Multiculturalism and secularism: Theoretical understandings and possible conflicts

Introduction

This paper analyses the relationship and the structure of possible conflicts between multiculturalism and secularism. This task is motivated by a development in discussions about multiculturalism, which have increasingly (especially in Europe) focused on religious minorities (particularly Muslims). Multiculturalism denotes positions about how to handle societal diversity that argue for accommodation of minority groups. When multiculturalism focuses on religious groups, the accommodation in question is of religious practices, and claims for accommodation increasingly appeal to the religious importance of such practices. Examples of this include cases such as exemptions from animal welfare requirements for halal or kosher butchering, the permissibility of religious symbols and clothing in public settings, public support for minority faith schools, permissions for the building of mosques and minarets, and public consultations with, and recognition of, organised minority faith communities. In such cases, multiculturalists argue in favour of exemptions, permissions, support and recognition.

This development threatens to conflict with secularism requiring separation of politics and religion. This possible conflict leads to a criticism of multiculturalism different from conservative or nationalist objections, since secularism is often a liberal view appealing to human rights and enlightenment values (Levey, 2009; Triandafyllidou et al., 2012: 7–10).

The aim of this paper is to use the examination of this possible conflict to focus on the general understandings of multiculturalism and secularism, and to show how multiculturalism and secularism can be, and are, understood in quite different ways. I argue that some of these understandings are more plausible and fruitful than others. I focus on the general structure of theoretical understandings of multiculturalism and secularism and show how some of these provide richer theoretical resources for understanding and discussing accommodation of religious minorities. The aim is to explicate general theoretical understandings of multiculturalism and secularism that do not foreclose normative debates about religious accommodation and provide us with theoretical categories and tools that can be used to structure and conduct the normative debate.

I start by explaining why there might be a conflict between multiculturalism and secularism. I then consider two common understandings of multiculturalism – one as a label for a specific set of policies and one as a view defined as ‘going beyond’ liberalism – and note some problems with them. I then propose a more general understanding of multiculturalism as accommodation of minorities, which need not be understood as a view going beyond liberalism but, rather, as an interpretation of basic liberal values and principles. I go through a similar exercise in relation to secularism, during which I also reject understandings of secularism as simply a label of policies and as a specifically liberal view. Again, I suggest a more general and structural understanding of secularism. In the conclusion, I sketch how the two proposed understandings of multiculturalism and secularism provide a framework for discussing possible conflicts. This approach shows that the conflict between multiculturalism and secularism is not a general one, but rather one that can appear in very specific respects depending on which particular claims are inserted into the general understandings of the two positions, and how the conflicts might not be as many or as intractable as one might otherwise think.

Why there might be conflicts between multiculturalism and secularism

The idea of separation, i.e. that politics and religion should operate in their respective spheres without either unduly influencing the other, is prominent in the history of liberal political thought and in the self-understanding of many contemporary liberals. But at this level of description, the idea of separation is extremely vague and mostly metaphorical – everything here depends on what we understand by ‘politics’, ‘religion’ and ‘separation’. But the idea of separation is also connected to more specific political principles dear to liberals, such as freedom of religion, non-discrimination and state neutrality. Principles like these concern the relationship between religion and politics and give more practical expression to the idea of separation (Audi, 2000: 32–33). Religious freedom, for instance, requires the state not to interfere in religious practices, within certain limits, which is one way to cash out the idea of separation in one particular respect.

Another idea about politics and religion is that the state should engage with, actively make room for, and even support, religion. Religion is extremely important to many people, for whom it has political implications, and politics cannot but affect people's opportunities to live in accordance with their religious beliefs. The state should therefore actively recognise religion. The general idea of recognition is also very broad – for present purposes it signifies a quite open category of possible relations between state and religion involving active engagement by both parties, public expression of acknowledgement or even affirmation, and acts of support, or even delegation of rights and powers (Lægaard, 2012: 199–201). Just as the idea of separation, the idea of recognition is also associated with more concrete political expressions, e.g. state consultation procedures with religious organisations (faith communities etc.), exemptions of religious minorities from rules that interfere with their religious practices, and state subsidies for certain religious activities. In such cases, the state does not keep its distance from religious actors or ignore how policies affect religious groups, as the idea of separation seems to require. Rather, the state actively engages with, and supports, certain groups, not despite the fact that they are religious but because they are religious, and gives special consideration to the category of religion in the assessment of the impact of its policies.

Given these two ideal types of politics–religion relations, the possibility of conflicts between secularism and multiculturalism appears if we consider how multiculturalism has developed recently, especially in Europe. Whereas multiculturalism in countries like Canada originally concerned indigenous groups and national minorities defined on the basis of language, ethnicity and culture (Kymlicka, 1995), multiculturalism in Europe is different. Euro-multiculturalism has, first of all, been concerned with diversity due to immigration. And this diversity has increasingly been understood, not in terms of culture, but in terms of religion – ‘multiculturalism’ in Europe is now primarily a label for discussions about Muslims (Lægaard, 2014a; Levey, 2009; Triandafyllidou et al., 2012). ¹ When we combine this development with the idea of recognition, the possible conflict between secularism and multiculturalism emerges: Multiculturalism seems to require a kind of recognition of religious minorities and practices that might contradict the separation required by secularism (Kymlicka, 2015: 30).

The policy approach to multiculturalism

We should distinguish between multiculturalism as a description of the demographic diversity of societies and as a label for how societies respond politically to this diversity. One common understanding of multiculturalism in the second of these senses is as a set of policies concerning diversity. A version of this understanding can be found in the Multiculturalism Policy Index. ² According to the Multiculturalism Policy Index concerning immigrant groups, multiculturalism is a matter of the extent to which a society has implemented policies like official affirmation of multiculturalism, multicultural school curricula, exemptions from dress codes, dual citizenship, funding of ethnic group organisations and bilingual education or mother-tongue instruction, and affirmative action.

A policy approach to multiculturalism operates by enumerating specific types of policies, i.e. by providing a list. The question, therefore, is what the guiding criterion for counting as a multiculturalism policy is? Definitions of multiculturalism in terms of ‘constitutional, legislative, or parliamentary affirmation of multiculturalism’ or ‘adoption of multiculturalism in school curricula’ evidently presuppose, rather than provide, an understanding of multiculturalism. The other items on the list do provide independently specifiable criteria for multiculturalism. Here the question is what (if anything) they have in common? This shows that a policy approach is inadequate – we need a theoretical understanding of multiculturalism to tell us which policies should be on the list in the first place and why.

The relational understanding of multiculturalism

A standard understanding of multiculturalism is that principles and policies count as multicultural if, and to the extent that, they go beyond standard liberal rights and duties. Multiculturalism has been understood, both by proponents (e.g. Kymlicka, 1995; Taylor, 1994) ³ and opponents (e.g. Barry, 2001), as the view that standard liberal rights and policies are not sufficient to secure justice for minority groups. Multiculturalism claims that justice for minority groups requires more than standard liberal rights, e.g. group-differentiated rights or policies of recognition. A policy can then only be multicultural if, and to the extent that, it is not part of the standard repertoire of liberalism.

This understanding of multiculturalism is relational in the sense that it defines multiculturalism with reference to liberalism. This relational definition explains much of the criticism levelled at multiculturalism from liberal quarters – what is taken to be a point in favour of multiculturalism by its proponents, namely that it accommodates minorities in ways that standard liberalism cannot, is taken by some liberal critics as a conclusive reason against multiculturalism. This conflict between liberalism and multiculturalism only emerges on some understandings of liberalism, however. Since the relational understanding of multiculturalism refers to liberalism, it is important to be clear whether the implied understanding of liberalism is what I, for present purposes, will call restrictive or non-restrictive. ⁴

Restrictive understandings conceive of liberalism not only as a claim about what is required for justice but also about what exhausts justice. Liberalism then says that a given set of principles, rights and policies are necessary and sufficient for justice – but also that any way of ordering society beyond these requirements is unjustified and ruled out by justice. ⁵ Restrictive liberalism might, for instance, take the form of claims that only individuals matter and that rights are only individual rights. Such claims are restrictive in the sense that they rule out assigning intrinsic value to collectives and rule out appeals to genuine group rights.

Given a restrictive understanding of liberalism, the relational definition implies that multiculturalism is necessarily illiberal. The combination of a restrictive understanding of liberalism and a relational understanding of multiculturalism accordingly makes ‘liberal multiculturalism’ a contradiction in terms. Liberal multiculturalists therefore cannot understand liberalism in this way, whereas more radical multiculturalists (e.g. Young, 1990) can hold both a relational definition of multiculturalism and a restrictive understanding of liberalism, because they do not see their multiculturalism as liberal.

But liberalism might also be understood non-restrictively. Then liberalism is only a claim about what is required for justice. This is compatible with holding that liberal requirements (e.g. standard civil, political and social rights), though necessary, may not be sufficient for justice. Under certain circumstances, further measures might be required for justice and other values than those on which liberalism relies might be legitimate policy aims within liberal constraints.

Given a non-restrictive understanding of liberalism, multiculturalism can be an addition to, or an extension of, liberalism and there is no inevitable conflict between liberalism and multiculturalism. ⁶ So the notion of liberal multiculturalism presupposes a non-restrictive understanding of liberalism. ⁷ But the mere conceptual possibility of compatibility is no guarantee that specific multiculturalist principles or policies are compatible with liberalism – it is an open question whether any particular expansion is acceptable from a liberal point of view.

The relational definition of multiculturalism makes everything hinge on the definition of liberalism: We do not know what multiculturalism means if we do not know what liberalism is. Since questions about what liberalism means, permits and requires are hugely controversial, and perhaps even essentially contested issues, the relational definition of multiculturalism is arguably theoretically unhelpful in at least two ways: (a) it presupposes a settled definition of liberalism, which might not be had at all, and (b) it might be trying to define something relatively less unclear (what multiculturalism is) in terms of something relatively more unclear (what liberalism is). These are reasons for looking for another understanding of multiculturalism.

The accommodationist understanding of multiculturalism

A broader understanding of multiculturalism that avoids the problems of the relational definition is that multiculturalism denotes ‘policies designed to provide some level of public recognition, support or accommodation to non-dominant ethnocultural groups’ (Kymlicka, 2007: 16). Multiculturalism is about ‘the legal and political accommodation of ethnic diversity’ justified as ‘a vehicle for replacing older forms of ethnic and racial hierarchy with new relations of democratic citizenship’ (Kymlicka, 2012: 1). Such definitions make accommodation the central point of multiculturalism and are not relational in the way previously discussed.

On an accommodationist definition of multiculturalism, the question becomes what ‘accommodation’ means? Whereas, the relational definition of multiculturalism took accommodation to mean something beyond what liberalism offers, an accommodationist definition need not say this. The following example illustrates the difference.

On a relational definition, liberal rights must be taken as settled, since the meaning of multiculturalism will otherwise be indeterminate. Liberal rights are normally taken to be individual rights that everybody enjoys in the same form and to the same extent. So whatever protection a liberal right affords is extended equally to everybody and is not a matter of multicultural accommodation. The right to freedom of religion, for instance, is an individual right to freedom of conscience, belief and observance, including the freedom to assemble with others for worship and to manifest one's religion in public and private. The right is constrained in the same way for everybody by respect for the rights of others and by sufficiently weighty public aims, e.g. concerns of security or public order, which have the same weight in all cases. Any kind of accommodation not provided by this uniform liberal right to freedom of religion, e.g. forms of communal autonomy regarding family law or exemptions from generally applicable laws, constitutes multicultural recognition because it goes beyond liberalism.

On the accommodationist understanding of multiculturalism on the other hand, liberal rights to freedom of religion are an important form of multicultural accommodation (of religious minorities) rather than a liberal residue that multiculturalism defines itself as having progressed beyond. Contrary to what relational definitions must assume (on pain of indeterminacy), the liberal right to freedom of religion is not settled – it is rather not only (as a matter of fact) an object of recurring political controversy but also (as a matter of principle) an object of continuous legal specification, extension and reinterpretation in light of new empirical developments. Religious freedom might at one point (e.g. in a historically predominantly Protestant country) be taken simply to concern freedom of belief and the right to assemble with others for worship (i.e. ‘going to church’), but increased presence of Muslim immigrants raises new issues about what religious freedom means in new respects, e.g. in relation to obligatory rules of dress in schools, workplaces and public functions, or animal welfare legislation. On accommodationist definitions of multiculturalism, the main question becomes how liberal rights and principles, e.g. freedom of religion, non-discrimination and state neutrality, should be understood in relation to new minorities – to what extent existing rights and principles should accommodate new minority groups.

The understanding of multiculturalism makes a difference. The same policy or rule regarding religious minorities, e.g. revisions of dress codes for public officials to accommodate Muslim headscarves, can often be described in two different ways:

As a matter of special accommodation of religion, e.g. because religion is especially important or valuable, or because religion plays a particularly important role for group identity. The policy or rule is then seen as an exemption or departure from generally applicable rules, which therefore requires a special justification.

These two representations connect to the relational and the accommodationist definition of multiculturalism, respectively. But they also differ importantly, both at the level of framing and at the level of theory, in ways that speak in favour of the second representation and thereby for the accommodationist understanding of multiculturalism.

Representations of rules or policies as special accommodation connect to the relational definition of multiculturalism because accommodation only is ‘special’ in relation to some settled standard. The salient standard is the rights, rules and distributional mechanisms already in place, or which should be in place according to liberal theory. That multicultural accommodations are ‘special’ is then just another way of saying that multiculturalism goes beyond liberalism.

The interpretative representation rather connects to the accommodationist definition of multiculturalism insofar as it emphasises the common ground for how different groups are treated; what according to the first representation is ‘special’ treatment is according to the second the result of an application of the same right or rule under different circumstances. Such treatment is naturally categorised under the heading of multiculturalism insofar as the application of rights and rules happens in an accommodative way, i.e. in a way that is minimally disruptive and as supportive for minority groups as possible.

At the level of framing, the labelling of a policy or claim as ‘special’ exposes it as something standing in need of special justification and the recipient as someone making special claims. ⁸ This will often place claims for accommodation at a rhetorical disadvantage, since minority groups can then be seen as making demands for benefits that others do not get or for being relieved of burdens that others must keep on shouldering. Some proponents of multiculturalism seem aware of this challenge – Tariq Modood, for instance, represents multiculturalism as simply a new form of claims for accommodation similar to those already claimed by, and to a great extent acknowledged and granted to, groups like women, gays and blacks (2007: 69–70). Relative to these claims for accommodation, there is nothing special about claims for accommodation for cultural and religious minorities. But this is often not the most salient standard of comparison. Therefore multicultural claims for accommodation often remain represented as ‘special claims’ and face accompanying opposition.

The interpretative representation permits the framing of multicultural accommodation as ordinary and unexceptional – it suggests that the same rights and rules that everybody else already enjoys and has to respect also apply in the case of new minority groups, but that differences in circumstances might mean that the practical implications are different.

The most important difference, however, is at the theoretical level. Here, the two representations express different ideas about the object and structure of justification:

The special accommodation characterisation suggests that accommodation provides particular groups with some benefit (e.g. a permission and the resulting freedom) denied to others. Such special benefits stand in need of justification and such justification presupposes that there is something special about the group. This is the case both at the level of accommodation of religion in general (the debate about whether ‘religion is special’) and specific minority groups (debates about whether Muslims make special claims).

The interpretative characterisation rather presents justification as application of a general rule to a particular case. The object of justification is a decision regarding how to treat a specific group. In that sense the object of justification is particular. But it is not ‘special’ if the justification for the particular treatment of the group makes reference to a general right or rule. Here is a possible representation of such a process of interpretation and the corresponding structure of justification: ⁹

Basic values: The system of rights and rules is assumed to be based on values of a fundamental and general nature which (although often implicit) explain what is normatively significant, e.g. appeals to ‘human dignity’ and axiological statements about value.

Even though this structure of justification is general and can be projected on to a broad range of normative systems regardless of their specific object and content, it is relevant to multiculturalism: The point of the accommodationist understanding precisely is that multicultural accommodation need not be something special – and that the same justificatory structure therefore applies here as to other normative claims. This is what is denied by the relational understanding and the accompanying representation of multiculturalism as special.

The structure of justification clearly does not say anything about minority groups or accommodation (it is, after all, just a structure and in that sense formal). But ideas central to multiculturalism fit into, and can be represented within, the structure of justification.

Debates about multiculturalism are responses to the presence of diversity and difference. While this is sometimes represented in multicultural theories and rhetoric as something of intrinsic importance, a more plausible way to think of it is as a change in empirical background conditions and in the resulting effects of rules on people: accommodation has become an issue because established ways of ordering the relationship between politics and religion are challenged by the presence of religious minorities due to immigration. Another central multicultural idea is the claim that established rules are not neutral. Such claims usually concern the actual effects of specific rules, e.g. language requirements, clothing regulations, animal welfare requirements or criteria for public subsidies.

My suggestion is that multicultural calls for accommodation can be plausibly represented as claims about the empirical effects of rules under changing circumstances, or as claims that the empirical presuppositions of rules do not hold in new cases. The central justificatory moves in multiculturalism can be represented in the interpretative model sketched above. Even though the structure of justification does not mention accommodation of minority groups, minorities fit nicely into the model if we understand minorities as individuated on the basis of empirical circumstances, e.g. the fact that they are seen as different from prevailing standards and norms in society and have relatively little political power to affect the formulation and implementation of standards in formal rules (Lægaard, 2008: 294–295). Accommodation is then simply an application of a rule to a case in a way that takes account of such changes in empirical circumstances so as to make the implications of the rule less intrusive or more supportive relative to aims and interests that are relevant within the model. ¹⁰

The liberal understanding of secularism

Since the paper concerns the relationship between multiculturalism and secularism, I now turn to the understanding of secularism.

The first understanding of secularism, prominent in the history of liberal political thought and in the self-understanding of many contemporary liberals, is that secularism is the liberal view of what the relationship should be between politics and religion. Robert Audi for instance defines ‘principles of separation’ as necessary components of liberal democracy and as interpretations of the fundamental values underlying liberal democracy (2000: 31–32).

Given this understanding, the question about the relationship between multiculturalism and secularism becomes a more specific version of the question about the relationship between multiculturalism and liberalism discussed above in connection with the relational definition of multiculturalism. So the question about the relation between multiculturalism and secularism can be quickly dealt with if we understand multiculturalism according to the relational definition and secularism as the liberal view of the relationship between politics and religion. In that case, everything turns on whether liberalism is restrictive or non-restrictive in the senses discussed above. Given a restrictive understanding of liberalism, there is a straightforward incompatibility between multiculturalism and secularism. Given a non-restrictive understanding of liberalism (and hence of secularism) the question is open, but the structure of the discussion will be that each proposed multicultural accommodation has to be compatible with liberal values and principles, which constrain permissible forms of accommodation. Secularism then denotes liberal constraints on accommodation of religious groups or practices.

I have already noted some problems with the relational definition of multiculturalism. But the picture of the relationship between multiculturalism and secularism just painted is also problematic for further reasons having to do with the assumed understanding of secularism. First, the understanding of secularism as the liberal view about politics and religion is unhelpful in the same way as the relational understanding of multiculturalism: it does not tell us what secularism is unless we already know what liberalism requires regarding politics and religion, which is something liberals disagree sharply on. Identifying secularism with reference to liberalism offers little explanation, since liberalism is just as contested as secularism. Secondly, the identification of secularism with a liberal position is furthermore unwarranted; non-liberal positions (e.g. republicanism, socialism, some versions of conservatism and specific forms of political theology) also have views about separation of politics and religion, and there is no reason not to think of these as including, or giving rise to, conceptions of secularism (Lægaard, 2013b). For these reasons, we should not define secularism merely as the liberal view about politics and religion.

The policy approach to secularism

Another understanding of secularism rather defines it as a label for specific policies or institutional mechanisms regarding politics and religion, e.g. the separation of church and state. This policy approach leaves it open what the reason for adopting such policies or mechanisms might be. Therefore it can accept the point noted above that we encounter non-liberal secularist claims.

Tariq Modood's discussion of multiculturalism and secularism, which focuses on institutional links between state and organised religions, apparently assumes a policy approach. Modood argues for extending the privileges already granted by the British state to the Anglican Church to organised minority religions. This ‘institutional integration approach’ (2007: 79) can be squared with secularism, according to Modood, if we think of secularism not in terms of ‘absolute’ separation but as ‘moderate secularism’, which is defined as ‘the relative autonomy of politics so that political authority, public reasoning and citizenship does not depend on upon shared religious conviction and motivation’ (2007: 73).

The policy approach understands secularism mainly as a label for policies and institutional mechanisms of separation. Modood's moderate secularism, as an instance of this approach, consists in a gradualist interpretation of the requirements of institutional separation, motivated as a response to the problem that ‘secularism simpliciter appears to be an obstacle to integration and equality’ (2007: 78), i.e. the earlier noted problem that multicultural accommodation of religious minorities seems to clash with requirements of separation of politics and religion. Modood handles this problem by down-scaling requirements of separation until they allow for multicultural accommodations of religious minorities. On this approach, then, the relationship between multiculturalism and secularism is the inverse of the one resulting from the liberal understanding of secularism. Modood adjusts the requirement of separation until it is compatible with the advocated multicultural policies, so the permissible sense of secularism is constrained by the requirements of multicultural equality.

The policy approach to secularism is problematic for a number of reasons: First, secularism as a mere label for specific policies or institutional mechanisms does not provide a criterion for what secularism means in new cases; it consists in a list of examples and does not explain what secularism means for questions concerning politics and religion not presently on the list, e.g. in relation to how the state should deal with new religious minorities. So this understanding is unhelpful in much the same way as the first understanding of secularism, namely that it does not tell us, but rather presupposes that we already know, what secularism requires.

Secondly, the policy approach is a theoretically uninteresting understanding of secularism, because it does not tell us why church and state should be separated. Modood's discussion mainly diagnoses an important problem and offers a descriptive typology of understandings of secularism but does not systematically discuss normative justifications for secularism. Modood's gradualist solution to the conflict between multiculturalism and secularism therefore remains theoretically unsatisfactory since it presupposes that nothing of normative importance is lost by down-grading requirements of separation – this might be true, but we cannot know without knowing why and to what extent politics and religion should be separated in the first place.

For these reasons we should not understand secularism merely as a label for specific policies or institutional mechanisms either. But on the basis of the faults with the liberal understanding and the policy approach, we can now see some desiderata for a helpful and theoretically interesting understanding of secularism: (a) it has to define secularism in general terms and (b) in a way that includes the justificatory structure of conceptions of secularism.

The structural understanding of secularism

The two desiderata suggest that secularism has to be understood as a multi-level view. On the one hand, secularism has as its object the regulation of the relationship between politics and religion; on the other hand, secularism includes a justification for the proposed regulatory policies or measures. Our understanding of secularism should be sensitive both to the normative implications and to the justifications offered for them. But neither uniquely and conclusively define secularism since there can be justifications for secularist policies from a range of different theoretical premises (e.g. liberal or non-liberal) and the policies and measures that can figure as objects of justification are not fixed by a list of existing policies. The understanding should make sense of both liberal and non-liberal justifications for separation as instances of secularism, and it should be open and allow for development of secularist policies in new cases on the basis of the justifications.

Here I will sketch such an understanding of secularism ¹¹ and show its relevance for the discussion of possible conflicts with multiculturalism. My suggestion is that we should think of secularism as a justificatory structure concerned with a specific object of justification, namely the relationship between religion and politics. ¹² To fit with our most general conceptual intuitions and common usage, a view about the relationship between politics and religion should only be counted as a form of secularism if it requires some kind of separation. But what ‘separation’, ‘religion’ and ‘politics’ mean is a fairly open matter for at least two reasons: First, just as Forst (2013) has argued with respect to the concept of toleration, secularism is a ‘normatively dependent concept’; the meaning of secularism reflects normative justifications, so without appeal to such justifications, we do not really know whether, why and in what sense a certain view is a secularist view. Secondly, since there are different normative justifications for claims that politics and religion should be separated, we should think of secularism as a general concept that is then fleshed out with different justifications and resulting meanings of ‘separation’, yielding different conceptions of secularism. My proposal is that the general concept is defined by the general object (the relation between politics and religion), and the general structure of justification. Particular conceptions are defined by inserting particular values and principles in this structure, which generate specific normative implications, e.g. specific policies and institutional mechanisms. Normative discussions of secularism concern these particular conceptions.

This understanding of secularism is analogous to the earlier proposed accommodationist and interpretative understanding of multiculturalism. This is not a coincidence, since the reasons for structural–interpretative understandings are similar in the two cases. The understandings are similar both in focusing on the justificatory structures and by mainly limiting the extension of multiculturalism and secularism, respectively, by a general specification of the object and the kind of normative implications of the justifications: a justificatory structure counts as multicultural if it generates policies of minority accommodation and as secularist if it yields requirements of separation of religion and politics.

Are these understandings too broad and permissive? Should our understandings of multiculturalism and secularism not only be anchored by limits on the normative ‘output’ but also on the ‘input’, i.e. the kinds of values and other premises that can be inserted into the structures? One proposal might be that multiculturalism is always an egalitarian position, which simply insists that equality requires minority accommodation (cf. Lægaard, 2013a). This holds for many prominent theories of multiculturalism (e.g. Kymlicka, Taylor, etc.). But if one can sensibly talk of non-egalitarian forms of multiculturalism, we should not include egalitarianism as a conceptual requirement. Some critics evidently conceive of multiculturalism as non-egalitarian; at least this is a common objection to multiculturalism. It would therefore be a theoretically unfruitful conversation stopper to stipulate that such objections are misconceived since multiculturalism by definition is egalitarian. Similarly, one proposal for specifying the meaning of secularism further on the input end is, as noted, that this is a liberal view. But as already argued, this is an unwarranted and unhelpful claim. These are reasons for not limiting our general understandings of multiculturalism and secularism by requirements on the kinds of values that can be inserted in the structures.

A further similarity between the proposed understandings of multiculturalism and secularism is that empirical assumptions can play an important role in secularism just as in the interpretative understanding of multiculturalism. The relevant empirical considerations will, in the case of secularism, usually be claims about the actual role of religion in society. A prominent recent formulation of secularism, namely Cécile Laborde's ‘justificatory secularism’, illustrates this. Laborde explicates a normative justification for specific requirements of separation in which empirical circumstances relating to religion are crucial. Laborde's justificatory secularism is ‘meta-theoretical’, i.e. not a first-order normative policy prescription but a constraint on state reasons (2013: 167). This is a requirement of separation – it merely operates, not directly at the level of institutions or policies, but at the second-order level of legitimate reasons for public policies or institutional measures.

Laborde's justificatory secularism fits into my structural understanding of secularism, but my understanding is more general than hers. Laborde offers a conception of secularism in one specific respect, not a general definition of secularism. Whereas she points to one particular type of constraint on justifications, the ‘non-imposition norm’, in one specific respect, namely exercise of state power, I suggest that the same general structure can be used to represent views as secularist irrespective of whether they concern state actions and irrespective of whether the central justificatory principle is the non-imposition norm. While the values at the basis of secularism of course have implications for state action, there is no reason to rule out that they might also have implications in other respects and for other actors. And while state action is central to our understanding of secularism, we should be wary of defining secularism exclusively in statist terms, which would replicate the rigid perspective of the policy approach.

Conclusion

The structural–interpretative understandings proposed here have implications for the question regarding the possible conflict between multiculturalism and secularism. Both understandings involve a distinction between a general and mostly formal concept and more specific and normatively substantive conceptions of multiculturalism and secularism. So potential conflicts between multiculturalism and secularism are not general or conceptual but only obtain for specific conceptions and in specific respects. This preliminary result moves the debate forward relative to the general and necessary conflict suggested by the relational and liberal understandings.

The proposed understandings further suggest that, even though multiculturalism and secularism under some circumstances may have the same objects, namely religious groups, practices and claims, it does not follow from this that the normative claims of the two positions necessarily conflict – this depends on how the two understandings are fleshed out in terms of values, interests, aims and derived normative implications. So the proposed understandings move the debate from a focus on abstract positions articulated in metaphorical terms to more specific discussions about which particular normative claims follow from particular justifications.

The focus on structures of justification suggests a programme for how these more specific questions should be addressed and investigated. The first step should be an investigation of the internal coherence of conceptions of multiculturalism and secularism focused on assessing whether (a) underlying values really justify specific claims that (b) might be in conflict with each other. In cases concerning accommodation of religious forms of dress in public institutions, this would for instance require an investigation of whether the values supposed to underlie multiculturalism really justify claims for such accommodation, and whether the values underlying secularism really require institutional religious neutrality in a form incompatible with such accommodation? The interpretative structure provides a framework for this investigation, several features of which might contribute to dissolving some apparent conflicts already at this point, either because the investigation of justifications reveals that claims are not, in fact, in conflict or because the underlying values turn out not to justify the claims in question. Since the application of rules happens partly by taking empirical circumstances into account, and the rules themselves come with empirical preconditions as part of their justification, inspection of the implications of a given right or rule might reveal relevant empirical differences that change the implications of multiculturalism and secularism in ways easing tensions between the two. The weighing of competing considerations and the general assessment of the entire system of rules in light of the basic values is a further point at which conflicts might turn out to be merely apparent. ¹³

The second step should be a normative assessment of conceptions. Given that two conceptions do have incompatible implications, the question is whether they are plausible. This question calls for a standard reflective equilibrium discussion, where the plausibility of a conception is assessed on the basis of how the implications of the underlying principles square with considered judgments about other cases (Rawls, 1999). At this point it might simply turn out that one of the conceptions involved in a conflict is implausible for other reasons.

The final step would be a constructive attempt at reinterpreting conceptions of multiculturalism and secularism which both are internally coherent and independently plausible, but which nevertheless have conflicting implications. Here the interpretative structure directs focus to whether there are alternative formulations of rights and rules that do not have conflicting implications.

All of these questions concern substantive normative argument, which the proposed formal structures do not answer. But the suggested understandings deliver a framework within which to ask the relevant questions and discuss the issues in a constructive way.