For the reasons given by Porter J, I would infer that hot work was the cause of the fire and regard it as far more likely that the hot work of Mr Maine, rather than the hot work of the appellant's employee Mr Cooper, caused the fire. Consideration of the whole of the evidence in this case leads me to conclude that, on the balance of probabilities, an "inference of causal connection" between Mr Maine's hot work and the respondent's breach of duty and the occurrence of the fire should be drawn.
I also agree with Porter J that a breach of duty by the appellants has not been proved.
[2]
[1]40 Duty in relation to general workplace facilities
[3]
A person conducting a business or undertaking at a workplace must ensure, so far as is reasonably practicable, the following:
[4]
[2] As I understand Mr Thomas' evidence, "left" equates with the eastern side.
[4] Mr Barnes thought that the British text, Cooke & Ide, Principles of Fire Investigation, was no longer authoritative as being "too old", but he agreed with the propositions from it which were put to him.
[7]
[5] A metal nut heated to 900oC so that it is "red hot".
[8]
[6] This pile seems to have first been mentioned in the cross-examination of Mr Thomas. The pile was in the extreme north-eastern corner of the main building. It was close to but not within the area of origin identified by Mr Muench and Mr Thomas. Mr Thomas said that he could not recall what it was, and that it was severe low level burning but not as severe as further into the building. He said the burn indicators were that the fire extended back to the area in which it was. It was not specifically raised with Mr Muench or Mr Thomas as a cause, or the point of origin, of the fire. Mr Barnes did not suggest that it was either of those things, and it was not the subject of closing submissions. In the appeal, the respondent merely noted that the trial judge had not excluded it as a possible cause. In my view it has no consequence.
[9]
[7] Kirby P dissented as to the outcome, but this statement does not seem controversial.
(1) An employer is to appoint a responsible officer for each workplace at which the employer carries on business.
[12]
(2) If an employer fails to appoint a responsible officer for a workplace, the person responsible for the direction and management of the business of the employer at that workplace is taken to have been appointed as the responsible officer for that workplace.
The Langmaids claimed damages for breach of contract. If a plaintiff proves a breach of contract, but not that any damage was occasioned by the breach, nominal damages are recoverable: Luna Park (NSW) Ltd v Tramways Advertising Pty Ltd[1938] HCA 66; (1938) 61 CLR 286. It follows that the learned primary judge had a duty to determine whether there had been a breach of contract on the part of the respondent. She erred in law by failing to make any such determination. If she made no other error, her failure to determine whether there was a breach of contract may have no practical consequences.
So far as the appellants' claims for damages for breach of statutory duty and negligence are concerned, the position is different. In my view there is nothing impermissible in a trial judge making findings of fact first, and considering the legal consequences of those findings, to the extent necessary, after those findings have been made. Thus, in Amaca Pty Ltd v Booth[2011] HCA 53; (2011) 246 CLR 36, a case about liability for damages for negligence, Gummow, Hayne and Crennan JJ said at [64]:
However, because there was a claim in contract, the appropriate course is for this Court to make its own determination as to whether the respondent breached any contractual duty asserted in the statement of claim.
The respondent admitted that it had entered into a contract with the Langmaids for, amongst other things, the performance of work to upgrade the entire grading area. No doubt it was an implied term of the agreement that the respondent would exercise reasonable care and skill: Chin Keow v Government of Malaysia[1967] 1 WLR 813; Costa Vraca Pty Ltd v Berrigan Weed & Pest Control Pty Ltd[1998] FCA 693; (1998) 155 ALR 714 at 720. That implied term no doubt required the respondent, by its employees, to take reasonable care to avoid starting a fire, and to undertake a reasonably thorough inspection after their hot work ceased as a precaution against the risk of fire.
There was no doubt a foreseeable risk of harm within the meaning of s11(1)(a). The respondent's employees undertook hot work in the vicinity of various combustible items - the wooden box, the plastic bucket, and the rubber components of the grading line. The risk that a spark or some hot spatter would start a fire was not insignificant. The risk of fire could have been obviated by using protective guards to prevent hot work coming into contact with combustible material, undertaking a proper inspection at the site after the hot work had been completed, and wetting down the combustible material in the area of the hot work. Those precautions were required by Australian Standard AS 1674. The risk of a fire was probably very small, but the possible seriousness of a fire was enormous. The precautions required by the Australian Standard were not burdensome. In the circumstances, I am satisfied that there were breaches of contract on the part of the respondent. It follows that the Langmaids must at least be entitled to recover nominal damages.
For the purposes of the appellants' claim for damages for negligence, I accept that the respondent breached its duty of care by failing to take the precautions that I have mentioned.
At the relevant time, reg20(b) of the General Fire Regulations 2000 provided, "A person must ... in performing the hot work, observe the requirements of AS1674". That regulation created an offence punishable by a fine. By virtue of reg3(1), "hot work" had the same meaning as in AS 1674. The respondent admitted in its pleadings that the work undertaken by its employees or agents amounted to "hot work" within the meaning of AS 1674. I accept that a contravention of reg20 could form the basis of a cause of action for breach of statutory duty: O'Connor v SP Bray Ltd[1937] HCA 18; (1937) 56 CLR 464 at 478. And I accept that the respondent breached the statutory duty imposed by reg20 by failing to comply with AS 1674 in the respects that I have mentioned. The critical question is whether the evidence establishes on the balance of probabilities that the fire was caused or contributed to by any such breach of statutory duty, or any other pleaded breach of duty.
There was nothing to prevent an intruder from entering the building that housed the grading line and starting a fire there. However neither Mr Muench nor anybody else found any evidence suggesting the use of any liquid accelerant, nor any evidence to suggest that the first material to start to burn was something not in the vicinity of the grading line. It would be a very surprising coincidence if, in those circumstances, an intruder had started a fire in the very area where hot work had been undertaken during the afternoon of 28 March.
Having regard to all the evidence as to the possible causes of the fire, I consider that the most likely explanation - or, more accurately, the least unlikely explanation - is that something produced by the hot work operations - either a piece of hot metal, a piece of hot slag, or a spark - smouldered undetected until after Mr Maine, Mr Nalder and Mr Andrew Langmaid had all left the premises, and subsequently caused the fire.
But the critical question is not whether it has been established on the balance of probabilities that hot work caused the fire. The critical question is whether it has been established on the balance of probabilities that the respondent's hot work - that is, the hot work undertaken either by Mr Maine or Mr Nalder - caused or contributed to the fire. One cannot rule out the possibility that the ignition source was a spark generated by Mr Cooper's hot work. One cannot rule out the possibility that the fire was unrelated to anybody's hot work. Having regard to the various possibilities and the evidence relating to them, I am not satisfied on the balance of probabilities that the fire was caused or contributed to by the respondent's hot work.
Counsel for the appellants argued that if there was a breach of duty on the part of the respondent, and if a fire could be caused by that breach, then it should be inferred that the fire was caused by the act or omission constituting the breach. He relied on passages in the judgment of Dixon J (as he then was) in Betts v Whittingslowe[1945] HCA 31; (1945) 71 CLR 637 at 649, and in the judgment of French CJ in Amaca Pty Ltd v Booth (above) at [49]. In Betts v Whittingslowe at 649, Dixon J said:
That case concerned an injury to the arm of a boy who was operating a machine. There were various theories as to how the accident happened. The High Court concluded that it should be inferred that the injury was caused by a breach of statutory duty on the part of the boy's employer. The case is not authority for the proposition that evidence of a breach of duty coupled with damage that might thereby be caused is always enough to justify an inference that the damage resulted from the breach. Whether such an inference should be drawn will depend on the circumstances of the particular case. A value judgment always needs to be made as to whether the inference should be drawn or whether, in the words of Dixon J, there is "sufficient reason to the contrary".
In this case, I consider that the evidence establishes breaches of duty on the part of the respondent, and that the fire could have resulted from those breaches of duty. However, having regard to the unlikelihood of the respondent's hot work having caused something to smoulder without being detected, and having regard to the strength of the chance that the respondent's hot work did not cause the fire, I do not think it appropriate to infer that the respondent's hot work did cause the fire.
Since writing the above paragraphs I have had the advantage of reading the draft reasons for judgment of Porter J, which provide an extremely thorough analysis of the evidence and the case law relating to causation. As Dixon J (as he then was) said in Briginshaw v Briginshaw[1938] HCA 34; (1938) 60 CLR 336 at 361:
Strong v Woolworths Limited[2012] HCA 5; (2012) 246 CLR 182 establishes that probabilistic reasoning can be a sufficient basis for actual persuasion of the occurrence of an alleged fact. The respondent's hot work may well be the most probable cause of the fire. However I do not "feel an actual persuasion" that it was the cause of the fire.
As to the "inadequate inspection" issue, I also agree with the reasons of the Chief Justice, subject to one matter which affects the approach but not the conclusion. That matter is the standard of care in relation to the allegation of breach of statutory duty; that is, what is required to discharge the obligation imposed by AS 1674, and in turn, by reg20 of the General Fire Regulations.
Unlike, for instance, reg40 of the Work Health and Safety Regulations2012[1], reg20 of the General Fire Regulations was not conditioned or qualified by any notion of reasonableness. Regulation 20 says that a person must observe the requirements of AS 1674. The Standard itself contains two separate directions that after hot work has been completed, inspections of the site shall be carried out. The first direction mandates a "thorough" inspection "to ensure that the site is safe"; the second requires an inspection to be carried out "to ensure that no smouldering materials remain".
I am unable to agree that what would be needed to satisfy the inspection obligations is a reasonably thorough inspection in the circumstances.
In the context of health and safety legislation, the word "ensure" has been given its ordinary meaning of "to make sure or certain" or "to warrant or guarantee": Director of Public Prosecutions (NSW) v Illawarra Cashmart Pty Ltd[2006] NSWSC 343; (2006) 67 NSWLR 402 at 412 [65] - [66]; Wilton v Cole & Allied Operations Pty Ltd[2007] FCA 725; (2007) 161 FCR 300 at 371 [145] (both citing at 470); The use of the directive "to ensure" that a state of affairs exists or does not exist indicates a high standard of care: ; at 480.
By virtue of the Civil Liability Act, s13(1)(a), the appellants need to establish that it was more probable than not that the respondent's breach of duty was a necessary element of the fire's occurrence. That involves the drawing of inferences. The discharge of the civil burden of proof was discussed by McDougall J (with whom McColl and Bell JJA agreed) in Nguyen v Cosmopolitan Homes (NSW) Pty Ltd[2008] NSWCA 246. At [55], his Honour summarised the position as follows:
In Jones v Dunkel (1959) 101 CLR 298 Dixon CJ at 304 said that an inference must not be drawn where it is but "a choice among rival conjectures". There must be "... evidence supporting some positive inference ... 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". His Honour referred to the above extract from Bradshaw (then unreported but the passage was set out in Holloway v McFeeters[1956] HCA 25; (1956) 94 CLR 470 at 480 - 481), and observed:
It follows, of course, that "a court is not authorised to choose between guesses, even on the ground that one guess seems more likely than another or others": ACCC v Metcash Trading Ltd[2011] FCAFC 151; (2011) 198 FCR 297 per Buchanan J at 306 [31]. However, a court is entitled to draw an inference from "even slim circumstantial facts that exist so long as that goes beyond speculation": Progressive Recycling Pty Ltd v Eversham[2003] NSWCA 268 at [7] per Young CJ in Eq (Ipp JA and Davies AJA agreeing); Condos v Clycut Pty Ltd[2009] NSWCA 200 per McColl JA at [68] (Campbell and Macfarlan JJA agreeing).
Expert opinion of a possibility can be used as circumstantial evidence, and a finding of factual causation may be made where the expert evidence "does not rise above the opinion that a causal connection is possible"; the evidence will be sufficient if, but only if, the materials justify an inference of probable connection: Fernandez v Tubemakers of Australia Ltd[1975] 2 NSWLR 190 per Glass JA at 197. See also Seltsam Pty Ltd v McGuiness (2000) 49 NSWLR 262 per Spigelman CJ at 274 - 275 [79] - [83] and, in the context of a "fire" case, McDonald v Girkaid Pty Ltd[2004] NSWCA 297 per McColl JA (with whom Beazley JA and Young CJ in Eq agreed) at [103], [107].
Evidence of possibility is capable of supporting a probative inference, and expert evidence of possibility may, as circumstantial evidence, alone or in combination with other evidence, establish causation: Seltsam (above) per Spigelman CJ at 276 [89]; McDonald v Girkaid Pty Ltd (above) at [104].
As to the ultimate question of proof on the balance of probabilities by inference, in Lithgow City Council v Jackson (2011) 244 CLR 352 at 386 [94], Crennan J noted that whilst a more probable inference may fall short of certainty, it must be more than an inference of equal degree of probability with other inferences so as to avoid guess or conjecture. Her Honour continued:
As I have noted, the appellants must, of course, prove on the balance of probabilities that the respondent's breach of duty was a necessary element of the occurrence of the harm. That involves processes of inference. In this case the exercise may involve a two-stage factual inquiry as to the probabilities of hot work causing the fire, and then as to whose hot work did so, but the ultimate question is whether it has been shown on the balance of probabilities that the respondent's omissions caused the fire.
As to whether hot work caused the fire, the relevant omissions are the same irrespective of whose hot work it was. The appellants call in aid the approach described by Dixon J in Betts v Whittingslowe (above). The question is whether the omissions amounting to the breach, coupled with an accident of the kind that might thereby be caused, is enough to justify a factual inference about a causal link. In Amaca Pty Ltd v Booth (above) at 57 [49], French CJ highlighted the words "that may thereby be caused", and explained that the logic in the approach "encompasses the case of an ex ante probability, of accident given breach, supported by a causal explanation linking breach and accident". Causal explanations can lie in expert evidence. No doubt those explanations can be general, as in that case, or specific.
Whilst in some cases the inference might be drawn on that basis, that is not a rule of law; "It is an aspect of an available process of drawing of conclusions about causation": Bauldserstone Hornibrook Engineering Pty Ltd v Gordian Runoff Ltd[2008] NSWCA 243 per Allsop P (Beazley and Campbell JJA agreeing) at [240]. In Roads and Traffic Authority v Royal[2008] HCA 19; (2008) 82 ALJR 870, Kirby J at 888 [88] said that Dixon J in Betts had helpfully explained the way in which a finding of a breach of duty "may open the way for (whilst not compelling) an inference of causation-in-fact."[7] See also Kiefel J at 897 [139] - [140]. For that process to alone justify drawing the inference, "the facts proved need to be compelling": Stitz v Manpower Services[2011] QSC 268 at . "The facts must warrant no other inference inconsistent with a defendant's liability": per Dixon J at 649.
Lack of evidence about the precise mechanism of how injury or damage came to be suffered does not necessarily prevent the inference being drawn about the causal link. To begin with, a plaintiff does not have to establish that the precise sequence of events leading to the particular damage was foreseeable. It is sufficient if the general kind or type of injury is reasonably foreseeable: Chapman v Hearse[1961] HCA 46; (1961) 106 CLR 112 at 121; Rosenberg v Percival[2001] HCA 18; (2001) 205 CLR 434, per Gummow J at [64]; Metrolink Victoria Pty Ltd v Inglis(2008) 25 VR 633 per Neave JA at 636 [12][8].
As to causation, the law does not insist on the need to demonstrate the mechanism of loss or damage "to a fine degree": Duma v Mader International Pty Ltd[2013] VSCA 23 per Neave JA at [3]. See also Strong v Woolworths Ltd (above) at 196, [34], and Kuhl v Zurich Financial Services Australia Pty Ltd[2011] HCA 11; (2011) 243 CLR 361 per French CJ and Gummow J at 381 - 382 [53] - [54].
It was alleged in the statement of claim that one or more hot products came into contact with combustible materials. Those combustible materials were identified as the wooden bin, one or more PVC buckets, a wooden pallet, wood offcuts, or the conveyor belt. The ultimate focus seemed to be on the conveyor belt, the roller, the bin and the buckets. However, there was no legal requirement, nor any factual necessity, to establish that any one of these things in particular provided the initial fuel source.
(1938) 61 CLR 286
(2011) 246 CLR 36
(1998) 155 ALR 714
(1937) 56 CLR 464
(2003) 214 CLR 118
(2003) 77 ALJR 1449
(2004) 78 ALJR 572
(2004) 79 ALJR 452
(2005) 79 ALJR 1104
(2008) 82 ALJR 870
(1945) 71 CLR 637
(1938) 60 CLR 336
(2012) 246 CLR 182
(2006) 67 NSWLR 402
(2007) 161 FCR 300
(1985) 11 IR 467
(1999) 89 IR 464
(1979) 142 CLR 531
(1951) 217 ALR 1
(1959) 101 CLR 298
(1956) 94 CLR 470
(2011) 198 FCR 297
(2000) 49 NSWLR 262
(2011) 244 CLR 352
(1961) 106 CLR 112
(2001) 205 CLR 434
(2008) 25 VR 633
(2011) 243 CLR 361
(1995) 185 CLR 410
(1988) 164 CLR 535
In my view there is no reason to doubt the correctness of Mr Muench's evidence ruling out the switchboard, electric welder and computer equipment as sources of ignition.
Something had been destroyed in the extreme northeast corner of the building that housed the grading line. Mr Thomas was cross-examined about that, but was unable to say what had been destroyed in that area. The learned trial judge, who no doubt had more of a feel for the case than I have, commented at [119] of her reasons that an unidentified burnt object in that corner was a possible cause of the fire. At [134] of her reasons she said:
Even accepting, however, that the standard of compliance is higher than what is reasonable in the circumstances, I share the Chief Justice's view that the appellants have not established that a failure to undertake an inspection after the hot work was of any consequence. The way in which the fire must have started, and the slow pace with which it must have developed, becomes more plainly apparent when considering the next question. It is possible that an inspection of the required standard may have revealed combustion, but is equally possible that it may not.
The next question is the one described by Blow CJ as the "critical question" of whether the evidence establishes on the balance of probabilities that the respondent's hot work - that of Rodney Maine or perhaps Scott Nalder - caused or contributed to the fire.
For the reasons which follow, I think that the trial judge erred in her approach to the resolution of the question, and in my view the evidence does establish, on the balance of probabilities, the requisite causal connection between the Mr Maine's activities and the respondent's breach of duty, and the fire. I am conscious that I have not had the advantages of the trial judge. They include the "feeling" of the case which a reading of the transcript cannot fully provide: Fox v Percy[2003] HCA 22; (2003) 214 CLR 118 at 125 - 126 [23]. However, in deciding what is the proper inference to be drawn, I have given respect to the conclusion of the trial judge, but I have reached a different conclusion, and I should give effect to it: Warren v Coombes[1979] HCA 9; (1979) 142 CLR 531 at 551.
In this case, the fire was very obviously an outcome of the kind that might well be caused by the failures to comply with cls 2.7, 2.8 and 2.10 of AS 1674. The opinion of both Mr Muench and Mr Thomas, expressed as a matter of probabilities, was that it was caused by hot work.
Arson and electrical faults in various pieces of equipment were excluded as causes. The trial judge found accordingly, and there is no reason to suggest that those findings were unreasonable or for some other reason ought not to have been made. The factual basis for the opinions was not undermined, and the trial judge seems to have accepted their evidence. Even if expressed as a possibility, these opinions, as I have shown, form part of a circumstantial case and may even of themselves form a foundation for an inference to be drawn on the balance of probabilities.
The concept of a hot metal particle causing fire which ultimately led to flaming combustion was explored in the evidence. It was dealt with by Mr Thomas in some detail. Both men were of the view that the fire was of slow build-up. The expert opinion and the body of knowledge showed that smouldering ignitions from hot works have been known to take hours. I do not believe it is correct to say that there was no evidence as to how the "phenomenon might occur", nor do I see that there is any significance in the fact that there was no evidence about the precise number of hours over which the type of smouldering might occur. Different factors would operate in different situations.
Both Mr Muench and Mr Thomas were quite comfortable with the notion that in this case, hot works commenced a combustion which went undetected for a period during which people were present, and which later slowly built into a destructive fire. Mr Bassett accepted the possibility of the fire starting in those sorts of circumstances, and agreed that his testing had its limitations in attempting to assess what did or might happen in a particular environment. It might be noted that his testing did not establish that it was not possible to cause a flame in any of the combustibles he tested.
When considering the question of discovery, the activities of the men involved need to be borne in mind. There was no evidence as to whether Mr Cooper did or did not do anything. Mr Maine said that he did not carry out any inspection for signs of fire. Mr Nalder said that after packing up they inspected the grading line "pretty thoroughly", but that this was because they had to work out what they were going to do the next day. Mr Langmaid was also assessing the stage of the work and was not looking for signs of fire. The failure to detect any smell is explicable by a combination of the stage of combustion, and the expert evidence that the smell might well have been disguised by odours generated by the hot work.
It might be thought that it is unlikely a fire would occur in the circumstances established in this case, but there is a clear distinction between the prospective assessment of a risk of something happening, and the retrospective exercise of making findings of fact about what has happened. In Amaca Pty Ltd v Booth (above) French CJ said (omitting references):
The precautions adverted to in AS 1674 relate to the protection of places where there is any possibility of starting a fire, and to materials which might give rise to a fire. The particulars of negligence and breach of agreement are, in part, directed to those general things. The expert evidence was that the ignition source was hot work particles but the likely initial fuel source was destroyed in the fire. The inability to precisely identify what it was is no impediment to the drawing of the inference and being affirmatively satisfied that, on the balance of probabilities, hot work caused the fire.
In terms of the competing possibilities as to whose hot work caused the fire, the following matters make it more likely than not, in my view, that it was that of Mr Maine. Mr Cooper used a grinder; Mr Maine used a welder. The particles produced from the welding process are generally larger and significantly hotter. The sparks from a grinder tend to be thrown directly in line with the rotating grinding disc. Subject to the placement of guards, welding spatter can go in any direction. The evidence, and what can be seen from the footage of Mr Bassett's tests, show that although seconds are involved, welding spatter retains its "red heat" for longer.
Mr Maine's welding was done at a more elevated level than Mr Cooper's cutting. Further, it may be of little weight, but although the evidence of the two men would tend to suggest that they were each working at roughly about the same distance from the wooden bin, an assessment of the photographic evidence - bearing in mind the difficulties associated with interpreting photographs - would tend to suggest that Mr Maine's work was closer to the area of origin.
A significant aspect to be considered is the time at which the work was performed. Mr Maine's work was done about three hours after that of Mr Cooper. Mr Maine's welding finished at about 7pm. Mr Maine and Mr Nalder were in and about the premises up until about 7.45pm. At about the same time, although apparently after Mr Maine and Mr Nalder had gone, Mr Langmaid came in and looked at the grading line to assess the progress of the work.
A combination of the expert evidence, the reference materials and common sense would strongly suggest that the longer the period of time between contact between a hot particle and a fuel source, the greater the likelihood of a fire or fire potentiality would be discovered. It is far more likely that a fire or the potentiality of one, would have been discovered in the 3¾ hours between Mr Cooper's hot work and when the last person left, than in the ¾ of an hour which went by after Mr Maine's work.
In my respectful view, the trial judge's concluding remarks reveal errors in the approach to whether or not the relevant inference could be drawn. No reference was made to the reasoning process explained in Betts v Whittingslowe. The accepted expert opinion as to cause seems to have been overlooked as part of the evidence from which the inference could be drawn. These matters are made more acute when put alongside the fact that her Honour seems to have accepted the real possibility that hot work caused the fire. I am satisfied that the trial judge erred in that her Honour ought to have been satisfied on the balance of probabilities that the respondent's hot work caused the fire, and that the established breach of duty was a necessary element of the occurrence of the damage.
In my view, the whole of the evidence justifies the drawing of the inference. Obviously, I do not consider that this is mere speculation, nor that it is a matter of choosing between guesses. On the evidence, I am affirmatively persuaded that the respondent's breach of duty was a necessary element of the occurrence of the damage.