András Sajó is a former judge at the European Court of Human Rights. He is currently a University Professor at Central European University. He previously was a visiting professor and guest lecturer at University of Toronto, Columbia University, University of Chicago Law School, NYU School of Law, Cardozo School of Law (New York), Harvard Law School, Bocconi University (Milano), and other institutions. Professor Sajó received the Hungarian Academy of Sciences Book Award (1989) and The Blackstone Lecture at Oxford University (1993). He holds Honorary Doctorates from the European University Institute (Florence) and Ilia State University (Tbilisi). Professor Sajó has been involved in legal drafting throughout Eastern Europe. In addition, he participated and/or advised in drafting the Ukrainian, Georgian, and South African constitutions.
Professor Sajó has published extensively on constitutionalism, the rule of law, theory of democracy, the secular state, human rights, and media regulation. His recent publications include Routledge Handbook of Illiberalism (2022, co-edited with Renáta Uitz and Stephen Holmes), Ruling by Cheating: Governance in Illiberal Democracy (2021), Constitution of Freedom: An Introduction to Legal Constitutionalism (2017, co-authored with Renáta Uitz), and The Oxford Handbook on Comparative Constitutional Law (2012, co-edited with Michel Rosenfeld). In 2020, professor Sajó was appointed to the Meta Oversight Board, an independent body to which people can appeal if they disagree with corporate decisions about Facebook or Instagram content.
András Sajó was interviewed by Dmytro Vovk.
Do you see any comprehensive doctrine of freedom of religion or belief (FoRB) in European Court of Human Rights jurisprudence? Or does the Court use a more case-by-case approach and, as you said in one of your works with respect to national constitutional systems, has the Court “avoided arriving at unequivocal answers to the question of church-state or religion-state relations?”
If you try to make an analysis of ECtHR jurisprudence, you will see that it is significantly circumstantial. The Court got into the FoRB area relatively late, in Kokkinakis v. Greece (1993). It has a rather strong and relatively coherent approach when it comes to individual exercise of religion. However, when it comes to areas where FoRB conflicts with other rights like freedom of speech or non-discrimination, the Court’s jurisprudence becomes somewhat less clear. Think about employment and labor rights. Although the Court has decided on the matter, I’m not so sure that this is the final word, particularly because of the latest developments in the European Court of Justice, which seems to go in a different direction. The same is true regarding church-state relations. Public secondary education is an obvious example here. Compare Folgerø and Others v. Norway (2007) with Lautsi v. Italy (2011), and you will see the same principles resulted in very different conclusions. So there are some guiding principles, but the way they apply is to a very great extent determined by national differences. That makes it much more akin to a case-by-case approach.
In your keynote speech at the 2017 ICLRS Law and Religion Symposium, you mentioned the shift in the freedom to manifest religion from a matter of liberty toward being a matter of personal identity. Can you explain this shift and its consequences for ECtHR jurisprudence?
Traditionally, religion was thought as something to some extent objective in the sense that most religions had a hierarchy and doctrine—something that was easily identifiable. (I mean religions exercised in Europe; this is not applicable, for example, to Hinduism.) People understood religion as a collective exercise; manifestation of religion was considered as a matter of belief and relation to the religious community, the organized church, and, at the top of this structure, the deity. This approach has already changed. To illustrate the difference, I would like to refer to SAS v. France (2014), the burqa case, where the applicant says that she’s wearing the burqa with the aim of “satisfying herself and her conscience.” Her wish to wear this attire was not based on a religious command but on her personal identity, which allowed certain behavioral modifications. Thus, her religious conviction was understood as a matter of identity. The ECtHR reflected upon that because the Court accepted that wearing the burqa was not simply a freedom of religion issue but also a personal life issue, which is technically correct.
However, this individualization or personalization of religious convictions deviates from a fundamental historical trait of freedom of religion. Basically, religious freedom was probably the first ever fundamental right recognized, a right that received specific protection in Europe because religion emerged from or resulted in social conflicts. Whenever religion dictated commitments strong enough to provoke violence, the whole issue became particularly important. It’s hard to imagine someone going to war for an identity issue, although debates over identity may go very far and provoke conflicts. But religion had a very specific characteristic of possessing anonymous emotional spiritual power, and exactly because of this power, freedom of religion became so strongly protected. That’s where the difference lies, to my mind. From a social perspective, it’s hard to imagine that identity requires a protection of the same intensity. It’s just a matter of personal choice, a part of “I want to live the way I like to live,” while traditionally freedom of religion meant the obligation to live according to some kind of impersonal, divine or other external, command. I’m not arguing that it’s good or bad, but this is something that is now increasingly reflected in ECtHR case law and, I suppose, in a number of other jurisprudences.
People understood religion as a collective exercise; manifestation of religion was considered as a matter of belief and relation to the religious community, the organized church, and, at the top of this structure, the deity. This approach has already changed.
Religious Freedom and Constitutionalism
In one of your papers, you point out that “a stronger version of the secularist constitutionalism theory would assume that there can be no constitutionalism without some level of secularism, nor can it exist outside a secularized (modern) society.” Does that mean there is no (real) constitutionalism beyond Europe, the U.S., and some other countries?
Let me start with one clarification. In the American usage, secularism is basically understood as militant atheism or even anti-religion. This is not the standard meaning of secularism, in my view. It is not equal even to French laïcité. Secularism means a public sphere that is not dominated or decided on religious premises, especially if they are defined by one church. Under the secularist regime, religion is present and has legal possibility to be present in the public discourse. But it cannot obtain or occupy a privileged position. That’s the meaning of secularism that I have in mind.
Now, if we accept that constitutionalism is per se liberal in the traditional European sense and reflects the primacy of the individual, then we inevitably have to conclude that it’s very difficult to imagine a liberal and therefore constitutional, or constitutional and therefore liberal, society without a rather high level of secularism. It doesn’t mean that this is to be reflected in the text of the constitution. You can have a state church and still have a secular society with a secular legal system that does not grant privileges to specific religions’ belief systems. The main issue is whether religion has power to determine the public sphere. If it does, this is detrimental for constitutionalism, especially for such a fundamental tenet of constitutionalism and democracy as the supreme power of the people. If there is an alternative to the people being the source of power, then you have a different system. This system can be tolerant and reflect the understanding of the public role of religion prevailing in the society. Besides, I’m not advocating that secularism is necessary for a good life. But its absence is very problematic for constitutionalism.
The main issue is whether religion has power to determine the public sphere. If it does, this is detrimental for constitutionalism, especially for such a fundamental tenet of constitutionalism and democracy as the supreme power of the people.
To what extent is your thought about links between secularism and constitutionalism (including protection of human rights) relevant to the ECtHR, whose jurisdiction covers some countries with a different or even opposite approach to this issue (Turkey, Russia , and some other countries, for example)?
The ECtHR does not ignore national differences. Think about the examples of Russia, Austria, and Turkey. Austria has a very restrictive regime of church registration. The Court took some steps to attenuate it but was basically reluctant to touch the heart of the matter, which was the registration of minority religions, and by minority, I mean both numerical presence and social recognition. In Russia, restrictions imposed on religious organizations are much more obvious. They target the heart of free exercise, while in the Austrian case we still have free exercise with some hurdles. Hence, the Court has an easier task dealing with Russian cases. What strikes me in Turkey is that this is a constitutionally secular country where, in the name of secularity, millions of Alevi people are treated by Turkish administrative law as second-class believers. In a number of judgments, the Court claimed that is not acceptable and constitutes discrimination in the context of freedom of religion. (Very little has happened in Turkish law, however, since then.) The Alevi case is very different because it occurs within the frame of a secular constitutional system. In Russia the constitution is formally neutral, but there are still certain elements of support of Christian Orthodoxy. The Turkish constitution is clearly a version of laïcité. Can you apply the same principle to these cases when the context is so different? Yes, but you have to take very different routes to argue violations of FoRB.
Let’s talk more about Turkey. In the Refah Partisi case (2003), the Court stressed the importance of secularism for the survival of a democratic regime and, based on this argument, justified the ban of this political party imposed by the Turkish Constitutional Court. This decision was criticized by some observers as non-democratic. From the perspective of 2022, was the Court right in confirming that the state can ban a political party merely for its program, even if this program is incompatible with secularity?
There is a theory called “militant democracy” that emerged in the 1930s when Hitler and other far-right parties were already close to coming to power in Europe. The ECtHR has basically applied a version of militant democracy, which means that the state is not supposed to allow enemies of democracy to come to power using democratic tools—this is the underlying concept, which was not invented by the Court itself. The Refah Partisi case was a complex one. It’s not a religious freedom case per se, but it deals with a fundamental problem of religion, namely, to what extent religion or dictates of religion, which are not or may not be democratic, should be tolerated by the state. In other words, is such a religion an appropriate basis for a political movement? While answering this question, you are getting into the realm of religious freedom because these non-democratic premises are to some extent based on or commanded by religion and manifested in religious practices by those who would like their religion to be more present in society. If they turn it into a political platform, is this compatible with democracy and constitutionalism? Is the politicization of religion through the establishment of a fundamentally religious political movement consistent with democracy and constitutionalism? Those were the main questions in the Refah Partisi case, in my view.
To answer those questions is a very difficult matter because it is based on a certain forecast of what the banned party would do if it had not been banned. Courts are very poor in forecasting the future. It’s more a political issue than a judicial issue. But in the Refah Partisi case the Court was in a comfortable situation as it could side with the Turkish Constitutional Court. It was the latter that applied a political evaluation, so the margin of appreciation doctrine became a helpful tool for the ECtHR in this case. (I am not a great friend of this doctrine, although I know this is a minority position.) In the judgment delivered by the Turkish Constitutional Court there was some analysis that the operation of Refah Partisi created a real political risk. Thus, the ECtHR position in this case is defensible, in my view. It’s a risky legal tool, of course. But that’s the nature of militant democracy.
Religious Freedom in the Private Sphere
One of the important divisions in constitutional law and human rights law is the division between private and public spheres. How does this division play out in the context of constitutional secularism and religious freedom?
There is very convincing scholarship arguing that the borderline between public and private is disappearing and this division is not so strong anymore. I am a bit old fashioned with respect to this issue and like to quote Benjamin Constant, a major French liberal philosopher of the beginning of the nineteenth century. Constant argues that there is a fundamental sphere of private life, which is off limits to the state. In my view, this sphere still exists even though its boundaries are not always clear. Generally, the state has nothing to do with religion as an element of the private sphere. Of course, there are some disputable areas. Think about parents’ right to indoctrinate their children. The European Court recognized that there are certain limits to what parents can do. The UN Convention on the Rights of the Child goes way beyond that and clearly separates the parents’ and the child’s rights to freedom of thought, conscience, and religion. Hence, the indoctrination of children is not just a matter of the private sphere, although the scope of permissible state interference still remains disputable. I am very concerned about state paternalism and excessive interference, but certainly even the relationship with children is not an area absolutely protected from this interference.
But what is most interesting and problematic is the area between public and private spheres. Think again about employment and labor relations. Today they are clearly not a matter of private law only. They can’t be regulated merely on the basis of the will or arbitrariness of the employer. To what extent private considerations prevail over public interests is an open question, especially when we are dealing with discrimination issues. For example, in the realm of FoRB the question arises: at which point does a ban on wearing religious garb at the private workplace hamper free exercise and turn into discrimination? If wearing religious attire is an identity issue, then a significant protection to a person claiming to have a specific religious identity ought to be granted. There is also an alternative approach: if the duty to wear religious attire is based on a clear religious command, then this duty shall be accommodated to the maximum possible extent. If there is no clear command, then the employer enjoys more discretion. The ECtHR, as well as other courts, are very reluctant to use this approach because they are afraid to side with a specific meaning of a certain religious command. They are afraid to take sides in such issues. Even in centralized religions like Catholicism there are some divisive issues. The Pope said one thing, but the bishop says something else, or the previous Pope said something else, or the Pope said nothing, or it’s not clear what he said. Then a court, deciding a case, finds itself in a difficult situation of being potentially more papist than the Pope. Of course, courts would like to avoid this. This can be done by not going into an objective analysis of religious convictions and accepting personal beliefs as such: whatever you believe is welcome. This approach is very much based on judicial practicalities, but it also reflects the fact that religion has been turning into a very personal, individualized matter: I believe the way I like to believe. I am neither criticizing nor praising this. It’s simply a social fact that courts have to deal with, although they may not recognize it expressly and justify their reluctance to provide objective analysis of religious convictions by reference to church autonomy.
The Pope said one thing, but the bishop says something else, or the previous Pope said something else, or the Pope said nothing, or it’s not clear what he said. Then a court, deciding a case, finds itself in a difficult situation of being potentially more papist than the Pope.
I should also say that employment and free exercise of religion constitutes a major problem in a number of European countries where churches are important employers. Consider Germany where various churches and denominations provide employment for about 1.3 million people. What is the nature of these relations? Don’t forget also that these relations have a third-party effect, as churches run huge numbers of charities and hospitals; many schools and universities across Europe are denominational. So, as I said, these relations are neither public nor private; they are something in-between. And there is much that is ad hoc in the regulation of these relations, if I may use this term. Thus, there is no surprise that the ECtHR, the EU Court of Justice, and many national courts take different positions while considering cases on this matter.
Let’s elaborate a bit more on the legal nature of churches and religiously affiliated businesses as employers. The Court made a few judgments on this issue (see paragraphs 241–42 of the ECtHR Guide on Article 9). Although this jurisprudence is not fully consistent, it looks like the Court takes the approach that, in contrast to secular employers, churches and religiously affiliated employers can impose some additional (ethical or organizational) requirements on some employees based on their teachings and doctrines. Do you think that this distinction between secular and religious employers is justified?
Well, the Court’s answer seems to be, “It depends on the circumstances.” Partly because of margin of appreciation and partly because of respect for church autonomy, the Court is inclined to accept that the church as an employer can have specific ethical requirements for its employees. If the position of the church on a certain issue is not clear or the employer abuses its power, it violates the rule of law, and the Court can accept it as such. But if religious ethical rules are clear—and by “clear rules” I mean certain very important standards of behavior—this is a matter of loyalty to the church, and the Court can accept their imposition on employees and recognize that this is more or less a matter of the church’s discretion.
The ECtHR approach based on church autonomy is not in line with the EU Court of Justice anymore, and most likely EU members (for example, Germany) will have to follow the Luxemburg Court’s approach.
However, this approach is challenged by recent EU Court of Justice judgments like Egenberger (2018) and an increasing number of other cases. In these cases, the Luxemburg Court concluded that the discretion of the church as an employer is more limited. (This contradicts the American approach where the ministerial exception, as interpreted in Hosanna-Tabor (2011), covers, in my view, almost everything.) Then, regarding religious freedom in the workplace, the EU Court of Justice seems to have a different perspective. In one recent case, this court says that secular private employers do have the right to restrict freedom of religious manifestation when necessary, to put it simply, for business purposes. Thus, in its reasoning the Luxemburg Court seems to utilize the functional approach: while the religious employer can require loyalty or conformity (but the issue of discrimination can arise), the secular private employer also enjoys this functional autonomy of regulating religion at the workplace. This means that, according to the Luxemburg Court, EU law limits religious employees’ discretion, with some anti-discriminatory requirements prevailing over national regulation.
Thus, the ECtHR approach based on church autonomy is not in line with the EU Court of Justice anymore, and most likely EU members (for example, Germany) will have to follow the Luxemburg Court’s approach. Then the question will arise whether, when following the EU Court of Justice’s approach, these countries will violate Article 9 of the European Convention of Human Rights as it is interpreted by the ECtHR. In my view, this is one of the most interesting developments of multilevel human rights law. Maybe it’s not the most important aspect of religious freedom, but it’s certainly a very important problem in terms of social functions of religion.
FoRB and Gender Issues in the Strasbourg Court’s Jurisprudence
A significant part of European jurisprudence on freedom of religion or belief are the headscarf and veil cases. In S.A.S. v. France the government and some third parties argue for the ban on wearing the niqab or burqa, particularly because they symbolize repression of women and gender inequality. A similar argument is also utilized by many feminist organizations, while some religious organizations and women of faith, as well as the UN Human Rights Committee, consider such bans discriminatory and criticize them as “selective pluralism” hostile to religious minorities. What do you think about this debate?
I find the UNHRC’s argumentation reasonable. However, the problem that the French government and other governments refer to in order to justify this sort of ban seems to be of relevance. We see some instances where the practice of wearing the burqa is oppressive and supported by social sanctions imposed by women’s families. In such circumstances women might feel cognitive dissonance and think that the only way out is to claim that this attire is part of their identity. This claim seems not to be based on God’s command or the conviction that wearing the burqa is the only way to be a good believer but rather is based on the identarian feeling of pride, which, by the way, they may really feel. From a purely legalistic perspective, if the state can reliably demonstrate that the practice of forcing women to wear the burqa is widespread in the society and those who refuse to wear it suffer from social ostracism or even worse consequences, then there will be a strong argument in favor of such a ban. Can the state demonstrate it? I think it would be very difficult to do it, and the jury is still out on this matter. Let me also add that the burqa case has become a symbolic issue. I might be naïve, but it’s not in any way a central issue either for French “living together” or for freedom of religion.
The analysis will be different if we take the case of Leyla Şahin (2005), a Turkish Muslim female student who was prohibited from wearing a headscarf on the premises of her university. The ban was imposed through a circular issued by the Vice-Chancellor of Istanbul University, which, of course, is a very questionable source for imposing restrictions on human rights. However, if you go beyond the legalistic aspect of the case, the role of political Islam in Turkey is again at stake. No doubt, wearing a headscarf at Istanbul University would not constitute a problem at that time. But think about wearing headscarves at universities in eastern Turkey, where the majority of the population was rather pious and put significant pressure on the minority of secular women to compel them to wear headscarves. I think the Court took this argument in consideration while confirming the restriction imposed on Leyla Şahin. In other words, the majority of judges agreed that the restriction imposed on Şahin enables freedom for other women. I am not defending this argument. A Belgian judge, Françoise Tulkens, wrote a famous dissenting opinion with strong counterarguments based on a feminist perspective. But this is also a part of the story. One can have an ad hoc solution for Leyla Şahin allowing her to wear a headscarf at Istanbul University 30 years ago, but one must also consider how to protect those women in eastern Turkey who did not want to wear a headscarf. Is it possible to have a general rule in these circumstances? Or are you going to proceed with local rules for each situation? I’d naïvely say that local rules are better, but law is not about local rules.
One can have an ad hoc solution for Leyla Şahin allowing her to wear a headscarf at Istanbul University 30 years ago, but one must also consider how to protect those women in eastern Turkey who did not want to wear a headscarf.
It seems that religious freedom issues divide Western and Eastern Europe. Even within the European Union approaches to anti-discrimination policies, family rights, and religious exemptions, the interference into religious expression in the interest of non-discrimination and tolerance are rather different. Compare Germany with Hungary or the United Kingdom with Poland, for example. What is the role of the ECtHR in addressing issues that divide Europe?
Before I try to answer this very difficult question, let me mention a curious fact. Hungary is one of the countries that has been increasingly promoting traditional Christian doctrines on family, sexual relations, childbirth, and, to a significantly lesser extent, abortion. And this trend is clearly shared by most post-communist countries. At the same time, Hungary was one of the very first countries in the region to introduce civil registration of marriage in 1895. France introduced this institution only 10 years later, in 1905. So there is a strong liberal tradition of regulating family relations in Hungary. But notwithstanding this liberal legal tradition, something has changed. That’s why these processes do not merely reflect path dependence; they have something to do with internal politics.
The ECtHR is of the view that, in the most contested areas, the Court doesn’t have a mandate to become a final arbiter. I am not sure if it is the right approach, but this is how the Court perceives its role.
Going back to your question, the Strasbourg Court will definitely face it at some point. But the Court is very slow on these issues. As you know, contrary to existing case law claiming that European consensus matters, the Court invented the idea of “a very sensitive issue” and, with certain reservations, rejected recognizing that the Irish constitutional ban on abortion violates the European Convention on Human Rights. The same is true with respect to a very restrictive Polish law. The Court is still considering that case. Overall, the ECtHR is very reluctant to endorse reproductive rights. I am not saying that the Court is right or wrong. I am trying to explain that the Court doesn’t feel the legitimacy to determine these fundamental human rights issues, which are still under discussion in European societies. The ECtHR is of the view that, in the most contested areas, the Court doesn’t have a mandate to become a final arbiter. I am not sure if it is the right approach, but this is how the Court perceives its role.
 Russia was still a member of the Convention system at the time of the interview.