1 This appeal raises some questions as to the proper limits of contractual interpretation. The parties collaborated to assist a Chinese company in its effort to get listed on the Frankfurt Stock Exchange. They agreed to expend efforts and to bear the expenses associated with the listing in exchange for share capital that would be issued in the listed entity. The company was listed and a quantity of shares was issued, a portion of which was registered in the names of various companies owned by the Appellant. A part of the shares registered in the names of the Appellant’s companies was to be transferred to the Respondent pursuant to a share allotment agreement (“Allotment Agreement”). The main issue before us centres on how the Allotment Agreement is to be construed.

2 Both parties put forward flatly inconsistent accounts of what happened in the lead-up to the listing and, on that basis, advanced vastly different constructions of the Allotment Agreement. On the Respondent’s construction, she was entitled to more shares than she had thus far received; under the Appellant’s construction, the Respondent had already received all the shares she was entitled to. The High Court judge (“the Judge”) saw problems with both parties’ accounts but held that the Respondent’s was, on the whole, to be preferred. On that basis, she accepted the Respondent’s interpretation of the key term in the Allotment Agreement and ruled in her favour for damages representing the value of the untransferred shares. The Appellant counterclaimed for some unpaid expenses. The Judge allowed the counterclaim, but only in part as she found in favour of the Respondent in relation to certain disputes of fact. Her written judgment was published as Ding Pei Zhen v Yap Son On [2015] SGHC 246 (“the Judgment”). Dissatisfied, the Appellant challenges the Judge’s construction of the Allotment Agreement as well as her findings on the counterclaim.

3 In this appeal, we take the opportunity to underscore the importance of conceptual clarity, evidentiary discipline, and procedural rigour in the process of contractual interpretation. In Sembcorp Marine Ltd v PPL Holdings Pte Ltd and another and another appeal [2013] 4 SLR 193 (“Sembcorp Marine”), we warned that the modern contextual approach towards contractual interpretation could lead to greater uncertainty and a needless increase in the cost of litigation if parties treated the contextual approach as giving a “licence to admit all manner of extrinsic evidence” (at [72]). This appeal is a case in point. The Allotment Agreement is an agreement in Mandarin of some 8 handwritten lines at the bottom of a page in a company prospectus. It could not have taken long to draft, but the parties spent eight days on a wide-ranging inquiry into the minutiae of the parties’ commercial dealings in an attempt to construe a single term of the Allotment Agreement. The result, however, was less, rather than more, clarity.

4 After careful consideration of the parties’ arguments, we allow the appeal in relation to the main claim in full and the appeal in relation to the counterclaim in part. We now give our reasons, beginning with a more detailed recitation of the relevant facts.

5 The background facts are largely undisputed and were summarised with clarity by the Judge at [5]–[17] of the Judgment. We will only reproduce those aspects which are germane to this appeal. The Appellant, Yap Son On, is a Malaysian businessman who resides in Singapore while the Respondent, Ding Pei Zhen, is a Chinese businesswoman who resides in the People’s Republic of China. They were business partners who agreed to work together with one Mr Xie, a business associate of the Respondent, to procure the listing of some Chinese companies on foreign bourses. Under this arrangement, the present parties would bear the expenses of the listing in return for share capital in the listed entities.

6 In 2010, the Appellant became acquainted with one Mr Li Wenwen (“Mr Li”), who was the owner of Jinjiang Goldrooster Sports Goods Co Ltd (“Goldrooster Jinjiang”), a sports apparel company operating in the People’s Republic of China. In July 2010, Mr Li contracted with One Capital Group Investment Limited (“One Capital”) – a company solely owned and controlled by the Appellant – to engage the Appellant as a consultant in the intended foreign listing of Goldrooster Jinjiang. This contract was known as the “Listing Agreement” and it provided that One Capital would be paid, among other things, 12% of “[Goldrooster Jinjiang’s] shares of the after-listing total capital.” Mr Li and the Appellant agreed that the Appellant would bear all expenses associated with the listing (“the Listing Expenses”). Separately, the Respondent and Appellant agreed that the Listing Expenses would be apportioned between them on a 60:40 basis and that this would reflect their entitlement to the shares they would eventually acquire in the Goldrooster listed entity (that is to say, the Respondent would be entitled to 60% of those shares while the Appellant would be entitled to 40%).

7 It was eventually decided that Goldrooster Jinjiang would be listed on the Frankfurt Stock Exchange through a listing vehicle incorporated in Germany (“Goldrooster AG”). On completion of all preparatory work for the listing, Goldrooster AG had an issued share capital of 20 million par value ordinary bearer shares. These shares were held by the following four companies, all of which were incorporated in the British Virgin Islands:

(a) Zhuo Wei Investments Limited (“Zhuo Wei”), which held 14.5m shares;

(b) Season Market Limited (“Season”), which held 3,005,000 shares;

(c) Xanti Investments Limited (“Xanti”), which held 1,247,500 shares; and

(d) Fortune United Investment Limited (“Fortune”), which held 1,247,500 shares.

Mr Li was the ultimate owner of Zhuo Wei whereas the Appellant was the ultimate owner of the remaining three companies. The three companies owned by the Appellant were referred to collectively as the “Yap Companies”. At this point, the Yap companies held a total of 5.5m shares in Goldrooster AG, which represented 27.5% of the issued share capital.

8 On 18 May 2012, Goldrooster AG was successfully listed and an initial public offering (“IPO”) of 5m shares was made of which 720,026 shares were subscribed for. This took the total number of shares in Goldrooster AG to 20,720,206 (“the post-listing shares”). Goldrooster AG’s listing prospectus (“the Goldrooster Prospectus”) is of particular significance to this dispute and we reproduce the relevant section of it below (we have added the shaded row to set out the percentage shareholding of the Yap Companies in each of the different scenarios listed in the prospectus):

9 Two points should be noted. The first concerns the third column of this table. If all the 5m shares made available in the IPO had been sold, Goldrooster AG would have had a total issued share capital of 25m shares (“estimated post-listing shares”) of which 22% would been owned by the Yap Companies. The second concerns the so-called “Greenshoe Option” which was referred to in the third and fourth columns of the table. This was an overallotment provision in the IPO which would be triggered if Goldrooster AG’s shares were oversubscribed. In such a scenario, the Yap companies would have to yield up to 750,000 shares of the 5.5m shares which they held at that time (3% of a projected maximum share capital of 25m shares) for public subscription, taking their shareholding in Goldrooster AG down to 4.75m shares, which would be 19% of the total. As will become clear in the course of this judgment, the figures of 22% and 19% are of great significance to this dispute. In the event, the Greenshoe Option was not exercised due to the low demand for the shares.

10 After the listing, the parties could not agree on how their respective shareholdings were to be distributed. On 15 June 2012, they met together with Mr Xie in China to work out how they would distribute the shares (“the June Meeting”). This was when they concluded the Allotment Agreement. It was literally an annotation written in Mandarin at the bottom of the page of the Goldrooster Prospectus containing the table reproduced above. There was no dispute as to its authenticity or as to its translation, which is as follows:

We pause to note that “Xinye, Zhong Yedian” were the names of two individuals who served as the auditors and accountants of Goldrooster AG and who were collectively known as the “Finance Team”. It is common ground that it had been agreed that they would be remunerated for their work in the listing through the issuance of share capital.

11 The parties took starkly different positions on how they came to agree on these terms. We will say more about this later, but for now it suffices for us to say that the nub of the dispute concerned the proper interpretation of the term “19%” in the Allotment Agreement. The respective positions of the parties may be summarised as follows:

(a) The Respondent contends that what was contemplated in the Allotment Agreement was the division of all the shares held by the Yap Companies. On this basis, the Respondent submits that the term “19%” was not in fact meant to be an expression of a percentage at all; rather, it was an idiosyncratic way of expressing the totality of the shares held by the Yap Companies, namely 5.5m shares. The Respondent then goes on to say that this lot of 5.5m shares was to be divided in the ratio of 10.35 parts to 6.65 parts to 2 parts between the Respondent, the Appellant, and the Finance Team, ostensibly in accordance with the remaining parts of the Allotment Agreement. On this interpretation, the Respondent would be entitled to 2,996,053 shares (which is calculated as follows: 10.35/19 x 5.5m).

(b) The Appellant contends the Allotment Agreement dealt (and was intended only to deal) with 19% of the post-listing shares in Goldrooster AG (approximately 20.7m). On this basis, he submits that the term “19%” does not refer to all the shares in the Yap Companies, but instead refers to a block of 3,936.839 shares, this being 19% of the post-listing total of 20.7m shares. On this interpretation, the Respondent would only be entitled to 2,144,541 shares, which is 10.35% of the 20.7m shares in Goldrooster AG after its listing.

12 On 27 June 2012, just less than a fortnight after the Allotment Agreement was entered into, the Appellant transferred his ownership of Xanti to the Respondent, and with it the 1,247,500 Goldrooster AG shares which Xanti held. In March 2013, the Appellant sold the 1,247,500 Goldrooster AG shares held by Fortune at a price of €1.05 per share and remitted approximately HK$6.4m to Mr Xie, to whom, according to the Appellant, the Respondent had assigned her rights to the balance 897,041 shares. This sum of HK$6.4m transferred to Mr Xie represented the proceeds of 897,041 shares in Goldrooster AG, less an amount which the Appellant said had previously been given to Mr Xie and which the Appellant said should be offset against the total sum owed (see [16(b)] below). Mr Xie eventually transferred this sum of HK$6.4m to the Respondent (see the Judgment at [106]). The net effect of all this was that the Appellant would appear to have transferred an interest in 2,144,541 shares in Goldrooster AG to the Respondent.

13 Because of their disagreement on how the Allotment Agreement should be interpreted, the parties disagree over whether the steps detailed in the preceding paragraph were sufficient to satisfy the Respondent’s entitlements under the Allotment Agreement. Unable to compromise, the Respondent commenced Suit No 558 of 2013 (“the Suit”), seeking the transfer of the 1,748,533 shares she contended was still owed to her or, in the alternative, damages. She arrived at the figure of 1,748,553 shares by subtracting the 1,247,500 shares formerly held by Xanti (which she had now received) from the total of 2,996,053 shares which she claimed she was entitled to (see [11(a)] above). At the time she commenced the Suit, the Respondent did not acknowledge that the sum transferred to Mr Xie should be considered as having been given in discharge of the Appellant’s obligations under the Allotment Agreement (though she later changed her mind and was willing to give the Appellant credit for the sum given to Mr Xie: see [17] below). The Appellant, on the other hand, claimed that he had fully discharged his obligations under the Allotment Agreement by transferring an interest to 2,144,541 shares to the Respondent or Mr Xie. The Appellant counterclaimed for a sum of approximately €795,000 which he claimed represented the Respondent’s 60% share of the Listing Expenses which remained unpaid.

14 In her amended statement of claim (“SOC”), the Respondent claimed that her entitlement to part of the 5.5m shares in Goldrooster AG which were held in the Yap Companies arose in various ways, aside from the Listing Agreement. She asserted that the parties’ entitlement to these shares had originated, and therefore fell to be divided, in the light of the following three agreements that had been concluded between 2007 and 2011. Before we summarise these agreements, we observe that their existence was only pleaded when the SOC was amended on 2 April 2015, four days before the date for the exchange of affidavits of evidence in chief (“AEIC”).

(a) First, sometime in October 2007, the Respondent’s sister concluded an agreement with Mr Li under which she acquired a 5% interest in Goldrooster Jinjiang (“the 2007 Agreement”). In December 2010, the Respondent and Mr Li agreed that the Respondent would hold a 5% interest in Goldrooster AG on her sister’s behalf and that this was to be in substitution of the 5% interest in Goldrooster Jinjiang that was promised under the 2007 Agreement.

(b) Second, under the Listing Agreement, 12% of the shares in Goldrooster AG were given to the parties. These shares were to be divided 60:40, in accordance with the 60–40 Agreement.

(c) Third, in January 2011, the present parties agreed to acquire a further 5% shareholding in Goldrooster AG and further agreed that the shares acquired pursuant to this agreement were to be divided equally. This was referred to as the “2011 Agreement”.

15 It may be noted that the first and third of these had nothing to do with the distribution of the shares that were to be issued in favour of the parties for the work they did in connection with the listing of Goldrooster AG. However, in her AEIC, the Respondent explained that these shares were to be held in three companies she set up but was advised against this by the Appellant, who offered instead to hold these shares on her behalf in the name of the Yap Companies. She claims that she trusted the Appellant implicitly at this time and agreed to his proposal. The Respondent averred that the parties were subsequently unable to agree on how the shares held by the Yap Companies were to be divided and it was for that reason that they entered into the Allotment Agreement, which she averred “provided that out of the [5.5m shares] of Goldrooster AG, the [Respondent] is entitled to … (2,996,053) shares”. However, the specific terms of the Allotment Agreement that would support this conclusion were not pleaded. All that was said was that the “Goldrooster AG shares to which the [Respondent] and [Appellant] were collectively entitled … were held by the [Yap Companies]”.

16 In his Defence and Counterclaim (“Defence”), the Appellant joined issue with the matters raised by the Respondent. He accepted that the parties had entered into the Listing Agreement but denied any knowledge of either the 2007 Agreement or the 2011 Agreement. As for the Allotment Agreement, he denied that it ever stated that the Respondent would be entitled to 2,996,053 shares in Goldrooster AG. Instead, he averred that it only provided that that Respondent would be “entitled to an interest in 10.35% of the shares of Goldrooster AG after listing”. Given that Goldrooster AG had 20,720,206 shares after listing, this would mean that the Respondent only had an interest in 2,144,541 shares. The Appellant claimed that the Respondent had already received her full entitlement to these shares since:

(a) the 1,247,500 shares in Xanti had been transferred to the Appellant; and

(b) the Appellant had assigned her rights to the balance 897,041 shares to Mr Xie, who had received value amounting to HK$9,513,119.81, which was the cash value of these shares, in two parts: (i) the Appellant had given Mr Xie a sum of around HK$6.4m in March 2013; (ii) the remainder of around HK$3.1m was to be offset against other amounts previously paid by the Appellant to Mr Xie.

17 For the most part, the Respondent’s Reply and Defence to Counterclaim (“Reply”) consisted of a bare denial of the Appellant’s case. The only point of note is that the Respondent now accepted that Mr Xie had received a sum of HK$6,389,910.21 on her behalf. However, she disputed the Appellant’s claim that Mr Xie had agreed for the balance to be offset against other sums which the Appellant had previously given to Mr Xie.

18 Even though the ostensible issue was one about the interpretation of a contractual term, the trial took the form of a contest between two competing narratives of why and how the parties had come to agree on the use of the term “Total 19%”. The Judge framed the issue as follows (at [54]):

19 After a comprehensive review of the evidence, she identified many problems with each party’s account of the events (see, for example, the Judgment at [64] and [97]). On the whole, however, she preferred the Respondent’s account which, she held, provided a more convincing picture of the relevant events.

20 In gist, the Respondent’s case in the court below may be summarised as follows:

(a)  She and the Appellant were collectively entitled to 22% of the shares in Goldrooster AG comprising 5% from the 2007 Agreement, 12% from the Listing Agreement, and 5% from the 2011 Agreement. These shares were, at the Appellant’s suggestion, held by the Yap Companies on their behalf. It was also stressed that it was no coincidence that the 5.5m shares held by the Yap Companies amounted to 22% of the 25m estimated post-listing shares in Goldrooster AG.

(b) During the June Meeting, the Appellant told the Respondent and Mr Xie that 3% of the shares were “gone” and so the Yap Companies only held 19% of the shares in Goldrooster AG, and not 22%. In her AEIC, the Respondent averred that she was confused by this as she knew that this would only be the case if the Greenshoe Option were triggered, but she had also been told by the Appellant at the same meeting that it had not. However, the Appellant was insistent that “the Greenshoe shares were ‘gone’” as there “may be investors subscribing for the Greenshoe shares in the next few months”.

(c) To get around this impasse, Mr Xie devised a way of preserving the ratios of distribution (as calculated based on their respective entitlements arising out of the three agreements detailed at [14] above) but using 19% as the base of distribution. What this entailed, the Respondent explained, was that they would “divide the Original 22% [which was equivalent to 5.5m shares and which were held by the Yap Companies] into 19 portions [such] that both the [Appellant] and [the Respondent] would be entitled to a number of these Goldrooster AG shares divided in 19 portions” [emphasis in original]. On this approach, “every percentage point in the Original 22% should be converted at the rate of 0.863 (as 19/22 = 0.863)”.

(d) The Respondent accepted that an application of this approach would not yield the figures in the Allotment Agreement. For example, it was the Respondent’s case that not all of the shares given to the parties under the Listing Agreement were for them. Instead, she claimed that 2% was for the Finance Team while the remaining 10% were to be divided 60:40. On this premise, the Respondent would rightfully have been entitled to 13.5% of the “Original 22%” (5% from the 2007 Agreement, 6% from the Listing Agreement, and 2.5% from the 2011 Agreement: see [14] above), which works out to approximately 11.65 portions (13.5 * 0.863). However, she was only given 10.35 portions in the Allotment Agreement. The Respondent explained that this was the case because it had been agreed that the Appellant and the Finance Team would each get more shares and that the increases for both were to come out of her entitlement.

21 The Judge noted that the Respondent’s account was far from perfect. For instance, she expressed deep misgivings regarding the authenticity of the 2007 Agreement (at [64]). On the whole, however, the Judge held that the Respondent’s account, though not problem-free, was “clearly preferable” to the Appellant’s, which made “even less sense” (at [97]). Chief among her reasons was the fact that the Appellant’s evidence was internally inconsistent (at [65] and [66]). During the trial, the Appellant proffered two contradictory accounts as to how the Yap Companies came to hold 5.5m shares (which was far in excess of 12% of the shares in Goldrooster AG that One Capital was entitled to under the Listing Agreement). The first was in his Supplementary Affidavit of Evidence in Chief (“SAEIC”), where he explained that Mr Li unilaterally agreed to revise One Capital’s entitlement from 12% to 19% after the listing was successful. However, no details were given to explain why Mr Li might have done this. The second was given during cross-examination where he admitted – contrary to what was said in his pleadings (see [16] above) and maintained in his SAEIC – that the 2011 Agreement was authentic. He explained that the 2011 Agreement related to a joint investment made by the Respondent and by a friend of his, one Ms Yu Yahong, in the listing of Goldrooster. He said that Ms Yu’s shares were registered in the names of the Yap Companies and that he held them on her behalf. However, apart from his testimony at trial, there was no evidence to support the assertion that any such arrangement existed.

22 In the premises, the Judge held that the term “Total 19%” in the Agreement “should be interpreted in this context as referring to all the Goldrooster AG shares held by the Yap Companies ie, 5.5 million shares” (at [99]). She therefore held that the Respondent was entitled to 2,996,053 shares of which 1,247,500 shares had already been transferred, leaving a balance of 1,748,533 shares. She granted the Respondent judgment in damages (quantified at the price of €1.05 per untransferred share). The Judge held that the HK$6,389,910.21 already transferred to Mr Xie was to be deducted from the judgment debt; however, she did not accept that the Appellant had consented for the remaining sum of HK$3,123,209.60 to be used as an offset for other purported debts owed by her to the Respondent and did not give the Appellant credit for this (at [102]). In the result, the Respondent was awarded a sum of €1,835,980.65 (1,748,533 * €1.05) less HK$6,389,910.21.

23 Turning to the counterclaim, the Judge noted at the outset that it was not disputed that (a) RMB 7,477,715.17 was the Respondent’s share of the Listing Expenses (at [113]) and (b) the Appellant had already paid the Respondent a sum of RMB 4m (at [129]). However, there were two issues to be decided.

(a) First, it was not clear if four transfers made by the Respondent to the Appellant between 10 June 2011 and 4 September 2011 totalling RMB 2.6m (“the subsequent transfers”) were meant to be applied towards the Listing Expenses or towards the parties’ other joint investments (at [129]). In relation to this issue, the Judge found that since the Appellant had not been able to establish that the payment of RMB 2.6m had been for the parties’ other projects, she found in favour of the Respondent that these were contributions towards the Listing Expenses.

(b) Second, it was not clear whether the sums of RMB 100,000 and S$300,000 paid to one Mr Teoh, an accountant, could be recovered (at [127]). She disallowed this claim as she found that the Appellant had not proved that Mr Teoh’s fees were paid for the Listing Expenses.

24 She therefore concluded that the Respondent was liable for a sum of RMB 877,715.17 (RMB 7,477,715.17 – RMB 4m – RMB 2.6m).

25 Insofar as the main claim is concerned, the Appellant submits that the conclusion reached by the Judge falls outside the scope of what the court can do in interpreting a contract. He also argues that the extrinsic facts relied on by the Appellant are inadmissible either because they were not pleaded and/or because they do not relate to a clear or obvious context. As for the counterclaim, the Appellant submits that the evidence shows that (a) the subsequent transfers were meant for investment in other companies and therefore should not be counted towards the discharge of the Respondent’s obligations in respect of the Listing Expenses; while (b) the work done by Mr Teoh was legitimate and the sums paid to him should be claimable as a legitimate part of the Listing Expenses.

26 The Respondent seeks to uphold the Judge’s decision in full. The key, she submits, was an examination of their competing accounts on how the term “Total: 19%” came to be agreed. On this basis, she submitted that the Judge was correct to hold that the Appellant’s account was inconsistent and contradictory and to be rejected. As for the counterclaim, the Respondent submits that the Judge had rightly placed the burden of proof on the Appellant. What was critical in this case, she contends, is that the Appellant had “unilateral control over how payments were to be applied” and therefore bore, but had failed to discharge, the burden of showing that the subsequent transfers were not meant for the Listing Expenses. As for Mr Teoh’s expenses, the Respondent argues that it should be disallowed because the evidence led by the Appellant on the issue was shadowy.

27 There are three broad issues in this appeal, which are as follows:

(a) whether the Judge had erred in the construction of the Allotment Agreement;

(b) whether the Judge was correct in finding that the subsequent transfers were meant to be applied towards the Listing Expenses; and

(c) whether the Judge was correct in finding that the sums paid to Mr Teoh cannot be claimed as Listing Expenses.

28 The judge applied the law governing contractual interpretation to the Allotment Agreement. The contextual approach to contractual interpretation in Singapore requires the court to proceed in two broad steps (see Zurich Insurance (Singapore) Pte Ltd v B-Gold Interior Design & Construction Pte Ltd [2008] 3 SLR(R) 1029 (“Zurich Insurance”) at [124]):

(a) The first step requires consideration of whether the extrinsic evidence sought to be adduced in aid of interpretation is admissible. This is a matter governed by the procedural rules of the law of evidence, which governs what and how facts may be proved.

(b) The second step is the task of interpretation itself, which involves ascertaining the meaning of expressions used in a contract, taking into account the admissible evidence. The rules which govern this process may be found in the substantive law of contract.

29 After careful consideration of the parties’ arguments, we agree with the Appellant that this appeal must be allowed because the extrinsic evidence sought to be admitted should not have been taken into consideration and because the interpretation preferred by the Judge is one which – in our judgment and with the greatest respect to the Judge – was not permissible in the present circumstances. We begin our detailed analysis in reverse sequence, namely, we will first consider the substantive task of interpreting the text, before turning to consider the admissibility of the extrinsic evidence relied on by the respondent and accepted by the Judge. We do so because, in our judgment, in the particular context of this case, there was simply no basis to go outside the terms of the Allotment Agreement and have recourse to extrinsic evidence.

30 A comprehensive summary of the law on contractual interpretation in Singapore may be found in our recent decision in Y.E.S. F&B Group Pte Ltd v Soup Restaurant Singapore Pte Ltd (formerly known as Soup Restaurant (Causeway Point) Pte Ltd [2015] 5 SLR 1187 (“Soup Restaurant”) at [30]–[42]. In gist, the purpose of interpretation is to give effect to the objectively ascertained expressed intentions of the contracting parties as it emerges from the contextual meaning of the relevant contractual language. Embedded within this statement are certain key principles: (a) first, in general both the text and context must be considered (at [2]); (b) second, it is the objectively ascertained intentions of the parties that is relevant, not their subjective intentions (at [33]); and (c) third, the object of interpretation is the verbal expressions used by the parties and so, the text of their agreement is of first importance (at [32]). In this judgment, we are concerned principally with the third key principle – the importance of the text.

31 In our judgment, the Judge’s interpretation must be rejected for the simple reason that the meaning ascribed to the term by the Judge is not one which the expressions used by the parties can reasonably bear. In Zurich Insurance at [122], we stressed, citing [63] of our previous decision in Citicorp Investment Bank (Singapore) Ltd v Wee Ah Kee [1997] 2 SLR(R) 1, that even under a contextual approach, the “meaning imputed by the court [must] be one which ‘the words are reasonably adequate to convey’”. While the court is entitled to depart from the plain and ordinary meaning of the expression used, there is a limit to what the court can legitimately do in the name of interpretation. Whether this limit has been transgressed is sometimes a matter of nicety but in our judgment, the line was plainly crossed here. The result the Respondent sought was one which she could achieve, if at all, only through a successful application for rectification.

32 For convenience, we reproduce the translated text of the Allotment Agreement below:

33 The term in question is “Total 19%”. It is clear that this refers to the “shareholdings after listing in Germany”. One cannot escape the fact that a percentage is, by any definition, a fraction out of 100. On its plain terms, this means 19% of the total issued shares after the listing of Goldrooster AG. There is no ambiguity in the terms of Allotment Agreement at all. In interpreting the term as a reference to a fixed allotment of shares the Judge had effectively read the expression “19%” out of the Allotment Agreement and substituted it with an integer – 5.5m. In our judgment, this crosses the line from interpretation to a variation of the contract. Indeed, this observation is not confined only to the term “Total 19%”; for in order to reach the ultimate result, the Judge also had to interpret the figures in the succeeding clauses of the agreement (10.35%, 6.65%, and 2%) as also giving rise to entitlements to defined numbers of shares. In particular, she had to read the Respondent’s “confirmed holding 10.35% in Goldrooster Co” as giving rise to an entitlement to 2,996,053 shares even though this was not 10.35% of the share capital of Goldrooster AG. How this was achieved was explained by the Judge at [49] of the Judgment:

34 In this extract, the full breadth of the decision below can be seen. With respect, this was not simply a case of expansive interpretation but something which amounted to a re-writing of the contract. This can plainly be seen in the fact that there can never be a situation in which the Yap Companies would both hold 5.5m shares and have a 19% shareholding in Goldrooster AG. Put simply, these are mutually exclusive events. That this is the case is abundantly clear if we refer to the table from the Goldrooster Prospectus which was reproduced at [8] above, which sets out the following two scenarios:

(a) If the Greenshoe Option was not exercised, the Yap Companies would own 5.5m shares. Assuming that the full free float of 5m shares was taken up in the listing, this 5.5m shares would amount to 22% of the shares in Goldrooster AG.

(b) If the Greenshoe Option was exercised, the Yap Companies would own 4.75m shares, which would represent a 19% shareholding in Goldrooster AG.

35 Under the Respondent’s interpretation, therefore, the Allotment Agreement purports to be a record of a portion of shares in a scenario that was never supposed to exist. The Yap Companies would only ever own 5.5m shares if the Greenshoe option was not exercised, but in that case it would have a 22% shareholding in Goldrooster AG, and not a 19% one. Conversely, the Yap Companies would only own a 19% shareholding if the Greenshoe option was exercised, but in that case, it would own 4.75m shares, and not 5.5m shares. It is also clear that the Respondent’s interpretation cannot be a reference to the scenario which did in fact eventuate – as it happened, the Yap Companies held approximately 26.6% of the post-listing shares in Goldrooster AG because of the underwhelming response to the public offering. Something had clearly gone wrong with the language if the Respondent is correct as to what had been intended.

36 In our judgment, what was done amounted to a variation of the express terms of the Allotment Agreement. This is not something that can be done under the guise of interpretation. The parties have not disputed that the Allotment Agreement represents the complete agreement between them. This attracts the application of s 94 of the Evidence Act (Cap 97, 1997 Rev Ed) (the “EA”), which bars the use of extrinsic evidence to contradict, vary, add to, or subtract from the terms of a written agreement if it can be shown that the parties intended to embody their entire agreement in writing. Thus, even leaving aside the matters we deal with below at [41]–[60], the Judge’s use of extrinsic evidence to arrive at the conclusion she did was not legally permissible.

37 In reaching this conclusion, we do not mean to signal a retreat to literalism. As we noted above, courts can and frequently do depart from the plain meaning of words. Among the powerful insights that the modern contextual approach provides, as stated by Lord Hoffmann in Investors Compensation Scheme Ltd v West Bromwich Building Society [1998] 1 WLR 896 (“ICS”), is that the meaning of a word should not be confused with the meaning that would be conveyed by the use of that word in a document. The former, as he said, was “a matter of dictionaries and grammars”; the latter is the proper domain of contractual interpretation, which is about discerning the meaning that the expressions in the document would convey to a reasonable person with the relevant background knowledge (at 913C).

38 However (and this is where we come back to the importance of the text), in ascertaining the meaning that the words of a contract would convey to a reasonable person with the relevant background knowledge, the words used by the parties occupy primacy of place. In Arnold v Britton and others [2015] 2 WLR 1593, Lord Neuberger of Abbotsbury PSC (with whom Lord Sumption and Lord Hughes JJSC agreed) explained the reason for this as follows (at [17]):

39 The verbal expressions in a contract are the vehicle through which meaning is conveyed. As we recently observed in Lucky Realty Co Pte Ltd v HSBC Trustee (Singapore) Ltd [2016] 1 SLR 1069 at [2], “absent the text, the contract cannot be constructed out of context alone”; and as we commented in Zurich Insurance at [41], citing the work of Prof Ronald Dworkin, interpretation is, by definition, an activity which is taken in relation to an object, and the shape of that object will be a constraint on the interpretation which may be applied. It therefore follows that the further away that one gets from the text of the provision, the less likely that one will be able to show that a proposed interpretation is one which the document would convey to a reasonable reader, endowed though he may be with the relevant context and background knowledge.

40 This is entirely in keeping with what we have consistently maintained. In Zurich Insurance, it was stressed that the meaning ascribed to an expression must be one which it is capable of bearing (at [122]) and that where the court “seeks to reject the actual words used by the parties altogether… the alternative of rectification exists and may be the more appropriate remedy” (at [123]). In Sembcorp Marine, it was stressed that the process of interpretation was one whereby “the parties’ intentions as expressed in the contract are objectively ascertained” [emphasis in original] (at [28]). Thus, where there is a gap in the contract arising from a mistaken belief that a document accurately records a transaction when it in fact does not, then “[t]he proper remedy… is the rectification of the instrument in equity” (at [96]). We will address the interaction between interpretation and rectification at [61]-[68] below.

41 This is sufficient to dispose of the appeal. But for completeness, we turn to what we have described at [28] above as the first step of the analysis, which concerns the admissibility of the evidence. As a starting point, the parol evidence rule continues to survive in Singapore and it is embodied in s 94 of the EA. As we noted at [36] above, s 94 of the EA is engaged here and it precludes extrinsic evidence from being admitted to vary the terms of the Allotment Agreement, as had been done here. However, assuming for the sake of argument that s 94 did not bar the admission of the extrinsic evidence here, the next point to consider is that under s 94(f) of the EA, extrinsic evidence of the surrounding circumstances accompanying the conclusion of a contract may be admitted, even in the absence of any ambiguity, to aid in the exercise of the interpretation (as distinguished from the contradiction, variation, addition or subtraction) of the expressions used by the parties (see Zurich Insurance at [132(c)]).

42 Even if it were accepted that proviso (f) to s 94 of the EA applies, the general admissibility of extrinsic evidence is subject to several restrictions which may be found in the EA and the common law (see Sembcorp Marine at [38] and [65(b)]). In the main, these are:

(a) The pleading requirements set out at [73] of Sembcorp Marine, which require the nature, particulars, and effect of the extrinsic evidence sought to be used to be pleaded with specificity.

(b) The exclusionary provisions of the EA, chiefly those found at ss 95 and 96, which act as an absolute bar to the admissibility of extrinsic evidence in certain cases (see Sembcorp Marine at [65(c)].

(c) The continued bar against the admissibility of parol evidence of the drafters’ subjective intentions at the time of the conclusion of the contract outside situations in which there is latent ambiguity (see Sembcorp Marine at [59] and [65(d)]).

(d) The general requirement that the extrinsic evidence sought to be admitted must be relevant, reasonably available to all the contracting parties, and relate to a clear or obvious context (see Zurich Insurance at [132(d)]).

43 These are not merely technical requirements of procedure, but integral “control mechanisms” that balance the need for certainty and cost-efficiency in the litigation process on the one hand and the goal of giving effect to the actual intentions of the contracting parties on the other. The second and third are of ancient vintage and reflect the position in the common law at the time of the enactment of the Indian Evidence Act of 1872 (Act I of 1872) and they were subsequently codified in our own EA. They embody the common law’s long-held preference for documentary evidence over the vagaries of subjective recollection. Despite the many amendments to the EA since its enactment, Parliament has not thought fit to abolish these provisions. The first and last requirements are more recent and represent the products of judicial development of the common law. They ensure that the contextual approach towards interpretation does not result either in greater uncertainty or an excessive increase in the time and cost of legal proceedings (see Zurich Insurance at [129] and Sembcorp Marine at [71] and [72]).

44 In our judgment, both of the common law control mechanisms had been transgressed here. Even assuming the bar to the admission of parol evidence could have been overcome on the basis that what was involved was not to vary the Allotment Agreement (which is what we have held it was: see [36] above) but to adduce evidence of relevant surrounding circumstances, this would have run into the separate bar against the admissibility of evidence of subjective intent, which is reflected in the third of these control mechanisms. The Judge had held that this was a case of latent ambiguity – that is to say, that the meaning of the expressions in the document were not of themselves ambiguous but became so when applied to existing facts – and that therefore recourse to declarations of subjective intent may be had (see the Judgment at [54], reproduced at [18] above). With respect, however, we cannot agree. As we noted at [33] above, the Allotment Agreement purported to deal with the division of the parties’ shareholdings in Goldrooster “after listing in Germany”. It is not in dispute that Goldrooster AG is the only Goldrooster entity listed in Germany and that it had approximately 20.7m shares after the listing had been completed. There is therefore no latent ambiguity here that would permit references to declarations of subjective intent to be made: the meaning of the words are plain and they apply to the existing facts. Moreover, we should also make reference to s 96 of the EA, which specifies that where the “language used in a document is plain in itself and when it applies accurately to existing facts”, evidence may not be admitted to show that it was not in fact meant to apply to these facts. In Zurich Insurance at [77], we referred to s 96 of the EA as “prescribing a common-sense limit on the use of extrinsic evidence which has been admitted under proviso (f) to s 94”. This is plainly a case in which s 96 applies, and it precludes the use of extrinsic evidence for the purpose of showing that the shares to be distributed was other than the 20.7m post-listing shares in Goldrooster AG.

45 Furthermore, leaving aside all the other objections against the use of the extrinsic evidence that was relied on by the respondent in this case, the evidence also did not, in our judgment, comply with the requirements set out at common law – that is to say, the extrinsic evidence sought to be admitted was not pleaded and it does not satisfy the Zurich Insurance criteria. Given that the parties devoted a large portion of their submissions to these two points, we shall cover both of these points in some detail.

46 In Sembcorp Marine, we observed that the shift towards a contextual approach to contractual interpretation had not been accompanied by developments in civil procedure to keep the process of litigation within reasonable bounds. For this reason, we considered it timely and essential that the following four requirements of civil procedure be instituted (at [73]):

(a)  first, parties who contend that the factual matrix is relevant to the construction of the contract must plead with specificity each fact of the factual matrix that they wish to rely on in support of their construction of the contract;

(b) second, the factual circumstances in which the facts in (a) were known to both or all the relevant parties must also be pleaded with sufficient particularity;

(c) third, parties should in their pleadings specify the effect which such facts will have on their contended construction; and

(d) fourth, the obligation of parties to disclose documents would be limited by the extent to which the evidence are relevant to the facts pleaded in (a) and (b).

47 At [74], we then said that “[i]n general, extrinsic facts that are placed before the court in a manner that is not consistent with the above requirements will not be accorded any weight when a court is construing a contract”. These four requirements were incorporated as para 35A(2) of the Supreme Court Practice Directions (1 January 2013 release) (“the PD”) on 15 August 2013. Para 35A(1) of the PD specifically enjoins advocates and solicitors to have regard to these pleading requirements whenever they are engaged in “disputes involving a contextual approach to the construction of a contract”.

48 Taken together, these pleading requirements out in Sembcorp Marine embody what has been described, in a speech delivered extra-judicially, as a “cards-up approach” towards commercial litigation (see Sundaresh Menon, “The Interpretation of Documents: Saying What They Mean or Meaning What They Say” (2014) 32 Singapore Law Review 3 (“The Interpretation of Documents”) at para 45). The basic idea is not novel – O 18 r 8 of the Rules of Court (Cap 322, R 5, 2014 Rev Ed) (“the ROC”) already contains a list of matters which have to be pleaded with specificity – and the underlying philosophy behind their introduction can easily be grasped. As Buckley LJ said in relation to the English progenitor to our O 18 r 8 of the ROC, “[t]he effect of the rule is … for reasons of practice and justice and convenience to require the party to tell his opponent what he is coming to court to prove” (see the decision of the English Court of Appeal in In re Robinson’s Settlement; Gant v Hobbs [1912] 1 Ch 717 at 728).

49 Adherence to these pleading requirements will bring about many critical benefits, which can conveniently be summarised under the headings of practice, justice, and convenience referred to by Buckley LJ:

(a) First, adherence to the pleading requirements will limit the scope of the materials that would be discoverable, reducing the practical burdens of litigation. When the relevant factual matrix is properly pleaded, both sides will be put on notice, from an early stage, of precisely what documents are relevant to the inquiry. This will help to ensure that only evidence which satisfies the requirements set out in Zurich Insurance of being “relevant, reasonably available to all the contracting parties and [which] relates to a clear or obvious context” will be admitted, reducing the burden on counsel at trial. This will also help to prevent the courts from being inundated by a flood of evidence, with the inevitable consequence that the proverbial wheat will be lost amidst the chaff (see Sembcorp Marine at [72]).

(b) Second, it will ensure procedural fairness and substantive justice by allowing parties to be put on notice of the case that they will have to meet. It is a cornerstone of our system of adversarial justice that no party (or the court) should be taken by surprise at the trial. Instead, the key points of dispute should be known in advance in order that the issues can be adequately ventilated at trial.

(c) Third, it will allow matters to proceed more expeditiously, to the convenience of all parties and the court. When the parameters of the dispute are defined in advance, the need for lengthy and protracted interlocutory tussles over such matters as discovery will hopefully be reduced. This will smoothen the trial process and might even facilitate a greater number of negotiated settlements (see The Interpretation of Documents at paras 45(a) and 45(d)).

50 In this light, in our judgment, the pleadings had not been sufficiently particularised. While it is true that the Respondent’s case was that the 5.5m Goldrooster shares would be apportioned between the parties, this falls far short of what was required in the circumstances. All that was said in the SOC was that the parties were collectively entitled, under the various agreements, to 5.5m shares in Goldrooster AG, and that these shares were held by the Yap Companies (see [15] above). However, there was no explanation of how this figure of 5.5m shares came to be expressed in the Allotment Agreement by the use of the term “Total 19%”. Indeed, we note that when the SOC was first filed, the terms of the Allotment Agreement could not even be found. It was only in response to a request for further and better particulars in October 2013 that a translation of the Allotment Agreement was set out and even then, no explanation was advanced as to how the figure of 5.5m shares could be derived from any particular term in the Allotment Agreement, let alone the interpretation the Respondent sought to place on the words “Total 19%”.

51 Indeed, even if the Respondent had pleaded this construction, it would not have made a difference because it is not enough for a party to allude, in a vague and general manner, to the result which it seeks to reach. Instead, as stated in Sembcorp Marine, each fact in the factual matrix, the circumstances under which it was known to both parties and, crucially, the effect that such facts would have on the intended construction must all be pleaded with specificity (see [46] above). Despite two further amendments to the SOC (the last of which, the fourth, was only filed two weeks before the commencement of the trial and introduced, for the first time, the existence of the three agreements: see [14] above) there was no mention of any of the contextual factors relied on by the Respondent at the trial, which are set out at [20] above).

52 We also reject the Respondent’s argument that it was sufficient for these details to have been set out in the AEICs. This runs contrary to the settled principle of law that defects in pleadings cannot be cured by averments in affidavits, much less those filed so late in the day (see Abdul Latif bin Mohammed Tahiar (trading as Canary Agencies) v Saeed Husain s/o Hakim Gulam Mohiudin (trading as United Limousine) [2003] 2 SLR(R) 61 at [7]). The consequences of the inadequacies in the pleadings were predictable. As the Judge observed on more than one occasion, the parties’ narratives were by no means easy to follow or reconcile, because the parties frequently failed to meet each other’s cases and many crucial pieces of evidence were only led at the trial, after a long, confusing, and often circumlocutious process of cross-examination. In the circumstances, we decline to place any weight on the contextual factors raised by the Respondent at [20] above in construing the Allotment Agreement.

53 In Zurich Insurance, we explained that in order to be admissible, extrinsic evidence had to be (a) relevant, (b) reasonably available to all the contracting parties, and (c) relate to a clear or obvious context. These three requirements were distilled from the case law and are to be understood as follows:

(a) Evidence is relevant if it “would affect the way in which the language of the document would have been understood by a reasonable man” (see Zurich Insurance at [125(a)], citing Bank of Credit and Commerce International SA v Ali and others [2002] 1 AC 251 at [39]).

(b) The requirement of reasonable availability is straightforward. Given that a contract is a bilateral (or multilateral) agreement involving one or more parties, it is plainly right that its terms can only be interpreted by reference to material which all the parties to the agreement would reasonably have had access to. In this connection, the availability of the material is measured with reference to the “situation in which [the parties] were at the time of the contract” (see ICS at 912 per Lord Hoffmann, cited in Zurich Insurance at [125(b)]).

(c) Evidence relates to a clear or obvious context if it would “allow the court to objectively ascertain a clearly defined or definable intention held by both parties with respect to how the contractual term in question should be interpreted” (see Smile Inc Dental Surgeons Pte Ltd v Lui Andrew Stewart [2012] 4 SLR 308).

54 We are satisfied that much of the evidence sought to be admitted in this case does not satisfy these three criteria. We do not propose to go through each piece of evidence but will only focus on the 2007 Agreement for the purposes of illustration. As a starting point, we note that there are significant question marks surrounding the authenticity of this document. First, it was a contract in Mandarin that had been handwritten on a piece of notepaper. This is highly unusual, given that the substance of the agreement – an investment by the Respondent’s sister in Goldrooster Jinjiang – is evidently one which involved a significant sum of money. Second, a copy of this agreement was only produced in the Respondent’s 3rd Supplementary List of Documents dated 8 April 2015 – less than a week before the commencement of the trial – even though it had allegedly been in the Respondent’s possession since July 2010. In the circumstances, we agree with the Judge that the circumstances under which the 2007 Agreement was produced were “highly suspicious” and that the document “may have been fabricated” (see the Judgment at [62]).

55 However, even if we were to accept, for the sake of argument, that the document was authentic, it is clear to us that it does not satisfy the tripartite Zurich Insurance criteria. First, it does not relate to a clear or obvious context. The 2007 Agreement was an agreement between Goldrooster Jinjiang and the Respondent’s sister. The operative clause reads that Respondent’s sister had “invested in 5% of [Goldrooster Jinjiang’s] company shares in October 2007 and became the shareholder of [Goldrooster Jinjiang]”. On its face, there is nothing which connects the 2007 Agreement with the later listing of Goldrooster AG in Germany. The Respondent tried to get around this by pleading that this was supplemented by another private agreement between the Respondent and Mr Li in 2010 in which the former was to hold the shares due under the 2007 Agreement for her sister. However, there was no evidence led as to the precise terms of this second agreement concluded in 2010; nor were the Respondent’s sister or Mr Li called on to testify in the trial.

56 Second, both these agreements are plainly not matters which would reasonably have been available to the Appellant at the time of the conclusion of the Allotment Agreement. Instead, they were private agreements to which the Respondent was not party. The Respondent argued that this should not matter because the Appellant “displayed knowledge of the 2007 Agreement” when he was on the stand. In our judgment, this is wholly inadequate. What transpired during the trial was that the Appellant asserted, in response to a question about how the Yap Companies had come to be given 5.5m shares, that Mr Li “did not acknowledge the [Respondent’s] sister’s 5 per cent” shareholding. Even taking the Respondent’s case at the highest, what this shows is that the Appellant was aware of the Respondent’s sister had asserted a claim to a 5% shareholding in Goldrooster AG, it does not show that he (a) knew the terms of the 2007 Agreement which allegedly gave rise to such an entitlement; (b) acknowledged this entitlement; or (c) that it was relevant to what the parties were discussing in respect of the Allotment Agreement.

57 The Judge was alive to these difficulties, but she held that the fact that “[t]he written agreement may have been fabricated after the event … does not mean that no oral agreement in those terms ever existed” (see the Judgment at [62]). With great respect, we think this was an erroneous approach. For a start, there was no evidence to support the existence of such an oral agreement. As the Appellant points out, it was never the Respondent’s case that there was an oral agreement; instead, it was always the Respondent’s case that the agreement was in writing. This was made clear by the Respondent in her closing submissions in the trial and she reiterates this in her case on appeal, where she specifically states that the 2007 Agreement was a “written agreement”. Indeed, we note that the 2007 Agreement specifically records that both parties had “signed this agreement which was specially made [sic] to avoid verbal statements without evidence in the future”.

58 The Judge also appeared to be swayed by the fact that the difficulties in the Respondent’s account were eclipsed by what she considered were the far greater difficulties with the Appellant’s account (see the Judgment at [64] and [96]). This was also the approach adopted by the Respondent on appeal, who submitted that whatever doubts that might exist over the authenticity of the 2007 Agreement could be “mitigated after the evidence was considered in totality, including the cogency of the Appellant’s evidence”. With respect, we cannot agree. While we accept that both parties’ put forward narratives which were riddled with inconsistencies, the difficulties with the Appellant’s narrative, numerous though they might be, cannot assist the Respondent with the proof of her affirmative case; nor do the difficulties with the Appellant’s case mean that it was therefore permissible to admit and accept any and all of the extrinsic evidence sought to be admitted by the Respondent in aid of her proposed interpretation. Ultimately, the question for the court at this stage is whether the evidence sought to be admitted satisfies the Zurich Insurance criteria. In our judgment, it did not and therefore should not have been admitted.

59 In our judgment, the only piece of evidence that could satisfy the requirements set out in Zurich Insurance is the Goldrooster Prospectus, at the bottom of which the Allotment Agreement was written. It is relevant as it related to the shareholdings in Goldrooster AG; it was reasonably available to both parties as they plainly had it before them at the June Meeting; and it relates to a clear or obvious context – namely, the shareholdings in Goldrooster AG – which is what the Allotment Agreement was concerned with. However, the Goldrooster Prospectus, as we have explained at [34]–[35] above, in fact points away from rather than towards the interpretation the Respondent prefers.

60 The result of our analysis in the preceding paragraphs is that the extrinsic evidence relied on by the Respondent to support its construction is very largely inadmissible. On this basis too, the appeal must be allowed because Respondent’s construction of the document is based almost entirely on the extrinsic evidence which was led. Without it, there would be no basis at all for interpreting the term in the way the Judge did.

61 In truth, it would only have been possible to achieve the outcome the Respondent sought if there had been a successful application for rectification. As a starting point, we clarify that what we mean by “rectification” in this context is the equitable doctrine of rectification, rather than the common law doctrine. The former refers to the reconstruction or amendment of a document where there is a “mismatch between the parties’ agreement and the instrument which purports to record it”; the latter is the use of the process of contractual construction “as a tool for remedying obvious drafting errors” (see Gerard McMeel, The Construction of Contracts: Interpretation, Implication, and Rectification (Oxford University Press, 2nd Ed, 2011) (“McMeel”) at para 17.01). As pointed out in Goh Yihan, “Clarifying Rectification in Singapore” (2015) 27 SAcLJ 403, the scope for what is known as “common law rectification” in Singapore is greatly circumscribed because of the provisions of the EA. The question of whether there is still space for common law rectification does not arise here and we leave that to be decided on another occasion.

62 The difference between interpretation and rectification may be explained as follows. “Interpretation is the ascertainment of the meaning which the expressions in a document would convey to a reasonable person having all the background knowledge which would reasonably have been available to the parties at the time of the contract” [emphasis in original] (see Sembcorp Marine at [33]). By definition, interpretation proceeds from the premise that the intention of the parties is captured in the expressions which they used. Thus, the process of interpretation is constrained by the expressions used in the document and where the court undertakes the exercise of interpretation, it is “assigning one of a range of possible meanings” [emphasis in original] to the expressions which have been used (see Zurich Insurance at [46]). Rectification, on the other hand, is a form of relief that involves “correcting a written instrument which, by a mistake in verbal expression, does not accurately reflect [the parties’] true agreement” (see Agip SpA v Navigazione Alta Italia (The “Nai Genova” and “Nai Superba”) [1984] 1 Lloyd’s Rep 353 at 359 per Slade LJ). It therefore proceeds from the premise that the words of the document are to be departed from because they are inadequate to convey the objectively ascertained intentions of the parties.

63 While it has been suggested that the remedy of rectification was in danger of being subsumed under the broader umbrella of the contextual approach towards interpretation (see Richard Buxton, “‘Construction’ and Rectification after Chartbrook” (2010) 69(2) CLJ 253 at 261), this has never been the position in Singapore. We note that even Lord Hoffmann, perhaps the most recognised proponent of the contextual approach towards interpretation in the common law world, recognised the continued existence of rectification as a “safety net” (see Chartbrook Ltd and another v Persimmon Homes Ltd and another [2009] 4 All ER 677 at [41]). The difference between the two doctrines is not just academic (even though we do think that clear conceptual thinking and terminological precision are important: see Sembcorp Marine at [26]), but it also has important practical ramifications. We mention three of them.

64  First, it affects the admissibility of extrinsic evidence. The scope of extrinsic evidence which is admissible in construction is, despite the shift towards a contextual approach, still greatly circumscribed. Even putting aside the contested areas of prior negotiations (see Xia Zhengyan v Geng Changqing [2015] 3 SLR 732 (“Xia Zhengyan”) at [62]–[69]) and post-contractual conduct (see Soup Restaurant at [73]), it is well-settled that evidence of declarations of subjective intent are not admissible in the interpretation of a contract save cases of latent ambiguity (see [42(c)] above). However, it has long been held that in an action for rectification all relevant evidence can be considered (see McMeel at para 17.31), even evidence of declarations of subjective intent (see ICS at 913B per Lord Hoffmann). Second, it affects the scope of the court’s discretion. In an exercise of interpretation, the court is simply determining the meaning of a document and there is no scope for the court to say other than what the law demands is the true meaning of the document; however, if the matter concerns rectification, then the court is being asked to exercise its equitable power to amend a document in order that it may bear a different meaning from that which it appears to have on its face, based on the terms which the parties have chosen to use. In such a case, the court has discretion to refuse rectification on the usual grounds on which equitable relief may be denied (for instance, delay, change of position, or third party reliance): see Marley v Rawlings and another [2014] 2 WLR 213 at [40] per Lord Neuberger PSC. Third, it affects the incidence of the burden of proof. It is well established that the party seeking the equitable remedy of rectification bears the burden of proving the facts essential for such relief to be granted. However, construction is inevitable whenever a contract is placed before the court and it does not make sense to say that either party bears a burden of proof, even though both of them will likely advance competing constructions (see McMeel at para 17-32).

65 As Mustill J explained in Etablissements Georges Et Paul Levy v Adderley Navigation Co Panama SA (The “Olympic Pride”) [1980] 2 Lloyd’s Rep 67 at 72, equitable rectification is available in two broad situations: (a) “where there is a mistake common to both parties, the mistake being the belief that the document accurately records the transaction” and (b) “where one party is mistaken as to the compliance of the document with the transaction and the other party knows of this mistaken belief but does nothing to correct it.” The former is referred to as “common mistake” rectification while the latter is referred to as “unilateral mistake” rectification (see Industrial & Commercial Bank Ltd v PD International Pte Ltd [2003] 1 SLR(R) 382 (“PD International”) at [24]). The burden is on the party seeking rectification to show “‘convincing proof’ not only that the document to be rectified was not in accordance with the parties’ true intention at the time of its execution but also that the document in its proposed form would accord with that intention” (see PD International at [29]).

66 Here, the respondent’s essential case was that the parties had a common intention to divide all the shares in the Yap Companies and sought to capture this common intention in the Allotment Agreement (see [20] above). Given what we have said at [31]–[40] above about the limits of the meaning that can be ascribed to the term “Total: 19%”, it is clear that the Allotment Agreement cannot give effect to this alleged accord. If the Respondent is correct as to what had happened, there was a shared mistake in the expression of a common intention – namely, that the parties were mistaken in their belief that the expressions used were able to convey the meanings which they intended them to bear. In the circumstances, the only way in which the Respondent would be able to achieve the result she sought would be to persuade us to depart from the expressions used in the Allotment Agreement through an application for common mistake rectification, which was not pleaded.

67 To obtain such relief, it would have been necessary for the Respondent to show the following (see the decision of the English Court of Appeal in Swainland Builders Ltd v Freehold Properties Ltd [2002] 2 EGLR 71 at 74 per Peter Gibson LJ):

(a) The parties had a common continuing intention, whether or not amounting to an agreement, in respect of a particular matter in the instrument to be rectified;

(b) there was an outward expression of accord;

(c) the intention continued at the time of the execution of the instrument sought to be rectified; and

(d) by mistake, the instrument did not reflect that common intention.

68 While the remedy of equitable rectification might be available even if not pleaded (see the decision of the Singapore High Court in Ku Yu Sang v Tay Joo Sing and another [1993] 3 SLR(R) 226 at [39], following the English High Court in Butler v Mountview Estates Ltd [1951] 1 All ER 693 (“Butler”)), this can only be done if the result which would otherwise be achieved is one which is plainly contrary to the parties’ agreement (see Butler at 700A). Given how disputed the facts were, we cannot conclude that the result which was achieved in this case was plainly contrary to the parties’ agreement. In our judgment, therefore, this is not a case in which the equitable remedy of rectification should be granted. Nor, for that matter, was this the basis for the Judge’s decision since she thought she could base her decision on what she considered was the true interpretation of the Allotment Agreement (see the Judgment at [99]).

69 In the final analysis, the Allotment Agreement falls to be construed, more or less, on its own terms. It is a short document but there are two points which stand out for comment. First, the subject of the Allotment Agreement is clearly stated in the very first clause: it was the Respondent’s “shareholdings after listing in Germany” [emphasis added]. Second, each succeeding clause in the Allotment Agreement makes reference to a percentage shareholding (10.35%, 6.65%, and 2%) in “Goldrooster Co”. This must be taken as a reference to Goldrooster AG, which is the only listed Goldrooster entity. In our judgment, having regard to these two points the Allotment Agreement can reasonably only be construed as being concerned with the distribution of 19% of the 20,720,206 post-listing shares in Goldrooster AG. On this premise, the Respondent is entitled to 10.35% of this block, which works out to 2,144,541 shares, of which she has already received 1,247,500 (see [12] above), leaving a balance of 897,041.

70 The Judge held that (a) damages should be quantified at a price of €1.05 per share; (b) the Appellant should be given credit for the HK$6,389,310.21 which he had transferred to Mr Xie; but (c) it had not been proved that the Respondent had ever agreed to offset the balance against other debts (see [22] above). There has been no appeal against these decisions and we do not, in any event, see any reason to depart from these findings of the Judge. Thus, we hold that damages should be assessed in the sum of €941,893.05 (897,041 * €1.05), with HK$6,389,910.21 to be deducted from this amount.

71 Before we conclude our discussion of the main claim, we will add a few words about our decision in Xia Zhengyan. The Respondent in that case was the owner of a chain of children’s education centres. The holding company, of which she was the sole shareholder, was named “Apple Plus School International Pte Ltd” (“the Company”). The Respondent also owned shares in a number of franchisees, the precise shareholdings of which ranged from 25% to 50%. After several months of negotiations, the Respondent concluded a sale and purchase agreement (“SPA”) with the Appellant under which she agreed to sell part of her interests in the businesses to the latter. The dispute in that case centred on the construction of the following clause in the SPA:

72 The question before the court was whether, on a true construction of the clause, the Respondent (a) need only transfer 50% of her interests in the Company or (b) in addition to transferring her 50% shareholding in the Company, the Respondent was also required to transfer 50% of her shareholding in the franchisees, the names of which were specifically listed in parenthesis. The trial judge held that it was the former. He placed great emphasis on the fact that the material words had been placed in parenthesis, which he said suggested that the parties had intended to draw a distinction between the shares in the Company, which had to be transferred, with those in the franchisees, which did not. We disagreed. Andrew Phang Boon Leong JA, delivering the judgment of the court, said as follows (at [50]):

73 In the Judgment, the Judge cited the above passage and opined that the “common sense approach” required “an investigation of the deeper background to the Allotment Agreement and not simply a consideration of the immediate context in which it was made as put forward” (see the Judgment at [44]). If what the Judge meant by this was that a trial of the competing narratives given by the parties was needed in order that the subjective intentions of the parties may be discerned, as she appeared to suggest at [54] of the Judgment (reproduced at [18] above), then we respectfully disagree. Given what was said, we think it important to clarify the ambit of the remarks that we made in Xia Zhengyan.

74 We clarify unequivocally that the cited passage should not be read as marking a dilution of the objective principle or as inaugurating a shift towards a subjective approach towards contractual interpretation. All we were saying there was that a reasonable person construing the SPA would have to bear two important contextual points in mind. First, the parties were laypersons and therefore could not be expected to have expressed themselves with the exactitude that might be expected of experienced legal draftsmen. Thus, a “common sense approach” rather than a technical and legalistic approach (with an excessive focus on the “structure and language” of the clause) should be adopted (see Xia Zhengyan at [50]). Second, English was not their first language so the parties could not be expected to be sensitive to the finer nuances of English grammar. The agreement had been drafted in Mandarin by lay persons. Of course, once the document has been translated, and provided there are no issues as to its translation, the court’s task is to construe the English language translation. But in Xia Zhengyan, the judge appeared to approach the construction of the agreement as if it had originally been drafted by lawyers fluent in the technical niceties of the English language. This, it seemed to us, was inappropriate. For example, a native speaker might conclude that a clause that is marked off with parenthesis should be treated as a non-essential insertion that is added for explanation or amplification and may therefore be removed or ignored without detracting from the meaning or grammatical completeness of the rest of a passage. However, this might not have been the natural conclusion of non-native speakers such as the parties in that case. Thus, we did not think that the fact that the disputed clause was placed within brackets should be determinative of the construction that was to be preferred. When we said that consideration should have been given to the “reasonable and probable expectations that parties would have had” (at [50]), what we meant by this was that a reasonable person with the relevant background knowledge would have construed the SPA while being mindful of the linguistic limitations the parties operated under. We emphasise again that nothing we said in Xia Zhengyan resiles from the objective principle of contractual interpretation.

75 In any event, we note that Xia Zhengyan is readily distinguishable on the facts because the clause in question was properly ambiguous and both the readings advanced by the parties fell comfortably within the ambit of the words used. It is quite different from the present case, where the interpretation advanced by the Respondent is not one which the expression used by the parties could reasonably bear.

76 The second and third issues relate to the Listing Expenses. At the outset, we observe that the following points are not disputed.

(a) At the material time, the parties made joint investments in a number of Chinese companies besides Goldrooster (“joint investments”).

(b) The Respondent would transfer sums of money to the Appellant at various times which were intended to be applied either in relation to the Listing Expenses or the joint investments.

(c) The Listing Expenses, excluding the fees paid to Mr Teoh, amount to RMB 12,462,858.62 and were to be apportioned 60-:40 between the Respondent and the Appellant. On this premise, it follows that the Respondent’s share of the listing expenses is RMB 7,477,715.17 (see the Judgment at [113]).

77 We turn to the second issue, which concerns the purpose of the subsequent transfers. The only relevant documentary evidence is a table of expenses which was marked “D1”. This document was prepared by the Appellant and a copy was given to the Respondent in June 2011. It sets out a breakdown of the expenses incurred in relation to Goldrooster AG as well as the parties’ joint investments. D1 records that the Respondent had contributed at least RMB 4m to the Listing Expenses as at June 2011. The Appellant accepts that this sum of RMB 4m was intended to be applied towards the Listing Expenses. The “subsequent transfers” referred to at [23(a)] above were so termed because there were four separate transfers for a total of RMB 2.6m which took place between 10 June 2011 and 4 September 2011, after D1 had been prepared. The Appellant accepts that he had received this sum of money but disputes that it was intended for the Listing Expenses and contends that it was for the joint investments instead. This is the principal dispute before us. Given the paucity of evidence, the resolution of this will turn on the incidence of the burden of proof.

78 In SCT Technologies v Western Copper Co Ltd [2016] 1 SLR 1471 (“SCT”) at [21], we explained that “the first port of call for determining the incidence of the legal burden of proof is… the pleadings” (at [21]). Referring to the pleadings, it is clear that the legal burden of proof rests on the Appellant to show that the Respondent has the obligation to pay the claimed debt, as he is the plaintiff in the counterclaim. However, the Respondent does not dispute the principle that she is liable to pay the Listing Expenses. Instead, in her reply, the Respondent seeks to avoid liability on the basis of her positive defence that the liability had been discharged (at least in part) by way of the subsequent transfers. It is settled law that the burden of proving the facts necessary to establish such a positive defence falls on the defendant (see SCT at [23]–[24]).

79 Ordinarily, this would be the end of the matter, and it would seem that the burden of proving that the subsequent transfers were meant to be applied towards the Listing Expenses rests on the Respondent. However, one important feature of this case is the fact that the Appellant occupied the position of an accounting party. The undisputed evidence is that the Respondent regularly transferred lump sums to the Appellant, who would then apply these sums towards the various joint investments which they had together. In his capacity as the recipient of these monies, he was the only one who knew how they were to be applied and he therefore undertook to give an account. Indeed, this was why D1 had been produced. This triggers the application of s 108 of the EA, which provides that when any fact is especially within the knowledge of any person, the burden of proving that fact is upon him. This is an exception to the general rule, as set out in s 103 of the EA, that the burden of proof lies on the party who asserts a particular fact. As Lai Siu Chiu J explained in Surender Singh s/o Jagdish Singh and another (administrators of the estate of Narindar Kaur d/o Sarwan Singh, deceased) v Li Man Kay and others [2010] 1 SLR 428 (“Surender Singh”) at [218] and [219], this is a rule of fairness which has as its object the alleviation the evidential difficulties that would be involved in the proof of facts which are solely within the knowledge of the defendant.

80 There are three important points to note about s 108 of the EA :

(a) First, it does not apply where the facts in question are known by others apart from the defendant. This is in keeping with the purpose of the rule, which is to maintain fairness between the parties (see Surender Singh at [217]).

(b) Second, it operates in a “commonsense way” and the “balance of convenience and the disproportion of the labour, that would be involved in finding out and proving certain facts” must be taken into account in deciding whether the incidence of the burden of proof has shifted (see Surender Singh at [219], citing Sir John Woodroffe & Syed Amir Ali’s Law of Evidence vol 3 (LexisNexis Butterworths, 17th Ed, 2002) at p 4223).

(c) Third, in order for s 108 of the EA to apply, a mere allegation that there are facts which are solely within the knowledge of the defendant is insufficient; instead, the plaintiff has to establish at least a prima facie case against the defendant. It is only after this has been done that s 108 of the EA operates to place the burden on the defendant to avoid liability by proving the facts which are especially within his knowledge (see Surender Singh at [221]).

81 On the facts, the Respondent has shown, and the Appellant has admitted, that she transferred sums of money to the Appellant on a regular basis. It is undisputed that the Appellant had control over how these monies were to be applied. In our judgment, this is a case in which s 108 of the EA should apply. The burden falls on the Appellant to account for how the sum of RMB 2.6m had been used. Indeed, this also appeared to be the approach taken by the Judge because she likewise observed that “D1 [was] the [Appellant’s] document” before going on to hold that the burden rested on the Appellant to show that the subsequent transfers were meant for the joint investments and that, in the absence of such proof, the facts would be construed against him and in favour of the Respondent (see the Judgment at [129]).

82 When we put the point to Mr Devinder K Rai, counsel for the Appellant, during the hearing he very fairly conceded that (a) his client was in control of how the money had to be used; (b) that the Appellant therefore had a duty to account; and (c) this burden had not been fully discharged in this case because the only set of accounts adduced at trial was D1, which did not cover events after June 2011. In Rhesa Shipping Co SA v Edmunds [1985] 1 WLR 948 at 955H–956A, Lord Brandon remarked that “[n]o judge likes to decide cases on the burden of proof if he can legitimately avoid having to do so. There are cases, however, in which, owing to the unsatisfactory state of the evidence or otherwise, deciding on the burden of proof is the only just course to take”. In our judgment, this is precisely such a case. On this basis, we dismiss the appeal in relation to the subsequent transfers.

83 The third and final issue concerns the Appellant’s appeal against the Judge’s decision to dismiss his claim for part of the sums of RM 100,000 and $300,000 in fees that had been paid to Mr Teoh for work he did in relation to the listing of Goldrooster AG. These payments were evidenced by two cheques for these sums which were issued in favour of Mr Teoh. After reviewing the evidence, the Judge found that the Appellant had not proved that “[Mr] Teoh’s fees as quantified should be taken into consideration for determining the counterclaim” (see the Judgment at [127]). It is not clear whether the Judge was of the view that the payments had not been made at all (as the Respondent now appears to contend) or that while the payments had been made, it had not been proved that they were for work done in relation to the listing of Goldrooster Jinjiang. The latter is more likely, as the authenticity of the cheques was never challenged and the Judge seemed to accept that they represented evidence of payment of the stated sums (see the Judgment at [125] and [127]).

84 The Judge reached this conclusion because of concerns over the evidence. First, she pointed out that there was “no document evidencing the terms of [Mr Teoh’s] engagement” (at [125]) and that apart from the cheques, the only documentary evidence adduced was a set of powerpoint slides (referred to as the “Information Memorandum”), which was used for marketing purposes. Second, she noted that there was “no invoice or receipt that showed that the cheques relied on were issued in relation to the listing of Goldrooster AG” (likewise at [127]). Third, she observed that the Appellant “could not recall during cross-examination exactly when he and Mr Teoh came to be acquainted” and that when Mr Teoh was asked why he was approached, he could “only point to his accounting background and contacts in Germany” (at [126]). Finally, in relation to Mr Teoh’s own evidence, she noted that his AEIC was short and did not contain details of how he came to be engaged by the Appellant on the Goldrooster listing (at [127]).

85 These are all valid points, and we can well understand why the Judge was troubled by the fact that a large sum of money appears to have changed hands without any formal documentation. It was also unusual that the Appellant and Mr Teoh would have dealt with each other so informally when this was the first time they were working together (see the Judgment at [126]). However, in our judgment, it is not possible to infer from these two points alone that Mr Teoh’s account should therefore be dismissed entirely. From our review of the record, there is sufficient evidence to conclude that Mr Teoh had been paid for the work he did in relation to the listing of Goldrooster AG.

86 First, Mr Teoh’s evidence is supported by the testimony of Mr Ashley Soh, who was the Chief Financial Officer of the Goldrooster entity in Hong Kong. Mr Soh had been called to give evidence on behalf of the Appellant in relation to the expenses which had been incurred in relation to the listing of Goldrooster AG. During the trial, he was asked if he knew Mr Teoh and he replied in the affirmative. He was then asked what Mr Teoh’s role was in the listing and his reply was that Mr Teoh was “more like [a] consultant, helping out with the preparation of the info memo, all this, before I came in” and that he had “sort of built the foundation for [the] whole project” [emphasis added].

87 In his AEIC, Mr Teoh explained that he had been engaged to assist in various matters including “introducing bankers, coordinating with various parties and preparing the Information Memorandum”. This is precisely the work which Mr Soh said that he did. When he took the stand, Mr Teoh was asked which banker he engaged and he unhesitatingly replied, “VEM Bank… So if you think my fee is [in]sufficient – pardon me – if not [for] me, the project could not have started”. There are two points of note here. One is that VEM Bank was the underwriter of the listing and they were paid for their work. This would appear to suggest that Mr Teoh was familiar with the listing. The next is that Mr Teoh’s comment about how the project could not have taken off without him, while somewhat immodest, resembles what Mr Soh said when he described Mr Teoh as the person who “built the foundation” for the listing.

88 The Judge gave short shrift to Mr Soh’s evidence on the ground that “Mr Teoh’s contribution predates Mr Soh’s employment on the listing of Goldrooster AG” (see the Judgment at [126]). With respect, we do not see how this justifies a rejection of his evidence out of hand. The fact that Mr Soh knew about Mr Teoh even though he only became involved in the listing after the latter had left adds to, rather than detracts from, the credibility of Mr Teoh’s account. It was never suggested that Mr Soh was in cahoots with Mr Teoh and the Appellant, or that he had any reason to lie. In the circumstances, we consider that his evidence should in fact be given weight because it constitutes independent attestation of the matters in issue.

89 Second, the timing of the cheques lends some support to Mr Teoh’s account. The first cheque was dated 10 February 2011; the second was dated 26 March 2013. During his testimony, Mr Teoh explained that his fees were structured in two parts. First, there would be a payment of RM 100,000 for his fees; second, there would a bonus of S$300,000 which would be paid if the listing proved successful. This is consistent with the fact that the first cheque was issued before the listing of Goldrooster AG in the middle of 2012 while the second was issued after the listing had been completed. Further, the second cheque was issued in the name of Fortune (ie, the third of the Yap Companies) and was dated 26 March 2013. It will be recalled that in March 2013, the Appellant liquidated the shares in Fortune and transferred part of the proceeds to Mr Xie (see [12] above). In our judgment, this was not merely a coincidence and it is a more than reasonable inference that Mr Teoh was being paid for work which he had done in relation to Goldrooster AG.

90 Taking these two points into consideration, we are of the view that the appeal in relation to Mr Teoh’s expenses should be allowed.

91 In summary, our decision is as follows:

(a) The appeal on the main claim is allowed in full. The Appellant is to pay the Respondent a sum of €941,893.05 less HK$6,389,910.21 instead of the sum originally ordered by the Judge.

(b) The appeal on the counterclaim is allowed in part. The appeal against the Judge’s decision on the subsequent transfers is dismissed. However, the appeal in relation to the sums claimed in respect of Mr Teoh’s expenses (RM 100,000 and $300,000) is allowed.

92 The parties are to file written submissions limited to 5 pages within 2 weeks of the date of this Judgment as to the appropriate costs orders to be made here and below. We also make the usual order for the release of the security for the costs of the appeal.