I pass on to a second explanation which also seems to me inadequate. Lord Greene M.R. said at pp. 329-330, that because such notices have unilateral operation, the conditions under which they may be served must be strictly complied with. I have already said that this principle is accepted on both sides. But, as an explanation of the method of construction used in Hankey v. Clavering, it begs the question. If the clause had said that the notice had to be on blue paper, it would have been no good serving a notice on pink paper, however clear it might have been that the tenant wanted to terminate the lease. But the condition in clause 7(13) related solely to the meaning which the notice had to communicate to the landlord. If compliance had to be judged by applying the ordinary techniques for interpreting communications, there was strict compliance. The notice clearly and unambiguously communicated the required message. To say that compliance must be strict does not explain why some other technique of interpretation is being used or what it is.

      A variation of this explanation is to say that the language of the notice must be strictly construed. But what does it mean to say that a document must be "strictly" construed, as opposed to the normal process of ascertaining the intentions of the author? The expression does not explain itself. If it operates merely by way of intensification, so that the intention must be clear, unambiguous, incapable of misleading, then I think that the notice in this case satisfied the test at that level. Likewise, as Lord Greene M.R. acknowledged when he said that the whole thing was obviously a slip, did the notice in Hankey v. Clavering. So the concept of strict construction does not explain the decision.

      A more promising clue to the explanation is Lord Greene's statement, in two places, that the notice must "on its face" comply with the terms of the lease. What does "on its face" mean? Clearly, the face of the document is being contrasted with the background, in law sometimes called the "extrinsic evidence", against which the language is ordinarily construed. But Lord Greene cannot have meant that the document must always be read without any background, because (even if, which I doubt, it were conceptually possible to interpret the use of language without the aid of any background) cases like the Cox case, 4 Esp. 185 show that some background, at least, can be used. It appears, therefore, that Lord Greene is referring to some principle whereby background can be used to show that a person who speaks of the Waterman's Arms means the Bricklayer's Arms, but not that a person who speaks of 12 January means 13 January. What principle is this?

      It is, I think, to be found in an old rule about the admissibility of extrinsic evidence to construe legal documents. In its pure form, the rule was said to be that if the words of the document were capable of referring unambiguously to a person or thing, no extrinsic evidence was admissible to show that the author was using them to refer to something or someone else. An extreme example is in In the Goods of Peel (1870) L.R. 2 P. & D. 46, in which the testator appointed "Francis Courtnay Thorpe, of Hampton . . . Middlesex" to be his executor. There was a Francis Courtenay Thorpe of Hampton, Middlesex. He was however only 12 years old and his father Francis Corbet Thorpe, of Hampton, Middlesex, was an old friend of the testator. Lord Penzance said, at p. 47, that these facts were inadmissible: "The testator makes use of a description which applies in fact to one person, and not to any other." A variation on this rule was In re Fish [1894] 2 Ch. 83 in which the testator left his residuary estate to his "niece Eliza." He had no niece called Eliza but his wife had an illegitimate grandniece called Eliza, to whom the evidence of their relationship showed that he must have intended to refer, and also, as it happened, a legitimate grandniece called Eliza. The Court of Appeal said that the estate went to the legitimate grandniece and that evidence of the relationship between the testator and the illegitimate grandniece was inadmissible. Lindley L.J. said, at p. 85:

 ". . . where the person most nearly answering the description is the legitimate grandniece of the testator's wife . . . no evidence can be admitted to prove that her illegitimate grandniece was intended."

      On the other hand, if there was no one to whom the description accurately applied, there was said to be a "latent ambiguity" and evidence of background facts which showed what the testator must have meant, notwithstanding that he had used the wrong words, was admitted.

      Let us compare this rule with ordinary common sense interpretation of what people say. If someone has gone to great pains, well in advance, to secure tickets for himself and a friend for a Beethoven concert at the Royal Festival Hall by a famous visiting orchestra on January 13th and says to the friend a week earlier "I'll see you at the Festival Hall concert on January 12th," it will be obvious that he is referring to the concert on January 13th. According to the old rules of construction, the law will agree if there is no concert at the Festival Hall on January 12th. In that case there is a latent ambiguity. But if there is a concert on that date (Stockhausen, say, played by a different orchestra) he will be taken to have referred to that concert.

      This extraordinary rule of construction is, as it seems to me, the only explanation for the decisions in Hankey v. Clavering [1942] 2 K.B. 326 and Cadby v. Martinez,11 Ad. & El. 720. The Cox case, 4 Esp. 185, was distinguished by counsel in Cadby, at p. 723 as involving a latent ambiguity: there was no Waterman's Arms in Limehouse, so evidence that the landlord would have been understood by a reasonable tenant as intending to refer to the Bricklayer's Arms was admissible. But Midsummer 1837, or 21 December 1941 (in Hankey v. Clavering) or 12 January 1995 (in this case) are all real dates to which the notices could have referred. Therefore evidence of background which showed that a reasonable recipient would have understood the person giving the notice as having intended to refer to a different date had to be disregarded. The effect is that apart from the exceptional case in which the date is obviously impossible on the face of the notice (as in Carradine Properties Ltd. v. Aslam [1976] 1 W.L.R. 442) the intention which the notice would convey as to date has to be determined without regard to the terms of the lease (or anything else) as background. There is an artificial assumption that the reasonable recipient does not know what would be the correct date. On this basis, the interpretation of the notices as referring to the wrong dates and therefore being invalid is, of course, inescapable.

      It is clear that this rule of construction has been applied to the interpretation of notices for at least 200 years and it is hardly surprising that Lord Greene M.R. and Lord Clauson felt obliged to apply it in Hankey v. Clavering and that the Court of Appeal applied it in this case. It is, however, highly artificial and capable of producing results which offend against common sense. Lord Penzance began his decision that the testator had appointed a 12 year old boy as his executor by saying:

 "If I am at liberty to look at the facts stated in the affidavits, I may possibly have no difficulty in deciding that the person meant is the father, but the question is, whether I am at liberty to do so." (In the Goods of Peel (1870) L.R. 2 P. & D. 46).

      In In re Fish [1894] 2 Ch. 83, 84, Lindley L.J. began his judgment by saying: "This is one of those painful cases in which it is probable that the testator's intention will be defeated" and A.L. Smith L.J. said, at p. 86, that if he could have admitted the evidence about the testator's relationship with his wife's illegitimate grandniece he would gladly have done so.
 I think that the rule is not merely capricious but also, for reasons which I need not develop at length, incoherent. It is based upon an ancient fallacy which assumes that descriptions and proper names can somehow inherently refer to people or things. In fact, of course, words do not in themselves refer to anything; it is people who use words to refer to things. The word "allegory" does not mean a large scaly creature or anything like it, but it is absurd to conclude, as judges sometimes do, that this is not an "available meaning" of the word in the interpretation of what someone has said. This is simply a confusion of two different concepts; as we have seen, a person can use the word "allegory", successfully and unambiguously, to refer to such a creature.

      Even in its natural habitat, the construction of wills, the rule has not been (and, I think, cannot be) applied with any consistency. In National Society for the Prevention of Cruelty to Children v. Scottish National Society for the Prevention of Cruelty to Children [1915] A.C. 207 this House refused to accept that a gift to the "National Society for the Prevention of Cruelty to Children" should go to the society of that name, which had its head office in Leicester Square. It relied upon the background facts that, as Earl Loreburn said, at pp. 211-214, the testator was "a Scotsman living in Scotland" who had made a "Scotch will" to construe the will as intended to refer to the "Scottish National Society for the Prevention of Cruelty of Children". Earl Loreburn refused to accept that there was "a rigid rule" that "once a persona is accurately named in a will" there is not to be "any further inquiry or consideration in regard to the person who is to take the benefit." The true rule, he said, was that "the accurate use of a name in a will creates a strong presumption against any rival who is not the possessor of the name." This demotes the rule to the common sense proposition that in a formal document such as a will, one does not lightly accept that people have used the wrong words. I doubt whether anyone would dissent from this principle, which would present no obstacle to a conclusion that the tenant in this case must have used the wrong words.
      If your Lordships are to follow this path, it will be necessary to say that Hankey v. Clavering and the older cases which it followed are no longer good law. It would be wrong, I think, to distinguish them on narrow grounds and leave them as wrecks in the channel, causing uncertainty and litigation in the future. Furthermore, the old rule of construction has been applied not only to notices exercising break clauses but also to notices to terminate periodic tenancies: Doe d. Spicer v. Lea (1809) 11 East 312. In his admirable submissions on behalf of the landlord, Mr. Patten Q.C. warned that a departure from the old rule would cause great uncertainty in the daily construction of notices to quit in county courts throughout the land. I confess that this prospect has caused me some anxiety and I think that it must be given serious consideration.

      The rule as applied to wills, which restricts the use of background in aid of construction, reflects a distrust of the use of oral evidence to prove the background facts. The people who could give evidence about the background to a will would in most cases be members of the family interested in the outcome of the case and until 1843, persons with an interest in the litigation were not even competent witnesses. No doubt the exclusion of background makes, in a somewhat arbitrary way, for greater certainty in the sense that there is less room for dispute about what the background was and the effect which it has upon the intention to be attributed to the testator. But, as the cases mournfully show, this certainty is bought at the price of interpretations which everyone knows to be contrary to the meaning which he intended.

      There are documents in which the need for certainty is paramount and which admissible background is restricted to avoid the possibility that the same document may have different meanings for different people according to their knowledge of the background. Documents required by bankers' commercial credits fall within this category. Article 13(a) of the Uniform Customs and Practice for Commercial Credits (1993 revision) says (echoing Lord Greene M.R.'s phrase in Hankey v. Clavering) that the documents must "upon their face" appear to be in accordance with the terms and conditions of the credit. But the reasons of policy which require the restriction of background in this case do not apply to notices given pursuant to clauses in leases. In practice, the only relevant background will be, as in this case, the terms of the lease itself, which may show beyond any reasonable doubt what was the intention of the person who gave the notice. There will be no question of the parties not being privy to the same background--both of them will have the lease--and no room for dispute over what the relevant background is.

      In the case of commercial contracts, the restriction on the use of background has been quietly dropped. There are certain special kinds of evidence, such as previous negotiations and express declarations of intent, which for practical reasons which it is unnecessary to analyse, are inadmissible in aid of construction. They can be used only in an action for rectification. But apart from these exceptions, commercial contracts are construed in the light of all the background which could reasonably have been expected to have been available to the parties in order to ascertain what would objectively have been understood to be their intention: Prenn v. Simmonds [1971] 1 W.L.R. 1381, 1383. The fact that the words are capable of a literal application is no obstacle to evidence which demonstrates what a reasonable person with knowledge of the background would have understood the parties to mean, even if this compels one to say that they used the wrong words. In this area, we no longer confuse the meaning of words with the question of what meaning the use of the words was intended to convey. Why, therefore, should the rules for the construction of notices be different from those for the construction of contracts? There seems to me no answer to this question. All that can be said is that the rules for the construction of notices, like those for the construction of wills, have not yet caught up with the move to common sense interpretation of contracts which is marked by the speeches of Lord Wilberforce in Prenn v. Simmonds [1971] 1 W.L.R. 1381 and Reardon Smith Line Ltd. v. Yngvar Hansen-Tangen [1976] 1 W.L.R. 989. The question is therefore whether there is any reason not to bring the rules for notices up to date by overruling the old cases.

      There can, I think, be no question of anyone having acted in reliance on the principle of construction used in Hankey v. Clavering [1942] 2 K.B. 326. The consequence of such a construction is only to allow one party to take an unmeritorious advantage of another's verbal error, an adventitious bonus upon which no one could have relied. In this respect, the case for rejecting the old authorities is at least as strong as it was in Sudbrook Trading Estate Ltd. v. Eggleton [1983] 1 A.C. 444, in which this House overruled cases going back to the early nineteenth century on the construction of contracts for sale at a valuation.

      Nor do I think that a decision overruling the old cases will create uncertainty as to what the law is. In fact I think that the present law is uncertain and that only a decision of this House, either adopting or rejecting the Hankey v. Clavering rule of construction, will make it certain. So, for example, in Carradine Properties Ltd. v. Aslam [1976] 1 W.L.R. 442, 444, Goulding J. said that the test for the validity of a notice was: "Is the notice quite clear to a reasonable tenant reading it? Is it plain that he cannot be misled by it?" and he went on to say that the reasonable tenant must be taken to know the terms of the lease. This test was approved by the Court of Appeal in Germax Securities Ltd. v. Spiegal (1978) 37 P. & C.R. 204, 206 and, as will be apparent from what I have already said, I think that it was the right test to adopt. It is, however, absolutely impossible to reconcile the application of such a test with the decision in Hankey v. Clavering, in which no reasonable tenant who knew the terms of the lease could possibly have mistaken the landlord's meaning. It is therefore not surprising that in Micrografix v. Woking 8 Ltd. [1995] 2 E.G.L.R. 32, Jacob J. felt free to dismiss Hankey v. Clavering as "much distinguished" and to ignore it, or that Rattee J. in Garston v. Scottish Widows' Fund and Life Assurance Society [1996] 1 W.L.R. 834 should be puzzled as to why the Court of Appeal in this case considered, as I think rightly, that they were bound by Hankey v. Clavering.

      In my view, therefore, the House should say unequivocally that the test stated by Goulding J. in Carradine Properties Ltd. v. Aslam [1976] 1 W.L.R. 442 was right and that Hankey v. Clavering and the earlier cases should no longer be followed. The notice should be construed against the background of the terms of the lease. Interpreted in this way, the notice in the present case was valid and I would therefore allow the appeal.

LORD CLYDE

My Lords,

      The question in this appeal is whether the two letters dated 24 June 1994 and sent by the appellant tenant to the respondent landlord qualify as effective notices to determine the leases to which each letter respectively referred. Their validity as notices has to be tested against the terms of the power under which they were served. It is accepted that the two cases are for present purposes indistinguishable and that the relevant terms of the power were set out in clause 7(13) and were as follows:

 "The Tenant may by serving not less than six months' notice in writing on the Landlord or its solicitors such notice to expire on the third anniversary of the term commencement date determine this Lease and upon expiry of such notice this Lease shall cease and determine and have no further effect . . ."

      The substance of the power is expressed by the words "The Tenant may . . . determine this Lease". The method of its exercise is specified by the intervening words. The tenant must give six months' notice; the notice must be in writing; the notice must be served on the landlord or its solicitors. The sub-clause also states that the notice is to expire on the third anniversary of the term commencement date. The significance of that statement is that the period of six months is to terminate on that date. This regulates the time for the giving of the notice. The third anniversary marks the end of the period prior to which a notice under clause 7(13) must be given. But it is not required that the notice should include mention of the date of the intended determination of the lease. That date is prescribed by clause 7(13) where it states that the lease shall determine on expiry of the notice.

      Where a notice of termination complies precisely and unambiguously with the provision which empowers the sending of the notice then its validity should be unquestioned. Where the terms of the notice do not altogether accord with the provisions of the contract that may or may not render the notice unenforceable. The problem then may come to be one of finding a fair and reasonable construction of the notice. But there can be cases where the validity of the notice cannot be saved by any construction and will have to be regarded as bad.

      In some cases it may be obvious from the notice by itself that an error has been made. In Carradine Properties Ltd. v. Aslam [1976] 1 W.L.R. 442 an expressed intention to determine the lease at a date in 1973 was obviously incorrect in a notice served in 1974. In other cases the discrepancy can only be seen from a study of the terms of the lease. One would need to be aware of the provisions of the lease in such a case to appreciate that the permitted date was inaccurately stated. I see no reason in principle why in each of these kinds of case, provided of course that the wording is not absolutely clear and unambiguous, a notice should not be equally open to construction with a view to its possible validation.

      In the present case the two letters in my view satisfy the formal and technical requirements of clause 7(13). But they go further and call for a determination of each lease one day before the day which the sub-clause identified as the date for the determination of the notice and for the determination of the lease. As I have mentioned that was not a formal requisite of the notice. Each notice proclaims at the outset that it is given "Pursuant to Clause 7(13)". This was a precise reference to the particular provision under which the notices were each being sent, as distinct from some general reference to the agreement between the parties. But it is evident from a consideration of that clause that there is a discrepancy between the date there indicated for the termination of the lease and the date stated in the notice. Whether that inaccuracy in the notice is fatal or not depends on the proper construction of the notices. The formulation propounded by Goulding J. in Carradine (p. 444) was "Is the notice quite clear to a reasonable tenant reading it? Is it plain that he cannot be misled by it?" Delta Vale Properties Ltd. v. Mills [1990] 1 W.L.R. 445 concerned a vendor's notice to complete which was in condition 23 of the conditions of sale, but I see no reason why any different principle of construction should apply. Slade L.J. (at p. 454) observed:

 "In my judgment, notices to complete served under condition 23, if they are to be valid, must be sufficiently clear and unambiguous to leave a reasonable recipient in no reasonable doubt as to how and when they are intended to operate."

The standard of reference is that of the reasonable man exercising his common sense in the context and in the circumstances of the particular case. It is not an absolute clarity or an absolute absence of any possible ambiguity which is desiderated. To demand a perfect precision in matters which are not within the formal requirements of the relevant power would in my view impose an unduly high standard in the framing of notices such as those in issue here. While careless drafting is certainly to be discouraged the evident intention of a notice should not in matters of this kind be rejected in preference for a technical precision.

      The test is an objective one. In circumstances where an estoppel might arise the actual understanding of the recipient may be relevant, but in general the actual understanding of the parties is beside the point. That the test is an objective one was recognised in Micrografix v. Woking 8 Ltd. [1995] 2 E.G.L.R. 32. It was held there that the landlords would not have been misled by the references to a wrong date both in the notice to terminate the lease and in the covering letter. Each document was expressly written pursuant to the particular break clause in the lease. The recipients would have observed the errors because they would be familiar with the terms of the lease and would have known that the only date of determination had to be 23 June 1995. They would know that there was no requirement to specify any date in the notice. They would see that the tenant wanted to leave. It was held that the notice was valid.

      In my opinion a like view should be taken in the circumstances of the present case. The notices were expressed to be "Pursuant to Clause 7(13)". It is plain from that that the tenant intended to invoke that clause. It is also plain that the tenant wished to determine the tenancy and that clause is the only clause under which the tenant could achieve that result. The landlord would be expected to know the terms of the lease and the date on which the lease fell to be determined under that clause. He would also be expected to know that there was no formal requirement for the tenant to specify in the notice the date of termination of the lease. There was no evident reason why the tenant should specify 12 January rather than 13 January. The close proximity of the 13th makes it the more evident that it was erroneous and that the date intended was the date which the parties had agreed for a determination of the tenancy under clause 7(13). While there is a discrepancy evident in the notices between the reference to the clause and the statement of the date it seems to me that the notices were sufficiently clear and unambiguous. No reasonable landlord would in my view be misled by the statement of a date which in the context of a clear intention to invoke clause 7(13) was inaccurate. The landlord would in my view recognise that in each case the reference to 12 January was to be read as a reference to 13 January and I would so construe the notices.

      In Hankey v. Clavening [1942] 2 K.B. 326 the court refused to disregard a slip even although the intention of the notice was sufficiently clear from its terms and the recipient could not reasonably misunderstand it. In my view that was too strict and too technical an approach. Counsel for the tenant sought to restrict the decision in Hankey to circumstances where the insertion of the date of termination is an essential requirement of a notice. I am, however, not persuaded that the decision in Hankey rested on the understanding that the specification of the date was an essential. I note that the corresponding provisions in Carradine and in Micrografix were not dissimilar and indeed in the latter case one element in the decision was the consideration that there was no requirement to specify a date. While to a considerable extent the cases in this field may turn upon their own circumstances I do not consider that the decision reached in Hankey was sound and in my opinion it should be overruled. In the circumstances of the present case I take the view that the notices were valid and effective. I agree with your Lordships that the argument based on Sidebotham v. Holland [1895] 1 Q.B. 378 is without merit. But for the reasons which I have explained I would allow the appeal.