Introduction
Can a declaration of non-recognition of a voidable overseas marriage, where at least one of the parties is domiciled in England and Wales, lawfully be made? The answer to this question depends on whether a voidable marriage is ‘a marriage at its inception void’ and thus caught by section 58(5)(a) of the Family Law Act 1986. This provides:
No declaration may be made by any court, whether under this Part or otherwise … that a marriage was at its inception void.
If a voidable marriage is not invalid at its inception, then the court is not barred from declaring that it has become invalid. But if a voidable marriage is invalid at inception, then the court may only make a declaration to that end in highly exceptional circumstances.
In NB v MI [2021] EWHC 224 (Fam), I reasoned at paras 53 and 58 that section 58(5)(a) of the 1986 Act clearly applies to voidable marriages. In Re SA (Declaration of Non-Recognition of Marriage) [2023] EWCA Civ 1003, at paras 85–90, Moylan LJ (with whom King and Dingemans LJJ agreed) took a contrary view.
At para 88 it was stated by Moylan LJ that in NB v MI the terms of section 16 of the Matrimonial Causes Act 1973, originally enacted as section 5 of the Nullity of Marriage Act 1971, had been overlooked.Footnote 1 This provided (when enacted) that:
A decree of nullity granted after 31st July 1971 in respect of a voidable marriage shall operate to annul the marriage only as respects any time after the decree has been made absolute, and the marriage shall, notwithstanding the decree, be treated as if it had existed up to that time.Footnote 2
In three separate places in his judgment (paras 82, 88 and 90) Moylan LJ stated that by operation of section 16 of the Matrimonial Causes Act 1973 a voidable marriage is not void at its inception and is therefore not caught by section 58(5)(a) of the Family Law Act 1986. Moylan LJ therefore contends that the effect of section 16 of the 1973 Act is that a voidable marriage starts off fully valid but only on making a decree absolute of nullity becomes invalid, and then only at and from that point. This is not only conceptually challenging and based, as I contend, on a misreading of the statutory language, but is directly contrary to long-established and powerful authorities.
I beg to differ and contend that a historical analysis of the concept of a voidable marriage shows that a decree of nullity in respect of a voidable marriage unquestionably declares it non-existent ab initio, but that for all civil purposes such a marriage is ‘treated’ (the word used in section 16) or ‘deemed’ or ‘regarded’ (the verbs are synonymous in this context) as existing up to the date of decree absolute. This analysis shows that the core property of initial invalidity of a voidable marriage remains intact but that there is a procedural bar preventing anybody from relying on that invalidity prior to the date of decree absolute. I considered this in some detail in Tousi v Gaydukova [2023] EWHC 404 (Fam). This article restates and expands the historical analysis in that decision and in NB v MI. It is to be hoped that it will not be peremptorily dismissed as a ‘gadfly attempt to challenge conventional wisdom’.Footnote 3
Concept of the voidable marriage
The current law concerning the annulment of marriages finds its roots in pre-Reformation Roman Catholic canon law, which came to rule all marriages following the adoption of Christianity as the state religion of the Roman Empire by the Edict of Thessalonica in 380. Prior to the Council of Trent in 1563 canon law essentially required only that the parties should have the capacity to marry (habiles) and give their free consent (consensus) per verba et praesenti. Footnote 4 Capacity to marry depended on there being no diriment impediments, all of which had their origin in theology. These included the prohibited degrees of consanguinity or affinity as described in the book of Leviticus; ligamen (a prior subsisting marriage); non-age (then 12 for females and 14 for males); and impotence (it being said that natural law demanded a capacity to generate). Free consent could be vitiated by duress, error or fraud, as well as by unsoundness of mind.
Canon law did not acknowledge a difference between a void and a voidable marriage. If an impediment existed, or if free consent was not given, the marriage was void ab initio and for all purposes.
Prior to the Reformation the ecclesiastical courts, applying canon law, and answerable ultimately to the canonical Court of Final Appeal – the Sacred Roman Rota – had the sole jurisdiction to grant a decree of nullity in relation to a void marriage. But issues as to the validity of a marriage could arise in the common law courts and the Court of Chancery. In Ross-Smith v Ross-Smith [1963] AC 280, 330 Lord Hodson explained:
Before the Reformation the line drawn in the civil courts was between marriages which could be proved invalid only by a sentence of the ecclesiastical courts and those which could be impeached by other evidence.
Where there was a sentence of the Ecclesiastical Court that was, before the Reformation, conclusive. However, after the Reformation the common law courts were not prepared to accept what they saw as the injustice of the ecclesiastical courts disinheriting children after the death of a parent by declaring the parents’ marriage to be void by reason of affinity or consanguinity. And so, the concept of the voidable marriage came into existence.
‘Voidable is void ab initio’
In Ray v Sherwood (1836) 1 Curt 173, 188 Dr Lushington in the Consistory Court explained how the concept came into being:
It was the interference of the Common Law Courts, which, in such cases, prohibited the spiritual courts from bastardising the issue after the death of one of the parties that created the distinction – the very unnatural distinction – of voidable and void: for voidable is void ab initio. [Emphasis added.]
On appeal, this development was elaborated by Sir Herbert Jenner, Dean of Arches, in Ray v Sherwood and Ray (1836) 1 Curt 193, 199–200:
Originally, as now, these marriages were void ab initio, when sentence was pronounced by the Ecclesiastical Court: and it appears that the Ecclesiastical Courts were in the habit of annulling these marriages, even after the death of the parties, after the death of both, or of one only. And this seem to have been the practice antecedent to the Canon of 1603, as will be evident from a reference to the Articuli Cleri (2 Inst. 614), by Archbishop Bancroft, in the 3d James I (in the year 1606), whence it appears that the practice had existed for a long time before, and that the Ecclesiastical Courts complained of the interference of the Temporal Courts in cases of ecclesiastical cognizance … The practice then clearly existed at that time of declaring these marriages void after the death of the parties, and the Temporal Courts interfered for the purpose of protecting the interest of the issue of such marriages, and not that of the guilty parties, for as it appears from the case of Harris v. Hicks (2 Salk. 548), in the 4th and 5th of William and Mary, where a man had married the sister of his deceased wife, and it was suggested that the second wife was dead, and a son, the issue of the second marriage, would be entitled to lands, the Temporal Court in that case issued a prohibition against these Courts proceeding to annul the marriage between the parties after the death of one of them, but it did not prohibit them from punishing the survivor for the incest committed during cohabitation.
In Brook v Brook (1861) 9 HLC 193, 204 the Attorney-General, Sir Richard Bethell (later Lord Westbury LC) put it thus:
Marriages within the prohibited degrees were, Hill v. Good (Vaugh. Rep. 302), void by the common law of England, which was founded upon God's law; but when the ecclesiastical courts attempted to enforce that law to the extent of declaring, after the death of the parents, the children to be illegitimate, the common law interfered to prevent that consequence, and hence grew up the distinction between marriages void and voidable. The latter word is not quite accurate. It should have been said, that the marriage was void, but that the law would not allow it to be so treated after the death of one of the parties. The ecclesiastical jurisdiction, however, continued with regard to the punishment of the survivor, as Harris v Hicks (2 Salk 547) expressly declares. In such marriages, the persons are inhabiles.
The development of the concept saw the creation of what were referred to as civil and canonical disabilities (i.e. impediments). Civil disabilities were those where a party was absolutely incapable of entering into any marriage whatsoever. These included non-age, ligamen (i.e. already married to another), and ‘idiocy’ (i.e. permanent unsoundness of mind). Canonical disabilities were those where the impediment was relative such as affinity and consanguinity, duress and error, or temporary, such as impermanent unsoundness of mind.
The concept of the voidable marriage was also applied to some (but not all) of those marriages which had been validly formed per verba et praesenti Footnote 5 but which had not been consummated.
These marriages – ratum sed non-consummatum – included those where the parties were advanced in years and lived tanquam soror vel tanquam frater (as sister and brother). They were not uncommon, and were, of course, valid: see A v B (1868) LR 1 P& D 559, 562 per Sir James Wilde JO, later Lord Penzance, where he said:
For although it has been said that the procreation of children is one main object of marriage, yet it cannot be doubted that marriages between persons so advanced in years as effectually and certainly to defeat that object, are perfectly legal and binding. The truth is, consensus non concubitus facit matrimonium.
If the cause of the non-consummation was impotence either quoad hanc or hunc such a marriage was void at canon law. If the marriage was not consummated for any other reason then it was, and remains, exceptionally, dissoluble at canon law, although the power to do so has always been reserved to the Pontiff. The grounds for dissolution were, and remain, stringent.
Following the break with Rome, a marriage defective for impotence was voidable only: see A v B (1868) LR 1 P& D 559 where Sir James Wilde stated at 561:
The gradual declension of spiritual authority in matters temporal has brought it about that all questions as to the intrinsic validity of a marriage, if arising collaterally in a suit instituted for other objects, are determined in any of the temporal courts in which they may chance to arise. Though at the same time a suit for the purpose of obtaining a definitive decree declaring a marriage void which should be universally binding, and which should ascertain and determine the status of the parties once for all, has, from all time up to the present, been maintainable in the ecclesiastical courts or the Divorce Court alone. How then, it may be asked, does it happen that the particular ground of nullity which is raised incidentally in this suit has not followed the fate of all other grounds of nullity, and become cognizable in the temporal courts? The answer is, that impotence does not render a marriage ‘void’ but only ‘voidable.’
Non-consummation due to impotence has been a ground for a voidable marriage since the early days of the concept of voidability.
For those marriages unconsummated for a reason other than impotence the pontifical dissolution power was not incorporated into domestic law after the break with Rome until 1937. Section 7 of the Matrimonial Causes Act 1937 added ‘new’ grounds of nullity to the existing grounds. Henceforward, a marriage would be voidable where there was a wilful refusal to consummate.Footnote 6 The legal construct of the voidable marriage was to be applied to such marriages, the predominant view being that a wilful refusal to consummate gave rise to a ground for nullity: see Ramsay-Fairfax v Ramsay-Fairfax [1956] P 115, 132–133 where Denning LJ stated:
I must say a word about Inverclyde v. Inverclyde [1931] P 29, 42 in which Bateson J. held that in a case of nullity on the ground of impotence the only court which had jurisdiction was the court of the domicile. The basis of his reasoning was that a case of nullity, whether for impotence or for wilful refusal, was much more like a suit for divorce than anything else; and that it should be equated, so to speak, with a suit for dissolution and be governed solely by the law of the domicile, and that the only courts which should have jurisdiction should be the courts of the domicile. He said, in a sentence: “To call it a suit for nullity does not alter its essential and real character of a suit for dissolution.”
I beg to differ from this view. Looking at the ground of wilful refusal from a legalistic standpoint, and treating marriage as a contract, the remedy of nullity does look like a remedy of divorce or dissolution, because it depends on events which occur subsequent to the marriage; but looking at it from a sensible standpoint, and having regard to the true ends of marriage, one of the principal aims of which is the procreation of children, it seems to me that the remedy falls more truly within the category of nullity. No one can call a marriage a real marriage when it has not been consummated; and this is the same, no matter whether the want of consummation is due to incapacity or to wilful refusal. Let the theologians dispute as they will, so far as the lawyers are concerned, Parliament has made it quite plain that wilful refusal and incapacity stand together as grounds of nullity and not for dissolution: and being grounds of nullity, they fall within the old ecclesiastical practice, in which the jurisdiction of the courts is founded upon residence and not upon domicile.
The alternative view is expressed by Joseph Jackson who says that while such a voidable marriage ‘wears the coat of nullity’, it ‘speaks with the voice of dissolution’.Footnote 7 It is quite difficult to see a marriage voidable for wilful non-consummation as void ab initio. But, in contrast to this curious ‘new ground’, no other type of voidable marriage has an intrinsic characteristic at odds with it being void ab initio.
Effect of a decree of nullity in a voidable marriage case
As has been seen, where a marriage had not been consummated the court could bring it to an end by a decree of nullity. Although de facto such a decree could be seen as a divorce by another name, that was not the case de jure. This is because the legal remedy that was devised for such cases was to annul the marriage.
The decrees nisi and absolute of nullity in the case of a voidable marriage have always stated that the marriage ‘be annulled’. For example, Form 10 in the Appendix to the Matrimonial Causes Rules 1973Footnote 8 provides:
Referring to the decree made in this cause on the […] day of […] 19[…] , whereby it was ordered that the marriage in fact solemnised on the […] day of […] 19[…] , at […] between ‘the petitioner and the respondent be annulled unless sufficient cause be shown to the court within from the making thereof why the said decree should not be made absolute, and no such cause having been shown, it is hereby certified that the said decree was on the […] day of […] 19[…], made final and absolute and that the said petitioner was from that date and is free from all bond of marriage with the said respondent. [Emphasis added.]
The decree does not declare the marriage was valid at its inception; on the contrary, by ordering that it ‘be annulled’ it declares it invalid at inception. In legal language to ‘annul’ can only mean to declare void. And if a thing is declared void, then it is declared never to have existed. In P v P [1916] 2 Ir R 400, 429 Ronan LJ put it thus:
Neither in the Spiritual Court nor the Civil Court was there such a theory as a marriage valid up to a certain time, and thenceforward void. This could only arise under a special Act of Parliament. Apart from such an Act, the parties were either never married or married for ever, and this was so both in the Spiritual and Temporal Courts.
Ronan LJ then recorded how Lord Cranworth in the House of Lords in Lewis v Hayward (1866) 35 LJP & M 105 (a voidable marriage due to impotence) had stated:
As we are to make a precedent in this case, I should rather recommend your Lordships not only to reverse the decree which has been made, but to decree that the marriage was absolutely null and void ab initio.
The view of Ronan LJ was followed in the Northern Irish case of Mason v Mason [1944] NI 134, 162 where Sir James Andrews LCJ held:
It cannot, however, be too clearly understood that, although the actual decree of nullity is requisite in the case of voidable marriages, this does not mean that the marriage was valid up to the date of such decree and only invalid thereafter.
However, before 1971 the old form of the decree was the same for both void and voidable marriages and would pronounce the marriage to have been and to be absolutely null and void to all intents and purposes in the law whatsoever: see de Reneville v de Reneville [1948] P 100, 106 per Lord Greene MR. He continued at 110:
It is perhaps unfortunate that a form of decree which was appropriate when a marriage was regarded as indissoluble and could only be got rid of by decreeing that it had never taken place is still used indiscriminately in the cases of both void and voidable marriages. It is particularly anomalous in the case of the new grounds of nullity laid down by the Act of 1937.
Therefore, for the reasons explored below, it was necessary, following the implementation of section 5 of the Nullity of Marriage Act 1971 on 31 July 1971, for the decrees nisi and absolute in a voidable marriage case to abandon this language and to state simply that the marriage was annulled and that the parties were only freed from all the bonds of marriage from the date of the decree. This is what was achieved by Form 10 in the Appendix to the Matrimonial Causes Rules 1973.
The difference between void and voidable marriages
For many years the definitive judgment on the difference between void and voidable marriages was that of Sir John Nicholl sitting in the Prerogative Court of Canterbury in Elliot v Gurr (1812) 2 Phillim 16, 18–19. He stated:
Now, the difference between void and voidable is so clear that no person who ever looked into any elementary book on the subject is ignorant of it. The canonical disabilities, such as consanguinity, affinity, and certain corporal infirmities, only make the marriage voidable, and not ipso facto void, until sentence of nullity be obtained; and such marriages are esteemed valid unto all civil purposes, unless such sentence of nullity is actually declared during the lifetime of the parties.
Civil disabilities, such as a prior marriage, want of age, idiocy, and the like, make the contract void ab initio, not merely voidable: these do not dissolve a contract already made; but they render the parties incapable of contracting at all: they do not put asunder those who are joined together, but they previously hinder the junction; and if any persons under these legal incapacities come together, it is a meretricious and not a matrimonial union; and, therefore, no sentence of avoidance is necessary. The present is not a void, but a voidable marriage; and, therefore, not having been declared void in the lifetime of the parties, is valid to all civil purposes; and to all such purposes the deceased died the wife of William Gurr, and he was her husband, and their issue are legitimate.
In P v P [1916] 2 Ir R 400, 430–4–1 (which concerned a voidable marriage due to the husband's impotence) Ronan LJ pointed out that the judgment in Elliot v Gurr was misleading. The second sentence of the first paragraph was a quotation from Blackstone, but it omitted a crucial qualification from the great man concerning the power of the common law courts to prohibit the ecclesiastical court from annulling a marriage after the death of one of the parties. Blackstone had written (as cited by Ronan LJ):
For after the death of either of them the Courts of Common Law will not suffer the Spiritual Courts to declare such marriages to have been void, and therefore when a man had married his first wife's sister and after her death the Bishops’ Court was proceeding to annul the marriage and bastardize the issue, the Court of King's Bench granted a prohibition quoad hoc; but permitted them to proceed to punish the husband for incest: Harris v Hicks (1693) 2 Salk 548. These canonical disabilities being entirely the province of the Ecclesiastical Courts our books are perfectly silent about them.
As mentioned above, in Harris v Hinks (1693) 2 Salk 548, Hinks and his wife Elizabeth were prosecuted in the ecclesiastical court for an incestuous marriage, as Elizabeth was the sister of the first wife of Hinks. Elizabeth died pending the suit, whereupon Hinks pleaded abatement. The ecclesiastical court rejected the plea and gave sentence that the marriage was ‘incestuous, void and null’. On an application for an order of prohibition Holt CJ held that:
Prohibition was granted quoad annulling the marriage only, but [while] the Spiritual Court might proceed against Hinks after the death of his wife pro salute animae, to punish him for the incest, they should not make the marriage void, so as to bastardize the issue, for that was against law.
According to Ronan LJ (at 431):
This shows clearly that the word “until” [in the second sentence of the judgment in Elliot v Gurr] does not mean that the marriage was valid up to the date of the decree, and only invalid from that date, but that the declaration of “such marriage to have been void” meant void ab initio, and that the effect of the declaration was to bastardize the issue of the marriage born at any time. There is no doubt that this is the effect of a … decree of nullity.
Ronan LJ then referred to the Marriage Act, 1835 (5 & 6 Will. 4, c. 54 – Lord Lyndhurst's Act). Its preamble stated:
WHEREAS Marriages between Persons within the prohibited Degrees are voidable only by Sentence of the Ecclesiastical Court pronounced during the Lifetime of both the Parties thereto, and it is unreasonable that the State and Condition of the Children of Marriages between Persons within the prohibited Degrees of Affinity should remain unsettled during so long a Period, and it is fitting that all Marriages which may hereafter be celebrated between Persons within the prohibited Degrees of Consanguinity or Affinity should be ipso facto void, and not merely voidable.
And Section II provided:
And be it further enacted, That all Marriages which shall hereafter be celebrated between Persons within the prohibited Degrees of Consanguinity or Affinity shall be absolutely null and void to all Intents and Purposes whatsoever.
Ronan LJ cited these terms and continued (at 432):
This clearly shows that the effect of the decree of nullity was to make the marriage void ab initio, not to dissolve or annul it as from the date of the decree.
It is noteworthy that in P v P the Court of Appeal in Ireland had the benefit of submissions from Serjeant Sullivan KC, which it followed. Sullivan submitted (at 416–417):
A decree of nullity is a declaration that there has been no marriage at all; that it was void ab initio. The decree in the present case declares that the form of marriage between the petitioner and the respondent “was and is null and void.” By sect. 13 of the Matrimonial Causes (Ireland) Act, 1870), the Court is to proceed and give relief on principles and rules as nearly as may be conformable to the principles and rules on which the Ecclesiastical Courts had theretofore acted. The Ecclesiastical Courts did not dissolve a marriage. They only declared that there had been no marriage at all: Napier v. Napier [1915] P 124, per Lord Cozens-Hardy MR at p.186.
The word “voidable” when used as it is in some of the cases in reference to a marriage of this kind, is not used in its ordinary sense. It merely means that until there has been a declaration of nullity by a competent court, the apparent marriage cannot, for reasons of public policy, he treated as void. It can never be avoided by third parties, and even the parties to it may lose the right to have a. declaration of nullity by having, with a knowledge of the facts and the law, approbated the marriage, or by reason of delay in applying for relief. But the refusal of the Court to make a declaration of nullity where there has been approbation or delay is not on the ground that the parties are well married, but on the ground of a personal equity against the party applying for relief.
…
The decree of nullity, when pronounced, relates back to the ceremony of marriage, and is conclusive proof that there never was a valid marriage.
In his judgment Sir Ignatius O'Brien LC first pointed out the confusion surrounding the legal language. He said (at 422):
The expressions “void” and “voidable,” when used with regard to ordinary contracts, have a well-settled meaning, but when applied to marriage contracts, their use leads to confusion and to the obscuring of the real position of affairs.
He then explained (as recorded above) how Lord Cranworth in the House of Lords in Lewis v Hayward (1866) 35 LJP & M 105 had proclaimed, explicitly as a precedent, that a decree of nullity in a case of a marriage voidable due to impotence should declare the marriage null and void ab initio. Sir Ignatius then concluded (at 424):
Once there has been a ceremony of marriage prima facie the marriage is good. If a person who has gone through that ceremony chooses to come to the Court to have it declared that there never was a marriage, the Court will consider the conduct of the person seeking its aid, and may, on general equitable principles, refuse to permit such person to obtain it decree of nullity. That, of course, recognizes that, so far as the public are concerned, it is presumed to be a good marriage until set aside, and that, as between the parties, one of them may not be able to have it set aside by reason of the conduct of such party adopting and taking advantages under it. But once the Court considers that the party seeking relief has not done anything which prevents him or her obtaining the relief sought, and the impotency of the other party is proved, it pronounces a decree declaring the marriage to be a form only, and to be absolutely null and void at the time it purported to be effected, and from that time on.
For Ronan LJ likewise, the true view of the matter was that where there was a canonical disability or impediment the ecclesiastical court might refuse to declare the marriage void if the applicant had by his conduct disentitled him to have it so declared. He stated (at 433):
This is analogous to the familiar practice of the King's Bench in refusing to quash a conviction admittedly void on its face because the applicant had by his conduct become disentitled to the order. The refusal of the court to quash the void conviction on certiorari would not prevent the applicant or anyone else from setting up its invalidity if it was relied on in any other proceeding, and the court would, in such proceeding, be bound to hold it void.
On this footing the curious difference between void and voidable marriages becomes much easier to rationalise. A voidable marriage is to be seen as a subspecies of a void marriage where the relief of a decree might be refused if the conduct of the applicant disentitled him or her to the order and it would be unjust to the respondent to make it. The difference is essentially procedural, not one of substantive law. Molony LJ summarised the position pithily (at 441):
Where impotency exists in either party at the time of the solemnization of the marriage, the marriage may be avoided by the party who is fit. There was a good deal of discussion in the present case as to whether such marriages were to be considered absolutely void, or only voidable. It is clear on the authorities that such marriages (though in fact void) are deemed valid for civil purposes unless and until a sentence of nullity is obtained.
The decree when obtained, however, decrees the marriage null and void ab initio, and this form of decree was carefully settled by the Home of Lords in Lewis v Hayward and has always in substance been followed since. The decree in the present case declares that the ceremony between the plaintiff and the defendant was and is null and void.
This characterisation of a voidable marriage as being void ab initio but treated as valid for all civil purposes has been followed consistently. It has been put on a statutory footing, as will be seen. Thus, Lord Greene MR in De Reneville v De Reneville [1948] P 100, 111 stated:
… a void marriage is one that will be regarded by every court in any case in which the existence of the marriage is in issue as never having taken place and can be so treated by both parties to it without the necessity of any decree annulling it: a voidable marriage is one that will be regarded by every court as a valid subsisting marriage until a decree annulling it has been pronounced by a court of competent jurisdictionFootnote 9. In England only the Divorce Court has this jurisdiction. The fact that in both cases the form of the decree is the same cannot alter the fact that the two cases are in this respect quite different.
In Ross-Smith v Ross-Smith [1963] AC 280 it was held by the House of Lords that for the purposes of jurisdiction no distinction should be drawn between marriages said to be void and those which are voidable. The Committee confirmed the historical characterisation of a voidable marriage. Lord Morris of Borth-y-Gest stated (at 316):
The long-established and long-spoken-of contrast between “void” and “voidable” marriages was recognised in the Marriage Act 1835, by which it was provided that marriages within the prohibited degrees (which previously to the Act were voidable) should be absolutely void. If an ecclesiastical court pronounced a sentence of nullity in the case of a voidable marriage the form of words used was in fact the same as that used in reference to a void marriage. That a form of words was made to do service in two separate and distinct situations did not in any way alter the circumstance that the situations were different. The ecclesiastical courts could not put an end to a marriage. The church held firmly to the indissolubility of marriage. If, therefore, there was a decree which annulled a voidable marriage it was retrospective: it related back to the date of the marriage. But that did not mean that for all purposes it could be treated as bringing it about that there had never been a marriage. In recent times there have been cases (in regard to which no occasion to express an opinion now arises) in which the courts have refused to set aside certain arrangements made and completed on the basis of the existence of a marriage after such marriage (being voidable) was annulled. … So also the incidents of married status have been held to apply …. The position of children is now provided for by section 9 of the Matrimonial Causes Act, 1950 , which reads: “Where a decree of nullity is granted in respect of a voidable marriage, any child who would have been the legitimate child of the parties to the marriage if it had been dissolved, instead of being annulled, at the date of the decree shall be deemed to be their legitimate child notwithstanding the annulment.
Lord Hodson held (at 329–330):
Whether the marriage is void or voidable the question to be determined is always the same, that is to say, was it or was it not a valid marriage. Leaving formality on one side the question will be, was there capacity to marry? The distinction between divorce and nullity proceedings is vital. The one seeks to destroy a valid marriage, the other seeks to establish that there was no marriage.
It is true that today the grounds of voidability are comparatively few. Apart from impotence, most were introduced by the Matrimonial Causes Act, 1937. One ground, “wilful refusal to consummate the marriage,” although relating to a state of affairs which follows upon marriage and thus logically more appropriate to divorce than to nullity proceedings, can be justified as a ground for nullity since it may be difficult to say in all instances of failure to consummate a marriage that in the case of a male it is a veil for impotence and in that of the female it is due to invincible repugnance amounting to impotence. The evidence on both questions will often be the same. The distinction between void and voidable is not new but may vary in its application to particular defects. For example, before the Act of William IV c. 54, referred to by my noble and learned friend, Lord Morris of Borth-y-Gest, marriages entered into between spouses within the prohibited degrees were voidable. In other countries the distinction may or may not be drawn. Mental infirmity is now a ground for making a marriage voidable in this country, but according to Wolff, Private International Law, citing the Swiss Civil Code, ss. 16, 97, 102 No. 2, the effect in Switzerland is to render the marriage void. The canon law did not itself recognise a distinction between void and voidable marriages, and before the Reformation the line drawn in the civil courts was between marriages which could be proved invalid only by a sentence of the ecclesiastical courts and those which could be impeached by other evidence, see P v. P [1916] 2 Ir.R. 400 at 434 per Ronan LJ After the Reformation when the common law courts interfered by prohibition to prevent the ecclesiastical courts from exercising jurisdiction in certain cases to annul a marriage after the death of one of the spouses the distinction between void and voidable began to be recognised. [Emphasis added.]
Prohibited degrees: some voidable marriages become void
Marriage within the prohibited degrees of affinity and consanguinity will render a marriage void: section 11 of the Matrimonial Causes Act 1973. It was not always thus. As has been seen, prior to the enactment of section 2 of Lord Lyndhurst's Act 1835 such marriages were voidable. In Sherwood v Ray (1837) 1 Moo. P.C.C. 353, 395 on the further, unsuccessful, appeal to the Privy Council from the decision of Sir Herbert Jenner, Dean of Arches, in a case where the marriage of the husband to his deceased wife's sister took place before 1835, Parke B giving the advice of the Privy Council held:
That marriage having been celebrated between persons within the Levitical degrees and so prohibited from intermarrying by Holy Scripture, as interpreted by the Canon law and by the statute 25 Hen 8, c 22, s 3 (1533: An Act concerning the King's Succession), was unquestionably voidable during the lifetime of both, and might have been annulled by criminal proceedings or civil suit. The suit in a criminal form is to punish the parties by public penance, and to declare the marriage void ab initio; it may be instituted by any one promoting the office of judge, with the permission of the court, though the prosecutor be wholly unconnected with the parties, and this is the only case of nullity of marriage to which it is applicable. In all other matrimonial causes the suit must be in civil form. Civil suits may, beyond doubt, be brought by either of the parties, who have each the greatest interest that their status should be determined and that they should be freed from the bonds of an incestuous intercourse. [Emphasis added.]
The Privy Council held that the civil suit could be brought by the wife's father in whose household she was living at the time of the marriage. Parke B continued (at 397):
The marriage, it must always be borne in mind, is incestuous according to the Divine law, and no complaint can be reasonably made if all persons are admitted to interpose, whose rights are affected, however remotely, by the existence of the marriage or its probable result, the birth of offspring; and it is to be recollected that this may be the only form in which any individual can question the marriage as a matter of right, for to promote the office of judge in a criminal suit requires the authority and consent of the court, and though this is obtained without difficulty in ordinary practice it cannot be demanded ex debita justitiae.
Thus the right of an affected third party to petition to annul a voidable marriage was the same as that for a void marriage. The legal effect of the order was the same for both types of invalid marriage. The order of the court in the case of a voidable marriage was to declare the marriage void ab initio.
In Ross-Smith at 396 and 348 Lords Reid and Guest opined in obiter dicta that the right of a third party to petition for the annulment of a marriage in which he had a pecuniary interest was confined to void marriages. Rayden on Relationship Breakdown, Finance and Children asserts this to be the law at para 1501.1.Footnote 10
The unsuccessful appeal in Sherwood v Ray was from the Dean of Arches, Sir Herbert Jenner, sitting in the Court of Arches: Ray v Sherwood and Ray (1836) 1 Curt 193. In his judgment Sir Herbert considered the terms of Section I of Lord Lyndhurst's Act of 1835 which provided:
I. Marriages before the passing of this Act of Persons within the prohibited Degrees not to be annulled.
Be it therefore enacted … that all Marriages which shall have been celebrated before the passing of this Act between Persons being within the prohibited Degrees of Affinity shall not hereafter be annulled for that Cause by any Sentence of the Ecclesiastical Court…
Did this purport to make the pre-1835 voidable-by-reason-of-affinity marriages valid ab initio? Sir Herbert was very clear that it did not, stating (at 204):
The Legislature has expressed, as strongly as it could do, that these marriages are still illegal and contrary to the law of God; although, for the protection of innocent parties, not to screen the delinquents, it has declared that those marriages shall be unquestioned which were celebrated at the time of the passing of the Act, under the particular circumstances and conditions mentioned in the first section. But this is not a question which the Court is called upon to determine, and it is therefore not necessary to proceed further into the consideration of it
Absence of consent: some void marriages become voidable
At canon law the absence of consent rendered a marriage void, and this principle transferred straightforwardly into domestic (ecclesiastical) law following the Reformation. However, it was a doctrine of Canon Law, adopted after the Reformation by English ecclesiastical law, that a marriage void on the ground that there was no consent at the time of the marriage ceremony could be ratified by consent voluntarily given subsequently, whereupon such consent was deemed to relate back to the time of the marriage (see NB v MI [2021] EWHC 224 (Fam) at para 101). Thus, if no valid consent was given at the time of the marriage in consequence of unsoundness of mind, and the marriage was therefore void, that void marriage could later be ratified following a later mental recovery.
Although unsoundness of mind was added as a ‘new’ ground for annulling a voidable marriage in 1937 (see above), along with those allied cases where consent had been deceptively procured (i.e. where there was at the time of the marriage an undisclosed pregnancy or an undisclosed venereal disease), unsoundness of mind had long been a ground that rendered a marriage void. In Countess of Portsmouth v Earl of Portsmouth (1828) 1 Hag Ecc 355; 162 ER 611 Sir John Nicholl sitting in the Court of Arches held (at 359–360):
The law of the case admits of no controversy, and none has been attempted to be raised upon it. When a fact of marriage has been regularly solemnised, the presumption is in its favour; but then it must be solemnised between parties competent to contract – capable of entering into that most important engagement, the very essence of which is consent. Without soundness of mind there can be no legal consent – none binding in law: insanity vitiates all acts. Nor am I prepared to doubt, but that considerable weakness of mind circumvented by proportionate fraud, will vitiate the fact of marriage – whether the fraud is practised on his ward by a party who stands in the relation of guardian, as in Harford v Morris which was decided principally on the ground of fraud – or whether it is effected his trustee procuring the solemnisation of the marriage of his own daughter with a person of very weak mind, over whom he has acquired arrest ascendancy. A person, incapable from weakness of detecting the fraud, and of resisting the ascendancy practised in obtaining his consent to the contract, can hardly be considered as binding himself in point of law by such an act. At all events, the circumstances preceding and attending the marriage itself may materially tend to show the contracting party was of unsound mind, and was, considered and treated by the parties engaged in fraudulently effecting the marriage.
When the Law Commission proposed codification of the law of nullity in 1968, it concluded that it would be anomalous to retain this curious exception to the absolute voidness of a marriage where there was no valid consent. Therefore, it suggested that it would be altogether more logical to categorise all cases of invalid consent as voidable marriages, and this was duly enacted in section 2(c) of the Nullity of Marriage Act 1971.
Putting the difference between void and voidable marriages on a statutory footing
In its 1970 Report the Law Commission proposed that in voidable marriage cases:Footnote 11
A decree of nullity of a voidable marriage should operate to annul the marriage only as from the date of the decree absolute …
This was duly enacted as section 5 of the Nullity of Marriage Act 1971 and is now section 16 of the Matrimonial Causes Act 1973. As enacted, it provided:
Effect of decree of nullity in case of voidable marriage
A decree of nullity granted after 31st July 1971 in respect of a voidable marriage shall operate to annul the marriage only as respects any time after the decree has been made absolute, and the marriage shall, notwithstanding the decree, be treated as if it had existed up to that time.
Its language was very carefully chosen by the Law Commission,Footnote 12 and was adopted by Parliament.
The language used in section 5 of the 1971 Act does not say that a voidable marriage was, as a matter of law, actually valid at inception. On the contrary, it leaves intact the old law as to the essential character of a decree of nullity. As explained above, a decree of nullity in a voidable marriage case orders that the marriage ‘be annulled’. Consequential to section 5 the wording of the decree therefore had to make clear that while the marriage was invalid at inception the decree only ‘operated’ to annul the marriage after the date of decree absolute and that the marriage was to be ‘treated’ as being valid for all practical and legal purposes (e.g. legitimacy of children) up to that date.
The terms of the Report make clear that the intention of the reform was to ensure that for all civil purposes a decree of nullity in a voidable marriage case only operated to annul the marriage from the date of decree absolute. This was because under the common law and ecclesiastical law the retroactive nature of the decree had caused some problems. For example, it had the effect of causing an ante-nuptial settlement made in anticipation of a valid marriage to fail. It was also unclear whether a person who remarried following a voidable marriage but without obtaining a decree of nullity, committed bigamy: in Wiggins v Wiggins [1958] 2 All ER 555, it was held that he had; in Mason v Mason [1944] NI 134 it was held that he had not.
The proposed reform resolved these questions and cleared up these anomalies.
There is nothing stated whatsoever, either in the 1968 Working PaperFootnote 13, or in the 1970 Report, or indeed in the language of section 5 of the 1971 Act itself, to suggest that the section was intended to overthrow the long-standing and deep-rooted doctrinal principle that it was impossible for a marriage to be valid up to a certain point in time but thenceforward void. Such a major change would have had to have been flagged up in the Working Paper and the Report. Instead, the Report merely said this:Footnote 14
We are, therefore, opposed to the abolition of the class of voidable marriages and think that it should be retained. But the effect of the decrees of nullity of a voidable marriage should be modified so as to make it clear that the marriage is to be treated in every respect as a valid marriage until it is annulled and as a nullity only from the date when it is annulled. [emphasis added]
In making this recommendation it is absolutely clear that the Law Commission was giving effect to Alternative (2) in the Working Paper which retained the essential character of a voidable marriage as being void ab initio.
The Law Commission further proposed, and Parliament enacted (in section 3(1) of the Nullity of Marriage Act 1971, now section 13(1) of the Matrimonial Causes Act 1973), that where a marriage is voidable no decree of nullity shall be made where the respondent satisfies the court that the applicant with knowledge that it was open to him to have the marriage avoided, so conducted himself in relation to the respondent as to lead the respondent reasonably to believe that he would not seek to do so and that it would be unjust to the respondent to make the order.Footnote 15 Again, this does not imply that the marriage was valid at its inception. On the contrary, for the reasons set out above, this provision reinforces the well-established principle that a decree of nullity of a voidable marriage acts retrospectively to annul the marriage ab initio, even if, for all practical and legal reasons, the marriage is treated as subsisting up to the date of decree absolute.
These reforms reflected the old ecclesiastical and common law, as explained above. They did not represent revolutionary change. The reforms did not seek to alter the substantive legal character of a voidable marriage by making them valid from inception up to the date of a decree of nullity.
Rather, the reforms do no more than to clarify the well-established procedural bar on a party, or any other person, relying for any civil purpose on the voidness of such a marriage at any time prior to decree absolute. The reforms mirror precisely the interpretation given by the Dean of Arches in Sherwood v Ray to section I of Lord Lyndhurst's Act 1835, namely that the pre-1835 marriages were not made valid by the Act but instead were procedurally protected from any penalty referable to their invalidity.
Summary of types of voidable marriage up to 2004Footnote 16
√, specifically designated as a voidable marriage; V, specifically designated as a void marriage; X, specifically rejected as a voidable marriage; n/a, not addressed.
It can be seen that the types of voidable marriage have altered very considerably over time. Only one type of voidable marriage – incapacity to consummate – has stood constant throughout the last 200 years.
Conclusion
It seems to me that the following propositions can be confidently asserted:
i. Whether the marriage is void or voidable, the question to be determined is always the same, that is to say, was it, or was it not, a valid marriage (Lord Hodson).
ii. The distinction between divorce and nullity proceedings is vital. One seeks to destroy a valid marriage, the other seeks to establish that there was no marriage (ibid).
iii. In the case of a voidable marriage, the decree which annuls a voidable marriage is retrospective: it relates back to the date of the marriage (Lord Morris).
iv. The effect of a decree of nullity in the case of a voidable marriage is to make the marriage void ab initio, not to dissolve or annul it as from the date of the decree (Lord Cranworth, Sir Ignatius O'Brien LC, and Ronan LJ).
v. A voidable marriage cannot be valid up to the date of the decree and invalid thereafter (Ronan LJ and Andrews LCJ)
vi. However, there is a procedural bar on any person relying for any civil purpose on the voidness of a voidable marriage until a decree annulling it has been pronounced by a court of competent jurisdiction (Sir Herbert Jenner, Sir Ignatius O'Brien LC, Ronan LJ, and Lord Greene MR).
vii. Section 5 of the Nullity of Marriage Act 1971 does not alter these elemental legal standards. It gives effect to Alternative (2) in the Law Commission's 1968 Working Paper which stated that ‘the decree should annul the marriage ab initio (as now), but the legal effect of the decree should be the same as it would be if the marriage had been annulled from the date of the decree’.
Therefore, it may be safely concluded that a nullity decree in respect of a voidable marriage (a) continues to amount to a retrospective order; (b) relates back to the date of the ceremony, declaring that there has been no marriage at all, i.e. that it was void at inception; but (c) is subject to the aforementioned procedural bar.
I respectfully suggest that it is a fallacy to assert that the Law Commission in 1973 (in its Working Paper),Footnote 17 and in 1984 (in its Report),Footnote 18 was referring only to void marriages when it proposed that there should be a bar on any court making a declaration that a marriage was invalid at inception (as opposed to issuing a decree of nullity). Such an argument fails to recognise the doctrinal nature of the remedy represented by a nullity decree in a voidable marriage case. Unless corrected by the Supreme Court, the Court of Appeal in Re SA seems to have annihilated centuries of canonical, theological as well as common-law jurisprudence concerning the status of voidable marriages.
The decision in Re SA has an arbitrary quality because the stated intention of the reform was to ensure that in all nullity cases the remedy should give rise to the right to claim ancillary relief (see the Report at paras 3.18–3.19). That objective would be emasculated if nullity proceedings could be bypassed in voidable marriage cases.
There is no logic in confining section 58(5)(a) of the Family Law Act 1986 to void marriages alone. This piece has shown how certain marriages have switched between the categories. As explained above, marriages within the prohibited degrees were at first voidable but were later made void by Lord Lyndhurst's Act 1835. In contrast, marriages without free consent were at first void marriages but were later made voidable by the Nullity of Marriage Act 1971. It is not as if the categories of void and voidable marriages are written in marble. What would be the point of confining the prohibition to certain types of invalid marriage but not others?
In his judgment at para 89 Moylan LJ stated that the 1973 Working Paper ‘only substantively addressed void marriages, namely those which were “initially” invalid and did not deal with voidable marriages at all’. This is not correct. The Working Paper addressed voidable marriages in para 43 (at footnote 95), in para 54, and in para 61 (at footnote 127). The 1984 Report referred to voidable marriages in footnote 206.
Parliament did not need to spell out in section 58(5)(a) that the proscription applied to both types of invalid marriage because even a cursory reach into history would demonstrate that it did. If Parliament had intended the proscription to be limited only to void marriages it could easily have said so. It does not follow that Parliament did not intend to include voidable marriages in section 58(5)(a) because it did not specifically mention them there. For the reasons I have given, it did not need to do so.
If section 58(5)(a) encompasses voidable marriages then it must follow that a declaration not to recognise the validity of an overseas marriage formed by a UK citizen in Pakistan, can only be made in highly exceptional circumstances, namely where recognition of the marriage would be directly at variance with the fundamental public policy of English law: Cheni v Cheni [1965] P 85. Forced marriage is an abominable practice, but it is hard to see in such a case (where time under section 13(2) of the Matrimonial Causes Act 1973 has not elapsed) what additional benefits or protections a declaration of non-recognition of an overseas forced marriage supplies over and above those inherent in a decree of nullity.
It is well-established that where the question of the invalidity of the overseas marriage turns on the personal capacity, or consent, of the parties, rather than on its due form, the law of the parties’ domiciles applies: section 14(1) Matrimonial Causes Act 1973; Sottomayor v De Barros (No.1) (1877) 3 PD 1, 5 per Cotton LJ; Berthiaume v Dastous [1930] AC 79, 83 per Viscount Dunedin; Tousi v Gaydukova [2023] EWHC 404 (Fam) paras 63–79. In Re SA nullity proceedings could have been pursued, and English law would have applied to the determination of the alleged ground of invalidity, namely want of free consent. By the time of the hearing before Newton J, the three-year time limit under section 13(2) of the Matrimonial Causes Act 1973 had not elapsed. The marriage ceremony had taken place in Pakistan. There was no issue concerning compliance with the due form required by Pakistani law. The ‘wife’ was at all material times domiciled in England.
By contrast, I accept entirely that it is possible to see in a forced marriage case, where time under section 13(2) has elapsed and nullity proceedings are therefore not possible, that a declaration of non-recognition of the validity of such a forced marriage would reflect the fundamental public policy of English law.
Acknowledgements
I acknowledge the very considerable assistance of Sir James Munby in the writing of this piece. He not only scrutinised it for legal, grammatical and presentational errors but also suggested a number of vital inclusions.
