Brennan v Bolt Burdon & Ors

Miss Margaret Brennan is the tenant of a flat of which the London Borough of Islington (Islington) is the landlord. She complains that she has suffered personal injuries from carbon monoxide exposure as a result of gas escaping from a faulty boiler. The exposure is said to have occurred during two periods of time, one between 1988 and November 1995, the other between August 1996 and October 1998. Islington has an obligation to maintain and repair the boiler. On 7 June 2001 the present proceedings were issued. Islington is named as second defendant. The first and third defendants, Bolt Burdon and Leigh Day & Co., are firms of solicitors alleged to have acted negligently when representing Miss Brennan between 1993 and 1996.

On Saturday 6 October 2001 the claim form, particulars of claim, a schedule of damage and expert reports were delivered to the several defendants. On 8 November 2001 Bolt Burdon and Leigh Day filed applications to set aside the claim form pursuant to CPR Part 11 Rule 1 because it had not been served within four months of the date of issue in breach of CPR Part 7 Rule 5 (2). The point sought to be taken was that, by reason of CPR Part 6 Rule 7, service was deemed to have taken place on Monday 8 October notwithstanding actual delivery on Saturday 6 October. Whereas service on the Saturday would have been within four months, service on the Monday was not. Islington did not file an application of that sort but learned of the applications of the other defendants on 22 November.

On 2 January 2002 Miss Recorder Plumtre made an order that the claim form had not been validly served within four months and struck out the claims against Bolt Burden and Leigh Day. In so doing she relied on the judgments of the Court of Appeal in Godwin v. Swindon Borough Council [2001] 4 All ER 641 and of McCombe J. in Anderton v. Clwyd County Council [2001] EWHC QB 161. On 7 January 2002 junior counsel then instructed on behalf of Miss Brennan orally advised her solicitor at the time against appealing the decision. Junior counsel had spoken to leading counsel who had informally opined:

“(1)

Rule 6.7 of the CPR was badly drafted and…Mr. Justice McCombe had incorrectly construed the rule in Anderton v. Clwyd County Council….

(2)

Lord Justice May’s comments in Godwin v. Swindon Borough Council…would make it extremely difficult to get permission to appeal. Lord Justice May was regarded as the ‘guru’ of the CPR and there would be a small chance of getting two Court of Appeal Judges to go against him.”

In the meantime Miss Brennan had commenced a second action against Islington and three firms of solicitors (including Leigh Day but, for limitation reasons, excluding Bolt Burdon) on 14 November 2001. On 13 February 2002 there was a hearing in Croydon County Court when Islington was seeking an extension of time for service of its defence in the second action. At this time Islington had still made no application in the first action. Whilst at the County Court in connection with the second action there was a conversation between Miss Brennan’s then solicitor (Mr. Scarles) and Mr. Cohen, a solicitor employed by Islington. It is described in Mr. Scarles’ witness statement as follows:

“Prior to the application…I spoke to Mr. Cohen who commented on the learned Recorder’s judgment and asked if, in the circumstances, I would be willing to discontinue the first claim. I said that I was provided that it was agreed that there would be no order for costs save for a detailed assessment of the Claimant’s costs. However I only agreed to this because of the learned Recorder’s decision. Islington was the alleged primary wrongdoer and the question of their liability had been investigated for over ten years. It was clear to me that in view of the learned Recorder’s judgment, the Claimant had no option but to discontinue the claim. If I had not agreed to Mr. Cohen’s suggestion, I have no doubt that, as I have already explained, Islington would have been able to strike out the claim and obtain an order for costs. As the Claimant was legally aided, I had a duty to the Legal Services Commission and I could not allow costs to be incurred unnecessarily. My conversation with Mr. Cohen about discontinuance lasted no more than a few minutes

After the hearing, I received a letter from Mr. Cohen in which he wrote

‘further to our meeting at court this morning, I understand that you wish to give notice of discontinuance of the first action against the Council and I consent to this on the basis that there will be no order for costs’

On 18 February 2002, I sent Mr. Cohen a draft consent order. On 21 February 2002 I received a signed consent order from him. On 25 February 2002 I sent this to the court but, by an oversight, I omitted to sign it. As a result, it was returned unsealed because the court refused to seal it.

On 12 March 2002, I received an e-mail from Mr. Thompson of St James Church Legal Advice Centre informing me that the Claimant did not agree to discontinue against Islington.

[I] ceased to act for the Claimant on 10 April 2002. ”

On 26 April 2002 Miss Brennan instructed Alison Trent, her present solicitor. On 3 July 2002 the Court of Appeal reversed the decision of McCombe J in Anderton v. Clwyd County Council. On 8 November 2002 Cox J gave Miss Brennan permission to appeal the order of Miss Recorder Plumtre out of time. In the light of the decision of the Court of Appeal in Anderton, Bolt Burdon conceded the appeal.

At some stage the action was transferred to the High Court. On 4 November 2002 an application was made on behalf of Islington to stay the first action until Miss Brennan had performed the agreement to discontinue on the basis that each party would bear its own costs or, alternatively, that the action be struck out on the basis that it had been compromised. On 11 June 2003 Deputy Master Eastman refused Islington’s application but granted permission to appeal. The appeal came before Morland J who, on 30 October 2003, dismissed it.

There is now before this court a second tier appeal for which Lord Justice Latham gave permission on 8 December 2003, observing

“both parties acknowledge that an issue of some importance is raised by this application and the appeal has a real prospect of success.”

In a nutshell, the important issue is whether the compromise of proceedings entered into by parties on the basis of a common mistake of law is void by reason of that mistake. (I shall refer to “common” mistake in circumstances in which purists might prefer “mutual” mistake because more of the references in the recent authorities have been to common mistake). The conclusion of Morland J was that, in the light of recent developments in the jurisprudence relating to mistake of law, the compromise in this case was vitiated. His judgment contains a fuller account of the history than it is necessary to set out for the purposes of this appeal. In it, the learned judge also observes that Miss Brennan is now being advised by at least the sixth firm of solicitors whom she has consulted in this matter. He added:

“an overall view of this litigation does little credit to the legal profession.”

He was specifically referring to the period ending on 26 April 2002 on which date Miss Brennan’s present legal advisors entered the fray. It is not suggested that they have lacked diligence.

Mistake of law

The turning point for the general principle came in Kleinwort Benson Limited v. Lincoln City Council [1999] 2 AC 349. Kleinwort Benson had made payments to a local authority under swap agreements which were thought to be legally enforceable. Subsequently, a decision of the House of Lords, Hazell v. Hammersmith and Fulham LBC [1992] 2 AC 1, established that such swap agreements were unlawful. Thereafter, Kleinwort Benson sought restitution of the payments on the basis of a mistake of law. The majority in the House Of Lords (Lords Goff of Chievely, Hoffman and Hope of Craighead) held that KleinwortBensonwas entitled to succeed upon that basis. The minority dissented, not on the issue of the ambit of mistake of law in principle but on the question of whether the declaratory theory of the common law required the case to be analysed in terms of mistake of law. Bilbie was overruled, along with other authorities to the same effect. Referring to the Law Commission’s Consultation Paper number 120 on Restitution of PaymentsMade Under a Mistake of Law (1991) Lord Goff referred to “the main criticisms” of the previously established principle. He described the distinction drawn between mistakes of fact and mistakes of law as producing results “which appear to be capricious”and to the exceptions and qualifications which “in truth betray an anxiety to escape from the confines of a rule perceived to be capable of injustice” with the result that “the law appeared to be arbitrary in its effect”. He added (all this being at page 372):

“As a result of the difficulty in some cases of drawing the distinction between mistakes of fact and law, and the temptation for judges to manipulate that distinction in order to achieve practical justice in particular cases, the rule became uncertain and unpredictable in its application”. ”

He concluded (at p 375H):

“…….the mistake of law rule should no longer be maintained as part of English law…..English law should now recognize that there is a general right to recover money paid under a mistake, whether of fact or law, subject to the defences available in the law of restitution.”

Although Kleinwort Benson concerned a restitutionary claim rather than a contractual one, it cannot be doubted that its effect now permeates the law of contract. In Pankhania v. London Borough of Hackney [2002] EWHC 2441 Ch the question arose as to whether a misrepresentation of law could now found a cause of action. In his lucid and trenchant judgment, Mr. Rex Tedd QC, sitting as a deputy judge, stated (at para 58):

“I have concluded that the ‘misrepresentation of law’ rule has not survived the decision in Kleinwort Benson. Its historical origin is as an offshoot of the ‘mistake of law’ rule, created by analogy with it, and the two are logically interdependent…..The distinction between fact and law in the context of relief from misrepresentation has no more underlying principle to it than it does in the context of relief from mistake…..The rules of the common law should, so far as possible, be congruent with one another and based on coherent principle. The survival of the ‘misrepresentation of law’ rule following the demise of the ‘mistake of law’ rule would be no more than a quixotic anachronism.”

The editors of the 28^th edition of Chitty on Contracts (1997) expressed the view that (para 5-018)

“a common mistake may now render a contract void even though the mistake is one of pure law.”

In the 29^th edition (2004), this sentence remains but with the addition of the words

“provided the mistake is such that it makes the contract adventure impossible.” (para 5-042).

The addition of that proviso is to accommodate the decision of the Court of Appeal in The Great Peace [2002] EWCA Civ 14

A compromise in civil litigation is a contract and it does not cease to be so when it is enshrined in a consent order, which is a “mere creature” of the contract: Huddersfield Building Company Limited v. Henry Lister & SonLimited [1895] 2Ch 273, 275, per Vaughan Williams J. In the unsuccessful appeal from that decision, the Court of Appeal concluded that a consent order

“can be impeached, not only on the ground of fraud, but upon any grounds which invalidate the agreement it expresses.”(per Lindley LJ at p 279; see also Lopes LJ at p 282 and Kay LJ at p 283)

Indeed, that case is authority for the proposition that a compromise enshrined in a consent order can be set aside by reason of common mistake.

The mistake there was, of course, one of fact. The next question is whether, post – Kleinwort Benson and its reception into the general law of contract, a compromise based on a mistake of law can be set aside, in particular where the mistake is in the form of a shared assumption about the meaning or correctness of an existing decision in a case involving different litigants. The present case is not the first one in which this question has arisen. It arose in the context of a consent order made in ancillary relief proceedings in S v S [2003] Fam 1. The parties had settled their differences and consented to an order on the basis of professional advice. Some weeks later White v. White [2001] 1 AC 596 was decided in the House of Lords. An application of it to the facts of S v S would have produced a more generous order in favour of Mrs. S than the one to which she had consented. Accordingly, she applied to set aside the consent order. Bracewell J refused the application. In so doing she founded her decision first on Barder v. Barder [1988] AC 20 in which the House of Lords held that a consent order in an ancillary relief case could be set aside on the ground that there had been a supervening event which had led to such a change of circumstances as to undermine or invalidate the basis of the consent order. Although Barder was concerned with supervening factual events, Bracewell J considered that a subsequent change in the law was also capable of amounting to a relevant supervening event. It did not do so in the case before her because, at the time of the consent order, it had been well known among family lawyers that the House of Lords might be about to change the law – a foreseeable event was not a supervening event. However, she then addressed the position on the basis of a mistake of law in the light of Kleinwort Benson. She concluded (at para 71):

“…..mistake of law as a vitiating factor ab initio has no place in consent orders for ancillary relief. In any event there would be public policy considerations against setting aside a consent order on such a basis by reason of the floodgates opening for all the orders made in the last quarter of a century or more. The principle that there should be an end to litigation must prevail.”

I say at once that I have some sympathy with that approach. As we have heard no submissions about special factors which may apply in ancillary relief cases, I shall not express any view about them. However, to the extent that Bracewell J sought to reconcile her conclusion with Kleinwort Benson, I have to say that, with respect, I cannot agree with her reasoning. She said that the decision of the House of Lords “is specific to the law of restitution and was not intended to apply across the board of every branch of law” (at para 70). Whether or not it applies to “every branch of law”, as I have already explained, I am satisfied that it is not confined to restitution. She purported to rely on a passage in the speech of Lord Hope of Craighead (at p 410H) but he was there dealing with judgments and orders made after adjudication and not by consent. I shall return later to the question of public policy. The fact is that in the present appeal Mr. Norman, on behalf of Islington, has chosen to approach his task on the basis that, in the light of Kleinwort Benson and its extension into contract, a common mistake of law may vitiate a compromise or consent order. His submission is that this is not such a case.

Before returning to the circumstances of this case, it is necessary to refer to some further relevant material from the authorities. The first one refers to compromises. In Huddersfield Banking, both at first instance and in the Court of Appeal, there is a recognition that, although compromises are susceptible to challenge on general contractual principles, there are limits. Thus, Vaughan Williams J said (at p 277):

“ if the arrangement come to was a compromise of doubtful rights and a give-and-take arrangement, parties to it could not afterwards have the compromise set aside because upon obtaining fuller information they thought they had made a bad bargain.”

On appeal, Kay LJ said (at p 284):

“A compromise takes place when there is a question of doubt and the parties agree not to try it out, but to settle it between themselves by a give-and-take arrangement. I quite agree that if this was a case of that kind it would be extremely difficult to interfere with the order.”

Mr. Norman submits that the present case involves a give-and-take compromise of that kind.

Although Huddersfield Banking was decided when mistakes of law could not generally vitiate contracts, it is pertinent to observe that when the House of Lords jettisoned the exclusion of mistakes of law in the context of restitution in Kleinwort Benson, a number of their lordships referred to potential defences to a restitutionary claim based on mistake of law. It is significant that the mistake of law there under consideration had not afflicted a compromise. There had been no compromise, indeed no litigation. The parties had proceeded on a common assumption as to the legality of the swaps which was later confounded by a case to which they were not parties. Lord Goff of Chievely said (at p 382G):

“I recognise that the law of restitution must embody specific defences which are concerned to protect the stability of closed transactions. The defence of change of position is one such defence; the defences of compromise, and settlement of an honest claim (the scope of which is a matter of debate) are others”

Lord Hope of Craighead added (at p. 412 F-G):

“Then there is the defence that the money was paid as, or as part of, a compromise. Brennan J in [David Securities Pty Ltdv. Commonwealth Bank of Australia (1992) 175 CLR 353], at p 395, said that, where a claim is satisfied by accord and satisfaction, a payment made in satisfaction is made in discharge of an obligation created by the accord; it is unaffected by any mistake as to the validity of the compromise. That must be so, irrespective whether the mistake is as to the facts or the law regarding its validity. In the Ontario Hydro case [1982] 1 SCR 347,380, Dickson J said that there was a head of public policy which recognised that there was a need to preserve the validity of compromises freely entered into with advice. I think that it is possible to find a more principled basis for the defence……But my main point is that it is available irrespective of the nature of the mistake.”

The final authorities to which I should refer relate to the issues of construction and risk in the context of mistake and compromises. The first is Bank ofCredit and Commerce International SA v. Ali [2002] 1 AC 251. This recent decision of the House of Lords was not cited to Morland J in the court below, nor indeed was it cited to us at the hearing. Our attention was drawn to it a week later by Mr. Bartle with the consent of Mr. Norman. Mr. Ali was a former employee who had been made redundant by BCCI. He made claims for appropriate statutory payments. Under the aegis of the Advisory, Conciliation and Arbitration Service he entered into an agreement to accept a sum “in full and final settlement of all or any claims….of whatsoever nature that exist or may exist” against BCCI. He did not know at the time that the business of BCCI had been and was being run in a corrupt and dishonest manner. After the company had gone into voluntary liquidation, amid a torrent of adverse publicity, the Liquidators sought to recover loans that had been made to Mr. Ali. He counterclaimed for damages for misrepresentation and breach of contract of employment. His case was that he had been disadvantaged and stigmatised on the labour market by reason of having been an employee of a bank the business of which had been conducted corruptly and dishonestly. The liquidators sought to rely on the ACAS agreement which was in the form of a general release. The House of Lords rejected a submission that special rules of construction apply to such an agreement and held that the normal rule of construction applies, namely the establishment of the intention of the parties, ascertained objectively in the context of the circumstances in which the release had been entered into and that, at the time, neither party could have realistically supposed that a claim for damages in respect of disadvantage and stigma was a possibility. Accordingly they could not have intended the release to apply to such a claim. The facts were not known and the legal potential had only become apparent after the decision of the House of Lords in Mahmud v. BCCI [1998] AC 20.

Lord Bingham of Cornhill stated that the general principles of construction summarised by Lord Hoffman in Investors Compensation Scheme Ltd v. WestBromwich Building Society [1998] 1 WLR 896, 912-913, apply. He added (at paras 9 and 10):

“A party may, at any rate in a compromise agreement supported by valuable consideration, agree to release claims or rights of which he is unaware and of which he could not be aware, even claims which could not on the facts known to the parties have been imagined, if appropriate language is used to make plain that that is his intention…

But a long and in my view salutary line of authority shows that, in the absence of clear language, the court will be very slow to infer that a party intended to surrender rights and claims of which he was unaware and could not have been aware.”

It is apparent from my somewhat meandering survey of the authorities that principles have been developed which may impact on the present appeal but that different principles have been articulated in different types of case, sometimes without cross-reference to each other. For example, KleinwortBenson, a seminal case on mistake of law, was not cited or referred to in Ali, which was decided on construction but perhaps also had the potential for consideration on the basis of a mistake of fact and law. Huddersfield Banking is helpful on the subject of the vitiation of compromises and consent orders but predates the step change in relation to mistake of law. What principles relevant to the resolution of the present appeal can be extracted from these rather diverse authorities? In my judgment, the following propositions emerge:

(1)

As with any other contracts, compromises or consent orders may be vitiated by a common mistake of law.

(2)

It is initially a question of construction as to whether the alleged mistake has that consequence.

(3)

Whilst a general release executed in a prospective or nascent dispute requires clear language to justify an influence of an intention to surrender rights of which the releasor was unaware and could not have been aware (Ali), different considerations arise in relation to the compromise of litigation which the parties have agreed to settle on a give-and-take basis (Huddersfield Banking)

(4)

For a common mistake of fact or law to vitiate a contract of any kind, it must render the performance of the contract impossible (The Great Peace).

Morland J’s conclusion that the compromise in the present case was vitiated by a common mistake of law was based on this analysis (at para 52):

“The common assumption of Mr. Scarles and Mr. Cohen was that the claim form had not been validly served, that mistaken belief arose from the judgment of the Recorder which was derived from McCombe J’s erroneous decision in Anderton’s case. That common mistaken assumption as to the law was the fundamental basis for and precondition of the compromise agreement, indeed its only springboard. Therefore I dismiss the appeal.”

Mr. Norman submits that Morland J and the Deputy Master were wrong to find that this is a case of a common mistake of law at all. He suggests that they assumed too much from the evidence of Mr. Scarles. In my judgment, there is force in this submission. Although the decision of McCombe J in Anderton undoubtedly presented a serious problem in relation to the validity of the service of the claim form in the present case, the view of leading counsel as relayed to Mr. Scarles was that McCombe J’s decision was incorrect but, to the extent that it was supported by May LJ in Godwin, it would be “extremely difficult” for Miss Brennan to obtain permission to appeal the Recorder’s decision and there would be “a small chance” of persuading the Court of Appeal to take a different view from that in Godwin. This, it seems to me, falls short of the unequivocal but mistaken view of the law which underlay Kleinwort Benson. As Lord Hope observed, the House of Lords was not dealing with the case where there is doubt as to the law – “a state of doubt is different from that of mistake” (p 410B). An appeal might have been correctly perceived as an uphill struggle but not as an inherently insuperable one – as subsequent events were to prove. I accept Mr Norman’s submission on this point.

The evidence does not disclose whether Mr. Scarles or anybody else enquired as to whether the decision of McCombe J was under appeal. His judgment had been handed down on 25 July 2001. The hearing of the appeal, and four consolidated appeals, commenced on 29 April 2002 – some ten weeks after the meeting between Mr. Scarles and Mr. Cohen at Croydon County Court. A simple enquiry would have disclosed the imminent appeals. It was knowledge that was readily obtainable. In the circumstances, it was incumbent on Mr. Scarles to obtain it. If he had done so, he could not prudently have agreed to the discontinuance of Miss Brennan’s action pending the outcome of the appeal in Anderton. By analogy with the approach taken by Bracewell J. in S v S, I would be reluctant to countenance as a mistake of law a situation in which it is generally known or ought to be known that the law in question is about to be reconsidered on appeal.

As I see it, the reality underlying this case is that, on behalf of his client, Mr. Scarles chose to discontinue rather than defend a further application by Islington of the kind that the two other defendants had made, having also chosen not to appeal the decision of the Recorder on those two applications. I appreciate that there may well have been funding difficulties but I do not consider that this analysis should be unduly inhibited by that. Although the agreement Mr. Scarles reached with Mr. Cohen was advantageous to Islington and not to Miss Brennan, I still see it as being in essence a give-and-take compromise of the kind referred to in Huddersfield Banking. Mr. Scarles was giving up the possibility of defending any application that Islington might make and of appealing any adverse decision. He was agreeing to finality forthwith, without awaiting the further consideration of the Court of Appeal in other cases about which he could and should have informed himself. He was securing in return an agreement on costs which was beneficial to his client and his firm (which had by then been subjected to wasted costs orders on the application of the other defendants). The consideration on each side was not of equal value but that does not necessarily mean that the contract of compromise was not a matter of give-and-take.

Another way of looking at all this is on the basis of risk in the context of impossibility of performance. As the Master of the Rolls said in The GreatPeace (at para 84):

“Once the court determines that unforeseen circumstances have, indeed, resulted in the contract being impossible of performance, it is next necessary to determine whether, on true construction of the contract, one or other party has undertaken responsibility for the subsistence of the assumed state of affairs. This is another way of asking whether one or other party has undertaken the risk that it may not prove possible to perform the contract, and the answer to this question may well be the same as the answer to the question of whether the impossibility of performance is attributable to the fault of one or other of the parties.”

In my judgment, this passage resonates in the present case for a number of reasons. First, this is quite simply not a case of impossibility of performance. The compromise has at all times remained performable, albeit to the disadvantage of Miss Brennan. It seems to me that, in the light of The GreatPeace, that is in itself sufficient to put it beyond the reach of common mistake of law. Secondly, although Mr. Bartle submits that it was for Islington to bargain for a term whereby the compromise would survive subsequent legal change (or, to be consistent with the declaratory theory of the common law which attracted the majority in Kleinwort Benson, further judicial consideration), I do not accept that that is the correct approach to the construction of compromises. Their essence in finality. There is a real difference between the situation where the compromise is agreed in ignorance of significant facts and the law which would be applicable to them (as in BCCI v. Ali) and the situation in which the compromise is agreed with no misapprehension of the facts at all (as in the present case), just an erroneous assumption about the law. This is not to reintroduce the distinction between mistake of fact and mistake of law. It is to require that, where a party wishes to reserve his rights in the event of subsequent judicial decision in a future case to which he is not a party, it is he who should seek and secure a term to that effect, not his opponent who should have to stipulate for protection notwithstanding the possibility of such a subsequent decision. Such a requirement is consistent with the policy of encouraging settlements and respecting their finality. I do not consider it to be in conflict with BCCI v. Ali. Thirdly, turning to what the Master of the Rolls said about fault, it is my view that, for the reasons I have given, in the present case the fault was on the side of Mr. Scarles who agreed the compromise without inquiry as to any appeal in Anderton, without appealing the decision of the Recorder and without defending an application on behalf of Islington. No fault lay at the door of Islington.

It follows from what I have said that I would allow Islington’s appeal on the grounds that the evidence does not establish a true mistake of law at all, more a state of doubt; that Miss Brennan’s solicitor was at fault in not inquiring about any appeal in Anderton; that the compromise was a matter of give-and-take which ought not lightly to be set aside; that it is not impossible to perform; and that, as a matter of construction, the risk of a future judicial decision affecting matters to Miss Brennan’s advantage was impliedly accepted and bargained away by her solicitor. In deciding the appeal on these grounds I have remained within the parameters of the submissions of counsel, Mr. Norman having disavowed a simple submission to the effect that a compromise in the course of litigation, entered into on professional advice, should never be vitiated by a subsequent judicial decision in a case to which the instant litigants are not parties, unless the compromise contains a suitable express provision. For my part I suspect that there is scope for a substantive exception to the ambit of mistake of law as a matter of policy in such circumstances and that it could live with what was said in Kleinwort Benson but, as we have heard no submissions about such an approach, this is not the case in which to say anything further about it.

In Kleinwort Benson Ltd v Lincoln Council [1999] 2AC 349, the House of Lords ruled that, in the law of restitution, there is no longer a distinction between payments made under a mistake of fact and payments made under a mistake of law.

In so deciding, the House of Lords overruled Bilbie -v- Lumley (1802) 2 East 9, the case which is taken to have first enunciated that distinction.

These various considerations more or less compel a conclusion that in the English law of contract the former distinction between mistakes of fact and mistakes of law no longer pertains. For a different approach to survive as between the law of restitution and the law of contract would seem illogical and difficult to justify.

So far so good. The difficulty arises, however, when there is added the ingredient of the ‘declaratory theory of judicial decisions’ whereby, when a judge states what the law is, his or her decision has a retrospective effect.

Once the position is (a) that a common mistake of law may vitiate a contract and (b) that the law may be changed retrospectively by judicial declaration of the law (conceptually creating a common mistake subsequent to the date of the contract, which was not a mistake judged according to the law as declared at the time of the contract) then an inevitable tension arises between, on the one hand, allowing the contract in question to be re-opened on the basis of the artifice of the common mistake of law and, on the other hand, adhering to the fundamental principle of contract law that parties should be held to their agreements.

The facts of this case, as summarised by Maurice Kay LJ. highlight this tension. The way in which I consider it falls to be resolved can best be set out under a number of headings.

In the context of restitution, the House of Lords in Kleinwort Benson clearly recognised the existence of certain defences as being available to a party responding to a case based upon an alleged common mistake of law. I refer to the passages already cited by Maurice Kay LJ from the speeches of Lord Goff of Chieveley and Lord Hope of Craighead, which refer to protecting “the stability of closed transactions” and recognising “the defences of compromise and settlement,” as well as the defence of “accord and satisfaction”.

Those passages alone serve to satisfy me for my part that the answer to the question “should the claimant’s claim against Islington succeed so as to release her from her agreement to discontinue the proceedings?” should be “no”.

I accept that all this is somewhat esoteric, probably far removed from the actual thought processes of lawyers (and more so of their lay clients) engaged in making contracts of compromise; but it is nevertheless utterly implicit in what a compromise is.

This sub-heading arises from a submission on behalf of the Claimant that this was not a “give and take” type of settlement (Huddersfield Banking Co Ltd-v- Henry Lister and Son [1895] 2 Ch 273). Rather it is said to have been a mere abandonment of the Claimant’s claim pursuant to her lawyer’s assessment (later retrospectively rendered erroneous) that her case was bound to fail by reason of what was believed at the time to have been late service. For this reason, it is submitted that the compromise is not caught by Huddersfield.

It is not in my view appropriate to seek to evaluate the adequacy of that consideration given by Islington, on the basis that “it did not give up very much”. The consideration which it did give was I think sufficient to create a “give and take” state of affairs, as per any compromise agreement properly so called.

It is for the Claimant to show that there was a common mistake, here of law (through the retrospective effect of the Court of Appeal’s in Anderton -v -Clwyd County Council reversing both Mr Justice McCombe’s decision in that case and Godwin –v- Swindon BC [2001] 4 All ER 641).

I agree with Maurice Kay LJ (leaving aside for the moment the retrospective effect of the Court of Appeal’s decision in Anderton) that there was at the time of the compromise here sufficient doubt as to the law as to service as to bring this case within the category of cases which Lord Hope excluded from recovery in restitution on the basis that a state of doubt is not a mistake.

In my judgment it cannot be right, given that the matter was compromised (notwithstanding that this was against the background of what has since emerged as a misunderstanding by the parties of the law on service) to allow the Claimant to reopen that compromise when there was doubt at the material time as to the proper construction of the CPR, yet the point was never tested as it might have been.

If one party to a contract can be said to have borne the risk that the parties might be mistaken in some respect, then there is no operative mistake. This is because the deal can be seen to have anticipated what actually occurred and to have provided for it.

As already stated, in contracts of compromise each party normally takes a risk that the law may change. Whichever party would be favoured by a subsequent retrospective change of law must therefore, so it seems to me, be taken to have assumed the risk that he or she would have done better by not compromising.

In such circumstances, there is no ground for seeking to re-open the agreement of compromise on the occurrence of the event in respect of which the risk was so apportioned.

What if any terms should be implied where the parties to a compromise do not expressly cover what is to happen if a subsequent declaration as to the law changes the law as it had appeared to be at the time of the contract?

By parity of reasoning with everything set out above, the answer in my view should be that the subsequent re-declaration of the law is by implication not intended to unpick that which was agreed. It follows that it must be for the party wishing to be allowed to reopen the compromise concerned to include in it an express terms to provide for such reopening, rather than for the other party thereto to do the opposite (i.e. to include a term to the effect that “this settlement is to stand even if the law should later change or be declared differently”).

Here, had the Claimant sought to include an express term for reopening in the event of a change of law about service, it is very difficult to envisage that Islington would have accepted such inclusion. There would have been no point. Islington would have not have achieved finality and it would have remained liable to a revival of the proceedings at any time in the future and as a result perhaps of some completely different proceedings.

I do not consider that my conclusion under this heading conflicts with the decision of the House of Lords in BCCI –v- Ali and others [2002] 1AC 251, where Lord Bingham of Cornhill spoke of the need for very clear language before a contracting party should be taken as having contracted out of rights which the parties could never have had in their contemplation. That case involved a novel cause of action arising many years after a contract of settlement and is to be contrasted with this case, where there was known to be doubt as to the reliability of the then state of the law as to service at the time the contract of compromise was entered into.

If one asks the question “would the Claimant have settled her claim if she had known that service of the proceedings on Islington had been good?” (as turned out to be the case) then the common sense answer is no. However, simple enquires would have revealed that Anderton was currently on its way to the Court of Appeal, but these were seemingly not made.

Furthermore, it is to be noted that the Claimant in fact sought to extricate herself from the agreement as early as 12 March 2002, whereas the Court of Appeal’s decision in Anderton was not handed down until 3 July 2002.

Islington, on the other hand, has done nothing to justify having the proceedings revived against it and wants nothing more than to uphold the compromise agreement entered into in good faith. In my view, the merits here lie more with Islington than with the claimant.

I acknowledge the irony that the two firms of solicitor defendants are now back in the Claimant’s original proceedings, by dint of judicial decision as referred to by Maurice Kay LJ, whereas (if this compromise stands) Islington will have got out of the action. However, this seems to me to be fortuitous and not as such a reason for treating Islington any differently than it would be treated if it had been the only defendant.

It seems to me that this case has demonstrated the considerable difficulties sometimes capable of arising by virtue of the declaratory theory of judicial decisions. If the theory did not apply, then those difficulties would evaporate.

Those difficulties, which I have called “practicalities”, include the problem of knowing just when the law has been definitively determined by judicial declaration. There is scope for the law to be declared at High Court, Court of Appeal and House of Lords level. I find it very difficult to see how contractual rights can potentially shift about according to which level of court the case concerned (or some other case on the same topic) has reached. This is the very opposite of the contractual certainty to which the law should aspire.

For these various reasons I agree with Maurice Kay LJ that this appeal should be allowed. So important is the principle of seeking to uphold contracts of compromise that in my view the court should not permit them to be reopened for mistake of law created by the retrospective impact of the declaratory theory of judicial decisions except where, for some truly exceptional reason, justice very clearly demands.

The degree of inconvenience, uncertainty and potential unfairness involved in reviving once-settled litigation is so obvious as not to need emphasis here. It leads me to agree with the concluding paragraph of Maurice Kay LJ’s judgment that, if this case had been argued on the basis of public policy, then that might well have provided a more simple answer to this appeal.

Because I share the concerns which animated Morland J, it is with some reluctance that I agree that this appeal succeeds.

The essential reason for her solicitor's agreement to withdraw Miss Brennan's claim against Islington was a straightforward mistake of law which was common to both solicitors: each believed on tenable grounds that the claim form had been served out of time. Both were proved wrong.

The element of give-and-take in the consequent compromise was marginal. It was that if Mr Scarles withdrew Miss Brennan's claim rather than force Islington to apply to dismiss it, Islington would seek no costs. But if, as both solicitors believed, the claim had been served out of time, it is unlikely that at the point of agreement Islington could have recovered any costs or were therefore forgoing anything by the compromise. Most of the giving was being done by Miss Brennan and most of the taking by Islington. But so long as there is some consideration, the courts are axiomatically not concerned with its adequacy.

I am also troubled by both the nature and the materiality of Mr Scarles' error. If the decision of the Court of Appeal reversing McCombe J in Anderton v Clwyd County Council had come out of the blue a year later in a different case, no criticism could be levelled at Mr Scarles for not anticipating it. The decision, by virtue of Kleinwort Benson, would relate back to the time of the compromise and would have the fictitious effect that both parties had misapprehended the law, when in practical reality they had got it right. Following Kleinwort Benson, and notwithstanding the remarks of Lord Hope (at 412) which Maurice Kay LJ has quoted, the present state of the law would seem capable of rendering the compromise, in that event, void.

Here, however, the news that Mr Scarles and his opponent had both misapprehended the law did not come out of the blue. The correctness of McCombe J's decision might have been tested in Miss Brennan's own proceedings (at the solicitor's risk as to costs, one would hope); and if not, it could at least have been ascertained that an appeal was pending against McCombe J's decision. The question is whether the failure to do either of these things has the character described by Steyn J in Associated Japanese Bank )International Ltd v Crédit du Nord SA [1989] 1 WLR 255 and approved by this court in TheGreat Peace (paras at 90-91), of rendering the mistake no more than "a belief which is entertained … without any reasonable grounds". I do not think it does. While it was arguably negligent to treat McCombe J's judgment as definitive without investigating the actuality or potentiality of a challenge to it, it was not unreasonable to accept leading counsel's advice that the decision, though probably wrong, would probably survive. There were, in other words, some reasonable grounds for Mr Scarles' belief that he had served Miss Brennan's claim out of time.

A further problem, in my view, lies in the formulation of the elements of common mistake set out in TheGreat Peace (at para 76). The fourth element is that "the non-existence of the state of affairs must render contractual performance impossible". Where the mistake is as to the existence of goods, or (as in The Great Peace ) as to the location of a vessel, this is straightforward. But what is the analogue in a case of mutual mistake of law?

The only kind of common mistake of law which is such that the true legal position renders performance of the contract impossible is, it seems to me, a mistake as to the legality of the contract. But Mr Norman, wisely as I think, has not sought to limit his concession to this class of mistake. He has accepted, as I would do, the broader proposition that a common mistake of law may vitiate an agreement by which litigation is compromised. The difficulty is in seeing how the effect of such a mistake can be equiparated with the impossibility of a contractual venture.

I think that in cases of mutual mistake of law a different test may be necessary. The equivalent question needs to be whether, had the parties appreciated that the law was what it is now known to be, there would still have been an intelligible basis for their agreement. This seems to me to come as close as one can come to what was identified as being at issue in The Great Peace (para 32) : a common mistaken assumption (in that case one of fact) which renders the service that will be provided if the contract is performed something different from the performance that the parties contemplated. It also echoes the question posed by Lord Atkin in Bell v Lever Bros [1932]AC 161 ,227. "Does the state of the new facts destroy the identity of the subject-matter as it was in the original state of facts?", if for "facts" one reads "law"

This, by itself, would permit the unravelling of a good many litigation compromises. But I see no principled way of distinguishing a compromise of extant litigation from a compromise of threatened litigation (which would presumably not attract the same policy considerations), nor a litigation compromise based on an error of law from one based on an error of fact. The generality of Mr Norman's concession, moreover, accommodated what Lord Bingham was shortly to say in BBCI v Ali, of which Maurice Kay LJ has cited the material paragraphs: that it requires clear and appropriate language to effect a surrender of rights of which the party concerned was not and could not have been aware. Thus there would be no difficulty in its becoming the practice for litigation compromises to spell out, if the parties are both willing, that the agreement which is being entered into is to be effective notwithstanding any future judgment by which the current understanding of the law may be changed. But that did not happen here.

But the overarching problem is the problem of public policy to which Lord Goff and Lord Hope drew attention in the passages of Kleinwort Benson cited (paragraph 14 above) by Maurice Kay LJ. We are concerned here with a compromise of extant litigation. What is to be done when anticipated or threatened or imminent litigation is compromised on the basis of a mutual mistake of law will have to await a case in which that is what has occurred. At that point the wisdom of Lord Goff's remark about the debatable nature of the scope of "settlement of an honest claim" will become very clear.

The possibility that extant litigation will be compromised on the basis of a mistake as to the current state of the law which is both mutual and non-negligent is not great. The typical case is going to be - as here - a shift in the understanding of the law subsequent to the compromise. It is not because mistakes of law have been added to mistakes of fact as grounds for undoing an agreement that this is now problematical. It is because the process is now required, by the majority decision of their Lordships' House, to incorporate the fiction that the law always was as it now (or at least for the time being) is. Were it not for this, the present problem would not have arisen. Instead we are forced to address it, as Lord Goff and Lord Hope anticipated we would be, by carving out a major exception almost at the birth of the new rule on what are essentially grounds of practicality and public policy - an exception which, for reasons I have mentioned, is necessarily fuzzy-edged and so destined to generate more uncertainty and more litigation.

But, like Maurice Kay LJ and Bodey J, I see no choice in the present case. The law must be taken to have been what it was only later declared to be, but the putative mistake created by this shift cannot be allowed to undo a compromise of litigation entered into in the knowledge both of how the law now stood and of the fact - for it is always a fact - that it might not remain so. While I am not happy about translating such knowledge into an implied term that the settlement is to stand notwithstanding any future change in the understanding of the law, I have less difficulty in recognising it as part of the matrix of fact in which a litigation compromise is cast.

On this footing, and without satisfaction, I agree that the appeal must be allowed.