My writing

YES Bank v Cantor Fitzgerald

Drafters fail their grammar test again, costing $7m

Lawyers interested in grammar  ̶   and shouldn’t we all be? ̶  have a new case to delight in. It goes under the name Cantor Fitzgerald & Co v YES Bank Ltd, and was heard in the Court of Appeal earlier in the year.[1] It raised, for drafters, this familiar conundrum: how far does modifying language in the contract extend? What, in grammatical terms, is the scope of that language?

The interpretation of a disputed contract is itself a question of law: that is why judges are able to resolve the dispute. But the grammar of the English language   ̶   the set of conventions that frame how we use the language to communicate   ̶  is a different matter.[2] It depends on ordinary English usage, which the courts need to establish as a core part of the process of determining the natural and ordinary meaning of the words used in the contract.[3]

All this gives the writer of a note like this (an ex-lawyer fascinated by grammar, in case you’re interested) freedom to comment without getting the law wrong. So, rather than getting into the details of the case at this point, let’s look first at the grammar issue.

In case it tempts you to read on, the botched grammar point in the drafting cost $7 million or so.

[1] [2024] EWCA Civ 695, on appeal from the High Court.at [2023] EWHC 745 (Comm).

[2] In giving the judgment of the Court of Appeal in YES Bank, Falk LJ says, after giving her view on the grammatical principles to be applied, ‘This is not a point of law for which any authority is required‘ (ibid at [36]).

[3] See for example Lord Neuberger in Arnold v Britton [2015] UKSC 36 at [15].

The grammar issue explained

The case exhibits a grammatical quirk of the language that often causes ambiguity in speech and writing. In contracts, the ambiguity often allows rival interpretations to explode into disputes[4]. The issue arises so often that it is surprising that drafting professionals seem to remain unaware of it.

[4] See chapter 7 of the author’s book The Language of Contracts (Bloomsbury, 2024), discussing cases such as Mid Essex Hospital Services NHS Trust v Compass Group UK and Ireland Ltd (t/a Medirest) [2013] EWCA Civ 200, and also discussing the High Court’s decision in YES Bank (the Court of Appeal decision arriving too late for inclusion).

Proposition 1: if an adjective precedes the first noun in a list of two or more nouns, it is a grammatically open question whether the adjective applies (a) to just the first noun, or (b) to all of them

The two crucial ingredients in the explosive mixture are what grammarians would call a modifier and a coordination. In the phrase

           cold ham and potatoes

the modifier is an adjective, cold. The coordination is the rest of the phrase, ham and potatoes. A coordination, at its simplest, consists of two words separated by a coordinator which is often and  or or.[5] The two words that make up the coordination are, in grammatical terms, of equal status within the sentence: neither is subordinate to, nor does it depend on, the other.

When these two elements, modifier and coordination, are present, English grammar does not dictate whether the modifier applies only to the part of the coordination which is adjacent to the modifier (only the ham is cold) or to both elements (both the ham and the potatoes are cold). Both are feasible grammatical readings.

A coordination may have more than two elements to which the modifier might apply. A list of three, such as

              royal palaces, parks and gardens

is an example. Indeed there is no theoretical limit to the number of items that might appear in such a list. A list that takes the form of [adjective] a, b, c, d and e is perfectly feasible, if a bit unusual.

As we shall see, the list in YES Bank comprised three items.

With a list of more than two items, the question is still how far the modifier applies. But it is still a question with a binary answer. In the above example, either royal applies just to palaces or it applies to all of the items on the list. There is no feasible reading that royal applies to palaces and parks but not to gardens, nor for that matter that it applies to palaces and gardens but not to parks. Either the whole list is modified, or just the item adjacent to the modifier. The same would apply to lists that are longer still.

The basic grammatical background is therefore that an adjective preceding a list of nouns may as a matter of grammatical principle qualify either all the nouns in the list or only the first of them. But is either answer more likely in practice than another?

[5] In the context of this issue, there is no difference between the behaviour of and and or, though the words are obviously fundamentally different in other contexts.

Proposition 2: with a two-item list, it is ‘more likely than not’ that ‘other things being equal’, the adjective applies to both nouns

It makes sense to begin with an authoritative work on the grammar of the English language. The Cambridge Grammar of the English Language, by Rodney Huddleston and Geoffrey K Pullum, fits the bill. The authors deal with our question at page 1286. They conclude that ‘other things being equal’, it is ‘in general more likely’ that the adjective relates to ‘the coordination’ (meaning, in our case, both nouns on the list). ‘Other things being equal’, they go on to say, ‘requires that the [adjective] could apply equally readily to both [nouns]’.

What do they mean by this? The question is best illustrated by considering first a case where the adjective cannot apply ‘equally readily’ to both nouns. In the sentence

                Has your child been vaccinated against German measles and mumps?

the adjective German plainly applies to measles only, not to mumps: German mumps is not known to medical science.

What you are doing when you read

                Has your child been vaccinated against German measles and mumps?

is subconscious and all but instantaneous, but important. As you move forward through the sentence, you instantly recognise, when you come across German measles, that this is something familiar: a ‘thing in itself’. This is no surprise: an adjective normally fits well with the noun it modifies precedes. If someone says to you

                I saw a flying horse today

you are immediately jolted into wanting more explanation: the adjective does not go with the noun. But with German measles, there is no such issue.

Having absorbed German measles, your eye moves on to and mumps. Instantly (and without thinking about it) you resolve the ambiguity inherent in the grammar by asking yourself: can the adjective German apply to mumps? Answer: no. That clear answer triggers the reading that German applies only to measles. You interpret the sentence as meaning that the two concepts that are in coordination are German measles and mumps.

What about the contrary case, where the adjective could ‘apply equally readily’ to both nouns so that, in accordance with the CGEL test, the reader takes the adjective to apply to both or all the nouns in the list? In the sentence

                The soldiers were wearing blue coats and caps

most readers would naturally take the adjective to apply to both nouns. But what is the process (too much submerged to call it thinking) that leads to this conclusion?

As with German measles, the reader naturally absorbs the concept of blue coats: this is not a flying horse example. The next words the reader encounters are and caps. Again, as with and mumps, the reader asks: can the adjective blue apply to caps as well as coats? Answer: yes, of course. A cap might be blue just as much as any other colour.

But that doesn’t determine how the reader ends up interpreting the sentence. The reader must then decide whether blue does indeed apply to the caps as well as the coats. And this can only be answered by reference to context. In this example, the reader knows that soldiers’ uniform often features caps that are the same colour as their coats, and so interprets blue as applying to both nouns. The concepts that are in coordination are coats and caps, with blue taken as applying equally to both.

The reader is also influenced by the lack of punctuation within the phrase. A sentence in the form

                  The soldiers were wearing blue coats, and caps

is naturally taken as leading to the opposite conclusion: that the caps were some other colour than blue. Otherwise a careful writer would not have used the comma. The items in the coordination are naturally taken, because of the comma, as blue coats on the one hand, and caps on the other.

So the CGEL test seems to need elaboration. If the adjective can apply ‘equally readily’ to both nouns, the conclusion may indeed be that it does apply to both, but only if the context supports that reading. Perhaps that additional contextual gloss is what the authors of CGEL mean when they caveat their conclusion by saying that the ‘adjective applies to both nouns’ conclusion is ‘in general more likely’ (emphasis added) where the adjective can apply ‘equally readily’ to both nouns.

Proposition 3: the ‘equally readily’ test is not met if the adjective is familiarly associated with the first noun

The two examples discussed above are from two ends of the spectrum. With German measles and mumps, the adjective does not apply ‘equally readily’ to both nouns: it applies readily to the first, but not at all readily to the second. With blue coats and caps, the ‘equally readily' test is plainly met: a cap is just as likely to be blue as a coat.

Let us consider an example that lies between the two extremes. If you read

                She bought French bread and wine for the picnic

would you conclude that the wine was French as well as the bread? Leave aside the feature that French bread is French in style but not necessarily in origin, whereas the opposite is true for wine: that is, if it is not confusing to put it like this, a red herring.

Here, though the basic ambiguity of the language leaves both possibilities open, it is surely more natural, if one reads this acontextually,[6] to read French as applicable to bread but not to wine. Why is this so? This is not, after all, a German measles and mumps case, where the reader’s first subconscious question is to ask whether the adjective can apply sensibly to the second noun. With German mumps, the answer is no; but with French wine, the answer is yes. Nonetheless, the reader’s initial feeling is that French does not apply to wine. Why?

The answer may be that French bread is a familiar concept, a ‘thing in itself’ in which the adjective is so naturally associated with the noun that the reader instinctively ring-fences the concept, and is correspondingly reluctant to see the adjective can be wrenched off and attached to the second noun wine. The ‘equally readily’ test is not met. Contrast blue coats, which are not so readily associated as part of a single concept.

As soon as context is added, that initial impression may be contradicted. If you know that the supplies are being bought for a French-themed picnic and that those invited have been exhorted to bring only French produce if they can, you may very well see

                She bought French bread and wine

as suggesting that the wine she bought was French as well as the bread. But in the absence of a contextual clue, the more natural reading remains the one that sees the adjective as qualifying only the noun to which it is adjacent.

[6] A word that this writer has been wanting to have the opportunity to use since seeing it used in the course of Lord Walker’s brilliant judgment in Chartbrook Ltd v Persimmon Homes Ltd [2009] UKHL 38 at [86].

Proposition 4: if the list is longer, the same principles apply

The grammatical issue is the same with longer lists, as the royal palaces, parks and gardens example above suggests. If our example  lists are extended from two to three items, the reader’s conclusions are unaffected. A doctor recommending vaccination against

                German measles, mumps and chickenpox

is not taken to suggest that the mumps or the chickenpox are German. Someone who reports that a group of soldiers were wearing

                blue coats, caps and trousers

is taken to signify that all three of the items of clothing were blue: it would be odd for anyone to feel it necessary to draw attention to their wearing trousers at all (of course they would be). An announcement that

                She bought French bread, wine and cheese for the picnic

would probably be taken as meaning that only the bread was French unless we knew that the picnic was French-themed.

The longer a list, the less readily will an adjective be found that can apply to all the items on it. But it is still possible: a four-item list in a sentence like

                The fund will concentrate on short-dated gilts, debentures, commercial paper and other debt instruments

may perhaps be taken by a reader to signify that all the proposed investment categories are short-dated. But as a list grows longer, the longer the gap between the first item on the list and the last, so that someone who intends the adjective to apply to all of the items might consider providing a clearer signal and say

                The fund will concentrate on gilts, debentures, commercial paper and other debt instruments that are (in the case of each                         investment category) short-dated.

How did this play out in YES Bank?

The drafting dispute in YES Bank turned on this very issue.

The facts are beguilingly easy to describe. The parties to the contract were YES Bank  ̶   an Indian bank  ̶   and the US-based financial house Cantor Fitzgerald. The contract was entered into at a time when YES Bank was facing financial difficulties. It urgently needed to raise new capital from equity investors. Under the contract[7], Cantor Fitzgerald agreed to use its contacts and expertise to solicit investment for YES Bank. For that work Cantor Fitzgerald was to receive a flat fee. But if it persuaded certain specified target investors, named on a list, to put capital into YES Bank under what the contract defined as a Financing, it would receive a performance-based fee as well.

In due course, YES Bank did raise further funds, including some from the specified Cantor Fitzgerald contacts. So the question whether Cantor Fitzgerald was entitled to a performance fee boiled down to whether the capital had been raised in a Financing.

The contract defined Financing as

                the private placement, offering or other sale of equity instruments in any form...

What had happened by the time of the dispute was that, following the date of the contract, YES Bank’s condition had become so parlous that it had had to be rescued under a scheme imposed on YES Bank by the State Bank of India. The State Bank had taken a 49% stake in YES Bank and organised a further injection of funds. Only later did the investors on the Cantor Fitzgerald list subscribe for YES Bank shares. And they did this through a further public offer, not a private arrangement.

The entitlement of Cantor Fitzgerald to its further fee therefore depended on whether

                (a)   as they argued, the adjective private in the definition of Financing related just to the word placement that it immediately                  preceded (leaving offering and other sale unmodified by private, and so capable of extending to the further public offer to                        which the specified investors had subscribed);

or

               (b)  as YES Bank argued,  private applied not just to placement but also to offering and other sale as well (so that the further public offering that actually took place was not a Financing at all.

So the point in issue was exactly the one discussed above. How did the courts approach the grammatical question?

[7] Though this is guesswork, the author thinks that the drafting was probably generated by Cantor Fitzgerald rather than YES Bank.

The High Court

In the High Court, Bright J had a go at explaining what he saw as the grammatical principle:

              Where an adjective or determiner is followed by a series of nouns in a list, the conventional understanding is that it modifies                 all the nouns in that list (unless a discordant adjective or determiner breaks the pattern).[8]

So, in the absence of a ‘discordant adjective’, the grammar points towards the adjective applying to all the items on the list.

The natural reading of this observation is that by ‘discordant adjective’, Bright J was referring to an adjective later in the list that conflicted with the adjective whose scope is under examination, such as 

                We were served cold ham, beef and hot soup.

In that sentence, hot is plainly ‘discordant’ with the initial adjective cold, so it must be that the adjective cold, which clearly applies to the ham, does not apply to the soup as well. (Whether cold applies to beef as well as to ham is a matter of conjecture: arguably, if the writer had intended it to, cold ham and beef, and hot soup would have been the more likely way of expressing the meaning.)

But Bright J does not seem to acknowledge what we might call the German measles point: that the adjective will be taken to apply

to the first noun alone if it is plainly so clearly associated with it that it would be perverse to apply it to the other items. In

                German measles, mumps and chickenpox

no ‘discordant adjective’ later in the list ‘breaks the pattern’. But readers take the adjective German to apply to measles alone because German measles is a familiar concept, but German mumps and German chickenpox are not.

Bright J’s analysis of the grammatical issue is thus, with respect, not a complete one. His belief that the grammar points towards the adjective applying to all three items on the list suggests to him that this is the ‘ordinary meaning of the words’.[9] So, even before considering the contractual context, he is predisposed to agree with the YES Bank side of the argument. And his careful consideration of the context does nothing to disturb this ‘provisional view’.[10]

[8] [2023] EWHC 745 (Comm) at [86].

[9] ibid at [90], and repeated at [95].

[10] ibid at [136].

The Court of Appeal

In the Court of Appeal,[11] Falk LJ expresses herself less trenchantly, but still tends to the same view:

                While I would agree with Mr Beltrami KC, for Cantor, that there is no firm grammatical rule to the effect that an adjective or                      determiner at the start of a list of nouns qualifies them all, the nature of the list may well indicate that it does. At the least,                        unless something in the content of the list or another adjective or determiner within the list suggests otherwise, the reader                      will naturally tend to assume that an adjective or determiner at the start of a list qualifies the entirety of it.[12]

So Falk LJ, while not going as far as Bright J, still assumes that the more likely  ̶  the ‘default’   ̶   answer is that the adjective applies across the list as a whole. It is telling that the way that the question is framed (‘the nature of the list may well indicate that it does’) presupposes that answer, rather than the contrary one. The type of case where the adjective relates to the initial noun alone is presented as the exception rather than the rule (‘unless something in the content of the list or another adjective or determiner within the list suggests otherwise, the reader will naturally tend to assume that an adjective or determiner at the start of a list qualifies the entirety of it’).

Falk LJ’s grammar analysis seems more convincing than Bright J’s. She seems to be alive to the German measles type of example: that would be one where she would say that ‘something in the content of the list’ suggests that what she thinks of as the normal rule is displaced. But is it truly the ‘normal rule’? The analysis in the earlier part of this note suggests otherwise, at least in cases where the adjective and the first noun are commonly associated, as in French bread, wine and cheese. With that type of case, the balance is, it is submitted, on the other side of the scale: the provisional expectation is that the adjective cleaves to its familiar companion alone, and only if the context strongly points the other way (the French-themed picnic) will the adjective be seen as relating to the list as a whole.

These may seem nuances rather than differences, but they may have been influential in the Court of Appeal’s reasoning. Indeed, this is apparent from Falk LJ’s more detailed analysis which begins ‘Given that natural assumption [namely, that the adjective applies to all the items on the list], it is notable that the parties did nothing to counter it...’ [13]

But is that truly a reader’s ‘natural assumption’? It seems arguable that the phrase private placement is a little more like French bread than blue coats: a familiar combination of words that relates to something readily recognisable. Indeed, the familiarity of the association is such that placement on its own rarely appears unchaperoned by its adjective private. If that is so, the reader’s natural expectation, before any context is added, may well be to expect that private applied to placement alone, not to offering or other sale. If that had been the interpretation, it would have pointed towards Cantor Fitzgerald having the better argument on the grammatical issue.

As it was, however, both the High Court and the Court of Appeal found that the grammatical arguments favoured YES Bank. That meant that before embarking on contextual arguments, Cantor Fitzgerald were already one-nil down at half-time.

[11] [2024] EWCA Civ 695.

[12] ibid at [36]: emphasis added.

[13] ibid. at [37]: emphasis added.

The context

A court should always, of course, think about the contractual context and not just the particular words in dispute. That is part of the iterative approach to construction that is now firmly established as a necessary part of the judicial task[14].

Both the High Court and the Court of Appeal found that the contextual arguments favoured YES Bank and supported what they thought was the answer indicated by grammar: that private applied to offering and other sale as well as to placement.

One particular argument struck home with the Court of Appeal. This was that the same clause in which the disputed language appeared went into some detail about how the regulatory regime would apply if the capital raising took the form of a ‘Qualified Institutional Placement’ (a term of art under Indian law). As Falk LJ explains, the focus on QIPs as a special category makes more sense if private placement, offering or other sale refers exclusively to private arrangements. It would make much less sense if the phrase could extend to public offerings and public sales, since the same regulatory problem addressed by the QIP drafting would arise, but not be catered for. 

This is a powerful point. It was in vain that Cantor Fitzgerald argued that QIPs were dealt with explicitly only because, given what the parties foresaw, they were the most likely method of raising funds for YES Bank. But Cantor Fitzgerald might perhaps have made more headway with that argument if the courts had not already concluded, perhaps a little controversially, that the grammatical arguments favoured YES Bank.

[14] See for example Sara & Hossein Asset Holdings Ltd v Blacks Outdoor Retail Ltd [2023] UKSC 2 at [29] (Lord Hamblen).

The drafting defect

The case shows how delicate the task of a court is when, as happened here, the worlds of English grammar and contract law intersect. But, assuming the drafting was generated on the Cantor Fitzgerald side, the drafter missed a trick. It would have been easy enough to make clear that the adjective was limited in scope to the first noun by defining a Financing as

                  the

                        (a)  private placement;

                        (b)   offering; or

                        (c)    other sale

                 of equity instruments in any form...

which might well have done the trick without adding any additional language. The drafter might then have done well to remember the eiusdem generis principle, under which general language at the end of a list is sometimes inferred to be limited by reference to the characteristics of earlier specific items on the list. To avoid this, some more adjustments might have been necessary. But something like

                the

                        (a)  private placement; or

                        (b)  offering or other sale (in either case, whether public or private)

                of equity securities in any form

would clearly have done enough to secure Cantor Fitzgerald the $7 million it felt it deserved.

Ben Staveley

November 2024