Heydon, Dyson --- "Is the Weight of Evidence Material to its Admissibility?" [2014] CICrimJust 22; (2014) 26(2) Current Issues in Criminal Justice 219

---

2014 Paul Byrne SC Memorial Lecture

Is the Weight of Evidence Material to its Admissibility?^[*]

The Hon JD Heydon AC QC

I do not think really that probative value is ever a question for the judge to decide conclusively. At all events I am not able to call to mind any conditions in which it would be.

Dixon CJ^[1]

Introduction

It is a great honour to have been asked to deliver this Memorial Lecture. Paul Byrne was an exceptionally able defence lawyer — in his day and generation one of the ablest criminal lawyers in the country. I had some involvement with him at both the start and the end of his career. At the start he had a deep interest in the law relating to identification evidence at a time when its acute problems were coming under close scrutiny. It was on that subject that he worked for his Master’s degree. The discrimination and precision he showed then reflected themselves later in his efforts for those clients fortunate enough to engage his services, especially in High Court appeals. His relatively early death was a great loss to the Bar.

The classical division between the functions of judge and jury

Was Dixon CJ’s statement forming the epigraph to this lecture correct? Is it correct now? He was not of course referring to those questions of fact which judges decide in jury trials in the light of evidence after it has been admitted (which are often called questions of law: Thayer 1969:202). Instead, Dixon CJ was concerned with the question whether a judicial conclusion about the probative value of tendered evidence was ever material to the judge’s consideration of whether to admit it or reject it. His statement reflects the classical division of functions between judge and jury. That decision requires a separation of two questions. The first question is: ‘Is this evidence admissible?’ That question is left exclusively to the judge. The second question is: ‘Is this evidence, which is admissible and has been admitted, evidence of sufficient weight to act on in resolving the controversy being tried?’ That question is for the jury, not the judge. The issue to be raised in this lecture is whether that classical division is breaking down.

Even if the answer to the second question is ‘Yes’, there are three qualifications to the finality and completeness of that answer. First, a judge can direct an acquittal in a jury case by accepting a ‘no case’ submission at the end of the moving party’s case. In criminal proceedings there is no case to answer unless, taking into account only that evidence which favours the prosecution, assuming that it is accepted without reservation by a jury, and ignoring any prosecution evidence which is so inherently unreliable that no reasonable person could accept it as being true, the evidence is capable of bringing satisfaction beyond reasonable doubt to the minds of a reasonable jury (Glass 1981:847; Wentworth v Rogers at 429). Secondly, a judge can decline to leave to the jury a case unsupported by any evidence (Doney v The Queen). Thirdly, a judge can tell a jury at the close of the prosecution case that the jury can then, or at any time thereafter, stop the trial because they think the evidence forms an unsafe basis for a conviction (R v Prasad at 163;

Glass 1981:845). But the judge cannot otherwise prevent a jury considering the weight of whatever evidence has been admitted, as distinct from giving directions and warnings about it. And a jury cannot exclude from its consideration evidence which has been admitted on the ground that it thinks the evidence should not have been admitted

(Sinclair v The King at 326).

Hence at common law it is wrong to tell the jury that they must be satisfied that

a confession received into evidence is voluntary (McPherson v The Queen at 522;

R v Blades; Ex parte A-G). But, as Rich J said in Sinclair v The King:

It does not follow from this that the evidence given before [the trial judge] on the voir dire on the question of whether the evidence should be admitted may not, in a proper case, be given again in its entirety as evidence in the trial, not of course for the purpose of inviting the jury to give a ruling on admissibility of evidence, but for the purpose of assisting them to consider whether, in their opinion, the evidence qualifies the weight of the evidence which the judge has admitted (at 326).

There is a fine statement by Lord Denman CJ of the classical division between the roles of judge and jury. He made that statement in relation to factual conditions to be satisfied before evidence can be admitted for the jury’s consideration. In Doe d Jenkins v Davies

(at 323; 125; see Wendo v The Queen at 572) he gave some examples: competency; an apprehension of immediate death in relation to dying declarations; a search in relation to secondary evidence of lost documents; consanguinity or affinity in relation to pedigree declarations of deceased relatives. Lord Denman CJ continued:

The judge alone has to decide whether the condition has been fulfilled. If the proof is by witnesses, he must decide on their credibility. If counter-evidence is offered, he must receive it before he decides; and he has no right to ask the opinion of a jury on the fact as a condition precedent (at 323; 125; emphasis added).

Lord Denman CJ’s examples have been to some extent overtaken by the modern statutory tide, but his general point still holds. Lord Denman CJ’s pronouncement does not contravene Dixon CJ’s epigraph. The evidence is not admitted as a result of a conclusive decision by the judge on probative value; it is admitted because some more specific condition for admissibility which may or may not involve probative value is satisfied. The evidence received by the judge in relation to a condition for admissibility can be retendered before the jury if it is relevant to an issue before the jury. Underlying s 142 of the Evidence Act 1995 (NSW) (‘Evidence Act’) is an assumption that Lord Denman CJ’s doctrine is still sound.

One question of fact preliminary to the reception of an admission is the question whether the admission was made. Lord Denman’s use three times of the verb ‘decide’ did not suggest that admissibility depended merely on a prima facie test, but on the actual decision of the judge on at least a balance of probability. Similarly, in Sinclair v The King, Rich J three times said that the judge ‘had to be satisfied’ (emphasis added). But there is more modern common law authority to the contrary.

This more modern stream of authority seems to begin with Cornelius v The Queen in 1936. Starke J, after quoting passages supporting the application of Lord Denman CJ’s test to admissions, then said that the judge ‘merely decides whether there is prima facie any reason for presenting the evidence at all to the jury’ (at 239). He spoke to the same effect in Sinclair v The King (at 328). In Wendo v The Queen (at 573: reproduced below), Taylor and Owen JJ quoted the latter passage. Can it be said that the prima facie test is a qualification on the satisfaction test only in the area of admissions? That suggestion is undercut by the fact that Rich J was speaking in the context of receiving admissions. The test adopted in the Evidence Act does not correspond exactly with either of the competing common law tests, but its test is similar to the prima facie test. Section 88 provides: ‘For the purpose of determining whether evidence of an admission is admissible, the court is to find that a particular person made the admission if it is reasonably open to find that he or she made the admission.’

Rules of admissibility enhancing reliability

It is desirable now to refer briefly to rules of evidence which make weight a condition of admissibility or which, if they are to be complied with, tend to compel parties to call weightier evidence. They fall outside the topic but a few examples may be considered by way of background.

A simple example centres on the fundamental rules of court process which operated at common law and still operate under ss 26–46 of the Evidence Act. Criminal trials, and many civil trials, depend on oral evidence being received in the form of answers to questions. In general, evidence in chief and evidence in re-examination can only be received if the questions are not leading. Evidence in cross-examination can be received even if the questions are leading. One key function of the rule about leading questions in chief and in re-examination is to ensure that the evidence given is that of the witness, not that of examining counsel. What witnesses say in chief or re-examination may or may not be partisan. But what counsel would ask of their own witnesses, if unrestrained by the ban on leading questions, is very frequently likely to be partisan. What is merely partisan is unlikely to be weighty. One function of leading questions in cross-examination is also to improve the weight of the evidence. That is because a cross-examiner who could not lead would have to put up with a lot of uncontrollable and often weightless speech-making by the witness, but a cross-examiner who is allowed to put carefully crafted leading questions can, if the witness is not a liar, obtain evidence of considerable value.

Then there are specific rules which exclude evidence that is likely to lack weight. Under both the common law and s 84 of the Evidence Act, admissions influenced by violent, oppressive, inhuman or degrading conduct or the threat of it are inadmissible. Pressure of that kind is likely to generate untrue evidence, for not everyone has sufficient fortitude to withstand the pressure. The common law and the Act diverge on a different point.

The over-technical common law voluntariness rule excluded many admissions that were reliable. That had been a strong trend in the years before 1852, when R v Baldry was decided. R v Baldry stopped the trend but did not reverse it. Section 85(2) cuts back the rule of rejection by requiring admissions falling outside other provisions like s 84 and s 90 to be received unless the circumstances were likely adversely to affect the truth of the admission. The burden of proof rests on the prosecution, but the criterion is reliability — or the absence of unreliability. So the weight of an admission can be a factor going to its admissibility — implicitly in s 84, explicitly in s 85.

Section 85 does not stand alone in the Evidence Act as a provision which explicitly makes admissibility turn on weight. Section 65(2) creates four exceptions to the hearsay rule in criminal proceedings. A firsthand hearsay representation is admissible if it ‘was made under a duty to make that representation or to make representations of that kind’: s 65(2)(a). It is admissible if it ‘was made when or shortly after the asserted fact occurred and in circumstances that make it unlikely that the representation is a fabrication’: s 65(2)(b). It is admissible if it ‘was made in circumstances that make it highly probable that the representation is reliable’: s 65(2)(c). And it is admissible if it was:

(i) against the interests of the person who made it at the time it was made; and

(ii) made in circumstances that make it likely that the representation is reliable

(s 65(2)(d)).

In different measures these four exceptions turn on criteria involving reliability. There was a common law hearsay exception for statements of a deceased declarant which the declarant was obliged to record or report. It rested on an assumption about weight: that the obligation to record or report something increased the chance that it would be recorded or reported accurately. Similarly, s 65(2)(a) assumes that the obligation to say something increases the chance that it will be said accurately. Section 65(2)(b) also rests on an assumption about weight: that the fact that a representation was made when or shortly after an asserted fact occurred increases the chance of its reliability, provided there is no likelihood of fabrication. Section 65(2)(c) does not rest on any mere assumption as to weight, but explicitly selects as the criterion of admissibility a high probability of reliability. And s 65(2)(d) rests on both an assumption and an explicit criterion. The assumption is that statements by people against their interests are likelier to be true than other statements by them. The explicit criterion is a likelihood of reliability — a test similar to s 85(2).

There are other rules, whether now found in the Evidence Act or formerly found at common law, which make admissibility turn on weight. One relates to expert opinion evidence. Admission depends on the witness having ‘specialised knowledge’ based on the witness’s ‘training, study or experience’. So s 79 provides, and so did a corresponding rule at common law. The proffered evidence of a witness without the relevant ‘training, study or experience’ will be rejected. And a claim of ‘training, study or experience’ which is suspicious can be tested on the voir dire and rejected — though sometimes waiver of this opportunity in favour of exposure of the false claim before the trier of fact can have a more damaging effect on the whole of the tendering party’s case. If a witness claims to possess a doctorate from the Harvard Medical School and to be in specialist practice in Sydney, and it is established that in fact the witness has never been near Boston and has been struck off the roll of medical practitioners, the criterion of admissibility is not satisfied. The point is that the criterion rests on weight: non-satisfaction of the criterion results in the exclusion of evidence having no weight.

The problem

This discussion of evidence which is excluded for non-satisfaction of rules going to weight is outside the field marked out by Dixon CJ’s statement. This lecture does not centre on a classification of particular rules of evidence by reference to their impact on the reception of reliable evidence. This lecture will instead centre on a different question. To what extent is it open to a party opposing a tender to say, either bluntly or as part of a more sophisticated submission: ‘It is true that the evidence meets the formal criteria for admissibility. But it should be rejected simply because it has too little weight to merit reception into the record’?

The starting point is a disagreement between Starke J and Dixon J. In Sinclair v The King, the accused suffered from mental illness. Starke J said:

It was contended that the mere possibility that the confessions were the result of a disordered mental condition was sufficient to exclude them from evidence. ... I am unable to agree. A judge is not bound to exclude a confession from evidence because of such a possibility. He is entitled and bound to consider the probability of the mental condition affecting the truth of the confession in all the circumstances of the case and to decide whether there is prima facie reason for presenting it to the jury (at 328).^[2]

Though Dixon J sat in Sinclair v The King, he did not deal specifically on that occasion with what Starke J had said. But 17 years later, in Wendo v The Queen, Dixon CJ offered the following reflections:

I do not quite understand what the late Starke J meant by the statement in Sinclair v The King when he said that a judge was entitled and bound to consider the probability of the mental condition (scil of the prisoner) affecting the truth of a confession in all the circumstances of the case and to decide whether there was prima facie reason for presenting it to the jury. It appears to me that once it was established that a prisoner understood what he was doing in making a statement which, if true, would amount to a confession, it is admissible in evidence quite independently of its probative value. See the discussion in Sinclair’s case [at 336–38]. I do not think really that probative value is ever a question for the judge to decide conclusively. At all events I am not able to call to mind any conditions in which it would be (at 562; one footnote omitted; emphasis added).

I stress the last two sentences, which form the epigraph to this lecture.

This passage is interesting for those who study Dixon CJ’s styles of approbation and disapprobation, developed in a more reserved age than ours. His approbation could be rather chilly, as when he said of Windeyer J’s masterly analysis of equitable assignments in Norman v Federal Commissioner of Taxation: ‘I have had the advantage of reading the discussion contained in the judgment of Windeyer J of the whole subject of voluntary equitable assignments and I do not know that there is anything contained in it with which I am disposed to disagree’ (at 9). In disapprobation he could be severe, as when he attacked a redefinition of murder by the House of Lords in Smith v DPP in sombre terms in Parker v The Queen (at 632). Now Dixon CJ admired Starke J. He said, five years before Wendo’s case and a few days after Starke J died, that he had ‘a forensic power as formidable as I have seen’.^[3] In the light of that one must read behind the guarded qualifying phrases in the passage from Sinclair v The King quoted above — ‘I do not quite understand’, ‘It appears to me’, ‘I do not think really’. What Dixon CJ said in Wendo’s case is in truth not equivocal, but strongly condemnatory.

What attracted Dixon CJ’s condemnation appears to have been Starke J’s suggestion that the confession of a mentally ill person could be excluded merely because there was a ‘probability’ of the mental condition affecting the truth of the confession.

In Sinclair v The King, Dixon J accepted that the confession of a mentally ill person might be excluded, but not on the sole ground that its truth might be affected by the mental illness. In passages to which he referred in Wendo v The Queen, he quoted with approval from a decision of the Supreme Judicial Court of Massachusetts in Commonwealth v Zalenski. That Court said:

The medical evidence falls far short of proving that the mental infirmities of the defendant deprived him of the faculty of consciousness of the physical acts performed by him, of the power to retain them in his memory, and of the capacity to make a statement of those facts with reasonable accuracy (at 128–9).

Dixon J then made the point:

The tendency in more recent times has been against the exclusion of relevant evidence for reasons founded on the supposition that the medium of proof is untrustworthy, in the case of a witness, because of his situation and, in the case of evidentiary material, because of its source (Sinclair v The King at 337).

He said that there was independent evidence supplying many reasons for supposing that the confession made by Sinclair was substantially correct. He then said (at 338): ‘Though this consideration is not relevant to the question of the legal admissibility of such statements, it provides an example of the inconvenience or undesirability of a rule of rigid exclusion.’ It is rather difficult to assess what Dixon J’s precise position was. But Dixon CJ’s test seems to turn on a type of involuntariness, while Starke J’s turned on reliability.

The epigraph to this lecture thus raises a question: Is probative value ever a question for the judge to decide conclusively? A conclusive decision would be made if the judge decided that otherwise admissible evidence had too little probative value to make it receivable. It would be conclusive because it would prevent the jury, which is often called the constitutional tribunal of fact, from ever coming to consider the probative value of the evidence and perhaps arriving at a higher view of its value than the judge did after considering fuller materials than would normally have been available to the judge.

The issue can arise in various contexts. Here only a few will be examined.

Relevance

One area where the problem could in theory arise is relevance. The most fundamental rule of the law of evidence, both at common law and under s 56 of the Evidence Act, is that evidence cannot be admitted unless it is relevant. Does the inquiry into relevance entail an inquiry into the weight of the evidence tendered?

James Bradley Thayer, in A Preliminary Treatise on Evidence at the Common Law stated:

There is a principle — not so much a rule of evidence as a presupposition involved in the very conception of a rational system of evidence, as contrasted with the old formal and mechanical systems — which forbids receiving anything irrelevant, not logically probative. How are we to know what these things are? Not by any rule of law. The law furnishes no test of relevancy. For this, it tacitly refers to logic and general experience, — assuming that the principles of reasoning are known to its judges and ministers, just as a vast multitude of other things are assumed as already sufficiently known to them (1969:264–5).

This passage assumes that at the time when it must be decided whether the things tendered are ‘forbidden things’, there is no inquiry into their truth. The only inquiry is whether, if they establish what the tendering party says, they are fit to be received.

Another relevant passage is this:

When one offers ‘evidence’ ... he offers, otherwise than by reference to what is already known, to prove a matter of fact which is to be used as a basis of inference to another matter of fact. ... [H]e offers testimony, oral or written, to prove a fact; for even direct testimony, to be believed or disbelieved, according as we trust the witness, is really but a basis of inference (Thayer 1969:263–4).

This assumes that the time for belief or disbelief is not when the offer of evidence takes place, but after the offer has been accepted, the evidence has closed and argument has finished.

There is no common law authority against these fundamental propositions.^[4]

Hence in assessing questions of relevance in relation to admissibility, it is not for judges to speculate about possible constructions of the evidence which are adverse to the interests of the tendering party. It is necessary to assess relevance by taking the proposed evidence at the highest level it can reasonably be put at from the tendering party’s point of view. It is not correct for judges in jury trials to assess the probative value of the evidence for themselves and reject it as irrelevant if they identify aspects of it which may make it unconvincing or not probative in the fashion which the tendering party alleges. The possibility or likelihood, even, that evidence is fabricated does not make it irrelevant. When it is said that judges in jury trials in determining the admissibility of evidence have regard to the weight of the evidence, what is meant is not that they determine for themselves whether it is to be or may be believed, but that they determine what weight it would have in the case as a whole if it were believed (BBH v The Queen at 532).

So much for the common law. The position under s 55(1) of the Evidence Act is equally clear. It provides: ‘The evidence that is relevant in a proceeding is evidence that, if it were accepted, could rationally affect (directly or indirectly) the assessment of the probability of the existence of a fact in issue in the proceeding’. So evidence is relevant if, assuming it were accepted (and it may not be), it could (not would) rationally affect (not determine), not just directly but also indirectly, the assessment of the probability of the fact in issue. This test is not restrictive. The requirement that the relevance of evidence be assessed on the assumption that it is accepted may be compared with provisions like s 65(2)(c) and (d). They treat reliability as a passport to the admissibility of hearsay. They depart from the classical model in calling for an assessment of probative value at the admissibility stage, but it is an assessment distinct from relevance, which is assumed not to call for an examination of reliability. The same is true of s 85, which treats unreliability as a ground for the exclusion of relevant evidence. In each case reliability is treated as being quite distinct from relevance.

So the principles relating to relevance do not contradict Dixon CJ’s epigraph.

Hoch v The Queen

The doctrine in Hoch v The Queen, however, does. Hoch v The Queen was a case on the common law test for the admissibility of similar fact evidence. Under that common law test the evidence is excluded unless there is no rational view of it consistent with the innocence of the accused in the context of the prosecution case. It is commonly known as the test in Pfennig v The Queen (at 483–5). Hoch v The Queen put a qualification on the Pfennig test. The case concerned a complaint by three boys that Hoch had indecently assaulted them. Two of the boys were brothers and the third was a friend of one of those brothers. Hoch was convicted and appealed on the ground that if there were ‘a real possibility of collaboration between the witnesses the similar fact evidence should be rejected’ and that no joint trial should have been ordered (Hoch v The Queen at 293). The majority (Mason CJ, Wilson and Gaudron JJ) went beyond that submission, at least in form. They said that even if similar fact evidence were otherwise admissible it should be excluded if there were a ‘possibility of joint concoction’. They did not require the possibility to be either a ‘real’ or a ‘reasonable’ possibility. They relied on Lord Wilberforce’s reasoning in R v Boardman (at 444), which rejected tests turning on the ‘probability’ of concoction or a ‘real chance of concoction’. They said: ‘the evidence, being circumstantial evidence, has probative value only if it bears no reasonable explanation other than the happening of the events in issue’ (Hoch v The Queen at 296). That stated what came to be known as the Pfennig test at common law for admissibility of similar fact evidence. They then stated what became known as the Hoch qualification on that test:

In cases where there is a possibility of joint concoction there is another rational view of the evidence. That rational view — viz. joint concoction — is inconsistent both with the guilt of the accused person and with the improbability of the complainants having concocted similar lies. It thus destroys the probative value of the evidence which is a condition precedent to its admissibility (at 296).

In passing it may be said that a ‘mere’ possibility, as distinct from a ‘reasonable’ possibility, does not create a reasonable doubt and does not create what the majority called ‘another rational view’ (emphasis added).

To establish the necessary possibility, the majority relied on the ‘close relationship between the complainants’, on the existence of an ‘opportunity to concoct their accounts’, and on the fact that one complainant was ‘ill disposed’ to the accused even before the time of the alleged offences.

The minority (Brennan and Dawson JJ) preferred a ‘real chance’ test. Their conclusion is preferable. But they also allowed the appeal (at 304).

Although the minority in Hoch v The Queen did not expressly advert to the ban pronounced by Dixon CJ, they appeared to have endeavoured to place their reasoning outside it in the following way. Brennan and Dawson JJ said:

In the ordinary course, credibility is a question for the jury. That is so even with similar fact evidence. But in determining the admissibility of certain special classes of evidence it is inevitable that the trial judge must make an initial determination of questions of fact which the jury may ultimately have to decide. If the evidence is admitted, its probative force becomes a matter for the jury who may form their own view as to the possibility of a conspiracy among the witnesses to concoct their allegations (at 303; emphasis added).

The words ‘an initial determination’ stand in contrast to Dixon CJ’s words: ‘a question for the judge to decide conclusively’. But if the ‘initial determination’ is that there is a real possibility of concoction, the judge will have decided that question conclusively: for the jury will never have to consider the question. They only consider the question afresh if the evidence is admitted, not if it is excluded.

The practical day-to-day significance of the Hoch qualification as a rule of the common law is at present nil. That is because the common law Pfennig rule, to which it is a qualification, has been abolished by legislation in all jurisdictions but Queensland. It has been abolished because the legislatures have seen it as unduly favourable to the interests of the accused. Thus in federal jurisdictions and in New South Wales, Victoria, Tasmania, the Australian Capital Territory and the Northern Territory it has been abolished by the Evidence Act 1995 (Cth) and its corresponding enactments. It has been abolished in Western Australia and South Australia by s 31A of the Evidence Act 1906 (WA) and s 34P of the Evidence Act 1929 (SA) respectively. Though the Pfennig rule has not been abolished in Queensland, the Hoch qualification to it was abolished by s 132A of the Evidence Act 1977 (Qld).

But the Hoch qualification survives in a different form. That is because in determining ‘significant probative value’ for the purposes of ss 97 and 98 of the Evidence Act — the successors to the common law similar fact rules — the judge is obliged to consider the real possibility of concoction, contamination or innocent infection.^[5] This importation of a Hoch qualification was a gratuitous judicial creation, unsupported by any explicit statutory language. Indeed the judicially generated rebirth of the Hoch qualification in, or its grafting on to, a statute was a rather astonishing event, because it took place in the teeth of the legislative murder of its common law existence.

A key practical question has survived the legislative killing of the Hoch qualification at common law and attached itself to its subsequent judicial rebirth. The question is: on what materials is the judgment about the ‘possibility’ of concoction to be made? Ordinarily the admissibility of similar fact evidence is determined at some point before the trial, or at least before all the other evidence has been received. Admissibility is usually determined before the relevant witnesses have given oral evidence. It is usually determined on the basis of what is recorded in the committal proceedings (if any) or in depositions or in statements to the police or in proofs of evidence or in accounts by the tendering party of what the witnesses, it is hoped, will say. The trial judge will only rarely have heard a witness orally testifying to the similar fact evidence, or being cross-examined, before a ruling on admissibility is made. Opportunities to assess demeanour will normally be insignificant. At the time of the tender the rest of the evidence in the case will be incomplete.

How, then, can the factual elements relevant to the Hoch qualification be established in a satisfactory way? This was not a problem which arose when the courts applied the Pfennig test independently of the Hoch qualification. That is because it was not necessary for the judge to determine the probative value of the evidence in the way the jury would. In turn, that is because the Pfennig test rested on applying two assumptions rather than assessing the actual reliability and credibility of evidence.

The first assumption was an assumption in favour of the tendering party. That assumption was that the similar fact evidence would be accepted. Thus in Hoch v The Queen, Mason CJ, Wilson and Gaudron JJ said:^[6]

The basis for the admission of similar fact evidence lies in its possessing a particular probative value or cogency by reason that it reveals a pattern of activity such that, if accepted, it bears no reasonable explanation other than the inculpation of the accused person in the offence charged (at 294; emphasis added).

This first assumption is reminiscent of the position for decisions about relevance both at common law and under s 55(1).

The second assumption, in contrast to the first, was in favour of the accused. It was that the evidence in the case other than the similar fact evidence was insufficient to exclude a reasonable doubt. The trial judge had to ‘accept at least the possibility of the truth of the accused’s account’ (R v Cahill (No 2) at [24]). That second assumption was a necessary part of the Pfennig test. That was because if the contrary assumption were made — that the other evidence excluded any reasonable doubt — it would be impossible to carry out the task, required by many authorities, of examining the admissibility of the similar fact evidence in the light of the other evidence, and in view of that examination assessing whether the evidence as a whole would remove a reasonable doubt.

The first judge who saw plainly how the two assumptions worked was the late, and much lamented, Hodgson JA. Hodgson JA rightly said that the Pfennig test operated as follows:

[I]f it first be assumed that all the other evidence in the case left the jury with a reasonable doubt about the guilt of the accused, the propensity evidence must be such that, when it is considered along with the other evidence, there will then be no reasonable view that is consistent with the innocence of the accused. That is, the propensity evidence must be such that, when it is added to the other evidence, it would eliminate any reasonable doubt which might be left by the other evidence (R v WRC at 29).

Hodgson JA also said that the Pfennig test:

certainly does not require the judge to reach the view that the jury acting reasonably must convict: the judge must form his or her own view as to whether there is no rational view of the evidence, as it then appears to the judge, which is consistent with innocence, and the judge does not need to speculate as to how precisely that evidence might be affected by the way it is presented at the trial or by cross-examination, or how other minds might view it (R v Folbigg at [28]).

It is plain from Hodgson JA’s analysis that the Pfennig test did not invalidate or operate as an exception to the principle stated in Dixon CJ’s epigraph. Hodgson JA in R v WRC

(at [27]) added that the test:

does not mean that the judge must look at the propensity evidence in isolation, and not admit it unless there is no reasonable view of the evidence so considered that is consistent with the innocence of the accused of the offence with which the accused stands charged. That approach would be quite inconsistent with the correct approach for considering circumstantial evidence ... and the ... [test is sourced in] the character of propensity evidence as circumstantial evidence.

So much for the way in which the Pfennig test itself was applied. That mode of approach reduced the need for a voir dire. But what of the common law Hoch qualification to it? Did it call for a voir dire in which the judge decided, as a matter of fact, whether there was a possibility of concoction? The minority judges in Hoch v The Queen (Brennan and Dawson JJ) criticised the trial judge for not having conducted a voir dire. Counsel for the prosecution in Hoch v The Queen warned that various of the tests debated ‘would require the judge to hear much of the evidence on a voir dire before the trial commences’ (at 293). The minority judges considered that the evidence disclosed in the depositions might render a voir dire unnecessary. The majority judges did not deal with this problem. They said only that the matters of relationship, opportunity and ill-will referred to above were ‘clear from the evidence’. But what evidence? Was it the ‘evidence’ which was received after the objection to it had been rejected by the trial judge, or was it other evidentiary material considered before the objection was rejected? If it was the former, no practical guidance was given to later courts in relation to what evidentiary material could be taken into account on admissibility and how, because evidence in the form of depositions or the like before the trial can be very different from testimony at the trial after the witness who gave it has been cross-examined. It seems that it was the former. That is because Brennan and Dawson JJ said that the trial judge did not rely on a voir dire (which can be very detailed: see, for example, Tasmania v W (No 2)). They said that the trial judge did not read the depositions taken on the committal proceedings. They said the trial judge’s reasons for rejecting the accused’s application for separate trials relied only on an outline of the evidence which the prosecutor proposed to adduce on the various counts. The trial judge’s reasons did not set out the factual considerations on which the majority relied. The reasons of Brennan and Dawson JJ thus suggest that the evidence about the relationships of blood and friendship among the complainants came from their testimony after it had been received into evidence, not before (Hoch v The Queen at 299).

These problems of establishing the factual preconditions to the Hoch test at common law arise also in relation to the application of the similar test to ss 97–98.

The initial question is: what are the factual preconditions for the Hoch qualification to ss 97–98? Some of the s 97–98 authorities speak of ‘the reasonable possibility of concoction’ (see R v Colby at [107]; R v OGD (No 2) at [74], [77]). Another speaks of a ‘real chance’ that concoction has occurred (BSJ v The Queen at [21]). Others speak of a ‘real possibility of concoction’ (R v OGD (No 2) at [70]; BSJ v The Queen at [16]). Another speaks of ‘the possibility (in practical terms) of concoction’ (R v OGD (No 2) at [112]; emphasis in original). Others speak only of a ‘possibility’ (AE v The Queen at [44]; BSJ v The Queen at [20]). Another speaks of ‘whether the similar fact evidence is capable of reasonable explanation on the basis of concoction’ (R v OGD (No 2) at [70]) — picking up a phrase of Gaudron J’s in Harriman v The Queen (at 164). Another speaks of ‘likelihood’ and ‘more than a theoretical possibility’ (Tasmania v W (No 2) at [30], [32]). Sometimes a single judgment uses more than one of these expressions. That may reveal some inter-changeability between them. The Hoch qualification at common law should have turned on something sufficient to raise a reasonable doubt, that is, a reasonable possibility, for that is how the majority judges in the Hoch case fitted the Hoch qualification into the Pfennig test. From that point of view the majority judges in the Hoch case put the barrier too high for the prosecution, and the minority were correct. The approach of the majority judges in the Hoch case was to seek to reconcile the Hoch qualification with the primary test for admissibility. The enterprise was, with respect, entirely rational, though its execution was not. Can one embark on a similar enterprise in relation to the Hoch qualification to ss 97–98? That is, can one work out the test for the concoction qualification by inference from the primary test for admissibility? The primary test for admissibility under ss 97–98 is an easier test for the prosecution to satisfy than the common law Pfennig test. The statutory test for admissibility turns on ‘significant probative value’. It is necessary to remember that the probative value of evidence is defined in Pt 1 of the Dictionary as ‘the extent to which the evidence could rationally affect the assessment of the probability of the existence of the fact in issue’ (emphasis added). It is also necessary to remember that in general the trial judge’s determination of ‘significant probative value’ for the purposes of ss 97 and 98 is to proceed on the assumption that the evidence will be accepted by the jury. Whether it actually will be accepted by the jury depends on issues of reliability and credibility, to be assessed by the jury at the end of all the evidence. That is, the judge takes the evidence tendered at its highest. At least that was the traditional position in New South Wales and Tasmania, and in some Victorian cases (R v Shamouil at [51]–[65]; Lodhi v The Queen at [174]–[177];

R v Mundine at [33]; DAO v The Queen at [182]; KRI v The Queen at [53]–[55];

BSJ v The Queen at [18]). On this approach, the court does not usurp any of the jury’s functions. It simply makes a decision of law about the reasoning processes which are open to a jury (R v Ford at [52]). The ‘significant probative value’ test established by ss 97–98 calls for evidence which may be less weighty than evidence of which there is no rational view consistent with the innocence of the accused in the context of the prosecution case. Obviously a mere possibility of concoction, contamination or innocent infection is not enough to justify exclusion. But does it follow that even a real or reasonable possibility may not be enough to justify exclusion either?

Why do the courts read ss 97–98 as subject to the Hoch qualification? The first case which endeavoured to answer this question seems also to have been the last. It explained the position thus:

If the reasonable possibility of concoction suggests that evidence of this nature may be contaminated, it must be withheld from the jury because that risk deprives the evidence of its significant probative value, regardless of its substantial and relevant similarity (R v Colby at [107] per Mason P, Grove and Dunford JJ concurring).

That does not follow. On any view a risk as low as only a ‘reasonable possibility of contamination’ where in the end a jury might correctly decide that there was no contamination at all does not necessarily deprive the evidence of significant probative value. And the reasoning begs the question: what does ‘significant probative value’ mean? If the phrase refers to the evidence taken at its highest, the risk does not diminish significant probative value.

So much for the question: what is the foundation of the Hoch qualification in relation to ss 97–98? It is now necessary to consider another question: does the history of the Hoch qualification to ss 97–98 cast any light on how the judge is to determine admissibility? The Victorian Court of Appeal has said that the process of deciding whether there is a real chance:

that concoction has occurred will ordinarily not involve any assessment of the reliability or credibility of individual witnesses (BSJ v The Queen at [21]). Rather, it entails a fact-finding exercise, in which the judge will consider what the objective record shows about matters such as relationship, opportunity and motive.

For this the Victorian Court of Appeal cited a New South Wales case, R v OGD (No 2) (at [70], [112]). That case certainly referred to relationship, opportunity and motive. Indeed the facts relied on by the majority in Hoch v The Queen itself included relationship and motive. But the New South Wales Court of Criminal Appeal in R v OGD (No 2) did not use the expression ‘objective record’. Perhaps what the Victorian Court of Appeal meant is that if it emerges from material like the record of committal proceedings, other depositions, a voir dire, proofs of evidence or other material relied on by the prosecution to indicate the similar facts it hopes to prove that matters like relationship, opportunity and motive are not in contest and in that sense are ‘objectively’ established, the trial judge can use the material to assess the risk of concoction. But if assessing the risk of concoction depends on believing some witnesses and disbelieving others by recourse to estimates of reliability and credibility, the matter ought not to be decided against the prosecution but left for assessment by the jury after the evidence has been admitted. Whether or not those propositions are what the Victorian Court of Appeal meant, there is much to be said for them in principle. They minimise, or even avoid, any transgression of Dixon CJ’s epigraph.

General exceptions to judge/jury division under the Evidence Act

In relation to the Evidence Act it has been suggested that there are exceptions to the strict judge/jury division in areas other than the Hoch qualification. Thus in R v Shamouil, Spigelman CJ referred to the definition of ‘probative value’ and said that in some circumstances ‘issues of credibility or reliability are such that it is possible for a court to determine that it would not be open to the jury to conclude that the evidence could rationally affect the assessment of the probability of the existence of the fact in issue’ (at 63).

Similarly, in PG v The Queen, Nettle JA (with whom Neave and Harper JJA agreed) said:

[T]he assessment of reliability is predominantly a question for the jury, except in cases where the circumstances are such that the issues of credit and reliability are so fraught that it is possible for the court to determine that it would not be open to the jury to conclude that the evidence could rationally [a]ffect the assessment of the probability of the existence of the fact in issue (at [62]).

To what circumstances are these passages referring? That is unclear. Do they relate to some provision in the Act or do they operate to some extent outside it? That too is unclear. It does seem, however, that the contemplated exceptions are narrow in scope and likely to be infrequent in occurrence.

Probative value and unfair prejudice

Questions about Dixon CJ’s epigraph also arise in relation to the exclusion of evidence the probative value of which is outweighed by the danger of unfair prejudice to the defendant. At common law the process of exclusion turns on a discretion known as the Christie discretion (R v Christie at 559, 564). The New South Wales Court of Appeal, speaking through Hunt CJ at CL, said of the common law position in R v Carusi:^[7]

The power of the trial judge to exclude evidence in accordance with the Christie discretion does not permit the judge, in assessing what its probative value is, to determine whether the jury should or should not accept the evidence of the witness upon which the Crown case depends. The trial judge can only exclude the evidence of such a witness where, taken at its highest, its probative value is outweighed by its prejudicial effect; whereas this Court may use its supervisory powers to set aside a verdict where, the issue having been left to the jury, this Court is satisfied — on the whole of the evidence — that the jury ought nevertheless to have had a reasonable doubt (at 65–6).

Section 137 of the Evidence Act is no doubt derived from the Christie discretion. But it creates a different and non-discretionary regime. Section 137 provides: ‘In a criminal proceeding, the court must refuse to admit evidence adduced by the prosecutor if its probative value is outweighed by the danger of unfair prejudice to the defendant’ (emphasis added). Many judges have agreed that what Hunt CJ at CL said of the common law position is also true of s 137 (see R v Singh-Bal at 403; Adam v The Queen at 115; R v Shamouil at [47]–[68]; Lodhi v The Queen at [174]–[177]; R v Mundine at [33]; Irani v The Queen (2008) at [25]; KMJ v Tasmania; cf Papakosmas v The Queen at [86]). The primary argument for that view was advanced by Spigelman CJ in R v Shamouil (at [61]–[62]).

It turns on the definition of ‘probative value’ in the Act.

61. In my opinion, the critical word in this regard is the word could in the definition of probative value as set out above, namely, ‘the extent to which the evidence could rationally affect the assessment ...’. The focus on capability draws attention to what it is open for the tribunal of fact to conclude. It does not direct attention to what a tribunal of fact is likely to conclude. Evidence has ‘probative value’, as defined, if it is capable of supporting a verdict of guilty.

62. This conclusion is reinforced by the test that evidence must ‘rationally affect’ the assessment. As Gaudron J emphasised in Adam [v The Queen at [60]], a ‘test’ of ‘rationality’ also directs attention to capability rather than weight.

Earlier Spigelman CJ referred to Gaudron J’s conclusion that the definition of ‘probative value’ in the Dictionary ‘must have read into it an assumption that evidence would be accepted on the basis that “evidence can rationally affect the assessment of the probability of the fact in issue only if it is accepted”‘ (R v Shamouil at [53]).^[8] The statutory definition of ‘probative value’ of evidence as construed by Gaudron J and Spigelman CJ is similar to the definition of ‘relevance’ of evidence in s 55 as depending on an assumption that the evidence is to be accepted. In the case of s 55 the assumption is explicitly mandated. In the case of the definition of ‘probative value’ it is seen as implicit in the language. Critics of Gaudron J contend that the existence of the express words ‘if it were accepted’ in s 55 and their absence in the definition of ‘probative value’ are fatal to her conclusion. But this fails to answer her point: how can one conclude that the evidence can rationally affect the assessment of the probability of the fact in issue unless one accepts the evidence for the purposes of that exercise?

However, the Victorian Court of Appeal has departed from the position in New South Wales as stated in R v Shamouil. In Dupas v The Queen, it said that the New South Wales approach was ‘manifestly wrong and should not be followed’ (at 63). The single judgment of the full bench of five judges in Dupas v The Queen is, with respect, learned, long and complex. Since the Victorian Court of Appeal in Dupas v The Queen upheld the decision of the trial judge refusing to exclude the evidence, it would seem that no difference in approach could have affected the result, and thus the statements in Dupas v The Queen so far as they attack R v Shamouil are obiter dicta. However, even if they are, the obiter dicta are certainly well considered. The response of the New South Wales Court of Criminal Appeal took place in R v XY. It contained five judgments, which taken together are of substantial length, though shorter than Dupas v The Queen. There is learning and complexity in them too. The New South Wales response did not accept the Victorian position. But it did reveal differences of opinion within the New South Wales judiciary. Indeed there have been differences within it about s 137 ever since that section was enacted.

The appeal in R v XY was an appeal against a trial judge’s exclusion of evidence under s 137. By three votes to two, the appeal was dismissed. The minority, and one member of the majority (Hoeben JA), expressed strong support for the traditional New South Wales position as stated in R v Shamouil. Of the other two members of the majority, one appeared to accept R v Shamouil. The other, in dicta, attacked it.

The minority comprised Basten JA and Simpson J.

Simpson J adopted the most extreme position, which represented a reversal of the roles she and Basten JA played in the s 97 case of R v Zhang, where she adopted a less pure approach at [139] proposition (v) than Basten JA did at [46]. Extreme or not, there is much to be said for it. It corresponds with Dixon CJ’s epigraph. In particular she said that questions of credibility, reliability and weight played no part in the judge’s assessment of probative value for s 137 purposes. She relied on Spigelman CJ’s reasons for this conclusion in R v Shamouil. In addition she pointed out that it is not ordinarily possible to determine the actual probative value of any piece of evidence at the moment of tender; that cannot be done until the evidence as a whole is complete, including cross-examination, and it is possible to weigh witnesses against each other, and against the circumstances in which the evidence came into existence (R v XY at [166]–[175]).

Basten JA stated that in determining admissibility under s 137, the judge should assess the evidence tendered by the prosecution on the basis of its capacity to advance the prosecution case. It was necessary for the judge to assess the probative value of the evidence, taken at its highest from the prosecution’s point of view, for the purpose of the s 137 exercise (at [46]). The judge should deal with the evidence on the basis of any inference or direct support for a fact in issue which could be available to a reasonable jury considering the tendered evidence. The judge should not speculate about whether the jury would in fact accept the evidence and give it particular weight. And a judge should not make his or her own findings about accepting any inferences or giving the evidence particular weight (at [67]). But he (at [50]), unlike Simpson J (at [171]), did not quarrel with the following passage in Dupas v The Queen:

In order to determine the capacity of the evidence rationally to affect the determination of a fact in issue, the judge is required to make some assessment of the weight that the jury could, acting reasonably, give to that evidence. Where it is contended that the quality or frailties of the evidence would result in the jury attaching more weight to the evidence that it deserved, the trial judge is obliged to assess the extent of the risk. That does not require the trial judge to anticipate the weight that the jury would or will attach to it. The judge is obliged to assess what probative value the jury could assign to the evidence, against which must be balanced the risk that the jury will give the evidence disproportionate weight (at [63] proposition (d); emphasis added).

This is a difficult passage. The caution of Simpson J in not accepting it is understandable. Two points should be made. First, strictly speaking, the judge need not assess the weight which the jury could, acting reasonably, give to the evidence — only the weight he or she personally gives the evidence at its highest. Secondly, any risk that the jury will give undeserved weight to the evidence which the judge perceives is to be dealt with not by examining ‘probative value’ but by seeing whether there is unfair prejudice.

The majority comprised Hoeben JA, Blanch J and Price J.

Hoeben JA agreed with Basten JA and Simpson J, and agreed that R v Shamouil should be followed, subject to one qualification. He said:

When assessing the probative value of the prosecution evidence sought to be excluded, ie, its capacity to support the prosecution case, a court can take into account the fact of competing inferences which might be available on the evidence, as distinct from determining which inference or inferences should be or are most likely to be preferred (R v XY at [88]).

These gnomic words involved an extension of what was said in an earlier New South Wales case, DSJ v The Queen (at 761, 775), a decision on s 98 of the Evidence Act, to s 137.

Blanch J seemed to agree with Hoeben JA’s approach. He considered that the tendered evidence lay open to competing inferences, and hence was of diminished probative value

(R v XY at 207).

The decisive point for Price J was that the exclusion of the evidence did not ‘substantially’ weaken the prosecution case and thus fell outside the statute regulating the jurisdiction to consider prosecution appeals: Criminal Appeal Act 1912 (NSW) s 5F(3A). But, obiter, he expressed a preference for Dupas v The Queen over Simpson J’s approach. He said that cross-examination of witnesses on the voir dire would enable actual probative value to be assessed.

On one view, the majority New South Wales position seems to be that stated by Hoeben JA: for Blanch J agreed with him, Hoeben JA is supported by three members of a five judge Court of Criminal Appeal in DSJ v The Queen, and there is strictly speaking no majority against him in R v XY. But Hoeben JA’s qualification is, with respect, a somewhat complex one. If in all other respects the evidence tendered by the prosecution is to be taken at its highest from the prosecution’s point of view, why should not available inferences from it be taken at their highest as well? Further, the extension of DSJ v The Queen from s 98 to s 137 has been criticised. There is another stream of New South Wales authority limiting

DSJ v The Queen to s 98 and not extending it to s 137 (R v Burton at [181]–[183]; see also

R v Sood at [38]–[39]). In R v Burton, that result was explained on the basis that the s 98 task is to measure how ‘significant’ probative value is in the light of the whole of the tendering party’s case, which permits consideration of an ‘alternative explanation arising on the evidence’. But the s 137 task is different: it requires the balancing against unfair prejudice not of ‘significant probative value’, but only of probative value. In principle it may be that instead of trying to distinguish the two streams of authority, it would be better to concentrate on elimination of DSJ v The Queen.

The disputation between and within the intermediate appellate courts of New South Wales and Victoria is detailed. The detail may obscure the possible fact that the gap is narrow. That was certainly Basten JA’s view in R v XY (at [51]). As Basten J demonstrated (in R v XY at [54]–[56]), the earlier decision of the Victorian Court of Appeal in R v Dupas (No 3) (at [259]–[262]) is very close to the New South Wales position in pointing to earlier Victorian cases holding that the exclusion of unreliable evidence pursuant to the Christie discretion are likely to be very rare: those cases did not give a single instance of where it had been done. There is another respect in which the gap may be narrow. Even if the evidence is to be accepted in the sense of being taken at its highest level, the circumstances surrounding the evidence may indicate that its highest level is not very high at all. One example would be an identification made very briefly in foggy conditions and in bad light by a witness who did not know the person identified and whose racial background differed from that of the person identified. Is it right to say: ‘Well, it is an identification, and we must take it at its highest — as high as any other identification’? Or should we say: ‘It is an identification, but rather a weak one?’ A very weak identification at its highest is not equivalent to a very strong identification — only a very weak one. From that point of view it does not matter whether one takes the Victorian approach, which would seek to isolate and evaluate in detail particular weaknesses in the evidence, or the New South Wales approach, which takes inherently unconvincing evidence at its highest, but treats it only as weak because it is inherently unconvincing.

However that may be, the Victorian position, with respect, appears unsatisfactory for several reasons.

First, it rests on a contention that at common law the trial judge was required to ‘evaluate the weight that the jury could relevantly attach to the evidence’ (Dupas v The Queen at [63] proposition (a)). Now the Victorian Court of Appeal devoted 33 pages — 74 paragraphs — to discussing the common law discretion. They quoted from and referred to many cases in many fields. But despite all that learning and all that labour, it must be said with respect that that contention about the common law rests on a growing fallacy in legal analysis. The fallacy is to take very general and ambiguous statements by judges whose eyes were not specifically directed to the present controversy and to read them as supporting one side of the present controversy. The statements are all obscure as to whether they are talking about the probative value alleged by the prosecution, the probative value which a jury might ascribe to the evidence, the probative value which it can be predicted the jury would ascribe to it, or the probative value perceived by the judge. Thus in R v Christie, Lord Moulton spoke of ‘true evidential value’ and Lord Reading of ‘evidence’ having ‘little value’ (at 559, 564). But ‘value’ by what standard, assessed through what forensic techniques, in the light of what materials and applying what tests? It must, however, be admitted that there are cases on s 137 in which appellate courts may have picked apart the probative value of evidence and not accepted it at its highest from the prosecution point of view.

Secondly, the Victorian position rests on the view that s 137 reflects the common law. That is what the Australian Law Reform Commission contended. But the common law created a discretion and s 137 does not. Section 137 requires the evidence to be excluded if its probative value is exceeded by its prejudicial effect. That is an outcome which the Australian Law Reform Commission never recommended. Its draft Bills provided for a discretion only (ALRC 1985:vol 2, App A cl 115; ALRC 1987:App A, cl 118).

Its discussions of the problem were brief (ALRC 1985:vol 1, [957], vol 2, App C [259]; ALRC 1987:[317]–[319]). Save in one respect, those discussions never remotely approached the problem under present consideration (ALRC 1987:[146]). In the one paragraph which did contend that the Bill permitted an examination of ‘reliability’ as going to ‘probative value’, speaking about the tender of the hearsay statement, the Commission said:

the judge can take account of the fact that the plaintiff’s evidence is hearsay, as that will go to the probative value of the plaintiff’s evidence. The judge can also look to the surrounding circumstances in which the statement was made to the plaintiff and other matters going to the reliability of the evidence, such as how recently after the event the statement was made, whether the person who made the statement had an interest or not in the matters referred to and whether the circumstances placed some obligation on the person who made the statement to tell the truth (ALRC 1987:[146]).

But, first, does the New South Wales approach exclude consideration of those factors? Depending what is known about the factors listed, they may indicate no more than that, taken at its highest, the statement does not have much probative value. Secondly, the passage treated the personal ‘interest’ in the matters referred to as going to reliability, which indicates that the passage is contrary to the Victorian assumption of the ‘truthfulness’ of the evidence but not its reliability, which is explained below. That is because personal interest would seem only to go to ‘truthfulness’, that is, credibility, not reliability. Dupas v The Queen relies (at [179]–[180]) on other particular assertions in the relevant ALRC reports. However, those assertions go not to the present point, but to unfair prejudice — the unfair prejudice that can arise from an inability to test evidence by cross-examination or from a likelihood that the jury would overestimate the reliability of the evidence (ALRC 1985:vol I, [957]; ALRC 2005:[16.45], [16.47]). The last of those reports in fact treated credibility and reliability as issues for the jury, because they arose during the trial, especially in cross-examination. The Australian Law Reform Commission said that if issues of reliability or credibility led to the likelihood that the jury would overestimate weight, that could lead to exclusion. That would be exclusion because of unfair prejudice.

Thirdly, the Victorian position, whether or not it rests on a narrow difference from the New South Wales position or something wider, is rather complex, and not wholly workable. The Court held that the trial judge, in conducting what it described as ‘the balancing task’ under s 137, ‘is only obliged to assume that the jury will accept the evidence to be truthful but is not required to make an assumption that its reliability will be accepted’ (emphasis added) (Dupas v The Queen at [63] proposition (c)). The Court also said:

We proceed upon the view that, like s 55, the definition of ‘probative value’ (although not containing the words ‘if accepted’) involves the assumption which was made at common law by the trial judge that the jury would accept the evidence is ‘truthful’. Nothing in the language or context of the statute or in the underlying policy suggests, however, that the assumption was to extend beyond truthfulness to reliability (at [184]; emphasis added).

The argument posits a distinction between ‘truthfulness’ and ‘reliability’. It is a distinction for which there was prior Victorian authority at common law (Rozenes v Beljajev at 560). By ‘truthfulness’ is apparently meant ‘sincerity’ or ‘an honest attempt to be accurate’: that is, ‘credibility’. The words ‘if accepted’ as explicitly appearing in s 55(1) and as implicit in the definition of ‘probative value’ do not distinguish between ‘truthfulness’ (that is, sincerity) and ‘reliability’. Nor does any other part of the legislation. A foe of this distinction calls it ‘obiter dicta only’ (Odgers 2014:[1.3.14760]); but the whole of Dupas v The Queen is in a sense highly considered dicta only. The distinction drawn in Dupas v The Queen between truthfulness (that is, sincerity or credibility) and reliability is not always easy to apply in practice. Evidence which seems questionable may be erroneous because of lying or because of poor powers of perception, recollection or expression. What the cause is can be difficult to isolate. The assessment of what it is often depends on cross-examination. Further, the Victorian Court of Appeal’s distinction in Dupas v The Queen between reliability and truthfulness (credibility) does not overcome the difficulty that in the Dictionary ‘credibility of a witness’ is defined to mean the credibility of any part or all of the evidence of the witness, and to include ‘the witness’s ability to observe or remember facts and events about which the witness has given, is giving or is to give evidence’.

The quoted part of the definition includes an aspect of reliability.

Fourthly, the Victorian position is vulnerable not only to the reasoning of Spigelman CJ in Shamouil v The Queen, based on textual considerations, but also to the reasoning of Simpson J in R v XY, based on the unworkable consequences of the competing construction. It would greatly increase the dependence of trial judges on the voir dire in order to assess reliability. It would tend therefore to a great increase in the length of trials and the delays of litigation. It would generate considerable pressure to abandon the system of trial which makes the distinction between admissibility and weight necessary. That would ignore the good rational arguments for the jury system of trial. And it would collide with the general support which that system of trial enjoys.

Fifthly, both the New South Wales and the Victorian positions are difficult to reconcile with the introduction into ss 97–98 of the Hoch qualification in relation to the possibility of concoction. The Hoch qualification involves the assessment of credibility for the purpose of deciding on ‘significant probative value’ under ss 97–98. That is contrary to what R v XY in New South Wales said was the general rule for assessing ‘probative value’ under s 137. The Hoch qualification in relation to the possibility of concoction for deciding on ‘significant probative value’ under ss 97–98 is also contrary to the view stated in Dupas v The Queen in Victoria that in assessing ‘probative value’ under s 137 the judge must ‘assume that the jury will accept the evidence to be truthful’ (that is, sincere) (at 63).

In Dupas v The Queen, the Victorian Court of Appeal adopted many points made in an article by two critics of the New South Wales position (Smith and Odgers 2011:304–5). It is desirable to deal with one or two of the points they made which have so far not been mentioned.

The critics contend that the classical distinction ‘between judge and jury roles is irrelevant to’ construing the statutory definition of ‘probative value’. They say that Ch 3 ‘demonstrates a deliberate and detailed full frontal assault’ on the distinction ‘because it gives trial judges an extensive battery of powers to interfere in the jury’s fact-finding process’. This is open to doubt. Chapter 3 certainly widens admissibility. The Act does create powers or duties to exclude evidence despite that widening: see, for example, ss 90, 101, 135, 137 and 138(3)(a). So far as admissibility is wider, this correspondingly widens the jury’s powers in its fact-finding process. The exclusionary powers and duties of judges can narrow the field available to the jury. But that does not point to any particular construction of the legislation creating those exclusionary powers or duties.

The critics say that the natural meaning of s 137 is the wide meaning. But all proponents of a particular construction say their construction is the natural meaning. In any event, the words ‘reliability’ and ‘credibility’ are not used in s 137. This creates an obstacle to contending that s 137 should be read as if they were. It creates an even greater obstacle to contending, as the Victorian courts do, that s 137 should be read as if one was used but not the other.

However, the real difficulty in the stance taken by the critics of the New South Wales position, and in the stance taken in the Victorian Court of Appeal, is that a stimulating acid question has not been faced and dealt with. That question is: how can reliability and credibility — including through the demeanour of witnesses both in chief and under crossexamination — be assessed at the moment of tender, save by a thorough voir dire? And if thorough voir dires became a standard practice in relation not only to s 137, but to s 135, s 97, s 98, s 101 and s 138(3)(a), all of which speak of ‘probative value’, how could they be prevented from swamping the main trial process? Will there not arise a system in which most controversial issues are dealt with first by having a voir dire and then, if the tender succeeds, by calling the evidence afresh before the jury? What impact would so cumbersome a development in criminal procedure — wasteful of time, irritating to jurors whom it renders idle, destructive of trial rhythm — have on the survival of trial by jury itself?

The Victorian Court of Appeal has recently, in Velkoski v The Queen (at [179]), suggested that the construction of s 137 was relevant to that of ss 97–98 as well. If so, it might be relevant to ss 101, 135 and 138(3)(a). Any consideration of that problem must take account of what Simpson J said in R v Burton.

There are other fields in which Dixon CJ’s epigraph could be investigated. And there are many other authorities which could be considered. But lectures, mercifully, have to end at some point. It is submitted that analysis of the above authorities leads to four conclusions. First, the Hoch qualification should not have been introduced into ss 97–98. Secondly, Dupas v The Queen is wrong in relation to reliability. Thirdly, whether the New South Wales position for s 137 is of general application to other sections turning on probative value — ss 97, 98, 101, 135 and 138(3)(a) — remains an open question. Fourthly, Dixon CJ’s epigraph is under stress. But it is revealing some strength. It is not yet buckling.

Cases

Adam v The Queen (2001) 207 CLR 96

AE v The Queen [2008] NSWCCA 52 (20 March 2008)

BBH v The Queen [2012] HCA 9; (2012) 245 CLR 499

Boardman v DPP [1975] AC 421

BSJ v The Queen [2012] VSCA 93; (2012) 35 VR 475

Burns v The Queen [1975] HCA 21; (1975) 132 CLR 258

Commonwealth v Zalenski, 287 Mass 125 (1934)

Cornelius v The Queen (1936) 55 CLR 235

DAO v The Queen [2011] NSWCCA 63; (2011) 81 NSWLR 568

Doe d Jenkins v Davies [1847] EngR 10; (1847) 10 QB 314; 116 ER 122

Doney v The Queen (1990) 171 CLR 206

DSJ v The Queen [2012] NSWCCA 9; (2012) 84 NSWLR 758

Dupas v The Queen (2012) 218 A Crim R 307

Harriman v The Queen [1989] HCA 50; (1989) 167 CLR 590

HML v The Queen (2008) 235 CLR 344

Hoch v The Queen [1988] HCA 50; (1988) 165 CLR 292

Irani v The Queen [2008] NSWCCA 217; (2008) 188 A Crim R 125

KMJ v Tasmania [2011] TASCCA 7; (2011) 20 Tas R 425

KRI v The Queen [2011] VSCA 127; (2011) 207 A Crim R 552

Lai v Western Australia [2012] WASCA 181; (2012) 225 A Crim R 218

Lodhi v The Queen (2007) 179 A Crim R 470

McPherson v The Queen [1981] HCA 46; (1981) 147 CLR 512

Palmer v The Queen [1998] HCA 2; (1998) 193 CLR 1

Papakosmas v The Queen [1999] HCA 37; (1999) 196 CLR 297

Parker v The Queen [1963] HCA 14; (1963) 111 CLR 610

Pfennig v The Queen [1995] HCA 7; (1995) 182 CLR 461

PG v The Queen [2010] VSCA 289 (19 October 2010)

Phillips v The Queen [2006] HCA 4; (2006) 225 CLR 303

PNJ v DPP (Vic) [2010] VSCA 88; (2010) 27 VR 146

R v Baldry (1852) 2 Den 430; 169 ER 568

R v Blades; Ex parte A-G [2001] QCA 384; (2001) 124 A Crim R 415

R v Boardman [1975] AC 421

R v Burton [2013] NSWCCA 335 (20 December 2013)

R v Cahill (No 2) [1999] VSCA 84; [1999] 2 VR 387 (27 May 1999)

R v Carusi (1997) 92 A Crim R 52

R v Christie [1914] AC 545

R v Colby [1999] NSWCCA 261 (26 August 1999)

R v Dupas (No 3) [2009] VSCA 202; (2009) 198 A Crim R 454

R v Folbigg [2003] NSWCCA 17 (24 October 2003)

R v Ford [2009] NSWCCA 306; (2009) 273 ALR 286

R v Mundine [2008] NSWCCA 55; (2008) 182 A Crim R 302

R v OGD (No 2) [2000] NSWCCA 404; (2000) 50 NSWLR 433

R v Prasad (1979) 23 SASR 161

R v Shamouil [2006] NSWCCA 112; (2006) 66 NSWLR 228

R v Singh-Bal (1997) 92 A Crim R 397

R v Sood [2007] NSWCCA 214 (19 July 2007)

R v WRC [2002] NSWCCA 210; (2002) 130 A Crim R 89

R v XY [2013] NSWCCA 121; (2013) 84 NSWLR 363

R v Zhang [2005] NSWCCA 437; (2005) 227 ALR 311

Rozenes v Beljajev [1995] VicRp 34; [1995] 1 VR 533

Sinclair v The King [1946] HCA 55; (1946) 73 CLR 316

Smith v Director of Public Prosecutions [1961] AC 290

Sutton v The Queen [1984] HCA 5; (1984) 152 CLR 528

Tasmania v W (No 2) (2012) 217 A Crim R 155

Velkoski v The Queen [2014] VSCA 121 (18 June 2014)

Wendo v The Queen [1963] HCA 19; (1963) 109 CLR 559

Wentworth v The Queenogers [1984] 2 NSWLR 422

Legislation

Criminal Appeal Act 1912 (NSW)

Evidence Act 1995 (Cth)

Evidence Act 1995 (NSW)

Evidence Act 1977 (Qld)

Evidence Act 1929 (SA)

Evidence Act 1906 (WA)

References

Australian Law Reform Commission (‘ALRC’) (1985) Evidence, Interim Report, ALRC Report No 26 <http://www.austlii.edu.au/cgi-bin/sinodisp/au/other/lawreform/ALRC/1985/26.html>

Australian Law Reform Commission (‘ALRC’) (1987) Evidence, Final Report, ALRC Report No 38 <http://www.austlii.edu.au/au/other/lawreform/ALRC/1987/38.html>

Australian Law Reform Commission (‘ALRC’) (2005) Uniform Evidence Law, Final Report,

ALRC Report No 102 <http://www.alrc.gov.au/publications/report-102>

Glass HH (1981) ‘The Insufficiency of Evidence to Raise a Case to Answer’, Australian Law Journal 55(12), 842–54

Odgers S (2014) Uniform Evidence Law (11^th ed), Law Book Co

Smith T and Odgers S (2011) ‘Determining “Probative Value” for the Purposes of Section 137 in the Uniform Evidence Law’, Criminal Law Journal 34(5), 292–305

Thayer JB (1898) A Preliminary Treatise on Evidence at the Common Law, first published 1898, republished 1969, Rothman Reprints Inc

---

^[*] The 2014 Paul Byrne SC Memorial Lecture, ‘Is the Weight of Evidence Material to Its Admissibility?’, was presented by the Hon John Dyson Heydon at Sydney Law School, The University of Sydney, on 15 October 2014.

^[1] Wendo v The Queen at 562.

^[2] This passage was quoted in Wendo v The Queen at 573 by Taylor and Owen JJ, but seemingly only on the standard of proof issue in the course of rejecting the ‘beyond reasonable doubt’ test.

^[3] (1958) 97 CLR iv.

^[4] In Palmer v The Queen at 24 McHugh J said: ‘[E]vidence of a relevant fact is excluded only when it infringes some policy of the law, one of which (even in civil cases) is that evidence of the relevant fact is not admissible if the probative value of that fact is so low that it cannot justify the time, convenience and cost of litigating its proof.’ But this does not say that evidence of low probative value is not relevant. It says only that an exclusionary rule may apply to it.

^[5] R v Colby at [107]; AE v The Queen at [44]; PNJ v DPP at [27] (accepted by Crown); BSJ v The Queen at [21]; R v OGD (No 2) at 445–6, 454. Section 97(1) provides that evidence is not admissible to prove that a person has or had a tendency to act in a particular way or to have a particular state of mind unless ‘the court thinks that the evidence will, either by itself or having regard to other evidence adduced or to be adduced by the party seeking to adduce the evidence have significant probative value’. Section 98(1)(b) makes a similar provision in relation to coincidence evidence.

^[6] See also Boardman v DPP at 457 (‘only an ultra-cautious jury, if they accepted it as true, would acquit in face of it’ (emphasis added)); Sutton v The Queen at 564 (‘before admitting the evidence the trial judge himself must conclude that a reasonable jury would, if they accept the evidence, regard it as being consistent with innocence’ (emphasis added)); Pfennig v The Queen at 487 (‘if accepted’ — twice); Phillips v The Queen at 323–4; HML v The Queen at 359, 362, 370, 385, 428–31.

^[7] See also BBH v The Queen at 536; Lai v Western Australia at 223.

^[8] Gaudron J was in dissent but not on that point. In Papakosmas v The Queen at 323, in contrast, McHugh J said that the definition of ‘probative value’ required an assessment involving ‘consideration of reliability’.