IN THE SUPREME COURT OF APPEAL
OF SOUTH AFRICA
REPORTABLE
CASE NO 488/03
In the matter between
ROAD ACCIDENT FUND
Appellant
and
WZ MGWEBA
Respondent
________________________________________________________________________
CORAM: BRAND, VAN HEERDE N JJA et ERASMUS AJA
________________________________________________________________________
Date Heard: 11 November 2004
Delivered: 26 November 2004
Summary
: Motor vehicle collision with pedestrian – Negligence – proof of
excessive speed – Circumstantial eviden ce based on physical evidence – limits
to reconstruction by court – distinction between conjecture and acceptable
deductive reasoning – Reaction time – Stopping distance – Evidential value
of fact that witness untru thful on collateral issue.
________________________________________________________________________
J U D G M E N T
________________________________________________________________________
ERASMUS AJA
2
AR ERASMUS AJA
[1] The respondent sustained bodily injuries when he was struck by a
motor vehicle while crossing a road. He claimed damages from the Road
Accident Fund, the appellant, in the Johannesburg High Court for his losses
arising from those injuries. The tria l proceeded on the question whether the
driver of ‘the insured vehicle’ was causally negligent in relation to the
collision; and, if so, whether there was contributory negligence on the part of
the respondent. The trial court declared that the collision was caused solely
through the negligence of the ‘insured driver’. An appeal to the full bench
succeeded to the extent th at it held that the accide nt was occasioned through
the negligence of both the insured driver and the respondent, which
negligence was apportioned 80% to the driver and 20% to the respondent.
The appellant was granted special leave to appeal further to this court.
[2] At the trial there was no dispute about the physical features of the
scene of the accident. The collision occu rred at the intersection of two major
urban roadways, Columbine Ave and Rifle Range Rd. The former runs west-
east, the latter north-south. Both roads, up to the point of their intersection,
are divided down the centre by traffic islands, with traffic proceeding in
opposite directions on either side of th e islands. It is of particular relevance
that Columbine Ave has three lanes for traffic passing through the
3
intersection from west to east. (For the sake of convenient reference, I
number these lanes 1, 2, and 3 from left to right.) Lanes 1, 2 and 3 continue
beyond the intersection, with lane 3 running immediately to the left (north)
of the centre traffic island. To the east of the intersection there is a fourth
lane, to the left of lane 1, for tra ffic filtering from Rifle Range Rd into
Columbine Ave from the north. The inte rsection is controlled by traffic
lights positioned on the centre traffic is lands on all four approaches, as well
as at all four corners of the intersecti on. Pedestrian crossings are demarcated
on the perimeter of all four sides of the intersection. There was, however, no
evidence regarding distances in respect of the physical and topographical
features outlined above.
[3] In the main, the facts of th e accident were common cause. The
collision occurred at the peak tra ffic hour of about 08:15 on a Monday
morning. The insured vehicle struck th e respondent at a point in Columbine
Ave immediately east of the intersection in the lane immediately north of the
centre island (lane 3). The vehicle had travelled from west to east across the
intersection. It was driven by Ms CR Bailey.
[4] The respondent described the accide nt as follows. On the morning in
question he was vending newspapers at the intersection. He kept his supply
of papers at the foot of the pole bear ing the traffic lights on the centre island
4
on the eastern side of Columbine Ave. Immediately prior to the accident he
had sold a newspaper to a motorist stopp ed in the traffic lane for vehicles
turning left (east) out of Rifle Range Rd into Columbine Ave. After he had
completed his sale, he proceeded to retu rn to his supply of papers on the
traffic island. He walked past the ve hicle, stopped, looked at the robot and
seeing that it was green for Rifle Range Road, commenced crossing
Columbine Ave. He proceeded along the pedestrian crossing. He was
walking fast but not running. He was about to put his foot on the island,
when he heard the sound of a motor ve hicle. It struck him and he lost
consciousness.
[5] The traffic officer who attended the accident, a Mr Jacobs, testified in
the respondent’s case. He had no inde pendent recollection of the incident,
and based his evidence – which was not contested – on a report completed
by him at the time. He arrived at the sc ene at 8:20 and was informed that the
collision had occurred at 8:15. He found the vehicle and the pedestrian that
had been involved in the collision. Th e left side of the windscreen was
shattered. The alleged point of impact was pointed out to him by the driver
of the vehicle. The road surface was ha rd and ‘normal’, and dry at the time.
He observed a set of skid marks. He pa ced off certain distances relevant to
the accident and thereafter prepared a report which included a sketch
5
depicting Columbine Ave immediately to the east of the intersection. It
indicated the following points:
E: The skid marks. These comme nce in the intersection and extend
eastwards more or less straight al ong lane 3, across the pedestrian
crossing.
X: The alleged point of impact. This was where the skid marks ended.
A: The motor vehicle. This was show n as a rectangle in lane 3, to the
east of X.
P: The pedestrian. He was shown as lying in lane 3, east of A.
The following was recorded by Jacobs:
E – X = 16 paces
(length of skid marks)
X – A = 16 paces (distance between impact point and vehicle)
X – P = 22 paces (distance between impact point and pedestrian)
It would seem to follow that the distance A – P was 6 paces. This
measurement is uncertain, however, be cause it is unknown from what point
of the motor vehicle (A) the measurement X – A was taken.
[6] Jacobs paced off the distances hims elf. A pace is of course a relative
measurement, and he was asked to indi cate the length of his stride. This was
agreed upon as being ‘just a bit more than half a metre’. On that basis, the
skid marks of 16 paces on the diagram were 9 to 10 metres in length.
6
[7] The evidence of the further w itness for the respondent can be
disregarded for reasons that are not relevant to adjudication of the appeal.
[8] The first witness for the appe llant was a police inspector, one
Madocks. He had stopped his vehicle in Rifle Range Rd north of the
intersection, because the traffic lights were red for him. As he was waiting
he heard the screeching of tyres. When he looked in that direction he saw
that a white Toyota had hit a newspape r vendor. He put on his police light
and cut across the flow of traffic in Co lumbine Ave to get to the accident. It
was clear from the evidence of this witness that, at the time of the collision,
the traffic lights were red for Rifl e Range Rd and therefore green for
Columbine Ave, and remained so for some seconds after the collision.
[9] Ms Bailey told the court that she was driving to work that morning in
an easterly direction in Columbine Av e. Her speed was between 50 to 60
kph. As she approached the Rifle Range Rd intersection, she was travelling
in the far right-hand lane (lane 3). The traffic lights were green in her favour.
As she entered the intersection, she observed two pedestrians running across
the road, the one behind the other. (The one in the rear proved to be the
respondent.) They were not in the pede strian crossing. When first she saw
the respondent, he was ‘maybe four car lengths’ from her. She could not say
how many metres that was. He was in th e far left hand lane (lane 1, it would
7
seem). She was still travelling at betw een 50 to 60 kph. She applied her
brakes strongly. The first pedestrian crossed the road successfully, but the
respondent ran into the left side of her vehicle. He fell onto the windscreen.
The car came to a dead stop as she st ill had her foot on the brake. Her
windscreen was damaged in the collision.
[10] Bailey stated that the accide nt occurred between 07:30 and 07:45,
because that was the normal time that she went past there. In cross-
examination, she was confronted with her police statement in which it was
recorded that the accident had happened at approximately 08:15.
[11] Counsel for respondent put it to her (without factual foundation for
the proposition) that the length of the substantial skid marks indicated that
she was not travelling at 50-60 kph. She replied:
‘Sir there is no ways that you can travel faster than 60 in that plac e at peak period, in
peak traffic, you cannot. There is no ways th at you can travel faster than that. I drive
there every day of my life, I have worked there for eight ye ars, eight solid years I drive
there every day of my life. … I always look at my speedometer when I drive that I do not
exceed speed limits. Because I travel in a company car and if I get any fines for the
company car I am liable for the fines.’
8
Counsel put it to her (again without f actual basis for the proposition) that
according to the diagram the pedestrian was ‘flung’ 22 paces from the point
of impact. The cross-examination continued:
‘… Can you explain why he was flung so far on collision? --- If I may ask how do you
mean by flung? ---Well possibly not flung, he fi nished up – perhaps flung is the wrong –
according to the diagram he finished up 22 paces down the road from the point of
collision. --- He rolled off my car, he was not flung nowhere, he rolled off my car. You
say that you carried him on your bonnet? --- On the impact of the accident when my car
came to a stop he rolled off the car.’
[12] On this evidence then, the trial court ruled that Bailey was solely to
blame for the collision. His reasoning is reflected in the following
comments:
‘I must also point out that the length of the skid marks as well as the distance from the
point of impact up to where the plaintiff fell must mean that Bailie was travelling at a far
greater speed than 50/6 0 km. It is simple logic that if indeed Bailey was travelling at
between 50 and 60 km per hour the vehicle woul d not have made skid marks that long
and could not have necessitated that the plaintiff land some 22 paces away from the point
of impact.
Based on Madock’s testimony, it appears more probable that the robot was green for
Baillie as she approached the intersection i.e. for vehicles travelling west to east in
Columbine Avenue. Probabilities point to th e robot having changed from green to amber
9
before Bailey entered the intersection. Because she was in a hurry she clearly must have
tried to cross the intersection as fast as po ssible before the robot turned red. … When the
robot changed from green to amber for Bailie at the same time the plaintiff must have
seen the robot change to green for him and st arted crossing the intersection from north to
south. …
In my view Bailie entered a busy intersec tion at high speed and when the robot had
already changed from green.’
The learned judge found that Bailey must have been rushing in order to
report for work on time.
[13] The court a quo held that the trial judge had erred in his finding that
the traffic lights were red for Bailey a nd green for the respondent. The court
accordingly held that the respondent wa s negligent in crossing the road. The
court, further, held that the trial cour t was correct in its finding that Bailey
(too) was negligent in the driving the insured vehicle. The learned judge
who delivered the judgment of the court reasoned that -
‘… the cumulative weight of the following factors:
(i) The distance from the traffic lights (whi ch were green for the insured driver) to
the pedestrian crossing used by the plaintiff at the other end of the intersection;
(ii) The length of the skid marks of the insured vehicle;
(iii) The distance between the point of imp act with the pedestrian and the point where
he fell;
10
(iv) The insured driver’s untruthful te stimony about the time of the accident; and
(v) The insured driver would have been seriously late for work.
compel the conclusion that the insured driver was travelling at an excessive speed in the
circumstances. Put differently, had she been travelling at between 50 and 60 kilometres
per hour, as she said she had, she would easily have been able to st op before hitting the
plaintiff. … Had she travelled more slowly she would also have a voided colliding with
the plaintiff. Therein lies her negligence. An examination of her evidence also reveals
that there were aspects of the failure to k eep a proper look out an d the failure to take
reasonable steps to avoid the collision in her negligence as well. It should be borne in
mind that she herself said that there was no vehicle in her lane be hind her. Accordingly
she could have applied her brakes sooner than she did.’
[14] Bailey’s evidence as to her sp eed was not controverted by direct
evidence. Both the tria l court and the court a quo relied on circumstantial
evidence for rejecting her version th at she was travelling at 50 to 60 kph
when she entered the intersection. In my view, both courts, with respect,
failed properly to analyse the evidencial material.
[15] I deal first with point (iii) of the court a quo: the distance between the
point of impact and the point where the pedestrian fell. Both the trial court
and the court a quo simply mentioned that th e distance between these two
points was 22 paces, without further comment or elaboration. Both
apparently assumed that the force of the impact somehow caused the
respondent to be propelle d this distance. This a ssumption does not however
11
accord with the evidence . On Bailey’s evidence ( see para [11] above) the
respondent was conveyed on the bonnet of the motor car for some distance
beyond the point of collision: the vehi cle was decelerating and came to a
stop; the respondent was a free-moving body that continued on its eastwards
course beyond and over the front of the vehicle, until his contact with the
road surface brought him to a stop in front of the car. This description
accords with the shattered windscreen and the relative positions of the motor
car and the pedestrian after the collis ion, as indicated by Jacobs. On his
report the distance between the motor vehicle and the pedestrian was 6 paces
(but see para [5] above). No thing in this scene conflicts with Bailey’s
evidence regarding her speed of travel . It follows that the trial court
misconceived the evidence on th is aspect, and the court a quo uncritically
adopted that misconception.
[16] I deal next with the skid marks ( point (ii)). The trial court, as well as
the full court, simply assumed that the lengt h of these marks constituted
evidence that Bailey was travelling at a speed in excess of 50-60 kph. In
doing so, with respect, they indulged in accident reconstruction without the
benefit of expert evidence. Accident re construction is a branch of dynamics
requiring special knowledge in the discipline of physics. A court may
venture into that field but only at a le vel that can properly be said to be a
12
matter of common sense falling within common human experience. It is
often a fine line (and frequently a deba table one) that separates unacceptable
conjecture from acceptable deductive reasoning based on proven physical
facts.
[17] A court could, in my view, take cognizance of the fact that skid marks
are caused by the wheels of the motor ve hicle locking in the application of
its brakes. It is, further, a matter of basic and obvious logic that there is some
correlation between the length of the marks and the speed at which the
vehicle was travelling when the skidding commenced: the higher the
velocity, the longer the marks will be. However, calculating the speed of the
vehicle from the length of the skid marks is beyond the ability of the non-
expert that is the court. That calcu lation will require evidence regarding the
stopping distance of the particular vehicl e at a given speed in the particular
physical circumstances. In this case, there was no evidence regarding the
facts relevant to that computation, no r of the mathematical formula based on
such information. In the circumstances, the fact that the insured vehicle left
skid marks of 9-10 meters does not have the precise probative value placed
upon it by the trial court; and the court a quo followed it into that error.
[18] Both the trial court and the full court found that, had Bailey been
travelling between 50-60 kph, she would have been able to stop her vehicle
13
before striking the appellant. Neither court, however, examined the factual
basis underlying this reasoning. An essential factor in such deductive
process is distance, here the distan ce between Bailey a nd the respondent
when she first saw him. That meas urement was, however, unknown. The
court a quo described this distance as ‘considerable’, which was an
imprecise measure on which to base the finding that Bailey could have
stopped her vehicle timeously.
[19] Furthermore, we are dealing with two moving objects converging at
right angles on a point of collision. Ba iley had to take evasive action. The
phenomenon of reaction time is frequen tly mentioned in motor accident
cases. In the above crisis situation, it would be an important factor in
assessing whether Bailey coul d have avoided the collis ion. However, in the
absence of physiological or empirical evidence regarding human reflexes, a
court should be hesitant to attribute a pr ecise time to that factor: in short,
Bailey’s reflex ability was unknown. Nevertheless, the court could and
should have had regard to what the re action process logically entails. In an
emergency, the motorist must first obser ve and then assess the nature of the
looming danger. She must thereafter decide upon the proper evasive action
to take. In Bailey’s case this in volved a choice between swerving and
braking. She decided upon the latter (whi ch appears to have been the most
14
appropriate action in the circumstances). She then had to take that action.
This involved removing her foot from the accelerator and thereafter
depressing the brake pedal. All this would of course happen very quickly:
but then we are here concerned w ith time measured in seconds, even
fractions of a second. In the meanti me, prior to the deceleration, Bailey
would, at 60 kph, have been pro ceeding at 16,7 metres per second,
decreasing at an unknown rate as the ve hicle decelerated. It was furthermore
unclear at what point the braking ac tion took effect, as this could have
occurred prior to the wheels locking.
[20] In view of the lack of precise information regarding Bailey’s stopping
distance and her reaction time, there wa s insufficient factual basis for the
conclusion that the reasonably competen t driver in her position would have
avoided the collision by timeous reaction and appropriate action.
[21] One can look at the position al so from the perspective of the
pedestrian’s movement. The respondent had sold a newspaper to a motorist
stopped at the corner of Rifle Range Rd and Columbine Ave. He was
standing – so it seems – in the filter la ne in Columbine Ave, presumably at
the driver’s window ( see paras [2] and [4] above ). From there he moved
towards the centre island. When Baile y first observed him (on her evidence)
he was in the extreme left-hand lane of Columbine Ave. He was running
15
across the road (according to her) or walking fast (according to him). He
proceeded as far as lane 3, wher e he was struck. The distance across
Columbine Ave was not measured, nor do we know precisely how long it
took him to traverse the road: it would have been a matter of seconds (on
either version of his actions). Clearly , there was no factual basis for holding
that Bailey could, let alone should, have brought her vehicle to a stop in that
time.
[22] That leaves the f act that Bailey was untruth ful about the time of the
collision and was in fact late for her wo rk (points (iv) and (v)). This aspect
impacts adversely upon her credibility. However, her evidence regarding the
collision was consonant with the estab lished facts. There was no evidence –
either direct or circumstantial – that in any way contradicted her version of
the accident. The fact th at she had a motive for travelling fast did not per se
constitute evidence that she was trave lling beyond the speed limit. By itself,
the fact that she was untruthful about being late for wo rk was devoid of
probative value.
[23] Finally, there is a suggesti on in the judgment of the court a quo that
Bailey failed to keep a proper look-out . The judge did not elaborate on the
issue. There is no factual basis for such a finding. On the respondent’s
evidence, there was nothing in his actions that would have suggested to an
16
onlooker that he was about to rush blindly against the red light across the
busy road, into the path of oncoming tr affic. In the circumstances, a careful
motorist would not have had reason to take evasive action prior to the
respondent commencing to cross Colu mbine Ave, which was when Bailey
noticed him.
[24] In view of my finding s on the question of negligence on the part of the
insured driver, I need not d eal with the further substa ntial hurdle in the path
of the appellant, ie causation.
[25] For these reasons, I find that both the trial court and the court a quo
erred in finding negligence on the part of the driver of the insured vehicle.
The trial court should ha ve found that the responde nt had failed to prove
negligence on the part of the driver of the insured vehicle, and should
therefore have granted the appellant absolution from the instance.
[26] In the result, the appeal succee ds with costs. The judgment of the
court a quo is set aside and the following order is substituted therefor:
‘1. The appeal succeeds with costs.
2. The whole of the order of the trial court is set aside, and
substituted by the order that the de fendant be absolved from the
instance with costs.’
17
____________________
AR ERASMUS
ACTING JUDGE OF APPEAL
CONCUR:
BRAND JA
VAN HEERDEN JA