[18]
I am prepared to accept, without deciding, that, despite her denials, Ms Martin probably did make the
statements to Sergeant Moolman and that she was probably telling the truth when she did so. Untruthfulness, however, is not the only
danger. The other danger is that she might have been mistaken. Particularly with reference to identification evidence, the danger
of mistake has been underscored by our courts again and again (see eg S v Mthetwa 1972 (3) SA 766 (A) at 768; S v Charzen 2006 (2) SACR 143 (SCA) para 11 at 147i-j). By its very nature, hearsay evidence cannot be tested in cross-examination. The possibility
of mistake can therefore not be excluded in this way. The result is, in my view, that hearsay evidence of identification can only
be admitted if the possibility of mistake can be safely excluded in some other way, eg with reference to objectively established
facts.
[19]
In this matter there is no way to test the accuracy of the observations Ms Martin deposed to in her statements.
On the contrary, according to her testimony in court it would, as a result of physical obstructions impeding her view, be virtually
impossible for her to make those observations from her garden where she stood. It is true, of course, that at that stage she was
trying her utmost to distance herself from the contents of the statements. Nonetheless, her evidence about the physical obstructions
remained uncontested. In the circumstances the identification evidence deposed to by Ms Martin in her statements appears to be of
the most unreliable kind. For these reasons we should not, in my view, accede to the state's request to admit these hearsay statements
under the provisions of s 3(1)((c) of the Act.
[20]
The next question is whether the evidence of Ramakgula, on its own, is sufficient to justify the appellant's
conviction. The court a quo found Ramakgula an honest witness. I have no reason to doubt the correctness of that finding. However, the danger that again looms
large, is the possibility of mistaken identification. The court a quo found reassurance in the fact that the witness had sufficient opportunity to make his observations in that he was looking directly
at the assailant when he fired the shot. This reassuring factor is, however, diluted to a material extent by the contents of two
statements which Ramakgula made to the police. According to these statements he told the police that both the assailant and his companion
were armed with firearms and that the companion was pointing a firearm at him when the killer shot the deceased. Although Ramakgula
distanced himself from these statements in evidence, it is difficult to conceive why the police would fabricate this version. It
almost goes without saying that if this version is to be accepted, Ramakgula's opportunity of observing the killer would be materially
reduced.
[21]
The court a quo also found reassurance in the fact that Ramakgula had seen the appellant on about four occasions prior to the incident. The problem
is, however, that on Ramakgula's own version he had heard the assailant being identified by Ms Martin shortly after the incident.
Although he could not remember the name that Ms Martin mentioned, it appears from the context that she most probably mentioned the
name of the appellant. Apart from the inherent danger of suggestion, any mistaken identification by Ms Martin would thus have poisoned
the evidence of Ramakgula as well. Additional support for the notion that Ramakgula's identification of the appellant may be the
result of suggestion, seems to derive from his own evidence that Moolman provided him with some description of the appellant long
before he testified in court. Confidence in Ramakgula as a witness is further diminished by the fact that, in a statement to Moolman,
Ramakgula referred to the assailant and his companion as 'two black men unknown to me'. In cross-examination Ramakgula ascribed this
to a misunderstanding between him and Moolman. But according to Moolman's testimony, Ramakgula was indeed unable to give a description
of the assailant 'because of the fear he was under'. As to how Ramakgula was then able to identify and describe the appellant at
a later stage, Moolman volunteered the following solution:
'[M]aybe that which he said in court is based on what he gathered from Evaton . . . on that which Ms Bessie [Martin] told him.'
[22]
In the light of all these difficulties, it is in my view self-evident that the appellant cannot be convicted
solely on the basis of Ramakgula's testimony. Lastly there is the appellant's mendacity as a witness. Though false denials by an
accused person will often strengthen the state's case, it cannot serve as the sole basis for conviction. It too often happens that
innocent persons cannot resist the temptation of putting as great a distance as possible between themselves and criminal offences,
even by deceitful means.
[23]
For these reasons:
The appeal is upheld and the convictions and sentences are set aside.
.......................
F D J BRAND
JUDGE OF APPEAL
Concur:
VAN HEERDEN JA
THERON AJA
SAFLII:
|
Terms of Use
|
Feedback
URL: http://www.saflii.org/za/cases/ZASCA/2007/58.html