The Principle of Non-Refoulement

Introduction

The Principle of Non-Refoulement is one of the most fundamental safeguards in contemporary public international law because it imposes a clear legal limit on the sovereign power of States to remove individuals from their territory, borders, jurisdiction, or effective control. At its core, the principle prohibits a State from expelling, returning, extraditing, transferring, rejecting at the frontier, or otherwise removing a person to a country or territory where there are substantial grounds for believing that the individual would face persecution, torture, cruel, inhuman or degrading treatment, arbitrary deprivation of life, enforced disappearance, or other forms of irreparable harm.

The rule is preventive in nature: it seeks to avert the occurrence of serious violations before they materialise, rather than merely providing remedies after the harm has occurred. For that reason, it has become one of the most operationally significant principles in refugee law, international human rights law, and broader international protection regimes.

The modern legal articulation of the principle is most prominently associated with Article 33(1) of the 1951 Convention Relating to the Status of Refugees, which prohibits States from returning a refugee “in any manner whatsoever” to territories where his or her life or freedom would be threatened on account of race, religion, nationality, membership of a particular social group, or political opinion (UNHCR, 1951). This provision established non-refoulement as the normative cornerstone of international refugee protection. Its significance, however, extends beyond the refugee framework.

Over the past decades, the principle has evolved into a broader legal doctrine that operates across multiple treaty systems, particularly through the absolute prohibition on removal to a risk of torture under Article 3 of the Convention against Torture and through the protection of life and physical integrity under international human rights instruments (UN, 1984; Human Rights Committee, 2018).

The legal importance of the Principle of Non-Refoulement lies in the fact that it gives concrete effect to other substantive rights recognised under international law. Rights such as the right to life, the prohibition of torture, and the right to seek asylum would be severely weakened if a State could simply avoid responsibility by removing the individual before the threatened violation occurs. The principle, therefore, functions as a bridge between substantive rights and State conduct. It transforms international legal protection from an abstract normative promise into a direct legal restraint on State action. Without it, the legal architecture of refugee and human rights protection would remain structurally incomplete.

From a doctrinal standpoint, the principle is no longer confined to the formal status of “refugee.” Contemporary international law increasingly adopts a risk-based rather than status-based approach. This means that the decisive legal question is not solely whether the person has already been recognised as a refugee by a competent authority, but whether credible evidence establishes a real risk of serious harm upon return. This development has been especially reinforced by the jurisprudence of international and regional human rights bodies, which have clarified that protection against removal may extend to asylum seekers, irregular migrants, stateless persons, detainees, and other non-citizens regardless of their formal legal status (Goodwin-Gill and McAdam, 2021).

The principle has also undergone substantial expansion in terms of territorial reach. Historically, debates have often centred on removal from within the territorial boundaries of a State. Modern legal practice, however, has made clear that non-refoulement may also apply at borders, in airport transit zones, during maritime interception, and in situations where a State exercises effective control outside its territory. This extraterritorial dimension has become particularly relevant in the context of migration control policies, pushbacks at sea, offshore processing arrangements, and return agreements with third States. These developments have generated some of the most significant contemporary legal controversies in the field, particularly concerning the extent to which States may externalise migration management without breaching their international obligations.

Another reason why The Principle of Non-Refoulement remains one of the most dynamic areas of international law is the doctrinal tension between its different legal branches. In refugee law, the principle is subject to narrowly framed exceptions, particularly where the individual constitutes a danger to national security or has been convicted of a particularly serious crime. By contrast, under international human rights law, especially in relation to torture and inhuman treatment, the protection is generally understood as absolute. This distinction is of considerable legal importance because it requires careful separation between treaty regimes that are often discussed together but do not always produce identical legal consequences.

The principle is equally central to contemporary debates on forced migration, counter-terrorism, climate-related displacement, and transnational human rights protection. Questions surrounding safe third country agreements, diplomatic assurances, collective expulsions, and returns to zones affected by armed conflict have intensified both scholarly and judicial attention. These developments demonstrate that the principle is not a static rule inherited from the post-war legal order, but an evolving doctrine continuously shaped by judicial interpretation, State practice, and changing patterns of human mobility.

This article examines The Principle of Non-Refoulement as a layered and evolving doctrine under public international law. It will analyse its treaty foundations, customary status, material scope, territorial application, procedural requirements, and doctrinal controversies, with particular attention to the distinction between refugee law and human rights law approaches. The objective is not merely to define the principle in abstract terms, but to explain its legal operation in practice, the reasoning of courts and treaty bodies, and the challenges posed by contemporary State practices of migration control and externalisation. Through this approach, the article aims to provide a substantive and doctrinally rigorous understanding of one of the most consequential principles in international legal protection.

1. Concept, Function, and Legal Structure

1.1 Meaning of non-refoulement

The legal meaning of non-refoulement must be approached broadly and functionally, rather than through the narrow categories of domestic immigration law. At its core, the concept prohibits a State from expelling, returning, extraditing, surrendering, transferring, rejecting at the frontier, intercepting, or otherwise removing an individual to a place where that person faces a real risk of serious harm. The breadth of this formulation is deliberate. International law is concerned not with the label attached to the State’s action, but with the practical consequence of exposing an individual to persecution, torture, or other irreparable violations.

The classical treaty formulation appears in Article 33(1) of the 1951 Convention Relating to the Status of Refugees, which provides that no Contracting State shall “expel or return (‘refouler’)” a refugee to territories where his or her life or freedom would be threatened (UNHCR, 1951). The phrase “in any manner whatsoever” is legally decisive. It demonstrates that the prohibition is not confined to formal deportation orders following completed administrative proceedings. Rather, the rule captures any conduct by which a State causes an individual to be placed in danger.

This broader interpretation has significant doctrinal consequences. Expulsion traditionally refers to the removal of a person already present within the territory or jurisdiction of the State, generally by administrative or judicial order. Return, however, is wider in scope and may include refusal of entry combined with physical redirection to another country. International law does not permit States to avoid responsibility merely because the individual has not formally crossed the border. If the person is under the effective control of the State and is sent back into danger, the obligation may still arise.

The principle also applies beyond migration control in the strict sense. Extradition and surrender proceedings in the context of criminal cooperation are equally subject to the prohibition where the requested individual faces a real risk of torture, politically motivated persecution, or arbitrary deprivation of life in the receiving State. The European Court of Human Rights made this clear in Soering v United Kingdom (1989), where the Court held that extradition could violate Article 3 of the European Convention on Human Rights if the individual faced inhuman treatment in the destination State.

Equally important is the contemporary application of the rule to maritime interception and transfer operations. Modern border enforcement frequently occurs at sea, through interceptions, pushbacks, and disembarkation arrangements with third States. Jurisprudence has made clear that the legal obligation may arise in such contexts where the intercepting State exercises authority and effective control over those concerned, as affirmed in Hirsi Jamaa and Others v Italy (2012).

The prohibition is not limited to direct transfer. It also extends to indirect or chain refoulement, where a State transfers a person to an intermediary country that may later send the individual onward to a territory of persecution, torture, or grave harm. In legal terms, responsibility may attach where such onward removal is foreseeable.

The decisive point is that the rule addresses substance rather than form. A State cannot evade its international obligations by altering the terminology of removal. It is irrelevant whether the action is described domestically as a relocation, return assistance measure, safe third-country transfer, or border management operation. What matters is whether the State’s conduct foreseeably exposes the person to a prohibited risk.

1.2 Why the principle is foundational

The foundational nature of non-refoulement lies in its function as the operational core of international protection. It is the mechanism that gives substantive rights practical legal force by preventing States from defeating them through removal.

Without this safeguard, the architecture of international protection would be structurally incomplete. A State could formally recognise the prohibition of torture, the right to life, or refugee status while simultaneously avoiding responsibility by transferring the person to another jurisdiction before the threatened harm occurs. The result would be a legal framework that is normatively ambitious but practically ineffective.

This is particularly evident in refugee law. Recognition of refugee status derives its legal meaning from the assurance that the protected individual will not be returned to the danger that justified protection in the first place. The status itself would lose much of its practical significance if the State remained free to remove the individual immediately after recognition. This is why the doctrine has long been described as the cornerstone of international refugee protection (Goodwin-Gill and McAdam, 2021).

The same logic applies in human rights law. The prohibition of torture under the Convention against Torture would be substantially weakened if a State could knowingly extradite or deport an individual to a place where torture is likely. Similarly, the right to life under the International Covenant on Civil and Political Rights would become largely ineffective if States could transfer persons to territories where there is a foreseeable risk of arbitrary execution or exposure to life-threatening violence.

Its foundational role also extends to procedure. Asylum procedures and risk assessments exist because wrongful removal may lead to irreversible harm. The procedural dimension is therefore inseparable from the substantive guarantee. Summary removals, accelerated border procedures without effective review, and collective expulsions raise serious legal concerns precisely because they increase the risk of erroneous transfer.

At a deeper doctrinal level, the principle functions as a direct limitation on sovereignty. It confirms that State authority over admission, borders, and expulsion is not absolute. Sovereign discretion in migration control must operate within the framework of international obligations owed to individuals under a State’s jurisdiction or effective control.

For this reason, the principle is foundational not only because of what it protects, but because it defines the limits of lawful State conduct in one of the most sensitive areas of sovereign power.

1.3 Rule, principle, or composite norm

A central doctrinal question is whether non-refoulement should be understood as a single trans-substantive norm applicable uniformly across international law, or as a family of related but distinct legal prohibitions. The stronger and more accurate view is the latter.

Although legal writing often refers to it as if it were one unified rule, treaty structure and judicial practice demonstrate that it is better understood as a composite norm composed of several related but distinct obligations.

Under the 1951 Refugee Convention, the prohibition protects against threats to life or freedom based on specific Convention grounds, namely race, religion, nationality, membership of a particular social group, or political opinion (UNHCR, 1951). This branch is closely linked to the refugee protection regime and contains narrowly framed exceptions under Article 33(2).

Under Article 3 of the Convention against Torture, the obligation is framed differently. Here, the focus is specifically on the risk of torture. The protected harm is narrower, but the legal protection is stronger in terms of absoluteness, as no balancing against national security or criminality is permitted (UN, 1984).

Under the International Covenant on Civil and Political Rights, the obligation is derived through interpretation of the rights to life and protection from torture and inhuman treatment, particularly Articles 6 and 7 (Human Rights Committee, 2018).

These obligations share a common protective rationale, yet they differ in source, threshold, protected interest, evidentiary standards, beneficiaries, and possible exceptions. The divergence is not merely theoretical. It has concrete legal implications. For instance, the refugee-law branch permits certain security-based exceptions, while the anti-torture branch is generally treated as absolute.

For this reason, it is analytically more precise to understand the concept as a normative family of interconnected prohibitions rather than a single uniform rule. International courts and treaty bodies rarely apply it as an abstract meta-principle detached from a legal source. Instead, they ground their reasoning in specific treaty provisions and jurisprudential standards.

This composite understanding is more faithful to treaty design and case law. It also allows for doctrinal clarity when analysing the differences between refugee protection, anti-torture obligations, and broader human-rights-based removal prohibitions. Such distinctions are essential for rigorous legal analysis and will remain central throughout the remainder of this article.

2. Historical Development of the Principle

2.1 From territorial asylum to legal obligation

The historical development of non-refoulement reflects a decisive transformation in international law: the movement from discretionary asylum practices grounded in sovereign choice to a binding legal prohibition on removal. This shift is essential to understanding why Article 33 of the 1951 Refugee Convention represents the modern turning point in the doctrine.

Before the codification of refugee protection after the Second World War, asylum was largely understood as an expression of State sovereignty and political discretion. States possessed the authority to grant refuge to foreign nationals, dissidents, or persecuted individuals, but they were generally not subject to a clear international legal duty preventing removal. Protection was often based on diplomatic practice, humanitarian considerations, or political convenience rather than on enforceable legal obligation.

Classical asylum traditions existed in religious, territorial, and diplomatic forms. Historically, religious sanctuary and political asylum reflected practices of hospitality and protection, but these mechanisms did not amount to a universal legal rule prohibiting return. The decision to admit or expel remained fundamentally within the discretion of the receiving authority.

The interwar period marked an important preliminary stage in the evolution of the modern doctrine. Under the League of Nations, specific arrangements emerged for the protection of certain categories of displaced persons, particularly Russian and Armenian refugees. The work associated with Fridtjof Nansen and the development of the Nansen passport system represented an early international response to mass displacement (Hathaway, 2021). However, these arrangements remained fragmented and category-specific. They did not yet establish a general legal prohibition on removal.

The decisive legal transformation occurred after the displacement crises generated by the Second World War. The scale of forced migration, persecution, statelessness, and mass expulsions revealed the insufficiency of purely discretionary asylum models. International law moved away from viewing protection as an act of sovereign grace and began to recognise it as a legal restraint on State conduct.

This transformation was codified in Article 33(1) of the 1951 Convention Relating to the Status of Refugees, which provides that no Contracting State shall expel or return a refugee to territories where life or freedom would be threatened (UNHCR, 1951). This provision is the doctrinal turning point because it converted asylum protection from a matter of political discretion into a binding treaty obligation.

The legal significance of Article 33 lies in three features. First, it imposes a direct obligation of non-removal. Second, it frames the prohibition in broad terms through the phrase “in any manner whatsoever.” Third, it links protection to risk rather than mere presence. These elements established the modern legal foundation of the principle.

This codification fundamentally altered the structure of international protection. Removal powers were no longer purely expressions of sovereignty; they became subject to external legal limits grounded in treaty law.

2.2 Post-1945 expansion beyond refugee law

Although Article 33 of the Refugee Convention established the modern legal foundation of the doctrine, the principle did not remain confined to refugee law. The post-1945 development of international human rights law significantly expanded its scope, transforming it from a status-based protection mechanism into a broader risk-based legal rule.

This expansion is one of the most important doctrinal developments in the field.

Initially, protection under the Refugee Convention depended on the individual falling within the Convention definition of refugee. The legal analysis, therefore, centred on status: whether the person qualified as a refugee within the meaning of Article 1A(2).

Post-war human rights treaties altered this structure.

The Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment introduced one of the clearest examples of this evolution. Article 3 prohibits the expulsion, return, or extradition of any person to another State where there are substantial grounds for believing that he or she would be in danger of being subjected to torture (UN, 1984).

This marked a decisive doctrinal shift.

The protection no longer depended on refugee status, nationality, or Convention grounds such as race or political opinion. Instead, it depended on the existence of a real risk of prohibited harm. The decisive question became the nature and likelihood of future harm rather than the formal legal classification of the individual.

The same evolution occurred through the jurisprudence of human rights bodies interpreting the right to life and the prohibition of torture or inhuman treatment. Under the European Convention on Human Rights, the European Court of Human Rights developed an extensive body of case law prohibiting removal where a real risk exists under Article 3, most notably in Soering v United Kingdom (1989) and subsequent cases.

Similarly, the Human Rights Committee has interpreted Articles 6 and 7 of the ICCPR as imposing removal prohibitions where transfer would expose an individual to arbitrary deprivation of life or prohibited ill-treatment (Human Rights Committee, 2018).

This development is doctrinally significant because it moves the legal inquiry away from who the person is toward what risk the person faces.

The shift from status-based to risk-based protection is now one of the defining characteristics of the modern doctrine. It explains why the principle now applies not only to recognised refugees, but also to asylum seekers, migrants, stateless persons, detainees, and other individuals who may not formally qualify under refugee law but nonetheless face serious danger upon removal.

2.3 The move from admission to jurisdiction

A further major development in the history of the doctrine concerns the shift from an admission-based model to a jurisdiction-based model.

Earlier legal thinking often implicitly assumed that protection obligations arose only after lawful entry into the territory of the State. Under this approach, the legal focus was admission: once a person was admitted, certain obligations attached.

Modern doctrine has moved decisively beyond this framework.

Contemporary international law increasingly focuses on jurisdiction, authority, effective control, and exposure to risk, rather than formal admission status. This development has been driven largely by judicial interpretation and the realities of modern border governance.

The legal question is no longer confined to whether the individual has crossed the territorial border in the formal sense. Instead, courts and treaty bodies increasingly ask whether the State exercises factual authority over the person.

This shift is particularly important in border zones, airport transit areas, maritime interceptions, and offshore control operations.

The European Court of Human Rights’ decision in Hirsi Jamaa and Others v Italy (2012) is particularly significant in this regard. The Court held that the Convention obligations applied to individuals intercepted at sea because Italy exercised effective control over them during the interception and transfer operation.

This reasoning reflects a broader doctrinal movement away from territorial formalism.

Jurisdiction is now increasingly understood through control and authority, rather than through the technicalities of immigration status or territorial admission.

This development has profound practical consequences.

It means that a State may incur responsibility for pushbacks at land borders, refusal of disembarkation after maritime interception, or transfer arrangements conducted outside its territory where effective control is established.

The principle is therefore no longer triggered only after lawful entry. Instead, contemporary law centres on whether State conduct exposes the individual to the prohibited risk while under its jurisdiction or effective authority.

This doctrinal evolution is particularly relevant in current debates surrounding border externalisation, offshore processing, safe third-country agreements, and migration control partnerships.

Historically, this marks the movement of the doctrine from a territorially confined removal rule to a broader principle regulating State conduct wherever effective control over persons exists.

3. Treaty Bases of Non-Refoulement

3.1 Article 33 of the 1951 Refugee Convention

The modern treaty foundation of the doctrine begins with Article 33 of the 1951 Convention Relating to the Status of Refugees. Article 33(1) provides that no Contracting State shall “expel or return (‘refouler’) a refugee in any manner whatsoever” to the frontiers of territories where that person’s life or freedom would be threatened on account of race, religion, nationality, membership of a particular social group, or political opinion (UNHCR, 1951). The phrase “in any manner whatsoever” is the most important textual signal in the provision. It indicates that the prohibition is not confined to a formal deportation order issued after residence within the territory has been regularised. The drafting instead captures the practical effect of State conduct. If a State’s action places a refugee in a situation where the relevant danger materialises, the legal form of the measure does not alter the prohibition.

At the same time, Article 33 is not unlimited in personal scope. Its primary beneficiary class is the refugee. That point matters because refugee-law protection is linked to a defined legal category rather than to all persons facing all forms of serious harm. The beneficiary, however, cannot be restricted only to persons already formally recognised as refugees by administrative decision. Such a reading would defeat the purpose of the provision, since the protection exists precisely to ensure that persons who may qualify are not removed before their claim is properly examined. In legal operation, therefore, Article 33 protects those who are refugees in substance, not merely those who have already been recognised in domestic procedure (Goodwin-Gill and McAdam, 2021).

The structure of Article 33(2) confirms that refugee-law non-refoulement is both central and internally qualified. The paragraph states that the benefit of the provision may not be claimed by a refugee who there are reasonable grounds for regarding as a danger to the security of the host country, or who, having been convicted by final judgment of a particularly serious crime, constitutes a danger to the community of that country (UNHCR, 1951). These exceptions are narrow, but they exist. This feature distinguishes refugee-law protection from the more absolute formulations that later developed under human rights law. The consequence is doctrinally important: Article 33 remains the cornerstone of refugee protection, but it does not exhaust the broader legal field. The wider doctrine cannot be reconstructed by reading refugee law alone, because later treaties created parallel and in some respects stricter prohibitions.

The persecution logic of Article 33 is also broader than a torture-only model. The protected harm is framed as a threat to life or freedom for specified Convention reasons. This means the refugee regime reaches harms that may fall short of torture in the strict sense, while remaining tethered to a specific nexus requirement. That combination of breadth and qualification explains why Article 33 is foundational but not comprehensive.

3.2 Article 3 CAT

Article 3 of the Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment marks a major development in treaty protection because it reframes the prohibition around a particular type of harm rather than around refugee status. It provides that no State Party shall expel, return, or extradite a person to another State where there are substantial grounds for believing that the person would be in danger of being subjected to torture (UN, 1984). Textually, this provision is narrower than Article 33 of the Refugee Convention because the protected harm is specifically torture. It does not extend, on its own terms, to all persecution, all threats to freedom, or all forms of discrimination-based danger.

Yet Article 3 CAT is stronger in another respect: it is widely understood to be absolute. The treaty contains no equivalent to Article 33(2) of the Refugee Convention. A person’s criminal record, perceived dangerousness, or national security profile does not displace the prohibition if the torture risk threshold is met. This is not a marginal doctrinal distinction. It means that a State may, in principle, seek to invoke Article 33(2) in the refugee-law framework, but cannot rely on a comparable balancing exercise under CAT. The CAT Committee has consistently treated Article 3 as requiring an individualized and forward-looking assessment focused on risk rather than on the moral or political character of the person concerned (CAT Committee, 2017).

The threshold in CAT is also framed differently. The treaty speaks of “substantial grounds for believing” that the person would be in danger of torture (UN, 1984). In practice, this has generated a body of jurisprudence centred on indicators such as the receiving State’s human rights record, the complainant’s personal profile, prior torture, political activity, detention history, and the credibility of the account. The focus is not on whether the person qualifies for a legal status, but on whether the evidence establishes a future danger of torture. That shift is one of the key moments in the doctrinal movement from status-based protection to risk-based protection.

The contrast with refugee law is therefore sharp. Refugee law asks whether there is a threat to life or freedom for a Convention reason and allows narrow exceptions. CAT asks whether there is a real torture risk and does not admit comparable exceptions. The CAT model is narrower as to the category of harm, but firmer as to legal consequence.

3.3 ICCPR and implied non-refoulement

The International Covenant on Civil and Political Rights contains no explicit non-refoulement clause. That absence is doctrinally significant because it demonstrates how removal prohibitions can arise not only from express treaty wording, but also from interpretation of substantive rights. The Human Rights Committee has read non-removal obligations into the Covenant, especially through Articles 6 and 7, which protect the right to life and the prohibition of torture or cruel, inhuman, or degrading treatment or punishment.

The interpretive foundation is visible in General Comment No. 31, where the Committee stated that States Parties must not expose individuals to the danger of torture or cruel, inhuman or degrading treatment or punishment upon return by way of extradition, expulsion, or refoulement (Human Rights Committee, 2004). This position was reinforced in General Comment No. 36 on the right to life, which explains that States Parties must refrain from removing a person where there are substantial grounds for believing that there is a real risk of irreparable harm, including harm contemplated by Articles 6 and 7 (Human Rights Committee, 2018).

The doctrinal importance of this development is considerable. It shows that international law does not require an express refoulement clause in every treaty in order for removal prohibitions to arise. Where removal foreseeably results in a violation of a protected right, the obligation not to remove may be derived from the object, purpose, and effective operation of the substantive guarantee itself. This is a particularly important feature of human rights treaty interpretation. The Committee’s reasoning is grounded in effectiveness: rights to life and bodily integrity would be gravely weakened if a State could escape responsibility by physically transferring the individual before the threatened violation occurs.

This implied-obligation model also broadened the substantive field. Whereas Article 3 CAT is limited to torture, and Article 33 of the Refugee Convention is linked to persecution for Convention reasons, the ICCPR approach extends to irreparable harm under the right to life and the prohibition of ill-treatment. The result is a more flexible but also more interpretive branch of the doctrine, one built not on a standalone refoulement article but on the logic of preventing foreseeable Covenant violations.

3.4 ICPPED and disappearance risk

Article 16 of the International Convention for the Protection of All Persons from Enforced Disappearance adds an important but often neglected dimension to the treaty framework. It provides that no person shall be expelled, returned, surrendered, or extradited to another State where there are substantial grounds for believing that he or she would be in danger of enforced disappearance (UN, 2006). This provision matters for two reasons.

First, it is explicit. Unlike the ICCPR, where the prohibition has been inferred through treaty interpretation, the ICPPED directly states a non-removal obligation. Second, it addresses a form of harm that is distinct from, though often connected to, torture or arbitrary killing. Enforced disappearance combines deprivation of liberty, refusal to acknowledge detention or disclose fate or whereabouts, and placement of the victim outside legal protection. The legal injury is therefore composite and particularly severe.

Article 16 is narrower than a general humanitarian protection clause because it is tied to the risk of enforced disappearance. Yet it is broader than is sometimes assumed in practice, because disappearance risk may arise in contexts of secret detention, counter-terrorism operations, authoritarian repression, and conflict-related abductions. The provision also mirrors the structure found in CAT by using a “substantial grounds for believing” threshold, thereby reinforcing the risk-based method of analysis.

Its underuse in scholarship tends to produce an incomplete picture of treaty protection. Discussions that move directly from refugee law to torture law overlook the fact that disappearance risk has been separately codified as a trigger for non-removal. Article 16, therefore, enriches the doctrinal map of the field by showing that treaty law has progressively attached removal prohibitions to different categories of grave harm.

3.5 Regional treaty anchors

Regional human rights systems have played a decisive role in refining the practical meaning of risk, evidence, and procedural protection. In Europe, the core development emerged through Article 3 of the European Convention on Human Rights, which states simply that no one shall be subjected to torture or to inhuman or degrading treatment or punishment. The text says nothing expressly about removal. Yet in Soering v United Kingdom, the European Court of Human Rights held that extradition may engage Article 3 where substantial grounds have been shown for believing that the person, if extradited, faces a real risk of treatment contrary to that article in the receiving State (ECtHR, 1989). This was a foundational step in the evolution of implied non-removal duties in regional law.

Subsequent European case law sharpened the doctrine further. In Chahal v United Kingdom, the Court made clear that Article 3 protection is absolute even in national security cases, rejecting the argument that danger posed by the individual could be balanced against the severity of the risk on return (ECtHR, 1996). In Hirsi Jamaa and Others v Italy, the Court extended the practical reach of the principle to interception at sea, confirming that jurisdiction may arise where a State exercises control over persons outside its territory and that return operations must account for both direct and indirect risk (ECtHR, 2012). Through these judgments, the European system clarified that the doctrine turns on real risk, individualized assessment, and the concrete reach of State authority.

In the Inter-American system, the treaty anchor is more explicit. Article 22(8) of the American Convention on Human Rights provides that in no case may an alien be deported or returned to a country, regardless of whether it is his or her country of origin, if in that country the person’s right to life or personal freedom is in danger of being violated because of race, nationality, religion, social status, or political opinions (OAS, 1969). This text closely resembles the refugee-law model while existing within a human rights instrument.

The Inter-American Court of Human Rights has developed the doctrine further through its case law and advisory opinions. In Pacheco Tineo Family v Bolivia, the Court emphasised the centrality of individualized evaluation and due process in removal decisions affecting persons in need of international protection (IACtHR, 2013). In Advisory Opinion OC-21/14, the Court linked removal protection, child-sensitive procedures, and non-rejection at the border, showing how regional adjudication can deepen the procedural dimension of the doctrine (IACtHR, 2014).

Regional instruments on asylum and migration have also reinforced these obligations by embedding them in broader systems of mobility regulation and human rights supervision. Their practical value lies less in abstract restatement than in judicial elaboration. Regional adjudication has made the doctrine more precise by identifying how risk is assessed, what evidence matters, when jurisdiction exists, and why access to procedures is part of lawful compliance.

These regional developments are not merely supplementary. They have become central to the contemporary legal understanding of removal prohibitions because they translate treaty language into concrete standards for decision-making.

4. Non-Refoulement Beyond Treaty Text

4.1 Customary international law

The legal authority of non-refoulement cannot be fully understood through treaty interpretation alone. A central doctrinal issue in public international law is whether the principle, at least in its core forms, has evolved into a rule of customary international law, binding even on States that are not parties to the principal treaties. This question requires careful legal analysis because claims of customary status are often made in academic and institutional writing without sufficient distinction between normative aspiration and doctrinal proof.

Under Article 38(1)(b) of the Statute of the International Court of Justice, customary international law arises through a general practice accepted as law, which requires both sufficiently widespread and representative State practice and the existence of opinio juris, namely the belief by States that such practice is carried out because it is legally required.

The argument for customary status is strongest in two specific branches of the doctrine: the prohibition on return to persecution in refugee law and the prohibition on transfer to torture under international human rights law.

In the refugee-law context, the case for customary status is frequently advanced on the basis of the near-universal acceptance of the 1951 Refugee Convention and the 1967 Protocol, combined with the widespread incorporation of the principle into domestic asylum legislation, constitutional protections, and judicial review standards (Goodwin-Gill and McAdam, 2021). Numerous States expressly prohibit return to persecution through immigration statutes, constitutional rights provisions, or administrative regulations, which may support both elements of custom.

However, doctrinal rigor requires more than citing widespread treaty ratification. Treaty participation may be strong evidence of opinio juris, but it is not identical to custom. A treaty binds only its parties by virtue of consent, whereas custom binds States through general legal acceptance. The doctrinal inquiry must therefore examine whether the rule is followed and recognised beyond the treaty framework itself.

This is where the analysis becomes more complex.

State practice in this area is not entirely uniform. Contemporary practices such as pushbacks, maritime interceptions followed by return, border externalisation agreements, and summary removals reveal persistent violations and inconsistencies. Yet inconsistency alone does not necessarily defeat a claim of customary status. In the law of custom, what matters is how States legally characterise their conduct.

A particularly important doctrinal indicator is that States rarely deny the existence of the obligation itself. More often, they seek to justify their conduct by arguing that the relevant risk threshold was not met, that the receiving State is safe, or that jurisdiction was absent. Such legal justifications may themselves support opinio juris because they demonstrate that States continue to engage with the norm as a binding legal rule rather than dismissing it as non-binding.

The case for custom is even stronger in relation to the transfer to torture. The prohibition on removal to a real risk of torture is supported by extensive State practice, consistent judicial interpretation, treaty body jurisprudence, and repeated affirmation by international institutions. In this narrower branch, the evidence of both practice and legal conviction is significantly stronger (Allain, 2001).

The more defensible doctrinal conclusion is therefore not that every dimension of the doctrine has crystallised into customary law, but that a core customary obligation exists at least in relation to return to persecution and transfer to torture, while the precise outer boundaries—particularly regarding procedural obligations, extraterritorial application, and indirect transfer—remain more contested.

4.2 Jus cogens claims

A more demanding and controversial question is whether at least part of the doctrine has acquired the status of jus cogens, that is, a peremptory norm of general international law from which no derogation is permitted.

This claim is most plausibly advanced in relation to the anti-torture branch.

The starting point of the analysis is the widely accepted status of the prohibition of torture itself as a jus cogens norm. International courts, treaty bodies, and leading scholarship have repeatedly treated the prohibition of torture as peremptory in nature, meaning that no treaty, domestic law, emergency measure, or claim of national security may lawfully justify torture (ICJ, 2012; Orakhelashvili, 2006).

The legal argument then proceeds by extension. If torture is absolutely prohibited, it may follow that a State must also be absolutely prohibited from transferring a person to another jurisdiction where torture is a foreseeable consequence. Under this reasoning, the prohibition on transfer to torture is not a merely procedural rule but an integral extension of the substantive jus cogens norm.

This reasoning is legally persuasive, but it must be approached with caution.

The existence of a jus cogens primary norm does not automatically elevate every associated or derivative obligation to the same hierarchical status. International law distinguishes between the substantive prohibition itself and the secondary rules necessary to prevent its violation.

The doctrinal question is therefore whether the obligation not to remove a person to torture has independently attained universal recognition as a non-derogable norm, or whether it remains a strongly binding but non-peremptory corollary.

Some scholars argue that separating the two obligations is artificial because transfer to torture would amount to indirect participation in conduct prohibited by a peremptory norm (Allain, 2001). Others argue that the threshold for jus cogens is exceptionally high and requires clearer evidence that the international community of States as a whole accepts the removal prohibition itself as non-derogable.

This distinction is especially important when moving beyond torture.

For example, the refugee-law branch is significantly harder to classify as jus cogens because Article 33(2) of the Refugee Convention expressly contains limited exceptions related to national security and serious criminality. The existence of textual exceptions makes it doctrinally difficult to argue that the refugee-law branch as such possesses a peremptory character.

For that reason, a legally cautious position is preferable. The strongest claim is that the prohibition on transfer to torture may derive peremptory force from the jus cogens nature of the torture prohibition itself, while broader claims extending peremptory status to the entire doctrine should be treated as unsettled and contested.

4.3 Relationship with international humanitarian law

The relationship between non-refoulement and international humanitarian law requires careful doctrinal separation. Humanitarian law does not reproduce the same legal structure found in refugee law or human rights law, yet it provides important complementary protections against forcible transfer, unsafe displacement, and exposure to grave harm in situations of armed conflict.

This relationship is particularly significant in cases involving war zones, occupied territories, civilian evacuation corridors, and conflict-induced cross-border displacement.

Unlike refugee law, humanitarian law does not use non-refoulement as its central doctrinal term. Instead, protection arises through rules governing forcible transfer, deportation, displacement, and the protection of civilians.

A key provision is Article 49 of the Fourth Geneva Convention, which prohibits individual or mass forcible transfers and deportations of protected persons from occupied territory, except where the security of the population or imperative military reasons so demand (Geneva Convention IV, 1949). This rule is not identical to refugee-law non-refoulement, but it serves a related protective function by restricting forced movement into dangerous or unlawful conditions.

The legal relevance of humanitarian law becomes even clearer in contemporary armed conflicts.

Where civilians are displaced internally or across borders, several branches of international law may operate simultaneously. Refugee law may apply to cross-border asylum claims. Human rights law may apply to the risk of torture, arbitrary killing, or inhuman treatment. Humanitarian law regulates the conduct of parties to the conflict, particularly in relation to civilians, occupation, and evacuation.

This overlap is not duplication. Each regime addresses a distinct legal dimension.

Humanitarian law contributes by regulating conduct during conflict and by prohibiting the unsafe forced movement of civilians. It is especially relevant in situations involving military occupation, siege operations, population transfer policies, and compulsory evacuations imposed by armed forces.

For example, when civilians are transferred from occupied territories under the pretext of security, the legality of such movement must be assessed not only under refugee and human rights law but also under the specific rules of occupation law and civilian protection under the Geneva framework.

This is particularly important in contemporary conflict settings where States or armed actors establish so-called humanitarian corridors, relocation zones, or evacuation routes. The mere label of humanitarian evacuation does not automatically render the transfer lawful. The decisive legal question is whether the movement is genuinely protective, temporary where required, and compliant with civilian protection obligations.

Humanitarian law, therefore, reinforces the protective logic of non-removal without reproducing the exact same doctrinal framework. It complements the broader legal architecture by addressing risks specific to armed conflict, occupation, and mass displacement.

A rigorous legal analysis must therefore treat these regimes as complementary but distinct, each contributing to the broader international legal objective of preventing forced transfer into situations of grave danger.

5. Personal Scope of Protection

5.1 Refugees, asylum seekers, and those not yet recognised

The personal scope of protection begins with the category most closely associated with the doctrine: refugees and those seeking recognition as such. Yet one of the most important doctrinal clarifications in this field is that protection cannot depend exclusively on prior formal recognition by domestic authorities. If it did, the rule would be rendered ineffective precisely at the stage where protection is most urgently required.

Under the 1951 Refugee Convention, Article 33 protects “a refugee,” but this term must be understood in light of the declaratory nature of refugee status. In international refugee law, an individual does not become a refugee because a State formally grants recognition. Rather, the individual is considered a refugee from the moment the legal criteria set out in Article 1A(2) are satisfied. The administrative decision merely confirms a status that already exists in law (Goodwin-Gill and McAdam, 2021).

This distinction is not merely theoretical. It has direct consequences for removal procedures. If the prohibition on return applied only after the conclusion of a formal refugee status determination process, States could avoid their obligations by removing individuals before that process is completed. Such an approach would defeat the very object and purpose of Article 33. The legal protection would collapse at the threshold stage, allowing States to circumvent substantive obligations through procedural timing.

This is why access to fair and effective status determination procedures is not simply an administrative matter but a core doctrinal requirement flowing from the substantive guarantee itself. A person who claims fear of persecution must be given a meaningful opportunity to present evidence, to have that evidence assessed individually, and to access remedies capable of suspending removal while the claim is examined.

The protection, therefore, extends to asylum seekers whose claims are pending, individuals awaiting appeal decisions, persons whose applications have not yet been formally registered, and those intercepted at borders who express a need for protection. The law protects not only recognised status holders but also those who may qualify and whose claims have not yet been adjudicated.

This approach is consistent with the preventive logic of the doctrine. Because the harm feared is often irreversible, the legal system cannot wait for formal recognition before activating protection. The obligation necessarily applies at the pre-recognition stage.

5.2 Migrants, stateless persons, and other non-citizens

A frequent legal misconception is the tendency to treat non-refoulement as a doctrine belonging exclusively to refugee law. Although its classical formulation emerged in the Refugee Convention, contemporary international law has developed a much broader protective framework that extends beyond the refugee context.

This expansion occurred primarily through human rights law.

Under treaties such as the Convention against Torture, the International Covenant on Civil and Political Rights, and regional human rights instruments, the protection against removal is not limited to recognised refugees or even to persons who qualify under the refugee definition. Instead, the legal analysis focuses on the risk that the individual would face if returned.

This means that the decisive legal question is not the migration category into which the person falls, but whether removal would expose that person to a real risk of torture, inhuman treatment, arbitrary deprivation of life, enforced disappearance, or comparable irreparable harm.

This doctrinal shift is significant because it corrects the mistaken assumption that persons who do not qualify as refugees automatically fall outside the protective framework.

For example, a migrant who entered a State irregularly and whose claim does not satisfy the nexus requirements of the Refugee Convention may nonetheless face torture upon return. In such a case, protection arises through Article 3 of the Convention against Torture rather than through refugee law. Similarly, a stateless person may not fall within the refugee definition but may face arbitrary detention, disappearance, or life-threatening treatment in the receiving State.

The legal protection in these contexts derives from the rights-based structure of human rights treaties. These treaties protect all persons within a State’s jurisdiction or effective control, regardless of citizenship, residence status, or mode of entry.

This point is doctrinally essential because contemporary migration governance frequently attempts to distinguish between refugees, migrants, undocumented entrants, and other categories as though legal protection depended entirely on such labels. International human rights law rejects that approach. The relevant inquiry is not who the person is in administrative terms, but what legal risk the person would face if removed.

Stateless persons deserve particular attention in this context. Because they lack the protection of any State nationality, they may be especially vulnerable to detention, repeated removal attempts, and exposure to serious harm. Their protection under non-refoulement is therefore often grounded in human rights law rather than refugee law.

The same applies to other non-citizens, including persons in immigration detention, visa overstayers, rejected asylum applicants facing new risks, and individuals subject to extradition. The legal protection against return in these cases arises from the prohibition on exposing any person to prohibited harm, irrespective of formal migration classification.

5.3 Children and family units

The application of the doctrine to children and family units requires a more nuanced analysis because vulnerability factors play a heightened role in the risk assessment. Importantly, this does not change the legal nature of the rule itself. The prohibition remains grounded in the same legal logic of preventing exposure to irreparable harm. What changes is the way in which risk is evaluated.

For children, the principle of the best interests of the child becomes central. This principle, reflected in Article 3 of the Convention on the Rights of the Child, requires that the child’s best interests be treated as a primary consideration in all actions concerning children (UN, 1989). In removal cases, this affects the assessment of both immediate and future harm.

The analysis must consider factors such as age, dependency, psychological vulnerability, access to caregivers, education, healthcare, and the reception conditions in the receiving State. A risk that may not reach the relevant legal threshold for an adult may do so in the case of a child because of the child’s specific vulnerability.

This is particularly relevant in cases involving unaccompanied minors, separated children, or children with medical or psychological needs. International and regional jurisprudence has increasingly recognised that removal decisions affecting children require heightened scrutiny and individualized assessment.

Family unity also plays an important role in the personal scope of protection. International law does not transform non-refoulement into a general mechanism for family reunification, but separation from close family members may be highly relevant to the risk analysis.

For example, where removal would separate a dependent child from parents or primary caregivers, the legal assessment must consider the practical consequences of that separation, including risks of neglect, destitution, trafficking, or severe psychological harm.

Similarly, removal to conditions where family members would face divergent legal treatment may raise issues under both the doctrine itself and related rights such as the right to family life.

The key doctrinal point is that vulnerability factors such as age, dependency, medical condition, and family structure do not create a different legal rule. Rather, they influence the evidentiary and substantive assessment of whether the relevant threshold of serious harm is met.

In other words, vulnerability does not alter the legal nature of the doctrine, but it materially affects how risk is evaluated in practice.

6. Material Scope: What Harms Trigger the Rule

6.1 Persecution in refugee law

Within refugee law, the material scope of protection is defined primarily by the risk of persecution linked to a Convention ground. This remains the classical and doctrinally distinct model under Article 33 of the 1951 Refugee Convention. The focus here is not on harm in the abstract, but on a legally structured concept that combines the seriousness of the feared harm with the reason for which that harm is likely to occur.

Article 33(1) prohibits return to territories where the refugee’s “life or freedom would be threatened” on account of race, religion, nationality, membership of a particular social group, or political opinion (UNHCR, 1951). This wording must be read together with Article 1A(2), which defines the refugee through a well-founded fear of being persecuted for one of these Convention grounds. The legal logic is therefore twofold: there must be a sufficiently serious form of harm, and that harm must be causally linked to one of the protected grounds.

The notion of persecution is not exhaustively defined in the Convention, which is why doctrinal interpretation and jurisprudence have played a central role in shaping its meaning. Persecution generally refers to serious violations of fundamental rights, including threats to life, liberty, bodily integrity, or severe discriminatory treatment that reaches a comparable level of gravity. The legal assessment is not limited to physical violence. Arbitrary detention, politically motivated prosecution, systematic denial of civil rights, forced disappearance, or sustained discriminatory exclusion may all amount to persecution where the threshold of seriousness is met (Hathaway, 2021).

What distinguishes the refugee-law model from broader human-rights protection is the nexus requirement. The feared harm must occur because of one of the Convention grounds. General hardship, economic deprivation, or widespread insecurity, without this connection, does not ordinarily fall within Article 33 as refugee-law protection. This doctrinal requirement is essential because it keeps the analysis anchored to the Convention structure rather than transforming refugee law into a general humanitarian protection regime.

For that reason, the refugee-law trigger is broader than torture in terms of types of harm, but narrower in terms of legal causation, since it requires the protected ground nexus.

6.2 Torture and ill-treatment

In human rights law, torture and ill-treatment constitute the strongest and least controversial triggers of the removal prohibition. This is the area where international jurisprudence is most settled and where the legal threshold has been most consistently articulated.

Article 3 of the Convention against Torture prohibits transfer where there are “substantial grounds for believing” that the person would be in danger of being subjected to torture (UN, 1984). This formula places emphasis on a forward-looking evidentiary assessment. The inquiry is predictive: the decision-maker must determine whether the available evidence establishes a sufficiently serious likelihood of torture upon return.

The CAT standard of “substantial grounds” has been interpreted as requiring concrete and individualized assessment. General concerns about the human rights record of the receiving State may be relevant, but they are not sufficient on their own. The analysis must include the individual’s personal circumstances, past torture, political activities, detention history, ethnicity, religion, or other factors that increase the likelihood of harm.

The European Court of Human Rights employs slightly different language, typically referring to a “real risk” under Article 3 of the European Convention on Human Rights. In Soering v United Kingdom (1989) and subsequent cases, the Court established that removal is prohibited where substantial grounds have been shown for believing that the person would face a real risk of torture or inhuman or degrading treatment.

Although the wording differs, the practical standards substantially overlap.

The CAT formula emphasises evidentiary sufficiency, while the ECtHR formulation emphasises the seriousness and credibility of the future risk. Both require individualized analysis and reject purely speculative fears.

This branch of the doctrine is the strongest because it is widely treated as absolute. Unlike refugee-law protection under Article 33(2), there is no balancing against national security, criminality, or public danger where the torture threshold is met.

This doctrinal absolutism is one of the clearest features distinguishing the human-rights branch from refugee law.

6.3 Risk to life and irreparable harm

The material scope of protection extends beyond persecution and torture to encompass risk to life and other forms of irreparable harm. This development is particularly significant in human rights law, where removal may be prohibited even in the absence of torture if the return would expose the individual to death, arbitrary killing, or conditions incompatible with the right to life.

The Human Rights Committee has interpreted Article 6 of the ICCPR as imposing non-removal obligations where substantial grounds exist for believing that there is a real risk of irreparable harm to life (Human Rights Committee, 2018). This includes exposure to the death penalty in circumstances contrary to international law, arbitrary executions, armed conflict situations involving indiscriminate violence, and certain conditions of extreme deprivation.

This doctrinal expansion is important but must be approached carefully.

Not every difficult or dangerous condition in the receiving State automatically engages the prohibition. International jurisprudence generally requires a threshold of seriousness that reaches irreparable harm. Severe violence, collapse of public order, absence of medical treatment essential to survival, or catastrophic humanitarian conditions may satisfy this threshold where the consequences are sufficiently grave and individualized.

For example, the European Court of Human Rights has recognised that removal may violate Article 3 in cases involving terminal illness or lack of essential medical treatment where the foreseeable consequence is severe suffering or death, as developed in cases such as Paposhvili v Belgium (2016).

This area also illustrates the doctrinal limits of expansion.

The prohibition is not a general mechanism for preventing return to all forms of hardship, poverty, or lower standards of living. The legal threshold remains irreparable harm of sufficient gravity. This distinction is necessary to preserve doctrinal coherence and to prevent the concept from dissolving into a broad humanitarian discretion standard.

6.4 Chain refoulement

A particularly important dimension of the material scope is chain refoulement, sometimes referred to as indirect refoulement. This occurs where a State does not return an individual directly to the country of persecution or torture, but instead transfers the person to an intermediary State that may subsequently remove the individual onward to the place of danger.

This issue is doctrinally essential in contemporary migration governance because many return systems are structured through readmission agreements, safe third-country arrangements, offshore processing centres, and transfer partnerships.

The legal obligation does not end simply because the immediate receiving State is not itself the source of the feared harm. International law requires an assessment of whether onward transfer is foreseeable.

If the intermediary State is likely to return the individual to a place where persecution, torture, or irreparable harm exists, the original transferring State may still incur responsibility.

This is particularly relevant where the intermediary State lacks adequate asylum procedures, has a documented practice of onward removal, or does not provide effective protection.

The Grand Chamber of the European Court of Human Rights addressed this logic in M.S.S. v Belgium and Greece (2011), where transfer within the Dublin system was scrutinised not only in terms of immediate reception conditions but also in terms of the practical availability of protection and risk of subsequent removal.

The doctrinal rationale is clear: the prohibition cannot be circumvented through intermediary transfers.

If the law prohibited only direct return, States could systematically evade responsibility by routing individuals through third countries that would later perform the final removal.

Chain refoulement, therefore, reinforces the preventive logic of the doctrine by addressing the full foreseeable causal chain of State action.

7. Territorial and Extraterritorial Reach

7.1 Borders, territorial seas, and transit zones

One of the most important doctrinal developments in the law of non-refoulement concerns the rejection of the older assumption that international responsibility arises only after an individual has formally entered the territory of a State. Contemporary international law no longer supports such a restrictive approach. Instead, responsibility may arise at land borders, within territorial seas, and in transit zones whenever the State exercises authority and control over the individual concerned.

This development is particularly significant because many modern migration control strategies are specifically designed to prevent formal entry. Practices such as pushbacks, immediate returns at border fences, refusal of access at checkpoints, and confinement in airport transit zones often rely on the assumption that the absence of formal admission limits the State’s legal obligations. Current doctrine increasingly rejects that view.

The decisive legal issue is not whether the person has crossed a technical territorial line in domestic immigration law, but whether the State has exercised sufficient control to bring the person within its jurisdiction for the purposes of international responsibility. Where border guards, coast guards, police authorities, or immigration officers physically restrain, redirect, detain, or summarily return an individual, the State is exercising direct authority. That authority is sufficient to trigger obligations under the prohibition on return to persecution, torture, or other irreparable harm (Goodwin-Gill and McAdam, 2021).

This point is particularly important in relation to pushbacks and summary rejection practices. Pushbacks typically involve the immediate removal or redirection of individuals at land or sea borders without any individualized assessment of risk. From a doctrinal perspective, such practices raise serious concerns because they eliminate the procedural safeguards necessary to determine whether removal would expose the person to prohibited harm. Since the rule is preventive in nature, the legal violation may arise not only from the actual exposure to danger but also from the absence of an effective mechanism capable of assessing that risk (ECtHR, 2020).

Transit zones present similar legal issues. Historically, some States have attempted to treat airport transit areas or border holding facilities as legally exceptional spaces where ordinary territorial obligations do not fully apply. This formalistic approach has been increasingly displaced by a control-based analysis. Where an individual is confined, restricted in movement, or otherwise subject to the direct authority of State officials in such zones, legal obligations may arise regardless of whether domestic law treats the person as having “entered” the territory (UNHCR, 2007).

The doctrinal shift is therefore clear: the law focuses on jurisdiction and control rather than on formal admission status. This is essential to prevent States from circumventing international obligations through procedural design or geographic technicalities.

7.2 Interdiction at sea

The maritime context provides one of the clearest illustrations of how jurisdiction in international law increasingly follows power rather than physical territory. Maritime interdiction cases are particularly important because they test the extent to which legal obligations attach when States exercise control over persons beyond their territorial waters.

When a State intercepts a vessel, boards it, directs its movement, rescues persons in distress, or transfers individuals from one vessel to another, it is exercising direct factual authority over those individuals. Contemporary jurisprudence treats this form of control as sufficient to establish jurisdiction for the purposes of international responsibility (ECtHR, 2012).

The Grand Chamber judgment in Hirsi Jamaa and Others v Italy remains the leading authority in this area. Italy intercepted migrants on the high seas and returned them to Libya without conducting any individualized assessment of the risks they faced. The Court held that Italy had exercised continuous and exclusive control over the individuals during the operation and therefore bore responsibility under the European Convention on Human Rights (ECtHR, 2012).

The significance of this judgment lies in its rejection of the argument that obligations are territorially confined. The Court made clear that jurisdiction may arise extraterritorially where the State exercises effective authority over persons, even on the high seas.

The legal reasoning in maritime cases is especially important because contemporary migration control frequently occurs through naval patrols, joint coast guard operations, and coordinated return arrangements at sea. In such contexts, the question is not where the vessel is located geographically, but who exercises effective authority over the persons onboard.

This has direct implications for rescue operations. Search and rescue obligations under the law of the sea cannot be separated from the legal consequences of subsequent disembarkation and transfer decisions. If individuals are rescued and then disembarked in a place where they face persecution, arbitrary detention, torture, or onward transfer to danger, the rescuing or transferring State may incur responsibility (Milanovic, 2013).

The maritime context, therefore, demonstrates with particular clarity that international obligations attach to control over persons, not solely to territorial location.

7.3 Offshore processing and proxy enforcement

One of the most significant contemporary challenges to the doctrine arises from the increasing use of offshore processing systems and proxy enforcement mechanisms. States increasingly seek to externalise migration control by transferring asylum seekers and migrants to third States, delegating interception to foreign coast guards, financing external detention centres, or relying on international organisations and private actors to manage transfer chains.

These arrangements are often structured to create legal distance between the initiating State and the final place of detention, processing, or return. However, international responsibility cannot be avoided merely by outsourcing operational steps.

The central legal issue is whether the State retains decisive control over the chain of transfer.

Where a State funds, directs, coordinates, authorises, or substantially controls the transfer architecture, legal responsibility may still arise even if another State or private actor performs the physical act of detention or onward removal. International law focuses on substance over formal delegation.

This is particularly relevant in offshore processing arrangements where asylum seekers are transferred to facilities located outside the territory of the transferring State. If the original State determines who is transferred, where they are sent, under what conditions they are detained, and whether they may be subsequently removed, that State may remain internationally responsible for foreseeable exposure to prohibited harm (Gammeltoft-Hansen and Tan, 2017).

The same logic applies to proxy enforcement through third-State coast guards or migration-control partnerships. Financial support, operational coordination, intelligence sharing, or joint command structures may establish sufficient causal and legal connection to trigger responsibility.

The legal argument is therefore straightforward: a State cannot contract away its obligations where it exercises decisive influence over the chain of transfer and the foreseeable risk remains legally attributable to its conduct.

This principle is increasingly important in contemporary migration governance, where externalisation strategies are often designed precisely to create jurisdictional ambiguity. A rigorous doctrinal approach requires looking beyond formal institutional arrangements and examining who exercises effective and decisive control over the movement of persons.

8. Exceptions, Limits, and the Question of Absoluteness

8.1 Exceptions under refugee law

The question of exceptions must begin with a careful and textually disciplined reading of Article 33(2) of the 1951 Refugee Convention. This provision states that the benefit of Article 33(1) may not be claimed by a refugee for whom there are reasonable grounds for regarding as a danger to the security of the host country, or who, having been convicted by a final judgment of a particularly serious crime, constitutes a danger to the community of that country (UNHCR, 1951).

This clause must be interpreted narrowly.

First, Article 33(2) is an exception to a core protection norm, and under general principles of treaty interpretation, exceptions to protective provisions should not be read expansively. The legal burden rests on the removing State to demonstrate that the threshold conditions are met. Mere suspicion, generalized concern, or political rhetoric is insufficient. The text requires either reasonable grounds relating to national security or a final conviction for a particularly serious crime coupled with present danger to the community.

Second, this provision must be kept analytically distinct from the exclusion clauses in Article 1F of the Refugee Convention. Article 1F addresses whether a person falls within the Convention’s personal scope at all, excluding individuals responsible for war crimes, crimes against humanity, serious non-political crimes, or acts contrary to the purposes and principles of the United Nations. Article 33(2), by contrast, applies to persons who are already within the scope of refugee protection but who may lose the benefit of the specific non-return guarantee under narrowly defined circumstances (Hathaway, 2021).

This distinction is doctrinally essential because exclusion concerns status, while Article 33(2) concerns removal despite status.

A further point requiring emphasis is that the refugee-law exceptions do not automatically carry over into human-rights non-refoulement. This is one of the most important fault lines in the doctrine.

The existence of textual exceptions in Article 33(2) reflects the treaty design of the Refugee Convention. Human rights instruments such as the Convention against Torture and the European Convention on Human Rights do not contain equivalent balancing clauses in relation to torture or inhuman treatment.

Therefore, it would be legally incorrect to assume that a refugee who falls within Article 33(2) may automatically be removed if the destination State presents a real risk of torture or ill-treatment. In such cases, the human-rights branch remains independently applicable and may still prohibit removal.

This separation between treaty regimes is fundamental to doctrinal accuracy.

8.2 Absolute character under human rights law

The most significant doctrinal contrast with refugee law arises in the human-rights branch, particularly under Article 3 of the Convention against Torture and Article 3 of the European Convention on Human Rights.

Under these frameworks, the protection against removal to torture or inhuman treatment is generally understood as absolute.

This means that once the relevant risk threshold is established, the State is not permitted to balance the severity of the risk faced by the individual against the danger allegedly posed by that individual to the host State.

This point cannot be overstated because it marks one of the clearest distinctions between refugee law and human rights law.

Under CAT, Article 3 contains no equivalent to Article 33(2) of the Refugee Convention. The treaty simply prohibits expulsion, return, or extradition where substantial grounds exist for believing that the individual would be in danger of torture (UN, 1984).

Similarly, the jurisprudence under Article 3 ECHR has consistently rejected balancing exercises.

The legal reasoning is that the prohibition of torture and inhuman or degrading treatment is itself absolute. If a State were permitted to weigh the individual’s dangerousness against the risk of torture, the absolute nature of the substantive right would be undermined.

This is precisely what the European Court of Human Rights clarified in Chahal v United Kingdom (1996). The applicant was considered a national security threat, yet the Court held that Article 3 prohibits removal where a real risk of ill-treatment exists, regardless of the threat posed by the individual.

The doctrinal importance of this principle lies in the separation between risk assessment and public interest balancing.

The only relevant inquiry is whether the risk threshold is met.

The gravity of the individual’s conduct, criminal record, political affiliations, or security profile may be relevant in other legal contexts, such as detention or prosecution, but not in determining whether removal to torture is permissible.

This absolute character is one of the defining features of the human-rights branch and must be kept analytically distinct from the narrower and exception-based structure of refugee law.

8.3 National security and terrorism cases

National security and terrorism cases represent the most difficult test of the doctrine because they place the legal rule under maximum political and institutional pressure.

These are the cases in which States most frequently argue that the protection should yield to collective security concerns.

However, the jurisprudence in hard cases has not weakened the absolute human-rights branch. On the contrary, it has clarified and strengthened it.

The leading authority remains Chahal v United Kingdom (1996), where the applicant was alleged to pose a threat to national security due to suspected links to Sikh separatist extremism. The United Kingdom argued that removal was justified because of the seriousness of the threat.

The Court rejected this argument unequivocally.

It held that Article 3 enshrines one of the most fundamental values of democratic societies and that its prohibition is absolute, irrespective of the reprehensible nature of the individual’s conduct or the danger posed to the community (ECtHR, 1996).

This principle was later reaffirmed in Saadi v Italy (2008), which involved terrorism-related allegations. Again, the Court made clear that the prohibition on removal to torture or inhuman treatment admits no balancing against security interests.

These cases are doctrinally significant because they demonstrate that hard cases did not create exceptions; rather, they clarified the rule’s boundaries.

The legal response to national security threats must therefore occur through lawful domestic mechanisms such as surveillance, criminal prosecution, detention subject to due process, or other preventive measures compatible with human rights obligations.

Removal to torture remains prohibited.

This doctrinal clarity is particularly important in contemporary counter-terrorism contexts, where States may be tempted to use transfer arrangements, diplomatic assurances, or third-country removals as substitutes for domestic legal responses.

The jurisprudence makes clear that such strategies remain constrained by the absolute prohibition where the relevant risk threshold is met.

9. Procedure as Substance

9.1 Individualised assessment

A central doctrinal feature of the law of non-refoulement is that compliance depends not only on the final decision to remove, but also on the quality and structure of the procedure through which that decision is reached. The prohibition may therefore be violated even before removal occurs if the procedural framework is so defective that it creates a substantial likelihood of wrongful transfer. This is why procedure in this area cannot be treated as a merely administrative matter. It forms part of the substantive protection itself. The preventive purpose of the rule requires that States identify and assess risk before exposure to harm takes place. Because the feared injury is often irreversible, such as torture, disappearance, persecution, or death, international law requires procedures capable of detecting that risk in a timely and reliable manner (Goodwin-Gill and McAdam, 2021).

For this reason, individualised assessment is legally indispensable. A State cannot satisfy its obligations through automatic return mechanisms, blanket assumptions that a destination country is generally safe, or accelerated procedures that deny the person a meaningful opportunity to present personal circumstances. The legal inquiry must focus on the specific individual and the concrete risks that person would face if removed. General country information is relevant, but it cannot replace examination of the applicant’s own profile. Factors such as political opinion, religious identity, ethnicity, previous detention, evidence of past torture, family associations, gender-related vulnerabilities, sexual orientation, medical condition, or public visibility may significantly alter the level of risk. Two persons who return to the same State may face entirely different legal consequences because their personal circumstances differ.

This requirement is not simply a matter of good administrative practice; it is a doctrinal necessity flowing from the nature of the obligation itself. Since the prohibition is aimed at preventing foreseeable harm, the legal system must be structured in a way that can actually identify foreseeable risk. If the procedure is incapable of doing so, the State may incur responsibility even before the transfer is carried out. This understanding has been reinforced by jurisprudence. In M.S.S. v Belgium and Greece, the European Court of Human Rights examined not only the substantive risks associated with transfer but also the procedural deficiencies that prevented effective access to protection (ECtHR, 2011). The judgment reflects the broader principle that procedural safeguards are integral to the effectiveness of the substantive rule.

9.2 Burden and standard of proof

The assessment of future harm necessarily takes place under conditions of uncertainty. Decision-makers are required to evaluate events that have not yet occurred and that depend on future conduct by authorities in another State. For this reason, the evidentiary framework in this area differs fundamentally from criminal or punitive standards of proof. The inquiry is predictive rather than retrospective and preventive rather than punitive.

The relevant question is not whether the individual can prove with certainty that torture, persecution, or death will occur after removal. Such a standard would be impossible to satisfy in most cases and would undermine the preventive purpose of the rule. Instead, international law asks whether the available evidence demonstrates a sufficiently credible likelihood that removal would expose the person to prohibited harm.

Different treaty regimes express this threshold using different formulas. Article 3 of the Convention against Torture refers to “substantial grounds for believing” that the person would be in danger of torture (UN, 1984). The European Court of Human Rights usually employs the language of a “real risk” under Article 3 ECHR (ECtHR, 1989). Although the wording differs, both standards operate within a closely related doctrinal framework. Neither requires certainty, nor do they require proof beyond a reasonable doubt. Rather, the decision-maker must assess whether the evidence, taken as a whole, establishes a sufficiently serious probability of harm.

This assessment typically combines general country conditions with individual evidence. Reports from UN bodies, treaty committees, reputable non-governmental organisations, academic sources, and prior judicial decisions may establish patterns of torture, arbitrary detention, or systemic persecution. At the same time, the applicant’s personal circumstances remain central. Evidence of past torture or persecution is especially significant because it may provide strong evidence of future risk, particularly where the political or institutional conditions in the receiving State have not materially changed (CAT Committee, 2017).

9.3 Interim measures and suspensive effect

The existence of a legal remedy is not sufficient in itself. In removal cases, the remedy must be capable of preventing the transfer while the claim is under examination. This is one of the most practically important procedural safeguards in the doctrine.

A legal challenge that becomes available only after the person has already been removed may be incapable of providing meaningful protection. If the feared harm consists of torture, enforced disappearance, persecution, or arbitrary killing, review after removal may arrive too late to prevent the violation. Because the doctrine is fundamentally preventive, the effectiveness of the remedy depends on its ability to preserve the individual within the jurisdiction of the reviewing authority until the risk assessment is complete.

This is why international law increasingly requires remedies in removal cases to have a suspensive effect. Suspensive effect means that the transfer is halted pending the outcome of the legal challenge. The European Court of Human Rights made this clear in Jabari v Turkey, where it held that a remedy must be effective both in law and in practice, which necessarily includes the ability to prevent removal while the Article 3 claim is being examined (ECtHR, 2000).

The same logic underpins interim measures issued by international courts and treaty bodies. These measures are not mere procedural formalities. They serve the essential function of preventing irreversible harm before the final legal determination is made. Without a suspensive effect, the legal system risks offering remedies that exist formally but fail to prevent the very injury they are designed to address.

9.4 Diplomatic assurances

Diplomatic assurances represent one of the most difficult issues in the practical application of the doctrine. They arise when the receiving State formally promises that the transferred individual will not be subjected to torture, inhuman treatment, arbitrary detention, or other prohibited harm.

The legal question is whether such assurances are sufficient to neutralise an otherwise established risk. International law does not treat diplomatic assurances as automatically determinative. Their legal value depends on their specificity, credibility, enforceability, and consistency with the receiving State’s actual record.

A general diplomatic statement that the person will be treated in accordance with international standards is rarely sufficient. The decision-maker must examine whether the assurances are tailored to the individual case, whether there are realistic mechanisms for monitoring compliance, whether access for diplomatic or independent verification is available, and whether the receiving State has a demonstrated record of compliance with previous undertakings.

The prior conduct of the receiving State is especially important. If that State has a documented history of torture, secret detention, or disregard for previous diplomatic commitments, the evidentiary weight of the assurances may be substantially weakened. This issue was analysed in detail in Othman (Abu Qatada) v United Kingdom, where the European Court of Human Rights examined whether the assurances provided by Jordan were sufficiently credible and practically enforceable to reduce the risk below the legal threshold (ECtHR, 2012).

The doctrinal significance of diplomatic assurances lies in the distinction between formal promises and actual reduction of risk. International law requires the latter. An assurance is only relevant insofar as it materially changes the factual likelihood of prohibited harm.

10. Non-Refoulement in Contemporary Governance

10.1 Pushbacks and externalisation

One of the most significant developments in contemporary migration governance is the increasing reliance on pushbacks and externalisation measures as instruments of border control. These practices are legally important because they are often deliberately structured to reduce, obscure, or distance State responsibility under international law. Instead of engaging with asylum procedures and individualized risk assessment, some States attempt to intervene at the earliest point of contact, frequently before formal admission, or through third actors whose conduct is then presented as legally separate from the removing State.

Pushbacks typically involve the immediate redirection, forced return, or summary removal of individuals at land borders, maritime frontiers, or transit areas without access to any meaningful procedure capable of assessing whether removal would expose them to persecution, torture, inhuman treatment, or other forms of irreparable harm. The legal flaw in such measures lies in the assumption that obligations arise only after formal entry into the territory. Contemporary international law does not support such a restrictive view. The relevant legal question is not whether the individual has been admitted under domestic immigration law, but whether the State has exercised authority and control over that person. Once border guards, police officers, coast guards, or military personnel physically intercept, detain, escort, or redirect individuals, the State is exercising jurisdiction in the relevant international legal sense, and the obligation not to expose them to serious harm may be triggered (UNHCR, 2007; Goodwin-Gill and McAdam, 2021).

Externalisation raises the same issue in a more complex form. Instead of carrying out the return directly, States increasingly rely on third-country authorities, offshore facilities, international organisations, or private contractors to intercept, detain, process, or transfer individuals. These arrangements are often framed as mechanisms that transfer operational responsibility to another actor. However, international law focuses on effective control and causal responsibility rather than on formal delegation. Where a State finances, coordinates, directs, authorises, or substantially influences the chain of transfer, responsibility may still attach because the harmful outcome remains legally connected to that State’s conduct. This is particularly true where the State knows, or ought reasonably to know, that the externalised mechanism may result in direct return to danger or onward transfer through an unsafe third State (Gammeltoft-Hansen and Tan, 2017).

This is what makes pushbacks and externalisation central to the contemporary doctrine. They test whether legal responsibility follows formal territorial admission or the actual exercise of power. Modern doctrine increasingly supports the latter approach, requiring courts and decision-makers to examine control, causation, and foreseeability rather than merely procedural labels.

10.2 Safe third country and return agreements

The use of safe third country mechanisms and bilateral return agreements has become one of the most common legal tools in contemporary migration governance. These arrangements are generally based on the proposition that an individual may be transferred to another State considered capable of providing effective protection or access to asylum procedures.

The legal difficulty lies in treating a general designation of safety as sufficient in itself. International law does not permit abstract safety labels to replace individualized risk assessment. A country may be generally regarded as institutionally stable and yet present serious risks for specific categories of individuals. Safety must therefore be assessed not only at the systemic level but also in relation to the particular characteristics of the person concerned.

This point is doctrinally important because many transfer regimes operate through presumptions. Once a country is designated as safe, the individual is often presumed not to require further protection against transfer. Such presumptions are legally problematic because they undermine the individualized nature of the inquiry that lies at the core of the doctrine. A political dissident, member of a persecuted minority, survivor of trafficking, or person with serious medical vulnerabilities may face risks that are not captured by a broad country designation.

In addition, even where the third State itself does not present a direct risk, the inquiry cannot end there. The removing State must assess whether the receiving State offers effective access to asylum procedures, adequate reception conditions, and protection against onward transfer. If those safeguards are absent, the transfer may give rise to chain refoulement, where the individual is subsequently removed to a place of persecution, torture, or arbitrary deprivation of life. This issue was central in M.S.S. v Belgium and Greece, where the European Court of Human Rights examined not only the receiving State’s conditions but also the practical reality of access to protection and the risk of onward harm (ECtHR, 2011).

The legal conclusion is therefore clear: safe third country labels and return agreements may be relevant, but they cannot displace the obligation to conduct individualized review. The decisive question remains whether this particular person will, in practice, have access to effective protection without exposure to direct or indirect harm.

10.3 Mass influx and emergency claims

A recurrent argument in contemporary migration governance is that large-scale arrivals, emergency situations, or sudden border pressure justify a dilution of protection obligations. States frequently invoke resource constraints, institutional overload, or exceptional migratory flows to justify accelerated returns, reduced procedural guarantees, or broader discretionary powers of removal.

This argument must be examined with doctrinal precision. The existence of large-scale arrivals does not, by itself, diminish the core legal obligation not to expose persons to persecution, torture, inhuman treatment, or other irreparable harm. Administrative burden, political urgency, or resource pressure cannot automatically justify weakening the substance of the rule.

This point is particularly clear in the human-rights branch of the doctrine. Where removal would expose an individual to torture, inhuman treatment, arbitrary deprivation of life, or comparable irreparable harm, the obligation remains non-derogable in its core forms. Neither the Convention against Torture nor the Article 3 ECHR framework permits derogation on the basis of migration pressure alone (ECtHR, 1996; UN, 1984).

This does not mean that emergency contexts are legally irrelevant. Certain legal frameworks may provide for temporary protection mechanisms, group-based processing models, or emergency reception systems designed to address mass displacement more efficiently. However, such mechanisms must be interpreted narrowly and cannot be used to bypass individualized assessment where the risk of prohibited harm remains present.

The doctrinal mistake frequently made in practice is to move from administrative difficulty to substantive dilution. International law does not support that transition unless the applicable legal framework expressly provides a narrowly construed basis. Even in mass influx situations, the human-rights branch of the doctrine remains fully applicable because its core purpose is to prevent irreparable harm regardless of the scale of arrivals.

This is one of the clearest examples of how the rule preserves its legal force precisely in moments of greatest political pressure.

Also read

• Climate Refugees and International Law

• How International Law Deals with Stateless Persons

• What Are Human Rights

• International Humanitarian Law Explained

• Ireland v. United Kingdom Case

• Customary International Law: How It Forms and Works

• Religion and International Law: Conflicts and Coexistence

• Crisis of Multilateralism and Global Order

11. Doctrinal Controversies

11.1 Is there a right to enter to claim protection?

One of the most contested doctrinal questions in the law of non-refoulement is whether the principle implies a right to enter the territory of a State for the purpose of seeking protection. This issue must be framed with precision because the legal debate is often incorrectly presented as a choice between absolute border sovereignty and a general right of migration.

The stronger doctrinal position is not that international law creates a free-standing and universal right of immigration. States continue to retain broad sovereign authority over admission, immigration control, and territorial borders. Non-refoulement does not convert international protection law into a general right to enter any State at will.

The legal issue is narrower and more precise. The relevant question is whether a State may lawfully prevent access to its territory, frontier procedure, or asylum mechanism when doing so foreseeably exposes the individual to persecution, torture, inhuman treatment, or other irreparable harm. The stronger doctrinal answer is that it may not.

This conclusion flows from the preventive structure of the rule itself. If border authorities block access, summarily reject individuals at the frontier, refuse disembarkation after maritime rescue, or otherwise deny any opportunity to present a protection claim, and the foreseeable consequence is exposure to serious harm, responsibility arises not because there is a general right to migrate, but because the State has created the conditions for refoulement.

This distinction is essential. The legal obligation is better understood as a duty not to obstruct access to protection where such obstruction makes unlawful return likely. In practice, this means that border control measures must preserve a real and effective opportunity for individuals to present claims that removal would expose them to prohibited harm.

This approach has become increasingly important in litigation concerning pushbacks and summary border returns. The legal focus is not on whether the person has a general entitlement to cross the border, but on whether the State’s conduct prevents the operation of the protective mechanism required to avoid refoulement (UNHCR, 2007; ECtHR, 2020).

11.2 Socio-economic deprivation and medical cases

A particularly difficult doctrinal controversy concerns cases in which the feared harm does not arise from persecution in the classical refugee-law sense or from deliberate torture, but instead from extreme socio-economic deprivation, lack of medical treatment, or broader humanitarian collapse in the receiving State.

This is one of the most nuanced areas of jurisprudence because international courts have generally approached it with caution and on a highly case-specific basis.

International law does not treat every return to conditions of poverty, social hardship, or lower living standards as unlawful. The doctrine is not a general guarantee against economic disadvantage or inferior welfare systems. Courts have been careful not to transform non-refoulement into a broad socio-economic equalisation mechanism.

However, the legal analysis changes where the deprivation reaches the threshold of serious and irreversible harm to life, bodily integrity, or human dignity.

This has been particularly developed in medical removal cases. In N v United Kingdom (2008), the European Court of Human Rights adopted a relatively restrictive approach, holding that removal of a seriously ill applicant to a country with substantially inferior healthcare did not automatically violate Article 3. The Court emphasised that only exceptional circumstances would engage the prohibition.

Later jurisprudence refined this approach significantly. In Paposhvili v Belgium (2016), the Grand Chamber held that removal may violate Article 3 where substantial grounds exist for believing that the person faces a real risk of a serious, rapid, and irreversible decline in health resulting in intense suffering or significant reduction in life expectancy (ECtHR, 2016).

This development is doctrinally significant because it clarifies that socio-economic and medical cases are not excluded as a matter of principle. Rather, the legal threshold remains one of irreparable harm.

The same logic may apply in cases of broader humanitarian collapse, such as severe food insecurity, collapse of essential services, or absence of access to life-sustaining shelter, provided the evidence establishes that the return would expose the specific individual to a sufficiently grave and foreseeable risk.

The case law remains cautious because courts seek to preserve doctrinal limits while still recognising that certain forms of deprivation may engage the right to life or the prohibition of inhuman treatment.

11.3 Climate-related displacement

Climate-related displacement is one of the most rapidly developing areas of contemporary legal debate. It is also one of the areas most vulnerable to doctrinal overstatement.

A careful legal analysis must begin by rejecting the assumption that climate mobility automatically generates refugee status under the 1951 Refugee Convention. The refugee definition remains grounded in persecution linked to a Convention ground such as race, religion, nationality, membership of a particular social group, or political opinion. Environmental degradation, sea-level rise, drought, or climate-related disasters do not, by themselves, automatically satisfy this legal structure.

The stronger doctrinal line is therefore found in the human-rights branch of the doctrine rather than in automatic refugee classification.

Where climate-related conditions create a real risk to life, physical integrity, or exposure to inhuman treatment, return may engage non-refoulement obligations under human rights law.

This position was significantly developed by the Human Rights Committee in Teitiota v New Zealand (2020). Although the Committee did not find a violation on the facts, it accepted the principle that environmental degradation and climate-related threats may, in sufficiently serious circumstances, create a risk of irreparable harm under Article 6 of the ICCPR (Human Rights Committee, 2020).

The doctrinal importance of this reasoning is substantial. It moves the analysis away from the mistaken assumption that climate displacement must fit within refugee law and instead grounds protection in the broader human-rights framework.

This does not mean that every climate-related movement engages the doctrine. The legal threshold remains one of sufficiently serious and foreseeable harm. Gradual environmental decline, economic hardship caused by drought, or future risk projections must be assessed carefully and individually.

The key legal inquiry is whether the return would expose the person to a real risk of death, severe deprivation incompatible with the right to life, or treatment reaching the threshold of inhuman or degrading treatment.

This area remains highly context-specific, but the stronger doctrinal position is now clear: climate-related exposure may engage non-refoulement where the risk reaches the level of irreparable human-rights harm.

12. Conclusion

The Principle of Non-Refoulement is best understood not as a single rule with a uniform content, but as a layered and interconnected protective architecture within public international law. A careful doctrinal analysis demonstrates that the principle operates through multiple legal branches, each grounded in distinct treaty regimes, interpretative practices, and thresholds of harm. The refugee-law branch, centred on Article 33 of the 1951 Refugee Convention, remains foundational because it provides the classical formulation of the prohibition against return to persecution linked to Convention grounds. However, the broader legal framework extends far beyond refugee law.

Human-rights instruments such as the Convention against Torture, the International Covenant on Civil and Political Rights, the International Convention for the Protection of All Persons from Enforced Disappearance, and regional human-rights systems have transformed the principle into a wider protective doctrine that applies whenever removal would expose an individual to torture, inhuman treatment, disappearance, arbitrary deprivation of life, or comparable irreparable harm. The legal significance of this evolution lies in the fact that protection no longer depends exclusively on formal legal status, but increasingly on the nature and seriousness of the risk that removal would create.

The doctrinal strength of this protective architecture rests on four propositions that have emerged consistently throughout treaty law, judicial interpretation, and scholarly analysis. The first is that risk, rather than formal classification, is the central legal trigger. Whether the individual is a recognised refugee, an asylum seeker, a migrant, a stateless person, or another non-citizen, the decisive question is whether the available evidence establishes a sufficiently real and foreseeable risk of prohibited harm upon removal. This risk-based logic explains why the doctrine has expanded beyond the refugee-law model and why human-rights law now plays such a decisive role in contemporary cases.

The second proposition is that jurisdiction in this field is broader than territory. Contemporary international law no longer treats the principle as applicable only after lawful admission into the territorial space of the State. Instead, responsibility increasingly follows authority, effective control, and causal influence. This is why obligations may arise at borders, in transit zones, during maritime interception, and in situations of offshore processing or delegated migration control. The law now looks beyond territorial formalism and focuses on whether the State, directly or indirectly, exercises sufficient control over the person or over the chain of transfer that may expose that person to serious harm.

The third proposition is that procedure is part of protection itself. The doctrine is not satisfied merely by avoiding wrongful removal in the final decision. It also requires procedures capable of detecting and assessing risk before removal occurs. Individualised assessment, fair access to protection claims, adequate evidentiary review, suspensive remedies, and effective judicial oversight are not merely administrative safeguards; they are structural elements of substantive compliance. A defective procedure that makes wrongful removal foreseeable may itself amount to a breach of the principle because the legal protection is preventive by design.

The fourth proposition is that the human-rights branch of the doctrine is stricter than the refugee-law branch. This distinction is of central doctrinal importance. Refugee law, while foundational, contains a narrowly framed exception structure under Article 33(2). By contrast, the anti-torture and human-rights branch, particularly under Article 3 CAT and Article 3 ECHR jurisprudence, is generally treated as absolute in its core forms. Once the threshold of torture, inhuman treatment, or irreparable harm is met, balancing against the danger posed by the individual is not permitted. This is one of the clearest fault lines in the legal architecture and remains essential for distinguishing the different normative branches of the principle.

Most current legal disputes no longer concern the existence of the principle itself. Its existence as a central norm of contemporary international protection law is firmly established. The real doctrinal and practical disputes concern the ways in which States attempt to narrow their reach. These efforts frequently appear through externalisation arrangements, pushbacks, offshore processing systems, accelerated procedures, safe third-country presumptions, and semantic relabelling of returns as transfers, disembarkations, relocations, or cooperation measures.

The contemporary challenge is therefore less about whether the principle exists and more about whether its protective force can be preserved against increasingly sophisticated techniques of avoidance. The enduring significance of The Principle of Non-Refoulement lies precisely in its ability to resist such formalistic attempts to dilute its substantive purpose: preventing States from exposing individuals to serious and irreversible harm through removal, in whatever form that removal may be framed.

References

1. Allain, J. (2001) ‘The jus cogens nature of non-refoulement’, International Journal of Refugee Law, 13(4), pp. 533–558.

2. Committee against Torture (2017) General Comment No. 4 (2017) on the implementation of article 3 of the Convention in the context of article 22. Geneva: United Nations.

3. Crawford, J. (2002) The International Law Commission’s Articles on State Responsibility: Introduction, Text and Commentaries. Cambridge: Cambridge University Press.

4. European Court of Human Rights (1989) Soering v United Kingdom, Application No. 14038/88, Judgment of 7 July 1989.

5. European Court of Human Rights (1996) Chahal v United Kingdom, Application No. 22414/93, Judgment of 15 November 1996.

6. European Court of Human Rights (2000) Jabari v Turkey, Application No. 40035/98, Judgment of 11 July 2000.

7. European Court of Human Rights (2008) N v United Kingdom, Application No. 26565/05, Judgment of 27 May 2008.

8. European Court of Human Rights (2011) M.S.S. v Belgium and Greece, Application No. 30696/09, Grand Chamber Judgment of 21 January 2011.

9. European Court of Human Rights (2012) Hirsi Jamaa and Others v Italy, Application No. 27765/09, Grand Chamber Judgment of 23 February 2012.

10. European Court of Human Rights (2012) Othman (Abu Qatada) v United Kingdom, Application No. 8139/09, Judgment of 17 January 2012.

11. European Court of Human Rights (2016) Paposhvili v Belgium, Application No. 41738/10, Grand Chamber Judgment of 13 December 2016.

12. European Court of Human Rights (2020) N.D. and N.T. v Spain, Applications Nos. 8675/15 and 8697/15, Grand Chamber Judgment of 13 February 2020.

13. Gammeltoft-Hansen, T. and Tan, N.F. (2017) ‘The end of the deterrence paradigm? Future directions for global refugee policy’, Journal on Migration and Human Security, 5(1), pp. 28–56.

14. Geneva Convention IV (1949) Geneva Convention Relative to the Protection of Civilian Persons in Time of War. Geneva: International Committee of the Red Cross.

15. Goodwin-Gill, G.S. and McAdam, J. (2021) The Refugee in International Law. 4th edn. Oxford: Oxford University Press.

16. Hathaway, J.C. (2021) The Rights of Refugees under International Law. 2nd edn. Cambridge: Cambridge University Press.

17. Human Rights Committee (2004) General Comment No. 31: The Nature of the General Legal Obligation Imposed on States Parties to the Covenant. Geneva: United Nations.

18. Human Rights Committee (2018) General Comment No. 36 on Article 6: Right to Life. Geneva: United Nations.

19. Human Rights Committee (2020) Teitiota v New Zealand, Communication No. 2728/2016, UN Doc. CCPR/C/127/D/2728/2016.

20. International Court of Justice (2012) Questions relating to the Obligation to Prosecute or Extradite (Belgium v Senegal). ICJ Reports 2012. The Hague: ICJ.

21. International Committee of the Red Cross (2005) Customary International Humanitarian Law. Volume I: Rules. Cambridge: Cambridge University Press.

22. Milanovic, M. (2013) ‘Extraterritorial application of human rights treaties: law, principles, and policy’, in Sheeran, S. and Rodley, N. (eds.) The Oxford Handbook of International Human Rights Law. Oxford: Oxford University Press, pp. 1–25.

23. Orakhelashvili, A. (2006) Peremptory Norms in International Law. Oxford: Oxford University Press.

24. Organization of American States (1969) American Convention on Human Rights. San José: OAS.

25. United Nations (1984) Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment. New York: United Nations.

26. United Nations (1989) Convention on the Rights of the Child. New York: United Nations.

27. United Nations (2006) International Convention for the Protection of All Persons from Enforced Disappearance. New York: United Nations.

28. United Nations High Commissioner for Refugees (1951) Convention Relating to the Status of Refugees. Geneva: UNHCR.

29. United Nations High Commissioner for Refugees (2007) Advisory Opinion on the Extraterritorial Application of Non-Refoulement Obligations under the 1951 Convention relating to the Status of Refugees and its 1967 Protocol [online]. Available at: https://www.refworld.org/policy/legalguidance/unhcr/2007/40854 (Accessed: 30 March 2026).

30. United Nations High Commissioner for Refugees (2025) Refugee Convention, 1951 [online]. Available at: https://www.unhcr.org/1951-refugee-convention.html (Accessed: 30 March 2026).