This is a really interesting article by a Supreme Court Justice in Australia about how our fact-finding processes are not up to date with current understandings in the field of psychology, etc. Wish more judges were this enlightened.
HE: I can remember everything as if it were yesterday. We met at nine.
SHE: We met at eight.
HE: I was on time.
SHE: No, you were late.
HE: Ah yes, I remember it well. We dined with friends.
SHE: We dined alone.
HE: A tenor sang.
SHE: A baritone.
HE: Ah yes, I remember it well. That dazzling moon …
SHE: There was none that night. And the month was June. 
This lyric illustrates the unreliability of human memory as well as the difficulties attendant on making findings of fact.
Fact-finding is labour on the factory floor of the judicial system. It is not glamorous work. Judgments on fact go unreported; they have no enduring fame. Nevertheless, justice depends on correct factual findings, and a fundamental measure of a legal system is the accuracy and skill with which facts are found. It is curious how little attention has been given to this facet of the judicial task . We accept with a knowing smile the virtual impossibility of remembering with accuracy past events in the ordinary course of life, such as those about which Maurice Chevalier and Leslie Caron so nostalgically sang. Nevertheless, the justice system has, as one of its basic foundations, the assumption that witnesses are capable of accurately describing events that took place years ago, and that judges can reliably tell what evidence is true and what is false.
Witnesses may be deliberately untruthful, but there may be many other causes of inaccuracy in a witness’s testimony. These include imperfect observation, faulty memory, an over-active imagination, emotional disturbances, self-interest, and other biases. Witnesses may be dishonest about only parts of their evidence. Little is more deceitful than half the truth. Judges must take care to differentiate between a witness that is being honest or deceitful, and a truthful witness who is giving accurate or inaccurate testimony. Distinguishing truth from intentional deceit is a different exercise from distinguishing between true and false memory. A liar intends to be deceptive, whereas the faulty rememberer tries to be truthful. 
Discerning what is accurate or inaccurate in the testimony of a truthful witness may be the most difficult task of all. Witnesses sympathetic to a litigant may, unconsciously, give a strong colouring to the facts, and remember things that were not seen, and misrepresent things that were seen. The renowned American judge, Jerome Frank, told of a trial judge who had to decide on the disposition of the family home by the deceased, Mr Humphrey. After hearing the testimony and the lawyers’ arguments, the judge announced:
“Gentlemen, provided Mr Humphrey said – in the light of these Missouri decisions – ‘daughter, if you’ll come and live with me, I’ll give you this house,’ then I’ll decide for the plaintiff”.
The court reporter, unfortunately, could not read his notes recording the crucial evidence. The judge accordingly asked the principal witness, the daughter’s maid, to take the stand and repeat her testimony. This is what she said:
“I remember very well what happened. It was a cold and stormy night. We were all sitting around the fire. Old Mr Humphrey said to his daughter: ‘Daughter, in the light of the Missouri decisions, if you’ll come and live with me, I’ll give you this house’”. 
Many distinguished judges have expressed deep scepticism about oral evidence. Apart from the telling of deliberate lies, observation and memory are often fallible, and humans have an unlimited capacity for honestly believing something that bears no relation to what really happened.
Recent psychological research has established that the more episodes of a certain type we experience, the harder it becomes to distinguish among them. On the other hand, repetition of episodes strengthens the overall memory for the entire class of event. A person who suffered many beatings as a child will find it difficult to recall details from any particular attack – unless something unusual occurred during it. But the person will never forget what it was like to be subjected to such violence. This suggests that the usual forensic technique of discrediting a witness by focusing on the peripheral detail of the crucial event may be unreliable and unfair.
Recent psychological research has also established that high levels of stress do not produce general impairments in memory. Emotional stress enhances memory for the central features of the stressful experience. An extremely stressed person, however, will remember central aspects of the experience but will fail to remember trivial detail. In the same way, distinctive, emotional stimuli often impair memory at the expense of other stimuli . Thus, a person experiencing a traumatic event will remember the main details of the event far better than peripheral details. This explains why robbery victims often remember details of the assailant’s weapon but not much about his appearance.
In one particular study, eyewitnesses who correctly identified, in a photo line-up, the person who stole a calculator in a staged theft, remembered fewer peripheral details about the event than did those who wrongly identified an innocent person as the culprit. In a worrisome secondary finding, because the mistaken witnesses provided many peripheral details about the ‘crime’, mock jurors regarded them as more credible than the witnesses who correctly identified the thief’s identity at the expense of peripheral detail.
Many experiments have shown that information provided to witnesses after an event affects how they later remember it. Studies have shown that persons who fall prey to misleading information consciously remember witnessing things that they have not seen. They hold these false memories with great confidence . An example of this is an occasion when an El Al cargo plane crashed into an apartment building near Amsterdam. One hundred and ninety three persons were asked whether they had seen television footage of the plane striking the building. In reality the crash had not been captured on film. Nevertheless, 55% claimed to have seen it on television. Two-thirds of a group of law students claimed to have seen this crash footage and some even provided details about what they had seen.
Researchers found that 44% of persons asked claimed to have seen a non-existent film of the crash that took the life of Princess Diana and were capable of providing details about their false memory. In 1998, a study revealed that nearly 30% of the members of the American Ex-POW Association had never been prisoners of war .
In a widely praised work, Remembering Trauma, Prof Richard J McNally, a professor of psychology at Harvard University, notes that a significant minority of people can unwittingly come to believe that they experience stressful events that never happened . To the casual observer, the notion that someone might remember having experienced a trauma that never actually occurred seems absurd. Yet people remember atrocities that never happened and experience emotional distress congruent with belief in their authenticity. Prof McNally concludes: “Suggestive, misleading information provided after an event routinely distorts witnesses’ memory reports of the event” . Prof Schacter, the chair of Harvard University’s Department of Psychology, has described recent psychological research showing that memory malfunction and suggestibility can result in a person mistaking fantasy for reality . He, too, asserts: “people can develop detailed and strongly held recollections of complex events that never occurred” .
Judges often disbelieve witnesses who have reconstructed their evidence. But, recent research has established that, unlike what most people believe, recollection is a reconstructive, not a reproductive, process. Recalling one’s past is not replaying a videotape of one’s life and working memory. When we remember an event from our past, we “reconstruct it from encoded elements distributed throughout the brain” . Accordingly, to admit that one’s testimony is reconstructed is no more than to admit a facet of human nature; all memory is reconstruction. The mere fact of reconstruction should not be a reason for rejecting evidence.
According to psychologists, internal contradictions and unsatisfactory evidence may be caused by the stress of the court atmosphere and the hostile demeanour of the advocate. Age can affect the confidence of witnesses. Young people are usually more confident witnesses than the middle aged and elderly. There are truly many pitfalls in deciding on the veracity of evidence.
Nevertheless, even though the traditional approach is by no means infallible, my observation after more than 40 years experience in the field, is that the time spent over many years observing evidence being led, and witnesses being questioned, is of great assistance to a trier of fact. One cannot help but develop antennae sensitive to deliberateuntruths. Although truthful but inaccurate evidence remains extremely difficult to detect, it is beneficial to be aware of this painful fact from one’s own practice of the law. The recent movement to broaden the reservoir of judicial appointments means that more and more judges will be devoid of this experience. There have undoubtedly been outstanding judges who have not come from busy practices at the Bar. But, as a general rule, the appointment of persons who have not spent many years observing the behaviour of witnesses in court will not do much for reliability in fact-finding. This is a problem that needs to be recognised.
The difference between success in life and ruin may turn on a single demeanour finding which another judge may not have made, or the same judge may not have made on a different day . Demeanour findings are often idiosyncratic and unpredictable.
All of us know that the way persons behave when they tell stories can often be helpful in determining whether they are being truthful or evasive. A person’s intonations, fidgeting, composure, yawns, facial and body movements, the use of eyes, hands, air of candour or of evasiveness, even complexion, may furnish valuable clues . But it is so difficult to put the right construction on them. An experienced English judge, MacKenna J, has put the problem with precision :
“I doubt my own ability, and sometimes that of other judges, to discern from a witness’s demeanour, or the tone of his voice, whether he is telling the truth. He speaks hesitantly. Is it the mark of a cautious man, whose statements are for that reason to be respected, or is he taking time to fabricate? Is the emphatic witness putting on an act to deceive me, or is speaking from the fullness of his heart, knowing that he is right? Is he likely to be more truthful if he looks me straight in the face than if he casts his eyes on the ground, perhaps from shyness or a natural timidity?”
Understandably McKenna J said that he relied on those considerations as little as he could help.
Cultural and ethnic differences cause serious difficulties to judges making demeanour findings. Sir Thomas Bingham (as he then was) illustrated these clearly when he said :
“[H]owever little insight a judge may gain from the demeanour of a witness of his own nationality when giving evidence, he must gain even less when … the witness belongs to some other nationality and is giving evidence either in English as his second or third language, or through an interpreter. Such matters as inflection become wholly irrelevant; delivery and hesitancy scarcely less so. … If a Turk shows signs of anger when accused of lying, is that to be interpreted as the bluster of a man caught out in a deceit or the reaction of an honest man to an insult? If a Greek, similarly challenged, becomes rhetorical and voluble and offers to swear to the truth of what he has said on the lives of his children, what (if any) significance should be attached to that? If a Japanese witness, accused of forging a document, becomes sullen, resentful and hostile, does this suggest that he has done so or that he has not? I can only ask these questions. I cannot answer them. And if the answer be given that it all depends on the impression made by the particular witness in the particular case that is in my view no answer. The enigma usually remains. To rely on demeanour is in most cases to attach importance to deviations from a norm when there is in truth no norm.”
A finding based on demeanour is essentially an intuitive exercise. The Oxford English Dictionary defines “intuition” relevantly as “The immediate apprehension of an object by the mind without the intervention of any reasoning process”. It is the absence of rationality that makes demeanour findings so arbitrary, so ephemeral, so uncertain, so personal and subjective, so susceptible to sub-conscious prejudice, so susceptible to error.
There are some cases where decisions have to be made on the basis of demeanour. A collision at a controlled intersection with similar damage to the vehicles and no witnesses is an example. But there are many cases that could be decided on the probabilities where judges choose to rely on demeanour. As Lord Rodger has said: “A weak judge will often try to cut off any appeal from his decision by structuring his opinion so that it is made to rest on his unappealable assessment of the witnesses” .
Cognitive illusions on the part of judges impair accurate fact-finding. Hindsight illusion can lead a judge to over-estimate what could have been foreseen by others and to regard what actually happened as inevitable. Overconfidence on the part of judges leads to illusions about the value and accuracy of their own judgment. What psychologists call “false-consensus bias” leads some to view their own behaviour and responses as typical and appropriate, and different behaviour as odd and inappropriate. Research has demonstrated that the way in which a question is framed can significantly affect the answer (that is, from truthful witnesses) and judges need to be aware of this.
All these illusions are capable of erroneously, and profoundly, influencing a judge’s opinion as to the veracity of evidence. Judge Frank said, aptly: “There can be no greater hindrance to the growth of rationality than the illusion that one is rational, when one is the dupe of illusion”  .
Another aspect of fact-finding that needs to be examined is the inevitable existence of prejudices and biases, or tendencies towards such attitudes and feelings, in each and every judge.
We have all encountered judges who do not transgress the boundaries of apprehended bias, but who appear to favour or disapprove of plaintiffs, defendants, landlords, tenants, women, black persons, immigrants, workers, employers, police, government bodies, and so on . Other judges are more disciplined and control themselves so that no partiality can outwardly be detected. But that does not mean that they do not have feelings, both conscious and sub-conscious, that are capable of influencing their decisions.
In the controversial Canadian case of R v RDS the trial judge disbelieved a police officer, saying that in her experience white police in Halifax lied when giving evidence against black youths. Four judges of the Canadian Supreme Court held that this was an acceptable approach and it was quite in order that the differing life experiences of judges be used to assist them in the decision-making process . As a general proposition this is debatable. For my part it is a licence to give free reign to personal prejudice. Moreover, during a trial, parties would not be informed of the judge’s personal experiences and beliefs and would not be able to combat what might be tendencies to prejudice and bias.
In some ways, RDS raises issues of judicial notice . The established rule is that a court may judicially notice a fact whenever it “is so generally known that every ordinary person may be reasonably presumed to be aware of it” . A judge’s life experience may cause him or her to form assumptions about certain groups of people that are not held by all. There are great dangers in relying on assumptions about the behavioural characteristics of particular groups of people. Judges are generally required to act on evidence actually adduced, and should be conservative about taking judicial notice of matters of supposed notoriety. Judges should make every effort to suppress feelings and attitudes that are neither impartial nor neutral, when deciding whether a witness is telling the truth.
An inter-disciplinary approach
I suggest that there is a need for appropriate self-understanding and self-knowledge on the part of judges so as to combat the harmful affect of cognitive illusions and judicial bias. Some form of education for judges in this area would not go amiss. Without a proper understanding of self, judges cannot be aware of all their prejudices or be able to guard against them .
As recent advances that psychologists have made in understanding memory become better known, judges (and, I suspect, barristers) will be expected to be cognisant of the psychological factors that influence memory and perception (and, if possible, of signs that those factors are inappropriately affecting a witness).
There is no organised attempt by judges and lawyers to learn about the developing techniques in understanding memory, and how it works, and the factual illusions that beset so many. There is no training in exploring the mind. Other disciplines, including economic, political, and medical, have made considerable advances in studying these cognitive illusions . I suggest that accurate judicial fact-finding requires an understanding of them and this may require an interdisciplinary approach. Although continuing judicial education is the order of the day, fact-finding is not subject to the rigor of scientific analysis. Judges are left to apply common sense or intuition. This is a highly subjective platform and leaves considerable scope for irrational decisions that are difficult to overturn on appeal.
What then are the techniques that a careful judge should apply in deciding which witness or what version should be believed?
Apart from developing self-awareness, I suggest that judges should focus on probabilities and inconsistencies, rather than demeanour. Close attention should be paid to contemporaneous documents. Where oral testimony is in conflict with contemporaneous documents, it may be unreasonable to believe what the witness says. The probabilities and consistency of the witness’s version should be measured against the incontrovertible and agreed facts as well as the remainder of the witness’s evidence (including that given on other occasions). Regard should then be had to the other facts found. Other facts relating to credit may be relevant. But the probabilities, together with external and internal consistency, should always be the touchstone of factual findings. Finally, attention must be given to the demeanour of the witness. I would consign demeanour to the bottom of the list .
Where there are differences between expert witnesses that are capable of being resolved rationally by examination and analysis, and there is no suggestion or dishonesty or undue partisanship, a decision based solely on demeanour will not provide the losing party with a satisfactory explanation for his or her lack of success. A justifiable grievance as to the way in which justice was administered will then arise. A coherent reasoned opinion expressed by a suitably qualified expert should be the subject of a coherent reasoned rebuttal and judges must then enter into the issues canvassed before them and explain why they prefer one case to the other .
The approach of appellate courts
The rule in New Zealand is that an appellate court will only interfere with the trial judge’s findings of fact in exceptional circumstances. The traditional view is entrenched, namely, an appellate court should not reverse the decision of a trial judge on a question of fact unless that decision is shown to be wrong. The fact that the trial judge hears and sees the witnesses is regarded as being of paramount importance .
In Australia, the approach has been the same but there are signs of a more liberal approach. In Fox v Percy Gleeson CJ, Gummow and Kirby JJ emphasised the dangers of too readily drawing conclusions about truthfulness and reliability solely or mainly from the appearance of witnesses . They referred to remarks made by Atkin LJ : “I think that an ounce of intrinsic merit or demerit in the evidence, that is to say, the value of the comparison of evidence with known facts, is worth pounds of demeanour.” 
In CSR Limited v Della Maddalena  Kirby J, with the concurrence of Gleeson CJ, said  that Fox v Percy had brought about “an important change in the statement by this Court of the jurisdiction and powers of intermediate appellate courts”. His Honour said that the change “involved a shift to some degree from the more extreme judicial statements commanding deference to the findings of primary judges said to be based on credibility assessments”. The degree to which the shift in emphasis has occurred is not yet clear. Nevertheless, their Honours’ judgment indicates that reliance upon the subtle influence of demeanour requires careful consideration in each case before it is permitted to trump appellate intervention .
There are major policy reasons for limiting the power of appellate courts to overturn factual findings of trial judges. Were a complete rehearing on fact as well as law to be allowed, the burden on the appellate courts would be increased enormously. Apart from the increase in personnel and government cost that this would require, litigation would drag on even longer than it now does, appeals would last far longer, there would be great uncertainty in forecasting the prospects of success on appeal and the costs of litigation would increase a great deal. These problems explain judicial reluctance to widen the scope of appeals on fact. They explain the general approach that every litigant should be entitled to a full contest on the facts at the trial level only, and that the facts should be open to review thereafter only if some obvious and significant error is demonstrated .
Controlling demeanour findings
I suggest, however, that, to keep pace with current knowledge of the mind and personality, the law must develop some means by which control is exercised over demeanour findings.
One way is by rejecting the approach of some trial judges who merely set out the evidence adduced by one side, then the evidence adduced by the other, and then assert that having seen and heard the witnesses he or she prefers or believes the evidence of the one and not the other. If that were to be the law, many cases could be resolved at the end of the evidence simply by the judge saying: “I believe Mr X but not Mr Y and judgment follows accordingly”. That is not the way in which our legal system should operate .Often important issues of credibility involve sub-issues. Often, objective facts, or facts that are probable, are capable of having significant bearing on the sub-issues. In cases of this kind, it is incumbent upon trial judges first to resolve the sub-issues and to explain, by reference to the relevant facts, the conclusions to which they have come. This having been done, they should then turn to the ultimate facts in issue and explain how their decisions on the sub-issues have assisted them in forming a conclusion on the ultimate issue. It is only when adequate reasons of this kind are given that an unsuccessful party will be able to understand why the judge has believed his or her successful opponent. A failure by a judge, when making a demeanour finding, to deal with the issues in this way should be regarded as an error of law .
I do not suggest that trial judges would not be entitled to make demeanour findings contrary to the probabilities – merely, that they should give properly reasoned decisions if they do so. In choosing between witnesses on the basis of probability, a judge must of course bear in mind that the improbable account may nonetheless be the true one. The improbable is, by definition, that which may happen, and obvious injustice could result if a story told in evidence were too readily rejected simply because it was bizarre, surprising or unprecedented.
Sir Richard Egglestone  gives a good example of this. A woman was considered to be suffering from hallucinations because of her complaint that the head of her late husband, an African American, had rolled down the steps into her kitchen and had been retrieved by the devil wearing a black cloak. In fact, the devil was an eminent scientist. He had been making a study of the heads of criminals. On the night in question, he had been carrying the head of an African American who had died in gaol. He dropped the head in the street and it rolled down the steps of the old woman’s house. As the removal of the head was unlawful, he had wrapped a black cloak around his face and, calling out “Where’s my head? Give me my head!”, had gone to retrieve it, confident that he would not be recognised.
An example closer to home is that recounted by Sir Zalman Cowen. One of his aunts, an aged but independent lady with a heavy Polish accent, suffered an unfortunate accident in a London street. She was severely injured, and was taken to hospital where she had difficulty in identifying herself to the hospital staff. They became concerned about her mental state and, in a valiant effort to establish her credentials, she persisted in telling her incredulous audience, “My nephew is the Governor General of Australia”.
So, to conclude, I suggest that – despite the compelling policy reasons to limit fact appeals – appellate courts should regard demeanour-based findings of fact, contrary to the probabilities, as appealable error if adequate reasons are not given for them. Such a rule would advance the administration of justice. The virtually untrammelled power of trial judges to make what, practically speaking, are final decisions affecting the fate of individuals, on the ground of what the judges happen to feel about witnesses’ physical reactions when testifying, is an anachronism in a system of justice that prides itself on objectivity and rationality.
1. Edited version of paper delivered at the Winter Conference of the New Zealand Bar Association on 2 September 2006 at Queenstown
2. Judge of Appeal, Supreme Court of New South Wales
3. Gigi, Alan Jay Lerner
4. Shortly before this paper was delivered, McClellan CJ at CL delivered his paper, Who is telling the truth? Psychology, common sense and the law, at the 2006 Annual Conference of the Local Courts of New South Wales. Both papers were prepared independently of the other, each author not knowing that another paper on the subject was being written by someone else. Perhaps this presages an awakening of interest in the topic.
5. See discussion in Goodrich Aerospace Pty Ltd v Arsic  NSWCA 187 at 
6. Courts on Trial, Princeton University Press (1973), at 19
7. Remembering Trauma McNally, Harvard University Press (2003) at 36
8. McNally, op cit at 50
9. McNally, op cit at 68
10. McNally op cit at 277
11. Harvard University Press (2003)
12. McNally op cit at 277
13. McNally op cit at 69
14. Schacter D L, How the Mind Forgets and Remembers, Souvenir Press (2003)
15. Schacter op cit at 111
16. McNally, op cit, at 35
17. See Dworkin R, Law’s Empire, The Belknap Press of Harvard University Press Cambridge (1986) at 379-381, (referred to in Sharp L, Cognitive Heuristics and the Law, (1995) 20 Bulletin of the Australian Society of Legal Philosophy 71 at 78).
18. Frank, Courts on Trial, op cit at 21
19. MacKenna J in an extra-judicial observation referred to in Bingham, The Judge as Juror: The Judicial Determination of Factual Issues, (1985) 38 Current Legal Problems 1 at 8
20. Bingham op cit at 10-11
21. Rodger, What Are Appeal Courts For? (2004) 10 Otago LR 517 at 521
22. Frank J, (1930), Law and the Modern Mind, 6th printing, Peter Smith (1970), p xx (referred to in Sharp op cit at 71).
23. Mason K, Unconscious Judicial Prejudice, (2001) 75 ALJ 676 at 684, Sharp op cit at 86-92
24.  3 SCR 484
25. See generally, Ipp D A, Judicial Impartiality and Judicial Neutrality Is there a difference? (2000) Australian Bar Review 212
26. Mason op cit at 679
27. Holland v Jones (1917) 23 CLR 149 at 153
28. Allison JR, (1994), Ideology, Prejudgment, and Process Values, 28 New Eng L Rev 657 at introduction, circa footnote 5, (referred to in Sharp op cit at 74)
29. Sharp op cit at 71
30. See the discussion in Evidence, Proof and Probability, Eggleston, Weidenfeld and Nicolson (1983) at 192-193
31. Wiki v Atlantis Relocations (NSW) Pty Ltd  NSWCA 174
32. See for example Rae v International Insurance Brokers Limited  3 NZLR 190 at 199
33. (2003) 214 CLR 118
34. At 128 –129,  – 
35. In Société d’Advances Commerciales (Société Anonyme Egyptienne) v Merchants’ Marine Insurance Co (The “Palitana”) [(1924) 20 Ll L Rep 140 at 152]
36. See discussion in Goodrich Aerospace Pty Ltd v Arsic at  to 
37. (2006) 80 ALJR 458
38. at 465, 
39. as Tobias JA points out in Walden v Black  NSWCA 170 at 
40. Bingham op cit at 11
41. Goodrich Aerospace Pty Ltd v Arsic at 
42. Goodrich Aerospace Pty Ltd v Arsic at 
43. See the remarks of Hayne J in Waterways Authority v Fitzgibbon (2005) 79 ALJR 1816 at 1835, 
44. Eggleston op cit at 200-201 (the story apparently comes from Wigmore)