Moyo v Minister of Justice and Constitutional Development; Sonti v Minister of Justice and Correctional Services
The SCA decision is significant because it is also relevant to the rights of employees who are faced with ‘charges’ and expected to plead ‘guilty’ or ‘not guilty’ and very often then expected to give evidence against themselves.
“An evidential burden does not impose a reverse onus, nor is it a per se case of a constitutional infringement. Nonetheless there is a constitutional problem with the section. It contravenes the provisions of s 35(3)(h) of the Constitution, not because it infringes the presumption of innocence, but because it places improper pressure on an accused to forego their constitutional right to silence and not to give self-incriminating evidence. That is inconsistent with the broader right to a fair trial, because it relieves the prosecution in the first instance from the need to lead evidence to show that the actions of the accused are without lawful reason and, after the close of the prosecution case, it constrains the accused to give evidence themself or to lead evidence from others. As such it infringes the constitutional right in a more insidious way in that it operates as a compulsion on the accused to disclose at an early stage of the proceedings what may be the key element of their defence. Indeed, as correctly pointed out in para 62 of the main judgment, it may go so far as to compel the accused to make admissions that relieve the prosecution of the obligation to prove certain facts. So it infringes the right to a fair trial and probably the right to silence. As Ms Sonti said in her replying affidavit it seeks to force the accused to break their silence before the close of the prosecution case.” [para 154]
Split decision and see the summary for details.
Discussion by GilesFiles
Constitutional law – constitutionality of s 1(2) of the Intimidation Act 72 of 1982 – s 1(2) presumes that accused’s actions or utterances are without lawful reason if such reason not advanced prior to close of prosecution case – whether presumption reverses the onus of proof or is merely evidential – s 35 of Constitution – right to a fair trial, to be presumed innocent and to remain silent.
Majority – presumption evidential – places pressure on the accused to disclose content of defence prematurely – infringes the right to a fair trial and the right to remain silent – no justification for limitation of rights in terms of s 36 of Constitution – section 1(2) unconstitutional and invalid.
Minority – presumption reverses onus of proof requiring the accused to prove the existence of a lawful reason for their acts or utterances – accused can be convicted even though no proof of guilt beyond reasonable doubt – infringes the right to be presumed innocent and the right to remain silent – no justification for limitation of rights – section 1(2) unconstitutional and invalid.
Constitutionality of s 1(1)(b) of Intimidation Act – whether infringes right of freedom of expression as contained in s 16(1) of Constitution – section capable of being interpreted in conformity with Constitution.
Majority – section to be interpreted in light of s 39(2) of Constitution – criminal provision to be construed in favour of the liberty of the citizen – presumption of mens rea in the absence of express provision negating presumption
Section 1(1)(b) to be construed as relating only to conduct that is intimidatory in character – section requires mens rea – conduct or utterances constituting intimidation must induce actual fear in target or inducing such fear would reasonably be the consequence of such conduct or utterances – mere anxiety, nervousness or apprehension not constituting fear within the meaning of the section – conduct that is lawful in terms of the Constitution or statute not unlawful – lawful expression in terms of s 16(1) of Constitution not falling within the section and not constituting intimidation.
Minority – section impermissibly wide – contravenes s 16(1) of Constitution – no justification for limitation of rights – section 1(1)(b) unconstitutional and invalid
Quotations from judgment
Note: Footnotes omitted and emphasis added
Wallis JA (Maya P and Makgoka AJA concurring)
 I have had the privilege of reading the careful judgment of my colleague Mbha JA (the main judgment). Unfortunately, I find myself unable to agree with his conclusion in Mr Moyo’s appeal that s 1(1)(b) of the Intimidation Act 72 of 1982 (the Act) infringes s 16(1) of the Constitution and falls to be struck down. In my view it is capable of being construed in a way that is compatible with the Constitution and serves the valuable purpose of providing the protection of the criminal law against intimidatory conduct that is abhorrent in any democratic society. I have in mind sexual harassment falling short of any of the crimes in the Criminal Law (Sexual Offences and Related Matters) Amendment Act 32 of 2007, stalking, trolling attacks on social media, cyber attacks and the like. As the author of the relevant section in a leading textbook notes: ‘the problem of intimidation in society, and the need for the law to intervene to prevent this from occurring, is generally acknowledged, even by critics of the Act.’
 If the section is struck down it will leave our police without any means to protect the people of this country against such conduct. It will also rob them of a weapon to be used against anyone making threats having a broader impact, such as a threat to release a poisonous substance into a city’s water supply, or a hoax warning that an explosive device has been placed in a football stadium or shopping centre. Accordingly, if I shared my colleague’s view that the section impermissibly infringed on forms of expression protected by s 16(1) of the Constitution, I would suspend any order of invalidity, subject to conditions that would prevent the section being used to prosecute people for constitutionally protected expression.
 My view can be shortly summarised. It is that the appellants’ submissions on the meaning of the section ignore fundamental rules in regard to the constitutional approach to the interpretation of statutes and other well-established principles of statutory interpretation, especially as applicable to provisions imposing criminal liability. Contrary to those submissions, I hold that, properly interpreted, s 1(1)(b) requires proof of both mens rea and unlawfulness; is only concerned with intimidatory conduct that induces or would induce fear properly so called in a reasonable person; and does not criminalise conduct that is otherwise lawful in terms of the Constitution and other legislation.
 On Ms Sonti’s appeal I agree with my colleague that s 1(2) of the Act is unconstitutional, but do so for materially different reasons from his. I also take a different view of the appropriateness of the procedure that has resulted in the prosecution of these two cases being delayed for nearly six years in the case of Mr Moyo and nearly five years in the case of Ms Sonti. I will deal with my reasons for holding that view in the closing section of this judgment. However, because matters have proceeded this far and dismissing the appeals on this narrow ground might be thought to leave the judgment of the high court unscathed and authoritative, I agree that it is in the interests of justice for us to adjudicate the case on its merits. I also agree with my colleague’s criticism of the reasoning of the high court. For ease of comparison between the two judgments I will follow the order adopted by my colleague and deal first with Mr Moyo’s appeal and s 1(1)(b) of the Act, then with Ms Sonti’s appeal and s 1(2), and lastly with the procedural history of these two cases.
Interpreting the section
 My starting point is the proper interpretation of s 1(1)(b) and the injunction in s 39(2) of the Constitution when construing legislation to promote the spirit, purport and objects of the Bill of Rights. Wherever possible, without straining the language of a statutory provision, it must be given an interpretation that is within constitutional bounds in preference to one that involves an infringement of constitutionally protected rights. The task must also be approached in the light of the summary of the proper approach to interpretation in Endumeni, a judgment that has been repeatedly cited and followed in this court and in the Constitutional Court. The words of the section are the starting point, but they are to be considered in the light of their context, the apparent purpose of the provision and any relevant background material. A sensible meaning is to be preferred to one that leads to impractical results.
 Two principles particularly relevant to the interpretation of criminal statutes need mention. Firstly, when dealing with a provision that creates a criminal offence it is to be construed in favour of the liberty of the subject. If there is more than one meaning available, the meaning that is least onerous should be adopted. Secondly, it is presumed that the commission of statutory offences requires intention (mens rea). Clear wording is required to exclude the need for intention because:
‘In view of such general maxims as nulla poena sine culpa and actus non facit reum nisi mens sit rea, the Legislature, in the absence of clear and convincing indications to the contrary in the enactment in question, is presumed to have intended that violations of statutory prohibitions would not be punishable in the absence of mens rea in some degree or other.’
When the penalties provided for the offence are heavy and the potential inroads into the rights of the citizen substantial that reinforces the need for the prosecution to establish dolus.
 Lastly, it is a basic principle of interpretation that internal inconsistency in a statute is to be avoided. So far as possible it is to be construed as a coherent whole. The need for internal consistency assumes particular importance when dealing with a crime such as intimidation that can manifest itself in slightly different ways involving the same central concepts. Otherwise differing standards for imposing criminal liability would be applicable to the same crime. This was the effect of the appellants’ argument, but it is inconsistent with principle.
The constitutional challenge
 The appellants did not, as I understood it, challenge the notion that criminalising intimidatory behaviour is legitimate in a democratic society. Stalking was put as an example to counsel and he accepted that it is covered by the section and ought properly to be criminalised. Nonetheless he argued that the section should be struck down as over-broad and having the effect of criminalising ‘a wide range of expression protected by section 16(1) of the Constitution’ and a ‘vast quantity of everyday political speech’. It was submitted that the language of the section ‘clearly evinced an intention to create no fault liability’, that is, that criminal intention was not required. The offence created by the section was deconstructed into two separate offences, the one subject to considerations of reasonableness and the other not. In summary it was said that the section ‘obliterates the distinction between “true threats” and “political hyperbole”’. I turn to consider whether the language of the section supports these arguments.
The offence of intimidation
 The Act creates thee offence of intimidation, but provides that it may manifest itself in different ways. That is apparent from s 1(1), which reads:
‘1. Prohibition of and penalties for certain forms of intimidation.—
(1) Any person who—
(a) without lawful reason and with intent to compel or induce any person or persons of a particular nature, class or kind or persons in general to do or to abstain from doing any act or to assume or to abandon a particular standpoint—
(i) assaults, injures or causes damage to any person; or
(ii) in any manner threatens to kill, assault, injure or cause damage to any person or persons of a particular nature, class or kind; or
(b) acts or conducts himself in such a manner or utters or publishes such words that it has or they have the effect, or that it might reasonably be expected that the natural and probable consequences thereof would be, that a person perceiving the act, conduct, utterance or publication—
(i) fears for his own safety or the safety of his property or the security of his livelihood, or for the safety of any other person or the safety of the property of any other person or the security of the livelihood of any other person …’
 Intimidation is a single offence. It may occur in various ways, but that does not detract from the fact that all of its manifestations, under both ss 1(1)(a) and (b), deal with the same thing, namely, intimidation. They do not, as suggested by the appellants, give rise to several separate offences. Whether under sub-section (a) or (b) the offence is the same and attracts the same penalties.
 That leads to the next point, which is that the nature of the offence is derived from its name, in the same way as the general nature of theft or murder are derived from their names. The offence is directed at behaviour constituting intimidation and the statutory purpose should be understood as having that goal. A construction that captures conduct that is not intimidatory in character is incorrect as it disregards the very essence of the offence. In argument we were given some examples of speech and conduct, such as colourful political rhetoric, that lacked the essential element of being intimidatory. Far from demonstrating that the section was overbroad, they demonstrated that the interpretation being urged by the appellants was overly literal and inconsistent with the principles set out in paras 89 to 91.
 Intimidation is committed by acts or conduct, or through the spoken or published word. I refer to these compendiously as ‘intimidatory acts’. The proper interpretation of s 1(1) requires that the offence retain the same character in each of these manifestations, that is, it must in all cases be intimidatory. Some intimidatory acts, in the form of threats to person, property or livelihood, are captured in s 1(1)(a), against which no constitutional complaint is levelled. Ms Sonti is charged under sub-section (ii) of that section with threatening to kill someone and burn their house down if they did not withdraw criminal charges against a third party. I hasten to point out that she denies making any such threat. Not all intimidatory acts take this simple form of threats to life, limb or property. Section 1(1)(b) addresses more complex cases. Threats of violence directed at the general public are dealt with in s 1(1A) of the Act, which was introduced at the same time as the amendments to s 1(1)(b). There is plainly some overlap between these two sections, but that need not concern us here.
 Intimidatory acts may manifest themselves in various ways. Seeking to persuade a person to vote for a particular political party, or in favour of strike action, by standing at the entrance to the polling station, catching their eye and drawing one’s hand across one’s throat, simulating a knife cutting their throat, is an example of intimidation by act or conduct. A bank manager who threatened to withdraw a customer’s overdraft if they did not vote for a particular political party, or against a strike at the bank, is an example of intimidation by utterance or, if the threat is in writing, publication. Respondents’ counsel proffered the example of someone in dispute with their neighbour sitting outside the neighbour’s house night after night, ostentatiously loading and unloading a firearm. The writing of anonymous threatening letters of the ‘If you don’t co-operate, I know where you live and where your children go to school’ variety is another obvious example.
 Examples of intimidatory conduct that are particularly apposite to current issues in the world are stalking and harassment. These are specific criminal offences in many parts of the world and stalking and harassment go hand in hand with intimidation and conduct directed at inducing fear in the victim. A good example is the Crimes (Domestic and Personal Violence) Act 80 of 2007 of the state of New South Wales in Australia, which provides in s 13(1) that:
‘A person who stalks or intimidates another person with the intention of causing the other person to fear physical or mental harm is guilty of an offence.’
Fear of physical or mental harm includes fear of physical or mental harm to another person with whom the victim has a domestic relationship. The intention to cause fear of physical or mental harm is established if the accused knows that the conduct in question is likely to cause fear in the other person.
 That statute defines both ‘intimidation’ and ‘stalking’. The former is constituted by:
‘(a) conduct amounting to harassment or molestation of the person, or
(b) an approach made to the person by any means (including by telephone, telephone text messaging, e-mailing and other technologically assisted means) that causes the person to fear for his or her safety, or
(c) any conduct that causes a reasonable apprehension of injury to a person or to a person with whom he or she has a domestic relationship, or of violence or damage to any person or property.’
and the latter is defined as including:
‘the following of a person about or the watching or frequenting of the vicinity of, or an approach to, a person’s place of residence, business or work or any place that a person frequents for the purposes of any social or leisure activity.’
 There is a similar offence in the state of Victoria in terms of s 21A of the Crimes Act 1958. The basis of the offence is that the conduct in question could reasonably be expected to cause physical or mental harm to the victim or to arouse apprehension or fear in the victim for their own safety or that of any other person. The intention to bring about that result is established by showing that the accused knew that this course of conduct (which may be physical or verbal) would be likely to cause such harm or arouse that apprehension or fear, or ought to have understood that it would be likely to have that result.
 My researches have shown that legislation directed at harassment and stalking has been passed in India, Singapore, New Zealand, Scotland and the rest of the United Kingdom. There is specific legislation on stalking in 21 member states of the European Union and Article 34 of the Council of Europe Convention on Preventing and Combating Violence against Women and Domestic Violence obligates all signatory states to enact legislation criminalising stalking. Seven states have not yet introduced legislation. Only Denmark has entered a reservation against this provision, preferring non-criminal remedies. There is similar legislation in all states in the United States of America and in at least some provinces in Canada.
 A common thread running through all this legislation is that conduct that operates to induce fear, or that is directed at inducing fear, in the victim is rendered criminal. It appears to be unusual to specify the subject of that fear, unlike the Act, which refers to fear in regard to personal safety, property or the security of a person’s livelihood. However expressed, it is the intimidatory nature of the conduct that gives rise to criminal liability. South Africa provides civil remedies against stalking and harassment by way of protection orders under the Protection from Harassment Act 17 of 2011, but only the Act, and specifically s 1(1)(b), imposes criminal penalties.
 The examples mentioned thus far involve conduct by individuals directed at other individuals, but intimidation may be aimed more generally at the population at large or specific sections thereof. Smearing pig’s blood on the entrance to a mosque or synagogue, accompanied by anti-Islamic or anti-Semitic slogans, provides an example. So does the example postulated by Justice Holmes of a person falsely shouting ‘Fire’ in a crowded theatre. Another is someone sending a threat to the media that, unless a prisoner is freed from gaol, the water supply of a city will be poisoned, an explosive device triggered, or fresh food, medicine or baby food in shops contaminated. Where such threats are made to obtain money they will usually constitute the offence of extortion, but when they are pursuant to social demands, such as the release of prisoners, or the closure of clinics providing reproductive help advice, they would not ordinarily do so. Nor would threats made by way of a hoax intended to induce fear and panic, such as a bomb threat at a sports stadium or a shopping centre, attract criminal consequences apart from the Act.
 All of these examples fall within the ambit of s 1(1)(b). The suggestion in argument that they are all encompassed by the crimes of crimen injuria, assault and public violence is incorrect. At points there may be some overlap between them and intimidation, but none of the examples in paras 97 to 103 are covered by those common law crimes. I should mention crimen injuria in particular because of the suggestion that it encompasses stalking. In the absence of any suggestion of sexual impropriety, the only case I have found of the type commonly regarded as stalking resulting in a conviction of crimen injuria, is a 1923 decision in which an older man pursued a young woman around a public library. The court held that it was a marginal case and in the similar case of Ferreira, where the accused on five separate occasions followed women who were unknown to him, whilst making innocuous remarks, the convictions were set aside on appeal.
 The discussion of crimen injuria in the textbooks reveals it to be a crime of uncertain ambit, dependent on perceived infringements of the vague concept of dignitas. Its own vagueness may render it liable to constitutional challenge and it is not concerned with inducing fear but with infringements of personality rights. It cannot be said with any certainty that it encompasses stalking in all its many manifestations and it is preferable that this be dealt with by statute. The need for statutory intervention to deal with intimidation is apparent from the background to the introduction of legislation dealing with harassment and stalking in other countries. It was generated by the inadequacy of common law crimes to deal with intimidatory behaviour causing fear. As already mentioned, there is no challenge to the appropriateness of criminalising such behaviour.
 I have gone into this in a little detail in order to illustrate the multifarious ways in which intimidatory conduct can manifest itself and the necessity for appropriately broad language to be used in a statute criminalising such behaviour in order to encompass the full range of conduct that is intimidatory. Such language needs to be sensibly and sensitively construed bearing in mind its potential to limit constitutional rights, but neither its breadth nor its complex drafting is a reason to contend that it is constitutionally defective. It is only if, when properly construed, the provision infringes the protection of freedom of expression in s 16 of the Constitution that the complaint of over-breadth may be justified. There is no complaint in the present case of the section being invalid on the ground of being impermissibly vague.
The grounds for the constitutional challenge
 Three reasons were proffered in support of the proposition that s 1(1)(b) infringes s 16 of the Constitution. First, it was said that it criminalised any expression that induced subjective feelings of fear in the persons at whom such expressive actions or speech were directed, irrespective of whether that fear was reasonable. Second, it was submitted that the section created ‘no fault’ liability. Third, it was submitted that in any event the section criminalised political hyperbole; emotionally charged rhetoric in the context of both political and industrial action and legitimate public and social aims, such as, the advocacy of radical land redistribution, consumer boycotts; and campaigns for the removal of politicians, however errant, from office. I will deal with each in turn.
 The first submission misreads the section. An intimidatory act is, or intimidatory acts are, criminal if ‘it has or they have’ the effect of inducing fear, or ‘that it might reasonably be expected that the natural and probable consequence would be’ that it or they would induce fear in a person perceiving that behaviour. The language covers two general situations. The first is where someone complains that intimidatory conduct induced fear in them personally. The second is where the threat is not directed at individuals, but at the public generally or a section thereof, such as the general threats described in para 103. In the latter case the person who makes the threat can be prosecuted without it being necessary for any one person among the general body of the public, to say they experienced fear as a result of the threat. It suffices for the prosecution to establish that this would reasonably have been expected as the natural and probable consequence of the threat. Take the case of a threat to detonate a bomb communicated to a television or radio station, but not broadcast publicly, so as to afford the police time to find and arrest the perpetrator. The threat may not have induced fear in the staff of the station because they trusted the police to catch the perpetrator. In the absence of the second case it would not be possible to arrest or prosecute the perpetrator unless the threat was broadcast and induced actual fear in some citizens. The ‘reasonable expectation’ relates to that latter situation, where it can reasonably be expected that if the threat had reached its intended audience it would have induced fear.
 The appellants argue that, in a case where fear is induced in an individual, all that is required for a conviction is subjective fear on the part of the complainant, while in what I have called the second case the fear needs to be reasonable. I do not agree. The section requires either that fear be induced, or that it might reasonably be expected to be induced as the natural and probable consequence of the intimidatory act. Appellants accepted that the latter case postulates an objective test of reasonableness. Fear only qualifies if it would reasonably be expected to arise. This is reinforced by the requirement that the fear be expected as the natural and probable consequence of the intimidatory act. Subjective fear that might be induced because some people are of a nervous disposition or a ‘timorous faint-heart always in trepidation lest he or others suffer some injury’, would not qualify. Why should the position be any different when the nervous person or timorous faint-heart comes forward to say that the intimidatory act induced fear in them? I can think of no good reason for differentiating between the two situations and none is evident from the language of the section. It introduces inconsistency without reason.
 That reasonableness is the yardstick by which to measure the existence of genuine fear was the approach of Leach J in Holbrook. The appellant, a young man under the influence of liquor went for a swim at night at the flat where he lived. Regarding his neighbour’s cat as a nuisance he threw it into the pool. A row ensued with the cat owner who told him that she would phone the agent and have him evicted. She ignored his pleas for her to reconsider and the row escalated to the point where he said ‘I’ll kill you, you bitch.’ She ignored this and went to her own room, where she put the cat down, armed herself with a revolver and went back to confront Holbrook, not to protect herself, but to induce him to stop pleading with her not to have him evicted. Leach J held that Holbrook’s conduct did not fall within the section and was no more than verbal abuse uttered under the influence of liquor. It was incapable of inducing fear in a reasonable person and in fact had not reasonably induced fear in the complainant.
 The appellants submitted that the use of the word ‘or’ between ‘it has or they have the effect’ and ‘it might reasonably be expected’ justified an interpretation that the former dealt with subjective fear and the latter with reasonable fear. That was a slender reed on which to support so far-reaching a conclusion. Grammatically ‘or’ is the natural way in which to introduce a provision dealing with general rather than specific threats as explained earlier. To leap from there to the contention that its use clearly distinguished between two situations, one in which actual fear, whether reasonable or not, was induced, and another in which reasonable fears might reasonably be expected to be induced, is fanciful. No reason was advanced for initially creating an offence based on inducing subjective and potentially unreasonable fears and, in the next breath in the same section, restricting the alternative manifestation of the same crime to reasonable fears. That alone is improbable, but to suggest that it was achieved by using the common conjunction ‘or’ to separate the two situations was perverse.
 Creating a crime that depended on fear being induced in the mind of the victim gave rise to questions of interpretation. The fear must obviously be genuine. That much is common cause. Would any subjective, albeit fanciful, fear suffice? Even without the second part of the section the answer must surely be ‘No’. Only a fear that was reasonable qualified. Any other answer would create the possibility of prosecutorial manipulation of the charge.
Take the case of a speaker at a political rally saying: ‘The land is ours. Whites must give it back or we will take it.’ A prosecutor, concerned that this piece of political rhetoric would not reasonably be expected to induce fear, could seek out an individual who claimed to have seen the speech on the television news and feared for their property as a result. On the appellants’ argument the speaker could be convicted because of that person’s subjective fear, even though a conviction could not be obtained on the ground of a reasonable expectation that fear would be induced by the speech. That is not a sensible construction of the section. This possibility alone points strongly in favour of an objective construction requiring proof that the fear induced was reasonable in both circumstances.
 A closer examination of the second situation for which the section provides reinforces this conclusion. Here criminal liability arises if it can ‘reasonably’ be expected that the natural and probable consequence would be to induce fear. It can only be reasonably expected that this will occur if the fear is reasonable. One does not reasonably expect unreasonable fear. That is reinforced by the requirement that this be the natural and probable consequence of the act, conduct, utterance or publication in question. The connection required is direct and only a reasonable inference that fear will be induced suffices. The appellants accepted that this related to reasonable fears.
 I have already made the point that the section creates only one offence namely intimidation. The appellants’ argument requires that the section be read as saying that criminal liability attaches if a person subjectively, but unreasonably, fears certain consequences and also where it is reasonable to infer as the natural and probable consequence of the accused’s actions that reasonable fear would be induced. That gives rise to an internal contradiction, where the creation of sometimes reasonable, and sometimes unreasonable, fears would attract criminal liability. Such an interpretation must be rejected in accordance with basic principles. The proper interpretation of the section requires that the fear that is induced is fear that would be induced in a reasonable person by the actions in question.
 This conclusion is reinforced by the constitutional protection afforded freedom of expression. Ours is a society where debate is perforce vigorous, passions run high and language and expressive acts may be blunt to the point of abuse. The appellants in argument cited the decision in this Court in Hotz as an illustration of this and of how broad the parameters of constitutionally protected expression are. That being so there is no reason to hold, in the context of the Act and its prohibition on intimidatory conduct inducing fear, that a subjective, but unreasonable, fear suffices, making criminal liability dependent on the vagaries of the complainant’s personal predisposition. A conviction of intimidation should not depend upon the subjective feelings of the more timorous individuals among us and s 1(1)(b) should not be construed in this way. Properly construed both manifestations of the crime of intimidation provided in this section require the fear induced by, or reasonably expected as the natural and probable consequence of, the intimidatory act, to be reasonable fear, not subjective fear.
No intention to induce fear
 This argument built upon the proposition that the fear induced by an intimidatory act did not have to be reasonably entertained. It was then submitted that the section catches expressive acts that are not intended to create fear. It was difficult to ascertain the scope of this argument, but in the heads of argument appellants submitted that the section leaves no room for any mens rea requirement and that it clearly evinced an intention to create no fault liability. I will deal with it on that footing.
 We were not referred to any of the authorities cited in para 90 dealing with the requirement of intention (mens rea) in statutory offences. The appellants misstated the correct approach. Their heads of argument said that courts read down penal statutes where possible to require some sort of fault. That waters down the proper approach beyond recognition. The correct approach is that mens rea is presumed to be required in the absence of clear and convincing indications to the contrary in the enactment in question. That is all the more the case where the statutory offence is one attracting substantial potential penalties.
 There is nothing in the section to suggest that mens rea is not required. The only argument presented to us was that where the charge was based on what might reasonably be expected as the natural and probable consequence of an intimidatory act, absolute liability was intended because actual fear does not have to be induced in a specific person. But that misunderstands this requirement. The requisite inference can only be drawn if a specific group of people within the population or the population at large can be identified as the target of the intimidatory act. The trial court must be satisfied on reasonable grounds that the natural and probable consequence of the intimidatory act would be to induce fear in members of that group. In order for there to be a conviction, the court would have to be satisfied beyond a reasonable doubt that, if members of that group had perceived the intimidatory act, they would in fact as a result have feared for their personal safety, or that of their property or livelihood. So it is incorrect to say that actual fear is not required and this is a case of no fault liability without the need to prove mens rea. It must be remembered that mens rea has to do with the state of mind of the accused, not the consequences of the accused’s actions. An intention to induce a state of fear is entirely compatible with a failure to achieve that purpose, although that would raise the question whether a conviction of attempt, rather than intimidation, would be the proper verdict.
 The argument proceeded on the basis that no mens rea of any type was required, so we have not had the benefit of argument on whether dolus or culpa would be required. The serious nature of the offence and the potential severity of the sentences that can be imposed point strongly in the direction of it being dolus, an intention to induce fear or an anticipation that fear would be produced and continuing reckless of whether it was. That is reinforced by the potential effect on freedom of expression. The only factor pointing away from that conclusion is the use of language that is frequently encountered when dealing with culpa, that is, the foresight of the reasonable person, rather than the subjective foresight and reckless continuation with the conduct in question that may amount to dolus eventualis.
 This difficult question need not be resolved in the present case. It suffices to say that the offence is not one of strict liability. Intention, either in the form of dolus or culpa, is a requirement for conviction. Choosing between the two may raise constitutional issues that were not ventilated before us and it is preferable to go no further than saying that intention (mens rea) is a necessary ingredient of the offence of intimidation.
The section criminalises conduct protected by s 16(1) of the Constitution
 Section 16(1) of the Constitution guarantees the right of freedom of expression, including in particular the freedom to receive or impart information or ideas. That is subject to the qualification in s 16(2) that this freedom does not extend to propaganda for war; incitement of imminent violence; or advocacy of hatred based on race, ethnicity, gender or religion, and that constitutes incitement to cause harm. I need not emphasise the importance of this right in our democracy. The Constitutional Court has repeatedly asserted it in ringing terms. The question is whether the provisions of s 1(1)(b) infringe that right. That is the primary issue. A justification analysis is only reached if the answer to that question is in the affirmative.
 The injunction in s 39(2) of the Constitution is that we should interpret the section in accordance with the spirit, purport and objects of the Bill of Rights. The jurisprudence of the Constitutional Court says that we must, where the language of the statute fairly permits, choose a constitutional rather than an unconstitutional meaning. Bearing that in mind, the short answer to the appellants’ contention is that by definition constitutionally protected expression lacks the necessary quality of being intimidatory and is lawful. It is lawful and protected by the supreme law. It is not conduct directed at inducing fear in any of the respects referred to in the section. Neither the intention that I hold is necessary in order to commit the offence, nor the intimidatory purpose that is likewise in my view a requirement, is present. Let me expand on this.
 In common parlance the concept of ‘intimidation’ conveys various shades of meaning. A child may find their teacher intimidating and a university student may regard an examination as intimidating. Most junior advocates find their initial encounters with judges intimidating. Some people are intimidating because of their position, reputation, fame, or the fact that they hold high office or have achieved great things. But none of these instances is intimidation within the meaning of the section. It is intimidation only in its most general sense. That is made clear by the requirement that the offence is only committed when fear for physical safety, the safety of property or the security of livelihood is induced or might reasonably be expected to be induced. The understanding of ‘intimidate’ that informs the section is ‘to discourage, restrain or silence illegally or unscrupulously; as by threats or blackmail’. Intimidation is:
‘the action of intimidating someone, now esp in order to interfere with the free exercise of political or social rights; the fact or condition of being intimidated.’
 The appellants’ argument attributed to the concept of fear the meaning of a sense of worry, anxiety, nervousness, concern or apprehension, however, modest or restricted. Again I do not regard that as justified either by the language of the section or its context and purpose. The fear with which the section is concerned is a real belief that the individual concerned will suffer imminent harm in consequence of the intimidatory act.
 All human beings suffer from daily anxiety or concern about the state of the world, what Shakespeare referred to as ‘the slings and arrows of outrageous fortune’ and its ‘sea of troubles’, but, to construe a statutory provision dealing with intimidation inducing fear for life, limb and property on the basis that it is concerned with those anxieties, rather than belief that danger and harm is imminent is not in my view appropriate. Were the criminal standard that low, I would, in our constitutional disposition, regard it as overbroad. That brings me back to the basic principles of interpretation that we are enjoined to apply in this case namely, a reasonable interpretation consistent with constitutional norms and preferring a constitutionally compliant construction to one that is non-compliant.
 Turning to the various instances advanced by the appellants in support of their argument, they started with the example of a consumer boycott, first raised in the parliamentary debate when the Act was introduced in 1982. The section did not then contain section 1(1)(b). Mr A S K Pitman MP claimed that it criminalised any form of consumer boycott. The response was that this ignored the context of the Bill as a whole, which required an unlawful purpose and a very specific intention ‘with a view to the aims in this Bill’, that is, intimidation. A consumer boycott is generally speaking – I leave open the possibility of it being invoked for illegitimate reasons, such as xenophobic attacks on refugees – an entirely legitimate form of protest in pursuit of legitimate ends. It is not intimidatory, merely because it seeks to impose some level of coercion on the target of the boycott to alter their behaviour in some way, and it is not conducted with the intent to intimidate. It is trite that the field of labour relations involves the exercise of coercive power, especially by employers over employees, but also by trades unions against employers. Much commercial activity in society may have the same effect. Yet even the most extreme submissions did not suggest that the section encompassed these activities.
 Reference was made to three cases in which the section has come before the courts where comments were made about its apparent breadth. The first was Holbrook dealt with in para 110, but it does not support the appellants. Contrary to the submission it held that the fear induced by intimidatory acts needed to be reasonable. It rejected an unduly literal approach as bringing about absurd results. We were referred to a passage where Leach J referred to the tortuous language of the section. But obscure or complex or even meaningless language is not a ground of constitutional invalidity. The remedy is a rigorous and correct approach to its interpretation. When that comes from a court of binding authority it will operate as a salutary deterrent to prosecutions advanced on an insubstantial basis.
 Motshari was another case where the prosecution arose in circumstances far removed from the purposes of the Act. It was a domestic dispute where the accused discovered on his return from serving a prison sentence that certain furniture in the home had been damaged and other furniture repossessed, for which he blamed his partner. In ranting at her he threatened to kill her. Kgomo J, after analysing the history of the Act held that its provisions did not apply to that situation. As had occurred in Holbrook the judge adopted a sensible approach to the scope and ambit of the section and correctly held that it was inapplicable to the type of domestic dispute with which he was concerned.
 The judgment in Cele appears to be the source of the idea that the section creates several different offences. That led the court to conclude that on its literal meaning no intention to induce fear was required to commit the offence, and that it was irrelevant whether the fear, giving rise to the charge, was reasonable. For the reasons set out above both propositions are incorrect.
In fairness to the court, it went on to hold that a restrictive interpretation was called for and that intention to commit one of the acts specified in the section was a requirement. I agree. The case arose out of an industrial dispute where, in the course of a heated row, the accused had said to the complainants, their superiors employed at the prison, that ‘We will crucify you.’ The court acquitted the three appellants on the ground that this was not intended literally and that it could not reasonably be construed as conveying that physical harm to life, limb, or property was intended.
 In all three of these cases the courts acquitted the appellants on the basis of elements of the same kind of principled, sensible, constitutionally compliant interpretation of s 1(1)(b) as in my view should be given to the section.
Prosecutions under the Act should not have been pursued in any of them, but in each case sense prevailed when the matter came before the high court. The existence of occasional foolish prosecutions is not, however, a reason for holding the section to be unconstitutional.
 I venture to suggest that the same would have occurred if any of the examples of the offence postulated in an academic article cited in these cases had ever seen the light of day in a court. The first of these was that of a policeman monitoring a picket by singing and dancing workers. The postulate was that the policeman might unreasonably fear for the safety of passers-by and try to break up the picket, or call for reinforcements, or make a report to a superior officer. The example fails at every level. Unreasonable fear does not justify a conviction and it was so held in Holbrook. The policeman’s response to the picket, which was only a relevant consideration when s 1(1)(b)(ii) was part of the section, which it no longer is, would not have been induced by fear, but by the obligation to perform police duties and safeguard the public. Lastly, the picketers were not intending to induce fear for the safety of the passers-by, so they could not, as the authors suggest, be convicted of an offence they were unaware they were committing. That is why mens rea is a requirement of the offence. Lastly, in our constitutional dispensation, where there is statutory protection for picketing activities under the Labour Relations Act 66 of 1995 (the LRA), the example has ceased to have any relevance. One cannot construe as unlawful, conduct that is specifically sanctioned by law.
 The other example given by the authors was that of a motorist seeing a graffiti artist about to deface a wall and, fearing damage to property, hooting to alert the owner of the property to what was happening. It is unnecessary to spend much time on it. The conduct by the graffiti artist is not intimidatory or intended to be intimidatory. Like Holbrook and Motshari it is a case to which the Act does not apply.
 I am aware that in Holbrook, in the light of submissions made to the court that the section was overbroad with reference to the academic article just discussed, the court referred its judgment to the Law Commission for consideration. It expressed a general concern about the potential scope of the section, but, understandably, without undertaking the detailed analysis undertaken here. Those general concerns were echoed in Motshari and Cele, as well as Gabatlhole, none of which contains a detailed exercise interpreting the section as a whole as this court has been compelled to do. Gabatlhole was a case where a burglar apprehended by the householder repeatedly said that he would return with his ‘bandiet tjommies’ (criminal friends). The conviction was set aside on technical grounds that do not affect the present case. However, echoing what had been said in Motshari the court expressed doubt whether the case fell within the Act.
 In Gabatlhole Majiedt J drew attention to the fact that s 1(1)(a) refers to intimidatory acts performed ‘without lawful reason’ and suggested that the same requirement of absence of a lawful reason for conduct is also required by s 1(1)(b). This is in accordance with basic principles of criminal liability that if the accused has a lawful reason for acting there can be no criminal liability. If so it reinforces the conclusion that constitutionally protected conduct and conduct authorised by law cannot be intimidatory for the purposes of the crime constituted under s 1(1) of the Act. In the absence of unlawfulness there can be no criminal liability. This point is well made by Professor Snyman when saying:
‘The mere fact that there is an act which complies with the definitional elements does not mean that the person who performs the act is liable for the particular crime. Satisfying the definitional elements is not the only general requirement for the particular crime. The next step in the determination of liability is to enquire whether the act which complies with the definitional element is also unlawful.’
 There is no numerus clausus (closed number) of grounds of justification for conduct that would otherwise be unlawful. The fact that s 1(1)(b) does not contain the express provision in s 1(1)(a) that the conduct criminalised must have been undertaken without lawful reason, does not mean that if there was a lawful reason for such conduct that would not protect the perpetrator from criminal liability. Accordingly if the conduct said to constitute intimidation is objectively lawful for some reason, for example, it is conduct sanctioned by other legislation, such as the LRA or, to give another example close to the issues raised by the appellants, participation in a gathering or demonstration authorised in terms of the Regulation of Gatherings Act 205 of 1993, it cannot constitute the criminal offence of intimidation.
 Given that the appellants’ argument is that the section is constitutionally invalid because it infringes the right of freedom of expression under s 16(1) of the Constitution, the requirement of unlawfulness in order for the conduct to constitute intimidation provides an immediate stumbling block in the path of the argument. Conduct that is lawful, because it is constitutionally protected freedom of expression cannot at the same time be unlawful. The case of Watts in the Untied States, to which I will refer below, makes this clear.
 The heads of argument contain some general statements of the kind of expressive conduct that it was submitted would be within the ambit of the section. They said that the section:
‘… criminalises a vast quantity of everyday political speech. There are innumerable statements that might create a state of fear in another. Many struggle songs and political slogans are actionable under its terms. Threatening to sue someone, or to have them punished for something they did wrong, will also often constitute an offence. Pressing for land redistribution through expropriation would put many people in fear for their property. Even campaigning to remove an unpopular politician is actionable under section 1(1)(b), as that politician might reasonably fear for his livelihood if he was removed from office.’
It was submitted that the problem with the section is that it obliterates the distinction between ‘true threats’ and ‘political hyperbole’, so that emotionally charged rhetoric with no serious intent to cause harm comes within its compass. This encompassed, so the argument proceeded, anything from popular struggle songs, to trite political slogans, such as ‘there will be blood in the streets’ all of which are rendered unlawful.
 None of these examples fall within the ambit of the section when it is construed, as in my view it must be, as requiring mens rea; only applying to intimidatory conduct properly so called; requiring that the fear relied upon is both genuine and reasonable and based on a fear of imminent harm, not a general state of nervousness, concern or apprehension. Lastly, if the conduct is lawful in the sense of enjoying either constitutional or statutory protection then it is not intimidatory. All of the examples proffered by the appellants would escape criminal liability in accordance with this construction of the statute.
 Most surprising, in the light of the somewhat overheated rhetoric in which the appellants’ argument was couched, is that, so far as I am aware, no-one has suggested that charges of intimidation could or should be brought against persons singing ‘Umshini Wam’, anymore than in days past it was seriously thought that singing ‘We shall overcome’ would induce fear. Nor have any of the many politicians and public figures, who in recent times have been denounced as scoundrels, thieves, criminals and deserving of gaol, thought of laying a charge of intimidation against their critics.
The reasons are obvious. None of this is intimidatory. None of it is incompatible with the right of freedom of expression. None of it is intended to intimidate, as opposed to campaign, or persuade, or expose to public scrutiny, by open, noisy and public means. None of it can reasonably induce fear in its intended sense, of imminent harm to life, limb, property and livelihood, as opposed to nervousness, concern or apprehension about what the future may hold. When criminals apprehend going to gaol for their crimes or forfeiting their ill-gotten gains that is not fear in terms of the section. The suggestion that a threat to sue someone is intimidation, or that to expropriate their property for the purpose of land redistribution – something that is expressly provided for in sections 24(5) to (8) of the Constitution – can be treated as intimidation is, with respect to counsel who made that submission, far-fetched.
 Counsel sought to use the allegations in Mr Moyo’s case to illustrate their point. That was unwise and I refrain from dealing in detail with them because that will be a matter for the trial court. If his conduct does not constitute the offence of intimidation within the parameters I have outlined he will be acquitted. Likewise if the charge sheet is defective in the light of those parameters it can be set aside. It would be inappropriate for this court in its judgment to express any view as to whether the various allegations in the charge sheet are capable of constituting the offence as that would pre-empt the function of the trial court. Whether his conduct can be construed as protected expression or intimidatory, involving a threat of violence, depends upon the context in which it occurred, as indeed is almost inevitably the case with alleged intimidation. This was conceded. The argument was based on the propositions that a conviction does not require proof of intention or that any fear to which the complainants may testify be reasonable. For the reasons given earlier I regard both propositions as incorrect.
 The approach to the interpretation of s 1(1)(b) that I hold to be correct is the same as the approach adopted by the United States Supreme Court to the legislation in Watts.
The statute in question created the criminal offence of ‘knowingly and wilfully … [making] any threat to take the life of or to inflict bodily harm upon the President of the United States’. Mr Watts, aged 18, was participating in a small group of young people discussing police brutality during a public gathering at the Washington Monument, and said in the context of his having been drafted to serve in Vietnam: ‘if they ever make me carry a rifle the first man I want to get in my sights is L.B.J.’
 Mr Watts’ conviction under the statute was set aside by the Supreme Court in a per curiam opinion holding that the statute was constitutional, but needed to be interpreted ‘with the commands of the First Amendment clearly in mind’.
A ‘threat’ had to be distinguished from constitutionally protected speech. That is my approach, namely, that expression sanctioned by s 16(1) of the Constitution needs to be distinguished from intimidatory acts. The statute in issue in Watts referred to a threat, which seems to correspond with an intimidatory act as I have referred to the acts, conduct, words or publications in the Act, and the Court said that: ‘We do not believe that the kind of political hyperbole indulged in by the petitioner fits within that statutory term’. However, a similar statement by a member of a right wing militia, protesting against laws restricting the right to bear arms, might well have justified a conviction.
 The other American case referred to by the appellants is far removed from the present one on its facts. It concerned the imposition of tortious (delictual) liability for conspiracy in respect of business losses suffered during a consumer boycott of White owned stores. Its only relevance is the finding by the Supreme Court that such a claim could not succeed insofar as the boycott and the damages suffered by the claimants arose from speech, which in America encompasses all forms of expressive conduct, protected by the First Amendment. Liability could only arise as a result of criminal conduct or statements not qualifying for protection under the First Amendment. I agree that this would be the case in similar circumstances in South Africa if a charge were brought under s 1(1)(b) of the Act arising out of a consumer boycott or similar action protected by the right to free expression.
 It follows that I do not accept the submission that s 1(1)(b) encompasses cases of conventional and protected freedom of expression as suggested in the examples proffered by the appellants in support of that contention. That would only be the case if the section were interpreted to cover such cases, an interpretation that in my view is inconsistent with the applicable principles of statutory interpretation and the Constitution.
General and conclusion
 Some play was made in argument of the origins of the Act in our unsavoury past and the malign intentions of the amendments to s 1(1)(b) directed principally at organisations campaigning for an end to apartheid and the trade union movement. It is unnecessary for me to canvas that history as it is amply set out in the article by Plasket and Spoor and the judgment in Motshari. I have borne it in mind but do not think that it can be decisive. The only reported judgment in a case arising from a work stayaway is that of Mahomed J, as he then was, in Ipaleng, where the appellant’s conviction was set aside on the basis that there was inadequate evidence of any intimidation. There is no other evidence that the section has been used for the purposes that gave rise to concern in the labour law community at the time of its enactment.
 Were there evidence of widespread use of the section to stultify political debate or hamper trade union and worker activities, this would raise concern and possibly illustrate defects in the section that are not apparent to me from the argument addressed to us, but there does not appear to be any. The vague and inconclusive allegation expressed in identical words by Ms Sonti and Mr Moyo, save for the substitution of the Centre of Applied Legal Studies (CALS) for the Socio-Economic Rights Institute (SERI), that ‘SERI is regularly approached for advice and support by people who are charged with intimidation because their community organising or political activities have been alleged to intimidate others’ is unhelpful. That is all that is said on behalf of Mr Moyo with no indication at all of how extensive this is said to be. Ms Sonti goes further to say that courts have frequently set aside convictions for failure to explain the reverse onus to an unrepresented accused, but no detail is given of this and it is not reflected in the law reports. In the absence of greater detail it is impossible to conclude that there is a widespread use of the Act to stifle legitimate expressions of view.
 The Act was preserved by item 2(1) of Schedule 6 to the Constitution, subject to consistency with the Constitution. That not only meant that it was preserved to the extent that it was consistent with the Constitution, but reinforced the injunction that it should be construed, so far as possible within the limitations of the text, in a manner consistent with the Constitution. There has been no move to repeal it, and it features in other legislation, for example, as one of the offences in Schedule 1 to POCA. This may convey a limited measure of parliamentary consideration and approval, at least of the need generally for the crime of intimidation.
 For those reasons I conclude that s 1(1)(b) passes constitutional muster. I would dismiss the appeal by Mr Moyo. Were I to take a different view, however, my view is that there is a proper case to be made for legislation rendering criminal a range of conduct falling within s 1(1)(b) that is not suggested to be constitutionally offensive. To invalidate the section without leaving open the possibility for prosecutions of stalking and harassment, while parliament considered the possible introduction of amending legislation, would remove a protection that vulnerable people, especially women, enjoy at present. That is in my view undesirable.
I would therefore suspend the operation of any period of invalidity for a period of two years, subject to a provision that during the period of suspension it would be a defence to a charge in terms of the section that the accused was exercising the right to freedom of expression conferred by s 16(1) of the Constitution.
 Section 1(2) provides that:
‘(2) In any prosecution for an offence under subsection (1), the onus of proving the existence of a lawful reason as contemplated in that subsection shall be upon the accused, unless a statement clearly indicating the existence of such a lawful reason has been made by or on behalf of the accused before the close of the case for the prosecution.’
The main judgment holds that this is a reverse onus provision placing the onus of proof of lawfulness on the accused and hence constitutionally impermissible. I would agree were it not for the words:
‘unless a statement clearly indicating the existence of such lawful reason has been made by or on behalf of the accused before the close of the case for the prosecution.’
In my view the presence of those words prevents this from being a reverse onus provision of the type that has been condemned in a number of cases by the Constitutional Court.
 A reverse onus is constitutionally objectionable because it infringes the presumption of innocence that is part of the common law and now enjoys constitutional protection under s 35(3)(h) of the Constitution. Its corollary is that it is the prosecution’s task to prove the guilt of the accused to the applicable criminal standard of proof, usually beyond a reasonable doubt. Where the onus is reversed, so that the accused has to prove (or disprove) something in order to escape conviction, they are no longer presumed to be innocent. The possibility exists of their being convicted notwithstanding the presence of a reasonable doubt in regard to the elements of the offence. No citation of authority is necessary for the proposition that a presumption of that character is unconstitutional.
 Sub-section 1(1)(a) commences with the words ‘without lawful reason’ so it is apparent that s 1(2) applies in that case. There are no corresponding words in s 1(1)(b), but as discussed in para 134 there is merit in the suggestion in Gabatlhole that their absence does not mean that an offence can be committed under that provision even if the accused had a lawful reason for their conduct. Accordingly, I assume that in both instances the crime of intimidation is committed where the conduct is without lawful reason.
 If s 1(2) placed the burden of proof of lawfulness on the accused in all instances, I would regard that as a reverse onus provision. But the opening words of s 1(2) are qualified by the rider that there is no such onus if accused persons have at any time during the course of the prosecution’s case made a statement clearly indicating that they contend that they acted with a lawful reason and describing the nature of that reason. In that event in order to obtain a conviction the prosecution would have to prove that they acted without a lawful reason. The prosecution would have to prove that the reason was factually unfounded or, if factually correct, did not constitute a lawful reason for the accused’s conduct.
 I am unable to see on what basis it can be said that one and the same offence may sometimes require the prosecution to prove the accused’s guilt beyond reasonable doubt and sometimes not. What would happen in the case of two accused being accused of intimidation arising out of a single event and one disclosed a lawful reason before the close of the prosecution case, while the other did not until after the close of the prosecution case? If s 1(2) imposed a true reverse onus then, if the facts constituting the lawful reason were unclear from the evidence, the one accused would be acquitted and the other not. That is absurd and leads inevitably to the conclusion that this is not a true reverse onus provision. What then is it?
 The answer is that it is a provision addressing an evidential issue. If a lawful reason is disclosed before the end of the prosecution case, the prosecution will have to lead evidence to disprove it before closing its case. If that evidence is unsatisfactory the accused may then obtain a discharge at the close of the prosecution case without being put on their defence and without having to decide whether to give evidence. The effect of the presumption is that, if no lawful reason is disclosed prior to the close of the prosecution case, the accused will not be able to seek and obtain a discharge on the basis that the prosecution has failed to show that they acted without a lawful reason. The evidential burden will then be imposed upon them to produce evidence of the lawful reason.
 An evidential burden does not impose a reverse onus, nor is it a per se case of a constitutional infringement. Nonetheless there is a constitutional problem with the section. It contravenes the provisions of s 35(3)(h) of the Constitution, not because it infringes the presumption of innocence, but because it places improper pressure on an accused to forego their constitutional right to silence and not to give self-incriminating evidence. That is inconsistent with the broader right to a fair trial, because it relieves the prosecution in the first instance from the need to lead evidence to show that the actions of the accused are without lawful reason and, after the close of the prosecution case, it constrains the accused to give evidence themself or to lead evidence from others. As such it infringes the constitutional right in a more insidious way in that it operates as a compulsion on the accused to disclose at an early stage of the proceedings what may be the key element of their defence. Indeed, as correctly pointed out in para 62 of the main judgment, it may go so far as to compel the accused to make admissions that relieve the prosecution of the obligation to prove certain facts. So it infringes the right to a fair trial and probably the right to silence. As Ms Sonti said in her replying affidavit it seeks to force the accused to break their silence before the close of the prosecution case.
 I agree with my colleague that there is no basis upon which this constitutional infringement can be justified as a permissible limitation of rights under s 36 of the Constitution and that no purpose would be served by suspending the order of invalidity. The declaration of invalidity must be retrospective to the extent that the conviction in any pending trial or appeal is dependent upon the invocation of the provisions of s 1(2), but not otherwise. Cases where the appeal process has been exhausted should not be affected by the order of invalidity.
 I made the point in the opening paragraph of this judgment that as a result of these proceedings Mr Moyo’s trial has been delayed for some six years and will be delayed even further while these proceedings are taken further as is inevitable in the light of my colleague’s judgment. Ms Sonti’s trial has been delayed for about one year less. This is most unsatisfactory as it means that their criminal trials have not been brought and concluded without undue delay as required by s 35(3)(d) of the Constitution. It has not only created a situation where the criminal charges continue to hang over their heads, but is a denial of justice to those who made the allegations on which those charges rest. They are legitimately entitled to ask why their allegations have not been brought before a court and their complaints heard and determined by an impartial judicial officer.
 In s 35 the Constitution guarantees a range of rights to arrested, detained and accused persons. Section 35(3) guarantees to all accused persons the right to a fair trial. That is secured in practice by the provisions of the Criminal Procedure Act 51 of 1977 (the CPA). The appellants do not seek to impugn the provisions of the CPA in any way, yet they are seeking to assert their fair trial rights before a civil court. That should give pause for thought. Why are issues germane only in the context of criminal proceedings being canvassed and determined in civil proceedings and not in the constitutionally compliant forum, and in accordance with the constitutionally compliant statute, provided for the adjudication of criminal cases?
 The appellants’ response to this question is to say that the Constitutional Court has held in Savoi that they have standing to bring the present proceedings. Savoi involved confirmation proceedings where the Constitutional Court was obliged to accept jurisdiction. The issue arose indirectly because there was also an application for leave to appeal against the high court’s refusal of orders of constitutional invalidity in respect of certain portions of the legislation under consideration. In the present case the issue is not one of standing, but solely one of timing and procedure. At an appropriate stage and in appropriate proceedings a person charged with a statutory offence obviously has standing to challenge the constitutionality of the statute under which they have been charged. The concern in this case is that it has been done outside the ambit of the criminal proceedings, which is the only place where the constitutionality of the legislation is in issue. It is an abstract challenge and, as Madlanga J rightly said in par 13 of Savoi, courts generally and rightly treat abstract challenges with disfavour.
 The case of Ms Sonti shows how abstract the challenge to s 1(2) was. The charge sheet alleged that she without lawful reason and with the intent to compel the complainant to withdraw certain criminal charges threatened to kill the complainant and/or to burn her house down. There was no indication, as is required, of any intention on the part of the prosecution to rely on s 1(2) of the Act. That is hardly surprising, as proof that she made such a threat would, prima facie at least, be unlawful. Ms Sonti said in her founding affidavit that she knew the complainant, and had some interaction with her, the nature of which she explained. She denied threatening to kill her or burn her house down. Nothing in the explanation of her interaction with the complainant suggested that it would render lawful the alleged threats if they were made. She denied threatening or intimidating the complainant. If the State fails to prove the threats she must be acquitted. If it transpires that she did make them, it is difficult to see on what basis she could claim to have had a lawful reason for doing so. The challenge to s 1(2) is therefore wholly academic on the facts of this case and, in the absence of evidence that reliance is being placed on s 1(2) in other cases, it is academic there. That brings to mind what Ackermann J said in Ferreira v Levin:
‘… cases for relief on constitutional grounds are not decided in the air … The time of this Court is too valuable to be frittered away on hypothetical fears …’
 In the early days of our constitutional jurisprudence under the 1994 Constitution, where the grounds upon which cases could come before the Constitutional Court were different from the present grounds, a case similar to the present one was brought before the then Natal Provincial Division asking that the constitutionality of a reverse onus provision be referred to the Constitutional Court. The application was refused on the grounds that there was no indication that the prosecution intended to rely upon the provision so that it was inappropriate to refer it to the Court. That approach was subsequently approved by the Constitutional Court, which said that it illustrated how in practice deferring the determination of constitutional issues until they prove decisive promotes the interests of justice.
 Under the present Constitution similar preliminary litigation in a criminal case was considered by Langa ACJ and he said the courts:
‘… should discourage preliminary litigation that appears to have no purpose other than to circumvent the application of s 35(5). Allowing such litigation will often place the prosecutors between a rock and a hard place. They must, on the one hand, resist preliminary challenges to their investigations and to the institution of proceedings against accused persons; on the other hand, they are simultaneously obliged to ensure the prompt commencement of trials. Generally disallowing such litigation would ensure that the trial court decides the pertinent issues, which it is best placed to do, and would ensure that trials start sooner rather than later. There can be no absolute rule in this regard, however. The court’s doors should never be completely closed to litigants . . . If the trial is only likely to commence far in the future, the victim should be able to engage in preliminary litigation to enforce his or her fundamental rights. But in the ordinary course of events, and where the purpose of the litigation appears merely to be the avoidance of the application of s 35(5) or the delay of criminal proceedings, all courts should not entertain it. The trial court would then step in and consider together the pertinent interests of all concerned.’
 I am mindful of the fact that in Jordaan Cameron J, giving the judgment of the Constitutional Court, said that the initial approach of the Court that where possible it was desirable for cases to be disposed of without reaching the constitutional issue ‘has long since been abandoned in favour of its opposite, namely that constitutional approaches to rights determination must generally enjoy primacy.’ Like Savoi that was a case where the Court’s jurisdiction arose from confirmation proceedings where, apart from very unusual situations, the Court’s jurisdiction is always engaged. The persons raising the issue – the proper interpretation of s 118(3) of the Local Government: Municipal Systems Act 32 of 2000 – had a direct interest in the answer to the constitutional issue, which Cameron J described as ‘pressing’ and ripe for decision. In those circumstances the Court held that it was in the interests of justice that the issue be decided.
 None of those features characterise Ms Sonti’s case. If she is successful, as we hold she should be, she will return to the Magistrates’ Court to face the same charges under the same charge sheet as before, but the case will be further delayed until the Constitutional Court has decided whether to confirm the order for constitutional invalidity and, if so, on what terms. I am unable to see on what basis that can be said to be in the interests of justice. The justification given for following this route was that the Magistrates’ Court has no jurisdiction to strike down a statutory provision as unconstitutional. That is true, but it does have the obligation so far as possible, within the constraints discussed earlier in this judgment, to give the provision a constitutional interpretation. In any event, if it held that the threats allegedly made by Ms Sonti were not proven, an acquittal would have followed. Even with delays in the conduct of the proceedings it is hard to believe that the case would not have been resolved long ago. If, at the end of the day, a constitutional issue existed it could have been dealt with then.
 A curious feature of this case is that it reverses the usual role of a lawyer defending a client against a criminal charge. Usually the defence is conducted on the basis that facts are disputed and, if the ambit of the offence is relevant, defence lawyers argue for a narrow construction of the statute in the interests of their clients’ acquittal. It is the prosecution that advances a wider construction of the statute.
Here, by contrast, the legal representative for the two accused argued for the broadest conceivable interpretation of the section in pursuit of constitutional invalidity. That is an undesirable situation and it would not occur if the issues relevant to the constitutional challenge had been dealt with, as they ordinarily should have been, at the trial. If an issue of constitutional invalidity remained at the end of the trial that could have been resolved on appeal.
 Had Mr Moyo’s trial proceeded his defence would have been conducted by challenging the factual basis for the charge, in part by placing it in the appropriate context. Evidence would have been led to show that he was engaged in a legitimate act of political protest. Arguments could have been advanced that the section should be construed in the manner outlined in this judgment. The constitutional point would have been reserved for the appeal court assuming the case went that far. Assuming it was reached it would have been determined in the context of a live dispute and not on the basis of hypothetical examples. That is always preferable.
 Mr Moyo sought to bring his case within the principle stated in para 11 of Savoi that:
‘The applicants contend, amongst others, that the definitions of the very offences that they are charged with are so vague as to be unintelligible. Assuming for a moment that there is substance in that, it would be unfair to expect the applicants to plead to charges, the inner and outer contours of which they have no idea.’
But those words were written in relation to a challenge on the grounds of vagueness. That is not Mr Moyo’s challenge. He complains that s 1(1)(b) infringes his right to freedom of expression. That is not a complaint of vagueness, but of overreach. While accepting (at least in argument) that some of the conduct covered by the section is legitimately within its purview, his complaint, quoting from the heads of argument on his behalf was that it ‘obliterated the distinction between “true threats” and “political hyperbole”’. The quoted passage from Savoi provides no support for the course that this litigation has taken. It is also one that must be circumspectly applied. Otherwise it might be thought to justify a resort to the civil courts whenever there was doubt as to the parameters of an offence. Such questions are best dealt with by way of exception to the charge or in argument in the ordinary course of the criminal trial.
 The reason advanced by both Ms Sonti and Mr Moyo for not following the usual and ordinary course was the possibility of their being convicted and sentenced before the constitutional complaint could be adjudicated. This was advanced as the reason for raising the alleged constitutional invalidity before trial. I accept that if the trial had proceeded there might have been some risk of this occurring, but a magistrates’ court faced with a colourable constitutional challenge to the legislation under which it had just convicted someone, would grant leave to appeal and consider favourably an application for bail pending the appeal. That diminishes the risk considerably. In any event this argument creates a situation where the ability to have a constitutional challenge dealt with in advance of the criminal trial would depend upon whether the charge was laid in the magistrates’ court or the high court. If there was thought to be a real problem, the National Director of Public Prosecutions could have been approached to remove the trial to the high court, where the case would take its usual course, including disposing of the constitutional challenge. As a general rule departures from the procedures laid down in the CPA and the effective removal of criminal proceedings to the civil courts should not be countenanced.
 I do not understand Cameron J in Jordaan to say that magistrates’ courts should as a matter of routine postpone criminal proceedings in order to facilitate the bringing of constitutional challenges in the high court. Nor do I understand him to say that the high court should necessarily hear such cases. The delays that such litigation causes in the conduct of criminal trials and the manner in which it serves to defeat the speedy resolution of criminal cases contrary to s 35(3)(d) of the Constitution point in the opposite direction.
 The problem of delay caused by this type of procedure is illustrated by the preliminary litigation in Moodley, which had delayed the criminal trial for four years when it came to this court over a preliminary challenge to the charge sheet. According to the chronology in Mr Savoi’s bail appeal in this court the criminal investigations commenced in 2006 and the charges in two different courts were brought in 2011. The trial has not yet commenced. It is commonplace to see in the media that the first step in any criminal litigation involving a prominent person is that they will challenge the constitutionality of the charge, or the process leading up to the commencement of criminal proceedings. The term ‘Stalingrad defence’ has become a term of art in the armoury of criminal defence lawyers. By allowing criminal trials to be postponed pending approaches to the civil courts, justice is delayed and the speedy trials for which the Constitution provides do not take place. I need hardly add that this is of particular benefit to those who are well-resourced and able to secure the services of the best lawyers.
 All of this conveys to me that the wisdom of Langa ACJ remains pertinent. There are echoes of that in Madlanga J’s words in Savoi. The question in every case is one of the interests of justice. In my view the interests of justice in both of these cases demanded that the high court decline to hear them before the resolution of the criminal trials. Like my colleague I deprecate the fact that the trial judge failed to address the point. However, like him, given that the proceedings have reached this stage, I consider it in the interests of justice to deal with the appeal.
 I make the following order:
1 The appeal in Moyo and Another v Minister of Justice and Constitutional Development and Others is dismissed, with all parties to pay their own costs.
2 The appeal in Sonti and Another v Minister of Justice and Correctional Services and Others is upheld with costs, including the costs of two counsel.
3 The order of the court a quo is set aside and in its stead is substituted the following:
‘(i) It is declared that s 1(2) of the Intimidation Act 72 of 1982 is unconstitutional and invalid.
(ii) The order of invalidity is retrospective only to the extent that it affects pending trials or appeals and does not extend to any convictions where the right of appeal has been exhausted.
(iii) The matter is referred to the Constitutional Court in terms of s 172(2)(a) of the Constitution.
(iv) The Minister of Police is ordered to pay the costs of this application, including the costs of two counsel.’