. Such conflict may stem from a disciplinary divide between the principles, notably regarding autonomy and beneficence, for example. Epistemologically, autonomy refers to philosophy and politics, discussed foundationally in these disciplines by Kant, Mill, and Berlin ( Bourret et al., 2015). Beneficence stems more from the disciplines of medical ethics, care, and deontology, and was first discussed by Hippocrates ( Bourret et al., 2015). These two principles have different epistemic histories, and the interdisciplinary contribution made by different disciplines is still visible today: autonomy has ramifications with comparable concepts in other disciplines, such as self-determination, dignity, freedom and rights. Beneficence is sometimes called benevolence, sometimes confused with non-maleficence … much of this depends on our epistemic differences and disciplinary background.

Concretely, these differences are not solely confined to terminology: autonomy represents the core of an “individualistic sphere”, while beneficence is entirely relational ( Bourret et al., 2015, p.91). While respect for autonomy requires a particular relationship between doctors and patients, beneficence in its application is more straightforward. Respect for autonomy involves careful attention throughout the entirety of care to ensure that the patient’s autonomy is not infringed on, while beneficence focuses on the end goal: determining if the care provided is overall beneficial to the patient ( Bourret et al., 2015). As such, while autonomy and beneficence are interconnected, the relationship between the two principles is unbalanced and unequal. A doctor who wants to guarantee the principle of beneficence by providing care without respecting a patient’s will is likely to fall into a form of paternalism. These epistemic differences, stemming from the diverse disciplines from which the principles emerged, may thus make the principles conflict.

These differences are even exacerbated, as the literature rightfully warns about a subtle ‘hierarchy of the principles’ ( Bourret et al., 2015). The literature often mentions a dominant, though not clearly admitted, hierarchy among the bioethical principles, which usually tends to put autonomy first ( Bourret et al., 2015). This hierarchy may historically stem from philosophy and the Kant’s emphasis on how autonomy is essential, as it is what is typically held to differentiate us from (other) animals ( Bourret et al., 2015). The hierarchy can also result from the principle of beneficence’s original formulation, suggesting that beneficence acts more as a safeguard, an obligation to do good, rather than a principle. Beneficence, in medical ethics, is also strongly related to a form of paternalism, implicitly degrading its perceived value, which may help explain why autonomy, the antagonist of paternalism, comes at the top of this presupposed hierarchy in modern ethics ( Bourret et al., 2015).

The non-consensual nature of autonomy may stem from the emerging criticism of medical paternalism that overemphasises autonomy as a solution ( Rendtorff, 2019). Several European countries, including Italy, Spain, Greece, Ireland, Portugal, Germany, and Sweden, value dignity and integrity as human rights and explicitly define them in their constitutions. In contrast, other countries, including Denmark, the United Kingdom, Norway, Austria, and France, are more focused on autonomy ( Rendtorff, 2019). Varied interpretations of autonomy across different countries also account for the bioethical divide around autonomy. Some ethical committees, such as the French National Consultative Ethical Board, even explicitly define autonomy as a principle that has “priority over other ethical principles” ( Bourret et al., 2015, p.97). Such a statement is troublesome as it confirms the concerns expressed in the literature. Most of the committees, however, recognise that respect for individual autonomy should be balanced with respect for the principle of justice, which can justify placing limitations on autonomy.

Limiting autonomy could therefore be ethically justified if autonomy conflicts with the principle of justice. For example, it can be argued that mandatory vaccination for COVID-19 is ethical, as putting others at risk by refusing vaccination would infringe on the principle of justice ( Bowen, 2020). Furthermore, the principle of justice encourages authorities to proceed with mandatory vaccination, as such efforts will benefit all, while protecting those most vulnerable to the virus, such as the immunosuppressed and those with legitimate medical contraindications ( Bowen, 2020). Such examples demonstrate that restricting autonomy to ensure the respect of another principle is ethically defensible. However, this balancing process is still insufficient, as it neglects the two other principles: beneficence, and non-maleficence. In principlist theory, all four principles need to be considered, and autonomy could and should therefore be balanced with beneficence and non-maleficence as well ( Ravitsky, 2022). While most committees recognise this autonomy-justice exception as an “inherent paradox”, they conclude, without much detail, that limiting autonomy solely through the principle of justice is the best compromise found yet ( Bourret et al., 2015).

However, interestingly, from a juridical point of view, this compromise seems far better achieved. In complex cases where autonomy conflicts with other principles, notably for young children or teenagers, the notion, in some countries, of Gillick competence, and being made a ward of court, suggests a more balanced account of autonomy and beneficence ( Bourret et al., 2015). Gillick competence exists when a young person is presumed to have sufficient capacity to be autonomous, in the sense that they have enough intelligence, competence, and understanding to formulate a decision ( Griffith, 2016), for the particular matter in question – so that, for instance, an eight-year old might be deemed competent to decide on what to spend their pocket money but not whether or not, on their own, to consent to an operation. When autonomy conflicts with beneficence, such as when refusing treatment, Gillick competence can be overruled, and the Court of Protection in England and Wales can oversee the decision ( Bourret et al., 2015). In this case, it can be decided to prioritise beneficence over autonomy. This demonstrates that it is possible, indeed, crucial, to question such a prevalent and implicit hierarchy and the pre-assumed superiority of autonomy in the literature, notably when scholars tend to see autonomy as the source of “individualistic whims” ( Bourret et al., 2015).

On that note, the literature critiquing autonomy (e.g., Clouser and Gert, 1990) tends to forget that when autonomy conflicts with another principle, such as beneficence, non-maleficence, or justice, the principlist approach can offer solutions. In the principlist approach, the conflicting principles can either be conciliated by elaborating a compromise, or by deciding which principle will have priority by arbitration. Part of the literature considers this process to be too relativistic: the principles only make sense in a particular context ( Takala, 2001). Such an objection is legitimate, as it is impossible to construct abstractly, or a priori, a hierarchy of the principles, nor is it possible to arbitrate theoretically: arbitration requires a particular, concrete, situation ( Ravitsky, 2022). This limitation was, notably, addressed by Beauchamp and Childress ( Durand, 1999). Compromise or arbitration requires weighting the principles, in a process defined as “ponderation” ( Ravitsky, 2022). Such evaluation of the weight or the strength of each principle is not, however, a subjective or purely intuitive process: it requires formulating ethically justified reasons to compromise or arbitrate the principles ( Ravitsky, 2022). However, the source of disagreement regarding principlism is how and what those ethically justified reasons could be. To ease tensions on the nature of autonomy and on the principles, it is thus essential to enunciate the conditions in which pondering the principles is justified, as to find a consensus that the literature could agree on, such as: 1.

The chosen prioritised principle offers better ethical reasons to act than the violated one.

Concretely, with the following conditions listed above, finding common ground across the principles often involves considering a principle in the light of another – such as the imperative of beneficence in the light of the principle of autonomy ( Ravitsky, 2022) . Such an exercise requires considering what is beneficent by examining the person’s belief system. Some of the literature has also developed this aspect in an argument called “respect for narrative autonomy” ( Asveld, 2008). The imperative of beneficence in the light of the principle of autonomy consists in considering autonomous decisions through the prism of identity and beliefs: a religious person acts autonomously by dedicating their life to God, and may believe themselves to be punished by God if they do not respect those beliefs. By infringing on the person’s autonomy, such as by forcing them to get vaccinated or to receive a blood transfusion if they do not wish this, the person may believe that they will be punished for their sin; they may even be barred from eternal salvation ( Thompson, 1989). Considering these elements, the intervention is, therefore, from the perspective of the person who is meant to be benefitting, far from beneficial. In this sense, it may be best to respect autonomy, as, in light of subjective beneficence, the most beneficial action is not to proceed with either the blood transfusion or the vaccination ( Asveld, 2008). As such, an adult patient with proper decision-making capacity can refuse a blood transfusion based on religious convictions, and such refusal should (will often) be respected (cf. Thompson, 1989). However, it seems that considering a principle in the light of another principle, as discussed above, is successful when confined to a case-by-case basis, or, more particularly, to the individualistic sphere. Refusing a blood transfusion cannot harm the unrelated collective, or society in general. Refusing vaccination, on the other hand, can harm others by increasing the numbers who become infected. When what is at stake has ramifications for the collective, such as in vaccination, it is less straightforward to determine what is best, and further criteria are required.

Narrative autonomy attempts to list those requirements. First, claims against vaccination should genuinely correspond to the individual’s identity: the reason for refusal must be credible and sincere ( Asveld, 2008). Secondly, demanding the right to refuse vaccination should fit a “shared sense of reality within a society” ( Asveld, 2008, p.256). Refusing vaccination based on an unrealistic or uninformed risk perception cannot fit this shared sense of reality. Considering these criteria, if we were to extend autonomy as a freedom to not get vaccinated, then only those who cannot receive vaccination based on sincere religious or other convictions, including ones to do with their health, should have their vaccination views respected ( Asveld, 2008). The religious claim is seen as adequate as it is both deeply tied to the personal sense of identity as a religious person, and, arguably, is not based upon “unsubstantiated risk perception”, contrary to some anti-vaccination claims ( Asveld, 2008).

While such permissibility granted by the principle of autonomy may be a core disagreement in the literature, such issues around permissibility were discussed when the principle was first enunciated ( Durand, 1999). A person can, indeed, autonomously drift away from their first-order decisions, as in what is objectively good for them based on individual preferences (such as religious ones) and, in accomplishing so, practise what may seem to be Kantian heteronomy (the antagonist of Kantian autonomy). This decision may however result from second-order autonomy. Second-order autonomy refers to the capacity not only to make decisions in alignment with one’s immediate desires or preferences (first-order autonomy) but also to reflect critically on, revise, and potentially override those desires based on higher-order values, beliefs, or goals. Essentially, it implies an additional layer of self-governance where individuals can evaluate their own motives and preferences and make decisions that align with their more deeply held values or life plans, rather than just reacting to their immediate impulses ( Loughrey, 1998). For example, a patient may need a vital blood transfusion (which is a first-order need) but will renounce such a first order need to fulfil a second-order desire: complying with religious convictions. Second-order autonomy is, in this case, still an expression of the rational capacity of a person ( Loughrey, 1998). Childress has always advocated that respect for second-order autonomy is central, even though those first-order choices may be seen as heteronomous ( Childress, 1990).

What are therefore the ‘structural limitations’ ( Rendtorff, 2019) of autonomy that are discussed in the literature? In effect, principlism may be particularly relevant to inform what is at stake, and which principles can be called upon to navigate a conflict. However, as rightfully stated in the literature, it cannot offer a concrete methodology on how to act or how to solve a conflict. However, principlism never aimed to compete as a normative theory ( Mauron, 1996), and such argument seems to be the most prevalent misconception of principlism in the literature. Principlism should be seen, as rightfully qualified by its detractors, as a paradigm, and should be considered and critiqued as such, and not as a normative theory. In the pyramid of normative ethics tools, principlism should be seen as sitting between, on the one hand, pivotal cases and examples that provided empirical ethical insights and, on the other, normative theories such as consequentialism and deontology that deductively inform moral conduct ( Ravitsky, 2022). Principlism stands in the middle, as a selected assemblage of different normative elements (among principles, values, norms or maxims), that all originated inductively, with a capacity to provide partial guidance when applied deductively ( Ravitsky, 2022). Bioethicists had to develop principles when faced with past scandals and crimes in history: respect for autonomy was notably introduced in the Nuremberg Code and the Belmont Report ( British Medical Journal, 1996; National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). Elaborating guarantees and rules after these events involved analysing inductively, based on the injustice committed, which principle could best protect participants and patients in the future. Those principles were inductively formulated, later to be deductively operationalised. Contrary to utilitarianism or deontology that settled on idealistic, universal maxims (maximise utility, respect rights and duties), principlism, through observation of empirical cases, settled on four principles that seemed most adequate to approximately guide moral conduct ( Beauchamp and Childress, 1979).

This hierarchal position of principlism in the scale of normative tools can thus indeed present principlism as either too weak on the theoretical level or too little anchored in particular situations – and such limitation applies to the principle of autonomy as well. Strengthening autonomy beyond these limitations requires constructing new formulations of autonomy that may find consensus within the literature.

Relational autonomy can be summarised as the acknowledgement that interpersonal relationships are essential to understand autonomy, as these relationships inform decision-making ( Brown and Browne, 2022). Relational autonomy emphasises that persons “are socially and historically embedded, not metaphysically isolated, and shaped by factors” such as race, religion, sexuality, gender, and social class ( Brown and Browne, 2022). As seen above, the traditional account of autonomy, granting individuals a self-governing capacity largely unaffected by or independent of their relations and consequences, problematises healthcare decisions in practice. The ‘shift’ to relational autonomy enables one to consider an individual autonomous decision through relational elements that inhibit or encourage such decision-making ( Brown and Browne, 2022). It is a more appropriate conception of autonomy, as healthcare often stands as a structure “of power and privilege” with consequences for presumptions about decision-making capacity ( Brown and Browne, 2022, p.11). Therefore, supporting autonomous decision-making requires recognising “all personal, relational, and structural factors affecting autonomy competencies or capacities” ( Brown and Browne, 2022, p.65).

The existing literature makes a strong case for the need for such a “relational turn” in bioethics to reform and improve general theoretical concepts such as liberty and autonomy. The literature advocates that the relational perspective represents a renewed emphasis on bringing practice into theory: rather than deductively discussing how autonomy should be operationalised in real-life contexts, relational autonomy seems to be a more inductive concept, offering the possibility of applying it in existing ethical dilemmas to shape the theory ( Gómez-Vírseda et al., 2019).

Relational autonomy is seen in the literature as a fitting concept to consolidate discussions around autonomy, as: 1.

It is a theoretical concept, stemming from a wide range of new interdisciplinary contexts — bioethics, feminist studies, health studies, nursing, law, psychology, education, and elderly care. All relational components that can inform decision-making, no matter the disciplinary point of view adopted, are valuable in the context of relational autonomy.

Concretely, relational autonomy is a more “civic and active participatory condition” ( Jennings, 2016, p.2), maintaining that individuals do not own or possess their autonomy; rather, they enact it by producing and reproducing choices that reflect the forms of life with which they want to engage. Relational autonomy should be achieved through an effective social order that sustains ‘practices of responsibility’, ensuring that autonomy and responsibility are always kept in a harmonious balance ( Jennings, 2016). As such, the concept of relational autonomy is distinct from the ‘attainment of an idealised moral self’ that Kant valued ( Jennings, 2016), though such a conception of autonomy should be permanently embedded in a social and political system that stresses the responsibility of individuals for their autonomy and the consequence it may have on others. This relational turn may solve the reported ‘excessive individualism’ that reportedly plagues the principle of autonomy.

However, a crucial distinction the literature emphasises is that while relational autonomy may indeed move away from individualism, this new conception does not aim to devalue individuality. This means that relational autonomy should not be opposed to the relevance of individuality; rather, it is what makes individuality possible ( Jennings, 2016). While such a distinction is still poorly conceptualised, one of the possible interpretations available is that individuality stems from the freedom to act alongside others, not the freedom to act against them ( Jennings, 2016) . This argument is also a core aspect of relational autonomy ( Jennings, 2016).

Relational autonomy potentially represents an exciting new approach for solving ethical dilemmas, such as mandatory vaccination for COVID-19, for example. However, there is a literature gap regarding the applications of relational autonomy to specific bioethical issues ( Jeffrey, 2020). While the literature agrees that relational autonomy provides an opportunity to widen discussions and offer a practical positive change in healthcare, there is an absence of a consideration of relational autonomy’s application in real-life contexts ( Jeffrey, 2020). How then to use the concept of relational autonomy in practice?

First, with relational autonomy, the classic argument that one should be morally permitted to refuse vaccination based on one’s autonomy no longer stands: individuals do not possess such inherent autonomy that trumps all other considerations, but should instead act autonomously based on a sense of responsibility, notably the responsibility not to endanger other’s health by refusing to get vaccinated. Relational autonomy thus makes it impossible to separate the individual from their relational context and grant absolute individual autonomy. Relational autonomy eases the ethical dilemma of mandatory vaccination, as it proposes that individuals should make informed autonomous choices “not in isolation, but with the support of others” ( Williamson and Glaab, 2018). The literature is somewhat divided regarding whether the role of others should be a mere consideration, or a form of active support, and articles often alternate between these two understandings. In any case, such ‘support’ should not be seen as a form of coercion but rather as a possibility to develop informed decision-making. Relational autonomy encourages a more systematic way of engaging in a discussion with anti-vaccination individuals to help provide the consideration they may lack to make a proper, informed, and thus autonomous decision. With the current autonomy model, such discussion is often neglected. To counter this, relational autonomy calls on particular obligations to be followed: 1.

Medical professionals, doctors, health institutions, and, more broadly, society, have a duty to inform about vaccines, what the process of vaccination involves, what the risks are, and the benefits.

A minimal part of the literature maintains that providing information about consent is too demanding for medical professionals, on the grounds that their role is not to educate patients. Such ‘duty’ to inform consent, both currently, and in relational autonomy, may require giving “too much information” ( Dive & Newson, 2018). More broadly, a small part of the literature is concerned with the paradigm of ‘informed’ consent, deemed excessive, and judges that simply ‘consent’ may be enough ( Dive & Newson, 2018). To this extent, scholars critique autonomy as a principle with a ‘scalar property’, in the sense that the more information is provided, the more a decision is deemed autonomous ( Dive & Newson, 2018). This ‘scalar’ property of autonomy poses a problem as ‘informing’ consent would involve providing as much information as possible, slowing down the care process. However, this conception seems misguided, as a patient can also consent to not knowing (such as, for example, in genomics investigation); demonstrating informed consent requires appropriate, timely and suitable information, and not necessarily the greatest amount of information possible ( Satyanarayana Rao, 2008). Relational autonomy should thus have a duty to inform with appropriate, timely and suitable information.

As relational autonomy may answer most objections identified in the literature, the principle of autonomy could effectively be strengthened by including such a relational component, both solidifying the principle’s use and application and reuniting what is done in practice with what is discussed in theory.

Even if relational autonomy is presented as the most consensual formulation yet existing, it is important to consider that some feminist critiques have rejected the notion of autonomy altogether, including what they define as the “reconfiguration” of autonomy “in terms of relational accounts” ( Brown and Browne, 2022). For feminist scholars, such as Martha Fineman, the root of the problem remains that our social contract has “over-emphasised individual liberty and freedom of action, even as that freedom has resulted in a diminishing of options and autonomy for many” ( Fineman, 2013, p.14). Fixing the principle of autonomy may then require a change of such a systemic “social contract” and of our institutions ( Fineman, 2013, p.14). Such a conclusion is shared with other disciplines, such as critical race theory ( Brown and Browne, 2022). Indeed, it has been argued that autonomy “has historically encompassed a Western individualist worldview”, having globally trumped the other three bioethics principles, particularly when mitigating them altogether, while in other cultures, beneficence, through notions of vulnerability and empathy, has always been emphasised ( Behrens, 2018). Such a rise in the power of individual freedom may not represent society’s will, and the way autonomy was initially considered and formulated in bioethics. Through its implementation as a principle, autonomy may have emerged as a positivist and pragmatist model, strengthening toxic individualism. Solving such a problem would presuppose the need for a certain dose of relativism, involving the consideration of bioethics as an ethnoethics or even as a cultural system, like any other system of values ( Bouffard, 2003).

A minimal literature segment has turned to other bold, new conceptions to counter such an inherently individualistic nature encompassed in autonomy. Recognising the paradoxical nature of autonomy that gives rise to “profoundly different action-guiding principles in healthcare” ( Sjöstrand et al., 2013), an emerging conception, ‘paternalism in the name of autonomy’, sometimes referred to as ‘autonomy-enhancing paternalism’, seems promising. While it recalls some critiques that were examined previously, such a new form of autonomy is more relevant as it does not call into question the value nor the relevance of autonomy as a principle: on the contrary, it tends to reinforce it by suggesting that, as a principle, it is indeed approximate, and thus should not act as an absolute right. Patients’ decisions should sometimes be “overruled to protect or promote their own autonomy” ( Komrad, 1983). The first argument to support such a claim is that illnesses inherently represent a state of diminished autonomy, and the care process therefore necessarily involves a degree of medical paternalism ( Komrad, 1983). Given such diminished autonomy, a part of the literature advocates that a limited form of paternalism is entirely acceptable, and perhaps even indispensable, particularly when such efforts aim to restore or maximise autonomy. Considering autonomy’s scalar aspect, applying a degree of paternalism should also be adjusted based on the individual’s current autonomy. Discussing the correct balance between paternalism and autonomy is therefore essential as: •

Autonomy and paternalism are inversely correlated.

While it may appear paradoxical to consolidate autonomy with its antagonist, both notions may not be as mutually exclusive as might be assumed. For a start, both paternalism and autonomy aim to promote the individual’s higher interest, though they often conflict on the means to get to this, and on the actual nature of that interest. The most common higher interest is often the improvement of the patient’s health, where, in one case, the patient decides for themselves, and in the other, a doctor decides for the patient.

What would the general literature think of autonomy-enhancing paternalism, and could it genuinely fill the literature gap? Paternalism in care is a profoundly divisive notion: paternalism can be considered as either unambiguously coercive or possibly instructive. However, among those who discussed autonomy historically, a consensus seemed to have been reached. Paternalism may be moral, for the “immature” ( Mill, 1859), for the “inherently non-autonomous” ( Beauchamp and Childress, 1979) and, perhaps remarkably, for Kant, for “those who do not culture reason” ( Kant, 1785). To generalise, paternalism is a legitimate response to all with a diminished autonomy, provided such paternalism aims ultimately to contribute to the restoration of autonomy. Paternalism’s telos, then, should be to maximise autonomy. In other words, “restitution of diminished autonomy is the only rationalisation of paternalism that does not profane autonomy” ( Komrad, 1983). Such a model for autonomy also reconciles the historical disparity between autonomy and paternalism, allowing them to co-exist throughout the care process. Such discussion is sufficient on the theoretical level, though requires further attention when it comes to defining what exactly are those paternalist interventions that can promote autonomy.

Evidence from behavioural economics and psychology suggests that autonomy can vary depending on two different behavioural ‘systems’. The first is an automatic, intuitive decision-making system, “not always characterised by full rationality, perfect information processing, and complete self-control” ( Binder and Lades, 2015). The second system, the reflective system, is responsible for careful observation and analytics, fostering rational decisions. Consolidating autonomy therefore suggests that, through ‘behavioural interventions’, the reflective system could be strengthened while the intuitive one could be softened. The aim is not to eradicate intuition or emotive decision-making but rather to promote self-empowerment and free individuals from irrelevant (or inappropriate) influences ( Binder and Lades, 2015): “Behaviourally informed paternalism lies in its intention to help individuals overcome their biases and decision-making fallibilities and help them in making better thought-through, autonomous choices” ( Binder and Lades, 2015, p.5).

Proceeding with paternalism to enhance autonomy is therefore a similar task to de-biasing strategies, aiming to improve decision-making processes and outcomes. Such “behavioural interventions” are thus processes to simplify, de-biase, or reframe some choices or complex chunks of information with which individuals are confronted ( Binder and Lades, 2015, p.5). In the case of vaccination, it is about guiding towards legitimate scientific evidence, uncovering why there is a concern with vaccination in the first place, and an attempt to re-centre the discussion and ignite critical thinking. The literature recognises, though, that “discussing on what grounds policy interventions are justified or not” is a demanding and intricate task ( Binder and Lades, 2015, p.5). Autonomy-enhancing paternalism insists on interventions that help individuals do the right thing for the right, unbiased, reasons. These reasons must still come from a personal reflection but are motivated by more explicit, factual rationales. The literature supporting autonomy-enhancing paternalism claims that paternalism is sometimes needed to foster sovereign individuals, to nullify ill-intentioned manipulation, and that, through, again, appropriate, timely and suitable behavioural interventions, paternalism will no longer be required in the future: autonomy-enhancing paternalism “counteracts slippery slope arguments by decreasing the probability of future paternalistic interventions” ( Binder and Lades, 2015, p.5). However, such an account seems somewhat speculative.

The largest problem with the ‘paternalism in favour of autonomy’ model is that, once again, autonomy is here largely considered as a synonym for capability, which is not a consensual characteristic of autonomy in the literature. The literature advocating autonomy-enhancing paternalism intervention describes illnesses as an unequivocal infringement on the capacity to act autonomously: pain, uncertainty of the consequences, limited medical knowledge, or loss of biological function inherently “diminish autonomy” ( Komrad, 1983). However, for the literature that considers autonomy as an ideal or a right, a patient can still act autonomously even under the influence of pain or other externalities, and such an account seems reasonably legitimate. The second problem is that paternalism in favour of autonomy is a prominent deficit model, considering the patient solely as a receiver of undisputed medical information.

Paternalism in favour of autonomy, while rectifying the individualistic aspect of autonomy, may therefore not be the most fitting improvement to refine autonomy. On the contrary, as paternalism in favour of autonomy suggests imposing appropriate, timely and suitable information, would it not be wiser to entrust this task to relational autonomy, less paternalistic and more centred on duties rather than speculative interventions?

Among the three conceptual additions examined – narrative autonomy, relational autonomy, and autonomy-enhancing paternalism – some have proven to be more consensual than others. While narrative autonomy is considered too permissive, autonomy-enhancing paternalism is, oppositely, considered too invasive, even reductive as a deficit model. Relational autonomy, although imperfect in certain aspects (the vagueness of duties, and the more subtle deficit model that can stem from it), seems much more consensual than the two other conceptions examined above.