Abstract

Keywords

Statutory interpretation; history; methods of interpretation; textualism; contextualism; intentionalism; Endumeni.

……………………………………………………….

1 Introduction

To what extent is the ordinary meaning of a word in a statute determinative of its legal meaning? For most of the 20th century we believed that the plain or ordinary meaning of a provision is almost always determinative of its legal meaning and that the broader context of an enactment such as other provisions in the same Act, headings, titles, preambles and debates on the floor of Parliament1 are of secondary concern, and to be invoked only when a word is vague, absurd or ambiguous. Our courts went about their interpretative exercise methodically: first find the ordinary meaning of a word or phrase. If the ordinary meaning is clear, the word or phrase should be given that clear meaning. If it is vague, absurd or ambiguous, then we may depart from the ordinary meaning to give the word or phrase a meaning intended by the legislature. This was known as the textualist, or as Professor Lourens du Plessis calls it, the literalist-cum-intentionalist approach to statutory interpretation, and it was the primary method for interpreting statutes for most of our history.2 But throughout the 20th century there were small cracks in the foundations of this approach, culminating eventually in a fracture which challenged the textualist approach. The case was Jaga v Donges and the decision was a dissenting one by Schreiner JA in which he proposed a different approach to interpreting statutes - one where the context of the legislation and the word or phrase being interpreted should be considered together.

Schreiner's approach to statutory interpretation found intermittent approval in the latter part of the 20th century, but for the most part, our courts still relied on the old textualist approach. Today our courts seemingly embrace Schreiner's approach largely due to the intervention of Wallis JA in Natal Joint Municipal Pension Fund v Endumeni Municipality, where he calls on us to consider interpretation as a unitary exercise, taking into account the

context and the provision being interpreted together.3 The Constitutional Court has endorsed Wallis JA's approach,4 and the Supreme Court of Appeal has eschewed reliance on the old textualist approach, with Wallis JA holding in a later case on contractual interpretation that "[the old approach] is no longer consistent with the approach to interpretation now adopted by South African courts in relation to contracts or other documents, such as statutory instruments or patents".5 But what are the reasons for this strong rejection? Does the new approach do something that the old approach couldn't? If Schreiner JA believed, as he did, that regardless of which approach one follows the result should be the same,6 why the strong rejection of the old approach?

This article does not make the case for a return to the old approach because the old approach is flawed. But Endumeni has not provided respite to the incoherent chain-novel that is statutory interpretation in South Africa. The stated aim of Endumeni was to provide "greater clarity about the task of interpretation".7 Lawyers and courts are no longer required to show that a word has an ordinary meaning that is not absurd, vague or ambiguous. They simply have to point out the objective meaning of a word having regard to the context. But Endumeni has not had the stabilising effect on statutory interpretation that it hoped. And, I believe that the courts will sound this message in the near-future too. But all is not lost. For the goals of Endumeni to be achieved, two things need to happen: Firstly, a theoretical approach to determining the ordinary meaning needs to be provided. This requires more than merely suggesting "look at the context". Secondly, the contextual considerations that may be taken into account must be limited. This article addresses the latter concern, namely the use of an unlimited context in interpreting statutes. A theoretical approach to determining the ordinary meaning is provided elsewhere.8

My aim is to show that Endumeni has not solved the problems which have plagued statutory interpretation for more than a century and that its

emphasis on the context is flawed. My prediction is that our courts will caution the overuse of context in future cases, and they may even determine that there are certain contextual factors, such as legislative history, that are impermissible to consult. I begin in Part II by tracing the historical conflict between the words and the context in the interpretation of statutes in South Africa. This is done to show that even after Endumeni, the interpretation of statutes in South Africa is as inconsistent as it was during the 20th century. In Part III I argue that Endumeni should not be read as embracing the kind of contextualism offered by Schreiner JA in Jaga. Finally, in Part IV I argue that there is a limited role for the context to play, using the jurisprudential debate between HLA Hart and Lon Fuller in the pages of the Harvard Law Review.9

2 The text and the context: A brief history

Fidelity to the text over its context in South Africa has its roots in the 1875 decision of the Supreme Court of the Cape of Good Hope. In De Villiers v Cape Divisional Council, John Henry de Villiers CJ decided that the rules of statutory interpretation should be determined with reference to English law rather than Roman Dutch law. He justified the adoption of the English approach to statutory interpretation over the Roman Dutch law approach by remarking that:10

[I]n construing statutes made in this colony after the cession to the British Crown, this court should, in my opinion, be guided by the decisions of English Courts and not the Roman Dutch authorities … some of the older (English) decisions … lay down rules which bear a close similarity to those of the Civil law. 'Every statute' says Lord Coke, 'ought to be expounded not according to the letter but according to the meaning: qui haeret in litera haeret in cortice.' There seems no doubt, however, that the enlarged or extensive interpretation of statutes which was admitted in former times has given way (except it would appear in old statutes) to restrict observance of the literal and grammatical sense of the words employed. The current of modern decisions seems to be in favour of considering the literal meaning of words in which the statute is expressed as the primary index to the intention with which the statue was made, and to abide by the literal meaning even where it varies from other indications of the actual intention of the Legislature.

Twenty years after Cape Divisional Council was decided, the same approach was surprisingly taken by the Supreme Court of Transvaal. In Hess v The State,11 Kotze CJ, who had championed Roman Dutch law

during his long judicial career,12 cited with approval a number of English authorities in favour of the textualist approach to statutory interpretation.13 By the close of the 19th century and the beginning of the 20th century, textualism, as it had been embraced in England, had established its roots in South Africa. And these roots would continue to find nourishment for decades to come because of the 1907 locus classicus in statutory interpretation, Venter v Rex.14

Venter's case concerned the meaning of the words "any person entering" in terms of section 3 of Ordinance 20 of 1905. Section 3 provided that any person entering the Transvaal would be guilty of an offence if he had been convicted of a crime in any place other than the Transvaal. Venter was born in the Cape Colony but later moved to the Transvaal where he became a citizen and resided for six years. He then moved to Natal and later the Orange River Colony where he stayed for another six years. During his residence in the Orange River Colony he was convicted of theft before the High Court at Bloemfontein and was sentenced to a year's imprisonment. Having served his sentence, he decided to return to the Transvaal in January of 1907 and was arrested eight months later for contravening section 3 of the Ordinance. The question before the Supreme Court of Transvaal was whether "any person entering" included someone like Venter, who was not entering the Transvaal for the first time but re-entering it. Innes J stated what became the golden rule of statutory interpretation as follows:15

[W]hen to give the plain words of the statute their ordinary meaning would lead to [an] absurdity so glaring that it could never have been contemplated by the legislature, or where it would lead to a result contrary to the intention of the legislature, as shown by the context or by such other considerations that the Court is justified in taking into account, the Court may depart from the ordinary effect of the words to the extent necessary to remove the absurdity and to give effect to the true intention of the legislature.

So statutory interpretation in South Africa, as in England,16 required our courts to do four things:

Despite the fact that Venter remained the most cited case on statutory interpretation in the 20th century,19 shortly after it was decided cracks in the foundation of this textualist approach gradually emerged. In 1912 Jacob de Villiers JA refused to abide by the plain meaning of the words in an Act governing prescription and instead cited a number of Roman Dutch authorities for the proposition that20

… the enquirer must take account of … context, and the reason of the law (ratio legis) … the history of the law in general … and [the] particular legal institutions about which the law to be interpreted deals (logical, systematic, historical interpretation).

Eight years later he made the same argument in dissent in the leading company law case, Dadoo v Krugersdorp Municipality. Here de Villiers JA was unwilling to agree with the majority decision that an Act which prevented Indian persons from owning property in the Transvaal did not apply to a company, even where the shareholders of the company are Indian. So instead he desperately cited the Digest, Donellus, Dernburg and other Roman Dutch authorities to escape the plain meaning of the statute - a plain meaning which he conceded did not prohibit a company, even one where the shareholders were Indian - from owning property.21 According to him, statutory interpretation required more than attention to the letter of the

law; the spirit of the law was sometimes more important, especially where Indians "float themselves into a private company with limited liability for the purpose of acquiring land".22

The emphasis on Roman Dutch authorities was perhaps most authoritatively made by Dr (later Chief Justice) LC Steyn in his famous scholarly contribution, Die Uitleg van Wette. But even before the publication of its first edition in 1946, opposition to the strict textualism of English law in favour of the more purposive Roman Dutch law authorities could be found in the law reports. One such example is the decision of Davis J in De Villiers v Cape Law Society, where he remarked that:23

There are a number of authorities dealing with the construction of a Statute in accordance with its spirit rather than with the literal meaning of the words used. I may first usefully refer to the somewhat neglected Roman-Dutch Law upon the subject. Voet 1.3.20 says: "That the legislator wished to depart from the proper signification of the words can be gathered from the antecedent or subsequent words of the law, from its preface, its conclusion, and the like; also from the reason of the law underlying the law itself: also from the fact that the words, if accepted in their proper signification would involve an absurdity, an impossibility, a defect, or a meaning not sufficiently suitable for carrying out the thing intended: these points are too well known to need any greater confirmation.

Steyn's work on statutory interpretation attempted to restore Roman Dutch purity at a time when the rules of statutory interpretation had for the most part already developed along the lines of English law.24 Die Uitleg van Wette was the first legal textbook in South Africa to be published in Afrikaans and even though it went through many editions and for many years was the only textbook on this subject, it remained untranslated. Edwin Cameron argued that this was intended to make plain to the world that "Steyn was a Roman Dutch purist determined to resist and, if possible, eradicate the pervasive grasp that English law and legal concepts had gained on the South African legal system…"25 Evidence of this can also be seen in the fact that he exclusively quotes from Roman Dutch authorities to support all of the interpretive presumptions in his book, despite the fact that we had already embraced English authorities to do the same thing.26 He laments the introduction of English rules on statutory interpretation by de Villiers CJ in Cape Divisional Council, and makes the usual prosaic purist arguments to

show that the introduction of English principles to govern statutory interpretation was erroneous.27 Steyn acknowledges that the vast majority of Roman Dutch authority on statutory interpretation is anti-textualist, unlike its English counterpart. English law, however, had only embraced the formalist text-based approach during the 17th and 18th centuries. For much of England's early history, the approach to statutory interpretation was one based on equity rather than the letter of the law.

In the 12th and 13th centuries medieval judges had the same freedom in their interpretation of legislation as they did in the application of the common law.28 This was because the legislative text was seen as having no special authority in itself, largely because the sovereignty of Parliament had not established itself as it did in the late 17th century.29 Judges during this period were essential in the drafting of statutes and would often impose the underlying policy considerations of the statute rather than the letter of the statute.30 Indeed, it was once remarked by Hengham CJ to a litigant attempting to exposit a statute of 1285 that he should not "gloss the statute, for we understand it better than you: we made it."31 Similar expressions can be found in other cases of this period including that of Bereford CJ who, though not personally involved in drafting the legislation in question, determined that it was perfectly acceptable to read words into legislation as the drafters had negligently omitted to include what they meant.32 Equity was thus central to the interpretation of legislation in England, and was rooted in the Aristotelian idea that the spirit of the text informed the meaning of the text over general words which were inherently deficient in covering every case.33 So strong was the view of equity at the time that it was even believed by some that the common law could overrule statutes enacted by Parliament. In the Bonham's Case of 1610, Coke CJ observed that "It appears in our books that in many cases the common law will control acts of parliament and sometimes adjudge them to be utterly void; for when an Act of Parliament is against common right and reason, or repugnant, or

impossible to be performed, the common law will control it and adjudge it as being void."34

Another important case which reflects the powerful role of equity at the time in English Courts is one which has often been embraced by our courts and continues to find application: the mischief rule of the Heydon's Case.35 Heydon is generally seen as authority for a purposive approach to interpretation in England, and elsewhere, by allowing judges to consider any defects in law for which Parliament has provided a remedy, and adopt an interpretation which supresses the mischief and advances the remedy according to the true intention of the legislative drafters.36 Although this particular aspect of equity, namely correcting the mischief, survives in England today,37 other aspects of equity such as supplementing legislative text would not last, as the century following the Heydon Case brought about the Glorious Revolution of 1688, which would establish a sovereign Parliament and fundamentally alter the course of statutory interpretation in England and in turn in South Africa.

The late 17th and 18th centuries saw a rejection of the equitable approach to statutory interpretation in England.38 Blackstone's rejection of the power of English courts to overturn legislation enacted by Parliament and of the tradition that legislation should be construed within the bounds of the enactment would have a pervasive influence over statutory interpretation in England for centuries to come.39 When Lord Denning MR attempted to resurrect the equitable approach to statutory interpretation in a 1950 decision by filling in gaps in the words, he was rebuked on appeal by the House of Lords as "nakedly usurping the function of the legislature under the thin guise of interpretation."40 Equity thus yielded to fidelity to the text.

Roman Dutch law on the other hand has always preferred the spirit of the law over its black letter. Although one may find sporadic indications that the letter trumps the spirit in some Roman Dutch authorities,41 they are

overshadowed by the vast majority of authorities which place the spirit at the heart of all interpretive inquiries.42 In the Corpus Juris Civilis, it was said that "interpretation is not proper without taking into consideration an entire law, either to decide, or give an opinion on any particular portion."43 This is known as interpretation ex vercribus actus or interpretation from the "entrails or bowels of an Act", which looks towards the broader context of the legislation enacted rather than the words in isolation of its context.44 Similarly, in the period of the aequitas, the principle function of the interpreter was to seek the intention behind the word and the form.45 Even Steyn, in addition to his rebuke in Cape Divisional Council, notes that Roman Dutch law at its core is anti-textualist.46 I mention this only because it would seem natural and even inevitable that Steyn would embrace these anti-textualist traditions in his Uitleg van Wette. But Steyn does not embrace the natural law traditions that pervade Roman Dutch law.47 And although Steyn expresses disapproval of the textualist approach the first chapter of his book is entirely textualist, being dedicated to the primary rule of interpretation, which concerns determining the literal and grammatical meaning of words, and when they are ambiguous ascribing a meaning to them intended by the legislature. Perhaps it is the case that Steyn was reluctant to fully embrace the Roman Dutch traditions of statutory interpretation because they could not be reconciled with his "executive-mindedness" so famously captured by Edwin Cameron in his assessment of Steyn's contribution to our law.48 So Steyn does very little to truly revive the Roman Dutch traditions of statutory interpretation in South Africa. Instead, his work is more akin to the English textualist approach in Cape Divisional Council, except that this time it is covered in civilian drag.49 After Steyn's contribution, dependence on the Roman Dutch authorities for greater reliance on the context was almost non-existent and the next big break for the context in statutory interpretation would come from Schreiner JA's famous 1950 dissenting opinion in Jaga v Donges.50

Jaga's case concerned the interpretation of the words "sentenced to imprisonment". The two appellants, Jaga and Bhana, had pleaded guilty to a statutory offence and were sentenced by a magistrate to a fine of 50 pounds or three months in hard labour, and a further three months suspended for three years, conditional upon the appellants not being convicted of a similar offence. But there was a bigger problem for Jaga and Bhana. By law the Minister of the Interior could remove "undesirable inhabitants" from the Union if they had been sentenced to imprisonment. Jaga and Bhana argued that "sentenced to imprisonment" means that they must have been sentenced to serve a term of imprisonment. Because no jail-time was given, the Minister, according to them, had no right to remove them from the Union. For Centlivres JA, this was a simple matter of determining whether the ordinary meaning of "sentenced to imprisonment" includes a suspended sentence. The answer for him was yes. But Schreiner JA disagreed.

Schreiner began his dissent by pointing out that there are two ways to go about statutory interpretation. The first is to do it methodically as Venter's case does. He then proposes a second approach, where the context is not relegated to a secondary consideration to be utilised only when the word is vague, absurd or ambiguous. According to Schreiner our understanding of what a particular word means is contingent on its context - we do not understand words divorced from the circumstances in which they are used.51 So it logically follows that when we give meaning to words we should give meaning to them in the context in which they are used, instead of considering the context only at a later stage when we have doubts. But for Schreiner, the context is wider than merely the context of the enactment. It includes its purposes, its background, and the practical consequences of one interpretation in comparison with another.52

Schreiner's approach found intermittent approval in the second half of the 20th century. But it did not replace the old approach. Some courts cited it with approval, as can be seen from the concurring judgment of Joubert AJA in Ebrahim v Minister of the Interior;53 Wessels AJA in Stellenbosch Farmers' Winery v Distillers Corporation SA;54 and Rabie CJ in University of Cape Town v Cape Bar Council, to name but a few. But there were always decisions by the Appellate Division to the opposite, endorsing the old

textualist approach, as can be seen in Public Carriers Association, where Smalberger JA remarks "it must be accepted that the literal interpretation principle is firmly entrenched in our law and I do not seek to challenge it."55

So, after more than a century of inconsistent and grossly contradictory jurisprudence on the text and the context, it seems a respite that Wallis JA has solved the problem by adopting Schreiner's approach as the law to be followed in Endumeni, The back and forth has been ended with the Constitutional Court endorsing Endumeni, and any attempt to get the pendulum swinging again has been rejected by the Supreme Court of Appeal when Wallis JA himself later holds in Bothma-Bato that the old approach is dead.56 Well… this is not quite true.

In a growing number of judgments that cite Endumeni, it seems that it is often cited only for the proposition that it is the correct approach to the interpretation of statutes, wills and contracts. But immediately after this, our courts revert to the ordinary meaning as it was intended by the legislature or contracting parties, doing the exact opposite of Endumeni. In addition to this, even though Endumeni was decided in 2012 and first endorsed by the Constitutional Court in 2013,57 in a number of decisions the Constitutional Court, the Supreme Court of Appeal and some divisions of the High Court have endorsed Endumeni and the old approach to statutory interpretation simultaneously without recognising the contradiction. Here are a few examples that illustrate this point: in Grindstone Investments the Constitutional Court cites Endumeni as the authority for the approach to interpretation, but in the paragraph immediately following this it cites a decision by the Appellate Division calling for words to be given their ordinary grammatical meaning used by the parties in a contract.58 In Excellerate Holdings, Meyer J cites Endumeni as authority for the established principles of interpretation, but in the paragraph immediately following it cites authority for determining the intention of the legislation, doing the exact opposite of Endumeni.59 In Public Servants Association, Nkabinde ADCJ tells us that we may depart from the ordinary meaning of words when there is an

absurdity, contrary to her Court's endorsement of Endumeni.60 In Jordaan Fourie J begins the interpretive exercise by citing authority for finding the "intention of the rule maker" and then follows this with Endumeni, failing to recognise the contradiction between the two approaches.61 In Mitchell Baartman AJA cites authority for the old approach and Endumeni in the same paragraph without any reference to the fact that the old approach has been overturned.62 In 2016 Henney J tells us in Nteta that we may depart from the ordinary meaning of the words only when it is absurd - overlooking Endumeni entirely.63 In 2017 Mhlantla J cites both Endumeni and Cool Ideas as the approach to statutory interpretation despite the fact that the two cases offer contradictory approaches to statutory interpretation. And most recently, Mogoeng CJ overlooks Endumeni, and the fact that it has been endorsed by the Constitutional Court dozens of times, by describing the operation of the contextual setting in the same terms as the old approach:64

Some of those key interpretive aides that have by now become trite are the textual or ordinary grammatical meaning, context, purpose and consistency with the Constitution. Context comes into operation where the ordinary grammatical meaning is not particularly helpful or conclusive. And contextual interpretation requires that regard be had to the setting of the word or provision to be interpreted with particular reference to all the words, phrases or expressions around the word or words sought to be interpreted. This exercise might even require that consideration be given to other subsections, sections or the chapter in which the key word, provision or expression to be interpreted is located.

The cases referred to above are but a few of many more judgments that do the same thing. So it is clear that the problems that have plagued statutory interpretation for the last century have not gone away. And it is unlikely that things will change, but this doesn't have to be the case. It is almost certainly the case that the inconsistency is no longer a result of the conflict between Roman Dutch purists on the one hand and the modernists who have embraced English law on the other. Perhaps it is the case that our judges have inadvertently overlooked the demands of Endumeni, or that, as in many judgments from the 1950s onwards, they simply invoke the approach that achieves the outcomes they desire.65 Or could it be the case that the

demands of Endumeni are unclear? Does Endumeni provide us with the guidance we need when looking at the context, and is the approach of Schreiner JA in Jaga the same as that of Wallis JA in Endumeni?

3 Battle of the context: return of the Jaga?

In order to answer these questions, we need to look at what exactly Endumeni does, and to do that I begin (and end) with the text of Wallis JA's judgment. I quote extensively from the judgment with emphases in italics, so that the exact demands of Endumeni are clear. He begins in paragraph 18 by stating that:66

Interpretation is the process of attributing meaning to the words used in a document, be it legislation, some other statutory instrument, or contract, having regard to the context provided by reading the particular provision or provisions in the light of the document as a whole and the circumstances attendant upon its coming into existence. Whatever the nature of the document, consideration must be given to the language used in the light of the ordinary rules of grammar and syntax; the context in which the provision appears; the apparent purpose to which it is directed and the material known to those responsible for its production. Where more than one meaning is possible each possibility must be weighed in the light of all these factors. The process is objective not subjective. A sensible meaning is to be preferred to one that leads to insensible or unbusiness like results or undermines the apparent purpose of the document. Judges must be alert to, and guard against, the temptation to substitute what they regard as reasonable, sensible or businesslike for the words actually used. To do so in regard to a statute or statutory instrument is to cross the divide between interpretation and legislation. In a contractual context it is to make a contract for the parties other than the one they in fact made. The 'inevitable point of departure is the language of the provision itself', read in context and having regard to the purpose of the provision and the background to the preparation and production of the document.

Wallis JA then goes on to say at paragraph 19 of the judgment that:67

All this is consistent with the 'emerging trend in statutory construction'. It clearly adopts as the proper approach to the interpretation of documents the second of the two possible approaches mentioned by Schreiner JA in Jaga v Dönges NO and another, namely that from the outset one considers the context and the language together, with neither predominating over the other. This is the approach that courts in South Africa should now follow, without the need to cite authorities from an earlier era that are not necessarily consistent and frequently reflect an approach to interpretation that is no longer appropriate. The path that Schreiner JA pointed to is now received wisdom elsewhere.

And lastly at paragraph 20:68

Unlike the trial judge I have deliberately avoided using the conventional description of this process as one of ascertaining the intention of the legislature or the draftsman, nor would I use its counterpart in a contractual setting, 'the intention of the contracting parties', because these expressions are misnomers, insofar as they convey or are understood to convey that interpretation involves an enquiry into the mind of the legislature or the contracting parties. The reason is that the enquiry is restricted to ascertaining the meaning of the language of the provision itself. Despite their use by generations of lawyers to describe the task of interpretation it is doubtful whether they are helpful. Many judges and academics have pointed out that there is no basis upon which to discern the meaning that the members of Parliament or other legislative body attributed to a particular legislative provision in a situation or context of which they may only dimly, if at all, have been aware. Taking Parliament by way of example, legislation is drafted by legal advisers in a ministry, redrafted by the parliamentary draftsmen, subjected to public debate in committee, where it may be revised and amended, and then passed by a legislative body, many of whose members have little close acquaintance with its terms and are motivated only by their or their party's stance on the broad principles in the legislation. In those circumstances to speak of an intention of parliament is entirely artificial. The most that can be said is that in a broad sense legislation in a democracy is taken to be a reflection of the views of the electorate expressed through their representatives, although the fact that democratically elected legislatures sometimes pass legislation that is not supported by or unpopular with the majority of the electorate tends to diminish the force of this point.

So, Endumeni stands for three propositions:

But for Endumeni to achieve the clarity it aims to provide to the interpretation of statutes, it must, in my view, stand for two more propositions:

The most important contribution of Endumeni to statutory interpretation, in my view, is that it sounds the death-knell in our law for the intention of the legislature. Wallis JA considers the search for legislative intent as "unrealistic and misleading"69 because the process of legislative drafting is often riddled with difficulties that make it impossible to know what the intention was: legislation is drafted by legal advisors in a particular ministry, redrafted by parliamentary draftsmen, subjected to public debate committees, and very often passed by members of parliament who have not read let alone understood the Bill they are passing.70 An additional problem is that legislation by its nature is a product of negotiation, compromise and artifice, so it is impossible to know what the collective intention of Parliament is when they may have conflicting views on the meaning of a particular provision because it suits their party-political position. When Wallis JA speaks of an objective interpretive process, he means that we must interpret the language used in the document as it is and not on the basis of what Parliament thought, believed or intended it to mean. What does matter is whether Parliament said "yea" or "nay" - thereafter, the legislation takes on a life of its own, divorced from the will of Parliament.71 So, because Wallis JA speaks of an objective standard, it is clear that we must decide what the words mean on their most reasonable construction.72 Wallis JA does not speak of a reasonable reader standard, but it is in fact what our Courts do when they ignore the will theory. Consider for example the case of Democratic Alliance v African National Congress. Here the Court had to consider the word "stole" in an SMS sent by the DA to voters in Gauteng seven weeks before a national general election. The SMS read "[t]he Nkandla report shows how Zuma stole your money to build his R246mn home. Vote DA on 7 May to beat corruption. Together for change". Van der Westhuizen J says that we should understand the word "stole" in the way an ordinary reader would understand it reading the SMS, rather than in the technical sense, as the ANC proposed, which requires that one must first be convicted of theft. To support the conclusion that the reasonable reader would not understand the word to be used only in its technical sense, he quotes from the Concise Oxford English Dictionary (11th ed) to show its most

likely construction.73 A similar approach has recently been taken by Froneman J in Marshall, where he rhetorically asks "Why should a unilateral practice of one part of the executive arm of government play a role in the determination of the reasonable meaning to be given to a statutory provision?"74

There is another reason for doing away with the will theory, although it is not advanced by Wallis JA and is, in my view, the most important reason for dispensing with it - reference to the intention of the legislature is incompatible with the Constitution and the rule of law. Our courts are not the faithful agents of the legislature and the Constitution does not envision the courts as an organ faithfully searching for what Parliament meant or intended. This could perhaps be the case under a system of Parliamentary sovereignty, but it is repugnant to a system of constitutional supremacy. Our courts are faithful agents only to the text of the Constitution and the rights and values contained in it. This is what it means to have a "government of laws and not Parliamentarians".75 The only way we know what Parliament means is by what it actually says in the words it uses, reasonably interpreted. So even though one may find reference in our law reports to the intention of the legislature in the year 2018, we ought to be aware of Froneman J's condemnation of it in Marshall as a "rule originating in the context of legislative supremacy" which "misses our fundamental change from legislative supremacy to constitutional democracy".76

So Wallis JA clearly takes us away from searching for the intention of the legislature as required by Venter and later followed by Centlivres JA in Jaga. But does it "adopt … the second of the two approaches mentioned by Schreiner"?77 The answer is that it does for the most part, but not entirely. Although Endumeni says that it adopts Schreiner's approach in Jaga, Wallis' treatment of the context is different from that of Schreiner. For Schreiner the context is not limited to the statute only; it goes beyond it. In fact, he tells us that the "context, as here used, is not limited to the rest of the statute

regarded as throwing light of a dictionary kind on the part to be interpreted."78 Schreiner goes on to tell us that sometimes what is even more important than the words used is their purpose, scope and background. By all indications, he is quite happy to search for intent because it is part of the context.79 For Wallis, this does not appear to be the case. The only sensible reading of Endumeni, that is without contradiction, is that the context is limited to the enactment as a whole. In other words, when we interpret a word we do so in the light of the entirety of its written context and not its unwritten subjective context. We know in particular that Wallis excludes legislative history as a contextual consideration because he strongly objects to searching for the intention of Parliament, and because he describes the process of adopting legislation in unsparing terms: riddled with twisting processes, inattentive parliamentarians and partisanship which are clearly unhelpful to determining the most reasonable construction of a word.80 Although it remains open for our courts to determine whether legislative history should be considered as part of the context in statutory interpretation, it is difficult to see how it could be permissible in the light of this critique. But Wallis does leave one with a great deal of confusion as to the extent of the permissibility of the context, and a reading of both Endumeni and Bothma-Bato seem to envision a limited role for the context. Take for example, this statement from Bothma-Bato: "[w]hilst the starting point remains the words of the document, which are the only relevant medium through which the parties have expressed their contractual intentions, the process of interpretation does not stop at a perceived literal meaning of those words, but considers them in the light of all relevant and admissible context, including the circumstances in which the document came into being".81 If the written contract (or statute) is the only way to determine what is said - that is "the only relevant medium" - why is there a need to consider it in the light of all the "relevant admissible context"? And leaving aside for the moment whether there should be bright-line rules on the extent to which we can rely on the context, if what Wallis means by "the circumstances in which the document came into being" or "the material known to those responsible for its production" is the negotiation history of the contract (or the statute), then we are back to searching for the subjective, unwritten and fictitious mental state of what the parties or the legislature thought, meant or intended. So, if Endumeni is to be embraced without any internal logical contradiction, the relevant context can only be

the written context of the enactment as a whole.

4 Battle of the context: the two towers, HLA Hart and Lon Fuller

In 1958 a famous jurisprudential debate took place between Professors HLA Hart and Lon Fuller in the pages of the Harvard Law Review.82 In this part of the article I will use this debate to show that the linguistic context sometimes features in determining the legal meaning of words, but that it is less important than the ordinary linguistic meaning and should, therefore, be given a secondary role when determining the legal meaning of words. The debate concerned what is now considered the most famous hypothetical in the common law world: a rule that prohibits vehicles in a park.83 Hart's contribution was principally addressed to the claims of American Realists who saw and represented the law as indeterminate. He believed that their obsession with difficult cases on the fragile ends of the law misrepresented the everyday cases before courts where the law is determinate - in other words, taking the most difficult cases of the law does not represent its everyday operation. He later remarked in the Concept of Law that, "while they [the realists] throw a light which makes us see much in the law that lay hidden, the light is so bright that it blinds us to the remainder, and so leaves us still without a clear view of the whole".84 Fuller, who was no realist, responded to Hart because for him it is impossible for language to be a source of legal determinacy without regard to the context in which the language is used. The meaning of legal words, for Fuller, is always entirely a function of the context in which they are used.85 So which activities are proscribed by the rule, "no vehicles in the park"?

Hart distinguished between the core of determinate meaning and the penumbra of uncertain meaning. In the core, words have a settled meaning independent of their context, and the settled meaning is informed by something that all speakers of a particular language share, even when the context and circumstances are not known.86 This is why someone who is competent in the English language - or any of the other 11 official languages - may pick up the South African Constitution and understand that Parliament is made up of two chambers; that the National Assembly may not have more than 400 representatives; that no law may override the text of the

Constitution, and that when the President constitutes his or her cabinet, all the members of cabinet must be Members of the Assembly save for a maximum of two. One does not have to know anything about South Africa, its history of disenfranchisement, Parliamentary sovereignty or the circumstances in which the Constitution was adopted for this to be clear. The context does nothing to influence or determine the operation of these rules. So when a case falls within the core of the general terms of the text, the judge is obliged to apply the rule.87 But there are instances where this is not the case, and here we are concerned with the penumbra of debatable cases that could fall either within or outside of the rule. Let's assume that the same person turns to sections of the Constitution that concern the "best interest of the child", or "just and equitable" compensation when property is expropriated, or reasonableness as the standard against which to test government action in realising socio-economic rights. Here the kind of conduct or activity required by these terms is nebulous and creates debatable cases about what is included and excluded by the rule. In these cases, Hart argued, the rules run out and the judge should use his or her discretion and rely on other considerations when deciding such cases, including moral and political considerations.88 Unlike legal formalists, Hart did not deny the law's indeterminacy. He considered the formalist rejection of law's occasional indeterminacy as an ideological response grounded in the separation of powers rather than a social fact that judges also make law when law runs out.89 So because language plays a role in the law's determinacy, and because there are occasions where language might be indeterminate, the law in these cases will naturally be indeterminate.

But Fuller was not concerned with the indeterminacy of the penumbra alone. He contended that all interpretation involves indeterminacy, including that at the core. There can never be a settled ordinary meaning, for Fuller. So, he said to Hart, what if a group of local patriots construct a memorial by mounting in the park a working truck that was used during the Second World War? Clearly, it would fall within the core of its general meaning but serves a completely different purpose to the rule which prohibits vehicles in the park - assuming that the rule created was to prevent congestion and noise. Hart quite simply points out that it might be the case that a system's norms require looking to the purpose, but in doing so one is not concerned with what the law ought to be, but merely recognising a matter of social fact.90

So it may be the case that the vehicle used as a memorial in this instance falls outside the core. Looking to the purpose does not render the law indeterminate because language is not always indeterminate. And so the purpose does not exist as an independent reason to give meaning to words, as Fuller would contend.

To illustrate this point in practice, consider this: the no-vehicle rule is adopted in Green Point Urban Park to prevent noise pollution. If X were to drive through the park in a sports car with an engine that makes a loud noise, this would clearly fall within the core and be proscribed. If X, perhaps too eager to fulfil his civic responsibilities, were to decide to take his noisy lawnmower to the park to cut the grass, this might similarly be proscribed, although one might not at first instance consider the lawnmower to be a "vehicle". But if X, after a football match at the nearby stadium, blows his vuvuzela in the park, this is not proscribed by the rule. The reason for this is that the purpose of suppressing noise does not exist as an independent reason for the application of the rule, but is a subsidiary and reinforcing reason for it that must still have a relationship to the core.91 The core limits the purpose. This was not the case for Fuller, as his example prohibiting sleeping at the train station shows. Fuller says that we assume that the no-sleeping rule was adopted to prevent homeless persons from using the station as their residence. If the businessman who waits for his train happens to fall asleep, he is not considered to have broken the rule, but the homeless person who comes to the station with a blanket and pillow but remains awake is covered by the rule. For Fuller, the purpose always overrides the ordinary meaning of the rule.

It would be dishonest to suggest that Endumeni is authority for the purpose of always overriding the plain words. This is not so. In fact, Wallis JA tells us that sometimes either the context or the plain meaning of a word could predominate over the other element, depending on their level of clarity.92 But he goes on to tell us that when Courts claim that the ordinary meaning is clear in its context and that there is little ambiguity, they misunderstand how language works because for him, like Fuller, it is always context-specific. He tells us that seeing language as isolated from its context is "a product of a time when language was viewed differently and regarded as likely to have a fixed and definite meaning, a view that the experience of lawyers down the years as well as linguistics, has shown to be mistaken."93

And to support this view he cites a speech delivered by the former Chief Justice of New South Wales, James Spigelman, which in turn relies on the philosophy of Wittgenstein and Fuller.94 Assuming briefly that it is a "misnomer", as Wallis JA characterises it, to believe that words can be understood a-contextually, it does not follow that meaning can be determined only with reference to the full context in which words are used. If we knew nothing about the meaning of individual words, sentences, grammar and syntax, we would never be able to understand each other.95 The full context might give us clarity, but it will do nothing to help us understand what the sentence "the boy climbed the tree" means. We know what this means, divorced from its context, because we know that the "boy" means "a boy", the tree means "a tree" and that climbed defines an activity that is different from say "jumped" or "walked" or "ran". So when Wallis cites the speech of Justice Spigelman, which relies on Wittgenstein to prove that words cannot be understood in isolation, he suggests that we should be aware that Wittgenstein was not concerned with individual words as a unit of meaning. Instead he was concerned with how conventions are a function of language and meaning. So the word "boy" as it is used by a specific linguistic community determines its meaning. The community could decide over time that the word boy means something other than a male child, but it is the community that determines the unit of meaning.96

Consider the following example used by Spigelman in his speech and quoted by Wallis in a footnote of Endumeni:97 "[I]n an adaptation of an example originally propounded by Ludwig Wittgenstein, parents leave their young children in the care of a babysitter with an instruction to teach them a game of cards. The babysitter would not be acting in accordance with these instructions if he or she taught the children to play strip poker." Wallis says that this example "vividly" shows why context is always important. Does it? One does not need any context to know what the instruction "teach them a game of cards" means. We know what this means because we understand the ordinary meaning of each word in the sentence. It is true that the conventions of that community would probably consider it

inappropriate for a babysitter to teach children strip poker, but this has nothing to do with the unit of meaning and everything to do with how conventions inform the meaning of words. What matters, is that we know what their unit of meaning is. This is why, I imagine, we adopted the textualist rule, not because we didn't think that the context is important. We did, but we did not think that it was always important. Sometimes it helps us because the ordinary meaning is absurd, vague or ambiguous, but most of the time it is not. And the context does nothing to help us. Instead, it is likely to be used by litigants to cloud the most obvious and reasonable construction of words. Of course, we often got things wrong when applying the rule - sometimes because our judgments were outcomes-based - but this is not a reason to dispense with the rule. So the danger signalled by Wallis that courts should not give provisions a meaning that they would prefer over a meaning that they objectively have is all the more likely when we think that context is always important.

5 Conclusion

The first disagreement with Endumeni has come from a recent dissenting opinion of Majiedt JA and Davis AJA in CSARS v Daikin Air Conditioning South Africa.98 Although the dissent does not call for a return of the old textualist approach to statutory interpretation, it does argue that the context is fact-specific and can be applied to the interpretation of contracts, but not statutes. So it rejects the proposition in Endumeni that its unitary exercise can be applied to the interpretation of all legal documents and it does this based on the linguistic distinction between sentence meaning and speaker meaning. The dissent in Daikin is an interesting one because it seems to suggest that statutes communicate but do not converse. In other words, statutes are commands of an Austinian kind that must be interpreted formalistically. This approach contradicts Davis AJA's earlier approach to interpretation and adjudication.99 In any event, it is unlikely that we will reject Endumeni in the near future, but I predict that we will adopt rules limiting the context, especially where litigants seek to invoke all possible kinds of contexts in order to persuade courts that a word is opaque.

The aim of this article was to tell the story of the conflict between the text and the context in South Africa. In doing so, I hope that it is clear that the back-and-forth experienced in the 20th century on placing emphasis on either the text or the context has not gone away with Endumeni. The same

problem still plagues interpretation and it won't go away unless we are clear about the demands of Endumeni. Our courts also need to take statutory interpretation seriously. If the Constitutional Court truly embraces Endumeni, it should adopt its methodology rather than casually use the same terms as those used in the old approach, giving one the impression that there is an absence of method, and instead a desire to reach preferred outcomes. The same is true for courts around the country where it very often appears that something is said about the approach to statutory interpretation for the sake of the saying, rather than to embrace what is required. I hope that this article is not read as a call for the blanket rejection of Endumeni. I also hope that it is not seen as a pamphlet supporting the old approach. It is not. But I do hope that it has persuaded you that the emphasis put on the context by Endumeni is unwarranted and flawed. Endumeni has directed much-needed attention to the study of statutory interpretation, and we should appreciate that it has done away with the intention theory that has plagued interpretation for far too long in South Africa. But it is not without internal logical contradictions. Interpretation, as Wallis JA tells us, is the "process of attributing meaning to the words used in a document." My hope is that we stick to the document and give words a meaning that they reasonably have, and that when our courts tell us "we begin with the text", they end there too. But this is a battle for another day.

Bibliography

Literature

Baker Introduction to English Legal History

Baker JH An Introduction to English Legal History 4th ed (Butterworths London 2002)

Cameron 1982 SALJ

Cameron E "Legal Chauvinism, Executive-mindedness and Justice LC Steyn's Impact on South African Law" 1982 SALJ 38-75

Cowen 1980 THRHR

Cowen DV "The Interpretation of Statutes and the Concept of the Intention of the Legislature" 1980 THRHR 374-399

Cowen "Prolegomenon to a Restatement of the Principles of Statutory Interpretation"

Cowen DV "Prolegomenon to a Restatement of the Principles of Statutory Interpretation" in Cowen S (ed) Cowen on Law: Selected Essays (Juta Cape Town 2008) 97-142

Davis Democracy and Deliberation

Davis D Democracy and Deliberation (Juta Cape Town 1999)

Devenish Interpretation of Statutes

Devenish GE Interpretation of Statutes (Juta Cape Town 1992)

Devenish 1989 SALJ

Devenish GE "Interpretation 'From the Bowels of the Act': An Essential Methodology for Unqualified Contextual Interpretation" 1989 SALJ 68-82

Devenish 1991 De Jure

Devenish GE "Fundamental Concepts and the Historical Roots of the Interpretation of Statutes in South Africa" 1991 De Jure 77-93

Du Plessis Re-interpretation of Statutes

Du Plessis L Re-interpretation of Statutes (LexisNexis Durban 2007)

Fagan 2004 Acta Juridica

Fagan A "Section 39(2) and Political Integrity" 2004 Acta Juridica 117-137

Fagan 2010 SALJ

Fagan A "The Secondary Role of the Spirit, Purport and Objects of the Bill of Rights in the Common Law's Development" 2010 SALJ 611-627

Fuller 1958 Harv L Rev

Fuller L "Positivism and Fidelity to the Law: A Reply to Professor Hart" 1958 Harv L Rev 630-672

Hart Concept of Law

Hart HLA The Concept of Law (Clarendon Press Oxford 1961)

Hart Essays in Jurisprudence and Philosophy

Hart HLA Essays in Jurisprudence and Philosophy (Oxford University Press Oxford 1983)

Hart 1958 Harv L Rev

Hart HLA "Positivism and the Separation of Law and Morals" 1958 Harv L Rev 593-629

Le Roux and Perumalsamy Constitutional Perspectives on Statutory Interpretation

Le Roux W and Perumalsamy K Constitutional Perspectives on Statutory Interpretation (Oxford University Press Oxford forthcoming 2019)

Popkin Statutes in Court

Popkin W Statutes in Court: The History of Statutory Interpretation (Duke University Press Durham 1999)

Schauer 2008 NYU L Rev

Schauer F "A Critical Guide to Vehicles in the Park" 2008 NYU L Rev 1109-1134

Steyn Uitleg van Wette

Steyn LC Die Uitleg van Wette (Juta Cape Town 1963)

Van der Merwe Brown v Leyds

Van der Merwe D Brown v Leyds: Who has the King's Voice? (LexisNexis Durban 2017)

Waldron Dignity of Legislation

Waldron J The Dignity of Legislation (Cambridge University Press Cambridge 1999)

Case law

AfriForum v University of the Free State 2018 2 SA 185 (CC)

Aumeye's Case (1305) YB 33-5 Edw I

Belyng v Anon (1312) B & M 52

Bothma-Bato Transport (Edms) Bpk v S Bothma & Seun Transport 2014 2 SA 494 (SCA)

CSARS v Daikin Air Conditioning South Africa (Pty) Ltd 80 SATC 330

Dadoo v Krugersdorp Municipality 1920 AD 530

De Villiers v Cape Law Society 1937 CPD 428

De Villiers v The Cape, Divisional Council 1875 Buchanan 1980 50

Democratic Alliance v African National Congress 2015 2 SA 232 (CC)

Department of Land Affairs v Goedgelegen Tropical Fruits (Pty) Ltd 2007 10 BCLR 1027 (CC)

Dr Bonham's Case (1610) 8 Co Rep 114

Ebrahim v Minister of the Interior 1977 1 SA 665 (AD)

Grey v Pearson 6 H L Cas 106

Hess v The State (1895) 2 ORC 112

Hleka v Johannesburg City Council 1949 2 SA 842 (A)

Jaga v Donges 1950 4 SA 653 (A)

Jordaan v Tshwane City and Four Similar Cases 2017 2 SA 295 (GP)

KwaZulu Natal Joint Liaison Committee v MEC Department of Education, KwaZulu Natal 2013 4 SA 262 (CC)

Magor and St Mellons RDC v Newport BC [1952] AC 189

Marshall v Commissioner for the South African Revenue Service 2018 7 BCLR 830 (CC)

Mathiba v Moschke 1920 AD 354

Mavromati v Union Exploration Import (Pty) Ltd 1949 4 SA 917 (AD)

Mazibuko v City of Johannesburg 2010 4 SA 1 (CC)

Natal Joint Municipal Pension Fund v Endumeni Municipality 2012 4 SA 593 (SCA)

National Credit Regulator v Opperman 2013 2 SA 1 (CC)

Novaris South Africa (Pty) Ltd v Maphil Trading (Pty) Ltd 2016 1 SA 518 (SCA)

Olitzki Property Holdings v State Tender Board 2001 3 SA 1247 (SCA)

Principle Immigration Officer v Hawabu 1936 AD 26

Public Carriers Association v Toll Road Concessionaries (Pty) Ltd 1990 1 SA 925 (A)

Public Servants Association v Head of Department of Health, Gauteng 2018 2 SA 365 (CC)

R v Westenraad 1941 OPD 103

Reezen Ltd v Excellerate Holdings Ltd 2018 6 SA 571 (GJ)

Rex v Detody 1926 AD 198

S v Nteta 2016 2 SACR 641 (WCC)

Seluka v Suskin and Salkow 1912 TPD 257

Stellenbosch Farmers' Winery Ltd v Distillers Corporation SA Ltd 1962 1 All SA 485 (A)

Swart v Cape Fabrix (Pty) Ltd 1979 1 SA 195 (A)

Trinity Asset Management (Pty) Ltd v Grindstone Investments 132 (Pty) Ltd 2018 1 SA 94 (CC)

Tshwane City v Mitchell 2016 3 SA 231 (SCA)

Van der Westhuizen v Arnold 2002 6 SA 453 (SCA)

Venter v Rex 1907 TS 910

Legislation

Massachusetts Constitution, Part the First, art XXX (1790)

Internet sources

Spigelman 2005 http://www.lawlink.nsw.gov.au/lawlink/supreme_ court/ll_sc.nsf/vwPrint1/SCO_speech_spigelman180

Spigelman JJ 2005 The Principle of Legality and the Clear Statement Principle http://www.lawlink.nsw.gov.au/lawlink/supreme_court/ll_sc.nsf/vwPrint1/SCO_speech_spigelman180 accessed 10 July 2019

List of Abbreviations