What happened
The factual matrix arose from a fatal collision on the Hume Highway approximately twelve miles south of Berrima at a location known as Little Hanging Rock. The deceased, Mr Jones, was driving an International truck northward towards Sydney alone. It was after dark, rain had fallen, and the road was wet. The highway wound through wooded hills with a gradient of about one in twelve. At a left-hand curve (to the deceased's direction of travel) the International truck came to rest at an angle against the inner bank of a low cutting. The deceased's body was found pitched across the cabin with his right leg trapped and the off-side of the cabin crushed. Twenty paces behind it, also on the western side of the road and facing north, stood the southbound diesel truck driven by the second respondent, Hegedus, an employee of the first respondent, Dunkel. The diesel truck's near-side door had been torn off and was located down the embankment on the eastern side of the road together with what appeared to be fresh blood stains. No tyre marks were visible on the bitumen or shoulders except short marks suggesting the diesel truck had slid backwards slightly after stopping.
The only eyewitness who gave evidence arrived on the scene within approximately one minute. He had been following the deceased's truck at a quarter-mile distance after both men had stopped for food some four or five miles earlier. A second motorist and then police and ambulance attended. A police officer noted the blood stains and later took a statement from Hegedus in hospital three days after the accident. In that statement Hegedus said he had been travelling at about 35 miles per hour down a slight grade, had just taken a right-hand bend, saw an oncoming vehicle with bright lights, applied his brakes and remembered nothing further. The roadway was wet and slippery. No other direct evidence of the dynamics of the collision existed. The width of the road from the outer white posts to the inner bank was estimated at twenty-eight feet, the bitumen twenty feet, each truck seven or eight feet wide, and the road was banked with a rise of about one in ten from inner to outer side.
At trial the plaintiff closed her case after calling evidence of these physical facts. Counsel for the defendants announced he would call no evidence and sought a directed verdict. The trial judge refused that application. After addresses and summing-up, including further directions given after a juryman asked specifically about the significance of the defendants not calling Hegedus, the jury returned a verdict for the defendants. An application to the Full Court of the Supreme Court of New South Wales for a new trial on the ground of misdirection was refused. The plaintiff appealed to the High Court. The Court was divided. Dixon CJ and Taylor J would have dismissed the appeal on the ground that the evidence could not support any affirmative inference of negligence. Kitto, Menzies and Windeyer JJ held there was a case fit to go to the jury and that the summing-up contained a material misdirection concerning the use the jury might make of the failure to call Hegedus. The appeal was allowed, the orders below set aside, and a new trial directed with costs following the event.
Why the court decided this way
The majority (Kitto, Menzies and Windeyer JJ) reached their conclusion by a two-step process firmly grounded in the authorities cited by all members of the Court. First, they examined whether the circumstantial evidence could support a reasonable inference that the collision occurred on the deceased's correct side of the road, thereby implying prima facie negligence by Hegedus. Kitto J undertook a detailed reconstruction. The International truck was found with its near-side front wheel against the inner bank and its off-side rear wheel seven feet in from the edge of the bitumen. The diesel truck, twenty feet long, was twenty paces behind facing the same direction with its near-side wheels one foot off the bitumen. Its front was badly damaged and its near-side door missing. No marks appeared on the wet outer shoulder. From these facts Kitto J reasoned that the jury could infer the vehicles collided with their off-side headlights in contact, that the diesel truck slewed completely around behind the International while forcing its front against the bank, and that the absence of drag marks on the outer shoulder meant the rear wheels of the diesel never left the bitumen. This placed the front of the diesel at least fifteen feet from the outer edge when at right angles, implying the driver's door of the International (and therefore the point of impact) was at least five feet inside the centre line on the deceased's correct side. The alternative—that the diesel hurled the International more than five feet across the road in a fraction of a second while its own rear wheels were already skidding—was one the jury could legitimately reject. Menzies J and Windeyer J reached the same conclusion on substantially the same physical facts, emphasising that the camber, the inside running of the International, the final positions and the nature of the damage made it more probable the collision occurred on the deceased's side.
Second, and decisively, the majority held the trial judge's direction on the failure to call Hegedus was incomplete and therefore a misdirection. The judge had correctly told the jury that the defendants were entitled not to call evidence, that the onus remained on the plaintiff, and that the facts proved by the plaintiff could be accepted more readily because they were uncontradicted. However, after a juryman expressly asked whether he could regard the absence of Hegedus as a weakness in the defendants' case, the judge declined to add anything further. Counsel for the plaintiff referred the judge to the judgment of Jordan CJ in De Gioia v Darling Island Stevedoring and Lighterage Co Ltd. In his further direction the judge repeated that the jury could accept the plaintiff's facts as proved and then decide whether an inference of negligence ought to be drawn from those proved facts, but he drew a sharp distinction between "proved facts" and "inferences" and gave the impression that once the jury reached the stage of drawing inferences the absence of Hegedus could have no significance. All three majority judges held this omitted the critical proposition that, once an inference is open on the evidence, the circumstance that the defendants had not called the one man who could put the true complexion on those facts is a circumstance the jury could properly take into account in deciding to draw the inference. Kitto J said the jury should have been told it would be proper to conclude that if Hegedus had given evidence it would not have assisted the defendants by throwing doubt on the correctness of the inference. Menzies J cited Black v Tung for the proposition that the charge withdrew from the jury a matter they were entitled to regard as rendering more probable the inferences contended for by the plaintiff. Windeyer J, after an extensive examination of the authorities from R v Burdett through Australian decisions, concluded that the direction did not accord with the commonsense principle that failure to call an available witness who could elucidate the facts permits the inference that the evidence would not have helped the defendants' case.
Dixon CJ and Taylor J dissented on the first step. They considered the physical evidence left the cause of the vehicles coming into contact wholly unexplained. While it was possible to say an empty diesel truck descending on the outside of a curve at 35 mph was more likely to cross the centre line than an International ascending on the inside at 25 mph, that was no more than a choice among guesses. The law, as stated in the unreported Bradshaw v McEwans Pty Ltd and Holloway v McFeeters, does not authorise a tribunal to choose between conjectures on the ground that one seems more likely than another. Because no positive inference of negligence could be drawn, the question of the direction on the absent witness never arose for them. The majority view prevailed, producing the order for a new trial.
Before and after state of the law
Before Jones v Dunkel the law governing circumstantial proof of negligence was settled by the unanimous judgment of the Court (Dixon J, Williams, Webb, Fullagar and Kitto JJ) in Bradshaw v McEwans Pty Ltd (unreported, 27 April 1951), itself drawn from Lord Robson's speech in Richard Evans & Co Ltd v Astley [1911] AC 674. That principle, repeated in Holloway v McFeeters (1956) 94 CLR 470 at 480-481, required that the plaintiff adduce evidence that, according to the ordinary course of human experience, gave rise to a more probable inference that the injury was caused by the defendant's negligence. Circumstances raising conflicting inferences of equal probability left the matter in the realm of conjecture and the plaintiff failed. The same authorities made clear that inferences from proved facts are themselves evidence.
The treatment of a party's failure to call a witness was also not new. Jordan CJ in De Gioia v Darling Island Stevedoring and Lighterage Co Ltd (1941) 42 SR (NSW) 1 had discussed it. Black v Tung (1953) VLR 629 and May v O'Sullivan (1955) 92 CLR 654 had applied the idea that uncontradicted evidence may be accepted more readily and that failure to call an available witness may strengthen inferences otherwise open. What Jones v Dunkel supplied was an authoritative synthesis and clarification of the precise manner in which a jury should be directed when both elements coexist: circumstantial evidence permitting an inference of negligence, and the unexplained failure of the defendant to call the person best able to explain the critical facts. The majority made plain that the absence cannot create an inference where none otherwise exists, but once an inference is open the absence is a legitimate consideration rendering that inference more probable.
The lengthy obiter discussion by Windeyer J on the procedural history of non-suit versus verdict by direction under the Common Law Procedure Act 1899 (NSW) and the Supreme Court Procedure Act 1900 (NSW) confirmed that the established practice in New South Wales—that a defendant who seeks a verdict by direction at the close of the plaintiff's case thereby elects not to call evidence—remained unaltered. That procedural rule interacted with the evidentiary principle: because counsel had deliberately chosen not to call Hegedus, the jury were entitled to draw the inference that his evidence would not have assisted the defence.
After the decision the principle became a settled and frequently invoked rule of evidence in civil trials throughout Australia. The distinction drawn by the majority between filling gaps (impermissible) and choosing between available inferences (permissible) has guided countless subsequent applications. The emphasis on the necessity for a clear jury direction when the point is raised, especially after a juror asks a direct question, has equally become standard.
Key passages with plain-English translation
Several passages are central. Dixon CJ stated: "In an action of negligence for death or personal injuries the plaintiff must fail unless he offers evidence supporting some positive inference implying negligence and it must be an inference which arises as an affirmative conclusion from the circumstances proved in evidence and one which they establish to the reasonable satisfaction of a judicial mind... the facts proved must form a reasonable basis for a definite conclusion affirmatively drawn of the truth of which the tribunal of fact may reasonably be satisfied." In plain English this means a plaintiff cannot ask a jury to guess; there must be enough concrete evidence to make it more likely than not that the defendant was negligent.
Kitto J's key direction on the absent witness reads: "the jury should at least have been told that it would be proper for them to conclude that if Hegedus had gone into the witness-box his evidence would not have assisted the defendants by throwing doubt on the correctness of the inference which, as I have explained, I consider was open on the plaintiff's evidence." Translation: once the physical evidence allows you to think the defendant's truck was on the wrong side, the fact that the defendants kept the only living eyewitness out of the box lets you more readily accept that explanation.
Menzies J wrote: "where an inference is open from facts proved by direct evidence and the question is whether it should be drawn, the circumstance that the defendant disputing it might have proved the contrary had he chosen to give evidence is properly to be taken into account as a circumstance in favour of drawing the inference." Plain English: if the evidence already points one way, the defendant's choice not to call the obvious witness is another reason to go that way.
Windeyer J quoted Abbott CJ from R v Burdett (1820) 4 B & Ald 95: "when such proof has been given, and the nature of the case is such as to admit of explanation or contradiction, if the conclusion to which the proof tends be untrue, and the accused offers no explanation or contradiction; can human reason do otherwise than adopt the conclusion to which the proof tends?" Translation: if the proven facts call for an answer and the person who could give it stays silent, ordinary common sense lets you treat the silence as confirmation.
These passages together articulate both the limits and the legitimate forensic use of a missing witness.
What fact patterns trigger this precedent
The precedent is triggered when three conditions coincide. First, the plaintiff must adduce some direct or circumstantial evidence from which an inference favourable to the plaintiff is open as a matter of rational probability rather than conjecture; the physical layout of the accident scene, the final resting positions of vehicles, the location and nature of damage, road camber, gradient, weather and absence of contradictory marks are classic examples. Second, the defendant must be in a position to call a witness who could reasonably be expected to throw light on the matter—classically the driver who survived, an employee with direct knowledge, or someone who inspected machinery or a site immediately after an incident. Third, the defendant must give no adequate explanation for the failure to call that witness. When those elements exist the tribunal of fact may treat the silence as strengthening the plaintiff's inference but may not use the silence to bridge an otherwise fatal gap in the plaintiff's case. The principle applies only in civil proceedings; it does not operate to reverse the onus of proof or to permit criminal convictions on suspicion. The procedural posture in which counsel announces at the close of the plaintiff's case that no evidence will be called, as occurred here, heightens the significance of the omission because it constitutes a deliberate forensic choice.
How later courts have treated it
The judgment itself demonstrates how the Court treated the authorities it inherited. All five justices treated the statement in Bradshaw v McEwans Pty Ltd as authoritative and binding on the question of inference versus conjecture. Dixon CJ and Taylor J applied it strictly to hold that the circumstantial evidence never crossed the line from conjecture to inference. The majority applied the same passage but found the physical facts did cross that line. Holloway v McFeeters was cited by Menzies J and Windeyer J for the proposition that "inferences from actual facts that are proved are just as much part of the evidence as those facts themselves." The Court uniformly cited Richard Evans & Co Ltd v Astley for the same distinction. De Gioia was referred to by plaintiff's counsel at trial and discussed by Windeyer J as correctly stating the commonsense view that silence may amount to more than acquiescence in primary facts. Black v Tung was adopted by Menzies J as illustrating the consequence of withdrawing from the jury a matter they were entitled to regard as rendering the plaintiff's inferences more probable.
Windeyer J's treatment of R v Burdett and the older English authorities was not to overrule them but to confirm that the same underlying principle of rational inference from conduct applies equally to the failure of a party to call an available witness. The judgment therefore both followed and refined the pre-existing law without departing from any cited precedent. The procedural history of non-suit and directed verdict was examined at length by Windeyer J to demonstrate that the New South Wales practice remained consistent with the common-law authorities and was not displaced by English post-Judicature Act decisions.
Still-open questions
The judgments leave several matters unresolved. First, precisely what constitutes a sufficient "explanation" for not calling a witness so as to prevent the adverse inference. The majority noted that no explanation was offered here, but the point was not explored. Second, the exact degree to which the inference may strengthen an otherwise weak case remains a question of fact and degree for the tribunal in each case; the line between "more confidently drawn" and "used to fill gaps" is not mathematically precise. Third, the interaction between the rule and the defendant's election to seek a directed verdict without calling evidence is left to the procedural law of each jurisdiction; Windeyer J's lengthy excursus demonstrates the complexity but does not purport to lay down a universal rule beyond New South Wales practice. Fourth, the weight (if any) to be given to an unsworn out-of-court statement by the absent witness, such as Hegedus's hospital statement, is left open; all members regarded it as inconclusive on the critical issue. Finally, the extent to which an appellate court should interfere with a trial judge's assessment of the sufficiency of a summing-up when no specific exception was taken at trial is touched upon by Windeyer J but not definitively settled beyond the facts of this appeal. These boundaries continue to be worked out in subsequent litigation, but the core principle remains as articulated in 1959.