legal cases are those of probability”.
566
To put is simply, Step 2 is obviously
and inherently probabilistic.
567
Thus probability is
part of proof of facts because it deals with, and explains,
uncertain knowledge. In finding facts a court does not normally believe that it
has found the facts of a case for certain. Instead it finds that it believes the
facts to some level of probability
or degree of certainty. This is the position
because, as a matter of law, to find that a party has proved their case the court
has to be satisfied only that the party has proved their case to a sufficient
degree of certainty. This degree of certainty or level of probability is
determined by the rules that lay down the standard of proof for the case. In a
civil case these require a court to find facts on the balance of probability and
in a criminal case the court has to find facts proved beyond reasonable doubt.
With some specific decisions there may another standard laid down by statute
-
for example that the decision maker is “satisfied” or “reasonably certain” as
to the facts.
Ideally when a court estimates a probability it would give it a numerical
measure. However, in any situation there are often problems in putting a figure
on a probability with any sort of confidence. And with evidence there are
additional problems because considerations “for and against particular
findings of fact” generally involve
“a global assessment of a whole complex
array of matters”.
568
This has led, not surprisingly, to the view that it is
misleading to try to give these probabilities “numerical expression”.
569
In most cases putting a numerical measure on probability is an estimate at best
and probably no more than a guess.
570
Nevertheless, referring to numerical
probabilities rather than using mere verbal formulas is a simple and effective
way of explaining the uncertainty of fact finding and the reasoning that
underlies the task.
In fact it brings three specific advantages. First, it makes the position much
clearer. A probability of 75% is more telling than being “fairly certain”.
Second, it makes for easier comparison. We can see the large gap between a
566
Robertson and Vignaux (1993) p 478
567
See, for example,
Victor v Nebraska 511 US 1 (1994) 14, 22 per O’Connor J;
Ginsburg J concurred at p 25 while Blackmun J dissented at p 36. See also Re
Winship 397 US 358 (1970), p 370 per Harlan J.
568
Hodgson (1995) p 738
569
Hodgson (1995) p 738
570
To reinforce the point, in Briginshaw v Briginshaw (1938) 60 CLR 336 at 361-
362 the High Court said that there can be “indefinite gradations of certainty” that
the facts portrayed by evidence are true.
 |
 |
 |