an expert lexicographer and consultant from the United Kingdom, who is, inter alia, the EditorinChief at
Macmillan Dictionaries in London, and a visiting professor at various universities around the world. Thirdly,
OUP has called into service Prof Tim Dunne, a statistician from the University of Cape Town with more than 30
years' experience. There are further expert reports by Prof ElsabeTaljaard, a lexicographer from the University
of Pretoria, and Dr Philip Louw.
[46] Dr Rundell observes that both works are relatively small dictionaries, each with about 5000 words per side.
They are said to be pedagogical dictionaries for use by second language learners and contain relatively low
levels
Page 489 of [2015] 3 All SA 478 (WCC)
of language proficiency. Importantly, says the expert, both works are reference materials, unlike, for example
novels or songbooks. And, as already noted, both aim to attract the interest of the same target market. Dr
Rundell is of the view that repetition by OUP is often based on the selection of similar headwords for which
the most common example sentences have been chosen. I will cite just two of his explanations.
[47] The headword "broom" has the following example sentence in the Pharos work: "Doreen, please take the
broom and sweep out the kitchen". In the OUP work the example is "Take the broom and sweep the floor,
please". In respect of each of these headwords, there is only one example sentence. As Dr Rundell points out
"Both [example sentences] refer to using a broom to 'sweep': this is what brooms are for."
[48] Then, there is the headword "spring". In the Pharos work there is only one example sentence in respect of
"spring" used as a noun: "spring is the season between winter and summer." In the OUP work, "spring" is
cited both as a noun and a verb. And, in respect of the noun there are three example sentences one of which
also reads "spring is the season between winter and summer". As to this obvious similarity, Dr Rundell holds
the view that "in many monolingual English dictionaries this precise formula is used as a definition of "spring",
the clear implication being that this example sentence, if it has been copied, may have been sourced in a work
other than the Pharos work. And, it may be that Pharos, too, consulted the same source material as OUP.
[49] The other example sentences for "spring" in the OUP work are "a spring is usually made of metal", and "we
get water from the spring on the farm". I shall revert to the use of multiple example sentences, later but
merely observe at this stage that it does appear in a number of instances that OUP employs the technique of
the use of multiple example sentences more often than Pharos. Finally, in several of the instances of
plagiarism complained of by Dr Smith, the allegedly similar example sentences presented by OUP are located
amongst a number of others which are entirely different.
[50] Prof Dunne issues a warning regarding the use of what he calls "afterthefact determinations of statistical
causality". The process of analysis, he says, "needs to take into account all plausible factors that could have
affected the data, including those factors that might easily be overlooked."
[51] In relation to the coincidence of similar headword use, Prof Dunne makes a number of observations. Firstly, he
says he found, with reference to the letters B, D, I and S that 712 headwords were common to the Afrikaans
side of each dictionary. Then he found at least 589 headwords that were unique to the Afrikaans side of the
OUP dictionary and 349 headwords, which were unique to the Afrikaans side of the Pharos work. From a
percentage point of view, Prof Dunne found that 54,7% of the OUP headword list was shared while at least
45% of the list was unique to OUP. Conversely 67,1% of the Pharos headword list was shared with OUP while
at least 32,9 % of the list was unique to OUP. It seems as if this view of Prof Dunne lead to Pharos
reconsidering reliance on plagiarism in
Page 490 of [2015] 3 All SA 478 (WCC)
respect of headword lists, something with which Dr Smith dealt fairly extensively in the founding affidavit.
[52] Commenting on this, Prof Dunne highlights the following aspect:
"82.
This level of uniqueness in the Afrikaans/English (B D I S) sides of the two dictionaries has to impact upon the
levels and extent of any plagiarism mooted. The question is why would a system that can construct at least 589
words completely absent from the Aanleerderswoordeboek bother with selection of 712 words out of 1061, and
then painstakingly obscure matchings of example sentences, but nonetheless leave matchings of some 212 of
these 712 allegedly stolen entries open to designations of [similarity]."
[53] This reasoning seems to me to be quite persuasive. I need, however, say no more in regard to the allege
copying of headword lists since it was common cause during argument that Media 24 no longer relied on any
claims of plagiarism in that regard.
[54] As already suggested, at the heart of Media 24's complaint in the founding affidavit was the alleged copying
of its example sentences. This copying was said to consist of two distinct components: firstly, exact copying of
the relevant example sentence (as one sees in the "spring" example referred to in paragraph 48 above) and,
secondly, substantial copying of the Pharos work but with the substitution of different words chosen by OUP.
One such example appears in the "melk" and "sap" examples already referred to above.
[55] For purposes of discussion and evaluation, Prof Dunne refers to the similarities identified by Dr Prinsloo as the
"Prinsloosimilarities" and that those of Prof Kidd as the "Kiddsimilarities". Prof Dunne draws attention to the
fact that, in the main, the OUP work has close on two example sentences per headword in comparison to
Media 24's single example sentence:
"86
I note that the number of example sentences per headword within the Oxford Skoolwoordeboek entries
appears from a cursory inspection to be plausibly close to two on average, in contrast to 2601 Afrikaans
English and 2400 EnglishAfrikaans example sentences for 1680 and 1426 B D I S entries respectively in the
Aanleerderswoordeboek, which therefore have lower averages of example sentences per headword, in the