Fundamental tensions and productive instability: The refugee definition of the 1951 Refugee Convention at 70 through a queer theory lens

Along with the 1951 Convention, the definition of who qualifies as a refugee turns 70 this year. This blog post uses a queer theory lens to uncover various tensions at the very core of the refugee definition. Using the resilience of the problematic ‘discretion’ reasoning in refugee status determinations in Germany and France as an example, it illustrates how these overlapping tensions create instability, for claims based on sexuality as well as other grounds. The post argues that this web of contradictions can be used productively in favour of each individual claimant – and helps explain the adaptability and continued relevance of the 1951 Refugee Convention 70 years after it was drafted.


In June 2021, Der Spiegel, a major German news magazine, published an article outlining how German decision-makers regularly reject asylum claims of gay applicants on the basis that they can return home and live secretly in such a way that they would not come to the attention of the persecutors in their home countries – where homosexual acts may be criminalised, or homophobic violence rampant. The logic is that gay people can avoid persecution if the surrounding society, and hence potential persecutors, do not know that they are gay. Such decision-making practice persists regardless of a 2013 judgment by the Court of Justice of the European Union, holding that asylum seekers cannot be rejected on the basis that they could ‘reasonably be required’ to be ‘discreet’ in their country of origin. In France, in contrast, previous research revealed an ‘indiscretion’ requirement, where claimants would only receive protection if they had openly expressed their sexual orientation in the past – otherwise, they would be returned to (continued) ‘discretion’.

It is striking that this decision-making practice has developed based on a refugee definition that was drafted 70 years ago – a time when homosexuality remained criminalised in most countries and asylum claims by gay people were not even fathomed. So the refugee definition has proved to be flexible enough to encompass reasons for persecution that were not thought of by the drafters. But while sexual orientation was not foreseen as a Convention ground from the outset, ‘discretion’ logics are almost as old as the definition itself. In his groundbreaking 1966 study of the earliest refugee status jurisprudence, Atle Grahl-Madsen already discussed this issue in the context of cases relating to draft evasion and the political opinion Convention ground.

In this blog post, I argue that the two distinct ways in which ‘discretion reasoning’ emerges in German and French decision-making practice are each paradigmatic for the broader intricacies of ‘discretion’ logics which are in fact inherent in the refugee definition from the 1951 Convention. They thus serve to draw out some broader observations regarding the refugee definition, which are particularly acute in, but extend beyond, sexuality-based claims.

I will proceed as follows: I will first briefly outline decision-making practice in both jurisdictions and then argue that these ‘discretion logics’ respond to three specific tensions within the refugee definition: The clash between the fundamentality principle and the severity argument, the Convention’s double rationale regarding persecution and Convention grounds, and the shifting focus for defining the Convention ground, fluctuating between the claimant and the persecutor perspective. Finally, I will submit that these tensions are part of the explanation of why the Refugee Convention from 1951 remains live and kicking today, and argue that although they cannot be undone, they can be used productively in favour of each particular claimant.


‘Discretion’ reasoning is entrenched in both German and French decision-making practice

‘Discretion’ reasoning emerges from decision-making practice in both Germany and France. The comparison of these two countries is striking, since in both jurisdictions, there is a tradition of ‘discretion’ logics – but the underlying reasoning emerges as opposed: Whereas France has traditionally rejected the asylum claims of gay asylum seekers and sent them back to discretion based on an absence of public expression of their sexual orientation, Germany has a tradition of returning claimants if their sexual identity isn’t deemed to be sufficiently stable.

In France, the focus on the claimant’s behaviour was explicitly introduced in the first definition of social group relating to sexual orientation in the late 1990s. The definition required that the claimant had to ‘be gay and assert their sexual orientation or manifest it in external behaviour’.

The notion of external manifestation required that the person had publicly revealed their sexual orientation or presented external ‘signs’ that rendered the sexual orientation visible and evident in the eyes of the society in which they lived – ie, to the persecutor. Based on this requirement, in what has been termed ‘discretion requirement in reverse’, essentially requiring ‘non-discretion’, recognitions were granted to those claimants who had manifested and asserted their sexual orientation, whereas those who had maintained their sexual orientation secret were rejected.

In these cases, even though a claimant may be accepted to be gay, and persecution of gay people is established in the relevant country of origin, the social group was understood to extend only to those who also expressed their sexual orientation (to the persecutors).

In contrast to France, Germany has a tradition of favouring identity over acts. In that sense, German practice was closer to ‘classical’ ‘discretion’ reasoning. In a leading 1988 judgment, the German Federal Administrative Court developed the distinction between – variously – ‘irreversible’, ‘fateful’ or ‘inescapable’ homosexuality on the one hand and ‘latent’ homosexuality on the other (Urteil vom 15. März 1988, 9 C 278/86). In the former case, a person is thought to be determined by their sexual orientation to such an extent that they have no choice but to satisfy their same-sex sexual urges (‘homosexuelle Triebbefriedigung’), in the latter case, they have a ‘mere inclination’ (bloße Neigung) towards same-sex sexuality and can choose whether or not to engage in same-sex sexual acts. Based on this classification, for years, only those claimants who could convince Courts that their identity was sufficiently and irreversibly determined by their sexual orientation would regularly be granted protection – regardless of whether anyone, including a potential persecutor, might be aware of this.

In this reasoning, even though a claimant might engage in same-sex sexual behaviour, this would not warrant protection if their identity was not considered to be sufficiently determined by the same-sex sexual orientation. Here, the social group only extends to those claimants whose identity was deemed sufficiently stable.

So in both jurisdictions, without formulating a requirement of discretion (which was established in UK jurisprudence and led to the 2010 UK Supreme Court HJ (Iran) and HT (Cameroon) judgment) ‘discretion’ emerges from the act/identity dichotomy by default. In France, those who had not manifestly expressed their sexual orientation were sent back to (continued) ‘discretion’, and in Germany, those whose identities were not sufficiently shaped by their sexual orientation were also sent back to (continued) ‘discretion’.

Interestingly, the CJEU judgments that rejected the ‘discretion requirement’ from 2012 (Y and Z on religion) and 2013 (X, Y and Z on sexual orientation) have in substance had a negligeable impact on decision-making practice in both these jurisdictions. After a moment of uncertainty, both jurisdictions appear to have found ways to incorporate the new developments in jurisprudence and guidance from the CJEU into their decision-making cultures.

In Germany, courts continue to assess whether the claimant’s identity is sufficiently shaped by their sexual orientation. Decision-makers have (mis)interpreted paragraph 70 from the Y and Z judgment on religion, to the extent that in the assessment, it is required that ‘the behaviour concerned must be significant and particularly important for the identity of the claimant’. Although this paragraph does not appear in the subsequent X, Y and Z Judgment on sexual orientation, which is of course the much more relevant judgment for this type of claim, this reading has allowed decision-making practice to essentially remain the same – the notion of ‘irreversible determination’ was simply replaced by ‘particular importance’.

In France, while courts no longer explicitly require past public manifestation, they continue to make a backwards-facing risk assessment, where claimants will have difficulty in establishing risk if they have not suffered from persecution already in the past – which will obviously only occur if they have been outed as (or imputed to be) gay, and so, have not passed unnoticed. This is based on the notion that the social group is defined by the way in which these persons are viewed by the surrounding society or by institutions. However, in order to qualify for protection, the claimant has to explain why they personally are at risk, that is, they must explain why they would be (considered to be) a member of the group by the persecutor, and this is hard to do if they have not already been outed/imputed as a group member in the past because the question is not whether they in fact are gay, but whether they will come to the attention of persecutors – because it is their perception that is relevant.

The remainder of the blog post will argue that these opposing traditions in German and French decision-making practice are each paradigmatic examples for two different approaches to the intricacies of ‘discretion’ reasoning which arise from a wider web of tensions inherent in the refugee definition. I will illustrate this with regard to three different tensions.


1. The clash between the fundamentality principle and the severity argument

The first tension is the clash between the fundamentality principle and the severity argument. The debate on ‘discretion’ is characterised by two central notions of refugee law, which are in principle un-controversially shared in refugee law doctrine: I call these the ‘fundamentality principle’ and the ‘severity argument’.

The fundamentality principle consists in the generally accepted idea that no one should be required to hide, change or renounce the attributes or opinions that are persecuted – this can be found among others in UNHCR Guidelines and numerous judgments. The rationale is that if hiding were required, then there would be no need for refugee law at all.

The severity concern, in turn, is based on the shared understanding that refugee law provides surrogate protection for serious harms. The logic is that the task of refugee law is not to guarantee the same level of rights and freedoms everywhere in the world – as human rights law generally would – and, therefore, to remedy even minimal or trivial harms. Rather, refugee protection extends only to injuries of a certain quality.

Refugee lawyers generally subscribe to both these notions. The problem is: They clash. ‘No hiding’ (inherent in the fundamentaliy principle) on the one hand, but ‘not the same freedoms’ (severity argument) on the other is difficult to reconcile. ‘Not the same freedoms’ requires some measure of hiding. And ‘no hiding’ requires the same freedoms. This tension creates an intricate situation for the Convention grounds: hiding is required and forbidden at the same time.

We see that Germany and France each approach this tension differently. In France, the emphasis on the behaviour, the externalising act, appears to favour the fundamentality principle – there should be no requirement to hide; if they decide to act then they are protected. Whereas the German jurisprudence appears to favour the severity concern – only if the sexual orientation is of such importance to their identity that they cannot go without then it is severe enough that they are entitled to protection.

Thus, in a situation of conflict, both jurisdictions opt for favouring the opposing elements. This happens through the separation of the Convention Grounds into acts on the one hand and identity on the other. So the act/identity dichotomy serves to respond to this incoherence. The persecuted characteristic (sexual orientation) is split into act and identity. As we have seen for Germany and France, either at times can serve as the Convention ground – and therefore, to uphold the fundamentality principle and the rejection of ‘discretion’, while simultaneously allowing the other element to float free such that it can expand or limit the definition elsewhere, and thus ensure that the severity concern is met through a degree of ‘discretion’. The claimant is thus entitled to the Convention ground and restricted from it in conflicting ways.


2. The Convention’s double rationale: persecution and Convention grounds

Another tension that plays out in the jurisprudence is the Refugee Convention’s paradoxical double rationale. Upon close inspection, it is not exactly clear what the Refugee Convention intends to protect. It protects from persecution, carried out by perpetrators, on account of a reason, located with the claimant.

So is the main motivation of refugee law to be seen in the protection of the Convention grounds (given that persecution for other reasons is not protected)? Or is it the protection from persecution (given that other harms on the basis of the Convention grounds are not protected)?

The fact that both are laid out in the definition as necessary conditions creates a dilemma in situations of conflict: persecution is relevant only if it is due to the Convention ground and the Convention ground is relevant only if it is met with persecutory harm. When one is not given, the other also becomes irrelevant.

When looking at decision-making practice in German and French jurisprudence, it appears that here too, both jurisdictions prioritise opposing elements.

The French approach with its focus on externalising behaviour appears to be particularly concerned with protection from persecutory harm. It is not much fussed about protecting the sexual orientation as such, but rather zooms in on those situations which are likely to lead to harm – or rather, have already led to harm. People who have always remained in the closet and never been imputed a gay identity by others are of no concern to the French system.

In contrast, German jurisprudence with its focus on the identity is mainly concerned with protecting the Convention ground – acts from someone whose sexual identity might be considered to be ‘fleeting’, which could very well lead to persecutory harm, do not matter much because in their assessment it would not arise from a determined and stable gay identity. But someone who has secretly suffered due to their sexual orientation would be protected.

Accordingly, the positioning in response to the tension between Convention ground and persecution can differ with real consequences for the claimants concerned. This is also the case for the third and last tension I want to briefly highlight.


3. Focusing on the claimant vs. focusing on the persecutor for defining the Convention ground

The third tension that is built into the refugee definition and surfaces in this jurisprudence relates to the question of whether to focus on the claimant or the persecutor in order to understand what is persecuted. As stated above, the Convention protects from persecution, carried out by persecutors, on account of a reason, located with the claimant – and over which the claimant in many cases has a degree of control.

Thus, the claimant is in a position to reduce the prospects of serious harm precisely by exercising control over the reason which triggers this harm, for example by seeking to hide it entirely or by manifestly asserting it in external behaviour. But the claimant is never in full control. Because persecution must be for a reason, the persecutor will only submit the claimant to harm relevant under the Refugee Convention if they identify and take them to possess that characteristic, but it remains the persecutor who determines the parameters of what (and therefore who) is persecuted. It is by no means clear that the persecutor will correctly identify the ‘right’ persons as gay. Nor is it clear which signifiers they will use for that assessment. Therefore, the Convention does also protect those who are (wrongly) imputed to be gay.

If we try to define the Convention grounds by establishing first, in abstract terms (often by reference to human rights), what they entail, and then second, whether the claimant possesses these characteristics, this whole exercise risks being completely out of step with the persecutor’s notion of what is, say, political, or gay – and therefore, with the group that is in fact targeted by the persecutor. On the other hand, an approach that relies entirely on the persecutor’s perception, without establishing the claimant’s actual identity, has difficulty in identifying those individuals that would be at risk. Therefore, such an approach is prone to singling out logics – where a claimant has to show that they have been or would be singled out for persecution – which is often connected to a backward-looking analysis requiring past persecution.

This tension also visibly plays out in German and French decision-making practice. The French approach, which defines the persecuted group entirely by the persecutor’s perception falls prey to the backward-looking analysis, essentially requiring past persecution in order to establish group membership on the basis of outing or imputation. German practice, in turn, with its focus on the claimant’s deep inner identity to establish group membership applies a definition of the protected group which is potentially much narrower than the actually persecuted group, by excluding those who are not ‘sufficiently’ gay.


Conclusion: Productive instability?

The ways in which German and French jurisprudence rely on ‘discretion” reasoning reveal some fundamental tensions at the very heart of the refugee definition. While these tensions become acutely visible in sexuality-based claims, they also play out in other types of claim, such as those based on religion or political opinion – they emerged in Grahl-Madsen’s early struggles with political opinion claims and are also reflected in the contemporary competing approaches to interpreting membership of a particular social group. The malleability that this web of contradictions provides is not necessarily only a bad thing. Firstly, while one of the elements in each of the tensions is always preferred, the submerged element is also present at the same time and does some work, sometimes unnoticed. I have said this with regard to the act/identity dichotomy, but the same is true for the submerged elements in the other tensions. While this work may be exclusionary or reductive in many cases, there may also be quite some emancipatory potential in these tensions. Where it is possible to harness the submerged elements to counter a predominant approach in a given situation, that may enable broader protection. The instability inherent in the refugee definition can be used productively in favour of each particular claimant.

Secondly, the self-contradictory nature of the refugee definition and its consequent malleability may well be one of the reasons why the refugee definition has been able to withstand time and adapt to changing circumstances – and with it, the 1951 Refugee Convention as a whole.


The arguments outlined here are developed more fully in the author’s monograph ‘The Concealment Controversy – Sexual Orientation, ‘Discretion’ Reasoning and the Scope of Refugee Protection’ (CUP 2021). This blog post is also published on Refugee Law Initiative Blog.

Teilen Sie den Beitrag