Court name
Constitutional Court of South Africa
Case number
CCT 50 of 1995

S v Coetzee and Others (CCT 50 of 1995) [1997] ZACC 2 (06 March 1997);

Law report citations
1997 (4) BCLR 437
1997 (3) SA 527
Media neutral citation
[1997] ZACC 2

CONSTITUTIONAL COURT OF SOUTH AFRICA


Case CCT
50/95

THE STATE

versus

ABRAHAM LIEBRECHT COETZEE
HENDRIK
SCHALK COETZEE

PIETER LE ROUX DE BRUIN

JOHAN MARAIS


Heard on: 19 March 1996
Decided on: 6 March 1997


JUDGMENT






LANGA J:


[1] This matter is one of many which
have been dealt with by this Court, in which the constitutionality of provisions
of the Criminal
Procedure Act 51 of 1977 (the “Act”), have been
challenged. The Act plays a crucial role in the criminal justice system
of this
country; it is nonetheless legislation which was drafted and enacted in a
different constitutional era in which the legal
validity of its provisions could
not be questioned. The Constitution of the Republic of South Africa Act 200 of
1993 (the “Constitution”)
has brought about a drastic change, not
only in doing away with parliamentary sovereignty thus making all laws subject
to judicial
review, but also in the values which must now hold sway. The
problem is that important provisions of old legislation, and in particular
the
Act, are being struck down because they are inconsistent with the Constitution,
leaving gaps in the law which only the legislature
can fill. It is primarily
the task of the legislature, and not the courts, to bring old legislation into
line with the Constitution.
Although understandable because of the transitional
stage we are in, the continued operation of, and reliance by the prosecutors
on
provisions which do not reflect the new constitutional order is an
unsatisfactory state of affairs. Hopefully, it will not be
long before a
revised Criminal Procedure Act, consistent with the Constitution, is put in
place.

[2] The applicants are standing trial in the Witwatersrand Local
Division of the Supreme Court inter alia on twelve (12) counts of fraud.
At the conclusion of the prosecution’s case, Marais J acceded to the
applicants’ request
for the suspension of the trial and the referral to
this Court of the constitutionality of two provisions of the Act, namely,
sections
245 and 332(5).

[3] Although the propriety of the referral has
been challenged by the Attorney-General, it is clear from the judgment of Marais
J
that he has applied his mind to the issues relevant to the referral. It was
not in dispute that there were reasonable prospects
of the provisions being
found to be unconstitutional by this Court. Having decided, correctly in my
view, that the issues referred
could be decisive for the case and that it would
be in the interests of justice for him to refer the matter to this Court, I am
of
the view that the matter is properly before us. It will be convenient to
deal separately with the challenged provisions.

Section
245

[4] The section provides thus:

“If at criminal proceedings at which an accused is charged with an offence
of which a false representation is an element, it
is proved that the false
representation was made by the accused, he shall be deemed, unless the contrary
is proved, to have made
such representation knowing it to be
false.”



[5] The phrase “unless the contrary
is proved” means that the presumption may be rebutted by proof on a
balance of probabilities.[1] Absent
such proof, for example where the probabilities are evenly balanced at the end
of the trial, the court would be obliged to
convict, notwithstanding the
existence of a reasonable doubt regarding the state of mind of the
accused.

[6] The presumption falls into the class of “reverse
onus” provisions which have been held by this Court to infringe the
right
of an accused person to be presumed innocent as envisaged in section 25(3)(c) of
the Constitution.[2] The function and
effect of the presumption is to relieve the prosecution of the burden of proving
all the elements of the offence
with which the accused is charged.


[7] An essential element of crimes such as fraud and theft by false
pretences is knowledge of the falsity of the representation by
the person making
it. The effect of the provision is that once it has been proved that the
accused had made the false representation,
the presumption of knowledge comes
into operation and the onus of disproving it falls on the
accused.

[8] It is clear that the presumption is in conflict with the
long-established rule of the common law on the burden of proof that “it
is
always for the prosecution to prove the guilt of the accused person, and that
the proof must be proof beyond a reasonable
doubt.”[3] The provision
clearly infringes the presumption of innocence which is entrenched in section
25(3)(c) of the Constitution.

[9] The applicants contended, however, that
in addition to the presumption of innocence, the section also infringed what
was described
as “the cluster of rights associated with it,” namely,
the general right to a fair trial, the privilege against self-incrimination,
the
right not to be a compellable witness against oneself and the right to silence.
Because of the view I take that the presumption
infringes the right to be
presumed innocent that is protected by section 25(3)(c) of the Constitution, I
do not consider it necessary
to deal with the nature and scope of “the
cluster of rights” or how the impugned provision impinges on those rights.


[10] What remains to be determined is whether the infringement can be
said to be a permissible limitation to the right in terms of
section 33(1) of
the Constitution. In order to pass muster, a law which limits a right enshrined
in section 25 of the Constitution
must, in addition to being a law of general
application, be reasonable, justifiable in an open and democratic society based
on freedom
and equality, and
necessary.[4] The limitation must
also be such that it does not negate the essence of the right in
question.[5]

[11] It has been
held that this inquiry involves a weighing up of competing values and ultimately
an assessment based on proportionality.
The relevant considerations in this
balancing process include “the nature of the right that is limited, and
its importance
to an open and democratic society based on freedom and equality;
the purpose for which the right is limited, and the importance of
that purpose
to such a society; the extent of the limitation, its efficacy, and particularly
where the limitation has to be necessary,
whether the desired ends could
reasonably be achieved through other means less damaging to the right in
question.”[6]

[12] The
provision has its origins in section 280bis of the Criminal Procedure Act
56 of 1955 which was added in 1959.[7]
Section 245 in the current Act is substantially similar. Its purpose is to
facilitate the task of the state in the prosecution of
crimes such as fraud and
theft by false pretences by relieving the prosecution of the need to prove that
the accused knew that the
misrepresentation was false at the time that he or she
made it. The presumption has been held to be applicable to instances in which
the representation relates to facts which are objectively
ascertainable.[8]

[13] There is
no doubt a pressing social need for the effective prosecution of crime.
Kentridge AJ, speaking for the Court in
Zuma,[9] noted that reasonable
presumptions may be required by the prosecution, in relation to certain
categories of offences, to assist in
this task. It must be accepted that
section 245 was instituted by the legislature to facilitate the attainment of
its objective
to protect society. The measures taken were, however, enacted
before the Constitution was in force; they must now be weighed against
the
rights that are guaranteed by the Constitution which puts a high premium on the
values of freedom and equality.

[14] In a number of cases decided by
this Court, we have emphasised the importance of the rights entrenched in
section 25(3)(c) of
the Constitution, which include the right to be presumed
innocent, in an open and democratic society based on freedom and
equality.[10] Underlying the
decisions in those cases is the recognition that a consequence of the value
system introduced by the Constitution
is that the freedom of the individual may
not lightly be taken away. Presumptions which expose an accused person to the
real risk
of being convicted despite the existence of a reasonable doubt as to
his or her guilt are not consistent with what is clearly a fundamental
value in
our criminal justice system.

[15] The rationale for the provision is
that it deals with matters which are peculiarly within the knowledge of the
accused. Indeed,
the accused is in the best position to know why he or she made
a representation. It may well be that proving the state of mind of
the accused
in the context of a false representation presents the state with more
difficulties than in other cases. However, the
touchstone for justification,
where section 33(1) of the Constitution requires the prevailing state interest
to render a provision
not only reasonable but necessary as well, is not simply
the fact that an obligation to prove an element of an offence which falls
peculiarly within the knowledge of the accused makes it more difficult for the
prosecution to secure a conviction. The question
is whether it makes it so
difficult as to justify the infringement of the accused’s right to be
presumed innocent on the grounds
of necessity. I am not persuaded that this
difficulty is, in itself, sufficient to outweigh the importance of the right
infringed
and to justify the reversal of the onus. It is a difficulty, moreover,
which is not peculiar to offences in respect of which section
245 is applicable.
Discharging the burden of proof is a function which the criminal justice system
requires the prosecution to perform
in the normal course with regard to many
common law and statutory
offences.[11] It was not claimed
that if all the circumstances surrounding the false representation are fully and
properly investigated and presented
in evidence the prosecution cannot obtain
the conviction to which it might be entitled.

[16] It has not been
contended that other open and democratic societies based on freedom and equality
have found it necessary to resort
to such an unqualified presumption for the
proper enforcement of the criminal law in relation to all offences of which a
false representation
is an element. I am not aware of, nor have we been
referred to any examples in comparable jurisdictions, where a general provision
in the same context is employed.[12]
No good reason suggests itself why it should be necessary in this country to
have such a provision if, in general, crimes involving
misrepresentations are
adequately dealt with in other jurisdictions without the expedient of a reverse
onus provision.

[17] Section 245 makes severe inroads on the right of
those who fall within its ambit to be presumed innocent. No grounds were
advanced
in argument to justify the infringement and I have been unable to find
sufficient justification for this limitation to the constitutionally
protected
right. Because the provision fails to comply with the requirements of
reasonableness, justifiability and necessity as
required by section 33(1) of the
Constitution, it follows that it is unconstitutional by reason of its
inconsistency with the presumption
of innocence as enshrined in section 25(3)(c)
of the Constitution.

Section 332(5)
[18] Section 332(5) of the
Act has its origins in the Criminal Procedure Act 31 of 1917 where it was
inserted as section 348(5) of
that Act. The provision has since been part of
successive Criminal Procedure Acts in substantially the same form. Its wording
overlaps
considerably with that of the subsection dealing with the liability of
members of an association, other than a corporate body, in
the relevant Criminal
Procedure Acts.[13] It is part of a
comprehensive set of provisions contained in section 332, designed to facilitate
the criminal prosecution of corporations,
their directors and servants and
members of associations. Section 332(5) provides as follows:

“When an offence has been committed, whether by the performance of any act
or by the failure to perform any act, for which
any corporate body is or was
liable to prosecution, any person who was, at the time of the commission of the
offence, a director
or servant of the corporate body shall be deemed to be
guilty of the said offence, unless it is proved that he did not take part
in the
commission of the offence and that he could not have prevented it, and shall be
liable to prosecution therefor, either jointly
with the corporate body or apart
therefrom, and shall on conviction be personally liable to punishment
therefor.”



[19] The applicants attacked the
provision on the basis that it requires a director or servant of a corporate
body that has committed
an offence to prove, on a balance of probabilities, that
he or she did not take part in the commission of the offence and could not
have
prevented it. It was argued that the onus cast upon the accused relates to an
essential element of the offence created by the
section and that the reversal of
the onus meant that the accused could be convicted despite the existence of a
reasonable doubt with
regard to his or her guilt. This reverse onus was
therefore said to violate the right to be presumed innocent as enshrined in
section
25(3)(c) of the Constitution as well as the “cluster of rights
associated with
it.”[14]

[20] The
Attorney-General did not resist this line of attack but a different stance was
adopted by the Government of the Republic
of South Africa (the Government),
which had been granted leave to intervene as a party. In the event, this
intervention facilitated
a fuller ventilation of the issues, which turned out to
be more complex than had at first appeared to be the case.

[21] The
attitude of the Government was that the proper construction of the provision had
to take into account the effect of the decisions
in R v Van den Berg and
Another
[15] and S v
Klopper.
[16] It was claimed
that the import of the subsection, in the light of those decisions, was that in
respect of crimes of which intent
was an element, the prosecution carries the
burden of proving the elements of the offence created by section 332(5),
including the
fact that at the time when the offence was committed by the
corporate body, the accused had knowledge of it, or, if not, that he
or she
deliberately refrained from acquiring that knowledge. This left the accused
with the onus to disprove the presumption that
he or she had taken part in the
commission of the offence and that such accused could have prevented it. It was
argued that because
the prosecution had to prove that the accused had knowledge
of the commission of the offence, the effect of the violation on the
right to be
presumed innocent is not severe and the limitation of the right is in any event
justifiable.

[22] I should mention immediately that I do not agree with
the Government’s contention that the section bears a meaning which
places
the burden of proving the accused’s knowledge on the prosecution. That
view finds no support from the language used
in the subsection. On the
contrary, the plain meaning of the words is that once the prosecution proves
that an offence has been
committed by a corporate body of which the accused was
a director or servant at the time of commission, the latter can escape
conviction
only by proving that he or she did not take part in and could also
not have prevented the commission of the offence. This is made
plain by both
Schreiner JA[17] and Steyn
JA,[18] in separate judgments in
R v Limbada and Another.[19]
The passages we were referred to in Van den Berg’s and
Klopper’s cases do not support the Government’s contention
either. In the former, Greenberg JA
stated:[20]

“[I]t is twice conceded that the first appellant may have been unaware of
the act of the second appellant in causing the fire,
and if he was so unaware,
then he has proved, in terms of the sub-section, that he did not take part in
the commission of the offence
and could not have prevented it . .
.”.


The learned judge does not purport to deal with the
burden of proving knowledge on the part of the accused. On the contrary, the
statement is based on the premise that the accused had been proved to be unaware
of the offence. From that, so the judge held, it
followed that he did not take
part in its commission and could also not have prevented it. In
Klopper’s case,[21]
Kotze AJA stated the following:

“Na my mening behoort sub-art. (5), wat . . . ’n vorm van
strafpligtigheid oplΛ, op die mins verswarende wyse uitgelΛ
te word.
Ten einde ’n objektiewe vertolking te regverdig, behoort ’n
kwalifikasie, soos bv. “rederlikerwyse”
of “sonder
nalatigheid” voor die woorde “kon verhoed het nie” ingelees te
word. Sonder so ’n kwalifikasie
in te voeg - waarvoor ek in ‘n
strafbepaling, soos hierdie, geen regverdiging kan sien nie - is dit onmoontlik
om te beslis
dat die Wetgewer ‘n objektiewe uitleg wou voorskryf. Dit
geld veral in ’n geval soos die onderhawige waar die aanklag
poging tot
bedrog is - ’n misdryf wat op opset berus. ’n Bevestigende antwoord
op die voorbehoude regsvraag, sou inhou
dat strafaanspreeklikheid op grond van
culpa opgelΛ kan word t.o.v. ’n misdryf waarvan opset ’n
essensiΝle element is. Dit is moeilik om te
aanvaar.”


The issue there was the proper test to be
applied in determining whether or not it had been proved that an accused could
not have
prevented the commission of the offence, where such accused had not
taken part in its commission. It was held that the test was
subjective and that
where dolus was an element of the offence, mere negligence would not be
sufficient to warrant a conviction. Again the judgment was not concerned
with
the burden of proof with regard to the knowledge of the accused and did not
purport to deal with that question. None of the
other cases we were referred to
supports the Government’s contention in this
respect.[22] I am accordingly not
persuaded that the provision bears the meaning attributed to it by the
Government.

[23] During argument the question was raised with counsel
whether section 332(5) is reasonably capable of being interpreted as creating
statutory criminal liability, subject to a special defence and, if so, whether
it would still be inconsistent with the Constitution.
Since the point had not
been canvassed in the arguments filed, the parties were invited to submit
supplementary written argument,
which they subsequently did.

[24] The
applicants persisted in their initial
argument[23] but contended that even
if, on a proper construction of the provision, the onus related to an excuse,
exemption or exception, the presumption of innocence would still be
violated as long as the onus is concerned with an element that is essential to
the verdict.
It was contended in addition that inasmuch as the provision might
expose the accused director or servant to the risk of a fine or
imprisonment for
an offence not committed by such accused but by another person, it infringed two
constitutionally protected rights,
namely, that of freedom and security of the
person, which is protected by section 11(1) of the Constitution, and the right
to property
which is enshrined in section 28 of the Constitution.

[25] In
its supplementary argument, the Government came out in support of the view that
the effect of the provision is to create statutory
criminal liability and that
the onus on the accused does not relate to an essential element of the offence
but to an exemption, exception,
or excuse. It was contended that because the
accused is only called upon to prove a defence after the offence has been
established,
the presumption of innocence is not breached. In the alternative,
it was argued that any infringement there might be was a permissible
limitation
in terms of section 33(1) of the Constitution.

[26] Before considering
whether section 332(5) of the Act creates liability on the part of natural
persons for the offences committed
by such corporate bodies, I should mention
that no one was prepared to defend the extension of liability to servants, as
the provision
does. I agree that there is no justification for including the
category of “servants” in the provision and I shall proceed
on the
basis that the section refers only to directors. It will be convenient to deal,
in the first place, with the meaning and
purpose of section 332(5).


[27] In Limbada’s case, Steyn JA, delivering the judgment
of the court observed:

“The sub-section does not purport to create a new species of offence by
superimposing the elements mentioned in it upon those
of whatever offence is
alleged against a member of an association of persons other than a corporate
body. What it does is to deem
an accused, in the circumstances described
therein, to be guilty of an offence committed by another, if he does not prove
that he
had no part in that offence and could not have prevented it. In the
circumstances so described it casts an onus of proof upon the accused and
in effect directs the Court to find him guilty if he does not discharge that
onus. It is essentially, therefore, an evidential provision . . . and
does not bring into existence a distinct though mutable offence,
having as one
of its essentials the commission of some other
offence.”[24]


And
further: [25]

“. . . what the prosecution was going to rely upon was not only the
active participation of each in the conduct specified
but also his or her
liability under sec. 381(7) in relation to the conduct of the
other”.



[28] Schreiner JA in his concurring
judgment however took the view that the section establishes a separate statutory
offence.[26] This was based on the
fact that the provision requires that the accused, who did not take part in the
commission of the offence,
should be convicted if such accused fails to prove
that he or she could not have prevented it. The two views regarding the nature
of the offence were the subject of comment in Klopper’s case.
Kotze AJA, speaking for the court, was of the view that neither was inconsistent
with the notion that an accused,
who was not involved in the commission of the
offence, nevertheless incurs liability for the offence unless such accused
proves that
he or she could not have prevented
it.[27]

[29] The effect of
Schreiner JA’s construction is that an accused who has not taken part in
the commission of the offence has
to prove what amounts to an element of the
offence created by the subsection, namely, that he or she could not have
prevented the
commission of the
offence.[28] A reverse onus of this
type would of course be a clear breach of the presumption of innocence enshrined
in section 25(3)(c) of the
Constitution.

[30] Two questions of
constitutionality arise from the view, reflected in the majority judgment in
Limbada, that the subsection establishes liability of the director for
the conduct of another and that an onus of proof is cast upon the
accused, on pain of conviction if he or she fails to discharge it. The first
question is whether the onus
provision is a violation of the presumption of
innocence protected by section 25(3)(c) of the Constitution. The second is
whether
the form of liability imposed on the director is an infringement of the
right to freedom and security of the person, which is protected
by section 11(1)
of the Constitution, as well as the right to property which is enshrined in
section 28. I turn now to deal with
the first question.

[31] In a number
of cases in which the constitutionality of reverse onus provisions has been
considered, this Court has left open
the question of the effect which a
provision, which requires the accused to prove an exemption, exception or
defence, has on
the presumption of
innocence.[29] What was decided in
those cases was that presumptions which required an accused to disprove an
element of the offence violated the
right to be presumed innocent because they
exposed the accused to the risk of a conviction despite the existence of a
reasonable
doubt.

[32] Applicants and the Government relied extensively
in their respective arguments on decisions of Canadian
courts.[30] Through these,
applicants endeavoured to demonstrate that the presumption of innocence is
violated where the accused is required
to discharge an onus on a balance
of probabilities in order to avoid a conviction. What the Government set out to
show was that different considerations
apply where the accused is only required
to prove a defence or an exemption or excuse, after a complete case for
conviction has been
presented by the prosecution. Firstly, so it was argued,
the presumption of innocence is not implicated at all in such an instance.

Secondly, it was contended that the context of the legislation is important.
Where the legislation is regulatory, as distinct
from being purely criminal,
the strict standard employed in respect of the presumption of innocence which
is relevant to criminal
prohibitions is not applicable.

[33] It is
necessary to deal with some of the decisions which, I consider, might provide a
useful indication of the approach followed
by Canadian courts in dealing with
this aspect. The relevant provision in the Canadian Charter of Rights and
Freedoms
is section 11(d) which provides in part:

“11. Any person charged with an offence has the right . . .

(d) to be presumed innocent until proven guilty according to law in a fair and
public hearing by an independent and impartial
tribunal.”



[34] The Government relied on the
majority judgment in R v
Holmes
[31] for its submission
that there was no breach of the presumption of innocence where the onus relates
to a defence and not to an element
of the offence. The provision in issue in
that case made it an offence for anyone to be in possession, “without
lawful excuse,
the proof of which lies upon him” of “any instrument
suitable for the purpose of breaking into any place ... under circumstances
that
give rise to a reasonable inference that the instrument has been used or is or
was intended to be used for any such purpose.
. . .” The presumption was
therefore only triggered once the Crown had proved beyond a reasonable doubt
facts from which such
inferences could be drawn. McIntyre J, speaking for the
majority of the court (with Le Dain and La Forest JJ concurring), held that
the
words used in the provision, “while apt in certain circumstances to do
so,” could not be said in the particular context
of that provision to
amount to a reverse onus. In his view, the presumption of innocence was not
violated where the prosecution
is required to prove its case beyond a reasonable
doubt without the benefit of any presumption, before any need for defence
arises.
He ruled that the provision required the Crown to discharge the burden
of proof by putting before the court “evidence covering
every element of
the offence of such nature that, if believed by the trier of fact and not
answered, would warrant a
conviction.”[32] In his view,
if the accused is convicted in the face of such a defence, it would not be
because of any presumption of guilt but
because the excuse was rejected after
the commission of the offence had been
proved.[33]

[35] R v
Schwartz
,[34] which was also
cited in support of the Government’s approach, was concerned with a
provision which required an accused charged
with possession of a
“restricted weapon” to prove that he or she was the holder of a
registration certificate or permit
for such
weapon.[35] It was further provided
that a document purporting to be a registration certificate is evidence of the
statements contained therein.
By a majority of five to two, the court held
that, notwithstanding the fact that the accused had to bring himself or herself
within
the exemption and despite the words employed in the section, no reverse
onus was imposed on the accused and there was no danger that
he or she could be
convicted despite the existence of a reasonable doubt.

[36] We were
urged by the Government to follow the majority approach in Holmes and to
reject that adopted in the later case of R v
Whyte
.[36] The issue in the
latter case was a provision which required an accused, charged with having care
or control of a motor vehicle while
his or her ability to drive was impaired by
alcohol, to prove that he or she had not entered the vehicle for the purpose of
setting
it in motion, once it was proved that such accused had occupied the
driver’s seat of a
vehicle.[37] In its judgment
the court expanded the theme which had been articulated in R v
Oakes
[38] that a reverse onus
provision in relation to an essential element of the offence violates the
presumption of innocence because “it
would be possible for a conviction to
occur despite the existence of a reasonable doubt.” It accepted that what
the accused
was required to disprove was not an essential element of the offence
but that it was a fact “collateral to the substantive
offence.” It
was held, nevertheless, that the presumption of innocence had been violated even
though in this case such violation
was justifiable in terms of section 1 of the
Charter. Dickson CJC articulated the principle on which the finding of the
violation
of the right was based as follows:



“The short answer to this argument is that the distinction between
elements of the offence and other aspects of the charge
is irrelevant to the s.
11(d) inquiry. The real concern is not whether the accused must disprove an
element or prove an excuse,
but that an accused may be convicted while a
reasonable doubt exists. When that possibility exists, there is a breach of the
presumption
of innocence. The exact characterisation of a factor as an
essential element, a collateral factor, an excuse, or a defence should
not
affect the analysis of the presumption of innocence. It is the final effect of
a provision on the verdict that is decisive.
If an accused is required to prove
some fact on the balance of probabilities to avoid conviction, the provision
violates the presumption
of innocence because it permits a conviction in spite
of a reasonable doubt in the mind of the trier of fact as to the guilt of
the
accused.”[39]


[37] The approach in Whyte was followed in R v
Keegstra,
[40] where a section in
the Criminal Code provided that an accused, charged with the offence of
promoting hatred against an identifiable
group, shall not be convicted “if
he establishes that the statements were true”. The words “if
he establishes” were characterised as imposing a reverse onus to prove a
defence in the court a quo. In
reaching his conclusion that the presumption of
innocence was violated, Dickson CJC, writing for the majority of the court,
considered
various decisions on the presumption of innocence. Observing that
the judgment of the court in Holmes had caused some confusion, he
stated:

“ . . . since Whyte it is clear that the presumption of innocence
is infringed whenever the accused is liable to be convicted despite the
existence of
a reasonable doubt as to guilt in the mind of the trier of
fact.”[41]



This
approach has been followed in a number of other cases as
well.[42]

[38] I consider
that both Holmes and Schwartz are distinguishable from the present
case. In both, the majority of the court reached their respective decisions on
the basis that
the provisions they were dealing with did not impose a reverse
onus and that there was no danger that the accused could be convicted
despite
the existence of a reasonable doubt. Section 332(5) involves elements which
have to be proved by the accused and which form
the substance of the offence.
In the circumstances of this case, I am of the view that the approach in
Whyte is to be preferred in considering the effect of section 332(5) on
the presumption of innocence. The provision imposes an onus on
the accused to
prove an element which is relevant to the verdict. It should make no difference
in principle whether or not an offence
created by a statute is formulated in a
way which makes proof of certain facts an element of the offence or proof of the
same facts
an exemption to the offence. What matters in the end is the
substance of the offence. If a provision is part of the substance
of the
offence and the statute is formulated in a way which permits a conviction
despite the existence of a reasonable doubt in regard
to that substantial part,
the presumption of innocence is breached.

[39] The fact that section
332(5) requires that the accused director should, on pain of conviction, prove
that he or she did not take
part in the commission of the offence and could not
have prevented others from doing so, even if it is formulated as an exception,
has the same consequence as a reverse onus provision which relates to an
essential element of the offence. Such accused will be
convicted unless he or
she discharges the onus; this despite the existence of a reasonable doubt with
regard to such accused’s
participation in the offence and the ability to
have prevented it.

[40] In the final analysis, whether section 332(5)
creates a form of statutory liability, with a shift in onus in respect of a part
thereof or a new crime with a special defence, the proof of which rests on the
defence, the final effect is the same. The objection
which is fundamental to
the reversal of onus in this case is that the provision offends against the
principle of a fair trial which
requires that the prosecution establish its case
without assistance from the accused. In either event, the right of the accused
to be presumed innocent is breached.

[41] It was argued on behalf of the
Government that the context of the legislation is relevant to the effect which a
reverse onus
has on the presumption of innocence. What the submission amounted
to was that a provision which would offend against the presumption
of innocence
when applied in a truly criminal context would not necessarily do so in a
regulatory one. Section 332(5) was said to
be a regulatory provision which, on
that basis, did not offend against the presumption of innocence.

[42] The
distinction between the “truly criminal” and
“regulatory” offences has been discussed in various
judgments in a
number of jurisdictions. It is perhaps best articulated by Cory J in
Wholesale Travel Group
Inc.
[43] in the passage quoted
in paragraph 193 of the judgment of O’Regan J. In much the same vein and
with regard to the justification
for the difference in treatment between
different categories of offences, Jackson J delivering the judgment of the US
Supreme Court
in Morissette v United States referred to the hazards which
have become part of modern living as a result of industrialisation. According
to him, those dangers:

“ . . . have engendered increasingly numerous and detailed regulations
which heighten the duties of those in control of particular
industries, trades,
properties or activities that affect public health, safety or welfare. . . . The
accused, if he does not will
the violation, usually is in a position to prevent
it with no more care than society might reasonably expect and no more exertion
than it might reasonably extract from one who assumed his
responsibilities”.[44]


Indeed,
Canadian Courts seem to have made a number of important distinctions between
offences that are “ordinary” and
“regulatory”, between
“strict” and “absolute” liability and between offences
that provide for
imprisonment and those that do not. In the context of some of
these categories, what looks like reverse onus provisions have been
found not to
be objectionable. It is not necessary to deal with these distinctions in the
present context. Although some assistance
might be derived from the
categorisation in a proper case, I consider that it would not be safe to regard
it as anything more than
a broad guideline. What is clear is that section
332(5) has a very wide reach and is not limited to regulatory offences. It is
a
general provision of extremely broad application. It is applicable to any
director or servant of a corporate body that has committed
a crime; the crime in
question is any possible offence which might be committed by a corporate
body or any of its officials, be they directors or servants. It is applicable
equally to all types of offences, be they
serious or trivial, common law or
statutory, “purely criminal” or regulatory. It is applicable
whether the liability
is absolute or strict and whether it is based on intent or
negligence on the part of the perpetrator. Because of the virtually
uncircumscribed
ambit of the provision with regard to offences, the penalties
could also range from the trivial to the very serious and there is
nothing to
preclude the imposition of imprisonment. The section cannot therefore be said
to be regulatory.

[43] Further, I am by no means persuaded that the mere
categorisation of an offence as regulatory would necessarily have the effect
of
a lower standard of scrutiny as contended for by the Government. The
presumption of innocence is breached whenever the effect
of a reverse onus
provision is such that the accused could be convicted despite the existence of a
reasonable doubt as to guilt or
innocence. As pointed out by La Forest J in his
partially dissenting judgment in the case of Wholesale Travel Group Inc.,
“. . . what is ultimately important are not labels (though they are
undoubtedly useful), but the values at stake in the particular
context.”[45] Once such
breach has been established, the balancing process prescribed in section 33(1)
of the Constitution becomes decisive.
The real question in each instance will
then be whether the provision is reasonable, justifiable and necessary. It is
the substance
of the provision, not its form, that is decisive.

[44] In
the result, I consider that the first question posed in paragraph 30 must be
answered in the affirmative. The onus provision
in section 332(5) offends
against the right to be presumed innocent as contained in section 25(3)(c) of
the Constitution. The provision
therefore falls to be declared unconstitutional
unless it is saved by the provisions of section 33(1) of the
Constitution.

[45] It is to be noted that the true purpose of section
332(5) is not the creation of criminal liability without any fault on the
part
of the accused director. What is intended is the conviction of those directors
who either take part in the commission of the
offence or are in a position to
prevent it but fail to do so. Proof of fault is therefore essential to a
conviction under the section.
For the purposes of this judgment, it is
therefore unnecessary to consider whether the creation of absolute criminal
liability for
an offence would be constitutionally permissible. I accordingly
do not find it necessary to comment on the view expressed in paragraph
93 of the
judgment of Kentridge AJ, that if an offence of absolute liability had been
created, it would not in itself have given
rise to any question of the
unfairness of the trial in respect of such an offence.

[46] What the
legislator has in substance done is to place a positive duty on the director or
servant to disprove factors which are
central to the offence and made a
conviction the consequence of a failure to do so. The legislature is, in my
view, fully entitled
to place a positive duty on directors and to make the
omission to discharge that duty an offence. What is in issue here is how this
has been done. It is appropriate that the Court should have regard, not only to
the purpose and effect of the legislation, but also
to the means used to achieve
its objective. What causes the provision to fall foul of the presumption of
innocence here is the effect
of merely changing the form of the provision to
require the accused, rather than the prosecution, to prove elements which are
essential
to his or her guilt or innocence. There is manifest unfairness where
the legislature, having created an offence potentially entailing
very grave
penalties, goes on to subvert an important constitutionally protected right by
requiring crucial elements of the offence
to be proved or disproved by the
accused on pain of conviction should the onus not be discharged. As pointed out
in Morissette,[46] there
should be a limit to the power of the legislature “to facilitate
convictions by substituting presumptions for proof”.
In Patterson v
New York,
[47] the dissenting
judges (Powell J, Brennan J and Marshall J) objected cogently to the fact that
the test which was the basis of the
majority judgment:

“ . . . allows a legislature to shift, virtually at will, the burden of
persuasion with respect to any factor in a criminal
case so long as it is
careful not to mention the non-existence of that factor in the statutory
language that defines the crime.”



[47] It remains
to be determined whether section 332(5) is nevertheless justifiable in terms of
section 33(1) of the Constitution.
The question is whether this limitation on
the presumption of innocence is, in all the circumstances, reasonable,
justifiable and
necessary and not a negation of the essential content of any of
the affected rights.[48]


[48] The purpose of section 332(5) is to ensure that directors who could
have prevented the commission of crimes by the corporate
body concerned should
bear responsibility for such crimes. Directors, of course, occupy a special
position of responsibility, not
only in relation to the corporate body but also
with regard to the public in
general.[49] The state consequently
has an important interest in ensuring that the affairs of corporate bodies are
properly and honestly conducted.
The corporate body itself has to be protected
against the dishonesty and other criminal conduct of those in charge of its
affairs
or who are involved with them. It would not in itself be unreasonable
to provide special measures to enable the prosecution to
overcome the difficulty
of gathering evidence about corporate activities. This would be consistent with
the state’s duty to
protect society. The question in this case is whether
the state could adequately achieve these legitimate ends by means which would
not be inconsistent with the Constitution in general and section 25(3)(c) in
particular.

[49] The problem of proving elements of the offence is one
that is not peculiar to offences envisaged under section 332(5). It is
a
problem that is often encountered in the criminal justice system. Where,
however, special measures have to be provided to meet
specific difficulties
related to facilitating prosecutions, they must fit in with the requirements of
the Constitution. It is not
the function of this Court to prescribe to the
legislature how it should seek to achieve these ends. I can see no reason
however,
why the state could not, for example, impose appropriate statutory
duties on directors and other persons associated with the corporate
body aimed
at ensuring that its affairs are honestly conducted and that it is itself
protected against dishonest conduct. This could
be done in a variety of ways by
means of appropriate legislative provisions which might, for instance, impose
the duties of disclosure
and reporting on the corporate body, its directors,
servants and other persons involved with its affairs. There has been no
suggestion
that such measures, enforced through appropriate sanctions, could not
accomplish as effectively the ends sought to be achieved by
section 332(5) of
the Act. It has further not been contended that such objectives could not be
achieved without placing an onus
on the accused to prove any aspect of his or
her innocence in a criminal prosecution for a breach of such duty. I am
accordingly
not persuaded that the reverse onus provisions in section 332(5) are
necessary. It follows therefore that reliance on section 33(1)
of the
Constitution must fail.

[50] This conclusion with regard to the effect of
section 332(5) on the presumption of innocence makes it unnecessary for me to
deal
with the second question[50] as
to whether the form of liability imposed on the director by the impugned section
infringes the rights protected by sections 11(1)
and 28 of the Constitution.
Not much was said in argument with regard to section 28 and I propose to say
even less. With regard
to the effect of section 11(1), I have had respectful
regard to the views and extensive research contained in the respective judgments
of Kentridge AJ and O’Regan J on this difficult issue. I
however consider it unnecessary for me to canvass the issue in this
judgment.

The Order:
[51] I turn now to the appropriate order.
The issue of whether any part of section 332(5) can legitimately be severed in
order to
avoid striking down the whole provision was argued by the Government.
It was suggested that if the words “it is proved that”
were removed,
what remains would still give effect to the main objective of the statute. In
Coetzee v Government of the Republic of South Africa; Matiso and Others v The
Commanding Officer, Port Elizabeth Prison and
Others
,[51] Kriegler J, speaking
for this Court, articulated the test as follows: “ . . . first, is it
possible to sever the invalid provisions
and second, if so, is what remains
giving effect to the purpose of the legislative scheme?” Put differently,
the rule to be
applied is that “where it is possible to separate the good
from the bad . . . and the good is not dependent on the bad, then
that part of
the Statute which is good must be given effect to, provided that what remains
carries out the main object of the
Statute.”[52] It is indeed
true that if severance would achieve the effect of preserving the provision in a
form which is consistent with the
Constitution, that route must be followed.
The Government suggested that the excision of the words indicated would have the
effect
of removing the onus of proof from the accused and placing it on the
prosecution. In his judgment, Kentridge AJ proposes that the
effect of excising
the words “it is proved” would be to cast an evidentiary, rather
than a legal burden as is the case
at present, on the accused. He suggests that
the effect of this change would be to keep the provision within the bounds of
constitutionality.
O’ Regan J’s judgment goes somewhat further.
She suggests that the words “ it is proved that he did not take
part in
the commission of the offence and that” should be severed. The result,
according to her, would be a shift of onus
from the accused to the prosecution.
I agree with Kentridge AJ that the effect of the excision, as suggested by the
Government and
as proposed in the learned judgments of Kentridge AJ and
O’Regan J, would raise the issue of whether the interpretation of
the
provision would be affected by the provisions of section 90 of the Act. I
consider, with respect, that it would. The interpretation
of the provision
would have to be dictated by the language used. The effect of the
“unless” clause in the truncated
version would be to introduce an
exception and this would immediately place the provision within the purview of
section 90 of the
Act. I am of the view that it is not open to the Court to
assign an interpretation to the provision in order to make it constitutionally
acceptable, if that interpretation is not supported by the words used. Such an
exercise would introduce more uncertainty into the
interpretative task of the
courts. In this case, if the suggested words are excised, the ordinary meaning
of what remains would
still constitute a legal
burden.[53] It follows that once
the truncated version cannot escape the effect of section 90 of the Act, the
severance serves no useful purpose
and therefore cannot be resorted
to.

[52] The following order is accordingly made:

1. Sections 245 and 332(5) of the Criminal Procedure Act No 51 of 1977 are
inconsistent with the Republic of South Africa Constitution
Act 200 of 1993 and
are, with effect from the date of this judgment, invalid and of no force or
effect.

2. In terms of section 98(6) of the Constitution, this declaration of invalidity
shall invalidate any application of sections 245
and 332(5) of the Criminal
Procedure Act 51 of 1977 in any criminal trial in which the verdict of the trial
court was or will be
entered after the Constitution came into force, and in
which, as at the date of this judgment, either an appeal or review is pending
or
the time for noting such appeal has not yet expired.

3. The matter of the State versus ABRAHAM LIEBRECHT COETZEE, HENDRIK SCHALK
COETZEE, PIETER LE ROUX DE BRUIN and JOHAN MARAIS is
referred back to the
Witwatersrand Local Division of the Supreme Court to be dealt with in accordance
with this judgment.


Kriegler J concurs in the judgment of
Langa J.


CHASKALSON P:

[53] I concur in the judgments of
Mahomed DP and Langa J, and in the order proposed by Langa J. I also agree with
Ackermann J’s
analysis of the purpose of section 332(5) and with his
conclusion as to the consequences this has for the severance proposed by
O’Regan
J.


MAHOMED DP:

[54] I have had the benefit of
reading the judgments prepared by my colleagues in this matter. I agree with
the order proposed by
Langa J. However, in view of the different views and
nuances which appear from these judgments I consider it desirable to set out
very briefly my approach to some problems which have manifested themselves
during and after argument in this case.

[55] I have nothing to add to the
judgment of my colleague Langa J with regard to section 245 of the Criminal
Procedure Act. That
section is manifestly and demonstrably unconstitutional for
the reasons articulated in his judgment. I also have nothing to add
to the
unanimous view of my colleagues that the reference to “a servant”
contained in section 332(5) of the Act is indeed
unconstitutional.

[56] With regard to the remaining part of section
332(5) there are two separate questions which arise. Firstly, does the section
constitute an invasion of section 11 of the interim Constitution (‘the
Constitution’)? Secondly, does the section properly
interpreted
constitute a breach of section 25(3)(c) of the Constitution?

[57] In his
analysis of the subsection Kentridge AJ considers that section 332(5) could have
been enacted without creating the qualification
introduced into the subsection
beginning with the word “unless”. He believes that the subsection
can therefore not be
open to any constitutional attack based on section 25(3)
because the qualification effectively gives to an accused person an opportunity
to escape the consequences which would have ensued without the qualification.
On this approach section 25(3)(c) becomes irrelevant
to any enquiry into the
constitutionality of the impugned section. I am respectfully unable to agree
with this approach. If section
332(5) was enacted without the qualification it
would in my view not have survived constitutional attack. It would have been
vulnerable
to challenge under section 11 of the Constitution because it
constituted a fundamental breach of the right to freedom and security
of the
person. Every director of a corporate body which had committed an offence
would, himself or herself, irreversibly have been
deemed guilty of the same
offence, however remote be her or his connection with the offence, and however
difficult it was for such
a director to have knowledge of the commission of the
offence or to prevent it. Such a provision would so seriously have undermined
the freedom and security of every director as to offend the basic guarantee of
freedom secured by section 11 of the Constitution.
It would have operated in
an invasive and unacceptable manner to discourage, to deter and to impede people
in this country from
continuing to engage in valuable entrepreneurship through
the mechanisms of corporate bodies which are crucial to the effective direction
of modern industrial society.

[58] If section 332(5) is to be saved from
attack under section 11 it must therefore be because the qualification contained
in the
subsection, to which I have referred, prevents the consequences which
would otherwise have operated to invade the guarantees contained
in section
11.

[59] The qualification introduced by section 332(5) is therefore
fundamental to its structure and central to the proper identification
of the
mischief or the evil which is sought to be addressed by the legislation. The
mischief targeted is not persons being directors
of corporate bodies which have
committed offences: it is directors who take part in the commission of the
offence and directors
who knew of the offence and fail to prevent it when they
are able to do so. This is the real heart of the offence, the real target
of
the legislation, its raison d’Λtre.

[60] Section 332(5)
achieves this by requiring the accused to prove that he or she did not
participate in the offence and could not
have prevented it. The result is this:
if at the end of the case the court has a reasonable doubt as to whether or not
the accused
took part in the commission of the offence by the corporate body, or
a reasonable doubt as to whether or not the accused could have
prevented the
commission of that offence, the court would nevertheless be required to convict
such an accused. Prima facie this
seems to me to be a breach of the presumption
of innocence contained in section 25(3) of the Constitution.

[61] On my
analysis of section 332(5) there is therefore a clear breach of section 25(3)(c)
of the Constitution. The only way in
which its constitutionality could be
upheld would be if it could be justified under the limitations authorised by
section 33(1) of
the Constitution. In my view there is no justification for any
such limitation because of the wide ambit of the purported operation
of section
332(5). The offence by the corporate body for which the accused director can be
held liable is not limited at all. No
attempt is made to confine its operation
to a limited class in which there may be sound grounds of public policy for
directors of
corporate bodies to maintain a high degree of circumspection,
diligence and vigilance in order to protect members of the public against
offences committed by corporate bodies which can have prejudicial effects on the
health and wellbeing of the general community.
There is no attempt in section
332(5) to limit such offences to offences of a ‘regulatory’
character. All offences are
included whatever be their nature, whatever be
their purpose and however remote be their connection with the ordinary purposes
and
activities of the corporate body.

[62] These conclusions make it
unnecessary for me to decide whether or not the impugned subsection, on my
interpretation, can successfully
be challenged under section 11 of the
Constitution. I would prefer to leave that question open. The question whether
a statute can
legitimately provide for the criminal liability of an accused
person without requiring mens rea in the form of either dolus or culpa
and, if
so, the circumstances under which this might be permissible, raise different
questions of interpretation and policy which
might have to be considered in the
future in the appropriate case, namely, where a decision on these issues is
necessary and the
Court has heard full argument.

[63] Notwithstanding the
conclusion to which I have come on my analysis of section 25(3)(c), and the
limitations clause in section
33(1), I do appreciate the need for proper
legislation to protect large sections of society from the injurious consequences
of the
conduct of corporate bodies engaged in fields of activity crucially
impacting upon society and the need for effective deterrence
against such
activities, often conducted by directors operating under the protective shield
of the corporate body. Inter alia for
this reason, I have given consideration to
the suggestion by Kentridge AJ (in paragraphs 107-9 of his judgment) that the
impugned
section can be saved simply by deleting the words “it is proved
that” within the section. Kentridge AJ suggests that
the effect of such a
severance could be to put an evidential burden on the accused. O’Regan
J, (in paragraphs 202-3 of her
judgment) in turn suggests severing the words
“it is proved that he did not take part in the commission of the offence
and
that” which, so she holds, would result in the prosecution bearing the
onus of proving that an accused, who did not take part
in the commission of an
offence, could have prevented it.

[64] I cannot agree with either
suggestion and subscribe to the view espoused by Langa J in paragraph 51 of his
judgment. In arriving
at that conclusion I am persuaded by the reasoning of
both Langa J and Didcott J. Excision of neither set of words would materially
affect the linguistic meaning or the legal effect of the subsection. It does
not say: “unless it is proved by the accused”;
yet, it has always
been construed by our courts as meaning this and not as meaning “unless it
has been proved by the
prosecution.”[1] Saying that a
director is deemed to be guilty unless he did not or could
not have done something, does not result in the onus of proof - nor even
an evidentiary burden - being cast on the prosecution in respect
of
participation by a director in the commission of the corporation’s crime
or the director’s inability to have prevented
such commission. I am in
full agreement with Didcott J’s analysis of the effect of the use of the
conditional word “unless”
followed by the negative. The negative
formulation of an exclusion from liability which follows the deeming provision
contained
in the main clause is structurally an exception to the main clause,
and because of this the words “it is proved” have
been consistently
construed by the courts as meaning “it is proved by the
accused”.[2] Deletion of the
words “it is proved that” or of the words “it is proved that
he did not take part in the commission
of the offense and that” cannot
logically have the result of converting an exception into an essential element,
and of changing
the meaning of the subsection from “it is proved by the
accused” to “it is proved by the
prosecution.”

[65] There is moreover much to be said for the view
that the subsection, on either truncation proposed, would trigger the
application
of section 90 of the Act, which in itself casts a full onus on an
accused in respect of exceptions, exemptions and the like. Reading
the
subsection as it stands, or reading it subject to either proposed deletion,
results in an accused director having to disprove
the one, the other or both of
the factors introduced by “unless”, in order to avoid being struck
by the deeming provision.
In order to attain an acquittal the director bears
the onus. Failure to discharge that burden, notwithstanding the existence (and
persistence) of a reasonable doubt as to guilt may result in a conviction.
Therein lies the unconstitutionality.

Kriegler J concurs in the judgment
of Mahomed DP.

ACKERMANN J:
[66] Save for the ultimate positive
conclusion which O’Regan J reaches on the question of severance in respect
of section 332(5)
of the Criminal Procedure Act I fully agree with her judgment
in this matter. My disagreement with my learned colleague relates
solely to her
application of the second part of the severance test, namely, whether what is
left after the severance proposed by
her still gives effect to the main object
of the section and her affirmative conclusion in this regard.

[67] This
is not the sort of case, referred to by Kriegler J in Coetzee v Government of
the Republic of South Africa; Matiso and Others v Commanding Officer, Port
Elizabeth Prison, and Others
, 1995 (4) SA 631 (CC); 1995 (10) BCLR 1382 (CC)
at para 16, which might require special treatment on the issue of severability.
Here, as there, the trite test can properly be applied. What is left after the
severance is a provision which in substance imposes
criminal liability on the
director of a corporate body for an offence committed by that corporate body in
circumstances where the
director, although aware of the commission of the
offence and able to prevent its commission, desists from doing so. This, in my
view, is for all practical purposes the same offence as that embodied in section
332(5) before severance, on the construction given
in S v Klopper 1975
(4) SA 773 (A) at 780-1; the effect of severance merely being the removal of the
reverse onus provisions.

[68] In determining what the main object of
section 332(5) is, it is crucial to determine whether its provisions (ignoring
the evidential
effect of the reverse onus) constitute a new offence or a new
substantive basis for imposing criminal liability on directors, or
whether they
are substantially the same as the common law. In my view they are substantially
the same as, if not identical to,
the common law. If an employer, being able
to control a physical act of a servant of which he or she is aware and which
would constitute
a crime, forbears to prevent it, such forbearance constitutes
an implied authority to commit the act; the employer is guilty as a
socius
criminis and the element of mens rea is provided by the employer’s own
mental condition (R v Shikuri 1939 AD 225 at 230-1; R v Bennett and Co
(Pty) Ltd and Another
1941 TPD 194 at 199-200; R v Van der Merwe 1950
(4) SA 124 (0) at 128F-129A; S v Claasen 1979 (4) SA 460 (ZRA) at
463H).

[69] This principle must apply, a fortiori, to the case of a
director who, having a fiduciary duty to the company in question, is
aware of
acts being performed which would render the company criminally liable, is in a
position to prevent such acts but forbears
from doing so. By such forbearance
the director must likewise be taken to authorise impliedly the commission of the
acts and is
liable as socius criminis. In R v Blackmore and Another 1959
(4) SA 486 (FC) at 490H-491A the court, in an obiter dictum, was prepared to
extend the principle to company directors but found
it unnecessary to do so for
purposes of its decision. In my view there would be a duty on the director to
act to prevent the commission
of acts which would render the company liable to
criminal prosecution and his intentional failure to prevent the commission of
these
acts, if he were in a position to do so, would render him criminally
liable as a socius criminis (compare Burchell and Milton Principles of
Criminal Law
(Juta, Cape Town 1991) 84-7; Snyman Criminal Law 2ed
(Butterworths, Durban 1989) 50-2, 266-8; S v Timol and Another 1974 (3)
SA 233 (N) at 235G and S v Williams en ’n Ander 1980 (1) 60 (A) at
63D-E). The dearth of authority on the extension of this common law principle
to directors is no doubt due to
the fact that since the introduction into the
Criminal Procedure Act 31 of 1917 of the precursor to section 332(5) prosecutors
have,
because of the reverse onus provisions in the statutory provisions, not
found it necessary to rely on the common law to secure the
conviction of
directors.

[70] From the above conclusion it must follow that the main
(if not the exclusive) object of the section is limited to the reversal
of the
onus of proof in respect of material elements of an offence and principles of
criminal liability of directors which exist
at common law. It is therefore
impossible to conclude that a severance which does no more than to do away with
the reverse onus
provisions gives effect to the main objects of the section; in
fact it does the reverse.

[71] In my view it is therefore not possible to
sever the onus provisions from section 332(5) and I accordingly concur with the
order
proposed by Langa J.


DIDCOTT J:
[72] I concur in the grant
of the order proposed by Langa J and in the judgment which he has written in
support of that. I also
agree with the judgment prepared by Mahomed DP,
especially the fundamental part pertaining to the effects that sections 11(1)
and
25(3)(c) of the interim Constitution have on the present case and their
relationship there with each other, which I find wholly convincing.
Some
comments of my own will nevertheless be added to the criticisms levelled already
at section 332(5) of the Criminal Procedure
Act, the one that has become
controversial. They concern its overbreadth and the consequences of that, as I
view those topics.

[73] An obvious point at which section 332(5) goes too
far, and to which several earlier judgments written in this matter have drawn
attention, is found where it includes within its ambit every
“servant” of the corporate body in question, a description
not
restricted to an employee with managerial functions or responsibilities but one
so wide that it embraces the humblest and most
menial worker. The word lends
itself readily and on all counts, however, to severance from the rest of the
subsection. So, by simply
striking it out, we could have repaired the
subsection if its presence there had been the only fly in the ointment. But
that is
not the case.

[74] The allusion to an “offence” also
contributes to the width of section 332(5). The offences which it encompasses
are not confined to those committed either peculiarly or mainly by corporate
bodies, to the sort that have been created for instance
in order to control or
regulate their affairs or the activities familiarly undertaken by them. All
other crimes are covered as well,
crimes which anybody else may happen to
perpetrate in contravention of the common law or some statutory decree. The
extra coverage
is both superfluous and foreign to the store which Kentridge AJ
has set in his judgment by the frequent need for such control and
regulation.
We cannot constrict the subsection there by using the tool of severance or
resorting to a limited interpretation. A
remedial qualification would have to
be introduced, one tantamount to an amendment of the wording which lay beyond
our competence.

[75] I come next to the respect in which the overbreadth
of the section 332(5) troubles me most. Here too no narrower meaning can
be
achieved by either severance or a restrictive reading. Here too nothing but an
amendment can accomplish that. I refer to the
mention made of “any
corporate body”. No distinction is drawn between companies incorporated
with limited liability
and other corporate bodies. Nor, within the field of
companies, does the subsection differentiate between public and private ones,
between companies which solicit and receive money from investors and those that
never do, or between companies engaged in trading,
manufacturing or any other
business with the public and the types that are not. All are treated alike, and
in the same deep breath.
The impact of the subsection is then spread further by
section 332(10) of the statute, which declares that:

“In this section the word ‘director’ in relation to a
corporate body means any person who controls or governs that
corporate body or
who is a member of a body or group of persons which controls or governs that
corporate body or, where there is
no such body or group, who is a member of that
corporate body.”


I agree with Kentridge AJ that it is
quite fanciful to imagine the vice-chancellor of a university or the chairman
of its council
being prosecuted for the negligent driving by one of its servants
of a vehicle belonging to it which was driven in the course of
his or her
employment there. But, between the extremes of that hypothetical situation and
those much more realistic where a high
level of personal responsibility can
rightly be expected, lies a large area in which the prosecution of a director as
defined of
“any corporate body” is less far-fetched and would often
be oppressive.

[76] One example will suffice, I hope, to illustrate that
point. It has to do with the private ownership of individual units in blocks
of
flats and complexes of cluster housing that stand on single plots of land. The
legal ownership of such units can be acquired
only through separate sectional
titles. Rights that are less than but comparable in their effects with those of
legal ownership,
bestowing some of its important advantages, can and have to be
obtained by means of shares held in the shareblock companies which
own the land
and buildings. The two alternative schemes have become so popular that they
amount nowadays to a substantial and significant
feature of our housing scene,
especially in the bigger and more affluent urban areas. Their popularity is
likely to grow in the
future. The result has been, and in all probability will
continue to be, the proliferation throughout the country of the bodies
corporate
which manage the properties owned by sectional titles and the shareblock
companies controlling the properties enjoyed through
them. In paragraph 98 of
his judgment Kentridge AJ has expressed the view that:

“Those who choose to carry on their activities through the medium of an
artificial legal persona must accept the burdens as well as the
privileges which go with their choice.”


The owners or
quasi-owners of the properties which I now have in mind make no such choice,
however, since none but Hobson’s
sort confronts them. For in no other way
can they gain titles to their dwellings. Nor do I see why that state of affairs
should
affect adversely those whom they appoint, by and large from their ranks,
as trustees of their bodies corporate or directors of their
shareblock
companies. Neither kind of structure conducts any business besides the running
of the property in the interests of its
residents. No funds are handled by
either but the ones which the holders of the units or shares are required to
contribute. No
dividends are paid in turn to them. Nor is any interest. Such
nett profit as may accrue from the administration of the property
is invariably
appropriated to future expenditure or to the reduction of future contributions.
Yet, despite the predominantly domestic
nature of their activities, bodies
corporate and shareblock companies may find themselves charged with misdeeds.
Their income tax
returns may not have been lodged timeously. A town planning
scheme may have been contravened. A malfunction in the lift at a block
of flats
may have caused its collapse and the death of a person using it for which the
body in question is said to be culpable.
Further examples are easily
conceivable. It seems hardly fair that part-time and unpaid trustees or
directors, as they certainly
are on the whole, should be exposed personally to
threats of criminal liability greater than the risks ordinarily run by
individuals.

[77] The overbreadth of section 332(5) in the second and
third respects which I have discussed, viewed alongside the analysis by Mahomed
DP of its other effects, satisfies me that it would have been incompatible with
section 11(1) without the qualification introduced
by the word
“unless”; that section 33(1) of the interim Constitution would not
have preserved it in that event from nullification
on the score of such
incompatibility; and that, once the qualification enters the reckoning, the
palpable inconsistency between it
and section 25(3)(c) which then arises is
likewise inexcusable under section 33(1) and accordingly invalidates
it.

[78] Kentridge AJ has suggested, however, that any inconsistency with
section 25(3)(c) could be remedied by the simple excision from
section 332(5) of
the words “it is proved that” which appear in the qualification
immediately after “unless”,
together with the next
“that”. The present reverse onus would then be converted, he
considers, into a mere evidential
burden resting on the defence which was
inoffensive to section 25(3)(c). I cannot unfortunately agree with him. The
deletion which
he recommends would not, in my opinion, produce the
transformation envisaged. It would make no difference at all that I can see
to
the meaning of section 332(5), to which the words in question contribute nothing
apparent to me. The qualification postulates
two material circumstances,
firstly the fact that the director who is prosecuted took no part in the
commission of the offence, and
secondly the fact that he or she could not have
prevented its commission. One side or the other has to prove either those
circumstances
or their converse. The prosecution would have been saddled with
the onus of proving the converse had the subsection decreed that
the guilt of
the director was deemed “if” he or she had participated in or could
have prevented the commission of the
offence. But the defence would still need
to prove both facts, to elevate them from postulated to actual ones, were guilt
to be
deemed in accordance with the severance “unless” the director
did not take part in and could not have prevented the commission
of the offence.
That strikes me as the obvious and inescapable effect of the attenuated wording,
and in particular of the conjunction
“unless”, followed by a notion
couched in negative terms, rather than the contrary conjunction
“if”, used
in relation to a thought expressed positively. It is a
meaning so clear, to my mind, that the subsection without the words scrapped
by
the severance would not even rank as a provision reasonably capable of bearing
an interpretation which substituted a mere evidential
burden for the prevailing
onus of proof. On that footing, if I am right there, section 35(2) of the
interim Constitution does not
enter the picture.

[79] The judgment
delivered in R v Shangase[1]
supports the construction which I have placed on the truncated version of
section 332(5) resulting from the deletion proposed. The
case had to do with a
statutory provision which forbade any “native” as defined to stay in
an urban area for longer than
a specified period “unless. . . permission
so to remain has been granted to him”, and declared a contravention of the
prohibition to be an offence. The Appellate Division
held[2] that the onus to prove the
grant of permission had thus been cast on the person charged with the offence.
And it did so, I underline,
notwithstanding the absence from the provision in
question of any such words preceding “permission” as “it is
proved
that”. The qualification examined then was classified in the
judgment, to be sure, as an exception or the like for the purposes
of the
statutory predecessor to section 90 of the Criminal Procedure Act. Perhaps the
one in issue now would be so rateable too
under section 90, in the event of and
after the surgery suggested for it. But that is neither here nor there. My
interpretation
does not depend on a recourse to section 90. It turns on the
particular wording of the residue surviving the surgery. I therefore
find it
unnecessary in this case to undertake a constitutional appraisal of section 90
itself.

[80] In the part of his judgment that deals with severance
Kentridge AJ has relied on some Canadian decisions and on one given in
an appeal
emanating from Hong Kong.
Those cases, especially the latter, may well be
distinguishable from the present matter on the grounds of material differences
between
the phraseology employed in the foreign legislation and the wording
which we must construe. If the decisions cited are in point
because that is not
so, however, I am unwilling to follow them.

[81] Severance is also
favoured by O’Regan J in the judgment which she has prepared. But she goes
distinctly further than Kentridge
AJ does, proposing the removal from section
332 (5) of the words “it is proved that he did not take part in the
commission
of the offence and that”. The adoption of her suggestion would
leave intact the phrase “unless... he could not have
prevented it”
and deem the director guilty otherwise. The effect of that exercise, she
maintains, would be no burden at all
resting on the defence, not even an
evidential one besides any imposed elsewhere by the need to answer a prima
facie
case, but a full onus transferred to the prosecution on the issue
whether the director could have prevented the offence from being
committed. She
obviously presupposes a factual situation where the director took no part in the
commission of the crime. For his
or her participation in that would have
amounted in any event to a personal guilt under the common law. No occasion
could then arise
for the same person’s guilt to be deemed on the score of
an ability to have prevented the commission of the offence. Nor would
it make
much sense to enquire whether its perpetration was preventable by a participant
in that very conduct. It follows from what
I have said already about the force
of “unless” combined with the negatively shaped phraseology which
comes next, however,
that I do not agree with O’Regan J either. I fail to
see how the abridged qualification for which she votes can rightly be
construed
by reading “unless” as “if”, by ignoring the
“not” in the clause that starts with the
conjunction, and
accordingly by treating the phrase as it would have run had its formulation been
“if he could have prevented
it”. That bold course is simply not
open to us, I consider, in the face of words so clear, so unambiguous and so
unequivocal
as those which were actually used. Such a departure from them
smacks more of redrafting the subsection than of interpreting
it.

[82] Nor, if the remnant left after her abridgement bore the meaning
ascribed to it by O’Regan J, would it appear to pass one
of the usual
tests set for severability, the test prescribed in Johannesburg City Council
v Chesterfield House (Pty)
Ltd
,[3] when Centlivres CJ
declared:

“The rule . . . is that where it is possible to separate the good from the
bad in a statute and the good is not dependent on
the bad, then that part of the
statute which is good must be given effect to, provided that what remains
carries out the main object
of the statute... In such a case it naturally
follows that it is impossible to presume that the legislature intended to pass
the
statute in what may prove to be a highly truncated form: this is a result of
applying the rule I have suggested and is in itself
not a
test.”


We used that test and found it to have been met
in Coetzee v Government of the Republic of South Africa; Matiso and Others v
Commanding Officer, Port Elizabeth Prison, and
Others.
[4] On that occasion
Kriegler J, who wrote the judgment endorsed by the majority of the Court, left
room for the future evolution of
some or other exception to the rule by adding
the rider that “severability in the context of constitutional law may
often require
special
treatment”.[5] No reason why
we should not apply the same test to section 332(5) was advanced in argument,
however, or occurs to me. What we have
to examine in then going ahead can
hardly be “the main object of the statute”, which is much too
general and broad for
our purposes once the Criminal Procedure Act happens to be
that statute. We must look instead at the particular object of subsection
(5),
viewed within the setting of the whole section and with the help of any light
which the rest of that may shed on it. Yet I
am far from clear in my own mind
how we are expected to identify the “main object” of a statutory
provision that apparently
has not just one object, but collateral objects of
equal importance which vary in their levels, directions and thrusts. Some may
consider that to be the case here. Ackermann J believes that the main object of
the subsection is reversing the onus of proof in
the prosecutions and on the
issues covered by it. O’Regan J accepts that as a major object, but
couples it with and attaches
an equivalent weight to another which she deduces,
the object of imposing a duty on directors to prevent their corporate bodies
from
committing offences whenever they can do so. I doubt that what are said to
be those objects amount in truth to such. The second
sounds like a consequence
attributed to the operation of the subsection rather than an object of the
subsection itself. The first,
it seems to me on the other hand, is not so much
an object as the way in which an actual and anterior one was intended to be
achieved.
Ends have been confused there, I venture to suggest in short, with
means. Nor, once the main object of a provision is defined as
simply the
enactment of that very provision, can any object ever remain after the provision
disappears for effect to be given to
it. The object which I therefore prefer
imputing to the subsection is this. It was designed to ensure or encourage the
disclosure
by persons prosecuted under it of information which had a bearing on
the charges, and to preclude them from hiding behind corporate
veils, when the
true circumstances pertaining to the internal workings of their companies or
other bodies were seldom known to outsiders
but usually ascertainable from them.
The construction placed by Kentridge AJ on the product of his severance
accomplishes that object.
But the one which O’Regan J puts on the outcome
of hers does not.

KENTRIDGE AJ
[83] I have had the advantage of
reading the judgment of Langa J in this case. I am in full agreement with his
reasons for holding
section 245 of the Criminal Procedure Act 51 of 1977
(“the Act”) to be unconstitutional, and with the order which he
proposes insofar as it applies to that section.

[84] The question of the
constitutionality of section 332(5) of the Act raises more complex issues.
Langa J has fully analysed the
elaborate written and oral arguments placed
before us both by the applicants and the South African Government and has given
a lucid
account of the complex Canadian case law - case law which counsel for
both parties cited generously and relied on heavily. I am
indebted to these
analyses, but my own approach to the issues raised differs from that of Langa
J.

[85] In the course of argument we were referred by counsel to the
judgments of the Appellate Division in R v Limbada and
Another,
[1] a case concerned
with a subsection in an earlier Criminal Procedure
Act[2] in substantially the same
terms as section 332(5). In those judgments a difference of view emerged
between Steyn JA, speaking for
the majority, and Schreiner JA, as to the nature
of the subsection. In brief, Steyn JA considered that the subsection was
“essentially
... an evidential provision”: it did not “bring
into existence a distinct ...
offence”.[3] Schreiner JA on
the other hand, took the view that the subsection did create a statutory offence
and was not merely evidential in
effect. This difference of opinion has,
however, little relevance to the present case. In S v
Klopper
[4] Kotzϑ AJA giving
the judgment of the Appellate Division, pointed out that the issue in R v
Limbada
[5] had been merely
whether the indictment against the accused had been properly drawn. He held
that the relevant subsection had the
effect of imposing vicarious criminal
liability on the directors or servants of a corporate body. Much earlier, in
R v Smith and Others,[6] De
Villiers J had regarded it as beyond question that the subsection imposed a
vicarious liability on the directors or servants of
a company. In De Wet and
Swanepoel, Strafreg
[7] the
learned author says that the subsection does not create a new type of offence,
but undoubtedly creates a new form of liability
for the offence of another.
This characterisation of the subsection must, with respect, be correct. That it
was the intention of
the legislature to create vicarious liability appears
beyond question from the language of the subsection. The “deeming”
provision does not create an evidential presumption but creates and defines the
new form of criminal liability. The applicants'
argument to the contrary is in
my view unsustainable.

[86] Liability for the crime of another is a form
of strict or absolute liability, i.e. a liability imposed on an accused without
personal fault on his part. As far as I am aware vicarious criminal liability
is unknown to the Roman-Dutch common
law.[8] But in modern statute law
it is not uncommon. In South Africa, as in other countries, the complexities
and the pervasiveness of
commercial and industrial endeavour, and the need to
control them in the public interest, have led in several instances to the
creation
of vicarious criminal
liabilities.[9] The main examples
which come to mind are to be found in statutes regulating the handling of
products which are potentially harmful
either
socially1[0] or
physically.1[1] In those instances
the statute permits the accused to escape an otherwise strict vicarious
liability by establishing a defence that
he or she had used due diligence to
prevent the commission of the crime. Although this seems to be a common
provision I would observe
that it is in each case a matter of legislative
policy. Parliament might have imposed an absolute vicarious liability for the
misdeed
of another without any defence or excuse being available to the accused.
In such a case the State would have to prove beyond reasonable
doubt all the
elements giving rise to the vicarious liability including the commission of the
original offence for which the accused
is to be held vicariously liable. In the
absence of any special defence there could be no question of any onus being
placed on the
accused: that would not arise. If such a statute were before this
Court for constitutional scrutiny there could therefore be no
question of any
infringement of section 25(3) of the Constitution. The accused would be
presumed innocent until the prosecution
had proved all the elements necessary to
give rise to the statutory criminal liability. The accused’s right of
silence would
remain. One could not point to any factor in the statute itself
impairing the fairness of the criminal trial for its contravention.
The
question of the legitimacy of such a statute would be a different matter. That
would fall to be tested not under section 25
of the Constitution but against
other provisions of the Constitution, such as section 11(1) which protects the
right to freedom and
security of the person.

[87] Section 332(5) of the
Act reads as follows:

“When an offence has been committed, whether by the performance of any act
or by the failure to perform any act, for which
any corporate body is or was
liable to prosecution, any person who was, at the time of the commission of the
offence, a director
or servant of the corporate body shall be deemed to be
guilty of the said offence, unless it is proved that he did not take part
in the
commission of the offence and that he could not have prevented it, and shall be
liable to prosecution therefor, either jointly
with the corporate body or apart
therefrom, and shall on conviction be personally liable to punishment
therefor.”


Accordingly, if section 332(5) had omitted
the words “unless it is proved that he did not take part in the commission
of the
offence and that he could not have prevented it”, the
applicants’ attack on the section, based as it was on the placing
of a
legal burden of proof on the accused, could not have been mounted. It could not
have been submitted that the accused was liable
to be convicted despite the
existence of a reasonable doubt as to his guilt. The prosecution would have had
to prove beyond a reasonable
doubt all the elements set out in the section. By
definition, no reasonable doubt as to the guilt of the accused could remain.
There could thus have been no basis for an attack on the constitutionality of
section 332(5) as being an impairment of the right
to a fair trial under section
25(3) of the Constitution. Indeed, I understood Mr Gilbert Marcus, who
presented the oral argument
for the applicants, to accept that that must be
so.

[88] The legislature did not in fact choose to create an absolute
vicarious liability. It chose to mitigate what would otherwise
have been the
harshness of the provision, by permitting an accused director or servant to
escape liability upon proof, on a balance
of probabilities, of the two exempting
factors which I have set out in the previous paragraph. As a matter of logic
and common sense
I find it difficult to accept that in thus rendering the impact
of the section less severe than it would otherwise have been, the
legislature
was thereby rendering a trial under the subsection less fair than it would
otherwise have been. With all respect to
the well-argued submissions of the
applicants’ counsel I would venture to say that this analysis of section
332(5) is a short
and complete answer to the attack based on section 25(3) of
the Constitution. Nonetheless, the applicants’ counsel have firmly
maintained that, whatever the position might have been in the absence of the
exempting provisions, the inclusion of those provisions
leads to the
infringement of section 25(3). They submit that its effect is to permit the
conviction of accused persons notwithstanding
the existence of a reasonable
doubt as to their guilt, a consequence which offends against the presumption of
innocence. They put
this argument on two alternative bases.

[89] Their
first submission is that the subsection created a new offence, or at least a new
criminal liability, and that an essential
element of that offence or liability
was that the accused must either have participated in the commission of the
offence or have
been able to prevent it. If, they say, that requirement were
“relegated to the status of a mere exception, exemption or excuse”,
then the provision would mean that it was a criminal offence to be a director or
servant of a corporation which had committed an
offence. That, they say, would
be “absurd”. I do not agree that that result must necessarily be
stigmatised as absurd,
given the policy of the legislature which seems to be,
broadly speaking, to ensure within limits that some natural person is liable
for
the criminal offence of a corporate body. But in any event, one must read the
subsection as it stands. It in fact expresses
non-participation in, and
inability to prevent, the corporation’s offence as matters of exception,
exemption or excuse. Reference
was made to cases on section 90 of the Act and
its predecessors, in particular to R v
Beebee
1[2] and R v Kula and
Others.
1[3] That section and
those cases deal with the essentials of charge sheets or indictments where the
offence charged is subject to a
statutory exception, exemption or excuse. Their
relevance to the present issues is in my view marginal. But whether one applies
the truncation test suggested by Watermeyer
CJ1[4] in the former case or the
broader method of construction preferred by Schreiner
JA1[5] in the latter, the language
of the subsection leaves me in no doubt that the “unless” clause
does not constitute an element
which the prosecution must negative, but in terms
creates an exemption or excuse which the accused may prove by way of defence.
No more need be charged than that the accused was a director or servant of a
corporate body which was liable to be prosecuted for
a specific offence. It is
then for the accused to bring himself or herself within the permitted defence.
I am accordingly not in
agreement with the applicants’ construction that
an essential element of the offence created by section 332(5) is that the
accused participated in its commission or could have prevented
it.1[6]

[90] The
applicants’ second and alternative submission requires more detailed
consideration. They contend, relying largely
on a line of cases in the Supreme
Court of Canada, that once a criminal statute contains a reverse onus provision
in the sense of
a provision requiring the accused to provide proof of some fact
in order to escape conviction, it is irrelevant whether that onus
relates to an
essential element of the offence or to a defence by way of excuse or exemption.
In either case the presumption of
innocence is destroyed and the fairness of
trial impaired. Section 11(d) of the Canadian Charter of Rights entrenches the
presumption
of innocence as an essential element in a criminal trial. The
Supreme Court of Canada has frequently had to consider whether reverse
onus
provisions violated that provision of the Charter. The essence of those
Canadian judgments on which the applicants rely is
perhaps to be found in the
following passage in the judgment of Dickson CJC in R v
Whyte
:1[7]

“The short answer to this argument [that the reverse onus provision did
not relate to an essential element of the offence]
is that the distinction
between elements of the offence and other aspects of the charge is irrelevant to
the s. 11(d) inquiry. The real concern is not whether the accused must
disprove an element or prove an excuse, but that an accused may be convicted
while a reasonable doubt exists. When that possibility exists, there is a
breach of the presumption of innocence.

The exact characterisation of a factor as an essential element, a collateral
factor, an excuse, or a defence should not affect the
analysis of the
presumption of innocence. It is the final effect of a provision on the verdict
that is decisive. If an accused
is required to prove some fact on the balance
of probabilities to avoid conviction, the provision violates the presumption of
innocence
because it permits a conviction in spite of a reasonable doubt in the
mind of the trier of fact as to the guilt of the
accused.”


That passage was quoted in full in the
judgment of this Court in S v Zuma and
Others.
1[8] Canadian courts
have followed and applied it in a number of other
cases.1[9] The applicants argue
forcefully that the principle stated by Dickson CJC applies to section 332(5).
They say, in paragraph 11 of
their supplementary written submissions:

“In the present case, the reverse onus provision introduces the
inevitability of conviction of the accused despite the existence
of reasonable
doubt whether he participated in the offence or could have prevented it. His
conviction despite this reasonable doubt,
violates the presumption of
innocence.”


[91] There seems at first sight to be much
force in this submission, particularly as in this field this Court has derived
much guidance
from the reasoning of the Canadian
courts.2[0] I should point out,
however, that in my judgment in S v Zuma, in which the other members of
this Court concurred, the constitutionality of reverse onus provisions in
exceptions, exemptions or
provisos to statutory offences, as referred to in
section 90 of the Act, was expressly left
open.2[1] Moreover, the above
quoted passage in the judgment of Dickson CJC, clear as it is, is not to be read
as if it were a praetor’s
formula, or a statute, to be applied to every
case that could be said to fall within its language. Judges of the Canadian
Supreme
Court have often pointed out that the protections to be found in their
Charter of Rights are to be interpreted and applied according
to the context in
which they may arise and not in the abstract. Thus, in Edmonton Journal v
Alberta (Attorney-General)
2[2]
Wilson J stressed the need for a contextual approach to Charter interpretation.
She said:

“One virtue of the contextual approach it seems to me, is that it
recognises that a particular right or freedom may have a
different value
depending on the context.”


In R v Wholesale Travel
Group Inc
.2[3] La Forest J said
that certain procedural protections may be constitutionally mandated in one
context and not in another.2[4] The
context in that particular case was a statute which made it an offence to
publish false or misleading advertising. The offence
was one of strict
liability subject to the defence of due diligence, with the legal burden of
proving due diligence being on the
accused. There were other aspects of the
statute which complicated the case, but one of the issues was whether that
reverse onus
provision was in conflict with section 11(d) of the Charter.
Iacobucci J, with the concurrence of at least a plurality of the Court
held that
it did.2[5] This was not a
unanimous view. Cory J with the concurrence of L’Heureux-Dubϑ J held
that it did not. The relevant context,
he said, was that the statute under
attack was designed for the regulation of industry and commerce in the public
interest, and was
a form of public welfare legislation. He said at 224:

“The reasons for ascribing a different content to the presumption of
innocence in the regulatory context are persuasive and
compelling. As with the
mens rea issue, if regulatory mechanisms are to operate effectively, the
Crown cannot be required to disprove due diligence beyond a reasonable
doubt.
Such a requirement would make it virtually impossible for the Crown to prove
regulatory offences and would effectively prevent
governments from seeking to
implement public policy through regulatory means.”

Later Cory J said:

“Criminal offences have always required proof of guilt beyond a reasonable
doubt; the accused cannot, therefore, be convicted
where there is a reasonable
doubt as to guilt. This is not so with regulatory offences, where a conviction
will lie if the accused
has failed to meet the standard of care required. . . .
If the false advertiser, the corporate polluter and the manufacturer of noxious
goods are to be effectively controlled, it is necessary to require them to show
on a balance of probabilities that they took reasonable
precautions to avoid the
harm which actually resulted. In the regulatory context, there is nothing
unfair about imposing that onus;
indeed it is essential for the protection of
our vulnerable society.”


I shall consider in due course
whether section 332(5) of the Act can be regarded as a “regulatory”
statute. I should
in any event point out (as did Langa J) that even outside the
regulatory context the Canadian cases do not speak with one voice.
In paragraph
34 of his judgment Langa J refers to R v
Holmes
.2[6] There a section of
a criminal statute provided that:

“Every one who, without lawful excuse, the proof of which lies upon him,
has in his possession an instrument suitable for the
purpose of breaking into
any place ... is guilty of a indictable
offence...”.


The Supreme Court held that the
presumption of innocence was not violated, because the prosecution was required
to prove its case
beyond a reasonable doubt without the benefit of any
presumption, before any need for defence arose. That, as I have pointed out
above, is the position with section 332(5). McIntyre J, at 706, giving the
majority judgment said:

“If he is convicted in the face of such a defence, it is not because he
has been presumed guilty or because the commission
of the crime has not been
shown, but because his excuse was rejected after proof of the commission of the
offence.”


I find this approach highly convincing and
very much in point. In section 332(5) of the Act the primary object of the
legislature
was to introduce a vicarious liability for corporate crimes. If an
accused is convicted under the section it will be because all
the elements
required by the subsection in order to give rise to that liability have been
proved beyond a reasonable doubt and because
the excuse provided for by the
subsection has not been established. That is not to be equated with a
conviction in the face of reasonable
doubt as to guilt. In such a case, as in
R v Holmes,2[7] the
prosecution must prove its case fully by factual evidence without the benefit of
any presumption. The applicants submitted that
the section created a
presumption that the accused took part in the commission of the company’s
offence and could have prevented
it. With all respect, no such presumption is
to be found in section 332(5). Nor does the prosecution require to invoke such
a presumption
in order to succeed.

[92] I have referred in some detail to
the Canadian authorities because, as I have said, they provided the main support
for the submission
of the applicants. I do not presume to state the law of
Canada. I merely point out that the Canadian authorities taken as a whole
do
not provide a sure and unequivocal foundation for the applicants’
submissions. Further, although they deal with offences
of strict liability
(which might in some cases lead to a vicarious criminal liability) I am not
aware that any of them deal directly
with a statute expressly imposing a
vicarious liability such as section 332(5). Their statutory context is very
different. I point
out further that the burden of proof imposed by section
332(5) upon the accused is substantially less than the burdens imposed upon
the
accused in such cases as R v
Whyte
2[8], R v
Keegstra
2[9] and R v
Downey
.3[0] Unlike the Canadian
statutes referred to in the above-mentioned cases, the subsection does not
require proof of due diligence on
an objective basis. At least in cases where
the company’s offence is one requiring a guilty intent, a director or
servant
will in practice escape liability on proof that he or she was genuinely
unaware of the commission of the
offence.3[1] In relation to section
25(3) of the Constitution I do not find it necessary to decide whether section
332(5) can be accurately characterised
as “regulatory”, although I
shall have to return to that question in a different context. It is sufficient
to say that
the object of the subsection is to control the activities of
corporate entities by imposing a responsibility on those who control
or conduct
their activities, and ensuring that they do not regard themselves as beyond the
reach of the criminal law if a crime is
committed in the course of corporate
activities. In that context, if guidance is to be found in the Canadian cases,
I consider that
the appropriate guides are Cory J and McIntyre J in the
judgments to which I have referred.

[93] In any event I consider that the
question of the constitutionality of the subsection is answered by the analysis
which I have
attempted in paragraphs 85 to 88 above. In brief, if an offence of
absolute liability had been created, it would not in itself have
given rise to
any question of the unfairness of the trial of such an offence. Where the
severity of such a provision has been mitigated
by allowing the accused to prove
a special defence it is in my view illogical if not perverse to say that this
destroys the fairness
of the trial. The constitutionality of section 332(5)
falls to be tested against other provisions of the Constitution, in particular
section 11(1). I venture to suggest that the underlying fallacy in the argument
of the applicants is that they have confused the
question of the fairness of
section 332(5) itself with the very different question of the fairness of a
prosecution under that provision.

[94] Section 11(1) of the Constitution
provides:

“Every person shall have the right to freedom and security of the person,
which shall include the right not to be detained
without
trial.”


The applicants submit that section 332(5) is an
infringement of that provision of the Constitution. They say, in paragraph 13
of
their supplementary written submissions:

“Such an offence which exposes the accused to a fine or imprisonment for
the conduct of others in which he did not participate
and which he could not
have prevented, would violate the rights to freedom and security of the person
in terms of section 11(1) and
property in terms of section 28 of the
Constitution.” (footnote omitted)


Although this
submission was presented virtually as an afterthought it is a serious contention
in relation to section 11(1) of the
Constitution.3[2] The imposition
of criminal liability in the absence of a criminal intention has for some
hundreds of years at least been regarded
as an abhorrent concept both in South
African law and in the Anglo-American common law.
Blackstone,3[3] in a much cited
passage, said:

“To constitute a crime against human laws, there must be first a vicious
will; and secondly an unlawful act consequent upon
such vicious will.”


In Morissette v United
States
3[4] Jackson J
said:

“The contention that an injury can amount to a crime only when inflicted
by intention is no provincial or transient notion.
It is as universal and
persistent in mature systems of law as belief in freedom of the human will and a
consequent ability and duty
of the normal individual to choose between good and
evil.” (footnote omitted)

Similarly in Sweet v
Parsley
3[5] Lord Pearce
said:

“The notion that some guilty mind is a constituent part of crime and
punishment goes back far beyond our common law. And at
common law mens rea is a
necessary element in a crime.”


In S v
Qumbella
3[6] Holmes JA
said:

“[T]he basic principle is that actus non facit, reum nisi mens sit
rea
. Current judicial thinking is recognising more fully the scope and
operation of this fundamental rule of our
law...”.


Holmes JA, at 364F, went on to refer to this
rule as a “fundamental principle of fairness”. It is on the basis
of this
principle that statutes creating criminal offences will, as far as their
language permits, be interpreted as requiring the element
of mens rea in some
form, either subjective guilty intent or at least negligence. So too, unless
the language clearly warrants it,
a statute will not be interpreted as enacting
vicarious criminal liability.3[7]
Indeed vicarious liability may entail conviction not only in the absence of a
guilty mind but even when the accused has not personally
committed the criminal
act.

[95] In Canada section 7 of the Charter of Rights provides that
“[e]veryone has the right to life, liberty and security of the
person and
the right not to be deprived thereof except in accordance with the principles of
fundamental justice.” In a number
of Canadian cases the Supreme Court has
measured offences of strict liability against that section. The
applicants’ reliance
on these cases calls for some analysis of them. I
should point out however, that in none of these cases is the statute in question
in anything like the terms of section 332(5). In Reference re s.94(2) of the
Motor Vehicle Act
,3[8] the
Supreme Court of Canada, through Lamer J, held that absolute
liability3[9] and imprisonment could
not be combined: to do so would not be in accordance with the principles of
fundamental justice. The degree
of mens rea required to comply with section 7
is related to the nature of the crime. In R v Wholesale Travel Group
Inc
.4[0] the same judge (now
Lamer CJC) said:

“In Reference re: s. 94(2) of Motor Vehicle Act, supra, this Court
held that the combination of absolute liability and possible imprisonment
violates s.7 of the Charter and will rarely
be upheld under s.1. This is
because an absolute liability offence has the potential of convicting a person
who really has done
nothing wrong (i.e., has acted neither intentionally
nor negligently).

In R. v. Vaillancourt, supra, I stated that whenever the state resorts to
the restriction of liberty, such as imprisonment, to assist in the enforcement
of a
law, even a mere provincial regulatory offence, there is, as a principle of
fundamental justice, a minimum mental state (or fault
requirement) which is an
essential element of the offence. Reference re: s.94(2) of Motor Vehicle
Act
inferentially decided that even for a mere provincial regulatory offence
at least negligence is required, in that at least a defence of due
diligence must always be open to an accused who risks imprisonment upon
conviction. The rationale for elevating mens rea from a presumed element
... to a constitutionally required element, was that it is a principle of
fundamental justice that the penalty
imposed on an accused and the stigma which
attaches to that penalty and/or to the conviction itself, necessitate a level of
fault
which reflects the particular nature of the
crime.”


What this indicates, applied to our own
Constitution, is that while it would in general be an infringement of section
11(1) to subject
a person to the risk of imprisonment on the basis of an
absolute liability without at least a defence of due diligence, nonetheless
the
constitutional standard may allow some degree of strict liability. In the
pre-Charter case of R v City of Sault Ste.
Marie
4[1] Dickson J identified
three main categories of criminal offence. The first is the category of
offences in which mens rea consisting
of some positive state of mind must be
proved by the prosecution. The third category comprises offences of absolute
liability where
it is not open to the accused to exculpate himself even by
showing that he was free of fault. In between these is the second category
of:

“Offences in which there is no necessity for the prosecution to prove the
existence of mens rea; the doing of the prohibited
act prima facie
imports the offence, leaving it open to the accused to avoid liability by
proving that he took all reasonable care.”[at
181]


Section 332(5), like most cases of vicarious liability,
would fall into this intermediate category. Whether any particular provision
of
this sort would be an infringement of section 11(1) of the Constitution must
depend on the nature of the particular statutory
provision under consideration,
and the weight of the burden on the accused.

[96] I have stated above
in broad terms, the purpose of section 332(5). It must be said at once that the
subsection does not fall
within the category of regulatory offences as that term
has been used by the Canadian courts. Typical examples of crimes falling
within
that category are offences created by statutes designed to prevent pollution of
waterways, the sale of adulterated food or
the distribution of dangerous drugs.
Section 332(5) by contrast covers every type of criminal offence which a company
might commit
including crimes such as fraud, theft or culpable homicide. Some
convictions under section 332(5) would carry a serious moral stigma.
Nor is
there any limit on the punishment which can be imposed upon a conviction under
section 322(5). It is nonetheless in my view
useful to examine the rationale
behind the holdings of Canadian (and American) courts that it is
constitutionally legitimate to impose
criminal penalties on certain forms of
conduct in the absence of criminal intent or even negligence on the part of the
accused.
The rationale appears to be a combination of the public interest in
preventing antisocial conduct, the belief that criminal penalties
will induce
those in responsible or controlling positions to take all possible steps to
avert such conduct, and the difficulty of
achieving the object of the
legislation if the prosecution has the burden of proving intent or negligence.
Thus in United States v
Dotterweich
4[2] the United
States Supreme Court had to deal with a statutory offence of shipping
adulterated and misbranded drugs. In the judgment
of the
court4[3] the following passage is
to be found:

“The prosecution to which Dotterweich was subjected is based on a now
familiar type of legislation whereby penalties serve
as effective means of
regulation. Such legislation dispenses with the conventional requirements for
criminal conduct - awareness
of some wrongdoing. In the interest of the larger
good it puts the burden of acting at hazard upon a person otherwise innocent but
standing in responsible relation to a public danger.” (My
emphasis)


In Morissette v United
States
4[4] the Court referred
to “what have been aptly called ‘public welfare
offences’”. Of these the Court (through
Jackson J) said:

“Hence, legislation applicable to such offences, as a matter of policy,
does not specify intent as a necessary element. The
accused, if he does not
will the violation, usually is in a position to prevent it with no more care
than society might reasonably
expect and no more exertion than it might
reasonably exact from one who assumed his
responsibilities.”4[5]


Another
relevant passage is to be found in the judgment of Cory J in R v Wholesale
Travel Group Inc.
4[6] He
referred to what he called “the licensing concept” in the following
terms:

“The licensing concept rests on the view that those who choose to
participate in regulated activities have, in doing so, placed
themselves in a
responsible relationship to the public generally and must accept the
consequences of the responsibility. Therefore,
it is said, those who engage in
regulated activity should, as part of the burden of responsible conduct
attending participation in
the regulated field, be deemed to have accepted
certain terms and conditions applicable to those who act within the regulated
sphere.
Foremost amongst these implied terms is an undertaking that the conduct
of the regulated actor will comply with and maintain a certain
minimum standard
of care.

The licensing justification is based not only on the idea of a conscious choice
being made to enter a regulated field but also on
the concept of control. The
concept is that those persons who enter a regulated field are in the best
position to control the harm
which may result, and that they should, therefore,
be held responsible for it.”


In the same case at 221c-d
Cory J returned to this theme. He said:

“As a result of choosing to enter a field of activity known to be
regulated, the regulated actor is taken to be aware of and
to have accepted the
imposition of a certain objective standard of conduct as a pre-condition to
being allowed to engage in the regulated
activity. In these circumstances, it
misses the mark to speak in terms of the ‘unfairness’ of an
attenuated fault requirement
because the standard of reasonable care has been
accepted by the regulated actor upon entering the regulated
sphere.”


At 225f-g the same judge said:

“Quite simply, the enforcement of regulatory offences would be rendered
virtually impossible if the Crown were required to
prove negligence beyond a
reasonable doubt. The means of proof of reasonable care will be peculiarly
within the knowledge and ability
of the regulated accused. Only the accused
will be in a position to bring forward evidence relevant to the question of due
diligence.”


As I have already pointed out, R v City
of Sault Ste. Marie
4[7] was a
pre-Charter case. The explanation given by Dickson J for strict, including
vicarious, criminal liability is nonetheless relevant
and compelling. He
said:4[8]

“The element of control, particularly by those in charge of business
activities which may endanger the public, is vital to
promote the observance of
regulations designed to avoid that danger. This control may be exercised by
‘supervision or inspection,
by improvement of his business methods or by
exhorting those whom he may be expected to influence or control’: Lord
Evershed
in Lim Chin Aik v The Queen [1963] A.C. 160 at p. 174. The
purpose, Dean Roscoe Pound has said (Spirit of the Common Law(1906)), is
to ‘put pressure upon the thoughtless and inefficient to do their whole
duty in the interest of public health or
safety or
morale’.”


Later, at 181 he said:

“The correct approach, in my opinion, is to relieve the Crown of the
burden of proving mens rea, having regard to . . . the virtual
impossibility in most regulatory cases of proving wrongful intention. In a
normal case, the
accused alone will have knowledge of what he has done to avoid
the breach and it is not improper to expect him to come forward with
the
evidence of due diligence. This is particularly so when it is alleged, for
example, that pollution was caused by the activities
of a large and complex
corporation.”


[97] These considerations can in my
opinion be properly applied to a provision such as section 322(5), designed as
it is to induce
those who control corporate bodies to ensure that those bodies
keep within the law. A corporate body can act and thus commit criminal
offences
only through human agents, but the identity of those agents cannot always be
ascertained. Moreover the agent through whom
the criminal offence is committed
may hold a lowly position. In view of the dominant role played by corporate
bodies in modern society
it seems to me to be a legitimate objective of
government to ensure that the persons who control such bodies are not entirely
immune
from criminal liability for offences committed by servants of that body
in furtherance of its objectives. An absolute liability
for the crimes of the
corporate body would be so extreme as to be regarded by reasonable persons as
unfair or oppressive. But the
subsection is not absolute. It provides a
defence for the controllers of the corporate body which, as I have already
pointed out,
is considerably less burdensome than the requirement of proof of
due diligence referred to in the Canadian cases. I see nothing
unfair in
placing that limited burden upon the controllers of the corporate body. They
are the ones who may be expected to be aware
of the internal workings of the
corporation. They are the ones in the best position to give evidence of their
own lack of participation
and knowledge. The prosecutor does not know what goes
on in the boardroom; the director does. The provision ensures or attempts
to
ensure that a person in the position of director of a company will understand
that he has responsibility for its conduct. The
inducement to responsible
corporate conduct is enhanced by placing personal criminal liability on the
shoulders of those in control,
subject to a burden of proof not unduly difficult
for the innocent to discharge. The corporation itself can be punished only by
a
monetary penalty, a penalty which may not seriously affect those in control.
Further, what Cory J called the “licensing
concept” is peculiarly
appropriate to the conduct of corporate bodies and in particular limited
liability companies. Counsel
for the Government, Mr Jeremy Gauntlett SC, in his
further written submissions said this:

“The conduct of a director contemplated by section 332(5) is proscribed
because the inadequately controlled (and criminal)
activity of the company to
which it relates redounds to the detriment of society at large. . . . Those who
choose to assume a directorship
have, in doing so, placed themselves in a
position of responsibility not only vis a vis the company but in relation
to the public generally. . . . They must accept the consequences of that
position of responsibility.
This is because they are in the best position to
control the harm which may result from the activities of the company. More
specifically,
they must accept (indeed, for all practical purposes, they are
deemed to accept) that the law requires them to control the corporate
body, and
otherwise discharge their duties as directors, in accordance with certain
minimum standards on pain of civil and criminal
liability.”


I agree with that statement of the
rationale of section 332(5).


[98] The application of the
“licensing concept” to the control of companies is supported by the
judgment of this Court
in Bernstein and Others v Bester and Others
NNO
.4[9] That case concerned
a constitutional attack on the provisions of the Companies Act 61 of 1973 which
provided, under sanctions, for
the interrogation of directors and others having
information as to the affairs of a company. In paragraph 85 of his judgment
Ackermann
J said:

“The establishment of a company as a vehicle for conducting business on
the basis of limited liability is not a private matter.
It draws on a legal
framework endorsed by the community and operates through the mobilization of
funds belonging to members of that
community. Any person engaging in these
activities should expect that the benefits inherent in this creature of statute,
will have
concomitant
responsibilities.”5[0]


Those
who choose to carry on their activities through the medium of an artificial
legal persona must accept the burdens as well as
the privileges which go with
their choice. That was also the approach of the European Court of Human Rights
in Fayed v United
Kingdom
.5[1] In another
context, but still pertinent to the present issue, in S v De Jager
and Another5[2] Holmes JA
forcefully stated that those who choose to carry on business through a limited
liability company must accept the burdensome
as well as the beneficial
consequences of their choice.

[99] Section 332(5) has been part of our
statute law since 1939. It is not unreasonable to regard those who take
positions of control
in corporate bodies as voluntarily subjecting themselves to
the regime of company and corporation law, which must be taken to include
the
provisions of section 332(5). Of course, in a broad sense, the imposition of
criminal liability under the subsection, a liability
which could conceivably
lead to imprisonment, is an impairment of their freedom. But, as O’Regan
J said in Bernstein and Others v Bester and Others
NNO
,5[3] not all regulatory
laws or criminal prohibitions are subject to constitutional challenge in terms
of section 11(1):

“A purposive approach to section 11(1) recognises that it is aimed not at
rendering constitutionally suspect all criminal prohibitions
or governmental
regulation. Our society, as all others in the late twentieth century, clearly
requires government regulation in
many areas of social life. It requires a
criminal justice system based on the prohibition of criminal conduct. . . . Only
when it
can be shown that freedom has been limited in a manner hostile to the
values of our Constitution will a breach of s 11(1) be
established.”


Similarly in McKinney v University of
Guelph
5[4] Wilson J said that
it was:

“ . . . untenable to suggest that freedom is co-extensive with the absence
of government. Experience shows the contrary, that
freedom has often required
the intervention and protection of government against private
action.”


[100] In the light of these considerations I
do not believe that section 332(5) can be regarded as an impairment of the
freedom protected
by section 11(1) of the Constitution. That legislation in my
opinion constitutes a legitimate function of government in a society
in which
the ubiquity of corporate personality creates the danger that the corporate
structure will unduly screen and protect individuals
from the consequences of
corporate crime. Section 332(5) is a measure designed to protect the public
against that danger. It may
be that section 332(5) goes further in imposing
vicarious criminal liability than the statutes of other countries which we would
regard as free and democratic. I would not regard that as a cause for concern,
nor as a ground for constitutional attack.

[101] I have referred in the
previous paragraphs to the directors or controllers of corporate bodies.
Section 332(5), however, imposes
vicarious liability not only on the directors
of corporate bodies but on any servant. The government has conceded in its
submissions
that in this respect the subsection is too wide. That concession
was rightly made. To servants as a class one cannot impute the
choice of
entering a regulated activity: nor is the element of control present. In
Reynolds v G.H. Austin & Sons
Ld
.5[5] Devlin J said:

“. . . a man may be made responsible for the acts of his servants, or even
for defects in his business arrangements, because
it can fairly be said that by
such sanctions citizens are induced to keep themselves and their organisations
up to the mark. . .
. But if a man is punished because of an act done by
another, whom he cannot reasonably be expected to influence or control, the
law
is engaged, not in punishing thoughtlessness or inefficiency, and thereby
promoting the welfare of the community, but in pouncing
on the most convenient
victim.”


That is sufficient basis for holding that the
reference to servants of corporate bodies is unduly wide and does constitute a
substantial
impairment of freedom under section
11(1).5[6]

[102] The term
“director” is defined in section 332 of the Act. It is defined to
mean:

“any person who controls or governs that corporate body or who is a member
of a body or group of persons which controls or
governs that corporate body or,
where there is no such body or group, who is a member of that corporate
body.”


That is a wider definition than that contained
in the Companies Act. In R v Mall &
Others
5[7] Caney J held that
this covered de facto directors, and also persons who usurp the functions of a
director: it is not limited to those
who have legal and constitutional control.
This view was endorsed in S v
Marks
5[8] and in S v
Vandenberg
and
Others
.5[9] That
interpretation does not give rise to any difficulty. I see no reason why those
who are actually in control of a corporate
body should not accept the
responsibilities placed upon them inferentially by section 332(5). Caney J held
also that the definition
would include the directors of a holding company even
though they were not directors of the subsidiary which committed the criminal
offence. I do not think that Caney J intended by this to bring within the net
all directors of a holding company. Such persons
do not necessarily in law or
in fact control the subsidiary. If, however, it is capable of that meaning it
is also capable of a
narrower meaning. Section 35(2) of the Constitution would
then come into play. This provides:

“No law which limits any of the rights entrenched in this Chapter, shall
be constitutionally invalid solely by reason of the
fact that the wording
used prima facie exceeds the limits imposed in this Chapter, provided
such a law is reasonably capable of a more restricted interpretation which does
not exceed such limits, in which event such law shall be construed as having a
meaning in accordance with the said more restricted
interpretation.”



[103] A question which is possibly
more difficult arises from the use of the term “corporate body” in
the subsection.
This term is not limited to bodies incorporated under the
Companies Act. It covers for example non-trading corporate bodies which
might
include local authorities, universities and many other entities which are given
corporate personality by statute. To some
it may seem extravagant that the
provisions of section 332(5) should apply to corporations which do not engage in
commerce. For
my part, I see no good reason why the considerations which would
justify the subsection in relation to limited liability companies
should not
apply to every statutory corporation. It means that those who are prepared to
assume positions of control in a corporate
body must accept the same
responsibilities in relation to criminal conduct in the course of that
body’s activities as do the
directors of limited liability companies. The
hypothetical case has been put forward of a servant of a university who, while
driving
a university vehicle in the course of his employment, commits the crime
of negligent driving. The university will be criminally
liable under section
332(1) of the Act, and it would follow that the Vice-Chancellor of the
university and the Chairman of the Council
of the university would prima facie
be vicariously liable for the university’s crime and thus put upon their
defence. As a
matter of interpretation that would no doubt be so. I accept
that if a provision of a statute plainly infringes the Constitution
it should
not be upheld simply because it is unlikely to be invoked, or because a person
prosecuted under such a statute will readily
obtain an acquittal. I do not,
however, think it right to test the constitutionality of a criminal statute by
positing an unrealistic
example of a prosecution that would undoubtedly
constitute an abuse of the process of the Court. Any criminal offence, even one
with no legal burden of proof on the accused, might be the subject of a
vexatious prosecution. I do not think that respect for the
Constitution and for
a culture of individual human rights is furthered by striking down legislation
thought by Parliament to be necessary
for the public welfare, on the basis of a
far-fetched possibility that it will be abused by the prosecuting authorities.
I am reminded
of the words of Lord Woolf in relation to the Hong Kong Bill of
Rights. He said:

“The issues involving the Hong Kong Bill should be approached with realism
and good sense, and kept in proportion. If this
is not done the Bill will
become a source of injustice rather than justice and it will be debased in the
eyes of the public. In
order to maintain the balance between the individual and
the society as a whole, rigid and inflexible standards should not be imposed
on
the legislature’s attempts to resolve the difficult and intransigent
problems with which society is faced when seeking to
deal with serious crime.
It must be remembered that questions of policy remain primarily the
responsibility of
legislature.”6[0] (citations
omitted)



[104] I have not overlooked the fact that the
term “corporate body” is capable of including non-statutory
corporate bodies.
In South African law, unlike English law, no Act of
Parliament or of the executive is necessary to bring an artificial legal entity
into being. A number of persons may combine together to form an association
with a constitution which gives the association the
power to hold property
separately from its members, and the capacity to sue and be sued in its own
name. If such a constitution
is adopted, and if the objects of the association
are lawful, the association becomes a body corporate or
universitas.6[1] In this way
sporting bodies or charities or unregistered trade unions, to take some examples
at random, can be brought into being
and endowed with separate legal
personality. The application of section 332(5) to these bodies does not in my
view give rise to
any special problem. The principle behind the subsection is
that where an artificial legal person exists, the activities of which
may be
conducted in a criminal manner, some responsibility should rest on those who
control it. I do not see why that principle
should not apply to a sporting or
charitable body. The privilege of directing the activities of an artificial
legal persona carries
with it a certain burden of responsibility, including some
responsibility for its crimes. In the present case, in any event, the
applicants were prosecuted as directors of a company registered under the
Companies Act. If in some future case a question arises
as to the
constitutionality of section 332(5) in relation to a common law body corporate,
amongst the matters that would fall to
be considered is the possibility under
section 35(2) of the Constitution of “reading down” the words
“corporate
body” so as to limit them to statutory
corporations.

[105] It follows that in my opinion section 332(5) survives
the constitutional attacks made on it both under section 25 of the Constitution
and under section 11(1).6[2] In
view of this conclusion the question of justifying any infringement of either of
those sections in terms of section 33(1) of
the Constitution does not arise.
If, however, I am wrong on the issue of the infringement of those sections of
the Constitution,
I consider that section 332(5) would be justifiable under
section 33(1). For the reasons which I have given for holding that the
subsection did not violate the Constitution I would also conclude that the
provisions are in terms of section 33(1) reasonable and
justifiable in an open
and democratic society based on freedom and equality. I would also hold that
they are necessary, as any lesser
burden of proof such as an evidential burden
of proof would not achieve the legitimate aims of the legislation. It would be
only
too easy for an accused, for example by a bare denial, to raise some doubt
whether he knew of the corporation’s offence and
could have prevented it.
The burden of proof which would then revert to the prosecution would be in most
cases well-nigh impossible
to discharge. I would refer again to the passage in
the judgment of Dickson J in R v City of Sault Ste.
Marie6[3]
which I have quoted
in paragraph 95 above. In fact, in most of the Canadian cases relied on by the
applicants, in which a reverse
onus provision was held to be a violation of
section 11(d) of the Charter, the violation was held to be justifiable under
section
1.6[4] In S v
Bhulwana; S v Gwadiso
6[5] this
Court said, in relation to the application of section 33(1) of the
Constitution:

“In sum, therefore, the Court places the purpose, effects and importance
of the infringing legislation on one side of the scales
and the nature and
effect of the infringement caused by legislation on the other. The more
substantial the inroad into fundamental
rights, the more persuasive the grounds
of justification must be.”


On this approach the balance
comes down in favour of preserving the legislation. I reiterate that the burden
imposed on accused persons
under section 332(5) is substantially lighter than
the burden imposed by the Canadian legislation which was the subject of the
cases
referred to above.
[106] In reaching this conclusion I am comforted by
the fact that O’Regan J has also held (albeit on different grounds) that
section 332(5) is not in conflict with section 11 of the Constitution. It is a
serious matter to strike down legislation which in
one form or another has, for
the greater part of the century, imposed vicarious criminal liability on those
who control corporate
bodies. Section 332(5) in its present form was introduced
into the Criminal Procedure and Evidence Act 31 of 1917 as section 384(5)
by
section 117 of the Companies Amendment Act 23 of 1939. It replaced and
broadened section 384(1) of the 1917 Act, which was limited
in its scope to the
secretary, the directors, the manager and the chairman of a company. The burden
of proof on each of these persons
was to prove “that he was in no way a
party” to the offence. That subsection applied to all offences committed
by the
company.6[6] The 1917 Act
had repealed section 33 of the Companies Act of 1909, (Transvaal), which was
limited in its scope to any director,
manager, secretary or other officer of a
company. The burden of proof imposed on these persons was to “satisfy the
Court that
the default or offence was made or committed without his knowledge,
authority or permission”. (The section was, arguably at
least, confined
to offences under the Companies Act.) Although the scope of these provisions
has broadened since 1909 it has been
consistent legislative policy to impose a
strict vicarious criminal liability on those who control companies or other
corporate bodies,
with a provision for exemptions which imposes on them a
limited legal burden of proof. There is nothing before us to show that the
operation of the present subsection or its predecessors has in practice given
rise to injustices. Nor, I should add, have the provisions
anything to do with
the history of racial and other discrimination in this country. They were
provisions enacted for the protection
of the public in a society in which
corporate entities played an increasingly pervasive and important role. To
strike out section
332(5) would leave a considerable gap in the mechanisms
available for ensuring the honest conduct of corporate institutions. I am
satisfied that nothing in Chapter 3 of the Constitution requires us to take that
step.

[107] In this conclusion I unfortunately differ from the majority
of my colleagues. I would nonetheless respectfully suggest that
even on the
majority view it is unnecessary to strike down the whole of section 332(5). The
applicants’ complaint is based
on the consideration that the subsection
imposes a legal burden of proof and not merely an evidentiary burden. This
legal burden
of proof arises from the words “unless it is proved
that”. The question therefore arises whether in the absence of those
words i.e. if the clause read “unless he did not take part in the
commission of the offence and could not have prevented it,”
the burden as
a matter of interpretation would be merely evidentiary. If so, the accused
would then have only to raise a doubt as
to those matters and the main burden of
proof would remain with the prosecution. There is persuasive Commonwealth
authority for
this approach. In R v Wholesale Travel
Inc.6[7]
the legal burden of
proof on the accused arose from the words “unless he establishes
that”. The view of Lamer
CJC6[8] was that the words
“he establishes that” led to a violation of section 11(d) of the
Charter. Consequently he held that
those words were of no force or effect and
should be struck out, reducing the legal burden to an evidentiary one. His
judgment was
concurred in by other members of the Court. In a separate judgment
McLachlin J6[9] also held that the
words “he establishes that” should be deleted from the statute so as
to reduce the onus on the accused
to an evidentiary one. That was also the
remedy advocated by Dickson CJC in his minority judgment in R v
Holmes
.7[0] The same remedy
was adopted by the Ontario Court of Appeal in R v Ireco Canada II
Inc
.7[1] In Attorney
General of Hong Kong v Lee
Kwong-kut
7[2] the Privy
Council had before it a Hong Kong statute penalising the handling of the
proceeds of drug trafficking. The relevant section
of the statute provided that
it was a defence “to prove” certain matters. That clause was held
to place a legal burden
of proof on the accused. The Privy Council, at 973,
held that in the context of the war against drug trafficking, for a defendant
to
bear that legal onus was “manifestly reasonable”. However, if they
had not come to that conclusion they would have
regarded only the words
“to prove” as being inconsistent with the procedural protection of
the Hong Kong Bill of Rights.
Lord Woolf said:

“[I]f the words ‘to prove’ are removed from section 25(4) the
second defendants would be no longer under a legal
or persuasive burden of proof
to establish the defence contained in section 25(4). Instead they would be
under an evidential burden
merely requiring them to raise the issue. This
burden could not conceivably contravene article 11(1).”
7[3]



[108] In
the present case the severance of the words “it is proved” would
meet the tests for severability which have been
laid down by the Appellate
Division and adopted by this Court. The offending words would be grammatically
severable and the provision
which would remain would be manifestly in accordance
with the main objective of the legislation. If the provision, so truncated,
were to be considered standing alone it would impose only an evidentiary burden
or would be at the very least be capable of being
so construed. Accordingly, it
should be so construed in terms of section 35(2) of the Constitution. It is
necessary however to
consider whether such an interpretation would be affected
by the provisions of section 90 of the Act. That section which “is
unusual in that it specifically enacts a rule of statutory
construction”7[4] provides
in relation to any statutory offence that:

“. . . any exception, exemption, proviso, excuse or qualification, whether
it does or does not accompany in the same section
the description of the
offence in the law creating the offence, may be proved by the accused but need
not be specified or negatived
in the charge and, if so specified or negatived,
need not be proved by the prosecution.”


It has been
authoritatively held that this provision places a legal burden of proof upon the
accused7[5] and, indeed, its
wording hardly allows of any other interpretation. Consequently, if this
section were to apply to the truncated
section 332(5) the proposed severance
would be ineffectual; a legal burden would remain. This is a problem which was
not addressed
in the Canadian cases or in the Privy Council case to which I have
referred. I do not know whether there is any equivalent provision
in Canadian
or Hong Kong statute law, but observe that section 90 and its South African
predecessors are based on nineteenth century
English and Scottish
statutes.7[6] Moreover, those
provisions
apparently embody a longstanding common law rule of statutory
interpretation.7[7] I nonetheless
do not regard section 90 as being an obstacle to the course which I have
suggested. Hitherto, the pre-constitutional
interpretations of section 90 and
its predecessors have concentrated on the question whether any term of a statute
creating an offence
is indeed an exception, exemption etc. or whether it creates
an element in the offence which it is for the prosecution to
negative.7[8] The nature of the
inquiry in every case is, however, whether the statutory provision under
consideration is one to which section
90 applies. To assume that section 90
would apply to section 332(5) in its truncated form is to put the cart before
the horse.
If properly interpreted, that section, in its truncated form, places
only an evidential burden on the accused it must follow that
section 90 would
have no application. The “unless” clause would ex hypothesi not
constitute an exemption, exception
etc which the accused must prove in terms of
section 90. That section is no more than a general rule for the interpretation
of statutes
and cannot alter the meaning of an otherwise constitutional
provision so as to render it unconstitutional.

[109] This conclusion
perhaps makes it unnecessary to consider whether section 90 itself is valid in
the light of the constitutional
principle stated by Langa J and accepted by the
majority in this case. The principle as stated in the Canadian cases of R v
Whyte
7[9] and R v
Chaulk
8[0] and approved by
Langa J is succinctly stated in paragraph 38 of his judgment in which he
says:

“The provision imposes an onus on the accused to prove an element which is
relevant to the verdict. It should make no difference
in principle whether or
not an offence created by a statute is formulated in a way which makes proof of
certain facts an element
of the offence or proof of the same facts an exemption
to the offence. What matters in the end is the substance of the offence.
If a
provision is part of the substance of the offence and the statute is formulated
in a way which permits a conviction despite
the existence of reasonable doubt
in regard to that substantial part, the presumption of innocence is
breached.”


Those words apply with full force to section
90. Its effect upon every statutory offence to which it applies is to impose on
the
accused an onus to prove an element relevant to the verdict. Failure to do
so may lead to a conviction notwithstanding a reasonable
doubt as to the
existence or otherwise of the exempting or excusing factor. If section 332(5)
in its present form, as analysed by
the majority of this Court, constitutes an
infringement of section 25 of the Constitution then a fortiori section 90 must
by its
very nature also infringe it. For this reason and for the reasons which
I have stated in the previous paragraph I do not believe
that it is necessary
for this Court to strike down the whole of section 332(5). The remedy of
severability is open to it and should
in my respectful opinion be
adopted.

MADALA J:
[110] I have read the several judgments prepared by
my colleagues in this matter. Langa J sets out the facts and the issues which
arise as well as the positions taken by the parties in argument before this
Court. I agree with Langa J’s findings that the
provisions of section 245
of the Criminal Procedure Act 51 of 1977 are unconstitutional. With regard to
section 332(5) of the Act,
however, I am unable to agree with him or my other
colleagues and, for different reasons from those stated by Kentridge AJ, I do
not find section 332(5) to be unconstitutional.

[111] I also agree with
the findings of Langa J that the inclusion of the words “or servant”
in section 332(5) is constitutionally
not defensible, and that these words pass
the severability test postulated by my brother Kriegler J in Coetzee v
Government of the Republic of South Africa; Matiso and Others v Commanding
Officer, Port Elizabeth Prison, and
Others.
[1] Because of the
general view I take on this matter, it is unnecessary for me to consider the
other portions of this section which
we are asked to declare
severable.

[112] The impugned section has a long history which dates back
to earlier Criminal Codes in South Africa. It is the successor to
section
381(5) of the Criminal Procedure Act 56 of 1955. Section 381(5), in turn, came
after a provision inserted into the 1917
Criminal Procedure Act in 1939 as
section 348(5) of Act 31 of 1917. There is no material difference between these
provisions. Indeed
the impugned section is identical to section 381(5) of the
1955 Act in all material respects. There is also a substantial overlap
between
the wording of the 1977 provisions and the equivalent provisions in the 1917 and
1955 Acts dealing with the liability of
members of an association, other than a
corporate body, for offences committed by other members in furthering or in
endeavoring to
further its interests. Granted, these sections were enacted
before the advent of the interim Constitution, and must, therefore,
be
re-assessed in order to determine whether they accord with the
Constitution.

[113] Section 332(5) provides:

“When an offence has been committed, whether by the performance of any act
or by the failure to perform any act, for which
any corporate body is or was
liable to prosecution, any person who was, at the time of the commission of the
offence, a director
or servant of the corporate body shall be deemed to be
guilty of the said offence, unless it is proved that he did not take part
in the
commission of the offence and that he could not have prevented it, and shall be
liable to prosecution therefor, ... and shall
on conviction be personally liable
to punishment therefor.”



[114] It was argued on
behalf of the applicants that the section is particularly far-reaching in its
effect. The applicants further
alleged that the section casts a burden of proof
on the accused, and that it creates criminal liability where none existed
before.
Because of the conclusions to which I come, it is not necessary for me
to consider the latter submission. It was further argued
that since an accused
was required to prove some fact on a balance of probabilities to avoid
conviction, the provision violated the
presumption of innocence. Counsel for
the applicants further submitted that by reason of the reverse onus allegedly
contained in
the section, an accused person was liable to be convicted despite
the existence of a reasonable doubt as to his/her guilt and therefore
the
section was unconstitutional. I do not agree with this last submission. The
mere fact that a section provides that an accused
person may be convicted in
circumstances in which there is a reasonable doubt is not in itself a sufficient
reason for regarding
such section as unconstitutional. There may be
circumstances in which the reverse onus provision is necessary and
justifiable.

[115] Indeed reverse onus provisions have been under
scrutiny in many democracies. Kentridge AJ put it thus:

“The legitimacy of such provisions has been considered by courts as varied
as the United States Supreme Court, the Canadian
Supreme Court, the Privy
Council and the European Court of Human Rights (and doubtless others) in the
light of provisions entrenching,
in varying language, the presumption of
innocence, the right to silence and the privilege against self-incrimination - a
privilege
not expressly referred to in section 25. The case law of these Courts
- which are undoubtedly Courts of open and democratic societies
- indicates that
reverse onus provisions are by no means uncommon and are not necessarily
unconstitutional. Reverse onus provisions in our own statute law are
also not
uncommon.”[2]



[116] In
a number of decisions handed down by this Court on the issue of reverse onus
provisions it has been stated that there may
be circumstances which warrant the
casting of a legal burden on the accused, i.e., where it is perfectly
justifiable to call upon
the accused to
answer.[3] This, in my view,
depends largely on the nature of the offence and, perhaps more importantly, on
the purpose or objective of the
provision in question.

[117] It seems
that to resolve the constitutional issue before us we have to answer two
questions:

(a) Does the impugned section infringe the rights guaranteed by section 25(3)(c)
of the Constitution?

(b) If the section does infringe the rights and freedoms guaranteed by section
25(3)(c) of the Constitution, is it justified in terms
of section 33(1) and
therefore not inconsistent with the Constitution?


[118] I
need not deal with the arguments presented in respect of the first question, but
on a consideration of all the arguments placed
before us, it must be accepted
that the reverse onus provision infringes the provisions of section 25(3)(c) of
the Constitution.

[119] The next question for consideration, and this for
me is the crucial one, is whether section 332(5), although it violates the
provisions of the Constitution, is nevertheless justified in an open and
democratic society based on freedom and equality; or whether
the limitation on
the presumption of innocence is, in all the circumstances of this case,
reasonable, justifiable, necessary and
not a negation of the essential content
of the right. The state, seeking to uphold such limitation, would bear the
burden of its
justification.

[120] In S v Makwanyane and
Another
,[4] this Court said,
that the inquiry involves the weighing up of competing values and ultimately an
assessment based on proportionality.
Chaskalson P stated:

“The limitation of constitutional rights for a purpose that is reasonable
and necessary in a democratic society involves the
weighing up of competing
values, and ultimately an assessment based on proportionality. This is implicit
in the provisions of s
33(1). The fact that different rights have different
implications for democracy and, in the case of our Constitution, for ‘an
open and democratic society based on freedom and equality’, means that
there is no absolute standard which can be laid down
for determining
reasonableness and necessity. Principles can be established, but the
application of those principles to particular
circumstances can only be done on
a case- by-case basis. This is inherent in the requirement of proportionality,
which calls for
the balancing of different interests. In the balancing process
the relevant considerations will include the nature of the right
that is limited
and its importance to an open and democratic society based on freedom and
equality; the purpose for which the right
is limited and the importance of that
purpose to such a society; the extent of the limitation, its efficacy and,
particularly where
the limitation has to be necessary, whether the desired ends
could reasonably be achieved through other means less damaging to the
right in
question. In the process regard must be had to the provisions of s 33(1) and
the underlying values of the Constitution,
bearing in mind that, as a Canadian
Judge has said, ‘the role of the Court is not to second-guess the wisdom
of policy choices
made by legislators’.”(footnotes
omitted)



[121] The requirement that the accused should
prove his or her innocence appears to run counter to the presumption of
innocence which
is foundational in our principles of the criminal law, and of
which Dickson CJC, quoting from R v
Oakes,
[5] stated in R v
Holmes
:[6]

“The presumption of innocence is a hallowed principle lying at the very
heart of criminal law. Although protected expressly
in s. 11(d) of the
Charter, the presumption of innocence is referable and integral to the
general protection of life, liberty and security of the person contained
in s. 7
of the Charter (see Reference re s. 94(2) of the Motor Vehicle
Act
, December 17, 1985, unreported, per Lamer J. [now reported [1985]
2 S.C.R 486, 18 C.R.R. 30, 23 C.C.C. (3d) 289] ). The presumption of innocence
protects the fundamental
liberty and human dignity of any and every person
accused by the state of criminal conduct. An individual charged with a criminal
offence faces grave social and personal consequences, including potential loss
of physical liberty, subjection to social stigma and
ostracism from the
community, as well as other social, psychological and economic harms. In light
of the gravity of these consequences,
the presumption of innocence is crucial.
It ensures that until the State proves an accused's guilt beyond all reasonable
doubt,
he or she is innocent. This is essential in a society committed to
fairness and social justice. The presumption of innocence confirms
our faith in
humankind; it reflects our belief that individuals are decent and law-abiding
members of the community until proven
otherwise.”



[122] I have no doubt in my mind that
the presumption of innocence is a fundamental right which plays a pivotal role
in our criminal
justice system. However, in my view, like all other rights and
freedoms guaranteed by the Constitution, this right is not absolute,
but that
its value and weight will differ according to a variety of factors and
circumstances against which it is pitted on the scales.

[123] In a case
such as the present one, the Court should determine whether the objective of the
impugned provision is of sufficient
importance that it warrants overriding a
constitutionally protected right or freedom. As Lord Woolf, quoting Lamer CJC
in R v Chaulk 62 CCC (3d) 193 at 216,
noted:[7]

“The objective . . . must relate to concerns which are pressing and
substantial in a free and democratic society before it
can be characterized as
sufficiently important.”



[124] The determination of
the objective should not be divorced from a consideration of the particular
context within which the provision
operates. Ackermann J, made a similar
observation in the case of Ferreira v Levin NO where he stated:

“. . . the problem cannot be resolved in the abstract but must be
confronted in the context of South African conditions and
resources - political,
social, economic and
human.”[8] (footnote
omitted)



[125] We are living in times of intense
corporate activity which unfortunately is often accompanied by a proliferation
of serious
crimes like fraud by both corporations and directors. It is, in my
view, of paramount importance that the social interest of seeing
to the
prosecution by the state of crimes such as fraud which are perpetrated by
corporate bodies and directors be facilitated.
As Ackermann J put it, the
directors and other officers of the company “are the brains, eyes and ears
of the company”,[9] and are
often the only persons who have knowledge of the workings of the company.


[126] In any democratic criminal justice system there is a tension
between, on the one hand, the public interest in bringing law breakers
to book,
and, on the other hand, the equally great public interest of ensuring that
justice is manifestly done to all. In my view,
it is of paramount importance
that the affairs of companies and other corporate bodies and associations be
conducted properly and
honestly. The section under attack assists in the
achievement of that objective. Windeyer J in Rees v
Kratzmann
1[0]
observes:1[1]

“The honest conduct of the affairs of companies is a matter of great
public concern to-day.”



[127] In support of this
view, Ackermann J, in his judgment in Bernstein and Others v Bester and
Others NNO
1[2] adds:



“This is particularly the case in South Africa at present. Such honest
conduct cannot be ensured unless dishonest conduct,
when it occurs, is exposed
and punished and ill-gotten gains restored to the
company.”



[128] I am also in agreement with Lord
Denning where he observes in the case of H. L. Bolton (Engineering) Co Ltd v
TJ Graham and Sons Ltd
1[3]
that:

“Others are directors and managers who represent the directing mind and
will of the company, and control what it does. The
state of mind of these
managers is the state of mind of the company and is treated by the law as
such.”



[129] This sentiment is shared by my
colleague Goldstone JA (as he then was) in the case of Howard v Herrigel and
Another NNO
,1[4] where he
states:

“At common law, once a person accepts an appointment as a director, he
becomes a fiduciary in relation to the company and is
obliged to display the
utmost good faith towards the company and in his dealings on its behalf. . . .
Whether the inquiry be one
in relation to negligence, reckless conduct or fraud,
the legal rules are the same for all
directors.”



[130] It is quite clear then, that in
assuming the office of director the incumbent consents to assuming the
responsibility that is
concomitant with such office. The steady growth of the
corporate world is burdened with ever increasing sharp business practices.
When
other factors such as the difficulty of obtaining evidence for an offence where
the information is usually within the exclusive
knowledge of a few, and the high
cost factor involved in implementing regulatory enforcement mechanisms are
considered, the impugned
provision becomes practically and objectively more
justifiable.

[131] In my view the objective of this impugned provision is
of such importance that it warrants the overriding of the guarantees
provided in
section 25(3)(c)of the Constitution.

[132] In the result I hold that the
impugned section is justified in terms of section 33(1) and is therefore not
unconstitutional.


MOKGORO J:
[133] I concur in the judgment of
O’Regan J that sections 245 and 332(5) of the Act are unconstitutional
because their reverse
onus provisions are unjustifiable breaches of the
presumption of innocence, guaranteed by section 25(3)(c) of the Constitution.


[134] As the precise analysis of the constitutionality of section 332(5)
has led to differing positions amongst my colleagues, I will
state very briefly
the reasons for my conclusions regarding that section.

[135] Section
332(5) provides in relevant part that where an offence has been committed by a
corporate body

“a director . . . of the corporate body shall be deemed to be guilty of
the said offence, unless it is proved that he did not
take part in the
commission of the offence and that he could not have prevented it . .
.”.[1]


It was argued by the applicants that the section breaches
the presumption of innocence because it places the onus on the accused to
establish that he or she did not take part in the commission of the offence, and
that he or she could not have prevented such commission.

[136] The
Government responded that section 332(5) does not breach the presumption of
innocence. First, the Government argued that
because the accused is required to
establish an excuse, exemption or defence, rather than an element of the
offence, the presumption
of innocence is not infringed. Second, it argued that
the presumption of innocence applies with lesser force to
“regulatory”
offences, as opposed to truly “criminal”
offences. Finally, the Government argued that even if the presumption of
innocence
has been breached, that breach is justifiable under section 33(1) of
the Constitution.

[137] As regards the Government’s first argument,
I agree with O’Regan J that the distinction between an excuse, exemption,
or defence, as opposed to an element of the offence is irrelevant to the
question of whether the presumption of innocence has been
breached. This Court
has previously held that the presumption of innocence is breached whenever an
accused can be convicted despite
the existence of a reasonable doubt as to his
or her guilt.[2] That doubt can
arise with respect to an excuse, exemption or defence, as well as with respect
to an element of the offence. Because
section 332(5) permits the accused to be
convicted when he or she cannot prove certain facts on the balance of
probabilities, it
necessarily follows that the accused can be convicted despite
the existence of a reasonable doubt as to his or her guilt. Section
332(5) is
therefore in breach of the presumption of innocence.

[138] As regards the
Government’s second argument, I agree that it is unnecessary to decide
whether the presumption of innocence
applies with lesser force to so-called
“regulatory” offences. The reach of section 332(5) extends greatly
beyond offences
which might be characterised as purely
“regulatory”.

[139] Finally, the Government has not shown
that the breach of the presumption of innocence is justifiable under section
33(1) of
the Constitution. As explained by Mahomed DP, the apparent aims
underlying section 332(5), namely the protection of the public from
risks
generated by the activities of corporate bodies, and the inculcation of high
standards of behaviour in directors with respect
to those activities, are
entirely legitimate. Besides, corporate business activity is crucial for
effective and much needed socio-economic
growth. Legislation should therefore
not be unduly invasive and act to stifle sound economic activity. Like my
colleagues, however,
I am extremely concerned also by the potential for harm
which can result from corporate recklessness, dishonesty and undue exploitation.
Illegal toxic waste dumping, corporate recklessness in the area of health and
safety in the work environment, particularly in situations
of underground
mining, are not remote possibilities in this country.

[140] Despite my
concerns about the dangers of corporate activity, I come to the conclusion that
the breach of the presumption of
innocence is not justified. The Government has
failed to show, for example, that requiring the prosecution to establish all the
elements of section 332(5) will result in particular difficulty in obtaining
convictions. That failure becomes more egregious when
one considers the
breadth of section 332(5). Section 332(5) applies to all offences, and
no attempt has been made to confine the operation of the section to situations
where there is a special risk to an
unsuspecting and vulnerable public.


[141] As regards the order in this case, I concur with O’Regan J
that severance of certain words from section 332(5), so that
the legal burden of
proof is removed from the accused, is an appropriate remedy in this
case.

[142] The test for whether severance is appropriate was stated by
this Court in Coetzee v Government of the Republic of South Africa; Matiso
and Others v Commanding Officer, Port Elizabeth Prison, and Others
:

“[I]f the good is not dependent on the bad and can be separated from it,
one gives effect to the good that remains after the
separation if it still gives
effect to the main objective of the
statute.”[3]



[143] As
recognized in Schachter v Canada, the decision to sever “rests on
an assumption that the legislature would have passed the constitutionally sound
part of the
scheme without the unsound
part.”[4] Further, the
purpose of severance is “to be as faithful as possible within the
requirements of the Constitution to the scheme
enacted by the
Legislature.”[5] That
insight was elaborated upon by Sachs J at para 75 of Coetzee:

“[I]n deciding whether the Legislature would have enacted what survives on
its own, we must take account of the coming into
force of the new Constitution .
. . . We must, accordingly, posit a notional, contemporary Parliament dealing
with the text in issue,
paying attention both to the constitutional context and
the moment in the country’s history when the choice about severance
is to
be made. It is in this context that we must decide whether the good can be
separated from the bad.”



[144] In my view, as
stated earlier, that test can be applied to make severance a viable remedy in
this case. The “bad”
is the imposition of a legal burden of proof
on the accused. The “good” is the imposition of criminal liability
on directors
who do not prevent commission of offences by their corporation,
when they are able to do so. The “bad,” i.e. the legal
burden of
proof on the accused, can be removed without detracting significantly from the
main objective of the section. The prosecution
will still be able to obtain
convictions under section 332(5) in the appropriate circumstances, although it
will become more difficult
(although not impossible) for them to do so.


[145] The severance proposed by O’Regan J is the removal of the
words “or servant” and “it is proved that
he did not take part
in the commission of the offence and that” from section 332(5). The
section, as severed, would read as
follows:

“When an offence has been committed, whether by the performance of any act
or by the failure to perform any act, for which
any corporate body is or was
liable to prosecution, any person who was, at the time of the commission of the
offence, a director
. . . of the corporate body shall be deemed to be guilty of
the said offence, unless . . . he could not have prevented it, and shall
be
liable to prosecution therefor, either jointly with the corporate body or apart
therefrom, and shall on conviction be personally
liable to punishment
therefor.”



[146] This truncated form of section
332(5) is capable of being interpreted as imposing an evidentiary burden
on the accused, or indeed shifting the burden entirely on the prosecution.
I do
not find it necessary to decide which of those interpretations is correct, as
neither infringes the presumption of innocence.

[147] Langa J considers
that even this truncated form of section 332(5) imposes a legal burden on the
accused. Indeed that may be
a possible interpretation, and there are certainly
examples in the case law of provisions prefaced by the word “unless”
being interpreted to impose a legal burden on the
accused.[6] Section 35(2) of the
Constitution, however, is quite clear that where there are competing possible
interpretations of a statute,
courts are now required to “read
down” the statute, i.e. to choose the interpretation which is consistent
with the Constitution.[7] In my
view, the course of severance, combined with reading down the severed section,
is a reasonably permissible route to follow
in bringing section 332(5) in line
with the Constitution.

[148] Further, I agree that we are not required by
reason of section 90 of the Act to interpret the severed form of section 332(5)
as imposing a legal burden of proof on the accused. Section 90 provides as
follows:

“In criminal proceedings any exception, exemption, proviso, excuse or
qualification, whether it does or does not accompany
in the same section the
description of the offence in the law creating the offence, may be proved by the
accused but need not be
specified or negatived in the charge and, if so
specified or negatived, need not be proved by the
prosecution.”



[149] The effect of this section is
that where a matter is an “exception, exemption, proviso, excuse or
qualification,”
the legal burden with respect to that matter is imposed on
the accused. Unlike Langa J, I do not consider the ability of directors
to
prevent the commission of offences by their corporations to be
“exceptions, exemptions” etc, so as to impose the
legal burden of
proof on the accused.

[150] It is well established that several factors
should be considered in deciding whether a matter is an “exception,
exemption,
proviso, excuse or qualification” for the purposes of section
90. As stated in R v Kula, those factors include “the grammatical
shape of the provision, its context, its apparent scope and object and the
practical
consequences of the competing
constructions”.[8]

[151] With
respect, in my view it is incorrect to conclude that, simply because section
332(5) contains the word “unless”,
everything which follows that
word must necessarily be an exception, exemption or qualification. That
conclusion, based simply on
the grammatical structure of the section, ignores
the other factors applicable. I consider the most important factor to be the
real
mischief aimed at by section 332(5), which is to penalise directors who
take part in the commission of an offence, or fail to prevent
its commission
when they are able to do so. Thus, as explained by Mahomed DP, ability to
prevent the offence is not an exception
to criminal liability, but lies at the
heart of criminal liability.

[152] There is support in foreign case law
for the conclusion that where a reverse onus provision breaches the presumption
of innocence,
the appropriate remedy is to amend the provision so as to remove
the onus from the accused. In R v
Laba
,[9] the Canadian Supreme
Court held that a provision making it an offence to trade in precious metals
“unless [the accused] establishes
that he is the owner or agent of the
owner or is acting under lawful authority” was an unjustified breach of
the presumption
of innocence.1[0]
The remedy ordered was an amendment of the provision (using a combination of
severance and “reading in”) so as to impose
an evidentiary burden on
the accused.1[1] In Attorney-
General of Hong Kong v Lee Kwong-kut
the Privy Council held the reverse onus
provision at issue to be in conformity with the presumption of
innocence.1[2] The court added
that had it reached the opposite conclusion, it would have ordered severance of
the provision so as to alter the
burden of proof on the accused from a legal
burden to an evidentiary
burden.1[3]

[153] I do not
agree with Ackermann J’s view that the main objective of section 332(5) is
the reverse onus provision. Ackermann
J’s view is premised on the
supposed concurrence between the common law and section 332(5), except for the
reverse onus provision.
He argues that because the reverse onus provision is
the only substantial change to the common law made by section 332(5), it follows
that “the main (if not the exclusive) object of the section is limited to
the reversal of the onus of
proof”.1[4] With respect,
that analysis mistakes the reason for enacting a statute with the
objective of a statute. Consider, for example, a statute enacted to
codify existing law of theft, in order to simplify and make more accessible
the
law of theft. The main objective of the statute must surely still be to
penalise theft, rather than to simplify the law of theft.
Similarly, the main
objective of section 332(5), like the main objective of the common law,
is to penalise certain forms of behaviour by a director. The difference in the
burden of proof between the statutory provision
and the common law does not lead
to a difference in their main objectives. Rather, the difference lies in the
precise method of
attaining that objective.

[154] Finally, the
Applicants also argued, in the following terms, that section 332(5) breaches the
right to freedom and security
of the person, and the right to property,
guaranteed by sections 11(1) and 28 of the Constitution respectively:

“If the requirement of participation or an ability to prevent the offence,
is relegated to the status of a mere exception,
exemption or excuse, then it
follows that the elements of the criminal liability created under section
332(5), are merely that the
accused was a director or servant of a corporation
of which another director or servant committed an offence in the exercise of
his powers or in the performance of his duties or in furthering or endeavouring
to further the interests of the corporation. Such
an offence which exposes the
accused to a fine or imprisonment for the conduct of others in which he did not
participate and which
he could not have prevented, would violate the rights to
freedom and security of the person in terms of section 11(1) and property
in
terms of section 28 of the
Constitution.”1[5]



[155] The
applicants seem to be arguing that participation of a director in the commission
of an offence, and the ability of the director
to prevent such commission,
should be ignored for the purposes of determining the compliance of section
332(5) with sections 11(1)
and 28 of the Constitution. Such an argument can be
disposed of on the simple ground that section 332(5) must be assessed as it
stands, rather than with certain parts ignored. Because the applicants have not
made the argument that 332(5), in its entirety,
breaches sections 11(1) and 28
of the Constitution, I will not rule on the merits of that
argument.

[156] I therefore concur in the order proposed by O’Regan
J.


O’REGAN J:
[157] I have had the opportunity of reading
the judgments of both Langa J and Kentridge AJ. I agree with the order proposed
by Langa
J in respect of section 245 of the Criminal Procedure Act 51 of 1977
(“the Act”) and with the reasoning that motivates
that order. I
cannot however support the order that he has proposed in respect of section
332(5) of the Act. My approach to the
question of the constitutionality of that
section is somewhat different from his and from that proposed by Kentridge AJ.


[158] Section 332(5) permits directors (and servants) of companies to be
punished where the state has established that the company
is guilty of an
offence and a director (or servant) of that company has failed to satisfy the
court upon a balance of probabilities
that he or she did not participate in the
offence and could not have prevented the commission of the offence. It raises
two separate
constitutional inquiries. The first is whether it is
constitutionally legitimate for parliament to impose criminal liability on
directors and servants of corporate bodies in the circumstances contemplated by
section 332(5). The second is whether it is legitimate
for parliament to impose
upon an accused the burden of proving that he or she could not have prevented
the commission of the offence
rather than requiring the state to establish that
the accused could have prevented its commission. This question, of course,
requires
us to revisit section 25(3)(c) of the interim Constitution (“the
Constitution”), which has already been the subject of
a series of
decisions by this Court.

[159] These are separate questions. They raise
two different aspects of freedom: the first is concerned particularly with the
reasons
for which the state may deprive someone of freedom; and the second is
concerned with the manner whereby a person is deprived of freedom.
As I stated
in Bernstein and Others v Bester and Others NNO 1996 (2) SA 751
(CC); 1996 (4) BCLR 449 (CC) at paragraphs 145-7, our Constitution recognises
that both aspects are important in
a democracy: the state may not deprive its
citizens of liberty for reasons that are not acceptable, nor, when it deprives
citizens
of freedom for acceptable reasons, may it do so in a manner which is
procedurally unfair. The two issues are related, but a constitutional
finding
that the reason for which the state wishes to deprive a person of his or her
freedom is acceptable, does not dispense with
the question of whether the
procedure followed to deprive a person of liberty is fair. With respect,
therefore, I cannot agree with
Kentridge AJ when he states at paragraph 93 of
his judgment:

“In brief, if an offence of absolute liability had been created, it would
not in itself have given rise to any question of
the unfairness of the trial of
such an offence. Where the severity of such a provision has been mitigated by
allowing the accused
to prove a special defence it is in my view illogical if
not perverse to say that this destroys the fairness of the
trial.”



[160] I accept that an offence of absolute
liability would not give rise to a challenge under the fair trial provision of
the Constitution,
although it may well give rise to successful challenge under
section 11. However an absolute liability offence which afforded a
special
defence to the accused which required the accused to establish certain facts
upon a balance of probabilities would give rise
to constitutional enquiries
under both section 11 and section 25. As I stated above, the Constitution
limits not only the reason
for which the legislature may deprive a person of
freedom, but also the manner in which it is done. Two separate constitutional
inquiries are raised and it does not seem to me that both inquiries will, as a
matter of logic and principle, reach the same result.
That is because different
considerations arise under each. It is my view that an affirmative answer to
the first question does
not dispense with any need to consider the second,
although similar considerations may deserve attention in addressing both
questions.
If under either inquiry the provision falls foul of the established
constitutional standard, and is not held to be justifiable in
terms of section
33, it will be held to be invalid.
Section 11(1)
[161] I turn now
to the first of the two inquiries required of us. Is section 332(5) of the Act,
to the extent that it imposes criminal
liability upon directors and servants of
corporate bodies in certain circumstances, in breach of section 11(1) of the
Constitution?
Much of counsel’s argument to us was concerned with the
question of whether the words in subsection 332(5) “unless it
is proved
that he did not take part in the commission of the offence and that he could not
have prevented it” constitute an
element of an offence, or an exemption or
defence to the offence established in the first part of the section. In my
view, this
line of inquiry has little purpose in the context of an inquiry
concerning a possible breach of section 11 of the Constitution.
In determining
whether the provision is in breach of section 11, we are required to look at the
provision as a whole to determine
the circumstances in which, according to the
provision, an accused person may be deprived of his or her liberty. Where the
legislature
creates an offence, with a special defence, the legislature is, in
effect, authorising the conviction of a person once the elements
of the offence
have been proved and no defence exists. The overall guilt therefore lies in a
consideration of both the elements
of the offence and the elements of any
special defence. The offence should be looked at as a whole to determine
whether, on an appraisal
of all the elements of the offence and the established
defences, a deprivation of liberty may occur which would be in breach of section
11. In my view, the question of the burden of proof itself is not a question
which should be dealt with under section 11. This
question falls for
consideration under section 25(3). Therefore, it does not seem to me to matter,
for the purposes of the constitutional
inquiry in terms of section 11, whether
section 332(5) contains an offence of liability coupled with a special defence,
or whether
all the elements contained in section 332(5) are seen as elements of
a new offence.

[162] I turn then to a consideration of section 11. The
general principle of our common law is that criminal liability arises only
where
there has been unlawful conduct and blameworthiness or fault (the actus reus and
mens rea). This principle is ordinarily expressed
in the Latin maxims actus
non facit reum nisi mens sit rea
and nulla poena sine culpa. At
common law, the fault requirement is generally met by proof of intent (dolus) in
one of its recognised forms, and, in rare circumstances,
by the objective
requirement of negligence (culpa). (See Burchell and Milton, Principles of
Criminal Law
(Juta, Cape Town, 1991) at 71 - 5; Snyman Criminal Law 2
ed (Butterworths, Durban 1989) at 25 - 9; De Wet and Swanepoel Strafreg
4e uitgawe (Butterworths, Durban 1985) at 103). As Kentridge AJ has mentioned
in paragraph 94 of his judgment, the requirement
of fault or culpability is an
important part of criminal liability in our law. This requirement is not an
incidental aspect of our
law relating to crime and punishment, it lies at its
heart. The state’s right to punish criminal conduct rests on the notion
that culpable criminal conduct is blameworthy and merits punishment. This
principle has been acknowledged by our courts on countless
occasions. For
example, in R v Wunderlich 1912 TPD 1118, De Villiers JP held that:



“There is no doubt that as a general rule a person is not criminally
liable unless he has what is called mens rea. This is usually expressed
by the maxim: actus non facit reum nisi mens sit rea. This is a sound
rule, for a person is not to be subjected to the stigma and other consequences
of a crime unless he had what is
sometimes called a guilty mind. And from this
it follows that in general a person is not criminally liable for an act or
omission,
unless he himself has committed or omitted the act or has authorised
it.” (At 1121; cited with approval in R v Weinberg 1939 AD 71 at
82; Ex parte Minister of Justice: in re R v Nanabhai 1939 AD 427 at
429.)



[163] In the last hundred years, however, the
legislature has enacted many provisions which stipulate that criminal liability
will
arise upon proof that a person has committed, or omitted to do, a
particular act. These provisions contain either no mental element
and are
referred to as offences of absolute liability or provide the accused with a
defence of due diligence, or something similar,
in which case they are referred
to as offences of strict liability (see Snyman, cited above, at 242 - 9;
Burchell and Milton, cited
above, at 315 - 8). The most important justification
for these new forms of criminal offence is that their purpose is to ensure
compliance with regulatory norms which may not otherwise be observed. Until
1994, because of the doctrine of parliamentary sovereignty,
it was plain that
the courts had no choice but to enforce these criminal
provisions.

[164] However, such criminal liability has been criticised by
academic writers in South Africa (see, for example, Burchell and Milton,
cited
above, at 317.) It was also disapproved of by the Viljoen Commission in its
report on The Penal System of the Republic of
South Africa (1976) at para
5.1.2.82:

“In spite of the recognition in certain legal systems of the so-called
strict liability offences, this Commission remains impenitent
and adamant in
expressing the view that strict liability offences cannot be justified in the
criminal law. If the “offender”
unwittingly commits an act which is
prohibited by the criminal law under circumstances which totally absolve him
from any blame,
what is the object in punishing or even penalising him? There
would, in the Commission’s view, be no sense in doing
so.”




[165] Repugnance to the notion of criminal
liability without fault is evidenced too in the reluctance of courts to
interpret statutory
provisions which contain no express mens rea
requirement as not requiring mens rea. In S v Arenstein 1967 (3) SA 366
(A) at 381D - E, Van Winsen AJA held as follows:

“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.”


(See, also, amongst other decisions, R v
H
1944 AD 121 at 125; S v Bernardus 1965 (3) SA 287 (A) at 296F;
S v Oberholzer 1971 (4) SA 602 (A) at 610H - 611A.)

[166] The
principle that fault is a prerequisite for criminal liability is also present in
the law of other jurisdictions. It has
been repeatedly recognised as a
fundamental principle of English law. In Harding v Price [1948] 1 KB 695
at 700, for example, Lord Goddard CJ held:

“The general rule applicable to criminal cases is actus non facit reum
nisi mens sit rea, and I venture to repeat what I said
in Brend v Wood
[(1946) 62 TLR 462 at 463]: ‘It is of the utmost importance for the
protection of the liberty of the subject that a court should
always bear in mind
that, unless a statute either clearly or by necessary implication rules out mens
rea as a constituent part of
a crime, the court should not find a man guilty of
an offence against the criminal law unless he has a guilty
mind.’”



[167] In the leading American
decision, Morissette v United States 342 US 246 (1952) at 250, Jackson J
held that:

“The contention that an injury can amount to a crime only when inflicted
by intention is no provincial or transient notion.
It is as universal and
persistent in mature systems of law as belief in freedom of the human will and a
consequent ability and duty
of the normal individual to choose between good and
evil.”



[168] Other jurisdictions, too, have
experienced the growth of legislatively imposed strict and absolute liability.
Their courts,
too, have displayed hesitance in interpreting statutory provisions
as imposing absolute or even strict liability. In England, the
courts have
taken the view that there is a presumption that mens rea is always a requirement
of a criminal offence, although that
presumption may be defeated by the language
of a provision. In the case of Sherras v De Rutzen [1895] 1 QB 918 at
921, the court held that:

“There is a presumption that mens rea, an evil intention, or a knowledge
of the wrongfulness of the act, is an essential ingredient
in every offence; but
that presumption is liable to be displaced either by the words of the statute
creating the offence or by the
subject-matter with which it deals, and both must
be considered.”


Renewed vigour has been afforded to
this approach by its recent restatement in a series of decisions by the House
of Lords and the
Privy Council. (See Lim Chin Aik v R [1963] AC 160 (PC)
at 172; R v Warner [1969] 2 AC 256 (HL (E)) at 271-2; Sweet v
Parsley
[1970] AC 132 (HL (E)) at 163; Gammon (Hong Kong) Ltd v
Attorney-General of Hong Kong
[1984] 2 All ER 503 (PC) at 507-8). Despite
the frequent restatements of the rule in Sherras’s case, however,
clear guidelines both as to the circumstances in which a court may interpret a
statute as not requiring mens
rea, and as to the nature of the fault requirement
that will be implied when a court is of the view that the legislature did not
intend to oust a fault requirement, remain elusive.

[169] The approach in
Sherras’s case has been adopted in several of the countries of the
Commonwealth. In Australia, in addition to the presumption against
strict
liability, the courts have developed a defence of reasonable mistake of fact.
Thus an accused, who could show that he or
she held an honest and reasonable
belief in the existence of circumstances, which if true would render the accused
innocent of the
charge, will be acquitted. (See Maher v Musson (1934) 52
CLR 100 at 104-5; Thomas v R (1937) 59 CLR 279 at 287-8; Proudman v
Dayman
(1941) 67 CLR 536 at 540; Iannella v French (1967-1968) 119
CLR 84 at 93-4; R v Bush (1974-5) 5 ALR 387; He Kaw Teh v R (1985)
60 ALR 449 at 455.)

[170] The New Zealand courts, too, have been
reluctant to accept that parliament intends an absolute liability in the absence
of an
express statement to that effect. In so doing, they have considered the
approach adopted in Canada, as well as that adopted in England
and Australia.
The difficulties that the courts in these jurisdictions have had in establishing
a uniform and coherent approach
to the problem are well described in a
relatively recent judgment of the New Zealand Court of Appeal, Millar v
Ministry of Transport
[1986] 1 NZLR 660 (CA). Cooke P categorised the
various responses by courts in England, Australia, Canada and New Zealand to
statutory
offences which contain no express fault requirement. What is clear
from his judgment is that although the courts have been uniformly
reluctant to
interpret a statutory offence which contains no express culpability requirement
as entirely dispensing with culpability,
they have been unable to forge a simple
approach to the difficulty. That problem is, however, only of indirect
relevance to our
current inquiry.

[171] The courts of the United States
have also shown resistance to accepting that a statutory offence dispenses with
the requirement
of culpability. The approach of the court was summarised by
Burger CJ in United States v US Gypsum Co 438 US 422 (1977) where he
remarked:

“While strict-liability offenses are not unknown to the criminal law and
do not invariably offend constitutional requirements,
see Shevlin-Carpenter
Co v Minnesota
218 US 57 (1910), the limited circumstances in which Congress
has created and this Court has recognized such offenses, see e.g., US v
Balint
258 US 250 (1921); US v Behrman 258 US 280 (1921); US v
Dotterweich
320 US 277 (1943); U S v Freed 401 US 601 (1970) attest
to their generally disfavoured status. ... Certainly far more than the simple
omission of the appropriate
phrase from the statutory definition is necessary to
justify dispensing with an intent requirement.” (At 437 -
8.)



[172] In Morissette’s case, cited above,
the Supreme Court was concerned with 18 USC  641 which provided that
“whoever embezzles, steals,
purloins, or knowingly converts”
government property is punishable by fine and imprisonment. Jackson J rejected
the argument
that the statute should be interpreted as dispensing with a mens
rea requirement:

“The Government asks us by a feat of construction radically to change the
weights and balances in the scales of justice. The
purpose and obvious effect of
doing away with the requirement of a guilty intent is to ease the
prosecution’s path to conviction,
to strip the defendant of such benefit
as he derived at common law from innocence of evil purpose, and to circumscribe
the freedom
heretofore allowed juries. Such a manifest impairment of the
immunities of the individual should not be extended to common-law crimes
on
judicial initiative.” (At 263.)



[173] The Canadian
Supreme Court in pre-Charter days also expressed a reluctance to endorse
absolute liability. In the case of R v City of Sault Ste. Marie (1978)
85 DLR (3d) 161 (SCC), the court interpreted a statute which contained no
express fault requirement as one which permitted
a defence of due diligence to
the accused.

[174] Since the adoption of the Charter, the Canadian
Supreme Court has held that where a statute imposes criminal liability without
any mens rea requirement (ie. absolute liability) which may result in
imprisonment, it will be a breach of section 7 of the Canadian
Charter of Rights
and Freedoms. (See Reference re section 94(2) of the Motor Vehicle Act
(1986) 24 DLR (4th) 536 (SCC); R v Vaillancourt (1988) 47 DLR (4th) 399
(SCC); R v Wholesale Travel Group Inc. (1992) 84 DLR (4th) 161 (SCC);
R v Nova Scotia Pharmaceutical Society (1992) 10 CRR (2d) 34 (SCC); R
v Burt
(1987) 60 CR (3d) 372 (Sask CA); R v Pellerin (1990) 42 CRR
292 (Ont CA); R v Sutherland (1990) 55 CCC (3d) 265 (NS CA).) Section 7
of the Canadian Charter is formulated differently to section 11 of our interim
Bill of
Rights. It provides that:

“Everyone has the right to life, liberty and security of the person and
the right not to be deprived thereof except in accordance
with the principles of
fundamental justice.”



[175] In Vaillancourt,
Lamer J (as he then was) held:

“In effect, Reference re Motor Vehicle Act acknowledges that,
whenever the state resorts to the restriction of liberty, such as imprisonment,
to assist in the enforcement of
a law, even, as in Reference re Motor Vehicle
Act
, a mere provincial regulatory offence, there is, as a principle of
fundamental justice, a minimum mental state which is an essential
element of the
offence. It thus elevated mens rea from a presumed element in Sault
Ste. Marie
[R v City of Sault Ste. Marie] to a constitutionally
required element. Reference re Motor Vehicle Act did not decide what
level of mens rea was constitutionally required for each type of offence,
but inferentially decided that even for a mere provincial regulatory offence
at least negligence was required, in that at least a defence of
due diligence must always be open to an accused who risks imprisonment
upon conviction.” (At 414.)



[176] The striking
degree of correspondence between different legal systems in relation to an
element of fault in order to establish
criminal liability reflects a fundamental
principle of democratic societies: as a general rule people who are not at fault
should
not be deprived of their freedom by the state. This rule is the
corollary of another rule which the same comparative exercise illustrates:
when
a person has committed an unlawful act intentionally or negligently, the state
may punish them. Deprivation of liberty, without
established culpability, is a
breach of this established rule. Where culpability is established, and the
conduct is legitimately
deemed unlawful, then no such breach
arises.

[177] What is also clear however, from an examination of our law
and that of foreign jurisdictions is that it is widely recognised
(both in our
common law and in the law of other countries) that the culpability required to
establish criminal liability need not
in all circumstances be evidenced by
direct intent (dolus directus) on the part of the accused to commit a criminal
act. In our
own law other forms of intent, such as dolus eventualis, have been
recognised as sufficient to meet the requirement of culpability,
and in certain
circumstances, the law has recognised that even negligence or culpa, can be
sufficient to give rise to criminal liability.
It is not necessary for the
purposes of this case for us to determine the level of culpability required by
section 11. Indeed the
appropriate form of culpability may well be affected by
the nature of the criminal prohibition as well as other factors. In addition,
it should be borne in mind that significant leeway ought to be afforded to the
legislature to determine the appropriate level of
culpability that should attach
to any particular unlawful conduct to render it criminal. It is only when the
legislature has clearly
abandoned any requirement of culpability, or when it has
established a level of culpability manifestly inappropriate to the unlawful
conduct or potential sentence in question, that a provision may be subject to
successful constitutional challenge.

[178] The question that arises then
is whether section 332(5) imposes a form of culpability, sufficient to justify
the deprivation
of freedom without giving rise to constitutional complaint. To
determine the answer to this question requires a careful analysis
of section
332(5).

[179] It provides that once the state has established that a
company has committed an offence, an accused will bear a burden of proof
to
establish upon a balance of probabilities that he or she did not take part in
the commission of the offence and that he or she
could not have prevented its
commission. The director will be liable for the offence, regardless of whether
the offence is one which,
at common law, contains a mens rea requirement, such
as fraud or theft. Liability will arise therefore even though there has been
no
positive act by the accused, but merely an omission.

[180] In S v
Klopper
1975 (4) SA 773 (A) the court was concerned with a forerunner to
section 332(5), section 381(5) of Act 56 of 1955, which was in identical
terms.
In that case, the accused had been charged with attempted fraud on the basis of
the provisions of section 381(5). At trial,
the accused gave evidence to the
effect that he was not aware of the offence committed. He conceded that his
ignorance was due in
part to his own negligence in failing to read sufficiently
carefully the documents which contained the fraud which were sent to him.
The
Witwatersrand Local Division convicted him of fraud, but referred to the
Appellate Division the question as to whether the court
had been correct in
concluding that the accused’s evidence did not absolve him of liability.
Kotzϑ AJA considered the
requirement imposed by section 381(5). He held
that:



“Na my mening behoort sub-art. (5), wat - soos reeds aangedui - ’n
vorm van strafpligtigheid oplΛ, op die mins verswarende
wyse uitgelΛ
te word. Ten einde ’n objektiewe vertolking te regverdig, behoort ,n
kwalifikasie, soos bv. ‘redelikerwyse’
of ‘sonder
nalatigheid’ voor die woorde ‘kon verhoed het nie’ ingelees te
word. Sonder so ’n kwalifikasie
in te voeg - waarvoor ek in ,n
strafbepaling, soos hierdie, geen regverdiging kan sien nie - is dit onmoontlik
om te beslis dat die
Wetgewer ,n objektiewe uitleg wou voorskryf.” (At 780
B - D.)



[181] Accordingly, the judge held that the
accused had met the onus imposed upon him by section 381(5) and set aside the
conviction
and sentence of the accused. In so holding, Kotzϑ AJA expressly
rejected the view that the section requires an accused to place
objective facts
before the court proving, not only that he or she was not aware of the
commission of the offence, but that he or
she had taken reasonable steps to
prevent the commission of the offence, an approach which had been adopted in
S v Salama Taxis (Pty) Ltd and Others 1964 (1) SA 371 (C) at 376C-H
and followed in S v Poole 1975 (1) SA 924 (N) at 934E. The rule
in S v Klopper has been repeatedly applied. (See S v Film Fun Holdings
(Pty) Ltd and Others
1977 (2) SA 377 (E) at 386H; S v Deal Enterprises
(Pty) Ltd and Others
1978 (3) SA 302 (W) at 314H-315A and 315H-316A; S v
Harper and Another
1981 (2) SA 638 (D) at 641G-H.)

[182] The nature
of the liability imposed upon directors by the provision then is that they will
be liable for any criminal offence
committed by the company unless they can show
that they did not have knowledge of it and/or that they could not have prevented
it.
According to the interpretation of the section in Klopper’s
case, the test for prevention is not an objective one. Directors need not show
that they exercised due diligence and care
to prevent the commission of the
offence, they need merely show that they could, as a matter of subjective fact,
not have prevented
it. In this sense, the provision as interpreted requires a
form of subjective blameworthiness.

[183] Imposing criminal liability
upon a director who knows of the commission of an offence by the company and who
is in a position
to prevent the commission of that offence but does not do so is
not in any sense egregious. Actual knowledge coupled with the ability
to
prevent the commission of the offence by a director who is in a position of
control in the corporate body renders the failure
to do so sufficiently culpable
to warrant criminal liability.

[184] In addition to directors, however,
section 332(5) also renders “servants” liable in circumstances where
they do
not participate in an offence of the company, but could have prevented
it. In my view, this is a legislative overreach. The class
of servants
includes a very wide range of persons, from those who act in a managerial
capacity to those who perform menial functions
only. I agree with Kentridge AJ,
for the reasons given by him at paragraph 101 of his judgment, that to impose
such liability upon
such a wide class of servants is in breach of section
11(1).

[185] The provisions of section 332(5) of the Act are therefore to
be distinguished from the provisions of section 6(6) of the Gambling
Act 51 of
1965, which we considered in Scagell and Others v Attorney-General of the
Western Cape and Others
1996 (11) BCLR 1446 (CC) at paragraph 33. In that
case the relevant section provided, in effect, that a servant who permitted the
playing of a gambling game would be guilty of an offence. The prosecution thus
bears the onus of establishing not only that a servant
permitted the
playing of a gambling game but also that he or she had the necessary mens rea.
The differences between that provision and the
one currently under consideration
are therefore material. No argument was addressed to us by the state to suggest
that the breach
of section 11(1) caused by the inclusion of the words “or
servant” in the subsection was justifiable. Indeed, it was
conceded in
argument that, to the extent that the section imposed liability upon servants of
corporations, it was constitutionally
intolerable. In the circumstances, I
would hold that the words “or servant” must be severed from the
section. Once
that has been done, in my view, section 332(5), as it has been
interpreted by the Appellate Division, is not in conflict with section
11(1) of
the Constitution. I reach the same conclusion as Kentridge AJ, therefore, but
for different reasons.

Section 25(3)
[186] The second question
for consideration, then, is whether section 332(5) in imposing a burden of proof
upon the accused is in
breach of section 25(3)(c) of the Constitution. It needs
to be emphasised that section 25(3) is an important constitutional right
in
itself. It recognises that before the state can impose a criminal sanction on a
person, that person must have been afforded a
fair trial. In so doing, the
section focuses on the importance of procedural fairness - long a cherished
value in democratic societies.

[187] In a series of cases, this court has
held that where a legislative provision imposes an obligation upon an accused to
establish
certain facts to avoid criminal liability it constitutes a breach of
the presumption of innocence as enshrined in section 25(3)(c).
(See S v Zuma
and Others
1995 (2) SA 642 (CC); 1995 (4) BCLR 401 (CC) at para 33; S v
Bhulwana; S v Gwadiso
1996 (1) SA 388 (CC); 1995 (12) BCLR 1579 (CC) at para
15; S v Mbatha; S v Prinsloo 1996 (2) SA 464 (CC); 1996 (3) BCLR 293 (CC)
at para 12; S v Julies 1996 (4) SA 313 (CC); 1996 (7) BCLR 899 (CC) at
para 3; Scagell and Others v Attorney-General of the Western Cape and Others
1996 (11) BCLR 1446 (CC) at para 7.) It is true that this court has left
open the question of whether, when a statute imposes a burden
upon an accused to
prove an element of a defence as opposed to the element of an offence, it will
be in breach of section 25(3)(c).
(See, in particular, S v Zuma, cited
above, paragraphs 41-2.) Counsel for the government argued that that question
arose for decision in this case.

[188] Counsel for the government argued
that section 332(5) imposes a burden upon an accused to prove facts to establish
a defence,
rather than an element of the offence. This argument does not assist
them. Even if the relevant portion of section 332(5) does
constitute a defence
rather than an element of an offence, it is my view that no distinction can
satisfactorily be drawn between
elements of an offence and elements of a defence
for the purposes of section 25(3)(c) and, in particular, the presumption of
innocence.

[189] We have stated on several occasions that the nub of the
protection provided by section 25(3)(c) is to ensure that people are
not
convicted of an offence where a reasonable doubt exists as to their guilt.
Guilt is only established when it is clear that the
accused has no defence and
that all the elements of the particular crime have been established. If an
accused person can be convicted
despite the existence of a reasonable doubt
either in relation to one of the elements of the offence or one of the elements
of a
defence and a court is compelled to convict because of a reverse onus
provision, the presumption of innocence is breached. As Dickson
CJC held in
R v Whyte (1989) 51 DLR (4th) 481 at 493:

“The short answer to this argument is that the distinction between
elements of the offence and other aspects of the charge
is irrelevant to the s
11(d) inquiry. The real concern is not whether the accused must disprove an
element or prove an excuse, but
that an accused may be convicted while a
reasonable doubt exists. When that possibility exists, there is a breach of the
presumption
of innocence.

The exact characterisation of a factor as an essential element, a collateral
factor, an excuse, or a defence should not affect the
analysis of the
presumption of innocence. It is the final effect of a provision on the verdict
that is decisive.”

(See also R v Keegstra (1991) 3 CRR (2d) 193 at
258-9.)



[190] I cannot accept, therefore, that section
332(5) would not be in breach of section 25(3)(c) of the Constitution simply
because
the burden imposed upon accused persons by that provision related only
to a possible defence and not to an element of the crime.
In my view, section
332(5) is in breach of section 25(3)(c) because an accused person may be
convicted of an offence despite the
existence of a reasonable doubt as to
whether he or she was in fact guilty.

[191] The question remains,
however, whether that breach may be justified in terms of section 33(1) of the
Constitution. For the
reverse onus to be held to be justifiable, we need to be
persuaded that section 332(5) is reasonable, justifiable and necessary in
an
open and democratic society based on freedom and equality. In determining
whether a provision is legitimate, the court must weigh
the infringement caused
against the purpose, effects and importance of the impugned provision. (S v
Makwanyane
and Another 1995 (3) SA 391 (CC); 1995 (6) BCLR 665 (CC)
at para 104.)

[192] Counsel argued that the purpose of section 332(5) was
to impose an obligation upon directors of companies to take every precaution
to
ensure that companies did not commit criminal acts. Imposing a legal burden
upon such a director, it was argued, would require
a director charged in terms
of the provision to present objective evidence to court of the steps he or she
had taken to prevent the
commission of offences by the company.
[193] It
may well be that one of the circumstances in which it will be justifiable to
impose a legal burden of proof upon an accused
is when the purpose of the legal
burden is to require a person diligently to take positive steps to prevent the
commission of certain
statutory offences. In many jurisdictions a distinction
has been drawn between criminal acts that are mala in se and criminal
acts that are mala in re prohibita, or as they have been, perhaps
inaptly, termed in English, criminal offences proper and regulatory offences.
(See for a discussion
of the distinction K R Webb Regulatory Offences, the
Mental Element and the Charter: Rough Road Ahead
in 21 Ottawa Law Review
419(1989).) The reason for the distinction has perhaps been most succinctly put
by Cory J in R v Wholesale Travel Group Inc. (1992) 84 DLR 4th 161 at
205-6 where he held that:

“It has always been thought that there is a rational basis for
distinguishing between crimes and regulatory offences. Acts
or actions are
criminal when they constitute conduct that is, in itself, so abhorrent to the
basic values of human society that it
ought to be prohibited completely.
Murder, sexual assault, fraud, robbery and theft are all so repugnant to society
that they are
universally recognized as crimes. At the same time, some conduct
is prohibited, not because it is inherently wrongful, but because
unregulated
activity would result in dangerous conditions being imposed upon members of
society, especially those who are particularly
vulnerable.

The objective of regulatory legislation is to protect the public or broad
segments of the public (such as employees, consumers and
motorists, to name but
a few) from the potentially adverse effects of otherwise lawful activity.
Regulatory legislation involves
a shift of emphasis from the protection of
individual interests and the deterrence and punishment of acts involving moral
fault to
the protection of public and societal interests. While criminal
offences are usually designed to condemn and punish past, inherently
wrongful
conduct, regulatory measures are generally directed to the prevention of future
harm through the enforcement of minimum
standards of conduct and
care.”



[194] Cory J then went on to hold that where
regulatory offences were concerned the shift of onus to the accused would not
constitute
a breach of section 11(d) of the Charter which entrenches the
presumption of innocence. Relying upon the distinction between criminal
offences and regulatory offences, he held that:

“[c]riminal offences have always required proof of guilt beyond a
reasonable doubt; the accused cannot, therefore, be convicted
where there is a
reasonable doubt as to guilt. This is not so with regulatory offences, where a
conviction will lie if the accused
has failed to meet the standard of care
required.” (At 224.)


A different, and in my view, a
more persuasive approach was taken in the same case by Iacobucci J who held
that the considerations
raised by Cory J concerning the distinction between
regulatory and criminal offences was relevant not to the question of whether
there had been a breach of section 11(d) but to the question of whether that
breach was justifiable in terms of section 1 of the
Canadian Charter which
requires any breach to be a “reasonable limit ... demonstrably justified
in a free and democratic society”.


[195] One of the main reasons
that Cory J gave for relying upon the distinction relates to the purpose that
regulatory offences are
meant to serve and, in particular, the purpose that
imposing a legal burden upon an accused to establish due diligence in the
context
of such offences may serve. Where an accused has to prove on a balance
of probabilities that he or she has exercised due diligence
to avoid the
commission of an offence, an accused will have to produce evidence of the
efforts that were made to exercise due diligence.
In many situations, this will
require an accused to show that steps were taken to avoid the offence. Where on
the other hand the
prosecution bears the legal burden to show absence of due
diligence, it is the prosecution which will need the detailed evidence
to
establish the absence of care. An accused would then need to raise only a
reasonable doubt. The implications of the different
burdens in a regulatory
context are clear. In the one, the accused must not only act with diligence to
avoid the commission of offences,
but he or she must have proof that such action
was taken. If such proof is not kept, then it may not be possible to establish
the
necessary defence. In the other, the prosecution will need to establish a
prima facie case of a failure to act diligently. Only
if the state discharges
such a burden, will the accused be required to rebut it by raising a reasonable
doubt. The need for an accused
to maintain evidence of his or her due diligence
will be considerably attenuated. In such circumstances an evidential burden
does
not have the same effect as a legal burden, and it may not be sufficient to
achieve the purposes of a regulatory statute. These
considerations may well
form the basis of a ground of justification under section 33(1) of our
Constitution.

[196] The distinction between regulatory offences and
criminal offences has been recognised in other jurisdictions as well. (See
Morissette v United States 342 US 246, 253-62 (1952); Sherras v De
Rutzen
[1895] 1 QB 918 at 922; Lim Chin Aik v R [1963] AC 160 (PC) at
174; Sweet v Parsley [1970] AC 132 (HL) at 163E-F; Gammon (Hong Kong)
Ltd v Attorney General of Hong Kong
[1984] 2 All ER 503 (PC) at 507-9;
Millar v Ministry of Transport [1986] 1 NZLR 660 (CA) at 668-9.) However
as these cases illustrate, it has also been the subject of judicial discord and
criticism.
It is not necessary for the purposes of this case for us to resolve
that discord or address those criticisms. A definitive pronouncement
of the
significance of such distinction for our law can await a case in which the
distinction has application.

[197] For what is clear in this case, as
both Langa J and Kentridge AJ note, is that section 332(5) is by no stretch of
the imagination
confined only to regulatory offences. It covers all offences,
common law and statutory. It cannot rely for justification therefore
on a
distinction drawn between regulatory and criminal offences as the government
sought to argue.

[198] Perhaps an even more cogent reason, however,
emerges from the analysis of section 332(5) which appears at paragraph 180 to
182
above: the defence which an accused person needs to establish in terms of
section 332(5) is not that he or she acted with due diligence.
An accused needs
to show that he or she did not have knowledge of the offence, or that he or she
could not have prevented it. Such
a defence does not require an accused to
produce evidence to show due diligence, but provides a defence to the negligent,
as happened
in Klopper’s case, above. It is hard in these
circumstances to accept that the onus contained in section 332(5) will
necessarily result in legislatively
desired diligent behaviour.

[199] For
these reasons, it is not necessary to consider any further the basis of the
distinction between regulatory and criminal
offences, nor the circumstances in
which a shift of onus in the context of regulatory offences will meet the
requirements of section
33. The government could not point to any other reason
or circumstance which would justify the infringement of section 25(3) of
the
Constitution that section 332(5) occasions. I am not satisfied that the
objectives of the provision could not have been achieved
by means less invasive
of the rights afforded by section 25(3). Nor was the government able to point
to a reason which would justify
the subsection's application to all offences,
regardless of their nature or purpose. In the circumstances, I have not been
persuaded
that there are grounds to find that section 332(5) is saved by the
terms of section 33(1) of the Constitution.

[200] It is for these
reasons, that I share the conclusion of Langa J that section 332(5) is
inconsistent with the provisions of the
interim Constitution. However, unlike
Langa J, I am in agreement with Kentridge AJ that the unconstitutionality of the
subsection
can be cured by severance. The form of severance I propose is
different from that proposed by Kentridge AJ and its result, in my
view, is not
an evidential burden as he suggests.

[201] In Coetzee v Government of
the Republic of South Africa; Matiso and Others v Commanding Officer, Port
Elizabeth Prison, and Others
1995 (4) SA 631 (CC); 1995 (10) BCLR 1382 (CC),
this Court held that:

“Although severability in the context of constitutional law may often
require special treatment, in the present case the trite
test can properly be
applied: if the good is not dependent on the bad and can be separated from it,
one gives effect to the good
that remains after the separation if it still gives
effect to the main objective of the statute. The test has two parts: first,
is
it possible to sever the invalid provisions and, second, if so, is what remains
giving effect to the purpose of the legislative
scheme?” (at paragraph 16
applying the rule in Johannesburg City Council v Chesterfield House (Pty)
Ltd
1952 (3) SA 809 (A) at 822D-E.)


It seems clear, too,
that the court should be loath to declare statutory provisions invalid, where
the result of such declaration
would be to invalidate aspects of a statutory
provision which are without constitutional blemish. If it is possible to
separate
the constitutional from the unconstitutional without giving rise to a
provision inconsonant with the original legislative scheme,
that is the course
the court should adopt. Two questions thus arise. Is it possible to sever the
constitutional aspects of section
332(5) from the unconstitutional aspects of
the section? If so, does the remainder give effect to the “purpose of the
legislative
scheme”? I will address each of these questions
separately.

[202] In my view the following words can be severed from
section 332(5): “it is proved that he did not take part in the commission
of the offence and that”. The section would then read:

“When an offence has been committed, whether by the performance of any act
or by the failure to perform any act, for which
any corporate body is or was
liable to prosecution, any person who was, at the time of the commission of the
offence, a director
... of the corporate body shall be deemed to be guilty of
the said offence, unless ... he could not have prevented it, and shall
be liable
to prosecution therefor, either jointly with the corporate body or apart
therefrom and shall on conviction be personally
liable to punishment
therefor.”


The first question to be considered in
determining whether this is a valid severance is whether what remains is indeed
constitutional.

[203] If the abridged provision still imposes a burden of
proof upon the accused, the severance will not have succeeded in separating
the
good from the bad. On a bare reading of the provision as severed, it does not
expressly impose a burden upon the accused. The
question then arises as to
whether section 90 of the Criminal Procedure Act 51 of 1977 applies. Section 90
governs the question
of whether, when a burden of proof is not expressly imposed
upon the accused, a provision such as the abridged provision will nevertheless
by read to impose a burden on the accused. It provides that:

“In criminal proceedings any exception, exemption, proviso, excuse or
qualification, whether it does or does not accompany
in the same section the
description of the offence in the law creating the offence, may be proved by the
accused but need not be
specified or negatived in the charge and, if so
specified or negatived, need not be proved by the
prosecution.”


The question that arises then is whether
the portion of the provision after severance that reads “unless he could
not have prevented
it” is an “exception, exemption, proviso, excuse
or qualification” within the terms of section 90. Section 90
has been the
subject of repeated consideration by the Supreme Court of Appeal. (See, for
example, R v Zondagh 1931 AD 8 at 16; R v Beebee 1944 AD 333 at
335-6; R v Kula and Others 1954 (1) SA 157 (A) at 159F-161H; R
v Shangase
1960 (1) SA 734 (A) at 735C-D; R v Moosa and Others 1960
(3) SA 517 (A) at 532F-H; Attorney-General, Cape v Bestall 1988 (3) SA
555 (A) at 568A-C.) The proper approach to the application of the section has
most recently been summarised in Nestadt
JA’s judgment in
Bestall’s case.

“What has to be decided is whether the negative element or excusing factor
forms a material part of the offence itself or whether
it is merely an exclusion
(to be established by the accused) from the general prohibition contained in the
provision. In each case
it is a question of construction of the relevant
legislation. Factors used as an aid in this regard, in addition to the form in
which the prohibition is cast, include the grammatical shape of the provision,
its context, its apparent scope and object and the
practical consequences of the
competing constructions. In addition, according to the so-called
‘truncation test’, assistance
may be derived from considering
whether, if the alleged exemption be excised, what remains looks like something
that the Legislature
might well have intended to make an offence.” (At
568A-C.)


To determine whether section 90 would have
application to the abridged provisions of section 332(5), one has to determine
whether
the words “unless he could not have prevented it” form a
material part of the offence created by section 332(5). It
seems to me that
these words are a material part of the offence. The offence does not lie in
being a director of a corporate body
that has committed an offence, but in
failing to prevent the commission of such an offence when it is possible to do
so. The purpose
of section 332(5) is to make it clear to directors that they
are under a legal duty to prevent the company from committing offences
of which
they have knowledge. Knowledge that an offence is being committed coupled with
the ability to prevent that commission are
the crux of the offence itself. In
my view, therefore, section 90 cannot apply to the abridged provision. As
section 90 has no
application in the case, it is not necessary to consider
whether section 90 itself constitutes a breach of section 25(3) of the
Constitution.

[204] Once it is found that the provision does not impose a
burden on the accused, whether expressly or as a consequence of the operation
of
section 90, it does not seem to me that it is open to an interpretation which
would impose such a burden. The Constitution now
requires that where a
provision is reasonably capable of an interpretation that is consistent with the
Constitution, that interpretation
should prevail (section 35(2)). As the
abridged provision seems reasonably capable of the interpretation that it
imposes a burden
upon the state, it seems to me that that is the interpretation
which should be given it. I come to the conclusion, therefore, that
the
severance would result in a constitutional provision and that the good may be
severed from the bad.

[205] The further question concerning severance is
whether what remains gives effect to the purpose of the legislative scheme. In
my view, it does. The purpose of the legislative scheme was twofold: to impose
a legal duty upon directors and servants to take
steps to prevent the commission
of offences about which they knew; and to impose a legal burden upon directors
and servants to establish
that they did not participate in an offence committed
by a company and could not have prevented it. I disagree with Ackermann J,
therefore, who adopts the view that the purpose of section 332(5) was merely to
shift the burden of proof to the accused in respect
of two elements of an
offence.

[206] The common law liability described in R v Shikuri
1939 AD 225 and relied upon by Ackermann J renders employers criminally liable
as accomplices because of the implied authority they
have given their employees
to act wrongfully in particular circumstances. This liability is a different
and narrower form of liability
than that imposed by section 332(5). The
liability created by that section is not based upon a doctrine of implied
authority but
upon the creation of a legal duty which binds directors and
servants to take steps to prevent the commission of criminal offences
by a
company, where they have knowledge of them. As such it renders an omission or
failure to act wrongful. This seems to have
been a clear purpose of the
legislature in enacting section 332(5). I cannot accept that such a duty
clearly exists at common law
as suggested by Ackermann J. In my view, the
purpose of the legislature in enacting section 332(5) was to avoid the
uncertainties
of the common law liability and to impose a firm duty upon
directors (and servants) of corporate bodies to take steps to prevent
the
commission of criminal offences by the company where they had knowledge of such
offences. If we were to declare the entire provision
invalid, we would render
void the duty created by the section. Whether the common law would fill the gap
thus created remains uncertain.
I cannot agree with Ackermann J therefore that
an order for severance would be inappropriate in this case.

[207] The
order I would propose in respect of section 332(5) is the following:

1. The words “or servant” and “it is proved that he did not
take part in the commission of the offence and that”
in section 332(5) of
the Criminal Procedure Act are inconsistent with the Constitution and are
declared to be invalid with effect
from the date of this
order.

2. In terms of section 98(6) of the Constitution, this declaration of invalidity
shall invalidate any application of section 332(5)
in its unabridged form in any
criminal trial in which the verdict of the trial court was or will be entered
after the Constitution
came into force, and in which, as at the date of this
judgment, either an appeal or review is pending or the time for noting of such
appeal has not yet expired.



SACHS J:
[208] For many
years a dispute existed as to whether a company, not having a mind of its own
capable of forming criminal intent, could
be prosecuted for a criminal
offence.[1] The matter was
resolved by the enactment of predecessor legislation to section 332, which
expressly made companies liable to prosecution
and laid down the conditions of
their culpability.[2] It was
felt, however, that the mere imposition of a fine to be defrayed from company
resources would be an insufficient sanction.
The Legislature accordingly decided
that

“. . . the public interests might require an additional sanction, viz.,
that the officers through whose delinquency or lack
of vigilance the company
became a defaulter should be made to suffer: hence the provision making the
officers liable in their individual
representative capacities unless they prove
absence of knowledge of the
contravention.”[3]


In
the striking words of an American court dealing with a similar matter, the
danger existed that fines established to deter crime
would become mere licence
fees for illegitimate corporate business
operations.[4]

[209] The
result was the enactment of section 332(5), the constitutionality of which is
under consideration in this matter. It reads
as follows:

“When an offence has been committed, whether by the performance of any act
or by the failure to perform any act, for which
any corporate body is or was
liable to prosecution, any person who was, at the time of the commission of the
offence, a director
or servant of the corporate body shall be deemed to be
guilty of the said offence, unless it is proved that he did not take part
in the
commission of the offence and that he could not have prevented it, and shall be
liable to prosecution therefor, either jointly
with the corporate body or apart
therefrom, and shall on conviction be personally liable to punishment
therefor.”



[210] Defending the constitutionality of
this provision, Mr Gauntlett argued that the vicarious liability it imposed
amounted to what
he called ‘regulatory’ liability and not true
criminal liability. He contended that persons who chose to assume a directorship
have, in so doing, placed themselves in a position of responsibility not only
vis-Β-vis the company but in relation to the public
generally.
Accordingly, they must accept (indeed, for all practical purposes, they are
deemed to accept) that the law requires them
to control the corporate body, and
otherwise discharge their duties as directors, in accordance with certain
minimum standards on
pain of civil and criminal liability. The realities and
complexities of the modern corporate world, coupled with the need to protect
society and all its members against the considerable harm which could result
from the reckless or inattentive stewardship of corporate
bodies, demanded that
the conduct of directors be effectively regulated. Directors must be impelled
to ensure that the conduct of
corporate bodies complied with the law. Seen in
this light, a provision which imputed to the directors of a corporate body
liability
for an offence committed by that body in cases where they did not
prove on a balance of probabilities that they could not have prevented
it,
served an important regulatory purpose which did not run counter to the
presumption of innocence, notwithstanding the fact that (which he accepted) in a
criminal context the same reverse
of onus would violate section 25(3) of the
Constitution.

[211] Mr Gauntlett’s alternative argument was that
if there was a breach of the presumption of innocence contained in section
25(3)(c), such breach was reasonable, justifiable and necessary, and thereby
saved by section 33(1). More particularly, the objectives
of section 332(5) -
protecting society and all its members against the considerable harm which could
result from the reckless or
inattentive stewardship of corporate bodies;
impelling directors to ensure that the conduct of corporate bodies complied with
the
law; and avoiding the loss of convictions of those who failed to do so
because the facts were particularly within the knowledge of
the accused - were
of sufficient importance to override the right guaranteed by section
25(3)(c).

[212] The essence of his argument was that the section was
intended to deter reckless or inattentive stewardship of companies, and
as such
essentially regulatory rather than punitive in character. If his
characterisation of section 332(5) is correct, then his
argument would
undoubtedly carry considerable weight. The legislative purpose would be not so
much to facilitate the punishment
of corporate crooks as to encourage the
prevention of corporate crookedness. Its aim would be to reinforce
accountability of company
directors, by ensuring that those in control of
companies took responsibility for both the conduct and the misconduct of those
whom
they hired, fired and directed. As the commanding brains of corporate
enterprises, they would, if called to account for their failure
to prevent
company misdemeanour, be subject not to capricious
targeting,[5] but, rather to
focused enquiry based on their responsible relation to the primary
perpetrators.[6] As the eyes,
ears and spokespersons of the corporation, it could possibly not be unreasonable
to infer or assume that they saw,
heard and spoke proven corporate evil, rather
than the reverse. Certainly it would not be unreasonable to hold them
personally to
account for the misdeeds of those obliged to do their
bidding,[7] provided that this
were done by penalising them for culpable lack of concern for keeping the
company on the straight and narrow,
rather than by punishing them by
attributing equal guilt when such could not be proven in the ordinary way.


[213] The first question before us, then, is whether it is possible to
read section 332(5) in such a way as to restrict its ambit
to regulatory
offences. This turns on the meaning to be given to the words “any act,
for which any corporate body is or was
liable to prosecution”. Attempting
to “read down” these words as I
may,[8] I can discover neither
textual nor contextual warrant for limiting them to acts or omissions relating
to regulatory offences only.
On the contrary, section 332(1) provides in
unqualified language that:



“For the purpose of imposing upon a corporate body criminal liability for
any offence, whether under any law or at common law
-

(a) any act performed, with or without a particular intent, by or on
instructions or with permission, express or implied, given by a
director or
servant of that corporate body; and

(b) the omission, with or without a particular intent, of any act which
ought to have been but was not performed by or on instructions
given by a
director or servant, of that corporate body,

in the exercise of his powers or in the performance of his duties as such
director or servant or in furthering or endeavouring to
further the interests of
that corporate body, shall be deemed to have been performed (and with the same
intent, if any) by that corporate
body or, as the case may be, to have been an
omission (and with the same intent, if any) on the part of that corporate
body.”


[214] The words “any act” are as
wide as they could be. The one limitation to the scope of this section is that
the corporation
would not be liable for prosecution in relation to crimes
committed by a director or servant for personal
interest,[9] but only for those
committed in furtherance of the corporation’s interests. Read with
section 332(5), the following further
qualifications could be made: Crimes such
as bigamy or rape which by their nature could only be committed by human persons
would
be excluded1[0] and the same
would apply to crimes which by statute can only be committed by natural
persons.1[1] Proof of absence of
knowledge of the offence could also enable the director to escape liability,
since a director could not have
prevented the offence if its existence was
unknown to him or
her.1[2]

[215] None of
these qualifications can be seen as limiting the application of section 332(5)
read with section 332(1), to regulatory
offences only. On the contrary, it has
been held that a corporate body can commit crimes based on intent and
negligence,1[3] and successful
prosecutions have been brought against companies for fraud, theft, and culpable
homicide.1[4] It is instructive
that the specimen indictment in a leading textbook on criminal law, reads as
follows:



“THAT X, charged in terms of section 332(1) of the Criminal Procedure Act
51 of 1977 in his representative capacity as a director
of the A Company
Limited, a corporate body, and charged in his personal capacity, in terms of
section 332(5) of the Act aforesaid,
and Y, the secretary and servant of the
said A Company Ltd, are guilty of the crime of fraud:



IN THAT, on or about . . ., at . . ., in the district of . . ., the said Y, in
the exercise of his powers . .
.”1[5]



[216] There
is, of course, no clear definition of what are regulatory offences. Yet,
whatever the term may cover, section 332(5),
even read in the most strained way
possible so as to favour constitutionality, cannot be limited to embracing such
offences only.
The typical matter prosecuted, namely, fraud, is clearly not a
mere regulatory offence, but a particularly ugly species of white
collar crime,
castigated as such by society, and carrying with it the prospects of heavy terms
of imprisonment.

[217] For present purposes what matters is not so much
the definition of regulatory offences, but an evaluation of the underlying
reasons for treating them in a special way, and thereby for permitting or even
requiring departure from the normal rules relating
to onus of proof. I would
like to refer to an American case which was cited by Mr. Gauntlett in favour of
upholding the section
in question, but which, in my view, goes emphatically the
other way. This is the matter of Morissette v United
States.
1[6] Although this
case dealt with the issue of whether statutory offences required intent or not,
the underlying questions of legal
principle were precisely the same as those in
the present matter, namely, of determining when the conditions of modern society
justified
legislative departure from time-honoured protections given to
defendants. Given the eloquence and pertinence of the opinion of Justice
Jackson (for the court), I trust I will be forgiven for quoting from it at some
length:1[7]



“Crime, as a compound concept, generally constituted only from concurrence
of an evil-meaning mind with an evil-doing hand,
was congenial to an intense
individualism and took deep and early root in American soil.

. . . .

[However,] [t]he industrial revolution multiplied the number of workmen exposed
to injury from increasingly powerful and complex
mechanisms, driven by freshly
discovered sources of energy, requiring higher precautions by employers.
Traffic of velocities, volumes
and varieties unheard of came to subject the
wayfarer to intolerable casualty risks if owners and drivers were not to observe
new
cares and uniformities of conduct. Congestion of cities and crowding of
quarters called for health and welfare regulations undreamed
of in simpler
times. Wide distribution of goods became an instrument of wide distribution of
harm when those who dispersed food,
drink, drugs, and even securities, did not
comply with reasonable standards of quality, integrity, disclosure and care.
Such dangers
have engendered increasingly numerous and detailed regulations
which heighten the duties of those in control of particular industries,
trades,
properties or activities that affect public health, safety or welfare.

. . . .

[“[P]ublic welfare offences”] do not fit neatly into any of [the]
accepted classifications of common-law offences, such
as those against the
state, the person, property, or public morals. Many of these offences are not
in the nature of positive aggressions
or invasions, with which the common law so
often dealt, but are in the nature of neglect where the law requires care, or
inaction
where it imposes a duty. Many violations of such regulations result in
no direct or immediate injury to person or property but merely
create the danger
or probability of it which the law seeks to minimize. While such offences do
not threaten the security of the
state in the manner of treason, they may be
regarded as offences against its authority, for their occurrence impairs the
efficiency
of controls deemed essential to the social order as presently
constituted. In this respect, whatever the intent of the violator,
the injury
is the same, and the consequences are injurious or not according to fortuity.
Hence, legislation applicable to such offences,
as a matter of policy, does not
specify intent as a necessary element. The accused, if he does not will the
violation, usually is
in a position to prevent it with no more care than society
might reasonably expect and no more exertion than it might reasonably
exact from
one who assumed his
responsibilities.1[8] Also,
penalties commonly are relatively small, and conviction does no grave damage to
an offender’s reputation.

. . . .

Stealing, larceny, and its variants and equivalents [on the other hand], were
among the earliest offences known to the law that existed
before legislation;
they are invasions of rights of property which stir a sense of insecurity in the
whole community and arouse public
demand for retribution, the penalty is high
and, when a sufficient amount is involved, the infamy is that of a felony,
which, says
Maitland, is “. . . as bad a word as you can give to man or
thing.”

. . . .

[W]e cannot accept [cases not requiring intent for public welfare offences] as
authority for eliminating intent from offences incorporated
from the common
law.

. . . .

The Government asks us by a feat of construction radically to change the weights
and balances in the scales of justice. The purpose
and obvious effect of doing
away with the requirement of a guilty intent is to ease the prosecution’s
path to conviction, to
strip the defendant of such benefit as he derived at
common law from innocence of evil purpose, and to circumscribe the freedom
heretofore
allowed juries.”



[218] In my view, this
is a convincing line of reasoning which is directly apposite to the question
before us. The same rationale
justifying departure from the normal rules
relating to proof of a guilty mind, could apply to deviation from the standard
practice
in respect of onus of proof. Conversely, the same considerations in
respect of honouring proof of guilty intent in relation to common
law offences
such as fraud, should operate with reference to onus of proof in such matters.
Yet, common law offences of a serious
nature, carrying heavy penalties and
severe social disgrace, are picked up by section 332(5) and imputed to persons
who are deemed
guilty unless they can prove their innocence in one of the
manners prescribed. The presumption of innocence is violated, not as
a matter
of overwhelming practical
convenience,1[9] or to prove
maintenance of standards for a licensed
activity,2[0] but simply to
facilitate conviction. Indeed, the very purpose of the strong deeming provision
is to invert the normal relationship
between prosecution and defence. A nexus
of easily inferred fact, which in practice would aid a finding of guilt
according to the
normal onus of proof criteria, is converted into a nexus of
law, opening up the very real possibility of a finding of guilt followed
by
severe punishment, even though the trial court had real doubts on the
matter.

[219] The problem with section 332(5) is that because of its wide
generic and mutable character, it serves as a deeming clause for
all seasons,
both for those where it is appropriate and for those where it is not. To the
extent that it covers regulatory offences,
it might well be justifiable,
balancing out all the relevant concerns and interests, to run the risk of
convicting persons about
whose guilt a doubt remains. Yet it embraces common
law crimes as well. Furthermore, given that regulatory offences are creatures
of statute that usually contain their own specific tailor-made legislative aids
to securing effective
implementation,2[1] prosecutors
could well prefer to rely on such statutes rather than on section 332(5) when
preparing their indictments. The main
function of section 332(5) could, then,
be to help the prosecution get round hallowed procedural protections normally
available to
the accused in criminal matters.

[220] Much was made during
argument of the importance of combatting corporate fraud and other forms of
white collar crime. I doubt
that the prevalence and seriousness of corporate
fraud could itself serve as a factor which could justify reversing the onus of
proof.
There is a paradox at the heart of all criminal procedure, in that the
more serious the crime and the greater the public interest
in securing
convictions of the guilty, the more important do constitutional protections of
the accused become. The starting point
of any balancing enquiry where
constitutional rights are concerned must be that the public interest in ensuring
that innocent people
are not convicted and subjected to ignominy and heavy
sentences, massively outweighs the public interest in ensuring that a particular
criminal is brought to book.2[2]
Hence the presumption of innocence, which serves not only to protect a
particular individual on trial, but to maintain public confidence
in the
enduring integrity and security of the legal system. Reference to the
prevalence and severity of a certain crime therefore
does not add anything new
or special to the balancing exercise. The perniciousness of the offence is one
of the givens, against
which the presumption of innocence is pitted from the
beginning, not a new element to be put into the scales as part of a
justificatory
balancing exercise. If this were not so, the ubiquity and
ugliness argument could be used in relation to murder, rape, car-jacking,
housebreaking, drug-smuggling, corruption . . . the list is unfortunately almost
endless, and nothing would be left of the presumption
of innocence, save,
perhaps, for its relic status as a doughty defender of rights in the most
trivial of cases.

[221] The second question that needs to be asked is
whether section 332(5) is capable of being read in such a way as merely to
penalise
what Mr Gauntlett referred to as reckless or inattentive stewardship of
corporations. I am in full agreement with O’Regan
J’s reasoning on
this point, as far as it goes. It should be remembered that interpretations of
the section by South African
courts in the pre-constitutionalism era were
directed simply to determining the “intent” of the legislature, and
then
to ensuring that indictments gave the accused adequate warning of the
precise charges they would have to meet. The courts then neither
had the
opportunity nor the obligation to opt for a reading which, although at first
sight was not the most convincing one, nevertheless,
was a reasonable one. The
fact that the interpretation proposed by so distinguished a jurist as Schreiner
JA in the Limbada2[3] case
was a minority one not supported in a later case by the Appellate Division, does
not make it an unreasonable one. If the approach
adopted is a reasonable one
which could lead to preserving the constitutionality of the provision, we have
no option: we are obliged
to adopt
it.2[4] Section 35(3) does not
give us an option. His approach goes a long way to establishing that the
section was intended to create
a new offence, or, rather, a new form of
liability based on failure of a director to prevent corporate crime when in a
position to
do so.

[222] It seems to me that the ambit of the section
was enlarged from serving as a means of ensuring, firstly, that corporation
directors
felt the sting when their companies broke the law, and secondly, that
they took appropriate measures to prevent possible offences
by those under their
command, to include a third element, namely, to see to it that prosecutors could
more easily get round the difficulties
of proof in relation to direct
responsibility for the principal offence. In my view, if a combination of
reading down and severance
rescues the achievement of the first two legitimate
legislative purposes, while eliminating the illegitimate third one, we should
adopt it. It would, of course, be absurd to require the prosecution to prove
beyond reasonable doubt that the accused was innocent
of the principal offence,
hence the words “it is proved that he did not take part in the
offence” should be deleted.
That would leave as the gravamen of the
offence the failure of the director to prevent the criminal activity. On the
assumption
that you can only prevent that of which you are
aware,2[5] this would not be a
very powerful weapon in the hands of the prosecution. Nevertheless, one can
envisage situations where there
would be sufficient evidence at least to call
for a reply, in terms of which directors could be compelled to testify on pain
of a
prima facie case against them being converted into a conclusive one by
their failure to answer the case made out against them. In
this respect, I am
persuaded that the legislature would have preferred half a loaf to no bread at
all, or, even, for that matter,
half of half a loaf, if preserving such a
portion kept alive one of its principal objects.

[223] As to the question
of onus of proof in relation to the “unless” clause relating to
failure to prevent the offence,
I support the approach of looking at the
substance of the offence, the mischief it was designed to combat, and the nature
of the
qualification, rather than relying purely on the grammatical forms
employed. In Kula’s
case,2[6] van den Heever JA had
the following to say on the subject:

“To postulate “the offence” as something divorced from its
actual definition is an unreal proceeding. If an area
is defined in terms of
what would otherwise have been a circle but by deduction of a certain segment,
it seems to me notionally impossible
to refer to the segment as either an
additional element of or an exception to the area defined.



The difficulty is that the Legislature has, in words which are deceptively
simple at first blush, provided extremely elastic criteria
which are purely
relative. In most statutory provisions creating offences it will be necessary -
save perhaps in regard to the simplest
matters - to describe and define the
scope of the prohibition by excepting, exempting, excluding, excusing or
qualifying the persons
of incidence or the corpus delicti, the facts.
Enlarge the segment to which I have referred and it becomes difficult to say
which is the notional circle, used as a
term of reference and which the
segment.



It cannot, surely, be suggested that the matter is governed by the set phrases
employed by the Legislature such as “unless”
and “provided
that”.”2[7]



[224] In
the present case, I would accept that the focus of the offence lies on the
segment rather than the circle from which it may
be detached. The culpability
of the director derives not from his/her position in the company, but from
his/her failure, once in
that position, to prevent a crime committed by those
under his/her control. This failure is more than a material element of the
crime, it is its essence. Even if that were to overstate the case, the
provision is at least reasonably capable of being read that
way. To construe it
as if the offence were merely to be a director of a company which has become
liable for prosecution, with an
escape route open to a director to prove
innocence, would raise problems both under section 11(1) and section 25(3)(c).
I agree
with O’Regan J’s illuminating analysis of the relation
between the two sections, and with the interpretation she arrives
at as a result
of applying the techniques of reading down, which is constitutionally incumbent
on us, and severance, its siamese
twin.

[225] If the matter had rested
there, I would have concurred fully in the judgment of O’Regan J, since I
regard my judgment
as being compatible with and intended to reinforce hers in
terms of overall approach, both in its substantive analysis and in its
treatment
of severance, the two being related. Unfortunately, I discern other problems
which compel me reluctantly to the conclusion
that the section is
irretrievable.

[226] The main one is the overbreadth of the term
“any corporate body”, to which Didcott J
refers.2[8] A supplementary one
is the complicated problems in terms of section 11(1) that could arise in
relation to the word “servant”:
should it be stitched back into the
legislative cloth again, or not? If I may be forgiven for switching my
metaphors from bread
to clothing, I do not regard snipping here and knitting
there as the exclusive function of the legislature. I believe the Constitution
requires us to be creative in saving the garment, or at least, a wearable part
of it, if we can do so in a manner consistent with
the purpose of the
legislature as expressed in the text. But, too much reading down of too many
terms, coupled with too many excisions
of the text, leaves something so tattered
and insecure, that it cannot be said that effect would be given to any of the
principal
objects of the legislature. In Coetzee v Government of the
Republic of South Africa; Matiso and Others v Commanding Officer, Port Elizabeth
Prison, and Others
2[9] case,
Kriegler J expressly left open the possibility that “severability in the
context of constitutional law may often require
special
treatment”.3[0] In my
judgment in that matter I said that

“[s]everability is an important concept in the context of the relations
between this Court and Parliament; like “reading
down”, it is an
instrument of judicial restraint which reduces the danger of producing an
overbroad judicial reaction to overbroad
legislation. ... [W]e must take account
of the coming into force of the new Constitution in terms of which we receive
our jurisdiction
...”3[1]


We
have not heard full argument on the matter, but my provisional view is that if
legislation mixes constitutionally legitimate with
constitutionally illegitimate
objectives, then, provided that the latter are not so fundamental and pervasive
as to vitiate the whole
legislative scheme, we should seek to preserve the part
that is constitutional. This would be consistent with the restrained
adjudicatory
posture enjoined on us by section 35(3) of the
Constitution.3[2] Yet, I feel
that even applying the test most favourable to severance would not save the
provision from being struck down in its
entirety.

[227] I come to this
conclusion with regret since I find myself in broad agreement not only with the
analytical thrust of O’Regan
J’s judgment, but with the overall
evaluation of the issues at stake so eloquently presented by Kentridge AJ.
3[3] I also wish to associate
myself with the forceful comments of Madala
J3[4] and Mokgoro
J,3[5] concerning the dangers
represented by white collar crime and corporate disregard for the public
welfare.

[228] This leaves me only to say that I agree with the
well-articulated analyses of Langa J with regard to sections 245 and 332(5)
respectively. I demur only to the extent that in my view his analysis of
section 332(5) stops short of dealing with problems I feel
this Court is obliged
to confront, hence my separate judgment. With these rather elaborate
qualifications, I concur in the orders
he proposes.

For the appellants: WH Trengove SC and GJ Marcus, instructed by Kok &
Hendrikse.

For the State: AJ Fourie and CTH McKelvey, instructed by the
Attorney-General of the Witwatersrand Local Division of the Supreme
Court.

For the Government of JJ Gauntlett SC and AM Breitenbach,

the Republic of South Africa: instructed by the State Attorney, KwaZulu-
Natal.


[1]R v Oliver 1959 (4) SA 145
(D) at 145H; S v Isaacs 1968 (2) SA 187 (D) at 191F; S v Van
Niekerk
1981 (3) SA 787 (T) at 789H-790A.

[2] See S v Bhulwana; S v
Gwadiso
1996 (1) SA 388 (CC); 1995 (12) BCLR 1579 (CC); S v Mbatha; S v
Prinsloo
1996 (2) SA 464 (CC); 1996 (3) BCLR 293 (CC); Scagell and Other
v Attorney-General of the Western Cape and Others
1996 (11) BCLR 1446
(CC).

[3]S v Zuma and Others 1995
(2) SA 642 (CC); 1995 (4) BCLR 401 (SA) at para 25 citing Woolmington v
Director of Public Prosecutions
[1935] AC 462 (HL) at 481; see also R v
Benjamin
1883 EDC 337 at 338; R v Ndhlovu 1945 AD 369 at 386.

[4] Section 33(1)(a) and 33(1)(aa) of
the Constitution.

[5] Section 33(1)(b) of the
Constitution.

[6]S v Makwanyane and Another
1995 (3) SA 391 (CC); 1995 (6) BCLR 665 (CC) at para 104. See also
S v Williams and Others 1995(3)SA 632 (CC); 1995(7) BCLR 861 (CC) at
paras 58-60.

[7] Section 6 of the Criminal Law
Further Amendment Act 75 of 1959.

[8] Courts have drawn a distinction
between instances in which the misrepresentation is one pertaining to a belief
on the one hand and
where it relates to an objective fact on the other. See S
v Andrews
1982 (2) SA 269 (NC) at 271F-G. The presumption is of no
assistance to the prosecution in the former instance. The prosecution has
to
prove affirmatively that the accused did not have the belief represented at the
time it was made. See S v Witbooi 1971 (4) SA 138 (NC) at 140E; S v
Hassim and Another
1976 (1) SA 508 (T) at 512A; S v Ostilly and Others
1977 (4) SA 699 (D) at 724B; S v Harper and Another 1981 (2) SA 638
(D) at 648H - 649B; S v Van Niekerk supra n 1 at 790A and F.

[9] Supra n 3 at para 41.

[10] In Bhulwana supra n 2
at para 24 ORegan J described the rights enshrined in section
25(3)(c) of the Constitution as a pillar of our system
of criminal
justice. See also Zuma supra n 3 at para 36; Mbatha supra
n 2 at para 19. The presumption of innocence is acknowledged as a
fundamental value in the criminal justice system of comparable democracies. See
the remarks of Dickson CJC in R v Holmes (1989) 34 CRR 193 at 212
and in R v Oakes (1986) 26 DLR (4th) 200 at 212-3.

[11] In Mbatha supra n 2 at
para 20, in a discussion on the difficulty of proving the mental aspect involved
on a charge of possession, this Court
stated: [T]he circumstances of
each case will determine whether or not the elements of possession have been
established.
. . . The evidence need not necessarily be direct. It may be, and
often is, circumstantial. . . . There will no doubt be cases
in which it will be
difficult to prove that a particular person . . . was in fact in possession. .
. . But this is inevitably a
consequence of the presumption of innocence; this
must be weighed against the danger that innocent people may be convicted if the
presumption were to apply. In that process the rights of innocent persons must
be given precedence.

[12] For the position in the United
States of America, see La Fave and Scott, Jr Criminal Law 2 ed (1989) 739
et seq; People v Ashley 42 Cal. 2d. 246 (1954) and Nelson v United
States
227 F.2d 21(1955); in Canada, see R v Theroux (1993)
100 DLR (4th) 624 at 635-8; R v Zlatic (1993) 100 DLR (4th) 642; in
England, see R v Landy [1981] 1 All ER 1172 (CA) and R v Ghosh
[1982] 1 QB 1053 (CA). With regard to Australia see Balcombe v De Simoni
[1971-1972] 126 CLR 576 at 593.

[13] Section 348(7) of Act 31 of
1917; 381(7) of Act 56 of 1955; 332(7) of Act 51 of 1977.

[14] See para 9 of this
judgment.

[15] 1955 (2) SA 338 (A) at
341A-B.

[16] 1975 (4) SA 773 (A) at 780A-D.


[17] 1958 (2) SA 481 (A) at 484G-H.


[18] Id at 486B.

[19] Id. The matter was concerned
with the nature of the liability imposed by section 381(7) of Act 56 of 1955, a
provision which is
the predecessor of the current section 332(7) of the Act.
The wording corresponds materially with that of section 332(5) of the
Act, save
that the former deals with the liability of members of associations and the
latter targets directors and servants of a
corporate body.

[20] Supra n 15 at 341A-B.

[21] Supra n 16 at 780B-D.

[22] See S v Film Fun Holdings
(Pty) Ltd and Others
1977 (2) SA 377 (E) at 386E-H and S v Harper and
Another
supra n 8 at 641F-G.

[23] See para 19 above.

[24] Supra n 17 at 486A-C.

[25] Id at 487A.

[26] Id at 484F-H.

[27] Supra n 16 at 779A-C.

[28] Supra n 17 at 484-5A.

[29] See Zuma supra n 3;
Bhulwana
and Mbatha supra n 2. Kentridge AJ, speaking for the Court
in Zuma at paras 41 and 42 stated: It is important . . . to
emphasise what this judgment does not decide. It does not decide that
all statutory provisions which create presumptions in criminal cases are
invalid. This Court recognises
the pressing social need for the effective
prosecution of crime, and that in some cases the prosecution may require
reasonable presumptions
to assist it in this task. Presumptions are of
different types. . . . Nor does it seek to invalidate every legal presumption
reversing
the onus of proof. Some may be justifiable as being rational in
themselves, requiring an accused person to prove only facts to which
he or she
has easy access, and which it would be unreasonable to expect the prosecution to
disprove . . . I would also make clear
that this judgment does not purport to
apply to exceptions, exemptions or provisos to statutory offences, referred to
in section
90 of the Criminal Procedure Act . . . .



[30] Kentridge AJ in Zuma at
para 21, described the judgments of Canadian courts on reverse onus provisions
as particularly helpful, not only because
of their persuasive reasoning,
but because s 1 of the Charter has a limitation clause analogous to section 33
of the South African
Constitution.

[31] Supra n 10 at 201. See also
the majority decision in R v Schwartz (1989) 39 CRR 260 at 268-9 and the
reasoning of Cory J (LHeureux-Dube J concurring) in R v Wholesale
Travel Group Inc.
(1992) 84 DLR (4th) 161 at 223-7.

[32] Id at 198.

[33] Id at 201.

[34] Supra n 31.

[35] Section 106(7)(1) of the
Canadian Criminal Code.

[36] (1989) 51 DLR (4th) 481.

[37] Section 234 read with section
237(1)(a) of the Canadian Criminal Code.

[38] Supra n 10 per Dickson CJC at
222.

[39] Id at 493.

[40] (1991) 3 CRR (2d) 193 at 205.


[41] Id at 258.

[42] See for example R v Chaulk
(1991) 1 CRR (2d) 1 in which the issue was a provision in the Criminal Code
which set out a presumption of sanity until the
contrary is
proved. Lamer CJC at 14, writing for the majority, concluded that the
insanity defence should be characterized
as an exemption to criminal
liability which is based on an incapacity for criminal intent. Adopting
the reasoning
in Whyte, he rejected the prosecutions argument
that because sanity was not an essential element of the offence, the presumption
of
innocence was not implicated. In the course of her separate concurring
judgment, Wilson J had occasion to consider the treatment
of the presumption of
innocence in Holmes and Schwartz and distinguished the latter on
the basis that it deals with regulated not prohibited
activity (at 49). She was however clearly of the view that the
approach followed in Whyte was the appropriate one in the particular
instance.

[43] Supra n 31.

[44] 342 US 246, 254-6 (1952).

[45] Supra n 31at 198.

[46] Supra n 44 at 275.

[47] 432 US 197, 223 (1977).

[48] Section 33(1)(AA) and (b) of
the Constitution.

[49]Bernstein and Others v
Bester and Others NNO
1996 (2) SA 751 (CC); 1996 (4) BCLR 449 (CC) at para
85; Ferreira v Levin NO and Others; Vryenhoek and Others v
Powell NO and Others
1996 (1) SA 984 (CC); 1996 (1) BCLR 1 (CC) at para
151.

[50] See para 30 of this
judgment.

[51] 1995(4) SA 631 (CC); 1995 (10)
BCLR 1382 (CC) at para 16.

[52] Per Centlivres CJ in
Johannesburg City Council v Chesterfield House (Pty) Ltd
1952 (3) SA 809 (A)
at 822C-E.

[53]R
v Shangase
1960 (1) SA 734 (A).

[1] See eg, R v Zondagh 1931
AD 8; R v Dekker 1931 AD 17; R v Beebee 1944 AD 333; R v
Kula
& Others 1954 (1) SA 157 (A); R v Limbada and Another
1958 (2) SA 481 (A); S v Klopper 1975 (4) 773 (A); Attorney
General, Cape v Bestall
1988 (3) SA 555 (A); S v Blaauw 1989 (1) SA
202 (A).

[2] Id.

[1] 1960
(1) SA 734 (A).

[2] at 735D-E.

[3] 1952(3) SA 809 (A) at 822C-F.

[4] 1995 (4) SA 631(CC); 1995(10) BCLR 1382
(CC).

[5] para 16: at 645B in the first report
cited and 1392H in the second one.

[1] 1958 (2) SA 481 (A).

[2] Section 381(7) of Act 56 of
1955.

[3] At 486C. See further R v
Milne and Erleigh
(7) 1951 (1) SA 791 (A) at 831.

[4] 1975 (4) SA 773 (A) at
779A-C.

[5] Supra n 1.

[6] 1960 (4) SA 364 (O) at 371D.

[7] De Wet JC De Wet en Swanepoel
Strafreg
4 ed (Butterworths, Durban 1985) at 61 n 83.

[8] See Burchell J & Milton J
Principles of Criminal Law, (Juta, Cape Town 1991) 322. There appear to
have been one or two minor exceptions to this general rule in the Anglo-American
common
law. See Halsbury's Laws of England, 4 ed, Volume 11(1) paras 18
and 54; Morissette v United States 342 US 246, 251 (1952).

[9] South African examples are given
in Burchell & Hunt South African Criminal Law and Procedure 2 ed
(Juta, Cape Town 1983) Volume 1, chapter 10.

1[0] See sections 165 and 166 of the
Liquor Act 27 of 1989.

[1]1 See section 7 of the Petroleum
Products Act 120 of 1977.

1[2] 1944 AD 333.

1[3] 1954 (1) SA 157 (A).

1[4] Supra n 12 at 336.

1[5] Supra n 13 at 161E-G.

1[6] As to the possible operation of
section 90 of the Act, see paras 108 and 109 infra.

1[7] (1989) 51 DLR (4th) 481 at
493.

1[8] 1995 (2) SA 642 (CC); 1995 (4)
BCLR 401 (SA) at para 23.

1[9] See for example R v
Keegstra
(1991) 3 CRR (2d) 193.

2[0] Supra n 18 at para 21. See para
109 infra.

2[1] Id at para 42.

[2]2 (1990) 64 DLR (4th) 577 at
583-4.

2[3] (1992) 84 DLR (4th) 161.

2[4] Id at 198. See also the
judgments of Cory J and of Iacobucci J in the same case at 210-11 and 233-4,
respectively.

2[5] But found that it was
justifiable under s 1 of the Charter.

2[6] (1988) 50 DLR (4th) 680 at
684.

2[7] Id.

2[8] Supra n 17.

2[9] Supra n 19.

3[0] (1992) 9 CRR (2d) 1.

3[1] See R v Van den Berg and
Another
1955 (2) SA 338 (A); S v Klopper supra n 4; Kriegler
Hiemstra Suid-Afrikaanse Strafproses 5 ed (Butterworths, Durban 1993)
877.

3[2] No argument was developed in
relation to section 28 and I propose to say no more about it.

[3]3 Blackstone Commentaries
IV.21.

3[4] Supra n 8 at 250.

3[5] [1970] AC 132 (HL) at 156E.

3[6] 1966 (4) SA 356 (A) at
364D.

3[7] Supra n 8 Burchell et al at
323-4.

3[8] (1985) 24 DLR (4th) 536 at
541.

3[9] Ie liability arising simply
from the doing of a prohibited act, without proof of mens rea, and with no
excuse available to the accused.

4[0] Supra n 23 at 179.

4[1] (1978) 85 DLR (3d) 161.

4[2] 320 US 277 (1943).

4[3] Id at 280-1.

[4]4 Supra n 8 at 256.

4[5] I should point out that the
Court there went on to say:

Also, penalties commonly are relatively small, and conviction does no
grave damage to an offender's reputation. Under such
considerations, courts
have turned to construing statutes and regulations which make no mention of
intent as dispending with it and
holding that the guilty act alone makes out the
crime.

4[6] Supra n 23 at 213d-g.

4[7] Supra n 41.

4[8] Id at 179.

4[9] 1996 (2) SA 751(CC); 1996 (4)
BCLR 449 (CC).

5[0] See also the judgment of
Ackermann J in Ferreira v Levin NO and Others; Vryenhoek and
Others v Powell NO and Others
1996 (1) SA 984 (CC); 1996 (1) BCLR 1 (CC) at
para 151.

5[1] (1994) 18 EHRR 393 at para
69.

5[2] 1965 (2) SA 616 (A) at
625.

5[3] Supra n 49 at para 151.

5[4] (1990) 76 DLR (4th) 545 at
582e-f.

[5]5 [1951] 2 KB 135 at 149.

5[6] See in this regard also
Lim Chin Aik v The Queen [1963] AC 160 (PC) at 174.

5[7] 1959 (4) SA 607 (N).

5[8] 1965 (3) SA 834 (W).

5[9] 1979 (1) SA 208 (D).

6[0]Attorney- General of Hong
Kong v Lee Kwong-kut, Attorney General of Hong Kong v Lo Chak-man and
another
[1993] AC 951 (PC) at 975.

6[1] See Morrison v Standard
Building Society
1932 AD 229.

6[2] In the applicants
written submissions, at para 31.3, the point is made that section 332(5) is
discriminatory because it does
not apply to those who carry on business on their
own account. This point was, with good sense, not developed or pursued.

6[3] Supra n 41. See also the
judgment of Cory J in R v Wholesale Travel Group Inc. supra n 23 at
224.

6[4] See for example R v
Keegstra
supra n 19; R v Chaulk (1991) 1 CRR (2d) 1; R v
Downey
supra n 30; R v Wholesale Travel Group Inc. supra n 23.

6[5] 1996 (1) SA 388 (CC); 1995
(12) BCLR 1579 (CC) at para 18.

[6]6 See R v Hewertson and
Others
(1) 1937 CPD 5.

6[7] Supra n 23.

6[8] Id at 195.

6[9] Id at 232.

7[0] Supra n 26.

7[1] (Ont CA) Action No 314/87
(1988); Quicklink [1988] O.J. No. 1258. In R v Laba (1995) 120 DLR (4th)
175 at 212 the Canadian Supreme Court went further: not only did it strike words
out of a reverse onus provision,
but read in words which clearly provided an
evidentiary onus in place of a legal one. Its justification was that it
furthered the
legislative objective while fully vindicating Charter
values. Moreover, this was no more an intrusion into the legislative
sphere than a complete striking down. Per Sopinka J at 210-214. We have not as
yet been called on to consider whether this Court
has powers to read
in as well as to read down.

7[2] Supra n 62.

7[3] Id at 974C-E.

7[4] See Hoffmann and Zeffert
The South African Law of Evidence 4 ed (Butterworths, Durban 1988)
517.

7[5]R v Shangase 1960 (1)
SA 734 (A). And see R v Hunt [1987] AC 352 (HL).

7[6] See Nimmo v Alexander
Cowan & Sons Ltd
[1968] AC 107 (HL) at 134-136.

[7]7 See Halsburys Laws
of England
4 ed Vol 11, para 358; R v Edward [1975] QB 27 at 39-40,
R v Hunt supra n 75 at 371-373. In the last-mentioned case the House of
Lords, like the Appellate Division in R v Shangase, supra n 75, held that
where a statute requires proof by an accused, a persuasive and
not merely an evidential burden
is imposed. See per Lord Griffiths at 376, per
Lord Ackner at 385.

7[8] See R v Beebee supra n
12; R v Kula supra n 13 and R v Zondagh 1931 AD 8.

7[9] Supra n 17.

8[0] Supra n 64.

[1] 1995 (4) SA 631 (CC); 1995
(10) BCLR 1382 (CC) at para 16, which states:

Although severability in the context of constitutional law may often
require special treatment, in the present case the trite
test can properly be
applied: if the good is not dependent on the bad and can be separated from it,
one gives effect to the good
that remains after the separation if it still gives
effect to the main objective of the statute. The test has two parts: first,
is
it possible to sever the invalid provisions and second, if so, is what remains
giving effect to the purpose of the legislative
scheme?(footnote
omitted).

[2]S v Zuma and Others
1995 (2) SA 642 (CC); 1995 (4) BCLR 401 (SA) at para 19.

[3]S v Bhulwana; S v
Gwadiso
1996 (1) SA 388 (CC); 1995 (12) BCLR 1579 (CC) at para 16-17 and
S v Mbatha; S v Prinsloo 1996 (2) SA 464 (CC); 1996 (3) BCLR 293 (CC) at
para 12-14.

[4] 1995 (3) SA 391 (CC); 1995 (6)
BCLR 665 (CC) at para 104.

[5] (1986) 19 CRR 308 at 322.

[6] (1989) 34 CRR 193 at 211.

[7]Attorney-General of Hong
Kong v Lee Kwong-kut, Attorney-General of Hong Kong v Lo Chak-man and
another
[1993] AC 951(PC) at 972B

[8]Ferreira v Levin NO and
Others; Vryenhoek and Others v Powell NO and Others
1996 (1) SA 984 (CC);
1996 (1) BCLR (CC) 1 at para 133.

[9]Bernstein and Others v
Bester and Others NNO
1996 (2) SA 751 (CC); 1996 (4) BCLR 449 (CC) at para
50.

1[0] (1965) 114 CLR 63 at 80.

[1]1 As quoted with approval by
Mason J in Hamilton v Oades (1988-1989) 85 ALR 1 (HC) at 5.

1[2] Supra n 9.

1[3] [1957] 1 QB 159 (CA) at
172.

1[4] 1991 (2) SA 660 (A) at
678B-D.

[1] Section 332(5) includes
servants as well as directors of corporate bodies. The Government conceded,
however, that such inclusion
of servants is unconstitutional.

[2] See S v Mbatha; S v
Prinsloo
1996 (2) SA 464 (CC); 1996 (3) BCLR 293 (CC) at para 9; S v
Bhulwana; S v Gwadiso
1996 (1) SA 388 (CC); 1995 (12) BCLR 1579 (CC) at para
15; S v Zuma and Others 1995 (2) SA 642 (CC); 1995 (4) BCLR 401 (SA) at
para 33.

[3] 1995 (4) SA 631 (CC); 1995
(10) BCLR 1382 (CC) at para 16.

[4] (1992) 93 DLR (4th) 1 at
12f.

[5] Id at 14f.

[6] See, eg, R v Shangase,
1960 (1) SA 734 (A); R v Kula and Others 1954 (1) SA 157 (A).

[7] Section 35(2) of the
Constitution provides:

No law which limits any of the rights entrenched in this Chapter, shall
be constitutionally invalid solely by reason of the
fact that the wording used
prima facie exceeds the limits imposed in this Chapter, provided such a
law is reasonably capable of a more restricted interpretation which does
not
exceed such limits, in which event such law shall be construed as having a
meaning in accordance with the said more restricted
interpretation.

[8] Supra n 6 at 161A.

[9] (1995) 120 DLR (4th) 175.

1[0] Id at 210a.

[1]1 Id at 214c-h. It is not
necessary to decide whether this Court also has powers to read
in, as done in Laba, as the desired result can be reached in this
case by a combination of severance and reading down.

1[2]Attorney-General of Hong
Kong v Lee Kwong-kut, Attorney-General of Hong Kong v Lo Chak-man and
another
[1993] AC 951 (PC).

1[3] Id at 974c-e.

1[4] Supra para 70.

1[5] Applicants
Supplementary Submissions at 17.

[1] Du Toit et al Commentary on
the Criminal Procedure Act
(Juta, Cape Town 1996) chapter 33-5.

[2] Section 384(1) of Act 31 of
1917.

[3]Alper and Alper v Rex
1931 NPD 429, per Lansdown J at 445.

[4]United States v Wise
370 US 405, 409 (1962).

[5]S v Mbatha; S v
Prinsloo
1996 (2) SA 464 (CC); 1996 (3) BCLR 293 (CC) at para 24.

[6]United States v Park
421 US 658, 671 (1975) per Burger CJ for the court:

It was enough [for liability] that, by virtue of the relationship he
bore to the corporation, the agent had the power to
prevent the act complained
of . . . [Thus] in order to make distributors of food the strictest
censors of their merchandise,
the Act punishes neglect where the
law requires care, or inaction where it imposes a
duty.

[7] It appears that in the USA,
statutes imposing criminal liability upon the innocent employer for the illegal
conduct of employees
are generally upheld as constitutional, although an
occasional case has said that such a statute would be unconstitutional.
However,
these latter cases appear to be based upon the proposition that
vicarious liability may not be imposed for the conduct of other persons
when the
defendant has no control over these persons, which is regarded as an appropriate
limitation. LaFave and Scott Handbook on Criminal Law (West
Publishing, St Paul Minnesota 1972) 227.

[8] Following the precepts of
section 35(2) of the Constitution.

[9] Burchell and Milton
Principles of Criminal Law (Juta, Cape Town 1991) 328.

1[0] Id.

[1]1 Id; see also S v
Sutherland
1972 (3) SA 385 (N) at 387C-G.

1[2] In S v Klopper 1975
(4) SA 773 (A) Kotze JA pointed out that in the matter before him, the appellant
had only heard about the misrepresentation
contained in a circular some months
after they had been issued. He cited van den Heever J in R v Couvaras
1946 OPD 392 at 397, in support of the proposition that you cannot prevent
something about which you know nothing:

Hoe hy dit kon verhinder gaan my verstand te bowe, want selfs die Gode
kan gedane sake nie ongedaan maak nie. at
780H

In R v Van den Berg and another 1955 (2) SA 338 (A) at 341B-C Greenberg
JA stated:

... ignorance caused through deliberately abstaining from making
inquiries may not avail a member of an association but this
position does not
arise here.

1[3] See R v Bennett and Co
(Pty) Ltd & Another
1941 TPD 194 at 199- 200; S v Joseph Mtshumayeli
(Pvt) Ltd
1971 (1) SA 33 (RA) at 35D-E.

1[4] See Meskin Henochsberg on
the Companies Act
vol II 5 ed (Butterworths, Durban 1994) Appendix I 8 who
gives the following examples:

It can be convicted of fraud (see eg R v Frankfort Motors
(Pty) Ltd
1946 OPD 255 at 266-267; R v Wege 1959 (3) SA 268 (C)) or
of theft (see eg R v Markins Motors (Pty) Ltd 1959 (3) SA 508 (AD)) or of
bribery (see eg S v Deal Enterprises (Pty) Ltd 1978 (3) SA 302 (W)); or
of culpable homicide (see eg R v Bennett & Co (Pty) Ltd 1941 TPD 194;
S v Joseph Mtshumayeli (Pvt) Ltd 1971 (1) SA 33 (RAD)) or of a statutory
offence (unless such cannot be committed by a corporate body) (see eg S v
Longdistance (Natal) (Pty) Ltd
1990 (2) SA 227 (AD); Ex parte Minister
van Justisie: In re S v Suid-Afrikaanse Uitsaaikorporasie
1992 (4) SA 804
(AD)).

In R v Bennett & Co (Pty) Ltd 1941 TPD 194 at 199 Murray J stated
that it was even suggested in this case that one of the startling
results of the clear
language of the section is that if a servant,
however humble, of a corporation in an endeavour to further the interests of
his
employer decided to murder a business competitor of such employer, and carried
such decision into effect, the corporation would
be guilty of
murder.

1[5] Burchell et al South
African Criminal Law and Procedure
Vol I 2 ed (Juta, Cape Town 1983)
405.

1[6] 342 US 246 (1952).

1[7] Id at 251-63 (citations
omitted).

1[8] This is the sentence relied
upon by Mr Gauntlett. It will be seen that the passages which follow it
undermine its use to support
the shifting of the onus of proof in the case of
prosecutions for fraud. On the contrary, they expressly point in the opposite
direction.

1[9] Per Kentridge AJ in S v
Zuma
1995 (2) SA 642 (CC); 1995 (4) BCLR 401 (SA) at para 36 where he
states:

Nor has it been shown that it is in practice impossible or unduly
burdensome for the State to discharge its onus. See also para
41.

2[0] Per Cory J in R v
Wholesale Travel Group Inc.
(1992) 84 DLR (4th) 161 at 213.

2[1] See eg, section 7 of the
Petroleum Products Act 120 of 1977; sections 165, 166 of the Liquor Act 27 of
1989. I make no observations
on the constitutionality of these provisions,
merely on their prosecutorial convenience.

[2]2 See Amalgamated Beverage
Industries Natal (Pty) Ltd v Durban City Council
1992 (3) SA (N) 562
at 567C-F.

2[3]R v Limbada and Another
1958 (2) SA 481 (A).

2[4] Schreiner JA wrote a separate
judgment precisely to underline his differing perspective on the section,
arguing that:

[A]lthough in form [the accused] is charged with what may fairly be
called the main offence, the substance of the matter
may well be, I think, that
he is being charged with a statutory offence under the sub-section, read with
the main offence. The reason
why I think that this may be the more acceptable
view is that the accused is not merely saddled with the burden of proving that
he
did not take part in the commission of the offence, but he must also prove
that he could not have prevented it.
A partner who proves that
he was not a socius criminis in the main offence is nevertheless
convicted if he cannot prove the second requirement. It seems somewhat
artificial to say that
he is guilty of the main offence, although it is
established that he was not a party to it. (At 484G-H).

In my view, the problem arose because the section had diverse objectives, hence
the diverse opinions.

2[5] I am not totally convinced of
the correctness of this proposition. In the days when, in order to prevent
manifest injustices, courts
went in for the limited equivalent of reading down
available to them by artful use of the canons of construction, interpretations
were given which might not be necessary in the present context of constitutional
control. It might well be, however, that for purposes
of legal certainty we
should not rush to revisit settled interpretations of this kind.

2[6]R v Kula and Others
1954 (1) SA 157 (A). This case dealt with the interpretation of the infamous
section 10 of the Native Urban Areas Act 25 of 1945.
Its context and outcome
indicate how different adjudication has become today even if aspects of legal
methodology survive.

2[7] Id at 162H-163B.

2[8] At para 75.

2[9] 1995 (4) SA 631 (CC); 1995
(10) BCLR 1382 (CC).

3[0] Id at para 16.

3[1] Id at para 75.

3[2] As Hogg points out,
Constitutional Law of Canada (Carswell, Ontario 1992) in Canada the
pith and substance approach to severability is not applicable to
Charter review,
where a presumption in favour of severance applies, rather than
the presumption of the existence of a single statutory scheme with
the parts
inextricably related (chapter 15.7). He describes severance as an important
tool of judicial restraint, and likens it
to reading-down as a technique to
mitigate the impact of judicial review (chapter 15.6). In principle, it should
make no difference
if the unconstitutionality can be cured by deleting words in
a provision or by excising a provision from a statute. Nor should it
matter if
the rescuing is done by severance, reading-down or a combination of both. At the
same time, we should not forget that our
armoury includes section 98(5) of the
Constitution which enables us to keep a provision alive pending rectification by
Parliament.
We would, accordingly, have to exercise a considered judgment in
each case as to which option was applicable.

[3]3 At para 97 of this judgment.
I find myself unable to concur with his conclusions, not because of the elements
he put in the scales,
but because of the different weights I think each should
possess.

3[4] At para 130.

3[5] At para 139.