happens, though, that there are a number of different conceptions of probability
and that there is disagreement about precisely what probability means in the legal
context.2This is not, however, a serious stumbling block for present purposes:
whether forensic probability is cardinal or ordinal, objective or subjective,
Pascalian or Baconian there is little doubt that it comes in degrees. This is certainly
the accepted understanding of the difference between the criminal and civil
standard of proof: in order to win a criminal case, the prosecution must prove its
case to a higher degree of probability than need the plaintiff in a civil case. Though
nothing much will turn on this, it will be simplest to interpret the concept of
probability in law as requiring the fact-finder to have a certain degree of belief in a
case before she accepts that it has been proved. Degrees of belief vary along a scale
between 0 and 1; a person who has a degree of belief of 0.1 in a proposition
believes that that proposition is unlikely to be true, whereas if her degree of belief
is 0.99 she regards it as almost certainly being true.
The civil standard of proof has been described in a number of different ways.
One of the clearest explanations has been offered by Lord Denning: ‘[i]f the
evidence is such that the tribunal can say ‘‘we think it more probable than not’’,
then the burden is discharged, but if the probabilities are equal it is not’.3Other
formulations use expressions such as ‘preponderance of probability’, ‘preponder-
ance of the evidence’ and ‘balance of probabilities’. The second of these is the
favoured term in the United States, while the third is the usual English formulation.
All seem to imply the same thing: that the plaintiff’s case need not be very
probable, but that it needs to be more probable than its negation. This suggests that
the best interpretation is that a degree of belief of just over 0.5 is required.4The
only phrase which might be interpreted differently is the expression ‘preponder-
ance of the evidence’, which might be thought to imply that the plaintiff needs to
produce more evidence than the defendant, but this interpretation does not appear
to have any current support. Indeed, it would be odd if it did, because it would
imply that there was some way of quantifying evidence, as opposed to the effect
that evidence has on belief.
2 See Colin Howson, ‘Theories of Probability’ (1995) 46 British Journal for the Philosophy of Science
1; Philip Dawid, ‘Appendix: Probability and Proof’ in Terence Anderson and William Twining,
Analysis of Evidence (London: Weidenfeld and Nicolson, 1991) 404–416; William Twining,
‘Debating Probabilities’ (1980) 2 Liverpool L Rev 51.
3Miller vMinister of Pensions  3 All ER 372, 373–374. Although Lord Denning states that it is
‘well settled’ that this is a correct description of the civil standard of proof, this passage has become
the standard citation for the assertion that the civil standard of proof is proof on the balance of
probabilities; before this, the law does not seem to have been clear and it is difficult to say when the
balance of probabilities became the dominant concept. Although we find the concept of founding a
verdict ‘upon that which is most probable’ as early as 1571 (Newis vLark Plowd 403, 412), we should
be wary of reading too much into statements like this one, made at a time when ‘probability’ may
have had a rather different meaning than it does today. Newis certainly did not settle the civil
standard: in 1858 we find the House of Lords, while declaring that ‘in civil cases the preponderance of
probability may constitute sufficient ground for a verdict’, also noting that ‘a difference of opinion
exists on this question’ (Cooper vSlade (1858) 6 HL Cas 746, 772). My suspicion is that, because the
civil standard of proof may require a more sophisticated notion of probability than does the criminal
standard, the modern concept only became established at a fairly late date. Although Barbara J.
Shapiro’s Beyond Reasonable Doubt and Probable Cause: Historical Perspectives on the Anglo-
American Law of Evidence (Berkeley: University of California Press, 1991) concentrates on the
development of the criminal standard, some of the sources she cites suggest that the courts did not
differentiate between proof in civil and criminal cases until well into the 19th century (see eg, ibid,
27, 38). Indeed, in this century we find judges professing not to understand how there can be more
than one standard of proof (eg Lord Goddard in RvHepworth and Fearnley  2 QB 600, 603).
4 It might be argued that ‘preponderance of probability’ leaves open the possibility that the plaintiff’s
case must be substantially more probable than its negation. But if this is so, the phrase is too vague to
serve as a description of a standard of proof.
The Modern Law Review [Vol. 62
168 ßThe Modern Law Review Limited 1999