UNDOING THE PAST THROUGH

STATUTORY INTERPRETATION:
THE CONSTITUTIONAL COURT AND
MARRIAGE LAWS OF APARTHEID
Wessel le Roux
BA BLC LLB LLD
Associate Professor
University of South Africa (UNISA)

SUMMARY
Recent legal scholarship has increasingly focussed on the way in which the law can
be employed as a strategy of memory in post-conflict societies. War crimes trials, civil
reparation or restitution claims, and the like, have all been extensively discussed as
memorial practices. However, the way in which ordinary process of statutory
interpretation can be employed to undo the past has thus far received little or now
attention. In this article I investigate how the South African Constitutional Court has
approached the interpretation of a number of apartheid marriage laws under sections
35(2) of the interim and 39(2) of the final Constitutions. I argue that, far from adopting
a purposive approach to statutory interpretation as is often claimed, the court has in
fact adopted a very narrow textual approach to statutory interpretation. I claim that
this approach to the legislative legacy of the past is both hermeneutically and
politically suspect. What is more, it discloses a predilection for monumental, as
opposed to, memorial practices of memory.

1 INTRODUCTION
For the greatest part of South Africa’s modern history our patriarchal
common law was supplemented by a series of statutes which were all
designed to impose a narrowly understood Christian Nationalist (some
claimed civilised) conception of marriage onto a diverse and multi-cultural
population.1 Many of these statutory measures were designed to financially
reward members of this population who willingly entered into officially
recognised marriages.2 The inequities and indignities which resulted from
this discriminatory use of the law scarcely need recounting.3

1
A marriage was accordingly only recognised if it was concluded in terms of the Marriage Act
25 of 1961 between two sexually potent (Joshua v Joshua 1961 1 SA 455 (GW)) and fertile
(Van Niekerk v Van Niekerk 1949 4 SA 123 (W)) persons of different sexes but of the same
race (s 1(1) of the Proh bition of Mixed Marriages Act 55 of 1949).
2
For the purpose of this discussion it is sufficient to refer to those laws which have recently
formed the subject of the constitutional litigation which will be discussed later. See for
example the Intestate Succession Act 81 of 1987, the Maintenance of Surviving Spouses
Act 27 of 1990, Aliens Control Act 96 of 1991, and the Judges’ Remuneration and

526
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 527

When the new constitutional order came into operation at midnight on 27

April 1994, the interim Constitution explicitly stated that all apartheid laws
(including the above-mentioned oppressive marriage laws) would remain in
force until such time as these laws were either repealed or declared
unconstitutional.4 However, it soon became clear that a comprehensive
revision of the apartheid statute book was not on the cards. It was effectively
left to the courts to gradually dismantle the apartheid statutory framework
through a process of constitutional and statutory interpretation.
My interest in this article centres on the way in which the Constitutional
Court has approached the process of statutory interpretation in this specific
post-apartheid context. As a means of getting a manageable grip on the
question, I will restrict my attention to the way in which the court has
approached the interpretation of a number of the apartheid marriage laws
mentioned above. In order to gain a critical perspective on this jurispru-
dence, the approach or approaches to statutory interpretation which can be
distilled from these cases will be compared to the purposive approach which
JMT Labuschagne had developed during the late 1970s and early 1980s in
anticipation of a future era of constitutionalism. The gist of my argument is
that the court has not yet assimilated the basics of this approach and that
this leaves its jurisprudence both hermeneutically and politically suspect.
The broader context of the discussion is the growing jurisprudential
interest in the role which the law can play as a strategy of memory in post-
traumatic societies. War crimes trials, reparation suits and quasi-judicial
commissions have all grabbed the attention of contemporary legal scholars.5
However, little attention has thus far been paid to the role which ordinary
processes of statutory interpretation can play in this regard. I hope that this
preliminary investigation of the way in which the South African Constitutional
Court has responded to the legacy of apartheid by interpreting its marriage
laws (or not), will broaden the limited scope of the present jurisprudential
interrogation of the relationship between law and the politics of memory.

2 STATUTORY INTERPRETATION AS A
JURISPRUDENCE OF RESTRAINT
It should be noted right at the outset that the South African courts were not
left at sea by the last apartheid legislature about the way in which apartheid
legislation should be interpreted in the future. The interim Constitution, which
was adopted by the last apartheid parliament in 1993,6 contained a specific

Conditions of Employment Act 88 of 1989, which all conferred pecuniary or other benefits on
legally registered married couples only.
3
Part of this history is recounted in more detail in the minority judgments in Volks v Robinson
2005 5 BCLR 446 (CC) par [107]-[122] and [163]-[174].
4
S 229 of the (interim) Constitution of the Republic of South Africa Act 200 of 1993.
5
See eg, Felman The Juridical Unconscious: Trials and Traumas in the Twentieth Century
(2002); Douglas The Memory of Judgement: Making Law and History in the Trials of the
Holocaust (2001); and Osiel Mass Atrocity, Collective Memory, and the Law (1997). The
German confrontation with the Nazi past after unification has proved to be one of the more
fertile areas of research. A comparison between the German post-unification and the South
African post-apartheid experiences remains to be written.
6
Constitution of the Republic of South Africa Act 200 of 1993.
528 OBITER 2005

injunction in section 35(2) to deal with the issue. The section reads as
follows:
“No law which limits any of the rights entrenched in [the Bill of Rights], shall be
constitutionally invalid solely by reason of the fact that the wording used prima
facie exceeds the limits imposed in [the Bill of Rights], provided such a law is
reasonably capable of a more restrictive interpretation which does not exceed
such limits, in which event such law shall be construed as7 having a meaning
in accordance with the said more restrictive interpretation.”

This does not mean that section 35(2) amounted to no more than an
unique and pragmatic arrangement typical of the South African constitutional
transition. The section was also sold to the new constitutional community on
the basis that it embodied a well-known and established interpretive
procedure. This procedure found wide-spread application in many
constitutional democracies, including those of Germany (where it was known
as the principle of verfassungskonforme Auslegung)8 and Canada (where it
was known as “reading-down”).9 Lourens du Plessis, who was partly
responsible for the inclusion of this section in the interim Constitution, has
gone to great lengths to unpack the jurisprudence behind this principle. His
understanding of this jurisprudence, seen through the eyes of German
constitutionalism, provides a useful starting point for our discussion.
The German principle of verfassungskonforme Auslegung (interpretation
conforming to the constitution) does not merely express a technical or
methodological concern. It is incidental to the broader constitutional principle
of judicial restraint.10 In fact, constitutionalism itself is understood as no more
than the embodiment of a jurisprudence of restraint. The principle of judicial
restraint can in turn not be divorced from the principle of subsidiarity in
general, and the principle of adjudicative subsidiarity in particular. The
principle of subsidiarity obliges a more encompassing superordinate body or
community to refrain from taking for its account matters that can be dealt
with by a subordinate body or community.11 As an adjudicative principle, it

7
Du Plessis and Corder Understanding South Africa’s Transitional Bill of Rights (1994) 121-
122 explain the reason for the inclusion of this section as follows: “The ultimate object of
these provisions, which curb judicial over-activism in str king down legislation, is to enhance
legal certainty during the transition. The multi-party negotiators agreed on the inclusion of
section 35(2) without any difficulty.”
8
Currie The Constitution of the Federal Republic of Germany (1994) 28-29 explains
verfassungskonforme Auslegung (interpretation conforming with the constitution) as a
pragmatic tool to reduce the friction inherent in constitutional review. The result is often the
same as if the statute had been struck down, but legislative feelings are spared by indulging
the presumption that the lawmakers are sensitive to their constitutional obligations. The
principle boils down to a presumption of statutory interpretation, namely, that the legislature
always seeks to meet its constitutional obligations.
9
Hogg Constitutional Law of Canada 3ed (1992) 809 formulates the principle as follows:
“Where the language of a statute will bear two interpretations, one of which would abridge a
Charter right, and one of which would not, the Charter can be applied simply by selecting the
interpretation that does not abridge the Charter right.” The process of reading-down is
subject to an actual or at least potential ambiguity in the language of the statute (906). As he
explicitly puts it: “It is only available where the language of the statute will bear the (valid)
limited meaning as well as the (invalid) extended meaning: it then stipulates that the limited
meaning be selected” (393). Hogg also regards the principle as embodying a “presumption
of constitutionality”, ie the enacting legislative body is presumed to have meant to enact
provisions which do not transgress the limits of its constitutional powers (394).
10
Du Plessis Re-interpretation of Statutes (2002) 30 141.
11
Du Plessis 29.
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 529

implies that legal disputes should be decided as far as possible by avoiding

the constitutionalisation of the issues involved.
Du Plessis suggests that this approach to constitutional interpretation also
makes good constitutional sense in the post-apartheid context. This is so
because of (i) the fact that the Constitution is not interpreted by the
Constitutional Court alone but by an open community of interpreters with a
plurality of perspectives and meanings;12 (ii) the weak institutional position of
the Constitutional Court within this community of interpreters;13 (iii) the
opportunities which this approach creates for a mediated dialogue between
the legislature and judiciary;14 and (iv) the fact that the Constitution is a
written text which is in need of constant interpretation and reinterpretation.15
The idea of statutory interpretation as a form of judicial restraint (that is,
as a means of avoiding constitutional confrontations with the legislature)
embodies the essence of what Du Plessis elsewhere called the memorial (as
opposed to the monumental) constitution.16 Monuments and memorials
share a concern with memory but differ significantly in the way in which they
remember. Monuments celebrate victories while memorials commemorate
the dead. Monuments institute new understandings and meanings of past
events. Memorials mark the limits of our ability to understand and give
meaning to past events and traumas. Monuments are bold and self-
confident. Memorials are reserved and tentative.17 Du Plessis argues that
the Constitution contains both monumental and memorial elements.
Where apartheid legislation has to be interpreted, Du Plessis’s distinction
assumes an added significance. One could say that the principle of restraint
as it is embodied in section 35(2) recognises the need to deal with the
apartheid past in a memorial and not a monumental fashion. The injustices
of the past are called to memory by incrementally and interpretively
reworking the apparently fixed, apparently natural, apparently legal
meanings which were authoritatively assigned to events, things, words and
persons in the past. The fact that the very same legislative instruments can
gradually but constantly be given new interpretive twists serves both to
remind us of the arbitrary and unjust nature of their application in the past,
and to warn us against repeating the same neutralisation, naturalisation and
legalisation of meaning in the present. It is not only the substance of
meaning but also the way in which legal meaning is produced that requires
our attention. This point is crucial to the rest of our discussion.
Section 35(2) requires that we produce new legal meanings out of the
material of the past in an interpretive fashion (by contrast, for example, to
the production of new meaning through remedial “reading-in”). We are called

12
Du Plessis 136. See also Du Plessis “Legal Academics and the Open Community of
Constitutional Interpreters” 1996 SAJHR 214.
13
Du Plessis 171.
14
Du Plessis 28.
15
Du Plessis 29.
16
Du Plessis “The South African Constitution as Memory and Promise” 2000 Stell LR 385.
17
The differences between these two practices or politics of memory cannot be explored here
in any more detail. For powerful yet tentative attempts to do so see Snyman “Interpretation
and the Politics of Memory” 1998 Acta Juridica 312; and Van Marle “Lives of Action,
Thinking and Revolt – A Feminist Call for Politics and Becoming in Post-apartheid South
Africa” 2004 SAPR/PL 605.
530 OBITER 2005

upon to respond interpretively to the injustice of the past, that means, not by
monumentally imposing new incontestable and closed meanings in the place
of the old, but by constantly reminding ourselves of the fluid, contestable,
interpretive and therefore political nature of all legal meaning. If Du Plessis’s
distinction is applied in this context, one could say that the principle of
constitutionally conforming interpretation forms part of the memorial
constitution, while the constitutional remedies (like reading-in) form part of
the monumental constitution.
The discussion of the jurisprudential background to section 35(2) will have
to be restricted to these brief comments. In the following section I begin to
explore how the courts have distilled a specific method of statutory
interpretation from the section. The question whether this method gives
proper expression to the jurisprudence of restraint, or allows for a proper
memorialisation of apartheid legislation, will be addressed a little later.

3 SECTION 35(2) OF THE INTERIM CONSTITUTION

AND THE BIRTH OF THE NEW TEXTUALISM
As soon as the jurisprudence of restraint is understood as an incidence of
the principle of subsidiarity, it becomes clear that it does not exclude a
potentially aggressive form of judicial activism. Restraint in this context
simply means keeping that activism to its own domain. It is therefore not
surprising that the question of limits has from the outset been part of the
jurisprudence which has come to surround section 35(2) of the interim Con-
stitution. How far can the interpreter go in assigning new meanings to old
legislation in order to prevent that legislation from being declared in-
operative, or to promote the spirit, purport and objects of the Bill of Rights in
terms of section 35(3)? Du Plessis himself grants a very limited degree of
judicial activism in this regard. The principle of verfassungskonforme Ausle-
gung can only be invoked “where more than one interpretation of a provision
is grammatically achievable”.18 This means that the alternative meaning
must not grammatically distort (let alone change or alter) the original wording
of the provision in question.19 Whatever the reasons for Du Plessis’s limited
understanding of the principle might be, it is this understanding which was
right from the outset adopted by the Constitutional Court.
The court held in S v Bhulwana; S v Gwadiso20 that the words “until the
contrary is proved” could not be read down to mean “unless the evidence
raises a reasonable doubt” because the “unambiguous language of the
phrase” left no room for such a reading.21 Whether or not this implied that
section 35(2) could only be invoked where the language of the legislation
was ambiguous, it made clear that the “reasonably capable” test of the
proviso in section 35(2) referred to the wording of the legislation (as opposed
to the scheme of the legislation as a whole). This is a questionable but
extremely important assumption. It gave the jurisprudence of restraint an

18
Du Plessis “The Jurisprudence of Interpretation and the Exigencies of a New Constitutional
Order in South Africa” in Bradfield and Van der Merwe (eds) Meaning in Legal Interpretation
(1998) 8 13.
19
Du Plessis 141.
20
1996 1 SA 388 (CC).
21
Par [28].
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 531

unnecessarily narrow textualist slant and effectively precluded the

development of an “authentic” purposive approach to statutory interpretation
under the interim Constitution.22 The development of such an approach
would have required that the reference to “law” in the proviso to section
35(2) be read as if it referred to the legislative scheme of the statute as a
whole, rather than to the words of a particular section in the statute. The
question would then have become whether the unconstitutionality of the
statute could be avoided by giving the statute a new purpose and not
whether the existing words of the statute could be given a new meaning.
This is, however, not what happened. The judgment of Marais JA in
Minister of Safety and Security v Molutsi23 underscores this fact. Marais JA
argued that section 35(2) limited the scope of statutory interpretation under
section 35(3) of the interim Constitution (the injunction that statutes should
be interpreted with due regard to the spirit, purport and object of the Bill of
Rights). It followed that a court was not empowered by the Constitution to
“assign to either a pre-constitution or post-constitution statute a meaning
which its language could not reasonably bear or which was in flat
contradiction to the ordinary and plain language used in the statute”.24 On
this basis, Marais JA held that the words “cause of action” could not be read
down to mean “proximate cause of action”. More telling is his remark that the
“intractable” language of the legislation “fortunately” saved him from “having
to consider what the true import” of the provision was.25 A sufficiently clear
text could still preclude an investigation into the purpose of the statute in
question. What would happen if the clear text was obviously in violation of
the Constitution is not clear.
When the final Constitution was adopted in 199626, it did not again contain
a provision similar to section 35(2) of the interim Constitution. Section 39(2)
of the final Constitution, however, included a slightly amended restatement
of section 35(3) of the interim Constitution. The new section demanded that
statutory interpreters “promote” the spirit, purport and objects of the Bill of
Right (as opposed to merely interpret statutes “with due regard to” that
spirit).
For a brief moment of time there was uncertainty about the implication of
these changes. In De Lange v Smuts27 Ackerman J held that the fact that the
final Constitution did not include a section similar to section 35(2) was not of
any true significance as the old section merely expressed a sound principle
of constitutional interpretation which was recognised in other open and
democratic societies whose constitutions, like the South African constitution,
did not contain an express provision which incorporated this principle.28 The
same interpretive approach should be adopted under the 1996 Constitution
22
The term is used by Devenish Interpretation of Statutes (1992) 36 to distinguish between a
“qualified purposive approach” where a purposive reading of legislation is only employed to
resolve cases of textual ambiguity, and an “authentic purposive approach” where legislation
is interpreted according to its ratio in all cases, even if it requires the alteration of the
wording of the text.
23
1996 4 SA 72 (A).
24
85D-F (my emphasis).
25
85J.
26
The Constitution of the Republic of South Africa, 1996.
27
1998 3 SA 785 (CC).
28
Par [85].
532 OBITER 2005

in terms of section 39(2) as that under the interim Constitution in terms of

section 35(2).29 What Ackerman J failed to explain is why the sound principle
that statutory interpretation should avoid unconstitutionality should continue
to be understood in a narrow textualist sense (as the proviso to section 35(2)
had seemingly required). In a minority judgment, Mokgoro J expressed some
well-founded reservations about simply equating the two interpretive
regimes. She described section 35(2) as particularly “benevolent” or “benign”
to the legislative text.30 Section 39(2), on the other hand, was less
permissive and not subject to the limitations which section 35(2) had
imposed on the old section 35(3). In her mind the new section 39(2) thus
allowed for a far more robust and activist process of statutory interpretation.
With hindsight these comments seem to mark a watershed in the recent
development of our law. If the more robust approach to statutory inter-
pretation under section 39(2) had been followed as Mokgoro J suggested it
should, the door would probably have been opened to the development of a
fully fledged purposive approach to statutory interpretation. This approach
would no longer have been hemmed in by an unqualified fidelity to the gram-
matical possibilities of the text. The need to create the constitutional remedy
of reading-in might never have arisen in the form that it did. However, as the
recent marriage law cases of the Constitutional Court clearly show, this is
not the road which the Constitutional Court chose to follow.
In the next section I provide a close reading of these cases in order to
gain a better understanding of the new textualist method of statutory
interpretation which the court had first distilled from section 35(2) of the
interim Constitution, and then assimilated into section 39(2) of the final
Constitution. At the same time I explore a few traces of resistance from
within the ranks of the court to the new textualism which dominates the
interpretation of statutes under section 39(2) of the Constitution.

4 REINTERPRETING APARTHEID’S MARRIAGE

LAWS

41 National Coalition for Gay and Lesbian Equality

v Minister of Home Affairs
In National Coalition for Gay and Lesbian Equality v Minister of Home
Affairs31 the Constitutional Court had to decide whether the special provision
which was made in section 25(5) of the Aliens Control Act 96 of 1991 for the
“spouse” of a permanent resident, discriminated unfairly and unjustifiably
against the unmarried same-sex life partners of permanent residents. The
Minister of Home Affairs argued that it did not (or that it was not ripe to
decide whether it did) because, on a proper interpretation of the section in
terms of section 39(2) of the Constitution, the term “spouse” included
permanent same-sex life partners.32 The regional committee which was

29
This approach was endorsed by the court in Investigating Directorate: SEO v Hyundai Motor
Distributors: in re Hyundai Motor Distributors v Smit 2001 1 SA 545 (CC) (par [23]).
30
Par [135].
31
2000 2 SA 1(CC).
32
Par [22]-[23].
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 533

responsible for the administration of the act could still very well interpret the
section in this fashion. The court should give the committee the chance to
first explore the interpretive possibilities which the Constitution created.
The court unanimously rejected this argument. Ackerman J, who wrote
the judgment of the court, held that the argument on behalf of the Minister
confused the interpretation of legislation under section 39(2), and the
process of altering the words of a statute through the constitutional remedies
of reading-in and severance under section 172 of the Constitution. According
to Ackerman J, these two processes were “fundamentally different” in
nature.33 Both the fact that Ackerman J insisted on drawing this distinction
and the way in which he then proceeded to attach significance to each of the
two processes, have turned out to be of decisive importance for the later
development of the court’s jurisprudence.
The principle that statutes should as far as possible be interpreted to
conform to the Constitution was accepted and restated by Ackerman J to
mean that “where it is reasonably possible to construe a statute in such a
way that it does not give rise to constitutional inconsistency, such a
construction should be preferred to another construction which, although
also reasonable, would give rise to such inconsistency”.34 The limits of
reasonableness in this context are reached when the meaning of the words
in question becomes distorted.35 Because the process is an interpretive one,
it is limited to what the text is reasonably capable of meaning.36 In order to
illustrate what he meant by an “interpretive” process (as opposed to an
amending process) Ackerman J explained, with an appeal to a dictionary,
that the ordinary meaning of the word “spouse” connoted a married person,
either husband or wife.37 The word “spouse” could therefore not reasonably
be interpreted (by a regional committee or otherwise) to include unmarried
same-sex life partners. To do so would be to distort the linguistic meaning of
the term.
In as far as the section restricted its special advantage to “spouses” only,
it was unconstitutional and had to be declared so in terms of section 172 of
the Constitution. Only then was it open to the constitutional interpreter to
remedy the defect in the legislation by altering its wording to include the
phrase “or partner, in a permanent same-sex life partnership” after the word
“spouse” in the legislative text.38
Ackerman J cited no authority for the statement that the interpretation of
legislation must, hermeneutically speaking, stop short before the
modification of the legislative text begins. His only point of reference seems
to be the Constitutional Court’s own earlier interpretation of section 35(2) of
the interim Constitution. From what was said above, it should be clear that
this section provides no or very little textual support for the distinction
between the interpretation of legislation on the one hand, and the alteration

33
Par [24].
34
Par [23].
35
Ibid.
36
Par [24].
37
Par [25].
38
Par [86].
534 OBITER 2005

of legislation, on the other, which Ackerman J here sought to introduce into

our law.
There is equally no or very little support to be found amongst South
African writers for the “fundamental difference” between the two processes
which Ackerman J so boldly announced. Even a conservative writer and
apartheid judge like LC Steyn, who has frequently been taken to task for the
regressive nature of his hermeneutics and jurisprudence, had argued
against literalists that woordwysigende (word-altering) interpretation formed
an essential and legitimate interpretive activity.39 In the preface to the 1981
edition of his work Die uitleg van wette, the new authors conceded that this
principle was now universally accepted in our law. The continued inclusion of
Steyn’s arguments in this regard was nevertheless justified as follows:
“Even if the walls of Jericho had already fallen in a jurisprudential sense, there
is, with respect, no reason why students should not be taught that there
indeed had been such walls, why they had to fall, and why 40 they should guard
against any attempt to rebuild these walls out of the rubble.”

That Ackerman J managed to start and finish the rebuilding of the walls in
three short paragraphs without challenge is one of the remarkable
intellectual feats in the history of our law. Steyn was of course not alone in
his defence of word-altering statutory interpretation. Writing at the beginning
of the 1990s, Devenish also insisted that “an amendment of a statute by
judicial interpretation” was merely an application of the ordinary technique of
extensive interpretation. Both processes rested logically on the same
hermeneutic principle.41 However, it was JMT Labuschagne who, during the
1980s and early 1990s, developed the most sophisticated and comprehen-
sive hermeneutical support for word-altering statutory interpretation.

39
Steyn Die Uitleg van Wette (1981) 55-68. Steyn argued that words should be read to be
other words or as not appearing in legislation at all if the intention of the legislature cannot
be achieved in any other way (55). This was the logical result of the so-called literalist-cum-
intentionalist approach which he defended against the pure literalists. Steyn insisted that
there was no logical difference between extensive interpretation, restrictive interpretation
and modificative interpretation (58). Steyn also responded to the old claim that the
interpretive modification of legislation amounts to an act of legislating or law-making itself.
According to him this statement rested on a misunderstanding of the nature of legislation. It
is the will of the legislature (the spirit of the laws) and not the words of the legislature (the
letter of the laws) that makes legislation (68).
40
Steyn xv.
41
Devenish 93-98. Devenish noted in this regard that the South African law at the time was
beset by a number of anomalies and inconsistencies. If the modification of the language of a
statute is permiss ble in our law, as he claimed that it was, then the court should also be
allowed to fill a gap or omission in the legislation (98). Devenish pointed out that the degree
of modification which “will be acceptable to the courts will depend on the theory of
interpretation implicitly adopted by the courts”: “The literal approach permits a modification in
exceptional cases of absurdity and ambiguity respectively … A fortiori a purposive or value-
orientated theory of interpretation permits a greater measure of modification of language”
(93). Devenish himself favoured a consistent purposive approach to statutory interpretation
and argued that “[t]he teleological theory of interpretation by its very nature necessitates
acceptance of the principle of modification of language in order to bring the meaning of the
words into line with the purpose of the law in general, which according to Voet, is ‘justice
and reason’. Modification of language is essential in order to develop a consistent and
dynamic jurisprudence of statutory interpretation” (97).
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 535

42 A short interlude (JMT Labuschagne and the end of

statutory interpretation as Ackerman J understands it)
Labuschagne insisted for many years that we do away with both the term
and the practice of statutory interpretation. In one of his last publications
before his untimely death, he wrote that “the use of the phrase ‘legal
interpretation’ as an independent concept is becoming increasingly
frustrating because a legal norm, regardless of its source, is not interpreted
but formed. The term ‘interpretation’ has a too mechanical connotation”.42 In
so far as Ackerman J not only insisted on using the term, but on
distinguishing it from any process by which the legislative text is altered or
even strained, the frustration which was expressed by Labuschagne is
particularly apposite and understandable.
Labuschagne claimed that all word cultures were undergoing a series of
universal evolutionary processes.43 As far as the law was concerned, this
evolution essentially involved a shift from iconic justice to communicative
justice.44 Applied to the field of statutory interpretation, this shift could be
traced through four distinctive phases of development.
The first phase was characterised by the holy origin and status of
legislation and the infallible nature of those who made and declared the
meaning of that legislation.45 The second phase was entered with the rise of
independent judiciaries in the West. Within the early trias politica the
legislature retained its infallible and sovereign status and judges their
apparently declaratory function. The legislation and legislative text of the
(previously holy) legislature equally retained its holy or iconic status. First the
literalist and then the intentionalist approaches to the legislative text arose in
this context. Interpretation was judged on the basis of the logical or formal
relationship which it showed in the abstract to the text which was interpreted.
Yet, the door was opened for the beginning of judicial intervention where
gaps existed in legislation or in cases where logic led to absurdity.46 In the
third phase of development, the literal meaning of the text and the intention
of the legislature were discarded in favour of the purpose of the legislation.
This new phase was highlighted by the development of the functionalist and
sociological jurisprudence of the American Realists during the first decades
of the 20th century. The fourth and final phase of development was reached
with the arrival of the constitutional state. In this phase judges are given the
task, not only to give functional effect to the purpose of legislation, but also
to ensure that the legislative purpose conforms in all cases to basic human
rights standards.47 The task of the interpreter of statutes is to bring the

42
Labuschagne “Die Opkoms van die Rasionele en van Individualiserende Geregtigheid by
Regsuitleg-en Vorming: Opmerkings Oor die Disintegrasie van Refleks- en
Ikongeregtigheid” 2004 THRHR 614 616 (my translation).
43
Labuschagne “Die Heilige Oervader, Regsevolusie en Redematige Administratiefregs-
pleging” 1995 SAPR/PL 444.
44
Labuschagne 2004 THRHR 615.
45
Labuschagne “Die Leemtebegrip by Wetsuitleg as Anachronisme: ’n Regsantropologiese
Verklaring” 1999 THRHR 300.
46
Labuschagne “Die Opkoms van die Teleologiese Benadering tot die Uitleg van Wette in
Suid-Afrika” 1990 SALJ 569 570.
47
Labuschagne “Die Leemtebegrip by Wetsuitleg as Anachronisme: ’n Regsantropologiese
Verklaring” 1999 THRHR 300 302.
536 OBITER 2005

legislative process to both a sensible (practical) and just (equitable)

conclusion. As the complexity of society increases, the scope of the judicial
discretion which is needed to achieve this result equally increases. It
becomes increasingly difficult to legislate in universal terms. The task of the
judge becomes increasingly to find novel ways of furthering the purpose of
the legislative scheme. This phase also brings an end to statutory
interpretation (as it has traditionally been understood) which is replaced by
the law of norm generation (regsnormvorming).48
Already during the purposive phase of development, the dynamic or
organic nature of the legislative process was accepted. The final,
constitutional, phase represents merely a specific application of that insight.
From this perspective, a legislative act is not brought to completion by its
embodiment in a legislative text. The legislative text merely represents a
“structural act”.49 The legislative process is only completed with the creation
or establishment of a “functional act”.50 To ensure that the structural act
becomes functionally operative, it is sometimes necessary to alter the
wording, punctuation and structure of a statute.51 The important point is that
the alteration of the “structural norm” (or legislative text) does not take place
beyond the ordinary process by which a “functional norm” for the specific
legal situation is actualised or formed.
The point of this brief discussion of Labuschagne’s purposive approach to
statutory interpretation is not to defend its merits (which are substantial), but
merely to highlight that a complex and rich body of work already existed in
our law when the Constitutional Court was called upon to develop an
approach to the interpretation of apartheid legislation under the new
Constitution. This body of work celebrated a purposive approach to statutory
interpretation and seemingly only waited for the crucial constitutional
moment to become actualised. Many writers have argued that this
actualisation was indeed effected with the introduction of section 39(2) of the
Constitution into our law.52 If the approach to statutory interpretation which
Ackerman J adopted in National Coalition is compared to the purposive
approaches which Labuschagne and Devenish had earlier advocated, it
becomes clear that these writers have completely misunderstood the
interpretive reach which the Constitutional Court is willing to allow
interpreters under section 39(2) of the Constitution. The Ackerman J of De
Lange has won out, not the Mokgoro J.53
48
Labuschagne “Regsnormvorming: Riglyn Vir ’n Nuwe Benadering tot die Tradisionele Reëls
van Wetsuitleg” SAPR/PL 205.
49
Labuschagne “Gewone Betekenis van ’n Woord, Woordeboeke en die Organiese Aard van
Wetsuitleg” 1998 SAPR/PL 145 148.
50
Labuschagne “Die Dinamiese Aard van die Wetgewingsproses en Wetsuitleg” 1982 THRHR
402 404.
51
Labuschagne “Op die Voetspoor van die Wetgewingsproses: Woord, Punktuasie- en
Struktuurwysigende Uitleg” 1986 SAPR/PL 58.
52
Mdumbe “Has the Literal/Intentional/Textual Approach to Statutory Interpretation been Dealt
the Coup de Grace at Last?” 2004 SAPR/PL 472; Kentridge and Spitz “Interpretation” in
Chaskalson (et al) Constitutional Law of South Africa (1999) 11-8; and De Ville
Constitutional and Statutory Interpretation (2000) 59.
53
Some of the leading writers on statutory interpretation in South Africa have implicitly
expressed their support for this outcome. Botha Statutory Interpretation: An Introduction for
Students (2005) 98-100 insists in so many words that the modificative interpretation of
legislation, what is often misconstrued as the law-making function of the courts, must not be
equated to the physical alteration of the text as promulgated. The text remains, it is only its
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 537

43 Satchwell v Minister of Home Affairs

The judgment in National Coalition has remained the framework within which
the Constitutional Court has subsequently developed its section 39(2)
jurisprudence and interpretive approach to apartheid legislation. In Satchwell
v President of the RSA,54 for example, the court held unanimously that
sections 8 and 9 of the Judges’ Renumeration and Conditions of
Employment Act 88 of 1989, which conferred certain financial benefits on the
“surviving spouse” of a deceased judge, was unconstitutional because it did
not confer those same benefits on the surviving “partner, in a same-sex life
partnership in which the partners have undertaken reciprocal duties of
support”. To counter the defect the necessary words were read into the
sections. As is required in our law, the remedy of reading-in was preceded
by the question whether the sections could be interpreted to conform to the
Constitution. Madala J, who wrote the judgment of the court, held that the
sections could not. His reasons all confirm the narrow textualist approach to
the matter which Ackerman J had introduced earlier. The “ordinary meaning”
of the term “spouse” referred to “a party to a marriage that is recognised by
law”.55 What is more, the term did not have any other more inclusive
linguistic meanings.56 The context in which the term was used in the
legislation, in the third place, in any case did not suggest anything but the
ordinary meaning of the term. As a result, the term “spouse” could not
reasonably be interpreted to include unmarried couples, whether of a
permanent heterosexual or same-sex life partnership.
The neatness of the process by which Madala J came to his conclusion is
beyond doubt. So is its conventionality. The ordinary meaning of the words
in the statute is the starting point. The legislative context in which the words
are used is called upon as a secondary aid to the interpretation of the text.
Where the ordinary meaning and the textual context mutually support each
other, the interpretive process is brought to a close. No further enquiry into
the purpose of the statute or the way in which the text of the statute should
be modified to promote that purpose is undertaken. Should any curative
constitutional action be needed after the textual interpretation of the statute,
the remedy of reading-in should be applied. A simple and workable two step
solution to the interpretation of apartheid legislation had seemingly
crystallised in our law.

meaning that is modified. The purposive approach involves betekeniswysigende (meaning

altering) interpretation not woordwysigende (word altering) interpretation (98-100). Du
Plessis likewise distinguishes between an extensive interpretation of a text and the
modification of that text by reading new words into it (143). Du Plessis claims that there are
good democratic reasons to maintain the distinction. Tampering with the words of legislation
threatens the horizontal division of power in the state and elevates the interpreter to
legislator (229). Du Plessis also suggests that in future the adaptation of legislation or
reading-in should be restricted to cases where “legislation is impugned on constitutional
grounds” (232).
54
2002 6 SA 1 (CC).
55
Par [9].
56
Ibid.
538 OBITER 2005

44 Daniels v Campbell
Less than two years after Satchwell the meaning of the term “spouse” again
arose for determination by the court in Daniels v Campbell.57 This time,
however, the earlier consensus in the court was put under pressure and,
surprisingly, gave way. The judges of the court began expressing differences
of opinion, both about what the term “spouse” in apartheid legislation meant,
and about the proper interpretive approach that should be adopted under
section 39(2) of the Constitution in order to arrive at that meaning. Madala J,
who was responsible for the judgment in Satchwell, distanced himself from
the meaning which the majority of the court was now attaching to his earlier
judgment.
The question before the court was whether the term “spouse” in the
Maintenance of Surviving Spouses Act 27 of 1990 could be interpreted to
include the partners in a Muslim marriage which had not officially been
solemnised in terms of the Marriage Act. The majority of the court held that it
could; the minority that it could not. In reaching this conclusion, the majority
significantly relied on not one, but two different interpretive approaches. The
first approach was predominantly textualist and in keeping with the approach
which the court had adopted in National Coalition and Satchwell. The
second approach was predominantly purposive, and introduced a potentially
far-reaching new dimension into the interpretation of statutes under section
39(2) of the Constitution.
The textualist argument can easily be reconstructed from the judgments of
Sachs J and Ngcobo J and went something like this: Section 39(2) places a
duty on the court to prefer an interpretation of legislation that falls within
constitutional bounds over one that does not, provided that such an
interpretation is not unduly strained and can be reasonably ascribed to the
legislative provision in question. In the case at hand, two unstrained
meanings can reasonably be given to the term “spouse”. The first meaning
dates from the era of apartheid and restricts the term to “partners in a legally
recognised and solemnised marriage”. The second meaning is wider and
extends the term to “partners in all monogamous marriages which have
been concluded under any tradition or system of religion, personal or family
law” (whether those marriages have been officially solemnised or not).
Which of the two meanings in fact reflects the ordinary meaning, dictionary
meaning or prima facie meaning of the term is open to debate, but actually
besides the point, as the term can reasonably bear both meanings.58
Because both the meanings can reasonably be ascribed to the term
“spouse”, the meaning which best conforms to the Constitution should be
adopted. That is the wider and more inclusive meaning. Partners to a Muslim
marriage were spouses.

57
2004 5 SA 331 (CC).
58
Sachs J, eg, suggests that the wider meaning reflects they way that the term is generally
understood and used. He even goes so far as to state that “it is far more awkward from a
linguistic point of view to exclude parties to a Muslim marriage from the word “spouse” than
to include them. Such exclusion as was effected in the past did not flow from courts giving
the word “spouse” its ordinary meaning. Rather, it emanated from a linguistically strained
use of the word flowing from a culturally and racially hegemonic appropriation of it. Such
interpretation owed more to the artifice of prejudice than to the dictates of the English
language” (par [19]).
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 539

The majority of the court did not restrict itself to this linguistic approach to
the interpretation of the statute but also endorsed a second, slightly different
approach to the matter. This second approach is particularly prominent in
the judgment of Sachs J. It is explicitly non-textual in character. The
reasoning went something like this: the case could not be decided by merely
looking at the linguistic meaning of the word “spouse” in the abstract (as it is
used in the dictionary). The true meaning of the word could only be
determined by looking at the socio-economic and legislative context of the
case as a whole. This context included the object or purpose of the
legislation. The aim or purpose of the legislation in question was to provide
some financial relief to widows (who would otherwise remain a marginalised
and vulnerable group in our patriarchal society). The term “spouse” should
be interpreted to ensure that this aim or purpose is indeed realised as far as
possible.
From this vantage point it is not difficult to conclude, as Sachs J in fact
did, that the “manifest purpose of the Acts would be frustrated rather than
furthered if widows were to be excluded from the protection the Acts offer,
just because the legal form of their marriage happened to accord with
Muslim tradition and not the Marriage Act”.59
Note that the question under this second approach is whether widows of
Muslim marriages are functionally (as opposed to linguistically) similar to the
widows of legally recognised marriages. The question is not whether the two
widows can both be included under the same conceptual definition. The rele-
vant context in question is patriarchy, not the dictionary. As Sachs J put it:
“The central question … is not whether it had been open to the applicant to
solemnise her marriage under the Marriage Act, [ie to bring her under the
definition of the term ‘spouse’] but whether, in terms of ‘common sense and
justice’ and the values of our Constitution, the objectives of the Acts60would be
furthered by including or excluding her from the protection provided.”

Although Sachs J carefully introduced this purposive reading of the term

“spouse” on the back of the dominant textualist approach (the purposive
reading was merely used to bolster the textualist reading), the groundwork
had been laid for the development of a distinctly new approach to the
interpretation of apartheid legislation. This approach would be far less
“benevolent” to the legislative text (as Mokgoro J had put it). However, the
roots of this alternative are not the only ones that can be discerned in the
judgment. In a minority judgment delivered by Moseneke J (supported by
Madala J) the outlines of exactly the opposite approach to the question were
traced. The critique of this minority was directed at the majority’s under-
standing of the boundaries which the legislative text imposed on the
interpretive process. In Moseneke J’s view, these boundaries were much
narrower than the majority believed. The dispute was therefore essentially
about the narrowness of the textualist approach which the court should
adopt.
Moseneke J held that the term “spouse” could not be read in terms of
section 39(2) of the Constitution to include all married couples, but had to be

59
Par [23].
60
Par [25].
540 OBITER 2005

restricted to legally married couples only. When read in this manner the word
“spouse” had a discriminatory effect. The legislative provisions in question
had accordingly to be declared unconstitutional. It could thereafter be
remedied by a process of reading-in. The starting point of this line of
reasoning lay in the confirmation that the method of construction which was
laid down in section 35(2) of the interim Constitution was now, without
reserve, to be found in section 39(2) of the final Constitution. This meant that
the proviso which was built into section 35(2) of the interim Constitution also
applied to section 39(2) of the Constitution:
“[The] affirmative duty to ‘read’ legislation in order to bring it within
constitutional confines is not without bounds. An impugned statute may be
read to survive constitutional invalidity only if it is reasonably capable of such
compliant meaning. To be permissible, the interpretation must not be fanciful
or far-fetched but one that reasonably arises from the challenged
61
text without
unwarranted strain, distortion or violence to the language.”

That this passage has been designed to give the principle of

verfassungskonforme Auslegung or reading-down its narrowest possible
meaning should be clear. The narrow textual or linguistic bounds within
which this interpretive method can operate is tautologically restated no less
than six times. All fanciful, far-fetched, unreasonable, strained, distorted and
violent uses of language are prohibited. Plato is banning the poets from the
city. The only difference is that Platonic conceptions of language have been
thoroughly discredited in the 20th century by Wittgensteinian pragmatics and
Saussurian and post-structuralist linguistics. It is Moseneke J himself who
ironically reminds us that “clarity of language does not rest on any ‘objective
quality’ of language itself, but on the reader and the context”.62 That he can
move from here to a finding that the majority misunderstood the ordinary
meaning of the word “spouse” is bewildering. Why in any case should
Moseneke J be so keen to insist on the correct interpretation of the ordinary
meaning of the word “spouse”, if he himself shortly before reminded us that
the interpretive approach laid down by section 35(2) of the interim
Constitution is still operative in our law? Was the whole point of that section
not to mandate constitutional interpreters like himself to move away from the
prima facie (read ordinary or dictionary) meaning of legislation?
The ordinary meaning argument therefore does not take Moseneke J very
far. When it runs dry, he slightly adapts it and begins to appeals to the “legal
meaning” of the text.63 What Moseneke J fails to make clear, however, is
where the legal (as opposed to the ordinary) use of language would rate on
his list of out-of-bounds language uses. Does the legal not amount to the
fanciful or even the violent?64 The question is made even more pertinent as
Moseneke J seems to imply that the present legal meaning of the term
“spouse” must be equated with the meaning of the term which courts had
developed during the long years of racial oppression in our land. This traps

61
Par [83].
62
Par [88].
63
Par [89].
64
Cover Narrative, Violence, and the Law: The Essays of Robert Cover (1992) 223 describes
statutory interpretation as “bonded interpretation” which takes place in a field of violence and
death.
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 541

us and the text of the legislation, he seems to imply, in a circle of violence

which can only be broken if the legislation is declared unconstitutional.65
The weakness of these two linguistic arguments prompts one to ask what
really was driving Moseneke J to prefer an outright declaration of
unconstitutionality to a nuanced reinterpretation of the statute in question.
Why was Moseneke J so vehemently opposed to the interpretive
engagement with the legacy of the apartheid past? He tries for a third time to
justify his position with an appeal to the legislative text:
“[The] main judgement conflates the meaning that the Acts can reasonably
bear with the constitutional remedy the applicant and others similarly situated
might be entitled to. These processes ought to be two separate enquiries; the
first goes to interpretation, and the second to remedy. Otherwise66the meaning
of the text becomes subservient to a preferred outcome of relief.”

However, we have already began to sense that something more lies

behind this rhetorical appeal to the “meaning of the text”. The following
remark confirms that suspicion:
“The problem of readily importing interpretations piecemeal into legislation is
the precedent it sets. Courts below will follow the lead and readily interpret
rather than declare invalid statutes inconsistent with the Constitution.
However, constitutional
67
reinterpretation does not come to this Court for
confirmation.”

The interpretive principle and approach encapsulated in section 39(2) of

the Constitution, the idea of judicial restraint and memorial constitutionalism,
poses a risk to the rule of law if it is not carefully delimited and controlled.
Law is made by a single centralised legislature. The application of that law is
undertaken by many different courts. If these courts were allowed and
encouraged to interpretively adapt legislation according to their reading of
what will be in conformity with the constitution, legal fragmentation and
uncertainty will arise and the rule of law will be threatened. The real problem
is that the interpretive development of legislation or the determination of
legal meaning will not be centralised. Constitutional reinterpretation of
legislation should be avoided because it is a process over which the
Constitutional Court has no direct control (unlike the process of reading-in
which has to be approved and certified by the Constitutional Court).
One should appreciate the boldness with which these considerations of
institutional politics are put forward. If the meaning of legislation depends in
the end on the readers of the legislation, as Moseneke J earlier suggested,
then the only way of controlling the range of meanings which is legally
allowed to circulate in society is by directly controlling the readers
themselves (if not by eliminating unwanted or uncontrollable readers from
the greater interpretive enterprise itself). The rule of law is therefore
ensured, not by the texts of the law, but by the institutional politics which
controls the reading of those texts. It turns out that the process of
constructing a non-violent, non-fanciful, unstrained and undistorted meaning

65
Par [99].
66
Par [86].
67
Par [104].
542 OBITER 2005

out of a legislative text requires far more violence and institutional politics
than Moseneke J himself initially cared to suggest.
We come away from the judgment in Daniels with a majority judgment in
which the court’s largely textualist understanding of the interpretive process
under section 39(2) of the Constitution has been reconfirmed; a minority
judgment in which we are exhorted to avoid that approach to apartheid
legislation as far as possible and, to that end, in which the approach is
restated in such narrow textualist terms as to make it practically useless; and
a second majority judgment in which the traces of a purposive alternative to
both these textualist approaches can be discerned. Even so, after all has
been said and done, all three of these approaches resulted in exactly the
same substantive relief to the applicant − she became entitled to the
legislative benefits which were in the past reserved for legally married
widows only. If this fact creates the impression that I am merely trying to
blow up a storm in a tea-cup by teasing out the differences between these
approaches, the recent Constitutional Court judgment in Volks v Robinson68
provides a powerful reminder to the contrary.

45 Volks v Robinson
The case again involved the meaning of the term “spouse” in the
Maintenance of Surviving Spouses Act 27 of 1990 (the same apartheid act
which entertained the court in the Daniels case). This time it was not the
surviving partner in a religious marriage who approached the court, but the
surviving partner of a monogamous heterosexual life partnership. In the
majority judgment of the court, Skweyiya J held that the term “spouse” was
not reasonably capable of being interpreted in terms of section 39(2) of the
Constitution to include heterosexual cohabitants.69 To include unmarried
persons under the term “spouse” would be to unduly strain the ordinary
meaning of the term (which, following Daniels, meant partners to a marriage
either by law or religion). In addition, the textual context in which the term
“spouse” was used in the statute made it clear that it was restricted to
married persons.70
This interpretation of the term “spouse” was also adopted in two minority
judgments. What divided the court was whether this interpretation
discriminated unfairly and unjustifiably against the surviving partner in a
unmarried life partnership. The majority held that it did not; the two minorities
that it did. While this question, strictly speaking, no longer involves the
nature of statutory interpretation under section 39(2) of the Constitution, the
different interpretive approaches which the court adopted in answering the
question before it go to the heart of our discussion.
In his minority judgment, Sachs J declared his total disagreement with
both the substance of the conclusion which the majority reached, and the
approach which was utilised to reach that conclusion.71 The second issue
finally brings the differences between the textualist and purposive

68
2005 5 BCLR 446 (CC).
69
Par [40]-[45].
70
Par [41].
71
Par [151].
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 543

approaches, which were still latent in the Daniels judgment, to a head.

Sachs J explained that he rejected the interpretive approach of the majority,
not because it was illogical, but because of the narrow definitional framework
in which the legal logic unfolded.72 The majority solved the dispute by
defining what a marriage is and highlighting which duties flowed logically
from that definition. It then interpreted the provisions of the act within this
definitional framework and found that a widow’s right to claim maintenance
was a logical extension of the duty of support which living spouses owed
each other.73 Because no equivalent duty of support arises by law in the
case of life partnerships, it was not illogical and thus not unfair to different-
tiate between married and unmarried partners for the purposes of the act.
Sachs J’s response to this approach to the issue is both terse and telling:
“I do not accept that it is appropriate to examine the entitlements of the
surviving cohabitant in the context of what the common law would provide
during the lifetime of the parties. To do so is to employ a process of
definitional reasoning which presupposes and eliminates the very issue which
needs to be determined, namely, whether for the limited socially remedial
purposes intended to be served by the Act, unmarried survivors could have a
legally cognisable
74
interest which founds a constitutional right to equal benefit
of the law.”

With this statement Sachs J essentially reconfirmed the purposive

approach which he began developing in Daniels. The legal dispute should
not be approached on the basis of the linguistic definition of the term
“marriage” in the act, neither on the basis of the common law implications
which follow logically from that definition, but on the basis of the social
purpose of the act and the best way to further that purpose. The rules of
logic which the majority employed only held as a result of the artificial (read
legal) way in which the framework for their operation had been chosen. What
was fair in a constitutional sense could not be determined within that
framework. What was fair could simply not be reduced to what was logical.
The “impeccable” logic of the legal argument became irrelevant as soon as
the framework itself was redrawn. This is the interpretive strategy which
Sachs J employed.
The interpretative framework of the act, he held, should not be located in
matrimonial law but in family law.75 The question then becomes, not how
“marriage” is defined and whether unmarried people could be included under
that legal category (which they obviously could not), but what the purposes
of marriage law are within the context of family law. Once the function of
marriages and matrimonial law within family law are identified, the question
becomes whether the same functions could be performed by other
functionally (as opposed to conceptually) similar relationships (which they
obviously could). In Sachs J’s own words, one should shift the focus from a
“decontextualised” to a “contextual” approach,76 or from a “definitional” to a
“functional” approach.77 Sachs J concluded that the purposes of the statute

72
Ibid.
73
Par [56] and [87].
74
Par [151].
75
Par [152].
76
Par [162].
77
Par [172].
544 OBITER 2005

(to protect widows and persons similarly positioned) will be frustrated if the
protection of the statute was not in principle, but always depending on the
facts of each case, extendable to the surviving partner of a life-partnership.78
The judgment in Volks finally and dramatically highlights the presence of
two clearly defined but diametrically opposing interpretive approaches in the
Constitutional Court. What is more, the two approaches resulted in two
different substantive outcomes to the dispute. Even so, the case did not
revolve directly around the meaning of the statutory provision in question. It
revolved around the constitutional implications of that meaning. The most
recent and controversial of all the marriage law cases, Fourie v Minister of
Home Affairs,79 takes us right back to the narrower question at hand.

46 Fourie v Minister of Home Affairs

Having found that the modern South African common law recognises same-
sex marriages, the interpretive question which confronted the Supreme
Court of Appeal in Fourie was whether the marriage formula in the Marriage
Act 25 of 1961 could be interpreted to accommodate those types of
marriages. The majority of the court held that it could not; the minority that it
could. The differences between the two judgments stem from the different
approaches to the interpretation of the statute which the respective justices
had adopted. For this reason the case is the most spectacular example of
the tensions which mark the Constitutional Court’s approach to statutory
interpretation.
Writing on behalf of the majority of the court, Cameron JA immediately
takes the questions back to the basic approach to statutory interpretation
which was formulated by Ackerman J in the National Coalition case.80 To
recall, the approach rests on (i) a clear distinction between interpretive
techniques and constitutional remedies; and (ii) a textualist understanding of
those interpretive techniques. As Cameron JA points out, in terms of that
approach, the court may assign “a broad meaning to a word whose purport
was not certain” but had to do so “without changing the word”.81 The
marriage formula which was inherited from apartheid required a public
declaration of the parties that they take each other as “lawful husband (or
wife)”.82 The words “husband” and “wife” could not reasonably be given a
wider meaning which would include the partners in a same-sex marriage.
This meant that the marriage formula could not be updated interpretively in
terms of section 39(2) of the Constitution to allow for same-sex marriages.
The marriage formula had to be declared unconstitutional and remedied by
an appropriate reading-in.83 (The technical difficulty with which the court was
confronted was that the constitutionality of the formula had, in an apparent
oversight, not been challenged before it.)

78
Par [190]; and [213]-[216].
79
2005 3 BCLR 241 (SCA).
80
Par [28]-[29].
81
Para [31].
82
S 30(1) of the Marriage Act 25 of 1961.
83
Para [33].
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 545

In a minority judgment, Farlam JA took a completely different approach to

the interpretation of the statutory marriage formula.84 He held that it was not
necessary to declare the present marriage formula unconstitutional (a
course of action which was in any case not open to the court), as the formula
could simply be updated in accordance with section 39(2) of the Constitution
to make provision for same-sex marriages. This required a slight change to
the wording of the legislation to include “or lawful spouse” after the words
“lawful husband (or wife)”.85 Although this amounted to an alteration of the
legislative text, the alteration was a legitimate part of the ordinary process of
statutory interpretation. This is so simply because of the nature of legislation
and the legislative process. Legislation is made to stay operational, or to
always remain speaking, for as long as it is on the statute book.86 To ensure
that legislation remains operative, the interpreter can safely work under the
presumption that the legislature intends the original wording of the legislation
to be updated where and when necessary. Farlam J calls this “the
presumption of updating interpretation” and claimed that it was a settled
presumption of ordinary statutory interpretation.87
This approach to the interpretation of apartheid legislation under section
39(2) of the Constitution departs radically, as Cameron JA rightly points out,
from the textualist approach which the Constitutional Court had thus far
adopted to the interpretation of statutes. It could at best be regarded as an
extension of the purposive approach which Sachs J began to develop in
Daniels and Volks. It is precisely this fact which commends it. Farlam JA is
not concerned with the meanings (ordinary or not, strained or not) of the
terms “husband” and “wife” and the question whether same-sex partners
could be incorporated into those meanings. All his attention is directed at the
purpose of the marriage formula in the context of the Marriage Act, the
common law, and the constitutional order as a whole. That purpose, he
claims, is to enable all persons who can get married in terms of the common
law to get married.88 The present marriage formula has clearly become
outdated in the light of the development of the common law and serves only
to obstruct and no longer to further that purpose. As an interpreter of the act,
Farlam J claimed that his duty was to see to it that the marriage formula was
updated, within the bounds set by the purpose of the act, to enable same-
sex couples to get married.89 In this way the statute could remain relevant in
the changing circumstances.
In support of this purposive approach to statutory interpretation, Farlam
JA relies on the work of Francis Bennion.90 Bennion in turn associates
himself, and has been associated with, the theory of dynamic statutory
interpretation which has been developed by William Eskridge.91 Both
Bennion and Eskridge insist that legislation is not something static and even
more, that legislation cannot be reduced to the textual form in which it has

84
Par [132]-[138].
85
Par [138].
86
Par [136].
87
Par [136]-[137].
88
Par [135].
89
Par [138].
90
Bennion Statutory Interpretation (2002) 762-782.
91
Eskridge Dynamic Statutory Interpretation (1994) 49.
546 OBITER 2005

been cast at the promulgation of the legislation. Legislation is essentially a

purposive undertaking by a legislature of which the initial textual
embodiment is merely a conditional or preliminary statement (a structural act
in the words of Labuschagne). The purposive undertaking in question might
require authoritative restatements as time goes on and the circumstances
and assumptions upon which the legislative text initially rested change.
Because the legislature cannot constantly adapt the language of its
legislation to the new circumstances, it is left to the courts in the most
pressing cases to provide the purposive undertaking of the legislature with a
new operational language. An absolute fidelity to the text misunderstands
both the nature of legislation and the legislative process.
One would have thought that the principle of updating statutory
interpretation would have been recognised as an essential part of statutory
interpretation after the fall of apartheid, where, as was mentioned above, the
legislature has not embarked on an extensive revision of the apartheid
statute; and where, so it is often assumed, a purposive approach to statutory
interpretation has been introduced by section 39(2) of the Constitution. From
the discussion above it should be clear that the opposite has in fact
happened. The point is underscored by Cameron JA’s rejection of the
interpretation of the marriage formula which Farlam JA proposes. Cameron
JA is quick to remind Farlam JA that in our law a clear distinction is now
drawn between the interpretation of legislation and the amendment of
legislation. He goes on to explain that updating statutory interpretation is
now merely of historical relevance to our law, as it is something that was
devised for jurisdictions which do not or did not have the ample remedies of
constitutionalism which South African courts now possess.92
If this is indeed the case, then the South African law of statutory
interpretation has paid a high price for the gains which have been achieved
by the introduction of the new constitutional order. Hermeneutically
speaking, our law of statutory interpretation is now closer to the textualism of
old than any sound and consistent purposive or contextual approach, such
as the ones which were advocated during the 1980s by writers like Devenish
and Labuschagne. The attempt by Cameron JA to deny the force of these
hermeneutical insights with an uncritical appeal to the positive law will not
do. The attempt by Farlam JA to broaden the debate by bringing the
theoretical work of other statutory interpreters to bear on the issue was much
needed and should be welcomed.

5 CONCLUSION
It is fair to say that the Constitutional Court has not been particularly creative
in exploring the opportunities which the principle of verfassungskonforme
Auslegung or reading-down offered the cause of legal and societal
transformation in post-apartheid South Africa. This has largely been as a
result of the textualist interpretation which the court has attached to that
principle in its interpretation of, first, section 35(2) of the interim Constitution
and then section 39(2) of the final Constitution. The court has from the
outset been more concerned with the bounds or limits of this principle than

92
Par [131].
UNDOING THE PAST THROUGH STATUTORY INTERPRETATION: 547

with its transformative potential. The fact that some of the senior justices of
the court have begun arguing for an even more restricted use of section
39(2) might in due course render statutory interpretation in post-apartheid
South Africa irrelevant as a strategy of collective memory. This would be an
unfortunate result both as far as the hermeneutics of our interpretive
practices and the democratic nature of our politics are concerned.
Hermeneutically speaking, the danger of the new textualism in statutory
interpretation is that it could easily spill over to constitutional interpretation
itself. The way in which Ackerman J introduced the new textualist approach
into our law in National Coalition suggests that this might already have
happened. The approach was introduced without discussing the merits of
competing hermeneutic approaches, as if the approach which Ackerman J
adopted followed naturally or logically from the text of the Constitution itself.
The approach itself is characterised by questionable conceptual distinctions
which quickly combine to form a fixed method to be followed in all cases,
even where the method results in a completely impractical and inequitable
result (as it did in Fourie). Cameron JA’s desperate and highly ironic appeal
in that case to the church to save him from his interpretive dilemma suggests
that the interpretive work around section 39(2) has been completed and that
the meaning of the section for statutory interpretation lies beyond
contestation. Cameron JA was certainly not willing to ask any probing
questions about the way in which he came to be painted into his interpretive
corner.
As Cameron JA’s remark about the historical relevance of updating
statutory interpretation makes perfectly clear, the new textualist
understanding of statutory interpretation under section 39(2) is completely
dependent on the existence of the section 172 constitutional remedies
(severance and reading-in). The statutory revision of apartheid legislation
can be limited to the extent that it has been by the court, only because the
necessary alterations to the legislation can be achieved through other
means. I have already suggested that this is a hermeneutically unsound
limitation (or even misunderstanding) of the interpretive process. It is an
equally unsound political limitation of the democratic process. The problem
is that the constitutional remedies (unlike the interpretive techniques) are
reserved for use by the Constitutional Court itself, or under its direct
supervision only. For somebody like Moseneke DP this is precisely what
commends the neat distinction between interpretive techniques and
constitutional remedies. However, the centralising tendencies which are built
into this distinction pose a threat to the democratisation of legal meaning in
post-apartheid South Africa. In fact, the centralised control of meaning in the
apartheid state is perfectly mirrored by this juricentric approach to legal
interpretation in its post-apartheid successor.
William Eskridge has remarked recently that legal interpreters would do
well to remember that courts are not the only institutions in society which
interpret legislation. He advocates a bottom-up approach to statutory
interpretation precisely because the political and contestable nature of legal
meanings are better accentuated thereby:
“Viewing statutory issues from the perspective of actors lower in the legal
hierarchy enables us to see more clearly how statutory interpretation reflects
the dynamics of political conflict and balance … The struggle for meaning
548 OBITER 2005

occurs when different communities of interpretation form around a statutory

issue. The communities of interpretation contest with one another about what
the statute should mean, each one jockeying for position within and without
the formal channels of93 government and developing arguments for their
different interpretations.”

By limiting the scope of what ordinary courts and other interpretive

agencies at the centre of local interpretive communities can effectively
achieve in the interpretive revision of our apartheid legacy, the Constitutional
Court has severely undermined the transformative capacity of the law in our
transitional society. The way in which Ackerman J introduced the new
textualist approach in National Coalition is a case in point. By prematurely
taking the interpretive issue out of the hands of the regional committee, the
court has reduced the places and opportunities for disruptive and
participatory democratic action and politics in our society. Little remains of
the jurisprudence of restraint which Du Plessis initially saw embodied in
section 35(2) of the interim Constitution after this centralisation of judicial
power. Where Du Plessis insisted on an open community of constitutional
interpreters in which the Constitutional Court resisted the temptation of
taking centre stage, Ackerman and Moseneke JJ seem to insist on a closed
and highly centralised community of constitutional interpreters in which the
Constitutional Court is the only stage.
The Constitutional Court’s engagement with the apartheid marriage laws
has now reached its final conclusion in the Lesbian and Gay Equality Project
v Minister of Home Affairs case.94 In this case the constitutionality of the
existing marriage formula in section 30(1) of the Marriage Act was directly
challenged. In bringing this challenge, the parties in question did not ask for
the marriage formula to be updated in terms of section 39(2) of the
Constitution (as was applied for on appeal in the Fourie case). The
Constitutional Court decided to hear the two cases together and took this
gap to proceed with the matter in both cases on the basis of a literal
interpretation of the existing marriage formula. This formula was duly
declared unconstitutional and remedied in terms of section 172 of the
Constitution by reading the words “or spouse” into the formula. Farlam JA’s
attempt in Fourie to creatively update the wording of the statute accordingly
does not receive any further attention. The fact that none of the
Constitutional Court judges even deemed it necessary to respond (even
disapprovingly) to Farlam’s JA approach to the interpretation of the marriage
formula is a dangerous sign of complacency. It simply and sadly reconfirms
the Constitutional Court’s drive to centralise the interpretive production of
statutory meaning. However liberating the Constitutional Court’s
engagement with the marriage laws of apartheid might in hindsight turn out
to have been, the unfortunate fact will remain that this same engagement
has, hermeneutically and politically speaking, also opened the door to what
might still turn out to be an extremely reactionary approach to constitutional
and statutory interpretation.

93
Eskridge 71-72.
94
Case number CCT 10/05 (as yet unreported). The case was heard and judgment delivered
together with the judgment in the Fourie case (case number CCT 60/04).