Weasel clause or protector of freedom? A new consideration of section 36 of the Constitution
Written constitutions are meant to limit government scope and power. Bills of rights are meant to safeguard individual freedom from government overreach. This is as true in South Africa as it is anywhere else.
Does the South African Constitution have a provision that allows government to ‘weasel’ its way out of strictly recognising and protecting our rights, or have we misconstrued section 36 entirely?
This article is a reworking of a portion of a recent paper I wrote for the Institute of Race Relations, ‘Civil Liberty in South Africa: Freedom Under Law Three Decades After Apartheid’.
Lochner and South Africa’s culture of justification
In his 2014 book Restoring the Lost Constitution, jurist Randy Barnett writes that in the early days of the Supreme Court of the United States, constitutional ‘due process’ was regarded as requiring ‘legislatures to establish to the satisfaction of an independent tribunal that its restrictions on liberty were necessary and proper.’ Barnett elaborates:
‘Given the pervasiveness of political motives in conflict with the original constitutional scheme of limited powers, the Supreme Court would not simply take the legislature’s word for its claim that some restriction of liberty was necessary to accomplish an appropriate end. The Court began requiring some proof that this was the case. It required states to show that legislation infringing upon the liberties of the people really was a necessary exercise of the state’s police power – a power that it held, quite expansively, to include the protection of the health, safety, and morals of the general public.’
This was what is today known as the controversial Lochner doctrine, named for the infamous 1905 case of Joseph Lochner v People of the State of New York.
In this judgment, a New York law setting maximum working hours for bakers was struck down, because the New York government had failed to adequately justify the law’s limitation on freedom of contract under the due process clause of the Fourteenth Amendment to the United States Constitution. Lochner was later reversed across multiple Supreme Court judgments in the 1930s (during the administration of Franklin Delano Roosevelt and his ‘New Deal’) and into the 1950s.
Today, American courts no longer scrutinise government infringements on (primarily economic) liberty as the US Constitution arguably requires.
Barnett summarises the Lochner doctrine as follows:
‘When the liberty of the individual clashes with the power of the state, the Court would not accept the “mere assertion” by a legislature that a statute was necessary and proper. Instead, it required showing that a restriction of liberty have [sic] a “direct relation, as a means to an end,” and that “the end itself must be appropriate and legitimate.” Having offered no such evidence, the State of New York lost.’
The Lochner doctrine has been misconstrued – even by South African jurists – as placing abstract, primarily economic freedoms (like the right to property and freedom of contract) above democratic will. In reality, it simply required government to justify its limitations upon (any) rights.
Like the regime established by the United States Constitution, South Africa’s constitutional scheme is also one of ‘limited powers’. What South African jurists would perhaps not like to admit is that section 36 of the Constitution, in effect, incorporates the Lochner doctrine into our formal constitutional law. Shed of its ostensibly tainted American context, even left-of-centre South African legal thinkers have endorsed the justification inquiry.
Indeed, before the Constitution was adopted, legal scholar Etienne Mureinik wrote in 1992 that ‘it will be impossible to fully undo apartheid without a legal order which makes every law, every government decision, indeed every decision having governmental effect, amenable to scrutiny; one which empowers the judges to demand to know the reasons for the law, or the decision’. If government were forced to justify itself in this manner before the courts, the quality of governance would necessarily be improved. Writes Mureinik:
‘The knowledge that any government programme could be summoned into court for searching scrutiny would force its authors closely to articulate their reasons for dismissing the objections and the alternatives to the programme, and precisely to articulate the reasons that link evidence to decision, premises to conclusion. The need to articulate those reasons during decision making would expose weaknesses in the programme that might force reconsideration long before the need arose for judicial challenge.’
The interim Constitution, and the current Constitution after it, adopted the spirit of Mureinik’s recommendation, primarily in the form of the limitations provision – section 33 in the interim and section 36 in the current Constitution.
Section 36
Section 36 of the Constitution – titled ‘Limitation of rights’ – provides as follows:
‘(1) 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 –
(a) the nature of the right;
(b) the importance of the purpose of the limitation;
(c) the nature and extent of the limitation;
(d) the relation between the limitation and its purpose; and
(e) less restrictive means to achieve the purpose.
(2) Except as provided in subsection (1) or in any other provision of the Constitution, no law may limit any right entrenched in the Bill of Rights.’
The free exercise of constitutional rights is the default position, whereas their limitation must always be the exception, reasoned Gerhard Erasmus in 1996. Limitations must therefore be exceptional, and for valid constitutional purposes only. This is what section 36 seeks to guarantee.
Section 36(b) and (d) of the Constitution seem to clearly encapsulate the Lochner doctrine as described above by Barnett.
Nonetheless, conceptually, the limitations provision is often regarded – even by the likes of Mureinik – as a way to ‘limit’ rights that are not ‘absolute’. This is precisely the wrong mentality with which to approach this provision.
In fact, limitation provisions, especially when they are included in a supreme bill of rights – as South Africa’s are – should be interpreted as a limitation on government’s authority to infringe on rights, rather than a limitation on the legal subject’s civil liberty. Section 36, then, is part of the regime for limiting the power of government, not part of a regime empowering government to limit civil liberty. Governments by their very nature already possess this power. An explicit limitations provision is about channelling and circumscribing that inherent power.
The fact that section 36 provides explicitly that a right may ‘only’ be limited as provided for in section 36(1) lends credence to this submission. The provision is there to protect rights, not open the way for government to limit them more easily.
There are, in my view, two conceptual approaches to section 36. While they deal with the same text and (hopefully) the same meaning of the text, they do vary what the outcome of a section 36(1) analysis would be. It all depends on the way in which the judge applies the provision.
The first approach – the present approach – understands section 36 as a provision that empowers government to limit the rights in the Bill of Rights. This approach, in other words, vests government with a power: it creates exceptions to constitutional rights. This traditional approach is identified with the idea that section 36 is a ‘weasel clause’ that allows government to weasel its way out of having to respect the rights of legal subjects as guaranteed in the Constitution.
The second approach – which one might term the ‘liberal’ approach – understands section 36 as a provision that constrains how government may infringe the rights in the Bill of Rights. This approach limits how government may exercise its power: it protects constitutional rights. This approach would hold section 36(1) out as a guarantor of civil liberty rather than a weasel clause.
One might argue that there is no difference between these two approaches. A judge must apply section 36(1) as written. However, how one reads section 36 very much determines how it is applied. Here follows an example.
Section 36(1)(e) provides that a court must take into account whether there are ‘less restrictive means’ available to achieve the purpose of a limitation of rights when determining whether the limitation is reasonable and justifiable. On the weasel-clause approach, a court would construe this leg of the analysis very narrowly, if at all. In the initial Constitutional Court judgment in Prince v Cape Law Society in 2002 (since overturned), for example, the Court acknowledged that there were, in fact means that were less restrictive than Parliament’s total ban on marijuana, but nonetheless upheld the limitation.
On the liberal approach, where the limitation provision is taken seriously as a guarantor of freedom, a court would emphasise whatever less restrictive means are available and require them be preferred over the more restrictive means, because section 36(1)(e) unambiguously and clearly requires the courts to do so.
A higher judicial standard
If one bears the nature of constitutionalism in mind – that written constitutions are meant to limit the scope and power of government, and bills of rights are meant to protect individual freedom against government interference – then the liberal approach is the most appropriate way to conceptualise section 36. South Africa’s courts must therefore insist on a higher standard of justification in terms of section 36 than is presently the case.
There are various ways to do this, but I explain only some of them here:
Justification must be compulsory
In the past, the courts have held that a limitation could be constitutionally justifiable even if government did not submit evidence or argument in favour of the limitation – that is, government did not even attempt to justify the limitation. The courts themselves will inquire into justification in such an event.The courts have also effectively taken ‘judicial notice’ that a certain restriction on civil liberty is justified without anyone – government or judge – applying any limitation analysis at all.
Both these approaches are misguided. The courts must always insist on a section 36 justification per se. If the government does not include a section 36(1) justification of a limitation on constitutional rights in its submissions before the courts, the courts must find against the limitation and declare it unconstitutional without further ado.
Justification must be comprehensive
The courts must never allow attempted justifications to skip any leg of the inquiry listed in section 36(1)(a) to (e). A common leg that is usually skipped is that of the aforementioned ‘less restrictive means’ in section 36(1)(e), in large part (according to jurists Iain Currie and Johan de Waal) due to the notion that the courts should not ‘second-guess the wisdom of policy choices made by legislators’.
However, insofar as inquiring into whether there are policy choices available to government that would have a less deleterious impact on the constitutional rights of legal subjects (compared to the avenue government has embarked upon), the Constitution unequivocally empowers the courts to ‘second-guess’. If the government wishes to avoid the courts doing so, it would do well to choose the least restrictive means reasonably at its disposal to achieve the purpose of the limitation, before the matter even gets to court.
Limitation must be strictly necessary
The courts must insist on strict necessity, not merely reasonable necessity, and certainly not expediency, for a justifiable limitation of rights. In other words, when section 36(1)(d) – ‘the relation between the limitation and its purpose’ – is considered, the courts must insist that the limitation be strictly necessary to achieve the lawful purpose for which it is being imposed, not merely expedient.
Legal subjects must enjoy the benefit of the doubt
The courts must always approach a section 36 justification in favorem libertatis. This means that at each leg of the justification analysis, the benefit of the doubt (the presumption) must lie with the liberty of the subject, not with the agenda of the government. Of course, this principle applies to the civil liberties guaranteed in the Bill of Rights, not the welfare entitlements. Where a welfare entitlement entails the limitation of civil liberty, the freedom of the subject must remain the court’s point of departure.
The Constitution must be the only source of justification
When considering section 36(1)(b) – ‘the importance of the purpose of the limitation’ – the courts must not consider any purpose outside of the Constitution as a lawful purpose. In other words, only if the purpose is directly concerned with giving effect to a constitutional obligation should government be able to justify its limitation. If the government submits a reason that cannot reasonably be traced back to a clear constitutional provision, it must be treated as ultra vires insofar as section 36 is concerned.
Substance must always weigh heavier than form
The courts must have regard to the substance of the limitation, not merely its form. In other words, the actual effect of the limitation on the constitutional rights of the legal subject must be the primary consideration, rather than how the limitation is stated in the text of the legislation or regulation, or how it is rationalised by the government.
This recommendation might seem trite; however, the courts have in the past engaged in purely formalistic limitation analyses. For instance, in the 2020 FITA v President of the RSA judgment, the High Court acknowledged that the lockdown prohibition on tobacco product sales had not led to the purpose of the limitation – widespread reduction in smoking – being achieved. In fact, the Court acknowledged that more widespread smoking of unregulated, black-market cigarettes was the result of the prohibition. Nonetheless, the Court believed the regulation was still rational because the prohibition was theoretically capable of achieving its intended purposes. This was a form-over-substance analysis that inquired only into the limitation as it appeared in the text of the regulation, rather than the irrational and counterintuitive consequences it was yielding in practice.
Harmonising the general with internal limitations
As Marius van Staden argues in a 2015 article, it is an accepted principle of interpretation that no legal provision may be assumed or construed as being redundant or of no effect. This applies doubly in the case of supreme constitutions. In other words, the constitutional text must be taken seriously.
There are three kinds of limitations in the Bill of Rights: logical limitations, internal limitations, and the general limitation. None of these limitations – part of the constitutional text – but in particular internal limitations, may be disregarded as pointless fluff that can be overridden by simply going straight to section 36.
Section 36(1) provides that ‘any’ right in the Bill of Rights may be limited by its application. This has been interpreted to mean that over and above the internal limitations, section 36 can also be applied. This approach has the result that internal limitations are effectively redundant. What would have been the point of including internal limitations in certain constitutional rights formulations if section 36 can simply be used for them all?
Another approach, to only apply internal limitations (an approach I advocated for some time ago), would have the result that section 36 is redundant vis-à-vis those rights.
To take an example: section 16, the right to freedom of expression, contains three exclusions from the protection of free expression in section 16(2): advocacy for war, incitement to imminent violence, and advocacy of hatred that is based on race, ethnicity, gender, or religion, and that constitutes incitement to cause harm. The constitutional logic is that if an instance of expression is not covered by the exclusions, it is constitutionally protected expression. Section 16(2) functions as an internal limitation on expressive freedom, as it allows government to regulate expression falling within the exclusions.
The first approach (section 36-over-internal limitations) would disregard these limitations and allow government to simply limit freedom of expression any way it desires by application of section 36(1). This renders section 16(2) superfluous – it might as well have been excluded from the Constitution entirely when it was drafted. The second approach (internal limitations-over-section 36) would tend to disregard section 36(1) – which clearly provides that ‘any’ right may be limited by that provision – and require government only to ask whether its limitation complies with the wording of section 16(2).
There is a third approach – recommended here – that gives both internal limitations and section 36 their due recognition.
This approach is as follows:
- Any right without an internal limitation may be limited by section 36(1);
- In the case of rights with internal limitations, those limitations themselves must also be tested against section 36(1); but
- Those internal limitations represent a closed list of allowable limitations vis-à-vis those rights.
In the example of section 16, using the third approach, the right to freedom of expression may only be limited under the circumstances stated in section 16(2), and when it is, those limitations must be justified in terms of section 36(1). Sections 36(1)(b) and (c) in particular – ‘the importance of the purpose of the limitation’ and ‘the nature and extent of the limitation’ – would allow government to cite section 16(2) as a clear constitutional justification for the contemplated limitation.
Under this approach, the function of an internal limitation is to indicate that the constitutional right may only be limited under the circumstances of the internal limitation, but that section 36 must still be utilised to justify such limitations.
Public (vs?) private interest
Officials and legislators often claim, explicitly and implicitly, that a system that prioritises the protection rather than limitation of freedom is inimical to the public interest. For instance, to insist upon an even stronger protection of the right to property in section 25 of the Constitution than South Africans enjoy today is portrayed as something that would undermine the government’s land reform agenda.
Even the courts have misconstrued section 25 in this way. In the 2002 judgment of FNB v SARS, Justice Ackermann says that the ‘purpose of section 25 has to be seen both as protecting existing private property rights as well as serving the public interest.’ In so doing, the judge creates a dichotomy that does not exist for, in fact, the overriding protection of civil liberty, including the right to property, is in the public interest. This is not least because this protection is the raison d’être of government in a constitutional democracy.