of application. The consequences of the matter proceeding thus is that the court is confronted by two sets of
opinion evidence each based on the facts put up by the applicant as synthesised (and amplified) by the facts
put up by the respondent. Where those expert views are based on sound reasoning and are equipoised how
is the court to determine preference applying the approach laid down in the Potato Board case? Most certainly,
the rule in PlasconEvans should not be applied, for that approach is to determine disputed facts not opinion.
[69] While much of what Prof Dunne has said is not disputed in reply, and while many of his conclusions make
emminent sense and follow logically, I am nevertheless reluctant to decide this matter of the strength of
expert opinion. The case involves technical issues of lexicography and the statistical relevance (or irrelevance)
of that science of coincidences and similarities in the content. There are a number of issues, which I would
prefer to be clarified before expressing a definite preference for an opinion one way or the other. For
example, how similar linguistically and lexicographically are the "melk" and "sap" examples? And how can one
determine without proper interrogation through questioning of the experts how much of the Pharos work was
procured lawfully from the same source as the OUP work. These are trenchant questions which remain
unanswered and which undoubtedly will be considered when the trial for damages proceeds.
[70] In the result I have come to the view that, given the technical nature of the issues involved, it would not be
appropriate for the court to decide the case on the strength of the expert reports on affidavit. The matter
must accordingly be determined in accordance with the PlasconEvans rule with due regard for the onus which
Media 24 has attracted by moving for final interdictory relief.
Sufficient objective similarity?
[71] In the passage in Galago to which I have already referred, Corbett JA stressed that it was for the party
claiming copyright protection to persuade the court that there was "sufficient objective similarity" between the
Page 495 of [2015] 3 All SA 478 (WCC)
alleged infringing work and the original, or more specifically, that "a substantial part" of the Pharos work has
been copied by OUP. But, how much is required for the plagiarism to be regarded as "substantial"? Prof Kidd's
evidence suggests a coincidence of example sentences at around 20%. That implies that one in every five
example sentences in the OUP work is the same (or substantially similar) to the Pharos work. However, as
Prof Dunne suggests the selection criteria for similarities relied upon by Prof Kidd may be flawed: it is certainly
open to criticism, as Prof Dunne has purported to do.
[72] In his objective evaluation, Dr Rundell seems to have conceded a similarity ratio of around 7%. In such
circumstances, is the repetition of every 14th example sentence then "substantial"? And, even if it is
substantial is it as a consequence of copying by OUP's lexicographers or is there some other similarly feasible
explanation? After all, as the leading authority19 in the English law of copyright points out, one is concerned
here with the quality or importance of copying rather than the quantity thereof.
[73] It is common cause that when preparing a new work lexicographers will have regard to a variety of other
dictionaries as source or reference material. In my view it is critical in this case to bear in mind, also, that we
are dealing here with a reference work for use by the general public, and in particular school learners, and
that there would therefore exist a fairly well defined body of knowledge (from traditional works to online
publications and computer software) available to be utilised by the compilers.
[74] On this score Mr Duminy SC referred the court to a conference paper written by John Williams, a freelance
lexicographer attached to the Open University at Milton Keynes in the United Kingdom, entitled "The Question
of Plagiarism and Breach of Copyright in the Dictionarymaking Process (with Particular Reference to the
UK)".20 The paper is manifestly not intended to be a legal treatise on the topic. Rather, it serves to highlight
the ethical and very real practical problems confronted by practitioners in this important area of learning, while
touching upon aspects of the English law of copyright. What the paper demonstrates is that at least 20 years
ago the question of plagiarism in respect of the compilation of dictionaries was anything but settled law in the
United Kingdom. And, given that neither party to these proceedings was able to adduce any subsequent
authority relating to the unlawful copying of dictionaries in either the United Kingdom or South Africa, I must
accept that the position today is no different to what it is was in 1992. I shall quote extensively from the
paper since it places the conundrum in its true perspective.
[75] In the opening section of his paper Williams sets the scene as follows:
"1.
POSING THE PROBLEM
The question of plagiarism or infringement of copyright poses particular problems in dictionary publishing,
distinct from those faced by the
Page 496 of [2015] 3 All SA 478 (WCC)
general publisher. Personnel at all levels of the dictionarymaking process from the company's lawyers to
project managers and even lowlevel freelance compilers may from time have to make decisions as to what
constitutes, and what does not constitute, an unacceptable degree of copying from previously published
dictionaries. Prima facie cases of plagiarism may manifest themselves in many different aspects of the
dictionary text: conventions of presentation, the headword list, definition style, not to mention the actual
content of particular definitions, examples, translations, and illustrations.
The core of the problem resides in the fact that all lexicographers working in a given language (or between two
given languages) are covering essentially the same body of knowledge. Given that this body of knowledge is
itself linguistic in nature, and that it is usually incumbent upon the lexicographer to produce a text which
documents often by way of example typical meanings and usages, it is probably inevitable that dictionaries of