(Yacoob ADCJ, Cameron J, Froneman J, Khampepe J, Maya AJ, Nkabinde J, Skweyiya J and van der Westhuizen J concurring)
This is an application for leave to appeal against a judgment of the Supreme Court of Appeal, which dismissed the appeal against a decision of the Western Cape High Court (High Court) by Hlophe JP. The High Court held that section 11(2) of the Regulation of Gatherings Act (Act) is constitutionally valid. This section provides a limited defence for an organizer of a gathering who is allegedly liable for riot damage resulting from that gathering. This liability is created by section 11(1) of the Act.
Section 11(1) and (2) is set out immediately to facilitate an understanding of the issues. Section 11(1) provides:
If any riot damage occurs as a result of—
shall, subject to subsection (2), be jointly and severally liable for that riot damage as a joint wrongdoer contemplated in Chapter II of the Apportionment of Damages Act, 1956 (Act No. 34 of 1956), together with any other person who unlawfully caused or contributed to such riot damage and any other organization or person who is liable therefor in terms of this subsection.
Section 11(2) provides:
It shall be a defence to a claim against a person or organization contemplated in subsection (1) if such a person or organization proves—
Two questions lie at the heart of this matter. The first is what section 11(2) means. In other words, does it create a real defence that meets the constitutional requirement of rationality? Assuming that the defence is rational, the second question is whether the defence nevertheless limits the rights contained in section 17 of the Constitution and, if so, whether that limitation is justifiable.
The first applicant is the South African Transport and Allied Workers Union (SATAWU), a trade union registered in terms of the Labour Relations Act. The second applicant is the Congress of South African Trade Unions (COSATU), a national trade union federation, which currently has 21 affiliated trade unions, with a combined membership of nearly two million. SATAWU is affiliated to COSATU.
The first and second respondents are Ms Jacqueline Garvas and Ms Thuraya Naidoo. At the time of the events which gave rise to this matter, both Ms Garvas and Ms Naidoo owned and operated small businesses in the Cape Town City Bowl. Ms Garvas’ business operations entailed the sale of bags and Ms Naidoo sold flowers. The third respondent is Chinatown (RSA) International Trading CC, which at the time of the incident that gave rise to these proceedings was the owner of a gift and stationery shop, situated in the Cape Town City Bowl. The fourth to eighth respondents are Mr Anees Soeker, Mr Andrew Njokwuemegi, Ms Dolores Rosanne Reitz, Mr Maurice Robertson and Mr Harold Burger, respectively. These respondents had financial interests in the vehicles allegedly damaged during the incident. The first to eighth respondents shall be referred to jointly as “the respondents”.
The ninth respondent is the Minister of Safety and Security (Minister). He was joined in the High Court proceedings after SATAWU had served a third party notice on him.
The City of Cape Town (City) joined the proceedings in this Court as an intervening party and the Freedom of Expression Institute was admitted as amicus curiae.
Our Constitution gives everyone the right to picket, present petitions, demonstrate and assemble, peacefully and unarmed.
In the exercise of this right, SATAWU organised a gathering of thousands of people to register certain employment-related concerns of its members within the security industry. This gathering was the culmination of a protracted strike action in the course of which some 50 people allegedly lost their lives. It is also alleged that during this strike action private property and property of the City was damaged.
In preparation for the gathering, SATAWU took steps to meet the procedural requirements set out in the Act. It gave notice of the gathering to the local authority and appointed about 500 marshals to manage the crowd. It apparently advised its members to refrain from any unlawful and violent behaviour and requested the local authority to clear the roads of vehicles and erect barricades along the prescribed route on the day of the gathering.
In spite of these precautionary measures, the gathering allegedly resulted in riot damage estimated at R1.5 million. Several people were injured and about 39 others were arrested. The first to third respondents claim that as a consequence of this gathering their shops were vandalised and looted. The vehicles in which the fourth to eighth respondents had pecuniary interests were reportedly damaged. As a result, the respondents instituted action for damages against SATAWU under section 11(1) of the Act, alternatively, the common law.
In its plea, SATAWU denied liability and delivered a conditional claim in reconvention. In that claim, it sought a declarator that the words “and was not reasonably foreseeable” in section 11(2)(b) of the Act are constitutionally invalid. The grounds for the challenge were that these words limited the right to freedom of assembly under section 17 of the Constitution and the right to fair labour practices under section 23 of the Constitution. The respondents and the Minister resisted this challenge.
In the High Court
By agreement between SATAWU, the Minister and the respondents, the High Court made an order that the merits be determined separately from the point of law. The essence of the order was that the constitutional validity of the words “and was not reasonably foreseeable” in section 11(2)(b) be decided first, and the proceedings relating to the rest of the issues be stayed, pending the determination of the point of law.
Before the hearing, SATAWU abandoned its reliance on section 23 of the Constitution. The only legal issue before the High Court was therefore whether section 11(2) unjustifiably limits the right to assemble peacefully and unarmed.
SATAWU advanced two inter-related arguments in support of its contention that section 11(2) unjustifiably limits the right to assemble peacefully and unarmed. First, that the word “and” between section 11(2)(b) and section 11(2)(c) imposes two requirements for an organization to avoid liability. One, the organization must prove that the act or omission which caused the riot damage was not reasonably foreseeable. Two, it must prove that it nevertheless took reasonable steps to prevent the occurrence of the act or omission that was not reasonably foreseeable. SATAWU contended that it is impossible for an organization to take reasonable steps to prevent an act or omission it did not and could not reasonably have been expected to foresee. It is for this reason that it contended that the words “and was not reasonably foreseeable” render the defence internally incoherent and self-destructive.
Second, that in all instances where there is an intended gathering, and a threat of violence, the content of the negotiations and consultations with the local authorities will deal with the potential for injury to persons or damage to property. It argued that in these circumstances organizations will almost always foresee the possibility of damage arising from a gathering and thus be held liable. For this reason, it concluded that section 11(2) does not provide a viable defence to a defendant who faces a claim for riot damage as a result of which organizations are exposed to the spectre of extensive liability. This would, in their view, discourage organizations from holding gatherings and demonstrations as a result of a chilling effect it has on the exercise of the right to assemble.
The respondents and the Minister argued that the right to freedom of assembly is conditional upon its being exercised peacefully and unarmed. Where riot damage occurs it would, in their view, mean that the gathering was not peaceful and therefore section 17 of the Constitution would not be implicated. They further contended that SATAWU had placed no evidence before the Court to prove that section 11(2) has a chilling effect on the exercise of the right, whereas the respondents had provided evidence to the contrary.
The High Court held that section 17 is not implicated by section 11(2)(b) because the right to freedom of assembly does not extend to gatherings which are not peaceful and section 11(2)(b) does not have a chilling effect on the exercise of the right. It also held that, even if section 11(2)(b) does limit the right to freedom of assembly, this limitation must be balanced against the rights of individual members of the public to dignity, freedom from violence and arbitrary deprivation of property, all of which are affected by riot damage. The Court concluded that the limitation is reasonable and justifiable in an open and democratic society based on human dignity, equality and freedom.
Aggrieved by this outcome, SATAWU appealed to the Supreme Court of Appeal.
In the Supreme Court of Appeal
The Supreme Court of Appeal dismissed the appeal on the ground that the statutory defence provided for in section 11(2) against a claim for riot damage is not illusory but real and capable of being proved. In support of this conclusion, the Court cited a number of hypothetical scenarios to illustrate the circumstances in which the defence could be raised successfully. It then reasoned:
Even though the conjunctive nature of the defence set out in s 11(2)(b) of the Act, on the face of it, seems burdensome one can only take reasonable steps in respect of conduct that is reasonably foreseeable. It does appear that unless the act complained of ─ leading to the riot ─ was reasonably foreseeable, a defendant would probably in all of the instances set out above escape liability. One can only take steps to guard against an occurrence if one can foresee it.
The Court also held that the scope of the right to freedom of assembly does not extend to persons who assemble in a manner that is not peaceful or unarmed. It found that the scheme of the Act, including section 11, is aimed at restricting unlawful, violent behaviour that violates the rights of others and ensuring that organizers of those gatherings are held liable.
The contention that section 11(2)(b) has a chilling effect on the exercise of the right to freedom of assembly was rejected and found to be unsubstantiated and inconsistent with the unchallenged evidence put forward by the respondents. It was concluded that any chilling effect section 11 has is on unlawful behaviour, which is not protected by the right.
In this Court
The applicants contended in this Court that the Supreme Court of Appeal was incorrect, that section 11(2) is contradictory and irrational, that it limits the right to freedom of assembly in section 17 of the Constitution and that the limitation is not justifiable.
There are three preliminary issues:
Condonation for the late filing of the application for leave to appeal.
COSATU’s application for leave to intervene in these proceedings.
The application for leave to appeal.
The issues of substance that arise are whether—
the words “and was not reasonably foreseeable” cause section 11(2)(b) to be internally inconsistent and irrational, thus rendering the section constitutionally invalid;12 and if not,
section 11(2) limits the right to freedom of assembly; and if so,
the limitation is justifiable.
The applicants seek condonation for the late filing of the application for leave to appeal, which they filed about three weeks late. The explanation for the delay is that SATAWU had to consult with its mother-body, COSATU, to decide on the best joint approach to the matter. As a prerequisite to finalising those discussions, COSATU also needed a resolution from its Central Executive Committee in relation to the application for leave to intervene and the best approach to the issues. Finally, the applicants submit that the respondents have suffered no material prejudice as a result of the delay.
The main question is whether it is in the interests of justice to grant condonation. I am satisfied that the delay is short, the explanation is adequate, the respondents suffered no prejudice and the question whether the Act limits the right to freedom of assembly is an important constitutional issue. It is thus in the interests of justice to grant condonation.
Leave to intervene
COSATU seeks leave to intervene on the basis that it has a direct and substantial interest in the outcome of this matter. This interest is said to lie in the impact of the Supreme Court of Appeal judgment on the federation’s ability to exercise the right to assemble peacefully and unarmed. The judgment, they say, has a materially limiting effect on their exercise of the right to assemble.
I am satisfied that COSATU has a direct and substantial interest in this matter and am therefore inclined to grant leave to intervene in these proceedings.
Leave to appeal
Two requirements must be satisfied before an application for leave to appeal to this Court may be granted. The application must raise a constitutional issue and it must be in the interests of justice to hear the matter.
This matter is about whether a statute that regulates public gatherings, by imposing liability for riot damage arising out of a gathering and then creating a defence to that liability that is narrower than would be available under the common law, imposes an unconstitutional limitation on the exercise of the right to freedom of assembly. It also concerns the rationality of this defence. And rationality is an incident of the principle of legality, which is a requirement of the rule of law.
The exercise of the right to assemble by trade unions and other organizations is an important constitutional issue. The riot damage allegedly caused by the gathering, which is said to have affected vulnerable people in the business sector, underscores the public interest in the matter. This judgment will have significant implications for the exercise of the right to assemble, not only for the applicants, but also for the public at large. The applicants have an arguable case and therefore have some prospects of success on appeal. It is thus in the interests of justice that leave to appeal be granted.
The meaning of section 11(2) and its alleged irrationality
The applicants do not challenge the constitutional validity of section 11(1) insofar as it provides that organizers of gatherings should be held liable for the riot damage resulting from their gatherings. It is the ambit of that liability as constrained by section 11(2) that they are particularly concerned about. The applicants regard the degree of that constraint as inadequate. They posit that the Constitution entitles them to a wider defence.
The first contention is that the defence is inconsistent with the principle of legality, in that it is irrational. The irrationality is said to lie in the plain impossibility of expecting the organization to take all reasonable steps to prevent a specific act or omission, even when that act or omission was not reasonably foreseeable. The argument appears to be that section 11(2)(c) can never find application. If the act or omission is reasonably foreseeable, then liability arises regardless of the steps taken. If the harm is not reasonably foreseeable, then there are no reasonable steps that can be taken to avoid it. In either case, section 11(2)(c) plays no role.
To decide this issue, it is necessary to examine section 11(2) and determine its meaning. The section is repeated for convenience:
It shall be a defence to a claim against a person or organization contemplated in subsection (1) if such a person or organization proves—
This Court has previously held that an interpretation of a statutory provision that gives rise to an absurdity or irrationality should be avoided where there is another reasonable construction which may be given to that provision. In other words, where a legislative provision is reasonably capable of a meaning that keeps it within constitutional bounds, a court must, through the use of legitimate interpretive aids, seek to preserve that provision’s constitutional validity. Thus, to the extent that it is possible, section 11(2) must be interpreted in a manner that yields a rational meaning, and preserves its validity so that the purpose it was enacted to serve is realised.
The somewhat unusual defence created for an organization facing a claim for statutory liability appears to have been made deliberately tight. Gatherings, by their very nature, do not always lend themselves to easy management. They call for extraordinary measures to curb potential harm. The approach adopted by Parliament appears to be that, except in the limited circumstances defined, organizations must live with the consequences of their actions, with the result that harm triggered by their decision to organise a gathering would be placed at their doorsteps. This appears to be the broad objective sought to be achieved by Parliament through section 11. The common law position was well known when section 11 was enacted. The limitations of a delictual claim for gatherings-related damage in meeting the policy objective gave rise to the need to enact section 11 to make adequate provision for dealing with the gatherings-related challenges of our times.
Parliament sought to ameliorate the impact of imposing liability on an organizer by providing for a viable, yet onerous, defence under section 11(2). The purpose was:
to provide for the statutory liability of organizations, so as to avoid the common law difficulties associated with proving the existence of a legal duty on the organization to avoid harm;
to afford the organizer a tighter defence, allowing it to rely on the absence of reasonable foreseeability and the taking of reasonable steps as a defence to the imposition of liability; and
to place the onus on the defendant to prove this defence, instead of requiring the plaintiff to demonstrate the defendant’s wrongdoing and fault.
In the context of what has been said in relation to the purpose of the provision and for the reasons that follow, the word “and” between subsections (b) and (c) of section 11(2) must be given its ordinary meaning. As illustrated by the purpose of the section set out above and the ensuing discussion, none of this yields an irrational outcome.
An organization will escape liability only if the act or omission that caused the damage was not reasonably foreseeable, and if it took reasonable steps within its power to prevent that act or omission. Subsections (a), (b) and (c) of section 11(2) must be read together. They all refer to the same act or omission that causes the damage, which must not be reasonably foreseeable, and in respect of which reasonable preventive steps within the power of the organization must be taken, when reasonably foreseeable, if the organization is to escape liability.
The applicants’ major contention is that any reasonable organizer who takes reasonable steps to guard against an act or omission materialising could never prove that it was not reasonably foreseeable. In these circumstances, the applicants say that it would be quite impossible to come within the purview of section 11(2) and that any organization would automatically be liable. The submission overlooks the interrelationship between subsections 11(2)(b) and (c), and proceeds from a conceptualisation of foreseeability that is grounded in the law of delict and not necessarily in the wording of the section itself.
There is an inter-relationship between the steps that are taken by an organizer on the one hand and what is reasonably foreseeable on the other. The section requires that reasonable steps within the power of the organizer must be taken to prevent the act or omission that is reasonably foreseeable. The real link between the foreseeability and the steps taken is that the steps must prove to have been reasonable to prevent what was foreseeable. If the steps taken at the time of planning the gathering are indeed reasonable to prevent what was foreseeable, the taking of these preventive steps would render that act or omission that subsequently caused riot damage reasonably unforeseeable. Both section 11(2)(b) and section 11(2)(c) would then have been fulfilled.
It must be emphasised that organizations are required to be alive to the possibility of damage and to cater for it from the beginning of the planning of the protest action until the end of the protest action. At every stage in the process of planning, and during the gathering, organizers must always be satisfied of two things: that an act or omission causing damage is not reasonably foreseeable and that reasonable steps are continuously taken to ensure that the act or omission that becomes reasonably foreseeable is prevented. This is the only way in which organizers can create a situation where acts or omissions causing damage remain unforeseeable. In such a case, the requirement of taking reasonable steps is not met simply by guarding against the occurrence of the damage-causing act or omission. The inquiry whether the steps taken were sufficient to render the act or omission in question no longer reasonably foreseeable might be very exacting.
An important qualification is that the steps that the organizers are required to take must be within their power. Where steps need to be taken that are not within their power, they must ensure that those who have the duty to take steps are notified of the need to do so. What reasonable steps within their power the organizers must take after this depends on the reaction to this notification. A defendant does, therefore, have a viable defence in terms of section 11(2).
To illustrate that the defence is viable, the Supreme Court of Appeal cited three examples of circumstances in which organizers would not be liable where there had been proper planning in the sense that all reasonable steps had been taken to prevent the foreseeable. The first concerned a policeman who mistakenly discharged his weapon causing a panic and stampede of participants; the second envisioned the sudden and unexpected infiltration of a gathering by a gunman, unconnected to the organizers, who fired indiscriminately because of a grudge against society; and the third involved a motorist who broke through barricades and drove into the marching crowd causing panic and a riot.
The applicants contend that these examples relate to acts and omissions that can only be classified as not reasonably foreseeable. They emphasise that their concern is in relation to foreseeable conduct that takes place despite the fact that they have taken reasonable steps to avert the danger of its occurrence. The answer to this contention is that the organizers are obliged to take reasonable steps to avert the reasonably foreseeable harm-causing act or omission to render that act or omission no longer reasonably foreseeable. Let us take the example of the car crashing through the barrier postulated by the Supreme Court of Appeal. The possibility of cars colliding with pedestrians during a march would have been reasonably foreseeable by any organization. If that organization had not provided barriers and a motor vehicle collided with the pedestrian either because the pedestrian strayed onto the road or because the vehicle veered slightly off its path, the organizer would be liable. The erection of the barriers would render a collision between a vehicle and a pedestrian no longer reasonably foreseeable. Put another way: reasonable preventive steps in relation to an initially foreseeable act or omission eventually would make the act or omission no longer reasonably foreseeable. If an act or omission were not at all reasonably foreseeable in advance, then taking no steps to guard against it will ordinarily be reasonable.
This is not the way in which foreseeable harm and reasonable steps to prevent it are expressed in the law of delict. Under the Aquilian action, when deciding what reasonable steps a person must take in relation to foreseeable harm there are two variables relevant to this case, the seriousness of the harm and the prospect of it happening. In Herschel v Mrupe, it was stated that—
the circumstances may be such that a reasonable man would foresee the possibility of harm but would nevertheless consider that the slightness of the chance that the risk would turn into actual harm, correlated with the probable lack of seriousness if it did, would require no precautionary action on his part.”
And in Ngubane v South African Transport Services, it is put thus:
[I]t is acknowledged that reasonable steps are not necessarily those which would ensure that foreseeable harm of any kind does not in any circumstances eventuate.
We are dealing with a statutory defence that must be given a rational meaning and this meaning does not have to equate to that of its delictual counterpart. Of course, that also means that the meaning given to reasonable foreseeability and the taking of reasonable steps in section 11(2)(b) and (c) should not be transposed unreflectively to the common law of delict.
In the light of what I consider to be the proper interpretation of section 11(2), set out above, the section is rational. The next question is whether or not this section limits the right to freedom of assembly.
Does section 11(2) limit the right to assembly?
“Everyone has the right, peacefully and unarmed, to assemble, to demonstrate, to picket and to present petitions.” That is what section 17 of the Constitution promises the people in South Africa.
This means that everyone who is unarmed has the right to go out and assemble with others to demonstrate, picket and present petitions to others for any lawful purpose. The wording is generous. It would need some particularly compelling context to interpret this provision as actually meaning less than its wording promises. There is, however, nothing, in our own history or internationally, that justifies taking away that promise.
Nothing said thus far detracts from the requirement that the right in section 17 must be exercised peacefully. And it is important to emphasise that it is the holders of the right who must assemble and demonstrate peacefully. It is only when they have no intention of acting peacefully that they lose their constitutional protection. This proposition has support internationally. As the European Court of Human Rights noted:
[A]n individual does not cease to enjoy the right to peaceful assembly as a result of sporadic violence or other punishable acts committed by others in the course of the demonstration, if the individual in question remains peaceful in his or her own intentions or behaviour.
This means that it is appropriate to proceed on the basis that section 17 of the Constitution means what it generously says.
The long title of the Act states that the purpose of the Act is “[t]o regulate the holding of public gatherings and demonstrations at certain places; and to provide for matters connected therewith.” The Act requires the appointment of persons responsible for giving and receiving notices to hold gatherings and to act at consultations or negotiations in relation to the holding of gatherings on behalf of the organizers, the police and the local authority involved.
The mere legislative regulation of gatherings to facilitate the enjoyment of the right to assemble peacefully and unarmed, demonstrate, picket and petition may not in itself be a limitation. Section 11(2), read with section 11(1), goes further than simply to regulate the exercise of the right in order to facilitate its full and appropriate enjoyment by those who organise and those who participate.
Section 11(1) holds organizers of a gathering liable for riot damage subject to section 11(2), which provides a limited defence to a claim of this kind. The effect of these specific provisions, in the context of the Act as a whole, is to render holders of a gathering organised with peaceful intent liable for riot damage on a wider basis than is provided for under the law of delict. This is all the more so given the extremely wide definition of riot damage in the Act. This means that proof of liability will, as indicated earlier, be easier in a large number of cases.
Compliance with the requirements of section 11(2) significantly increases the costs of organising protest action. And it may well be that poorly resourced organizations that wish to organise protest action about controversial causes that are nonetheless vital to society could be inhibited from doing so. Both these factors amount to a limitation of the right to gather and protest.
It must be emphasised that it is not the right of organizations alone that is affected. It is quite plausible that the organizer of a gathering who anticipates the involvement of, say, ten thousand people may be forced to cancel it because a few hundred of the participants would cause mayhem. In these circumstances, the right of thousands of people to protest peacefully and unarmed is affected.
It is true that the increase in costs and the wider basis on which there is civil liability will render organizations more reluctant to organise marches. But this is better dealt with in the section concerned with the extent of the limitation in the justification analysis to which I now turn.
It falls to be determined whether the limitation is constitutionally justifiable. To do so, regard must be had to section 36 of the Constitution, which provides:
A. The importance of the right to assemble
The right to freedom of assembly is central to our constitutional democracy. It exists primarily to give a voice to the powerless. This includes groups that do not have political or economic power, and other vulnerable persons. It provides an outlet for their frustrations. This right will, in many cases, be the only mechanism available to them to express their legitimate concerns. Indeed, it is one of the principal means by which ordinary people can meaningfully contribute to the constitutional objective of advancing human rights and freedoms. This is only too evident from the brutal denial of this right and all the consequences flowing therefrom under apartheid. In assessing the nature and importance of the right, we cannot therefore ignore its foundational relevance to the exercise and achievement of all other rights.
Under apartheid, the state took numerous legislative steps to regulate strictly and ban public assembly and protest. Despite these measures, total repression of freedom of expression through protest and demonstration was not achieved. Spontaneous and organised protest and demonstration were important ways in which the excluded and marginalised majority of this country expressed themselves against the apartheid system, and was part and parcel of the fabric of the participatory democracy to which they aspired and for which they fought.
So the lessons of our history, which inform the right to peaceful assembly and demonstration in the Constitution, are at least twofold. First, they remind us that ours is a “never again” Constitution: never again will we allow the right of ordinary people to freedom in all its forms to be taken away. Second, they tell us something about the inherent power and value of freedom of assembly and demonstration, as a tool of democracy often used by people who do not necessarily have other means of making their democratic rights count. Both these historical considerations emphasise the importance of the right.
There is also international support for this. The Special Rapporteur on extrajudicial, summary or arbitrary executions, Christof Heyns stated:
Supporting freedom of assembly implies a realisation that, as expressed so eloquently by the Spanish Constitutional Court, ‘in a democratic society, the urban space is not only an area for circulation, but also for participation’.
Moreover, the inter-relation between the various freedom rights and their importance to our democracy has been recognised by this Court. In South African National Defence Union v Minister of Defence, O’Regan J, writing for the Court, stated:
[Freedom of speech] is closely related to freedom of religion, belief and opinion (s 15), the right to dignity (s 10), as well as the right to freedom of association (s 18), the right to vote and to stand for public office (s 19) and the right to assembly (s 17). These rights taken together protect the rights of individuals not only individually to form and express opinions, of whatever nature, but to establish associations and groups of like-minded people to foster and propagate such opinions. The rights implicitly recognise the importance, both for a democratic society and for individuals personally, of the ability to form and express opinions, whether individually or collectively, even where those views are controversial.
In S v Mamabolo (E TV and Others Intervening) Kriegler J stated:
That freedom to speak one’s mind is now an inherent quality of the type of society contemplated by the Constitution as a whole and is specifically promoted by the freedoms of conscience, expression, assembly, association and political participation protected by ss 15-19 of the Bill of Rights. It is the right – idealists would say the duty – of every member of civil society to be interested in and concerned about public affairs.
Freedom of assembly is no doubt a very important right in any democratic society. Its exercise may not, therefore, be limited without good reason. The purpose sought to be achieved through the limitation must be sufficiently important to warrant the limitation.
B. The importance of the purpose of the limitation
The purpose of the limitation imposed by section 11 is very important. It is to protect members of society, including those who do not have the resources or capability to identify and pursue the perpetrators of the riot damage for which they seek compensation. When a gathering imperils the physical integrity, the lives and the sources of livelihood of the vulnerable, liability for damages arising therefrom must be borne by the organizations that are responsible for setting in motion the events which gave rise to the suffered loss. And that is what this important limitation is designed to achieve.
The fact that every right must be exercised with due regard to the rights of others cannot be overemphasised. The organization always has a choice between exercising the right to assemble and cancelling the gathering in the light of the reasonably foreseeable damage. By contrast, the victims of riot damage do not have any choice in relation to what happens to them or their belongings. For this reason, the decision to exercise the right to assemble is one that only the organization may take. This must always be done with the consciousness of any foreseeable harm that may befall others as a consequence of the gathering. The organizers must therefore always reflect on and reconcile themselves with the risk of a violation of the rights of innocent bystanders which could result from forging ahead with the gathering.
C. The nature and extent of the limitation
Whilst the Act does have a chilling effect on the exercise of the right, this should not be overstated. The Act does not negate the right to freedom of assembly, but merely subjects the exercise of that right to strict conditions, in a way designed to moderate or prevent damage to property or injury to people. Potentially, the exercise of the right also occasions deterrent consequences. One of them is the presumption of liability for riot damage, which can be traced back to the organization’s decision to exercise the right to assemble.
The effect of section 11 is thus to place the organizers in the first line of fire when riot damage occurs. The innocent victim need not look any further than the organizers for compensation. She does not need to prove negligence on their part. In this sense, the liability may be considered to be ‘strict’.
D. Apportionment of Damages
To make liability for riot damage less onerous, section 11 provides for apportionment of damages. That liability is, in terms of section 11(1), joint and several in that the organization is—
a joint wrongdoer contemplated in Chapter II of the Apportionment of Damages Act, 1956 (Act No. 34 of 1956), together with any other person who unlawfully caused or contributed to such riot damage and any other organization or person who is liable therefor in terms of this subsection.
The organizer’s right of recourse or contribution is further entrenched by the provisions of section 11(3) of the Act.
Is this legislative choice of making immediate and less restrictive amends for victims possible against organizations, and leaving recourse against other wrongdoers to the organizations, unreasonable? I think not. The organizations are intimately involved in the planning, supervision and execution of the gathering but the potential victims are not. Because of this the organizations would be in a better position than innocent victims to identify individuals or institutions that caused the damage. Although the primary liability is imposed on the organization, a soft landing is availed to it through the possibility of an apportionment of damages, as it is always open to the organization to track down the perpetrators and recoup its loss from them.
It is thus not unreasonable to allow the victim of riot damage to claim all compensation from the organizers of a gathering and then to leave it to the organizers to seek recourse or contribution from others who may have caused or contributed to the damage. It is, in reality, also the norm in our law of delict, as shaped by the Apportionment of Damages Act (Apportionment Act).
It is argued, however, that the organizers’ right of recourse or contribution is illusory, because the Apportionment Act is based on an apportionment of fault, in particular negligence, and is thus inapplicable to cases of strict statutory liability. It is true that case law under the Apportionment Act is to that effect in respect of apportionment of liability under section 1 of the Act. Commentators have, however, pointed out that Chapter II of the Act, dealing with joint and several wrongdoers, applies more generally to liability in delict, rather than to damages caused “by the fault” of someone, as is required by section 1. Earlier academic commentators emphasised that section 2 applies even to delicts that do not require fault, like the pauperien action, vicarious liability and breach of a statutory duty, whilst later writers tend to focus more on vicarious liability.
The difficulty cannot in my view be resolved only on these differences in the wording of the Apportionment Act. It is indeed correct that section 2(1) does not refer to fault as a requirement, but in the subsections dealing with recourse between wrongdoers, the apportionment is on the basis of an amount that the court “may deem just and equitable having regard to the degree in which that other joint wrongdoer was at fault in relation to the damage suffered by the plaintiff” (my emphasis). A textual argument can thus be made either way. But it is not to the Apportionment Act that we must look to determine whether the right of recourse or contribution is created. It is to the Act itself. And when one does that, the problem disappears.
The liability created by the Act is not liability where the organization’s own ‘fault’ is irrelevant. Fault is always relevant where an organization relies on section 11(2) as a defence, since section 11(2)(c) requires it to prove that it “took all reasonable steps within [its] power to prevent the act or omission in question”. In each case where it is able to prove this, and that the act or omission was not reasonably foreseeable, it would not be liable. But, where it is liable, its liability is not exclusive.
In that case recourse against, or contribution from, others would arise, for where the section 11(2) defence is raised, recourse and contribution would be relevant. In each case the organization’s degree of fault under section 11(2)(c) would have been in issue at the trial and determined by the judicial officer. To determine recourse or contribution against other wrongdoers, what would be left to establish would be the degree of fault of those wrongdoers in relation to the damages suffered by the plaintiff. And when that happens, the requirements of the Apportionment Act too would have been met.
Recourse and contribution under section 11(1) and (3) of the Act are thus not illusory. Nor is it unreasonable or unfair as between joint wrongdoers. Their eventual contributions would still be determined by their respective degrees of “fault” under the Apportionment Act. And as for making it easier for the innocent victim of riot damage to make her claim against the organizers, that is not unreasonable, because the organizers are not left without recourse against other joint wrongdoers.
The applicants submit that placing the onus on the organizers to prove the statutory defence renders the limitation unjustifiable. I disagree. If victims of riot damage were required to prove the elements section 11(2) obliges the organizers to establish, their position would be difficult. The facts are usually within the knowledge of the organizers. Victims’ damages claims could otherwise be rendered illusory.
E. The balance between the limitation and the purpose and less restrictive means
The purpose of the section is to ensure that a gathering that becomes destructive and results in loss to others does not leave its victims without recourse. It is thus to protect the rights of individuals who may be affected detrimentally by riot damage that takes place in the course of the exercise of the right to assemble.
There is a tight fit between the limitation and its purpose. The purpose is to achieve an appropriate balance between the right to assemble on the one hand and the safety of people and property on the other. That balance has been struck.
In assessing whether less restrictive means exist to achieve the purpose of the Act, I am mindful of the position adopted by this Court in Mamabolo, where it was stated:
Where s 36(1)(e) speaks of less restrictive means it does not postulate an unattainable norm of perfection. The standard is reasonableness. And, in any event, in theory less restrictive means can almost invariably be imagined without necessarily precluding a finding of justification under the section. It is but one of the enumerated considerations which have to be weighed in conjunction with one another, and with any others that may be relevant.
In view of the special nature and legitimacy of the purpose the Act seeks to achieve, and the fine balance that it seeks to strike between the conflicting rights and interests of organizations and members of the public, it cannot be said that any less restrictive means is available.
The limitation on the right to assemble is reasonable and justifiable in an open and democratic society based on human dignity, equality and freedom.
The respondents have succeeded in this Court. There is no reason to deny them costs, including the costs of two counsel.
In the result, the following order is made:
Condonation is granted.
Leave to intervene is granted.
Leave to appeal is granted.
The appeal is dismissed.
The applicants are ordered to pay the costs of the first to eighth respondents in this Court, including costs of two counsel.
(Zondo AJ concurring)
Does section 11(2) of the Regulation of Gatherings Act limit the rights in section 17 of the Constitution? If so, is the limitation justifiable? These are the questions which arise for determination in this matter. The second question will, however, be reached only if the answer to the first question is a positive one. The matter comes before this Court as an application for leave to appeal against the judgment of the Supreme Court of Appeal which dismissed an appeal against an order of the Western Cape High Court.
I have had the benefit of reading the judgment of Mogoeng CJ which holds that section 11(2) of the Act limits the rights in section 17 of the Constitution and that the limitation is justifiable. While I agree that condonation, the application to intervene, and leave to appeal should be granted, and that the appeal be dismissed, I respectfully disagree that the limitation contended for has been established. In my view, section 11(2) does not implicate nor does it limit any of the rights entrenched in section 17 of the Constitution.
The factual and statutory background
On 16 May 2006, the South African Transport and Allied Workers Union (SATAWU) organised a protest march in the centre of Cape Town (City). Participants in the march were members of SATAWU, who were engaged in a protracted labour strike that had turned violent. The acrimonious strike had led to the deaths of about 50 people before the protest march. The strike generated a highly volatile atmosphere which could turn violent upon the slightest provocation.
As required by the Act, SATAWU had to give notice to the City, notifying it about the march it intended to hold. Section 3 of the Act requires that notice be given at least seven days before the date on which a gathering is to be held. The Act defines a gathering to include “any assembly, concourse or procession of more than 15 persons in or on any public road”. However, if it is not possible to give notice seven days in advance, it may be given at least 48 hours before the commencement of the gathering. The notice must contain, amongst other details listed in the section, the name and contact details of the convener; the purpose of the gathering; the place where it is to be held; the time, duration and date of the gathering; and the route of the procession.
Upon receipt of the notice, a local authority must consult a relevant member of the South African Police Service, regarding the necessity to hold negotiations on the proposed gathering. If the relevant member indicates that negotiations are necessary, a meeting is convened between the local authority, the police and the convener of the gathering. The purpose of the meeting is to discuss and agree terms and conditions under which the gathering will be held.
If information comes to the attention of the local authority that the gathering is likely to cause serious disruption of traffic, injury to participants of the gathering or other people, or extensive damage to property, the local authority should meet with the convener and other relevant parties to discuss steps to be taken to avoid harm or disruptions. If not convinced that steps proposed to prevent harm will succeed, the local authority may prohibit the gathering. Where a gathering takes place and the police become convinced that it will result in harm to either participants or other people and that adequate protection is not possible, the police may demand that participants disperse.
But if a gathering takes place, and during the course of it damage is caused as a result of injury to a person or destruction to property, the Act imposes liability for such damage on the convener of the gathering. The convener may avoid liability only if it establishes the defence set out in section 11(2).
Section 11(1) of the Act provides:
If any riot damage occurs as a result of—
Although the section is not a model of clarity, its reading indicates that liability for damage caused as a result of a gathering is imposed on the convener of the gathering or an organisation on behalf of whom the gathering was held. The convener is held jointly and severally liable together with the person who caused or contributed to the damage. What is striking is the fact that “riot damage” is defined in wide terms and that liability does not arise in respect of violent and riotous gatherings only. Even a peaceful gathering does give rise to a claim against the convener. This is so because “riot damage” in the Act means “any loss suffered as a result of any injury to or the death of any person, or any damage to or destruction of any property, caused directly or indirectly by, and immediately before, during or after, the holding of a gathering.” It matters not that the gathering was peaceful or violent. As long as damage occurs as a result of the gathering the convener is liable, whether that damage was caused before or after the gathering or even if the damage was not directly caused by the gathering.
SATAWU complied with the Act by giving notice of the march to be held on 16 May 2006 to the City. This led to negotiations between the City, the police and SATAWU, taking place on 10 May 2006. The parties agreed on the route to be followed during the march and certain conditions were also imposed. Some of those conditions were that no undue obstruction would be caused to pedestrians or vehicular traffic; every precaution would be taken to ensure the safety and protection of the public, traffic and participants; and SATAWU would appoint sufficient marshals to ensure good order and compliance with conditions of the march. SATAWU acknowledged that it understood the provisions of the Act.
The negotiations and the conditions imposed were triggered by information that the march might result in damage being caused. The violence perpetrated in relation to the strike had already caused the deaths of a number of people. Apparently the City thought that SATAWU could take adequate steps to prevent damage, hence it did not prohibit the march. Indeed the march took place on 16 May 2006. Regrettably, it degenerated into chaos and extensive damage was caused to vehicles and other property along the route. Street vendors were robbed of their stock and quite a number of people suffered damage as a result of the march. To obtain compensation for their loss, the victims of the march instituted an action in the High Court.
In the High Court
The first to eighth respondents based their claim against SATAWU on section 11(1) of the Act. Apart from denying liability, SATAWU pleaded that should it be found liable, the words “and was not reasonably foreseeable” in section 11(2)(b) of the Act must be declared inconsistent with the Constitution and invalid.
Following the claim for constitutional invalidity, SATAWU lodged a notice in which this claim was raised. This is a procedural step required by the High Court rules. The notice articulates the constitutional challenge in these terms:
Upon application by the parties the High Court separated the constitutional challenge from the other issues and the determination of those issues was stood over to a later date. The Court considered the claim for constitutional invalidity. It held that section 11(2) does not limit any of the rights in section 17 of the Constitution. Based on the facts of the case, the Court held that SATAWU could not rely on section 17 because that section guarantees the right to assemble peacefully and unarmed whereas the present march was violent. The violent nature of the march, held the Court, placed it outside the ambit of assemblies protected by section 17.
In the event that this finding was held to be wrong, the High Court proceeded to consider whether the limitation is justifiable in terms of section 36 of the Constitution. It embarked on a justification analysis, during the course of which various factors were taken into account. It concluded that the limitation is justifiable.
In the Supreme Court of Appeal
Unhappy with the decision of the High Court, SATAWU appealed to the Supreme Court of Appeal. In that Court, SATAWU argued that if section 11(2)(b) could be read intelligibly to afford it a real defence to the claim, it would be unnecessary to consider whether the limitation was justifiable because in that event section 11(2) would constitute a reasonable and justifiable limitation. This is a novel approach to the determination of constitutional invalidity where the challenge is based on the assertion that an impugned provision limits a right in the Bill of Rights.
Acceding to the invitation, the Supreme Court of Appeal proceeded to interpret section 11(2) and held that despite the conjunctive nature of the defence it postulates, the section means that a defendant would escape liability under the Act unless the act giving rise to damage was reasonably foreseeable and the defendant has failed to take reasonable steps to prevent harm because “[o]ne can only take steps to guard against an occurrence if one can foresee it.” The appeal was dismissed.
In this Court
The question whether section 11(2) of the Act is invalid by reason of being inconsistent with section 17 of the Constitution must be determined with reference and be confined to the case pleaded by SATAWU. It is not proper to proceed on the footing that in general the Act constitutes a limitation to the right of assembly. The attack is not directed at the Act in its entirety nor is it a general challenge. By contrast, the attack is sharp and narrow. It is confined to section 11(2) and nothing more. Because section 11(1) was left out of the attack in circumstances where SATAWU was aware that it is this section that imposes liability, it is incorrect to read section 11(1) together with section 11(2) for purposes of determining whether the latter section is inconsistent with section 17 of the Constitution or not.
This Court has warned parties who challenge the validity of Acts of Parliament to plead their cases accurately, particularly, as in this case, where the constitutional challenge is raised through a Rule 16(A) notice. In Shaik v Minister of Justice and Constitutional Development (Shaik) this Court had to consider whether to grant leave to appeal against a judgment of the High Court declining to declare section 28(6) of the National Prosecuting Authority Act inconsistent with section 35 of the Constitution and for that reason invalid. The High Court had held that the words “any person” in the impugned section did not refer to an accused person. While that Court held that the impugned section limited the right to remain silent, it held that the limitation was justified.
The main reason for this Court to refuse leave was that the applicant, in his Rule 16A notice, had targeted the wrong subsection even though the basis of the attack was clear from the papers. The complaint was that “the s 28 procedure empowers the prosecuting authority to require a suspect to answer questions without giving the suspect full immunity from the consequences of such answers”. The similarities between the Rule 16A notice in that case and the present notice are striking. There the notice read:
Comparing the two notices reveals that both single out a specific subsection for attack. Each notice asserts that the impugned provision limits a specified right in the Bill of Rights. The sting of the attack in both cases is outside the impugned subsection but in other parts of the same section which are not mentioned in the notice.
This Court construed the notice in Shaik as excluding an attack on section 28(8) and (10) of the same Act. And because the notice made no reference to these subsections, the Court refused to grant leave directed at challenging the omitted subsections even though the parties themselves had understood the real challenge to have been directed at the omitted subsections and argument had been presented which covered them. The Court said:
The kernel of the applicant’s attack throughout has been that the section 28 procedure empowers the prosecuting authority to require a suspect to answer questions without giving the suspect full immunity from the consequences of such answers. This attack has been based on s 35 of the Constitution and has focused exclusively on subs 28(6) of the Act. Section 28(6) is, however, the wrong provision to target. It does no more than describe the Investigating Director’s powers and says nothing about the obligations of the examinee. It neither compels the examinee to heed the summons nor to answer any questions, nor does it stipulate what questions the examinee is obliged to answer, nor what use may be made of any answer, nor what the consequences might be if the examinee should fail or refuse to answer any question. The sting of the section – for purposes of the s 35 attack – is found in s 28(8) and (10). The punishment for the offence created by s 28(10) is not prescribed in the Act and, accordingly, the general enabling provisions of s 276 of the CPA – that empowers, amongst other things, the imposition of imprisonment – apply.
The compulsion to attend, to be sworn in or to make an affirmation, and to answer questions fully, are all stipulated in ss (10). The extent of examinees’ privilege to refuse to answer questions, and the manner and extent to which answers – that examinees are obliged to give – may subsequently be used against them, are detailed in ss (8). Indeed, the constitutional attack in the High Court and this Court focused on the alleged constitutional inadequacy of the direct use immunity provided for in ss (8)(b).
The minds of litigants (and in particular practitioners) in the High Courts are focused on the need for specificity by the provisions of Uniform Rule 16A(1). The purpose of the Rule is to bring [the case] to the attention of persons (who may be affected by or have a legitimate interest in the case) the particularity of the constitutional challenge, in order that they may take steps to protect their interests. This is especially important in those cases where a party may wish to justify a limitation of a chap 2 right and adduce evidence in support thereof.
It constitutes sound discipline in constitutional litigation to require accuracy in the identification of statutory provisions that are attacked on the ground of their constitutional invalidity.”
[emphasis added and footnotes omitted]
If it was proper to incorporate the omitted subsections into the attack, because the real complaint was located in them and section 28 was to be read as a whole, this Court could not have made the statements quoted above. Moreover it is clear from those statements that the Court regarded the mentioning of section 28(6) only in the Rule 16A notice to constitute an exclusive focus on it which precluded any reference to section 28(8) and (10). By the same token the mentioning of section 11(2) only in the present notice must mean that the challenge focused exclusively on this subsection. There is a difference between the principle that requires the whole section to be read for purposes of interpretation, on the one hand, and the process of determining the scope of a constitutional attack, on the other. The scope of the challenge is not determined with reference to the entire section where only one subsection is targeted. Shaik makes this plain.
In Shaik, this Court having laid down the principle that a constitutional challenge must be accurately worded, stated its reasons for refusing leave thus:
The wrong provision in the Act has been targeted for constitutional attack. The potential ambit of s 28 has been misunderstood, with the attendant consequences referred to above.... Under all these circumstances, it is not in the interests of justice to grant leave to appeal in which the thrust of the constitutional attack is not in substance against s 28(6) but against s 28(8) and (10).
It is significant, however, to note that this Court refused leave in Shaik despite the fact that it had serious concerns over the constitutional validity of section 28, albeit in respect of an issue that was not properly raised in the pleadings but still related to the section 28 procedure. The Court stated:
There is a concern about the constitutional validity of ss 28(6), (8) and (10) of the Act. It was not formally raised or dealt with in argument as a ground for attack under s 35 of the Constitution. While refraining from pronouncing on it, the Court cannot allow the concern to pass unmentioned. It relates to the fact that, under s 28(6)(b), the ‘Investigating Director or a person designated by him’ questions the person summoned under oath or affirmation, without the necessity of any other person being present, let alone a person who is independent of the Directorate of Special Operations.
The s 28 procedure raises the spectre of the interrogator and interrogatee alone in one room for days, the former asking the questions and making the record, the latter simply answering questions.
On this aspect the Court held that it was not in the interests of justice to determine the constitutional validity of section 28, based on a ground that was not pleaded even though it had serious concerns over the validity of the section.
In this case, unlike in Shaik where the applicant sought to expand his challenge by including section 28(8) and (10), SATAWU did not seek to impugn the validity of section 11(1). Reading section 11(1) together with the impugned provision will not only be inconsistent with Shaik but will also be tantamount to subjecting section 11(1) to an indirect or a collateral challenge in circumstances where no attack was raised against it. This approach is impermissible. Thus in Phillips v National Director of Public Prosecutions this Court held:
It is not ordinarily permissible to attack statutes collaterally. The constitutional challenge should be explicit, with due notice to all affected. This requirement ensures that the correct order is made; that all interested parties have an opportunity to make representations; that the relevant evidence can, if necessary, be led and that the requirements of the separation of powers are respected.
Accuracy in the pleadings is important not only for purposes of defining issues for parties involved in a particular litigation. Orders of constitutional invalidity have a reach that extends beyond parties to a case where a claim for a declaration of invalidity is made. But more importantly these orders intrude, albeit in a constitutionally permissible manner, into the domain of the legislature. The granting of these orders is a serious matter and they should be issued only where the requirements of the Constitution for a review of the exercise of legislative powers have been met. In section 2, the Constitution proclaims its supremacy and declares that law or conduct inconsistent with it is invalid. But the power to determine whether a particular law is indeed inconsistent with the Constitution is conferred on superior courts. Section 172(1)(a) obliges courts to declare law or conduct inconsistent with the Constitution to be invalid. The declaration must, however, be restricted to the extent of the inconsistency. The inconsistency delineates the scope of the judicial review and the consequent declaration of invalidity in respect of a particular challenge. It was in this context that in Shaik this Court observed:
It is constitutionally a serious matter for any Court to declare a statutory enactment of Parliament – or for that matter of any legislature – invalid, because it constitutes a serious invasion, albeit a constitutionally sanctioned one, by one arm of the State into the sphere of another. Moreover, an order by this Court that a statutory provision is constitutionally invalid does not operate between the litigating parties only, but is generally binding on all persons and organs of State.
Holding parties to pleadings is not pedantry. It is an integral part of the principle of legal certainty which is an element of the rule of law, one of the values on which our Constitution is founded. Every party contemplating a constitutional challenge should know the requirements it needs to satisfy and every other party likely to be affected by the relief sought must know precisely the case it is expected to meet. Moreover, past decisions of this Court have adopted this approach and in terms of the doctrine of judicial precedent we are bound to follow them unless we say they are clearly wrong. Judicial precedent serves the object of legal certainty. Following previous decisions constitutes not only compliance with the doctrine of judicial precedent but also accords with the principles of judicial discipline and accountability.
It is against this background that the question whether section 11(2) of the Act is invalid must be determined. Although in the Rule 16A notice SATAWU invoked both sections 17 and 23 of the Constitution as the benchmark against which section 11(2) must be tested, in the courts below and in this Court too, they did not persist in the assertion that the impugned provision infringes section 23. Therefore, we are called upon to weigh section 11(2) against only section 17 of the Constitution.
Does section 11(2) of the Act limit section 17 of the Constitution?
The proper approach to this question is to adopt a two-stage enquiry. During the first leg of the enquiry, the focus is on whether the impugned provision is inconsistent with the Constitution by way of limiting a right in the Bill of Rights. If the answer yielded at this stage of the enquiry is negative, then the enquiry comes to an end. But if the answer is in the affirmative, the Court has no option but to embark on a justification analysis with a view to determining whether the limitation meets the requirements of section 36 of the Constitution. For it is only a law that meets the requirements of this section that can legitimately limit a right in the Bill of Rights.
A court called upon to determine the validity of legislation may not base its decision on the mere say-so of the parties regarding whether or not a particular limitation is justified. This is so because section 36, when read with section 172, obliges courts themselves to determine whether a limitation “is reasonable and justifiable in an open and democratic society based on human dignity, equality and freedom, taking into account all relevant factors” including those listed in the section. Therefore, in my respectful view, once the Supreme Court of Appeal accepted that section 11(2) imposes a limitation of a right in the Bill of Rights, it ought to have done a justification analysis rather than accepting the argument that if the section is construed as affording a real defence, the justification analysis enquiry would be unnecessary.
Although it is apparent from its judgment that the Supreme Court of Appeal accepted that section 11(2) constitutes a limitation, the judgment does not expressly say so. Nor is it clear which right that Court found section 11(2) to be limiting and how the limitation arises. As stated earlier, the question whether a limitation is reasonable and justified arises only if indeed the court before which the claim for constitutional invalidity is made, is satisfied that the impugned provision limits a right in the Bill of Rights.
This leads me to the first leg of the enquiry, which is whether section 11(2) limits any of the rights in section 17 of the Constitution. It is proper to commence this part of the enquiry by first construing section 17 to determine what sort of rights it guarantees. It provides:
Everyone has the right, peacefully and unarmed, to assemble, to demonstrate, to picket and to present petitions.
What emerges from the plain reading of section 17 is that it guarantees four rights. These are the rights to assemble freely, to hold a demonstration, to hold pickets, and to present petitions. In democracies like ours, which give space to civil society and other groupings to express collective views common to their members, these rights are extremely important. It is through the exercise of each of these rights that civil society and other similar groups in our country are able to influence the political process, labour or business decisions and even matters of governance and service delivery. Freedom of assembly by its nature can only be exercised collectively and the strength to exert influence lies in the numbers of participants in the assembly. These rights lie at the heart of democracy.
In the apartheid era the exercise of these rights, even though they were not constitutionally entrenched, was the only means through which black people in this country could express their views in relation to government decisions that affected their lives. Now that they are guaranteed by the Constitution, the enjoyment of these rights may only be restricted in a manner allowed by the Constitution because the Constitution itself recognises that none of these rights is absolute and lays down conditions for their limitation.
Section 36 of the Constitution tells us that the rights in the Bill of Rights may be limited only in terms of a law of general application to the extent that the limitation itself is reasonable and justifiable. Implicit in this injunction is the fact that for a limitation to arise, it must be clear from the terms of the law that it limits a guaranteed right and to what extent. It is only if such law, when properly construed, clearly restricts the exercise of a right in the Bill of Rights that it can be said that it constitutes a limitation of the right in question. For this sort of limitation to survive constitutional scrutiny, it must satisfy all the requirements of section 36.
Plain reading of section 11(2)
This brings me to the interpretation of the impugned provision. Section 11(2) provides:
It shall be a defence to a claim against a person or organization contemplated in subsection (1) if such a person or organization proves—
The text of this section reveals that its object is to restrict the extent of liability imposed by section 11(1) by insulating defendants from its reach. But in order to enjoy the protection it affords, section 11(2) demands that the following requirements be met:
the defendant must show that it did not permit or connive at the act or omission that caused the damage giving rise to the claim;
it must further establish that the act or omission in question did not fall within the scope of the objectives of the gathering and was not reasonably foreseeable; and
in addition, it must prove that it took all reasonable steps to prevent the act or omission that caused damage.
Plainly read, the language of this section indicates that all three conditions must be met for a defendant to be entitled to protection. The difficulty is that there is a contradiction between conditions (b) and (c). The first requires that the defendant must prove that the act or omission which caused damage was not reasonably foreseeable. By contrast, the second implicitly requires that the defendant must have foreseen the wrongful act or omission and must have taken reasonable steps to prevent it.
Regarding condition (c), liability may be avoided if the steps taken to prevent the act or omission that caused harm are found to have been reasonable, for the section requires nothing more or less than reasonable steps to be taken to prevent the act or omission. This plainly does not mean that the convener must, in order to avoid liability, prevent the occurrence of the act or omission foreseen to be capable of causing harm. The convener only needs to take reasonable steps to meet the requirement. The failure to take reasonable steps signifies an action falling short of the requisite standard. Liability ensues due to a failure to meet the standard set.
If the section required the convener to prevent the damage causing act completely, all conveners would have been required to take steps to achieve that. And if such steps were taken, then perhaps the reasonably foreseeable damage causing act would cease to be foreseeable. But the obligation imposed by section 11(2) does not require conveners to take steps that completely prevent acts causing damage. It merely requires that reasonable steps be taken to prevent acts of that nature.
The reasonableness of the preventive steps taken is determined with reference to the nature and extent of the harm foreseen as at the time when the convener is expected to take steps to prevent it. The fact that the harm actually occurred is not relevant to the inquiry because the convener is not required to prevent it completely, but to take reasonable steps to guard against the act which may cause harm. If that act is not reasonably foreseeable as condition (b) demands, then the whole exercise of determining the reasonableness of the steps taken becomes impossible. The requirement that the act or omission that caused harm must not have been reasonably foreseeable lies at the heart of the conflict between conditions (b) and (c).
But this conflict does not translate into a limitation of the rights in section 17 of the Constitution. This is so because section 11(2) lays down an objective standard of what can reasonably be foreseen by every reasonable convener. A reasonably foreseeable act ought to be foreseeable to all reasonable conveners and a convener who fails to take reasonable steps to prevent it cannot enjoy the protection of the section. Whereas a merely foreseeable act may be foreseen by some reasonable conveners, it might not be foreseen by other equally reasonable conveners. Those conveners to whom the harm-causing act was not foreseeable cannot be expected to take reasonable steps to prevent what they could not foresee.
On a plain interpretation, it is impossible for any defendant to prove that it meets the three conditions collectively. It is clear from the Rule 16A notice lodged by SATAWU that this is what prompted it to challenge the validity of section 11(2). It asserted in essence that the unavailability of the defence despite compliance with other requirements of the section might discourage people from exercising their section 17 rights. This, however, is a speculative possibility because no evidence was led to support it and significantly section 11(2) does not impose liability on conveners of gatherings. It is difficult to imagine how a defence which restricts what is apparently a limitation of a section 17 right brought about by section 11(1) can itself become a limitation of the right it seeks to protect. It will be recalled that the effect of section 11(2) is to promote the exercise of the right of assembly by protecting conveners of gatherings from liability imposed by section 11(1).
SATAWU’s complaint is essentially that the defence afforded by section 11(2) is illusory and therefore unattainable. But if the impugned words are removed, it argued, then the defence becomes real. While this may be so, the difficulty that arises is that this Court and any competent court for that matter, can sever words from an Act of Parliament only if it finds them to be inconsistent with the Constitution. The condition precedent to severance is the finding of constitutional inconsistency. In present circumstances this finding can only be based on proof that the impugned provision limits the section 17 rights. Alive to this requirement, SATAWU contended in the Rule 16A notice that the words “and was not reasonably foreseeable” are incompatible with the Constitution and “they infringe on the rights contained in sections 17 and 23 of the Constitution.”
However, the obstacle standing in the way of the finding that the impugned provision infringes the section 17 rights is that section 11(2) does not, either expressly or impliedly, prevent anybody from exercising those rights. Its subject-matter is the defence to liability imposed by section 11(1) which falls outside the scope of the present challenge. It may well be that the defence afforded by section 11(2) is unattainable. But this deficiency does not translate into a limitation of section 17 of the Constitution. Section 17 does not mandate that national legislation that affords the defence we are concerned with here be passed. Put differently, there is no direct link whatsoever between section 11(2) and section 17 of the Constitution. A link between them comes into existence when section 11(1) is brought into the picture because it is the latter section that imposes liability for damage arising from the exercise of the section 17 rights, in circumstances where no wrongful acts can be attributed to the convener of a gathering. Thus section 11(1) creates a new form of liability which was not recognised in our law. At common law, vicarious liability is the only form where a defendant is held liable for damage it did not cause.
As stated earlier, section 11(1) cannot be invoked in order to buttress an otherwise inadequate claim for constitutional invalidity. SATAWU must be held to the case it has pleaded. If it cannot be shown that section 11(2) read in its own terms limits the section 17 rights, then the claim must fail for having not established that the impugned provision constitutes a limitation of those rights.
On its face, it seems that section 11(1) limits the rights in section 17 by imposing liability for damage caused as a result of the exercise of these rights. Furthermore, at face value section 11(1) appears to have been cast widely to the extent that a claim based on it may arise even when damage occurs as a result of a peaceful assembly or demonstration. This is evident from the wide meaning of “riot damage”. Therefore, I cannot agree with the Supreme Court of Appeal when it says that the word “riot” must be given its ordinary meaning. Where Parliament defines a word in a statute, the courts are duty-bound to uphold its defined meaning unless this meaning in the context of the provision interpreted leads to an absurdity not contemplated. Therefore the departure from the defined meaning of “riot damage” by the Supreme Court of Appeal is neither warranted nor justified.
Accordingly, it is incorrect to read section 11(2) as providing a defence to claims for damages which arise from violent gatherings only. Nor do I find support in the text of section 11 of the Act for the view that it was “designed to prevent unlawful violent behaviour” as the Supreme Court of Appeal held. If that was the case, then it could have meant that SATAWU’s claim fails at the starting line because section 17 of the Constitution guarantees peaceful gatherings only. A provision that prevents violent gatherings cannot be held to be limiting the right of assembly in section 17 of the Constitution.
Thus at a practical level, the application of section 11(2) is activated by a claim that a convener of a gathering be held liable in terms of section 11(1). The defence which section 11(2) affords may be invoked once there is a claim based on section 11(1) only. The existence of that claim must precede the invocation of the defence because it has to be a defence to a particular claim. Absent the section 11(1) claim, there can be no talk of calling section 11(2) to a convener’s aid. This illustrates plainly in my view, that the limitation of the right to assemble freely is not located in section 11(2) but in section 11(1) and possibly other provisions of the Act, which fall outside the boundaries of the present challenge. Therefore as presently formulated, the challenge for constitutional invalidity is ill-conceived.
In present circumstances I find that SATAWU has failed to show that section 11(2) constitutes a limitation of section 17 of the Constitution. What it has succeeded in demonstrating is that the conjunctive reading of the section renders it impossible to prove the three conditions it lays down for escaping liability. But this defect cannot be cured through the present constitutional challenge because no limitation has been shown to exist. As appears below, the solution to the present problem lies in the interpretation of the provision.
The main judgment finds that the requirements of section 11(2) “significantly increases the costs of organising protest action.” Added to this is the fact that “it may well be that poorly resourced organizations that wish to organize protest action about controversial causes that are nonetheless vital to society could be inhibited from doing so.” The main judgment then concludes that “[b]oth these factors amount to a limitation of the right to gather and protest.”
But, in my respectful view, the factors relied upon do not prove that section 11(2) limits the right to freedom of assembly. Apart from the fact that the second factor is stated in terms that are not definite, there is no evidence on record which supports these factual findings. The deponent to the affidavit filed by SATAWU asserted that without a meaningful defence afforded by section 11(2), a defendant like SATAWU “faces the spectre of extensive liability in terms of section 11(1) of the Act.” He stated that SATAWU is “run on an extremely tight budget” and this liability will bankrupt it. He concludes by stating:
In light of the obstacle caused by the inclusion of the words ‘and was not reasonably foreseeable’ in section 11(2)(b) of the Act to successfully raising a defence based on this section, the defendant will be precluded, in effect, from convening a gathering in all instances where there is a spectre of the defendant being held liable in terms of section 11(1) of the Act. The effect of the aforementioned words falls nothing short of entirely emasculating the defence contemplated in section 11(2).
SATAWU’s complaint was that since it has foreseen the damage causing act and has taken reasonable steps to prevent it, it would be impossible for it to prove that that act was not reasonably foreseeable. Nowhere in its evidence did SATAWU say the requirements of section 11(2) significantly increase the costs of organising protest action.
More significantly, this was not the ground on which SATAWU based its challenge for a declaration of constitutional invalidity. As earlier stated, in its Rule 16A notice SATAWU contended that it was the words “and was not reasonably foreseeable” which were incompatible with the Constitution in that they infringed the section 17 rights. According to the decision of this Court in Phillips, “[i]t is impermissible for a party to rely on a constitutional complaint that was not pleaded.” Furthermore, “[a]ccuracy in pleadings in matters where parties place reliance on the Constitution in asserting their rights is of the utmost importance.”  If a litigant is not allowed to rely on a complaint that was not pleaded, it must equally be impermissible for a court to base its finding on whether there is a limitation of a constitutional right, on a complaint that was not pleaded. Consistently with this proposition, this Court in Shaik declined to determine the constitutional validity of section 28 of the National Prosecuting Authority Act on a ground which was not raised by litigants but by the Court.
The finding that SATAWU has failed to show that section 11(2) limits the rights in section 17 of the Constitution makes it unnecessary to proceed to the other leg of the enquiry, namely, whether the limitation is reasonable and justifiable. Suffice it to say that the focal point of the analysis in the main judgment is the justification of the provision that imposes liability rather than the one that affords a defence to a claim for damages. In other words, it does not seek to justify the defence against a claim but offers reasons why the conveners of gatherings should be held liable.
Proper construction of section 11(2)
The removal of the obstacle standing in the way of raising the section 11(2) defence by SATAWU does not, in my view lie in a declaration of invalidity. In this regard I agree with the approach adopted by the Supreme Court of Appeal in construing the section. Conditions (b) and (c) must be read disjunctively. The word “and” between the two conditions must be interpreted to mean “or”. Construed in this way, section 11(2) means that a defendant is not required to do the impossible in proving both conditions (b) and (c). On this construction a defendant would be required to prove conditions (a) and (b) or (a) and (c). On this aspect of the case, the question that arises is whether this Court is empowered to depart from the literal meaning of “and” to interpret this word to mean “or”.
As far back as 1924 our courts accepted that sometimes a statute would reflect the word “and” when what was contemplated was “or”. Thus the Appeal Court held “and” to be an equivalent of “or” or vice versa. Recently, the Supreme Court of Appeal tabulated circumstances in which the word “and” may be read to mean “or”. In Ngcobo v Salimba CC; Ngcobo v van Rensburg, the Supreme Court of Appeal defined circumstances where one of these words could be interpreted to mean the other. The Court said:
It is unfortunately true that the words ‘and’ and ‘or’ are sometimes inaccurately used by the legislature, and there are many cases in which one of them has been held to be the equivalent of the other .... Although much depends on the context and the subject matter .... it seems to me that there must be compelling reasons why the words used by the legislature should be replaced; in casu why ‘and’ should be read to mean ‘or’, or vice versa. The words should be given their ordinary meaning ‘ .... unless the context shows or furnishes very strong grounds for presuming that the legislature really intended’ that the word not used is the correct one.... Such grounds will include that if we give ‘and’ or ‘or’ their natural meaning, the interpretation of the section under discussion will be unreasonable, inconsistent or unjust .... or that the result will be absurd .... or, I would add, unconstitutional or contrary to the spirit, purport and objects of the Bill of Rights.
[emphasis in original and references omitted]
It is clear from the statement quoted above that our courts are entitled to construe “and” to mean “or” in circumstances listed in the statement. Two of those circumstances present themselves when the word “and” between conditions (b) and (c) of section 11(2) is given its ordinary meaning. The ordinary meaning of “and”, as stated earlier, leads to an inconsistency if a defendant is called upon to establish both conditions simultaneously. The consequence of that is an absurdity which could never have been contemplated when section 11(2) was enacted. The purpose of this section is to limit the extent of liability imposed by section 11(1) on organisers of gatherings. It is plain from the language of section 11(2) that organisers must act prudently if they are to avoid liability under section 11(1).
It follows that the inconsistency and absurdity which flow from the natural meaning of “and” between the two conditions justify the reading of that word to mean “or”. This construction addresses the core of the complaint raised by the applicants. Indeed at the hearing in this Court their counsel conceded, correctly so in my view, that if section 11(2)(b) and (c) are read disjunctively, the problem would be solved.
Moreover, reading “and” to mean “or” achieves the purpose of section 11(2) which is to restrict liability imposed by section 11(1) and by so doing promote the right of freedom of assembly. This alone justifies the departure from the ordinary meaning of the word “and”.
This construction seriously undermines the claim for constitutional invalidity. As the main judgment correctly points out, if a provision is capable of two constructions: one interpretation rendering it inconsistent with the Constitution while the other does not, the interpretation that preserves the validity of the provision must be preferred. In Hyundai Motor Distributors, this Court reaffirmed as a sound principle of constitutional interpretation, the rule that in cases where the impugned legislation is capable of two reasonable constructions, courts must prefer the interpretation which conforms with the Constitution. This interpretive approach preserves legislation from constitutional invalidity.
Therefore, even if it were to be said that SATAWU has made out a case on the issue of limitation, reading section 11(2)(b) and (c) separately renders the enquiry into its constitutional validity unnecessary. This is so because the whole case is about construing the section in a manner that affords a defendant in SATAWU’s position, a real defence.
It is for these reasons that I would grant leave and dismiss the appeal.
 SATAWU v Garvis 2011 (6) SA 382 (SCA) (Supreme Court of Appeal judgment).
 Garvis v SATAWU (Minister for Safety and Security, Third Party) 2010 (6) SA 280 (WCC).
 205 of 1993.
 66 of 1995.
 Now the Minister of Police.
 This was done under Rule 13 of the Uniform Rules of Court. SATAWU joined the Minister because it took the view that it would be entitled to a contribution from the Minister in the event that it is liable to the respondents. Its view in this regard was premised on the argument that any losses sustained by the respondents were caused, at least in part, by the negligent conduct of members of the South African Police Service.
 Section 17 of the Constitution provides:
Everyone has the right, peacefully and unarmed, to assemble, to demonstrate, to picket and to present petitions.
 See sections 2, 3 and 4 of the Act.
 Set out in full at  above.
 Supreme Court of Appeal judgment above n 1 at paras 36-9.
 Id at para 41.
 In the High Court and Supreme Court of Appeal proceedings, SATAWU did not use the language of the rule of law, principle of legality or rationality. Rather, it argued that the defence created by section 11(2) is internally incoherent and self-destructive. Central to the applicants’ argument, however, is that the reasonable foreseeability requirement in section 11(2)(b) makes it impossible for a defendant to rely successfully on the defence. Nothing should be made of this difference in terminology. It follows that SATAWU sought to rely on rationality from the outset.
 Section 1(c) of the Constitution. See Pharmaceutical Manufacturers Association of SA: In re Ex parte President of the Republic of South Africa  ZACC 1; 2000 (2) SA 674 (CC); 2000 (3) BCLR 241 (CC) and Fedsure Life Assurance Ltd v Greater Johannesburg Transitional Metropolitan Council  ZACC 17; 1999 (1) SA 374 (CC); 1998 (12) BCLR 1458 (CC).
 S v Mhlungu  ZACC 4; 1995 (3) SA 867 (CC); 1995 (7) BCLR 793 (CC) at para 36. See generally Du Plessis “Statute Law and Interpretation” 25(1) LAWSA 2011 at para 334.
 Bertie van Zyl (Pty) Ltd v Minister for Safety and Security  ZACC 11; 2010 (2) SA 181 (CC); 2009 (10) BCLR 978 (CC) at para 20 and Investigating Directorate: Serious Economic Offences v Hyundai Motor Distributors (Pty) Ltd: In re Hyundai Motor Distributors (Pty) Ltd v Smit NO  ZACC 12; 2001 (1) SA 545 (CC); 2000 (10) BCLR 1079 (CC) at para 26.
 Section 11(2)(a).
 Section 11(2)(b).
 Section 11(2)(c).
 See Kruger v Coetzee 1966 (2) SA 428 (A) at 430E-G, where the test was formulated as follows:
For the purposes of liability culpa arises if—
 1954 (3) SA 464 (A).
 Id at 477.
  ZASCA 148; 1991 (1) SA 756 (A).
 Id at 776.
 Ziliberberg v Moldova ECHR (Application No 61821/00) (4 May 2004) at para 2. See also Cisse v France ECHR (Application No 51346/99) (9 April 2002) at para 50 and Christians Against Racism and Fascism v United Kingdom (1980) 21 DR 138 (Application No 8440/78) at para 4.
 Section 2, read with sections 3 and 4 of the Act.
 Riot damage is defined in section 1 of the Act as—
any loss suffered as a result of any injury to or the death of any person, or any damage to or destruction of any property, caused directly or indirectly by, and immediately before, during or after, the holding of a gathering.
 See Woolman “Freedom of Assembly” in Woolman et al (eds) Constitutional Law of South Africa 2 ed (Juta & Co Ltd, Cape Town 2008) at 43-4 to 43-6 for examples.
 See Ex parte Gauteng Provincial Legislature: In re Dispute Concerning the Constitutionality of Certain Provisions of the Gauteng School Education Bill of 1995  ZACC 4; 1996 (3) SA 165 (CC); 1996 (4) BCLR 537 (CC) at para 46.
 Woolman “My Tea Party, Your Mob, Our Social Contract: Freedom of Assembly and the Constitutional Right to Rebellion in Garvis v SATAWU (Minister for Safety & Security, Third Party) 2010 (6) SA 280 (WCC)” (2011) 27 SAJHR 346 at 348, captures the important, instrumental nature of freedom of assembly: “By creating space for crowd action, s 17 vouchsafes a commitment to a form of democracy in which the will of the people is not always mediated by political parties and the elites that run them”. On the importance of participation in democratic processes, see: Doctors for Life International v Speaker of the National Assembly  ZACC 11; 2006 (6) SA 416 (CC); 2006 (12) BCLR 1399 (CC) at para 115, where this Court held that participation by the public in a representative democracy enhances the civic dignity of those who participate, promotes a spirit of democratic and pluralistic accommodation, and is of special importance to those that are relatively disempowered; and Albutt v Centre for the Study of Violence and Reconciliation,  ZACC 4; 2010 (3) SA 293 (CC); 2010 (5) BCLR 391 (CC), where this Court held that the participation of victims in a special dispensation process, initiated to deal “with the ‘unfinished business’ of the Truth and Reconciliation Commission” (para 4), was necessary to achieve nation-building and national reconciliation (paras 58-9).
 Report of the Special Rapporteur on extrajudicial, summary or arbitrary executions, Christof Heyns (A/HRC/17/28) 26 May 2011 at para 33.
  ZACC 7; 1999 (4) SA 469 (CC); 1999 (6) BCLR 615 (CC).
 Id at para 8.
  ZACC 17; 2001 (3) SA 409 (CC); 2001 (5) BCLR 449 (CC) (Mamabolo).
 Id at para 28.
 Section 11(3) of the Act provides:
For the purposes of—
any person or organization held liable for such damage by virtue of the provisions of subsection (1) shall, notwithstanding the said provisions, be deemed to have been liable therefor in delict.
 34 of 1956.
 See Thoroughbred Breeders’ Association v Price Waterhouse 2001 (4) SA 551 (SCA) at para 17 of the majority judgment of Marais JA, Farlam JA and Brand AJA; Dlakela v Transkei Electricity Supply Commission 1997 (4) SA 523 (Tk) at 526G-J; and South British Insurance Co Ltd v Smit 1962 (3) SA 826 (A) at 835A-C.
 Section 1(1) of the Apportionment Act provides:
 See for instance McKerron The Law of Delict: A Treatise on the Principles of Liability for Civil Wrongs in the Law of South Africa 7 ed (Juta & Co, Limited, Cape Town 1971) at 109, 305 and 310; Lee and Honoré The South African Law of Obligations 2 ed (Butterworths, Durban 1978) at 229-32. For judicial support of this view see Mohamed J in Randbond Investments (Pty) Ltd v FPS (Northern Region) (Pty) Ltd 1992 (2) SA 608 (W) at 619E-G.
 Burchell Principles of Delict (Juta & Co, Ltd, Cape Town 1993) at 241; Van der Walt and Midgley Principles of Delict 3 ed (LexisNexis Butterworths, Durban 2005) at 247; Neethling et al Law of Delict 5 ed (LexisNexis, Durban 2006) at 245. This difference in emphasis may have been due to the slightly different working of the Apportionment Act. Nevertheless, the points made by McKerron id and Lee and Honoré id are still valid in the context of the current Act.
 Section 2(6)(a) of the Apportionment Act. See also section 2(7)(a), 2(8)(a)(ii) and 2(11)(a) of this Act.
 Smith v Road Accident Fund  ZASCA 15; 2006 (4) SA 590 (SCA) at para 10.
 Above n 33.
 Id at para 49.
 Act 205 of 1993 (Act). Provisions of the section are quoted in  and  below.
 Section 17 of the Constitution is quoted in  below.
 SATAWU v Garvis 2011 (6) SA 382 (SCA); 2011 (12) BCLR 1249 (SCA).
 Garvis v SATAWU (Minister for Safety and Security, Third Party) 2010 (6) SA 280 (WCC).
 Section 3 of the Act.
 Section 1 of the Act.
 Section 4 of the Act.
 Section 5(2) of the Act.
 Section 9 of the Act.
 Section 11(1) of the Act.
 The full text of section 11(2) is quoted in  below.
 Section 1 of the Act.
 See Rule 16(A) of the Uniform Rules of Court.
 SATAWU above n 3 at para 32.
 Id at para 41.
  ZACC 24; 2004 (3) SA 599 (CC); 2004 (4) BCLR 333 (CC) at paras 24-5.
 32 of 1998.
 In relevant part section 28 provides:
 Shaik above n 16 at para 23.
 Shaik above n 16 at paras 21-2 and 24-5.
 Id at para 33.
 Id at paras 38 and 39.
 Id at para 40.
  ZACC 15; 2006 (1) SA 505 (CC); 2006 (2) BCLR 274 (CC) at para 43.
 Sections 167-9 of the Constitution.
 Section 172(1) provides in relevant part:
When deciding a constitutional matter within its power, a court—
 Shaik above n 16 at para 23.
 Ex parte Minister of Safety and Security: In re S v Walters  ZACC 6; 2002 (4) SA 613 (CC); 2002 (7) BCLR 663 (CC) at para 57.
 See Coetzee v Government of the Republic of South Africa, Matiso v Commanding Officer, Port Elizabeth Prison  ZACC 7; 1995 (4) SA 631 (CC); 1995 (10) BCLR 1382 (CC) at para 9; S v Williams  ZACC 6; 1995 (3) SA 632 (CC); 1995 (7) BCLR 861 (CC) at para 54; S v Makwanyane  ZACC 3; 1995 (3) SA 391 (CC); 1995 (6) BCLR 665 (CC) at para 100; and S v Zuma  ZACC 1; 1995 (2) SA 642 (CC); 1995 (4) BCLR 401 (CC) at para 21.
 Section 36(1) provides:
The rights in the Bill of Rights may be limited only in terms of law of general application to the extent that the limitation is reasonable and justifiable in an open and democratic society based on human dignity, equality and freedom, taking into account all relevant factors, including—
 SATAWU above n 3 at para 32.
 SATAWU above n 3 at para 52.
 Hoban v ABSA Bank Ltd t/a United Bank  ZASCA 12; 1999 (2) SA 1036 (SCA) at paras 18-9.
 SATAWU above n 3 at para 52.
 Above n 24 at para 39-40.
 Above n 17.
 Shaik above n 16 at paras 38-40.
 SATAWU above n 3 at para 40-1.
 Barlin v Licensing Court for the Cape 1924 AD 472 at 478.
 1999 (8) BCLR 855 (SCA).
 Id at para 11, per Olivier JA, with Mahomed CJ, Grosskopf JA, Farlam AJA and Madlanga AJA concurring.
 Investigating Directorate: Serious Economic Offences v Hyundai Motor Distributors (Pty) Ltd: In re Hyundai Motor Distributors (Pty) Ltd v Smit NO  ZACC 12; 2001 (1) SA 545 (CC); 2000 (10) BCLR 1079 (CC). At para 23 Langa DP said:
Accordingly, judicial officers must prefer interpretations of legislation that fall within constitutional bounds over those that do not, provided that such an interpretation can be reasonably ascribed to the section.
W Trengove SC, H Maenetje SC and J Brickhill (instructed by Cheadle Thompson & Haysom Inc.) for the Applicants.
A Katz SC and D Cooke (instructed by Fairbridges Attorneys) for the First to Eighth Respondents.
Karrisha Pillay (instructed by the State Attorney) for the Ninth Respondent.
IV Maleka SC and N Bawa (instructed by Webber Wentzel (Cape Town)) for the Intervening Party.
AD Stein, IB Currie and NC Ferreira (instructed by Freedom of Expression Institute) for the Amicus Curiae.
all rights reserved