On 27^th February 2001 following a six month trial Lord Justice Rix handed down a judgment covering 655 paragraphs. He was asked to reconsider that judgment and on 21^st May 2001 he handed down a further judgment of 55 paragraphs refusing so to do, but granting permission to appeal certain aspects which raised points of law which the court of appeal would be able to deal with reasonably shortly. This court has granted permission to appeal certain other aspects, but directed that such issues as might be argued in a reasonably short compass, and which might dispose of the appeal as a whole should be heard first. It further directed that certain other matters should be dealt with at that first hearing in the hope of assisting the parties to deal with their appeals in a cost-effective way. This judgment deals with the matters in issue on that first hearing.

It is obvious that the litigation is complex, but the issues, which arise on this hearing, arise on the basis that the findings of fact of the judge are accepted. The background to the litigation can be taken directly from the judge’s first judgment. I simply append to this judgment that introduction and the three agreements with which this appeal is concerned referred to in that Introduction as appended to the judge’s judgment. For those unfamiliar with the litigation it will be necessary to read those documents at this stage. It will be seen that Rix LJ was trying as preliminary issues whether any of the claims being made in litigation between the three parties (Noga, FGN, and the SJ Berwin Defendants) had been settled, and if so which and on what terms by three “agreements” in writing dated 11^th, 13^th and 16^th August 1999.

Noga wished to establish that the document signed on 11^th August 1999 between itself, the SJ Berwin Defendants, and the FGN (the tripartite agreement) was binding. To establish that agreement as binding and indeed in order to recover the sum it suggested was due to it under that agreement, Noga had one major hurdle. In clauses 2 and 3 the agreement did not identify a sum which the SJ Berwin Defendants would pay Noga by way of settlement. The clauses simply referred to “a settlement amount” or “a settlement sum”. It was common ground that unless by parol evidence Noga could establish that as between the parties for “a settlement sum” and “a settlement amount” should be read $100 million the agreement would be void for uncertainty. A major part of the trial was concerned with the question whether in the settlement negotiations leading up to the signing of the tripartite agreement a sum of $100 million had been “agreed”. On this major factual issue the judge found in favour of Noga, but he found that the agreement of the figure was not such as to produce a binding contract containing that figure in the terms of the tripartite agreement.

The first issue on this appeal is whether on his findings of fact the judge was right to hold that there was no binding contract.

What followed the signing of the tripartite agreement were further negotiations between the SJ Berwin defendants and the FGN. That led to the signing of the 13^th August agreement. By the terms of that agreement, in return for the payment of DM 300 million the FGN waived all claims filed or contemplated against the SJ Berwin Defendants; that would include claims relating to the Ajaokuta bills of exchange and what were colloquially known as the “looting” claims. That agreement was however replaced by the 16^th August agreement, which in return for the payment of DM 300 million simply waived the Ajaokuta claims. The FGN relied on the 16^th August Agreement as being the only relevant agreement as between them and the SJ Berwin Defendants. The SJ Berwin Defendants argued that there was no consideration for their agreement to give up the benefit of the total waiver granted by the 13^th August 1991, and that thus the 16th August Agreement was unenforceable. The judge found that there was consideration for the 16^th August Agreement and the second issue before us is whether the judge was right in that conclusion.

The starting point has to be the terms of the document itself. Clause 1, 2 and 3 must be quoted because it is on their precise terms that the issue must be decided.

The judge concluded that if in place of the words “a settlement amount” in clause 2 and “a settlement sum” in clause 3 in each case a figure of $100 million had been inserted, on balance the agreement would have been certain and binding. At paragraph 586 he put his conclusion in these terms:-

He then also found, disregarding for this purpose any collateral stipulation relating to the agreement of $100 million itself, that there was no collateral representation, stipulation or agreement to upset his conclusion that by its terms and with $100 million actually stipulated as a figure in place of “a settlement amount” and “ a settlement figure”, the tripartite agreement was binding. [see para 587].

Obviously (as already indicated) without agreement of the quantum of the sum to be inserted in Clause 3 the tripartite agreement could not be complete and binding. As the judge recognised, without a binding agreement to pay $100 million, it is difficult to identify the consideration the SJ Berwin Defendants would have provided to bind them into the agreement. [see para 585].

The question accordingly is whether Noga established by parol evidence that when in Clause 2 and 3 the parties referred to “a settlement sum” and “a settlement amount” they can be taken to be referring simply to $100 million.

There is no issue between the parties that the law allows the use of parol evidence in order to identify the subject of a contract, or a common understanding as to what is meant by a word or phrase. Mr Gee QC cited to us many authorities for that proposition commencing with Maconald v Longbottom (1860) 1 E & E 977. But it is not the principle which is in issue in this case. It is its application. Each case depends on its own circumstances. In this case what has to be resolved is whether looking at the matter objectively it can be said that it was the intention of the parties that in the tripartite agreement the phrases “a settlement sum” or “a settlement amount” should be replaced by “$100 million”, and thus produce a binding agreement.

Mr Gee would suggest that the proper approach is simply to look at the document as it was originally drafted; establish from the parol evidence that $100 million was as between the SJ Berwin Defendants and Noga the settlement figure being offered to Noga; insert that figure in place of “a settlement sum” and “a settlement amount”; and then consider whether a binding contract was made.

The judge rejected that approach [see para 584 quoted below]. He as I read his judgment was concerned to see whether the parol evidence established that the parties were agreed that the figure of $100m should be inserted in place of “a settlement sum” or “a settlement amount”. May, for example the phrase “a settlement sum” have been used deliberately because there remained circumstances in which the SJ Berwin Defendants were not to be obliged to pay $100 million? Or possibly if Mr Bagudu had been forced to insert a figure would he have inserted that figure with conditions?

Once parol evidence is needed, it must be looked at as a whole. It could not be right for example to look at parol evidence in order to identify a subject or term if analysis of the totality of that evidence lead to the conclusion that there was still doubt as to what the subject or term was.

The judge examined the totality of the negotiations. The relevant parol evidence that the judge accepted was as follows. As regards the 10^th August:-

He also however concluded that the negotiations between Mr Schmidt and Mr Bagudu were critical. He found that $100 million had only been agreed between Mr Schmidt for Noga and Mr Bagudu for the SJ Berwin Defendants on condition that “FGN itself agreed to take nothing (other than the existing $50 million promise) for Ajaokuta, and lived by that promise……If the tripartite agreement was broken by the FGN, then Noga would not be paid.” He dealt with this in paragraph 138 as follows:-

The judges’ approach to the question whether the parol evidence was such as simply to lead to the $100 million being inserted into clause 2 and clause 3 appears from paragraph 584, 585 and 588 where he says:

Mr Gee QC is highly critical of the judge in a number of respects. He submits that the judge’s approach was to look to see whether there was some separate oral agreement as to the $100 million, and then see whether that agreement was subject to conditions. Having concluded that the agreement was subject to conditions, Mr Gee submits it was that conditionality which lead to the judge refusing to read “a settlement sum” as $100 million. He is also highly critical of the judge for being somewhat dismissive of the two face to face meetings where on the evidence Mr Gee would submit the judge has accepted that Mr Bagudu, Mr Gaon and Mr Schmidt have directed their minds to the tripartite agreement or its predecessor the MOU and to the clauses containing the very term “a settlement sum”, and identified $100 million as that sum.

Mr Gee submits that the judge did not direct his mind to what was the critical question. Whatever may have been the subject of negotiation right up until the moment the tripartite agreement was signed, when the parties actually signed what was the agreement they were actually making? Were they not intending to make a binding agreement that $100 million should be paid subject only to such conditions as were spelt out of the written document?

Mr Gee submits that the judge wrongly concentrates on what consensus Mr Schmidt may have reached with Mr Bagudu during the course of what were negotiations.

The judge’s conclusion in the result was as follows:

Mr Gee in his oral submissions before us reflected paragraphs 172 of the written submissions submitting that the judge had found a condition subsequent as a term of the tripartite agreement when no such term had been pleaded. In these submissions he was supported by Miss Selvaratnam QC who referred us to various authorities as to the proper approach of an appellate court where unpleaded issues were dealt with by a court at first instance. Mr Flint QC submitted that it was to misunderstand the judge's’ judgment and indeed the submissions of the SJ Berwin Defendants to suggest that there was any suggestion of the tripartite agreement containing a condition subsequent. The conditionality of any agreement to the settlement sum simply went to the enforceability of the tripartite agreement.

In relation to this aspect of Mr Gee’s criticisms, I would be sympathetic to Mr Gee’s complaint if it were a fair reading of the judge’s judgment that he was finding a binding agreement subject to a further term that the SJ Berwin Defendants obligation to pay would only arise if FGN actually withdrew their claims and demanded no money for Ajakouta, or if it was necessary for Mr Flint to establish some condition in addition to the terms of the written contract. But what the judge was examining was whether the tripartite agreement in accordance with its terms had become a final and binding agreement.

Reliance on Mr Schmidt’s evidence as to the conditionality of the agreement to the sum of $100 million, and its possible relevance to the question whether there was a final and binding agreement, had been well flagged in the SJ Berwin Defendants Closing submissions [see e.g. paragraph 267.3]. Furthermore it had been responded to by Mr Gee and Miss Selvaratnam without protest [see para 458-464 of their closing submissions]. It is thus reflected in Mr Gee’s submissions as recorded by the judge in para 579 of the judgment.

The case is not concerned with whether a further condition was agreed as part of the tripartite agreement. The question which the court has to resolve is whether there was a binding agreement at all.

I am also however a little sympathetic with Mr Gee’s criticism of the judge’s approach to this aspect which does seem to concentrate on the question whether some oral agreement had been reached between Mr Schmidt and Mr Bagudu, and on the conditionality of that agreement prior to the signing of the tripartite agreement; and on that being a reason why no binding agreement was reached by the signing of the tripartite agreement. It seems to me that Mr Gee must be right in saying that the critical question is - what must the parties be taken to be agreeing viewed objectively when they signed the tripartite document?

It is common ground that up until the moment the tripartite document was signed the parties were only in negotiation. This is reflected in Mr Gee’s written submissions at para 77 and was confirmed orally by him. Mr Flint QC would for obvious reasons not be arguing to the contrary. It follows that even though as between Noga and the SJ Berwin Defendants a figure of $100 million had been “agreed” during the negotiations, up until the moment of signature that figure could have been withdrawn, and a lesser sum offered. So for example if it were the SJ Berwin defendants intentions to pay only if the FGN did not seek money, and if the FGN had sought money from the SJ Berwin defendants, Mr Gee would accept that until the signing of the tripartite agreement on 11^th August, Noga could have had no complaint if the SJ Berwin defendants lowered the settlement offer. His case is that the tripartite agreement superseded negotiations.

In simple terms he submits that the document of 11^th August was produced in an attempt to conclude a final settlement; the document must then be construed and in its construction the parol evidence allows $100 million to be inserted in clause 2 and clause 3; signatures on the document demonstrate agreement to the terms of that document with the sum written in and such conditions as are to apply to SJ Berwin Defendants’ obligation to pay must be ascertained from the document. Such conditions in the written document included FGN withdrawing its claim against the defendants, and whatever that means that is the condition agreed as between the parties in the tripartite agreement.

The argument proceeds on the basis, correctly in my view, that the moment of time on which one must concentrate is the moment of signing of the tripartite agreement. It also correctly asks what it is objectively that the parties must be taken to be agreeing. Mr Gee on this basis submits that, even if it can be taken as established that in the negotiations Mr Schmidt and Mr Bagudu were at one that $100 million should only be payable if the FGN agreed to forego their claims against the SJ Berwin Defendants but also did not act in breach of that term by demanding money from the SJ Berwin Defendants, since at the conclusion of the negotiations a written document was produced resulting in the final signed document, it is that document properly construed which is intended to reflect the agreement of all three parties including FGN. “Properly construed” on Mr Gee’s submission brings in the parol evidence rule entitling the parties to identify what they meant by “a settlement sum”. He points to the fact that the term “a settlement amount” or “a settlement sum” was used in a previous document produced during the final negotiations containing identical terms to the tripartite agreement. He submits that the judge found that when the parties were face to face they identified “a settlement sum” as $100 million. The document on Mr Gee’s submission must be read as if that figure were inserted in place of those terms. He then relies on paragraph 586 of the judgment and says that a binding agreement results. He adds by way of alternative, that on the very day on which the document was signed, again as the judge he says found, Mr Bagudu for the SJ Berwin Defendants and Mr Gaon for Noga identified “a settlement sum” as $100 million. Indeed later the same evening he submits the judge found they identified “a settlement sum” as $100 million.. So he says the result is a binding agreement.

As I have previously said, it is not in question that parol evidence could identify the subject of a contract or what is meant by a term. But Mr Gee’s argument depends on it being clear that the parol evidence demonstrated that it was the intention of the parties that in place of the phrases “a settlement sum” or “a settlement amount” in clauses 2 and 3 the sum of $100 million was to be inserted so as to produce a binding contract. I stress this last because all the cases on identifying the subject of a contract by parol evidence to which we were referred were cases in which it can be seen that the parties clearly intended to conclude a binding contract. There is a danger of using the finding of the judge in para 586 that with the figure inserted he would on balance hold the document final and binding, as support for the conclusion that this is a case where clearly the parties intended to create a binding contract. Of course some of the points made by the judge in paragraph 586 can themselves be used to support an argument that such was the parties’ intention. Furthermore as Mr Gee emphasised there is the principle exemplified by such authorities as Hillas v Arcos (1932) 43 LL Rep 359 that where parties make a contract it should so far as possible be construed as valid. But why was the phrase “a settlement sum” or “a settlement amount” used? The fact is that the figure of $100 million was not set out in either of the clauses, and the clauses refer to “a settlement amount” and to “a settlement sum”. Adopting Mr Gee’s submission, which I accept, that what one is searching for is what the parties were intending to agree when they signed the tripartite document, there is an alternative to his analysis.

Before going to the parol evidence it seems to me the court should construe the document with the words in fact used by reference to its context in order to see whether there is any pointer to the need for parole evidence to exchange one comprehensible phrase for some other comprehensible but different phrase. The context was that FGN had no idea what sum the SJ Berwin Defendants were contemplating paying Noga, and in the negotiations the SJ Berwin Defendants had certainly been suggesting that the sum which they would agree to pay Noga would depend on the sum that FGN sought from the SJ Berwin defendants. The terms of the tripartite agreement are capable of the construction that FGN as Mr Gee suggested at one point in his oral submissions were simply agreeing to withdraw their claims against the SJ Berwin Defendants, and not agreeing not to demand any money from those defendants. Indeed if in addition to the uncertainty of the words “a settlement sum”, there is added the notion that FGN are not binding themselves not to demand money, the likelihood of the Agreement being meant to be the final legally binding settlement as opposed to an important step in continuing negotiations diminishes. Already I suggest it is questionable whether it was not deliberate that the phrase “a settlement sum” or “a settlement amount” was being used because of the contemplation that FGN might demand money, and the contemplation that whatever sum had been offered to Noga in the negotiations so far might not remain the sum if FGN did demand money.

One then turns to the parol evidence to examine whether there is a certainty that it was the intention that $100 million simpliciter should appear in place of the two phrases. In a case where the parol evidence arises from the negotiations which have resulted in the agreement at issue, the totality of the negotiations must be relevant. The totality of the negotiations show first on my reading of the judge’s judgment that he was not sure that (contrary to Mr Gee’s submission to us) the parties’ agreement not to refer to any figure was simply to keep that figure out of the public domain. In para 569 there is a reference to the possibility of keeping the figure out of the public domain but the judge also said in the last sentence of that paragraph:-

There is then the letter from Noga and the SJ Berwin Defendants requested by the Attorney General as confirmation that they had as between themselves settled. It was in these terms:-

Is it possible that, (just as Mr Gee accepted that the $100 million was a sum which up at least until 11^th August was being offered in negotiations, and was therefore withdrawable at any time), the intention was that the sum should not be referred to in the tripartite agreement or at least not referred to unconditionally because it was not the final sum in the eventuality that FGN demanded money? Is it not possible that all that was being identified as the settlement sum in the face to face negotiations was the sum which the SJ Berwin Defendants had offered on the eventuality that FGN would not demand more?That was certainly so as at 10^th August 1999, and there is no reason why it should not be so after the signing of the tripartite document if all that document is, is a continuation of, albeit an important continuation of, negotiations albeit with a better prospect of finalisation.

As I have said I think there is something in the criticism of Mr Gee of the judge’s approach to this aspect. The concentration should have been on the moment when the tripartite document was signed and consideration should have been given as to whether the draft should have been read so as to contain $100 million rather than the phrases “a settlement sum” or “a settlement amount”. Consideration should have been given to whether if that was so, the effect of the signed document was not to override any previous indications of conditionality. But it seems to me the judge was asking the correct question. Have Noga established that the parties intended £100 million simpliciter to be inserted in place of “a settlement sum” and “a settlement amount”? He concluded they had not and in my view he was right on his findings so to conclude.

This was a tripartite arrangement. It must be rare that parol evidence can be introduced to identify what is meant by a term as between only two of the parties to such an arrangement. The SJ Berwin Defendants and Noga had got very close to settling provided FGN would also settle without demanding money. They had not in fact settled as between themselves without FGN also being bound. The tripartite agreement did not prevent (as the facts showed) FGN demanding some payment. The likelihood is that the tripartite agreement was not intended to be final and binding and indeed that the phrase “a settlement sum” was used because in certain circumstances the SJ Berwin Defendants were not to be bound to pay $100 million.

I would dismiss the appeal on this aspect.

On the second issue as to whether there was consideration for the 16^th August Agreement, for the reasons given in the judgment of Tuckey LJ I would also dismiss the appeal.

I agree with Waller L.J. that Noga’s appeal should be dismissed.

If at the time they signed the tripartite agreement the parties had been asked: what settlement sum has been agreed? FGN would have said they did not know and, on the judge’s findings, Noga and the SJ Berwin Defendants would have said $100 million provided FGN do not ask more for Ajaokuta. In other words there was a string attached to the agreement of the settlement sum. As FGN did ask for more almost immediately after the tripartite agreement was signed, Noga can hardly complain that they have been denied the right to recover $100 million on a basis which ignores the string.

But the question on the appeal is whether this result withstands legal analysis. One obvious solution would be to say that the string was a condition subsequent to the agreement. But this was not pleaded, the judge did not so decide and Noga would have a justifiable complaint of unfairness if the case was now decided against them on this basis. The judge’s solution was to say that the string’s conditionality meant that the parties did not intend the tripartite agreement to be final and binding. I agree with Waller L.J. for the reasons he gives that the judge was justified in reaching this conclusion on the facts which he found. The references to “a” settlement sum/amount were deliberately inserted into the agreement because no final and binding agreement had yet been made.

Laws L.J.’s judgment, which I have read with admiration, shows that the situation can be analysed differently, although I think it produces the wrong result. His lexicon approach enables him to identify the amount, but ignore the string attached to it. I see no reason why the lexicon should not be capable of supplying a definition which includes the string, if that is what the parties have agreed. The fact they have done so in negotiation is not to the point. I see no reason why the string should become detached at the point when the parties sign the agreement. Laws L.J’s approach does not work anyway if the references to “a” settlement sum/amount were deliberately inserted into the agreement because no final and binding agreement had yet been made. The fact that the agreement was tripartite is a complication, but if it was not binding on Noga and the SJ Berwin Defendants it was not binding on FGN either whatever they may have thought. Their subsequent conduct however suggests that they did not think it was binding on them either.

There were two appeals and a cross-appeal raising issues about consideration but in the event we only heard the SJ Berwin defendants’ appeal against the judge’s conclusion that there was consideration for the agreement of 16 August. At the end of the argument we said that this appeal would be dismissed. My reasons for doing so follow.

The simple point taken by the SJ Berwin defendants at trial was that the 16 August agreement was no more than an entirely one-sided variation of the 13th August agreement by means of which they were deprived of their global release for which the FGN gave nothing. They relied on the leading case of Stilk v Myrick (1809) 2 Camp. 317 which established the principle that a promise to perform an existing contractual obligation does not constitute good consideration for a new agreement.

Relying on the decision of this court in Williams v Roffey [1991] 1 QB 1 the FGN submitted that Mr Bagudu had obtained a practical benefit from the 16 August agreement but the judge rejected this submission. His decision on this point was the subject of FGN’s cross appeal which they did not pursue after we announced our decision on the SJ Berwin defendants’ appeal, although they reserved their right to do so if the latter decision was reversed.

However FGN’s primary argument was that the principle in Stilk v Myrick has no application where the earlier agreement has been contractually rescinded. In such a case consideration was supplied by the mutual release of the executory promises under the earlier agreement as well as the provision of fresh consideration in the later agreement.

The judge considered the difference between variation and rescission by reference to a number of cases under the Statute of Frauds and concluded:

On the appeal Mr Flint accepted the judge’s conclusion that the earlier agreement was rescinded and replaced by the later agreement. It is not therefore necessary to examine further the judge’s reasons for reaching this conclusion.

But Mr Flint submits that one must look at the substance and not the form. The substance is that the FGN’s obligations remained the same under both agreements. The SJ Berwin defendants got no benefit under the later agreement by which they had to restrict the settlement to the Ajaokuta claims. If this had been done by varying the earlier agreement there would have been no consideration for the variation. It should not matter that it was done by rescission and replacement. He accepted however that there would be consideration if there was a real interval between the two. But here, as the rescission and replacement took place simultaneously or within a fictional scintilla temporis, it could not and should not provide consideration.

Mr Flint emphasised that although the principle in Stilk v Myrick has been criticised, it is still good law and has been confirmed by this court in D&C Builders Ltd. v Rees [1966] 2 QB 617 and in Williams v Roffey. The law will not enforce gratuitous promises. Foakes v Beer (1883/4) LR App Cas 605, which decided that there is no consideration for a promise to accept less than the full amount of a debt, is another example of this principle at work. There was in fact, Mr Flint submitted, no authority which clearly established the rescission and replacement theory of consideration.

Chitty (28^th edition) at paragraph 3-072 says:

The case cited in support of the first sentence of the text is Stead v Dawber (1839) 10 Ad & E 57. This was a Statute of Frauds case in which the question was whether an oral agreement to extend the delivery date in a contract of sale was merely a waiver of the contract term or a variation of it which had to be in writing. The court held that the extension of time was a variation and so the contract could not be enforced. But its analysis was that the parties intended to substitute a new contract for the old one, the same in all respects except for the date of delivery. Of this Lord Denman CJ said:

Mr Flint referred us to Harris v Carter (1854) 3 E & B 559 in which a sailor who had signed articles for a round trip to Melbourne was induced to make the home voyage after some of the crew deserted by fresh articles which promised increased pay. On the jury’s verdict that the plaintiff had not been discharged from his obligations under the original articles before entering into the fresh articles, his claim was non-suited because there was no consideration for the promise in the fresh articles. Mr Flint suggests this case casts doubt on the rescission and replacement theory. I do not agree for the simple reason that the jury’s finding meant that there was no rescission: the obligations under the original articles continued. There is in fact some support for the theory because the court suggested that if the plaintiff had been relieved from his obligations under the original articles he could have entered into a fresh contract for which there would have been consideration.

The only other case I need mention is a decision of the Supreme Court of New Hampshire in Watkins & Sons Inc. v Carrig (1941) 21 A. 2d 591. In that case builders had entered into a fixed price contract to excavate a cellar. When they encountered solid rock the building owner agreed to pay a higher price for the work. It was argued that there was no consideration for the later agreement. The trial judge found that the later agreement superseded the earlier one. In dismissing the building owner’s appeal the court said:

Mr Flint submits that we should not follow this decision. He refers us to The Restatement Contract 2d which at section 89 says:

No other authority has been cited to us which sheds light on the question we have to decide. Certainly there is no authority which compels a decision one way or another. It is helpful therefore to analyse the situation by reference to well-known general principles.

The essential difference between rescission and variation for present purposes is that a contract comes to an end when it is rescinded but continues if it is varied. If the rescinded agreement is replaced by a new agreement containing the same obligations, it is not the old agreement which compels the performance of those obligations but the new agreement. It follows that the principle in Stilk v Myrick has no application to this situation because it is premised on the continuation of the obligations in the old agreement. Mr Flint accepted this analysis in a case where there was an interval between the rescission and replacement, but I do not see that there can be any difference in principle between the two situations.

Consideration is however necessary for the agreement to rescind the old contract. There is no difficulty about this in the present case where at the time of rescission the agreement of 13 August was executory on both sides. Consideration is provided by the mutual release of those executory promises.

Consideration for the agreement of 16 August was provided not only by the mutual promises in clauses 1-5 of that agreement but by the mutual release from the earlier agreement. The later agreement expressly provides that it supercedes the earlier agreement.

This analysis leaves only Mr Flint’s point about timing which was really his main point. The problem with his submission is that it assumes that the obligations under the old agreement continue whereas the parties have expressly agreed to tear that agreement up. That is not appealing to the substance of the transaction, as Mr Flint submits, but creating a fiction which the parties have expressly disavowed. It is not necessary in my judgment to create a scintilla temporis for there to be a rescission and replacement. It can be achieved concurrently by the same document in the way it was done in this case. That is what the parties in this case intended when Mr Bagudu signed the agreement of 16 August in the presence and on the advice of his two Nigerian lawyers.

So for these reasons I reject Mr Flint’s submissions and agree with the judge’s decision that there was consideration for the agreement of 16 August.

I adopt with gratitude the account given by my Lord Waller LJ, whose judgment I have seen in draft, of the background to this appeal. I have, however, found myself driven to disagree with his conclusion that the document signed on 11^th August 1999 (“the tripartite agreement”) did not in the events which happened constitute a binding contract.

First, with respect I agree entirely with my Lord (paragraph 6) that the starting point has to be the terms of the document itself. So approaching the matter, it is to my mind beyond argument that the document has all the characteristics of a binding contract save only that nothing in its words identifies what is meant by the expressions (respectively appearing in paragraphs 2 and 3) “a settlement amount” and “a settlement sum”. It is I understand common ground, and I would certainly hold, that unless the connotation of those expressions can legitimately be supplied by evidence external to the document itself, the tripartite agreement (assuming it were otherwise binding) would fail as being void for uncertainty. But I think it no less plain that if their connotation can be so supplied, then the tripartite agreement took effect according to its terms as a binding contract unless there were evidence, accepted by the court, that at the moment of signing the document, despite its unequivocal language the parties to it were in fact agreed that they should not then and there be bound by it. It is of course elementary that all the parties must so agree if a contract, good and enforceable on its face and involving mutual promises between them, is not after all to be binding. One party, or some but not all, cannot as it were cross their fingers behind their backs as they sign: see the classic statement of Blackburn J in Smith v Hughes (1871) LR 6 QB 597, 607.

In my judgment it is of the first importance to keep separate these two quite different questions: (1) is there evidence to supply the connotation of “a settlement amount/sum”? and (2) is there evidence to prove that at the moment of signing the tripartite agreement the parties were in fact agreed that they should not be bound by it? I should say that while, as I have just indicated, affirmative evidence on question 2 would have to demonstrate agreement by all parties not to be bound, by contrast I see no reason, having regard to the particular shape of the tripartite agreement, why the connotation of the “settlement sum” (question 1) should not be supplied by proof of an understanding shared only by Noga and the S J Berwin defendants, to which the FGN were strangers.

Given the importance of keeping these two questions separate, as it seems to me the reasoning of Rix LJ at paragraph 584 of his first and principal judgment demands particular scrutiny. I will set out the paragraph in full:

There is no reason why the lexicon that gives $100 million for the settlement sum must also constitute the binding contract between the parties. Generally, there is no reason why the connotation of an uncertain word or phrase in a contract which, but for the uncertainty, would plainly be binding cannot be supplied by evidence of what was said at a stage when the parties were still in negotiation or had arrived at no more than a partial or conditional agreement. Indeed, reason demonstrates that it must be possible to supply the connotation of such a word or phrase by evidence from such a source. If it were not, the consequence would be that a written contract, by its language for all the world binding and enforceable save only for the vagueness of a particular expression, is not in principle a contract at all: the true contract between the parties, if there is one, has to be found in some prior meeting of minds. I know of no authority for this proposition.

In my judgment, the fact – if fact it be: I leave aside for present purposes Mr Gee’s reliance on what he called the “face-to-face” meetings when, it is said, the sum of $100 million was referred to or identified without qualification – that the context in which the answer was given to question 1 (is there evidence to supply the connotation of “a settlement amount/sum”?) was one of negotiation, does not of itself begin to supply an answer to question 2 (is there evidence to prove that at the moment of signing the tripartite agreement the parties were in fact agreed that they should not be bound by it?). If at the moment of signing the parties were in truth still in negotiation only, one would plainly have expected the document to reflect the fact. Instead, as I have said (and as the judge accepted: paragraph 586, set out by Waller LJ at paragraph 7), the document bears every appearance of a concluded contract save only for the want of connotation of “a settlement sum”. Once that connotation is supplied, the appearance becomes the legal reality. Its being supplied out of a context of negotiation is of no significance because the words of the tripartite agreement demonstrate that negotiation is past.

Waller LJ (paragraph 35) concludes that the trial judge asked himself the right question, viz. whether Noga had established that the parties intended $100 million simpliciter to be inserted in place of “a settlement sum” and “a settlement amount”, and rightly answered the question in the negative. It is at this point, with great deference, that I part company from my Lord. The possibility that “settlement sum/amount” referred only to a conditional agreement for $100 million depends in my judgment on a false premise, namely that the absence within the document of a concrete connotation for “settlement sum/amount” meant that it was an open question whether the tripartite agreement, given the whole of its language, was intended to constitute a binding contract. In my judgment, that was not an open question. The document’s words are the words of a binding contract; they are no less so because they include an expression whose connotation must be supplied by extrinsic evidence. If it was not to be binding, there would have to be hard evidence that at the moment of signing all three parties intended not to be bound. The fact of earlier negotiations, being the context in which the connotation of “settlement sum/amount” was found, is incapable of supplying or amounting to such evidence. Nor, in my judgment, is there anything in the byzantine events following the signing of the tripartite agreement which is capable of constituting such evidence.

For my part, I would allow the appeal on this aspect.

For the reasons given by Tuckey LJ I would dismiss the appeal on this aspect.