I Muslim divorce practice and English law: connections and gaps

It is now widely recognised that many British Muslims use religious legal constructs in ordering their family lives. Often, their doing so is constructed as adherence to Shari'a. Shari'a is the religious law of Islam in general as derived from the Qur'an (the revealed word of God) and the Sunna (the Hadith, or examples and legal verdicts of the Prophet Mohammed), and there is an abundant jurisprudential tradition dedicated to its exposition and interpretation (An-na'im, Reference An-Na'im2008, p. 24). We do not get very far by saying that some Muslims follow Shari'a. First, even the established classical lines of that tradition are inherently plural, albeit boundedly so. There are several schools of jurisprudential thought or fiqh. Each school plays hosts to minority viewpoints and struggles with competing readings of canonical texts (Mir-Hosseini, Reference Mir-Hosseini1997, p. 2; Abu-Odeh, Reference Abu-Odeh, Haddad and Stowasser2004). Second, the doctrine established by the schools of fiqh is subject to ongoing elaboration in the fatwas of contemporary scholars (MacFarlane, Reference Macfarlane2012, pp. 24−29). It has also been incorporated, often in adapted form, into the legal codes of jurisdictions where Muslims have a significant presence (sometimes channelled by colonialism) and has been central to the judgments of legal officials in those states (An-na'im, Reference An-Na'im2008, pp. 17−19). These contemporary interpretations and applications of legal principle may depart substantially from the classical texts. Third, scholars emphasise that Muslim families may draw on a range of sources in seeking solutions to family disputes, in accordance with their own preferences and commitments. These may include the opinions of religious advisors, family members and solicitors, online publications (Ali, Reference Ali2010), and personal or community understandings of English law and of the laws of other jurisdictions to which they maintain connections (Pearl et al., Reference Pearl and Menski1998, pp. 77−80; Shah, Reference Shah2005, p. 38). Rather than write about Shari'a, then, I will write about ‘Muslim divorce practice’. My doing so indicates that I do not make any substantive claims about the proper resolution of family disputes in accordance with the demands of Shari'a. Indeed, my focus is even narrower than the label ‘Muslim divorce practice’ suggests, because my interest is in the interface of that practice and English divorce law; particularly the law of ancillary relief. I will concentrate on arrangements which have actually been adopted by Muslims who have brought their divorces to English courts, and which they, together with their lawyers and the judges who have heard their cases, have described in a language of Islamic law. The argument is concerned with individual acts of legal agency rather than with debates on the internal coherence of anything which we might call by the name Shari'a.

We can find examples of Muslim divorce practice in the recent case-law, including the solemnisation of marriage in England in accordance with widely held Islamic tradition (called nikah),Footnote ² the pronouncement of talaq (a form of unilateral divorce ordinarily reserved to the husband)Footnote ³ on English soil,Footnote ⁴ and British Muslim couples’ agreement of mahr (a gift of property, generally payable in the event of a husband's death or of divorce, which a Muslim husband who wishes to contract a valid marriage must promise to the wife upon marriage).Footnote ⁵ Where one or both partners have ties to another jurisdiction, these practices may have a transnational aspect; for instance, a British citizen might attempt to marry a foreign resident by proxy,Footnote ⁶ or a British Muslim couple might marryFootnote ⁷ or divorceFootnote ⁸ abroad in a jurisdiction which draws on Islamic legal principle for its own family law. One couple's marriage and divorce may cut across multiple bodies of law − religious, domestic civil and foreign − as where a couple marries informally in England, attempts to divorce abroad, and then resorts to an English court to resolve the ensuing financial dispute.Footnote ⁹

Muslim divorce practice can depart from English legal principle. One of the most important differences is in the financial provision which a woman can expect to be made for her by her husband upon divorce. As Macfarlane notes, it is in the nature of informal legal pluralism that contrasts in principle between religious and civil law will not always play out in practice (Macfarlane, Reference Macfarlane2012, pp. 191−92). Thus, what I offer here is a very basic summary of the possible problem. Mahr is the core of the financial provision made for a woman upon divorce – it is effectively the price of divorce (Yassari, Reference Yassari, Mehdi and Nielsen2011, p. 200). Usually, the bulk of mahr is not paid upon marriage. The deferred mahr is a debt which the husband owes to his wife. If the mahr agreed upon marriage is relatively small, if the wife has agreed to waive or reduce the debt during marriage or upon divorce (al-Hibri, Reference Al-Hibri2000, p. 49), or if the husband will not pay, a wife who has not amassed assets or income of her own may be unable to support herself after divorce. As a matter of Islamic legal tradition, her husband's obligation to pay her maintenance after the marriage breaks down is usually minimal, and she has no claim on his other property (Mir-Hosseini, Reference Mir-Hosseini and Anwar2009, p. 31). Does English law permit a husband to leave his wife in this position?

If a Muslim couple have been civilly married in England or if their foreign marriage is recognised at English law, they will need a civil divorce to terminate that marriage: neither talaq Footnote ¹⁰ nor any ruling of a religious body can have the required effect.Footnote ¹¹ Suppose we imagine first that the couple came to a mahr agreement at the point of marriage. They are now divorcing, but do not now agree that the mahr agreement should determine their mutual financial obligations. Because the amount of mahr is agreed at the point of marriage, and is stipulated in the parties’ marriage contract, we might treat it as a nuptial agreement. No English court is required to give determinative effect to any nuptial agreement because parties cannot oust the jurisdiction of the court by agreement.Footnote ¹² The court will consider the agreement along with the universally applicable ‘requirements of fairness’ identified in Miller v. Miller and McFarlane v. McFarlane.Footnote ¹³ Assets will be divided between the parties according to their housing and financial needs, their need for compensation for any economic disparities wrought by the way in which the couple have conducted their marriage, and the principle of sharing, which presumes that each spouse is entitled to an equal share of the assets unless there is a good reason to the contrary.Footnote ¹⁴Radmacher v. Granatino makes clearFootnote ¹⁵ that courts can give due weight to agreements concluded before a marriage had actually broken down. The court will, of course, have regard to the parties’ intention at the time of the agreement that it should govern the financial consequences of marriage breakdown; to the couple's age, emotional and financial circumstances and experience; and to any material lack of disclosure, information or advice or to any other exploitative behaviour affecting the parties’ consent to the agreement.Footnote ¹⁶ It should not be assumed that a nuptial agreement will be determinative in all questions of ancillary relief. The court will also consider the Miller substantive fairness of the proposed distribution of property between the spouses, as it would do in any ancillary relief case. Although the fact of the agreement itself may alter what is fair, it seems unlikely that the courts will enforce a prenuptial agreement which makes very little provision for a wife's financial needs, particularly where she has made a significant contribution to her husband's success by her work within the home over the course of a long marriage.Footnote ¹⁷ If a court decides, in the exercise of its discretion, that an earlier agreement should determine a couple's financial obligations to one another upon divorce, then the agreement takes its force from the court's approval.

The position is slightly different where, during marriage breakdown, a couple come to a post-nuptial agreement or separation agreement based on their original mahr agreement, perhaps with the assistance of a Shari'a council. If the couple enter into a separation agreement in advance of divorce, it will carry considerable weight in any later proceedings for ancillary relief. The courts are reluctant to allow couples to depart from such agreements without compelling reason. Such reason might include a defect in the agreement's formation (such as the husband's exploitation of superior bargaining power such that his wife's will is overborne), or a change in circumstances which would justify variation of the agreement.Footnote ¹⁸ Alternatively, the couple may agree, under Section 33A of the Matrimonial Causes Act 1973, to compromise their mutual financial and property claims at the point of divorce. That agreement will be embodied in a draft consent order, and the court may agree to make an order on those terms. The court has an independent obligation to check the agreed arrangements for fairness in line with the usual principles.Footnote ¹⁹ Although it is stressed that the court is not to simply rubber-stamp the draft order, the information available to it where the parties arrive with an agreement already negotiated is less than when the case is contested: it only receives a two-page summary of the parties’ financial circumstances and will not investigate further unless something in that summary puts the court on inquiry. The court need not be informed that the order is the outcome of mediation at a Shari'a council. The point of the consent order procedure is to allow for negotiation without undue interference from the courts, and so the courts’ function in scrutinising such orders is limited.

As a matter of black letter principle, it seems clear that the door is closed to radical ‘contracting out’ of the core principles of the English law of ancillary relief. However, it is important to appreciate the extent to which Muslim women are effectively excluded from that law. First, Muslim couples, like other couples, may choose not to divorce in the civil courts at all, instead simply separating and coming to a financial settlement – on their own or with a Shari'a council − which never comes before the courts. Indeed, it seems clear that government policy is to positively encourage non-judicial settlement of divorce disputes (Eekelaar, Reference Eekelaar2011). Second, many Muslim couples are not married for the purposes of English law. This most commonly occurs where the marriage is not registeredFootnote ²⁰ as required by the Marriage Act 1949 (Beresford, Reference Beresford2011; Kennett, Reference Kennett2012). Many mosques are not registered for marriages or do not provide an official who is legally recognised as a solemniser of marriages.Footnote ²¹ Furthermore, it is not uncommon for Muslim couples to marry in private homes (Bano, Reference Bano and Mehdi2011, p. 200).Footnote ²² A couple can take steps to give formal notice of marriage to the Superintendent Registrar of the District where they are resident, and arrange to have a registrar present at the religious ceremony. Alternatively, they can have a civil marriage ceremony before or after the nikah, and some mosques encourage this practice. But many do not do so (Yassari, Reference Yassari, Mehdi and Nielsen2011, pp. 268 − 70). Following Gandhi v. Patel & Ors,Footnote ²³ a religious marriage which is unregistered in this way may be treated as a ‘non-marriage’.Footnote ²⁴ The court will consider whether the couple ought reasonably to have been aware of the requirements for contracting a legal marriage in England and, taking that into account, will ask whether they made efforts to contract a lawful marriage in accordance with the Marriage Acts.Footnote ²⁵ It is possible that a court will conclude that the couple might reasonably have been unaware of the marriage requirements, as where they were very young recent immigrants who had trusted in their elders to make the marriage arrangements for them.Footnote ²⁶ But an unreasonable mistaken belief in the validity of the marriage will not establish a marriage at English law. Footnote ²⁷ Neither will a firm religious conviction that the nikah is a ‘true’ marriage substitute for compliance with the formalities. Footnote ²⁸ Ordinarily, a Muslim couple who have been resident in the UK for some time and who undergo a nikah ceremony without any further effort to register it will find themselves in a ‘limping’ marriage, one which is recognised under Shari'a but not under civil law.Footnote ²⁹ The key problem then is that a court has no ancillary relief jurisdiction when such marriage breaks down.Footnote ³⁰ The couple are in the position of ordinary unmarried cohabitees, and in the absence of legislative provision since the Cohabitation Bill 2008 stalled, that is a relatively poor position for wives to occupy (Probert, Reference Probert2009). At best, a wife who does not want to fall back on the limited provision to which she is entitled at Islamic law may be able to establish some right in the family home, perhaps via the law of trusts.Footnote ³¹ However, a wife who has never had substantial savings or earnings is unlikely to be able to avail of this possibility, assuming her husband owned a family home at all (Diduck and Kaganas, Reference Diduck and Kaganas2006, pp. 266−67). Alternatively, she may come to court to enforce some private agreement which she has made with her husband. The mahr agreement may fulfil this role, acting as a kind of cohabitation agreement.

The permissibility of polygamous marriages in Muslim communities also generates problems. Where one party to the marriage is already legally married to another at the point of entering into a new legal marriage, the new marriage is void under Section 1(d) of the Matrimonial Causes Act 1973. This is true whether the first or second marriage was obtained at home or abroad, provided that either party to the second marriage was domiciled in Britain at the time of solemnisation. Polygamous marriages can be produced deliberately with the knowledge of both parties, or where an attempt to dissolve an earlier marriage fails, tainting subsequent marriages (Charlsley and Liversage, Reference Charlsley and Liversage2012, p. 69). Talaq divorces obtained abroad are especially vulnerable in the latter regard. In that event, the second wife can bring a nullity petition, and the court has the same powers to grant ancillary relief as if she were divorcing.Footnote ³² That said, since deliberately entering into a polygamous marriage amounts to the crime of bigamy, in making its award the court may exercise its discretion to consider the conduct of the parties under Section 25 of the Matrimonial Causes Act 1973.Footnote ³³ The upshot, perhaps unexpectedly, is that a party to a void polygamous marriage is in a better position in terms of ancillary relief than one in a monogamous non-recognised marriage. It remains possible for a man to obtain a polygamous marriage in England by registering one marriage but not the other; by entering into multiple religious marriages and not registering any; by divorcing for convenience and remarrying while maintaining his relationship with two women; or by maintaining another unrecognised marriage or marriages abroad. When a de facto polygamous marriage is procured by refusal to register a marriage, some women are absolutely excluded from the law of ancillary relief and only connect with common law via the law of trusts and contract.

A final set of issues arises from reliance on Shari'a councils. A Shari'a council is a tribunal of Muslim religious scholars, who provide advice and guidance on matters of Islamic law and morality, as well as mediating and informally resolving family and civil disputes (Douglas, Doe, Gillat-Ray, Sandberg and Khan, Reference Douglas, Doe, Gillat-Ray, Sandberg and Khan2012). Their assistance is generally sought by divorcing women since, by default, only men are empowered to pronounce talaq. Councils may be involved in brokering a divorce by mutual agreement. This is called mubaarat, where divorce is initiated by the husband and obtained with the wife's consent (Nasir, Reference Nasir2009, p. 130). Khula is initiated by the wife, and takes effect with the husband's consent (p. 134). Khula may entail the wife's waiver of the outstanding mahr obligation in exchange for the divorce (Bowen, Reference Bowen2011, p. 421). Some councils may also dissolve a marriage for appropriate reason. Faskh is the annulment of a marriage for some defect in the marriage contract (Esposito and DeLong-Bas, Reference Esposito and Delong-Bas2001, p. 33). Tafriq is a judicial dissolution of a marriage where the wife proves her husband's fault (Nasir, Reference Nasir2009, p. 134). The methodology adopted, and the types of divorce and financial remedy available, will vary from council to council, according to the scholars’ background and the schools of fiqh to which they subscribe. Now, the divorce itself has no effect at English law but, as part of the divorce process, the council may rule on their obligations to each other under Shari'a, and that ruling will likely focus on the agreed mahr, or on its potential waiver.

If the couple come to an agreement on that basis, it may take the form of a separation agreement, or be submitted for ratification as a consent order, as discussed above. It may be that, in some circumstances, the financial rulings granted by Shari'a councils in such cases could be binding as a matter of English law. That possibility is lent clarity by the launch, in February 2012, of a new family law arbitration scheme under the auspices of the Institute of Family Law Arbitrators (IFLA). The scheme allows for arbitration of family disputes in accordance with English law, or foreign law as appropriate. In theory, a religious scheme could follow its lead. Awards made under the scheme are binding on the parties under Sectio 58 of the Arbitration Act 1996, in virtue of the arbitration agreement. A court may summarily enforce an arbitral award in line with Section 66 of the Act. The rules of the IFLA scheme provide that awards may be appealed and challenged in accordance with the provisions of the Arbitration Act 1996, and that, where necessary, the parties must apply for a court order in the same terms as the award (IFLA, 2012). The precise status of an arbitral award under the scheme is unclear. However, the courts retain the power to set unfair awards aside. Certainly, the courts are deferential to contracts which contain religious content, though this is hardly unlimited.Footnote ³⁴ Moreover, while it may be that Radmacher v. Granatino signals a new judicial deference to nuptial agreements of all kinds, including agreements to arbitrate, it seems clear from that judgment that a court will not defer to an arbitration agreement procured by duress or which has brought about significant financial unfairness between the spouses.Footnote ³⁵ The balance to be struck is indicated in the recent case of AI v. MT,Footnote ³⁶ Baker J. gave further guidance on the approach to be taken to religious arbitration of marital disputes. He heard the case of a divorcing Orthodox Jewish couple who had agreed to refer all disputes between them to binding arbitration by a senior rabbi at the New York Beth Din. This was not a dispute about financial remedies as such, but the decision is nevertheless instructive. Baker J. confirmed that the Hyman principleFootnote ³⁷ against agreements attempting to oust the family law jurisdiction of the courts remains good law. Thus, in this case, the court was not willing to treat the arbitration agreement as binding. It retained the right to evaluate the outcome of the arbitration, and ultimately endorsed the decision of the Beth Din. In a judgment which refers to the IFLA scheme, Baker J. confirmed that an arbitration agreement of this kind should carry great weight with the court in those later proceedings, to be treated more as a lodestone than as a yardstick.Footnote ³⁸ The court also stressed its willingness to support the couple, both in choosing an alternative dispute resolution process, and in selecting a forum which could decide their dispute in accordance with the religious obligations.Footnote ³⁹ However, it confirmed that any arbitration of a civil dispute conducted in accordance with religious law is only enforceable by a court insofar as it complies with English law and public policy. Having scrutinised the principles applied by the Beth Din, Baker J. did not find any conflict between them and English law. This judgment indicates an intention to maintain a judicial role in the oversight of religious dispute resolution, albeit that role is supportive of religious practice and reasonably deferential to religious authority. Of course, as discussed below, this provision for judicial oversight only has bite insofar as dissatisfied parties can access the courts to challenge arbitral awards (Malik, Reference Malik2012, p. 44).

II Private ordering and feminist critique

Since the passage of the Divorce (Religious Marriages Act) 2002, successive governments have taken a standard position on the role of religious divorce at English law; that although the government may be aware of circumstances in which religious divorce law may oppress women, it:

The 2002 Act allows a civil court to compel a couple to obtain a religious divorce as a condition of granting a decree absolute under Section 10 A of the Matrimonial Causes Act. It is designed to assist women who would otherwise be refused a religious divorce or who could only obtain it on payment of a financial penalty. For our purposes, the legislation is not important in its own right, but the discourse around its passage and development helps sketch the boundaries of state engagement with religious divorce. First, the state implicitly delegates adjudicative space to religious leaders. The Section 10A order can only be granted in respect of religious groups the leaders of which have asked the Lord Chancellor to prescribe their group for that purpose. Such leaders will have determined that their own internal mechanisms for resolving disputes where a recalcitrant spouse refuses a divorce are inadequate. They cannot be compelled to bring themselves within the scheme. So far, only Jewish marriages are covered by the legislation, because the original bill was supported by all of the major representative bodies in Anglo-Jewry. Thus, the Act cannot assist a woman who might be refused khula unless she waives her mahr, because the legislation does not cover any Muslim organisations. Second, we may recognise that religious law conflicts with civil law, but only those injuries which are brought to the attention of the civil courts can attract a remedy: the Section 10A order can only be granted where the couple are able to seek a civil divorce in the first place, and the state can only inform them and encourage them to do so. Thus, where there are gaps between civil and religious law, they are to be closed by individual exercises of responsibility and religious leadership.

That position was famously restated in respect of Muslim divorce practice by Jack Straw as Lord Chancellor in a speech at the Islamic Finance and Trade Conference in October 2008 (Straw, Reference Straw2008). It runs as follows:

1. 1. The government acknowledges that Muslims living in England may order their family lives according to religious principle, but that religious principle has no legal force.

2. 2. Recourse to religious divorce practice occurs privately, on a basis of individual consent. It is no more troublesome than any other agreement or mediation. Individuals are permitted to regulate their lives in accordance with religious principle as an expression of religious and personal autonomy, and the government will not intervene to prevent them from doing so.

3. 3. If an individual's consent to religious divorce practice is in question, s/he can have recourse to the English courts, which will exercise a protective jurisdiction.

4. 4. Adherence to Muslim divorce practice does not raise the possibility of illegality even where Muslim tradition and English law conflict. This is because no English court will enforce a practice which runs contrary to English law or public policy.

5. 5. The government has no plans to change this legal position.

In its many iterations,Footnote ⁴¹ the policy ignores the fact that, on its terms, vindication of women's rights is contingent on their access to the courts. It assumes that failure to come to court is a matter of individual choice. It is not that law-makers are unaware of the problems posed by polygamy, unregistered marriages and reliance on non-state dispute resolution. Rather, they are relatively ‘hands-off’ in their approach: the responsibility of government, it seems, is only to inform women (in general), mosques and community groups of the requirements for legal marriage and of their entitlements under English law.Footnote ⁴² Even in October 2012, Lord Gardiner, responding to the second reading of Baroness Cox's Arbitration and Mediation Services (Equality) Bill on behalf of the coalition government, effectively restated the policy established under Labour, relying on the tropes of choice, voluntariness and notional judicial oversight to cure all ills.Footnote ⁴³ What is interesting is that the proposals in that Bill build upon rather than challenge the dominance of the ‘hands-off’ framing in British legislative discourse. The Bill, for instance, does not propose to do anything about marriage recognition. It will not ‘interfere in the internal theological affairs of religious groups’ and does not seek any radical limitation of the use of religious law in family proceedings.Footnote ⁴⁴ It would not extend the scope of Section 10A of the Matrimonial Causes Act. Neither does it challenge the government's support for ongoing privatisation of dispute resolution in family matters, or upset the IFLA arbitration scheme.Footnote ⁴⁵ The remedies offered to women by Parts 2 and 3 of the Bill follow the grain of the established civil law. They essentially add, perhaps unnecessary detail to judges’ existing statutory powers to set aside arbitral awards for bias (including evidence of gender discrimination), to refuse to enforce nuptial agreements for duress, or to refuse to enforce contractual terms which are discriminatory within the meaning of the Equality Act 2010. The Bill would still rely on women to challenge unfair settlements and awards; though it is noteworthy that Clause 5 would allow an appropriate third party, such as a local authority, to challenge a consent order on an affected party's behalf, following the model of Section 63C of the Forced Marriage (Civil Protection) Act 2007. The remainder of the Bill is designed to ensure that women in unrecognised marriages are better informed of the differences in their entitlements at civil and religious law.Footnote ⁴⁶ As things stand, then, the British legislative debate on regulation of Islamic divorce is turning in a small space between barely contrasting visions of ‘choice’.

One possible feminist critique of this position begins by explaining Muslim divorce practice as a ‘private ordering’ problem (McClain, Reference Mcclain2010; Bano, Reference Bano, Thiara and Gill2010). ‘Private ordering’ denotes arrangements by which government explicitly or tacitly (even if through neglect or refusal to govern) shares regulatory authority with private actors. Government policy on Muslim divorce practice reflects a familiar, if unsophisticated, understanding of the potential of private ordering. Tolerance of Muslim divorce practice is represented as a necessary corollary of religious autonomy of groups and of individuals: Muslim couples should be permitted to avoid some inappropriate or dysfunctional public norms (McMillan and Woodruff, Reference Mcmillan and Woodruff1999) and to generate private regulatory frameworks for facets of religious life which state law is ill-equipped to address (Ertman, Reference Ertman2007, p. 700; Davis, Reference Davis1999). Government policy demonstrates some appreciation of the limits of that position; it does refer to the possibility that some individuals might be bound to arrangements which they have not chosen. But its solution to that problem is a variant of the ‘right of exit’ argument (Mookherjee, Reference Mookherjee2009, pp. 85−87) – the argument that women's rights are adequately protected where they have the formal option to reject religious dispute-resolution mechanisms and choose civil law instead (Shachar, Reference Shachar2001, p. 41). A feminist critique of government policy would emphasise, as I have done, the high distributive stakes of private ordering of marriage in cases where reliance on Muslim divorce practice would leave a woman with meagre financial provision or entirely destitute. Given those stakes, it would interrogate the assumptions around the possibility of exit – of easy access to civil law – on which the legitimacy of government policy, by its own design, must stand or fall. It would point to the artificial subject which government policy allows to authorise significant distributive inequality between Muslim men and women, and between those Muslim women who are excluded from the law of ancillary relief and their included counterparts. This is a thinly drawn, autonomous, rational, responsibly bargaining subject, who finds herself in a given economic position on marriage breakdown because she has consciously chosen to occupy it. Her liberty consists in a sort of sovereign consumerism (O'Malley, Reference O'malley2000). If her access to justice is restricted it is because of other dimly understood but irresponsible choices; her own or those of other dangerous cultural agents. In order to posit this subject, government must ignore the relational nature of marital bargaining, whereby couples are hemmed in by the family's particular hierarchies, expectations, duties and orders of authority (Halley and Rittich, Reference Halley and Rittich2010, p. 758). It must ignore the experience of women who have been deliberately coerced or misled into unwanted ‘choices’ of law or forum (Bano, Reference Bano2008). It must ignore the particular patriarchal relations of domination which often characterise religious contexts, and which, even in the absence of actual coercion, may limit women's bargaining power upon divorce (Siddiqui and Patel, Reference Siddiqui and Patel2010, pp. 15−16; Suleiman, Reference Suleiman2012, pp. 54−56; Pettit, Reference Pettit1997, pp. 164−65; MacFarlane, Reference Macfarlane2012, pp. 180−86; Menski, Reference Menski2013, pp. 51 − 53). It must ignore the impact of women's poverty on their access to justice, and the related impact of the government's own recent cuts to legal aid (Genn, Reference Genn2012; Hunter, Reference Hunter2011). And, as Macklin argues in this isue, it must ignore the array of state-created ‘background rules’, derived from family law, other areas of law such as immigration, welfare and local government (Patel and Siddiqui, Reference Patel, Siddiqui, Thiara and Gill2010), and property law, which work to restrict individuals’ bargaining options: the law on recognition of marriages and on the property rights of cohabitees are cases in point (Halley and Rittich, Reference Halley and Rittich2010, p. 762; Fournier, Reference Fournier2012, pp. 61 − 62). A feminist critique of the government position would not ignore any of these influences on women's choices in divorce. It would draw out each of these interlocking aspects of what Ayelet Shachar has termed women's ‘multicultural vulnerability’ (Shachar, Reference Shachar2001), and would substitute for the responsible bargaining subject of the government imaginary, an encumbered subject who must negotiate that vulnerability.

The critique of private ordering has a rich heritage in both feminist work on family law and in feminist multiculturalism, and I do it something of an injustice to summarise it so briskly. I do so because, although I accept the crucial importance of this critique, it seems to me that something is happening in the politics of Muslim divorce practice which arguably strips the private ordering critique of any significant discursive purchase. Something is ‘in its way’. Specifically, I want to suggest that the figure of the responsible bargaining subject does some important work for the state, which attention to the purely distributive dimensions of private ordering cannot capture. In order to give a useful account of the function which the responsible bargaining subject performs, it is necessary to put government policy on the law of Muslim divorce practice in its political context.

III Law and loyalty: Britain's Shari'a debate, 2008−2012

Although British engagement with Shari'a has much older colonial and post-colonial dimensions, it has become common to date the beginnings of the domestic ‘Shari'a debate’ to February 2008. In that month, in a lecture delivered at the Royal Courts of Justice and chaired by the Lord Chief Justice, the Archbishop of Canterbury called for a re-evaluation of the interaction of state law and religious law in the United Kingdom, and explored the possibility of the state delegating limited jurisdiction over specific matters to Shari'a councils (Williams, Reference Williams2008). The Archbishop had argued that an ‘unqualified secular legal monopoly’ was no longer appropriate and that the English legal system should adapt, to reflect what he understood as the reality that people have ‘multiple affiliations’ beyond their allegiance to the state. In an interview with the BBC, he put this claim with rather less nuance, when he suggested that state recognition of Shari'a was effectively ‘unavoidable’; a word which, one journalist wrote, ‘triggers that little England panic that we are powerless as our world disappears in front of our eyes’ (Bunting, Reference Bunting2008). The Archbishop's speech inaugurated an anxious new British politics of Muslim divorce practice, at the heart of which is a subject quite different from the more or less successful bargaining subject already discussed in this article. This is the Muslim subject whose adherence to Shari'a indicates a compromised loyalty which places ‘our world’ at risk.

The discursive co-mingling of law and ‘our’ identity is an old one. As Hans Lindahl explains, in liberal political theory the legal order is grounded in collective acts of self-legislation. There can be no law without a group which constitutes and performs itself as a unity by reference to that law (Lindahl, Reference Lindahl2008). Legal orders begin with acts in which members of a polity refer to themselves as a unified ‘we’, which holds interests in common, and which enacts legal norms for the sake of those interests. These founding acts draw boundaries between the law's ‘we’, who share in common norms, and the ‘they’, who are outside the law; the boundary between selfhood and alterity, belonging and unbelonging (Yuval-Davis, Reference Yuval-Davis2011, p. 5). The community thus walls itself in by law (Lindahl, Reference Lindahl2006, p. 885). After that founding act, the community is known by its obedience to a shared body of law, and we pass new laws ‘as an expression of who we are as a people within our borders’. (Dudziak and Volpp, Reference Dudziak and Volpp2006, p. 595). But after this initial act in which the boundaries are posited, they must be policed and guarded against challenge. In a plural polity the ‘self’, or ‘we’ which has enacted those legal norms is always subject to challenge and resistance: there is an undecidability to the ‘we’ (Honig, Reference Honig2008). Its boundaries are perpetually drawn and redrawn in response to challenge (Lindahl, Reference Lindahl and Schaap2009). The state gives meaning to these boundaries, and to membership on the inside, by ‘rituals of entry, access, belonging and privilege’ (Benhabib, Reference Benhabib2004, p. 1). In this way, the legal community is the effect of legal practice.

It is telling that the then prime minister, the leader of the opposition, and other senior members of parliament, in rejecting the Archbishop's proposals, immediately cast his intervention as posing questions about national identity and emphasised the civil law's essential ‘Britishness’ (Butt, Reference Butt2008). This reaching for ‘Britishness’ is of no mere superficial importance. In contemporary Britain, discussion of membership in the national community is almost inevitably overlaid with questions of ‘cohesion’. The cohesion agenda is a strand of a New Labour approach to community relations – still current under the coalition government – which depoliticises social conflict by framing it in terms of ‘active citizenship’ and ‘community self-management’ (Lowndes and Thorp, Reference Lowndes and Thorp2011; Clarke, Reference Clarke2005). It is typically associated with a ‘retreat from multiculturalism’, born of the assumption that multicultural policies have produced divisions between minority citizens and the national group. Those divisions are, so the cohesion argument runs, exemplified by the British Muslim participants in the ‘race riots’ which occurred in Oldham, Burnley and Bradford in the summer of 2001. The politics of cohesion was, from its inception, a politics of risk (Jabri, Reference Jabri, Stephens and Vaughan-Williams2008). Initially, the risk posited was that of social division and violence, but post-2005, as cohesion became firmly enmeshed in the domestic counter-terrorism agenda following the 7/7 bombings, the threat was that of home-grown terror. Risk was said to flow from difference; from the actions of a ‘they’ who threaten the ‘we’: the problem was that migrant Muslims were culturally insular and had self-segregated from their majority neighbours (Cantle, Reference Cantle2001; Ouseley, Reference Ouseley2002). This explanation was swiftly uncoupled from acknowledgement of the effects of wider structures of poverty, racism and political exclusion (McGhee, Reference Mcghee2005, p. 54). In a series of reports devised as responses to the 2001 disturbances, it was determined that the solution to the problem lay in a programme of civic renewal, which would cultivate defective citizens’ primary loyalty to shared ‘British values’ (Denham, Reference Denham2001; Blunkett, Reference Blunkett2001; Cantle, Reference Cantle2004; Goldsmith, Reference Goldsmith2008). As prime minister, Gordon Brown expressed faith that a government-led search for common values would help forge common purpose and guard the nation from extremism (Great Britain Ministry of Justice, 2007, p. 11). The cohesion agenda notionally enlists Muslim citizens as ‘partners’ with government in securing the nation-state order and their own position within it (K. E. Brown, Reference Brown2010). On those terms, Muslims’ loyalty would be evident in their capacity to articulate and express their difference within the limits of shared civic culture. The cohesion agenda thus licensed a raft of targeted government projects which produced model Muslim subjects, identifiable by conspicuous lived loyalty to British values. To this end, government has aimed to direct Muslim community life. For instance, as part of its ‘hearts and minds’ counter-terrorism agenda, the government aimed to selectively identify and fund Muslim faith groups which would model loyal religious citizenship and counter ‘radicalising’ influences in their communities (Gutkowski, Reference Gutkowski2011, p. 352; Brown, Reference Brown2008). It has also sought to manage Muslim family life. For example, the government began a programme of legislative intervention around young adult marriage and immigration into which was embedded an opposition between civilised Muslim families who arrange marriages with their children's consent, and dangerous Muslim families who coerce them into undesirable marriages with foreigners (Enright, Reference Enright2009; Cameron, Reference Cameron2011).

I argue that, in seeking to govern Muslim divorce practice, as in these other policy contexts, government distinguished between a responsible, decent Muslim subject, committed to British values, and a dangerous Muslim other, affiliated to another values system, threatening by his disloyalty. The Archbishop's intervention met, by and large, with hostility from the establishment. Many responses from politicians and media commentators constructed adherence to Shari'a as foreign, barbaric and backward. Frequent reference was made to brutal punishments for criminal offences; a connection which was reinscribed even by those who were careful to note that the Archbishop had confined his remarks to family law (Moore, Mason and Lewis, Reference Moore, Mason and Lewis2008, pp. 32−33). Shari'a was contrasted with English law, which was presented as a civilised channel for steady upward progress in the vindication of human rights, and women's rights in particular.Footnote ⁴⁷ Others went further, linking the demand for state recognition of Shari'a institutions to support for terrorism; a link which government itself had already made on previous occasions. One newspaper, indeed, called the Archbishop's speech a ‘victory for terrorists’ (Moore et al., Reference Moore, Mason and Lewis2008, p. 33).

Public discourse had, thus, already identified a ‘they’ who desired state recognition for Shari'a councils. It was now open to government to reconstruct a ‘we’ in opposition. As we might expect, government fell back on the social cohesion agenda, using it to construct the possibility of an alternative, responsible mode of engagement between Shari'a and civil law which was conditioned by the overarching aim of safeguarding social cohesion. From the early days of the cohesion agenda, adherence to a shared law has been asserted as one of the core British values to which citizens must commit. For example, the ‘Life in the United Kingdom’ Advisory Group report (Crick, Brighouse, Brown, Coussey and Hira, Reference Crick, Brighouse, Brown, Coussey and Hira2003) suggested the following as part of its broader definition of Britishness:

Later, Lord Goldsmith's citizenship report repeatedly emphasised the more prosaic duty of citizens to obey the law of the land (Goldsmith, Reference Goldsmith2008). In the aftermath of the Archbishop's speech, the demands of that duty were made clear. Predictably, several of the Archbishop's opponents argued their case in the terms of the social cohesion agenda. His proposal, by undermining the connection between citizen and law, threatened social unrest, segregation and chaos. The emphasis was on the possibility of separate ‘parallel’ legal systems on one territory. The chair of the Equality and Human Rights Commission criticised the Archbishop's proposals as ‘old style divisive multiculturalism’ which would put ‘our common values’ at risk (Woodward and Butt, Reference Woodward and Butt2008). In a key speech, the then leader of the opposition, David Cameron, took his vision of the ‘legal apartheid’ which would result from the public recognition of Shari'a as ‘actually the logical endpoint of the now discredited doctrine of state multiculturalism’ (BBC, 2008). This was the context in which Gordon Brown's New Labour government came to formulate its policy on Muslim divorce practice; a policy which was mechanically repeated on the occasions when it was asked to explain or justify the relationship between English law and the operation of Shari'a councils, and which has been taken up by the current government.

The influence of that context was obvious in the 2008 speech in which Jack Straw explained that policy to the Islamic Finance and Trade Conference (Straw, Reference Straw2008). In it, he solidified the Muslim claim to resolve marital disputes in accordance with Islamic principle privately, at their own risk, and accepting responsibility for unsettling any resulting inequalities. Straw made the speech to a group of entrepreneurs, who he praised as model citizens who had made important contributions to life in Britain. The Conference was founded by the Muslim Council of Britain, which, at that time, was a relatively influential lobby group which supported the core elements of the government's cohesion agenda (Adamson, Reference Adamson2011, p. 909). The discussion of Shari'a is, predictably, couched in the discourse of social cohesion. In the section preceding the discussion of Shari'a family principles, Straw discusses the ‘positive influence of Muslims on this country’ and the desirability of ‘breaking down barriers between communities’. He outlined the government's broader project of engagement social cohesion agenda; its

In the section following his discussion of Muslim divorce practice, he moved to the theme of cohesion in the context of the ‘war on terror’. He dismissed ‘those who seek artificially to polarise the debate about violent extremism as being one between Muslims and non-Muslims’, and asserted that ‘the divide is between a tiny minority of criminals full of hatred, and a vast majority who must unite against them’. Later, he described this ‘vast majority’ as ‘us’: those who ‘must ensure our voices are heard clearly and … join together in finding a way to defeat violent extremists and the terrorism they espouse’. Straw acknowledged the association of Shari'a with ‘barbaric practices such as stoning’, but dismissed them as unacceptable exceptions which have parallels in other religions. He made no reference to contemporary concerns about women's rights.

Thus, crucially, Straw's speech does not deny the existence of the ‘dangerously disloyal other’. Indeed, it constructs an alternative Muslim subject by contrast. Government policy contemplates a responsible cohesion-minded subject who accepts the supremacy of English law. That subject's engagement with Shari'a is opposed to representations of Shari'a as irredeemably barbaric, archaic and foreign, and is congruent with preserving the stability and sovereignty of the ‘we’. What I want to suggest is that the responsible divorcing Muslim subject cannot be understood apart from the government's address to her disloyal other. Indeed, the first is constructed and engaged precisely in order to govern the threat posed by the second.

IV The contractual and the colonial

Gordon Brown's government did not respond positively to rough calls for a ‘ban’ on Shari'a councils. And yet, as feminists acknowledge, that government's policy is nevertheless exclusionary. In the second section of this article I listed a number of gendered exclusions at its heart. Neglect of these exclusions is the price to be paid for the coherence of the responsible bargaining subject in whom government is so deeply invested. Recent work on the political theory of boundary-drawing suggests how some Muslim women find themselves cut adrift from law as part of the process of separating law's ‘we’ from its dangerously disloyal other so that, although they remain subject to law, they are unable to make effective claims on it (Fraser, Reference Fraser2008, p. 21). Boaventura de Sousa Santos argues that law is an ‘abyssal thinking’. By this he means that law radically divides social reality into two imagined territories. On ‘this side of the line’ is all that can be thought of in terms of law, along a legal/illegal binary, from the centre to the periphery of the legal world. On ‘that side of the line’, in the abyss, is that which is entirely other than law (de Sousa Santos, Reference De Sousa Santos2007, pp. 1−2). In law's terms, it is absent, irrelevant, radically excluded and entirely negated. But it must also be law's constitutive outside. It is a zone of violence and appropriation, to be contrasted with the zone of regulation and emancipation inside the line (de Sousa Santos, Reference De Sousa Santos1987, p. 444). In the English discourse on Muslim divorce practice, the imagined foreign Shari'a of barbarous criminal punishments occupies this space. It is something which helps constitute English law by contrast, but cannot be thought in law's terms; it is entirely lawless (de Sousa Santos, Reference De Sousa Santos2007, p. 7). At the edge of the law's imagined territory, close to the line – in a zone which is poorly regulated and lacking in state institutional and symbolic resources − are those potentially illegal practices which come close to that lawlessness (de Sousa Santos, Reference De Sousa Santos1987). The colonial zone was the original ‘other side of the line’, but in the present time, Santos argues, we are experiencing a ‘return of the colonial’ (de Sousa Santos, Reference De Sousa Santos2007, p. 13). In this moment, it becomes apparent that the abyssal line is not fixed, albeit it may be heavily policed (p. 12). The metropolitan perceives the colonial's approach towards the line, and his potential demand to cross to the other side; to seize legal capital and be part of law's ‘we’. The colonial threatens to disfigure or unsettle law. And the metropolitan responds to this trespass and its attendant vulnerability by redrawing the line ‘as at close a range as is necessary to guarantee security’, shrinking the space on this side, and excluding the colonial (pp. 12−15; Balibar, Reference Balibar2004, p. 110). As Wendy Brown argues, the act of ‘walling out’ restores law's identity by cleansing it of its connection to the lawless outside (W. Brown, Reference Brown2010, p. 95). This is what has been happening in the UK since 2008: government met Shari'a's trespass onto the domain of family law, not by expelling it from the national territory, but by drawing a new line, closer to law's safe territory. We observe this, Santos argues, when the state withdraws from areas of social regulation. It becomes less concerned with regulation than with meta-regulation; with the distribution of power between private interests. It forsakes hard connections to state law in favour of ‘soft law’ and contractual obligation (de Sousa Santos, Reference De Sousa Santos2007, p. 17), or of adopting methods of indirect rule which leave group leaders with a significant measure of power over lower-status citizens at the margins (de Sousa Santos, Reference De Sousa Santos1987, p. 453); hence the effective ‘private ordering’ of Muslim divorce practice. The pressure of the colonial's advance may be such that some of the characteristics of the ‘other side of the line’ − the scope for violence and appropriation − now appear within law's territory (de Sousa Santos, Reference De Sousa Santos1987, p. 442) and some citizens find themselves in the grip of law's inclusive exclusions − its dis-locations (Balibar, Reference Balibar2006). Those who were already at law's margins by virtue of their association with Shari'a – in unrecognised marriages, in communities where Shari'a councils had influence – now find themselves strangers to law, even as they dwell in the middle of the ‘we’. They inhabit a ‘messy’, disorganised and ambivalent legal territory (de Sousa Santos, Reference De Sousa Santos2007, p. 15). This happens partly because, as Santos wryly observes, no aspect of this process ‘can be fully acknowledged, precisely because the other side of the line was from the beginning incomprehensible’ (p. 15). And, borrowing from Wendy Brown, we could argue that it happens partly because, in a globalised context, a final ‘walling out’ can only ever be a fantasy; the line will inevitably be breached again and again (W. Brown, Reference Brown2010, p. 88). So the government's policy has been to repeatedly state the simplicity of acceptable, responsible and contractual engagement with Shari'a, without ever admitting its potential shortcomings.

In the second section of this article, I suggested one route to feminist engagement with the government's privatisation of Muslim divorce practice; a critique of ‘private ordering’, which would deconstruct the responsible bargaining subject at its core. Feminist engagements might also benefit from stepping outside the private ordering framework which the state has inaugurated, and paying attention to how it is supported. I have argued, following Santos's analysis, that the responsible bargaining subject has been installed by a vulnerable state as a border guard against the incursion of a dangerously disloyal ‘other’; a colonial trespassing on law's territory. On the government discourse, the subjects are opposites – the responsible bargaining subject is so coherently constrained precisely because she must be the opposite of the dangerous colonial. Feminist interventions which wish to undo her myth might do well to begin with the myth of the dangerous colonial. Santos's analysis suggests the impossibility of the coherence of either of these subjects. Rather, it suggests the existence of messily regulated split subjects, dwelling at law's margins, in unacknowledged spaces where law has been unable to fully secure its borders. If we imagine that there is play in the borders, what new figurations or ways of being ‘loyal’ or ‘legal’ might emerge from our blind spots (Braidotti, Reference Braidotti and Passerini2010, pp. 27−30)? In the remainder of this article, I want to consider the ‘colonial’ suggested by Santos, who troubles law by her demand to cross to its territory from ‘the other side’, who aims to seize legal capital and be part of law's ‘we’. Can she be found where the state expects to be in control: in the courts during divorce litigation? Santos's analysis suggests that colonials as well as biddable divorcing subjects might inhabit the space of contract, councils and consent orders which the government designated as law's borderlands in the course of the Shari'a debate.

V Claiming disloyalty: beyond financial outcome?

Feminist multiculturalism is slow to embrace disloyalty. One of the most influential works in the field is Ayelet Shachar's (2001) Multicultural Jurisdictions. In that book, disloyalty emerges as a charge which might be levelled against women who are affiliated to religious groups, and used to justify their exclusion from law. Much of Shachar's effort is directed at circumventing that charge, and holding the state responsible for women's oppression even where they hold themselves at a distance from the state. Shachar achieves this in two moves. First, she downplays the significance of potential disloyalty, by emphasising both the state's role in producing and supporting the groups which make competing claims on women's loyalties (Shachar, Reference Shachar2009) and by insisting on the primary importance of supported access to the key state institutions as a fundamental aspect of citizenship (Shachar, Reference Shachar2001, p. 21). Second, she gives a limited account of the demands that women might make on law: her concern is that women should not be deprived of access to practical benefits; in particular, social and economic resources, because of their group membership (p. 21). What is key about the women who people Shachar's theory is that they will not be any trouble to the state. Although Shachar acknowledges that some disputes may have ‘symbolic’ dimensions insofar as they raise broader political questions about membership in the polity (p. 18), she seems to contemplate that those questions must be answered outside the courts as part of processes of institutional design. She builds her legal imaginary around women who are concerned only with the distributive outcome of judicial decisions in their own cause, and whose desires are calculable in simple monetary terms (p. 55). Now, Shachar's broader project is one of ‘transformative accommodation’ (p. 117). But this does not mean that the subjects of her theory are ‘colonials’ who seek to challenge, defy, transgress or remake English law. They only seek to take from it on its own terms. They may indirectly spark reform in either civil or religious law, primarily by choosing the former forum over the latter (pp. 141−43). The business of contesting, challenging and even transforming legal principle in the courts is left to judges. Women are willing objects of judgment; scarcely a threat to law at all. Shachar's conception of women's legal agency is thus limited by her radically depoliticised sense of the content of citizenship. If the ‘colonial’ appears here, it is in a very biddable form.

Shachar does not give an account of citizenship under a heavily culturalised politics of legal agency of the sort that obtains in the UK, or say very much about the citizen whose demands on law might be read as disloyal, irresponsible or threatening. The beginnings of an alternative, clearly political, conception of women's legal agency can be found in Karen Knop's recent work on private international law. Knop develops a theory of legal agency in new times which addresses both the proper conditions of women's access to law, and the purposes and desires which shape that access. First, as regards access, Knop begins from the same position as Shachar: she treats access to civil law's forums and remedies as a citizenship question (Knop, Reference Knop2008, p. 310). But Knop's discussion of that entitlement explicitly speaks to contested questions of membership which are only implicit in Shachar's Multicultural Jurisdictions. For Knop, citizenship is disaggregated so that political citizenship – in the sense of membership in a polity – and what she terms ‘legal citizenship’ are held apart from one another. One may be a legal citizen, and claim access to the law, even if one's membership of the polity is precarious because one is a foreigner, or a suspected enemy of the state (pp. 321−28). Legal citizenship might provide a point of purchase on the state for Muslim women in the face of the ‘jingoistic and parochial impulses’ of an unsettled government. Anika Anis's exercise of legal agency in Babar v. Anis may bear this analysis out. Footnote ⁴⁸ Anika was born in Karachi. She married Sohail Babar in a nikah ceremony in Pakistan in 1999. He was a reluctant bridegroom, marrying her in deference to his family, who had arranged the marriage for him. At the time he was still legally married to his first wife, but Anika married him in order to secure the opportunity to move to the UK. When his divorce was finalised in 2002, Sohail sponsored Anika's entry to the UK. However, he refused to enter into a civil marriage with her; they lived together but the marriage was not consummated. Sohail registered a talaq divorce with a Shari'a council and wrote to the Immigration and Nationality Directorate of the Home Office, stating that his marriage to Anika had been forced. The Shari'a council did not advise any provision for financial settlement, and the marriage was later annulled. Anika did not apply for financial relief. She applied for exceptional leave to remain in the UK, which was refused. She appealed. In the meantime, Sohail brought possession proceedings, compelling her to leave the flat where she had lived with him. In order to obtain a bargaining chip which would compel Sohail to support her in her immigration appeal, Anika defended the claim. She argued promissory estoppel – that Sohail had induced her to come to the UK with a promise of marriage and financial support, and that she had acted to her detriment in reliance on that promise. Her defence failed, but her claim to standing before the law illustrates Knop's point. Anika is subject to multiple exclusions: both those wrought by the state in terms of her immigration status, and those which Sohail is able to impose on her because of her precarious legal status. Her creative and strategic engagement with private law gave her an opportunity to enhance her bargaining power and resist his dominance for a time.

Second, Knop presents a richer picture than Shachar of women's purposes in accessing civil law. Shachar in Multicultural Jurisdictions presents state and religious law as separate entities inhabiting defined jurisdictions or spheres, albeit with some cross-fertilisation across the inside − outside divide (Shachar, Reference Shachar2001, p. 40). Knop, however, goes further. Like Albertyn in this issue, she emphasises that civil courts are a space where religious and civil law may encounter one another, albeit on civil law's terrain. She suggests that women themselves might exercise legal agency which goes beyond access, strategy and bargaining, beyond submission to law's protection, and which might extend to use of the civil courts as a ‘dislocated experimental space’ within which Muslim divorce practice can be contested. In court, she says, Muslim litigants can make concrete context-driven arguments about competing interpretations of a religious or foreign legal principle, and express dissent from dominant readings (Knop, Reference Knop2008, p. 340; Knop, Riles and Michaels, Reference Knop, Riles and Michaels2012, p. 631). She thus opens the possibility of transformative legal agency.

But although Knop's work provides a more politically engaged model of legal agency than that available on Shachar's distributive theory, it is lacking in some respects. First, Knop is writing about private international law, which has at its disposal mechanisms for circumventing concerns around foreignness and loyalty which are not available to domestic family law. That disciplinary context allows Knop to assume the possibility of a formally ‘open’ law to which individuals can have recourse in the face of political exclusion; a profoundly hospitable law. Our theory, on the other hand, must give an account of political circumstances in which, as in Britain, one's engagement with domestic law is freighted with exclusionary government expectation, so that one cannot simply retreat into law and let politics happen ‘elsewhere’. Second, Knop is part of a tradition in feminist multiculturalism which has emphasised ‘voice’: women's potential to work within their groups to ‘resist, revise and reinvent’ inherited traditions (Deveaux, Reference Deveaux2006, p. 7; Sezgin, Reference Sezgin2010) and mediate tensions between the demands of religious legal orders and the opportunities available at liberal law (Phillips, Reference Phillips2007, p. 157). More than that, she allies herself with emerging work which identifies the civil courts as a key forum within which women can do some of the work of renegotiating oppressive customs (Deveaux, Reference Deveaux2006, p. 5; Fishbayn, Reference Fishbayn1999, pp. 167−68), challenging cultural hierarchy from a position of ‘legal counterpower’ (Alexander, Reference Alexander2006). However, it is not clear that her theory embraces those who deploy their legal citizenship in a way which is unsettling to law itself. If our account of legal agency is thus confined, it must come up against the familiar criticism that feminists are only interested in women's agency before the law where it is transgressive of cultural norms, and bends religious law along an arc of liberalisation (Mookherjee, Reference Mookherjee2009, p. 21; Sandberg, Douglas, Doe, Gillat-Ray and Khan, Reference Sandberg, Douglas, Doe, Gillat-Ray and Khan2013, p. 26). On the contrary, it must make space for those other than ‘infantile citizens’, full of faith in the system (Berlant, Reference Berlant1997). It must make space for critique of the terms on which the state facilitates access to law.