Torts - Negligence - Essentials of action for negligence - Where nervous shock or mental disorder - Whether psychiatric injury caused by telephone call informing parents that child of 16 years was missing from his place of employment or learning of child perishing in desert months after death - Caution in expanding categories of negligence or classes of plaintiffs - Policy considerations - Distinction between grief and psychiatric harm - Duty of care to avoid psychiatric injury involves analysis of interrelated elements of duty, breach, damage - Foreseeability - Normal standard of susceptibility - Sudden shock - Proximity relevant to psychiatric injury claims - Causal proximity - Aftermath - Direct perception - No appropriate reliance by plaintiffs on employer

Abramzik v Brenner (1967) 65 DLR (2d) 651

Burnie Port Authority v General Jones Pty Ltd [1994] HCA 13; (1994) 179 CLR 520

Cormier v Dixon (1992) 127 NBR (2d) 358; (1992) 130 NBR (2d) 69

De Franceschi v Storrier (1988) 85 ACTR 1

Dulieu v White [1901] EWHC KB 1; [1901] 2 KB 669

Duncan v British Coal Corp [1997] 1 All ER 540

Farrugia v Great Western Railway Co [1947] 2 All ER 565

Hambrook v Stokes Bros [1925] 1 KB 141

Hinz v Berry [1970] 2 QB 40

Hunter v British Coal Corporation [1998] EWCA Civ 834; [1999] QB 140

King v Phillips [1953] 1 QB 429

Page v Smith [1995] UKHL 7; [1996] 1 AC 155

Quayle v State of New South Wales [1995] Aust Torts Reports 62,792

Ravenscroft v Rederiaktiebolaget Transatlantic [1992] 2 All ER 470

Reeve v Brisbane City Council [1995] 2 Qd R 661

Robertson v Forth Road Bridge Joint Board (No 2) [1995] ScotCS CSIH_1; 1994 SLT 568; 1996 SLT 263

Scala v Mammolitti [1965] HCA 63; (1965) 114 CLR 153

Stergiou v Citibank Savings Ltd (1998) 148 FLR 244

Strong v Moon (1992) 13 CCLT (2d) 296

Talibi v Seabrook (1995) 28 CCLT (2d) 254

Taylorson v Shieldness Produce Ltd [1994] PIQR P329

1 MALCOLM CJ: In my opinion, this appeal should be dismissed for the reasons to be published by Ipp J. I only wish to add a number of comments of my own.

2 I share his Honour's view of the problems associated with the determination of the question at issue on the basis of the facts admitted on the pleadings and the facts the Court was invited to assume to be true. In particular, problems arise in relation to the pleadings in par 8(c), (d) and (e), par 9 and par 9A of the statement of claim.

3 In my view, it is right to assume that the duty of care referred to in par 8(c) of the statement of claim was a duty to take reasonable care of the safety of the appellants' son James which duty extended to the respondent's duty as employer to provide a safe system of work. I would add in relation to this that I would be prepared to assume that the duty extended to ensuring that James was properly supervised. I agree with Ipp J that, as a result of what was agreed between the parties, the Court was required to assume that if the respondent breached the duty of care it owed to James, thereby causing him to die in the desert, the appellants would suffer "particular distress". It does not follow that it was reasonably foreseeable that they would suffer nervous shock in the form of psychiatric injury or illness.

4 It was pleaded in par 8(d) of the statement of claim that:

"The [respondent] knew of the ongoing concern of the [appellants] in relation to the supervision of [James] and were on notice by reason of their earlier and ongoing enquiries that if there was a breach of the duty of care owed by the [respondent] as the employer of [James] to [James] that there was a foreseeable risk that the parents would suffer not only a grief reaction but in addition a reaction extending beyond grief to an entrenched psychiatric condition of the type which has since developed."

5 There was a clear allegation that the respondent knew that the appellants were concerned that James should be properly supervised by the respondent. It does not follow, however, that because the respondent was "on notice" of such concern that, if the respondent breached that duty, there was a foreseeable risk that the appellants would suffer not only a grief reaction but "... an extended psychiatric condition of the type which has since developed".

6 There is a further question lurking in this pleading, namely, whether it ought to have been reasonably foreseen by the respondent that if James was posted to an isolated station property on his own, he would attempt to abscond by unlawful use of the respondent's motor vehicle, get bogged and fail to survive in the desert.

7 It is in this particular context that the comments of Lord Bridge in McLoughlin v O'Brian [1982] UKHL 3; [1983] 1 AC 410 at 432 are pertinent. Foreseeability must depend upon "what knowledge is to be attributed to the reasonable man of the operation of cause and effect in psychiatric medicine". The law has required the reasonable person in this context to be a person of "normal fortitude" and to conform to "normal standards of susceptibility": see Jaensch v Coffey [1984] HCA 52; (1984) 155 CLR 549 at 556 per Gibbs CJ and at 568, 570 - 571, 572 and 578 per Brennan J; and Morgan v Tame [2000] A Tort Rep 63,870. The question of reasonable foresight is more than a question of fact because it involves a value or qualitative judgment about the standard of reasonableness of human behaviour or capacity: cf Morgan v Tame, supra, at 63,872 - 63,874 per Spigelman CJ; at 63,887 - 63,888 per Mason P; and at 63,891 per Handley JA.

8 Although reasonable foresight is to a substantial extent a question of fact, there is a qualitative judgment which is required to be made by the Court, in the light of all of the evidence, to determine whether the respondent knew or ought to have known of the foreseeable risk of psychiatric injury in the circumstances. This issue was at the heart of the present case. It could only be determined in the light of all of the evidence and by making the qualitative judgment to which I have referred.

9 In these circumstances I agree with Ipp J that the parties could not have intended that the allegation that the respondent had knowledge (actual or constructive) of the foreseeable risk of psychiatric injury was a fact which this Court was invited to assume to be true. If it was, I am not prepared to act on the assumption for the reasons I have mentioned. In any event, it was an alleged fact the existence of which was a live issue and the subject of controversy before Heenan J. The same comments apply to the allegations in par 8(e) and par 9 of the statement of claim.

10 I also agree with the comments of Ipp J regarding par 9A. The proper approach to the matter is to deal with the case on the assumption that the appellants suffered psychiatric injury, both on learning that their son James was missing and believed to have run away, and later when they learned of his death, taken in conjunction with what they had been told earlier.

11 In my opinion, given the state of the authorities reviewed by Ipp J, it would not be appropriate for this Court to ignore the policy considerations referred to by Lord Steyn in White v Chief Constable of South Yorkshire Police [1998] UKHL 45; [1999] 2 AC 455 at 493 - 494; and the operation of the "control mechanisms" referred to by Lord Hoffman in that case at 504 - 506. See also Devji v Burnaby District (1999) 180 DLR (4th) 205 at 222 per McEachern CJ. Reasonable foresight alone is not enough. In Caparo Industries plc v Dickman [1990] UKHL 2; [1990] 2 AC 605 at 633, Lord Oliver said that a duty to use reasonable care to avoid reasonably foreseeable financial harm needed to be constrained by "some intelligible limits". Gleeson CJ agreed with this proposition in Perre v Apand Pty Ltd [1999] HCA 36; (1999) 198 CLR 180 at 192. The Chief Justice said that the consequence of the adoption of reasonable foresight as the touchstone of a general duty not to cause financial harm would be "intolerable".

12 I am indebted to Ipp J for drawing attention to the relevance in the present context of the warnings expressed by Rich J in Chester v Waverley Municipality [1939] HCA 25; (1939) 62 CLR 1 at 10 - 11. Additional reasons for rejecting the adoption of reasonable foreseeability as the sole criterion for the existence of a duty of care were stated by Brennan J in Sutherland Shire Council v Heyman [1985] HCA 41; (1985) 157 CLR 424 at 481. Given the current state of authorities it is not for this Court as an intermediate appellate court to extend the scope of the liabilities for psychiatric injury or nervous shock any further than the authorities to date have indicated. Such liability is dependent upon the existence of an independent duty of care and reasonable foreseeability of harm constrained by the tests of "normal fortitude" and "sudden shock". As to the requirement of "normal fortitude" see Bunyan v Jordan [1937] HCA 5; (1937) 57 CLR 1 at 14 per Latham CJ; and at 16 - 17 per Dixon J; Jaensch v Coffey, supra, at 556, 568 and 570 - 572 per Brennan J; at 557 per Murphy J and 610 per Deane J, as cited by Spigelman CJ in Morgan v Tame, supra, at 63,873; and by Mason P at 63,887. In that case, Handley JA agreed with both Spigelman CJ and Mason P. As to the requirement of "sudden shock" see Jaensch v Coffey, supra, at 566 - 567 per Brennan J; and the authorities referred to by Spigelman CJ in Morgan v Tame, supra, at 63,877; and Mason P at 63,889. Handley JA agreed with Spigelman CJ and Mason P in this respect. To the authorities there cited, Ipp J has added Mount Isa Mines Ltd v Pusey [1970] HCA 60; (1970) 125 CLR 383 at 407 per Windeyer J. In my opinion, with respect, Morgan v Tame was correctly decided and this Court should apply the law as stated in that case.

13 In relation to the present case, I agree with Ipp J that neither of the alleged consequences of the telephone call on 6 December 1986 when Mrs Annetts was informed James was missing, nor when Mr and Mrs Annetts were informed of his subsequent death, were reasonably foreseeable. Naturally, it was reasonably foreseeable that the appellants as parents would be deeply concerned when their son went missing and would suffer a deep sense of loss and grief upon learning of his death. It does not follow, however, that it was reasonably foreseeable that they would suffer psychiatric injury in the sense of a recognised psychiatric illness as a result of learning of either of these events. Over the period between 6 December 1986 when their son went missing and the end of April 1987 when Mr Annetts identified his remains from a photograph, persons of "normal fortitude" ought reasonably have been expected to realise that his prospects of survival would have become more and more remote, so that confirmation that he died would not constitute a "sudden shock" or, at least such a sudden shock that would make the onset of psychiatric illness reasonably foreseeable.

14 Even if that be wrong, Ipp J has demonstrated that the test of proximity in the context of proceedings for damages for nervous shock or psychiatric injury as formulated by Deane J in Jaensch v Coffey at 584 - 585 has not been satisfied. The test has been relevantly expressed as including physical proximity in the sense of space and time. As his Honour made clear at 606:

"The requisite duty relationship will not, on the present state of the law, exist in a case where mere psychiatric injury results from subsequent contact, away from the scene of the accident and its aftermath, with a person suffering from the effects of the accident."

15 In Victorian Railway Commissioners v Coultas (1888) 13 App Cas 222, the objection to recovery for nervous shock suffered by a person not otherwise physically injured was denied on the ground that such damage was too remote. Subsequently, however, in Bourhill v Young [1942] UKHL 5; [1943] AC 92 at 103, Lord MacMillan said:

"It is no longer necessary to consider whether the infliction of what is called mental shock may constitute an actionable wrong. The crude view that the law should take cognizance only of physical injury resulting from actual impact has been discarded, and it is now well recognised that an action will lie for injury by shock sustained through the medium of the eye or the ear without direct contact. The distinction between mental shock and bodily injury was never a scientific one, for mental shock is presumably in all cases the result of, or at least accompanied by, some physical disturbance in the sufferer's system. And a mental shock may have consequences more serious than those resulting from physical impact. But in the case of mental shock there are elements of greater subtlety than in the case of an ordinary physical injury and these elements may give rise to debate as to the precise scope of legal liability."

16 In that case, of course, a limitation in terms of the duty of a motor cyclist was owed to such persons as he could reasonably foresee might be injured by his failure to exercise reasonable care. A test of proximity was then adopted which required the plaintiff to be within the area of "potential danger" arising as a result of the negligence. This confined scope of the duty has, of course, been subsequently extended. Earlier approaches in such cases as Dulieu v White [1901] EWHC KB 1; [1901] 2 KB 669 per Kennedy LJ at 675; and Hambrook v Stokes [1925] 1 KB 141 suggested other arbitrary tests.

17 The cases, however, maintain a distinction between the primary victims in an accident or other event and claims by secondary victims who sustain psychiatric illness as a result of their association with the primary victim. In such cases, as Deane J explained in Jaensch v Coffey at 606 - 608, there has to be "causal proximity" between the relevant negligence and the resulting injury. In that case, there was a need for proximity in space and time in the sense that the plaintiff would need to be a witness to the accident or at least to its "aftermath". What a plaintiff was told in the context of the "aftermath" would also be relevant. Consequently, more remote consequences would be outside the scope of the requirement of proximity. In Jaensch v Coffey, while the respondent was not present at the accident, she suffered psychiatric injury as a result of the combined effect of what she saw at the hospital where her husband was in intensive care and what she was told. In that case, the element of proximity was present because "the aftermath" was regarded as including the journey by ambulance to the hospital and what took place at the hospital itself. Gibbs CJ at 551 agreed with Deane J; and Dawson J at 612 accepted that it was not necessary for a plaintiff to observe the consequences of the defendant's breach of duty at the scene of the accident. It was sufficient that they:

"... may be observed as part of the aftermath and the aftermath may extend to the journey by ambulance to hospital and to the scene at the hospital itself."

18 The approach adopted by Deane J was followed by the Queensland Full Court in Spence v Percy [1992] 2 Qd R 299 and the other cases referred to by Ipp J. These cases demonstrate that the appellants in this case were required to show that they were in a relationship of proximity to the respondent.

19 In this context, the appellants face another difficulty in that they did not directly perceive what happened to their son at the time when the events occurred or during any relevant aftermath. Direct perception of what happened or of the aftermath has been held to be an essential element of proximity by the House of Lords in England and in Canada at the intermediate appellate level. It is not a requirement in South Africa and the issue has not yet been finally resolved in Australia. The relevant cases are referred to by Ipp J. As his Honour points out, apart from the occasion in January 1987 when Mr Annetts saw a blood-covered hat belonging to James, and when he identified his remains from a photograph seen some five months after his death, there was no direct perception of the consequences of the respondent's breach of duty. I agree that this did not satisfy the direct perception requirement which, as Ipp J has demonstrated, is a requirement accepted by Windeyer J in Mt Isa Mines v Pusey, supra, and Brennan J in Jaensch v Coffey, supra, and required by the House of Lords in White v Chief Constable of South Yorkshire; and Alcock v Chief Constable of South Yorkshire Police [1991] UKHL 5; [1992] 1 AC 310. Finally, I associate myself entirely with the views expressed by Ipp J regarding the absence of proximity by reason of remoteness, upon the assumption that direct perception of the event or its aftermath is not an essential element of proximity.

20 PIDGEON J: I agree with the reasons to be published by Ipp J and with the further reasons of the Chief Justice.

The preliminary issue: the question as formulated

21 This is an appeal from a decision by Heenan J on a preliminary issue. The preliminary issue in question is whether, on certain facts admitted by the respondent, and on the assumption that certain other facts are true, the

respondent owes a duty of care to the appellants "to exercise reasonable care and skill to avoid causing them psychiatric injury". Heenan J answered this question in the negative and the appeal is against his Honour's decision.

22 The preliminary issue was intended to raise a question of law only and the grounds of appeal concern only the legal principles to be applied in answering the question posed. Unfortunately, there are two practical problems with the preliminary issue as formulated.

23 Firstly, certain of the facts to which the Court is required to have regard are not entirely clear. No evidence was led and the admitted facts and the facts to be assumed are, largely, facts pleaded by the appellants. The pleadings, as one would expect, do not go into the detailed particulars that would be helpful in clarifying many of the events and circumstances on which the duty of care depends. Accordingly, the Court is required to make do with a less than complete picture.

24 The parties agreed that "the facts pleaded" in certain paragraphs of the statement of claim are to be assumed to be true. Certain of the paragraphs in question, however, contain mixed assertions of facts, conclusions of law, rolled up statements containing arguments as to inferences and factual assertions. Others allege facts that are conditional upon alleged breaches of duty (which are not taken to be admitted). In these circumstances, it is not always clear what are "the facts pleaded" and what are the facts admitted or to be assumed to be true.

25 Secondly, by reason of the particular nature of the appellants' claim, it is at times difficult to separate the duty of care from the other elements of the tort of negligence alleged by the appellants. I refer below to the particular problems that, in consequence, arise.

26 James Annetts was the appellants' son. James was born on 13 March 1970 and lived with the appellants in their home in Binya in New South Wales. In August 1986 he left the family home and went to work for the respondent as a jackaroo at Flora Valley, one of the respondent's cattle stations. Flora Valley is situated about 40 kilometres southeast of Halls Creek in the Kimberley district of Western Australia.

27 The respondent knew that the appellants were James' parents. The respondent admits that the appellants relied upon it to supervise James "as a 16-year-old child" and that, to its knowledge, they "entrusted" it with his care and welfare.

28 Before James left home, the second-named appellant (Mrs Annetts) telephoned Mrs Loder, the wife of the respondent's station manager, and inquired about the conditions under which James would be living and the supervision that would be provided. Mrs Loder told Mrs Annetts that James would be working at Flora Valley under constant supervision and would be sharing a room with one to four other men, that all of his meals would be supplied for him, and that he would be well looked after. On the day after James arrived at Flora Valley, Mrs Annetts telephoned Mrs Loder to check that he had arrived safely. As a general proposition, the respondent admits that the appellants had made inquiries of its servants or agents in relation to arrangements for James' safety and they had received assurances in relation thereto.

29 In fact, James spent only seven weeks at Flora Valley. On 13 October 1986, Mr Loder, the respondent's station manager, sent him to work alone as caretaker at Nicholson Station, about 100 kilometres east of Flora Valley and about 270 kilometres north of Balgo. On 3 December 1986 the respondent learned that James was missing and on that date it had reason to suspect that James was in grave danger of injury or death. It was only on 6 December 1986, however, that the appellants were informed that James was missing.

30 On 6 December 1986, a police officer at Griffiths, New South Wales, telephoned the first-named appellant (Mr Annetts) and told him that James was missing from his place of employment and was believed to have run away. Mr Annetts collapsed and Mrs Annetts took over the telephone conversation. No other facts are known to the Court concerning these conversations.

31 At some time, not revealed by the facts before the Court, an intensive search was begun for James and another teenager, Simon James Amos, who had been employed by the respondent as a jackaroo on another station. Thereafter, the appellants had a number of telephone conversations with police officers at Halls Creek police station, Mr Loder, and numerous other persons in the Halls Creek area concerning the whereabouts of their son. In January 1987, the appellants went to Halls Creek where they remained for some four to five days. They were then shown some of their son's belongings, including a hat covered in blood. Thereafter, on several occasions until the end of April 1987, the appellants went to the Halls Creek area in attempts to obtain information about James.

32 On 26 April 1987, Mr Annetts was informed by telephone that the vehicle driven by James had been found bogged in the desert but there were no signs of any people around it. Later that day, he was told that two sets of remains had been found nearby. On 28 or 29 April 1987, Mr Annetts, alone, returned to Halls Creek. At the police station, he was shown a photograph of a skeleton and he identified it as being that of James.

33 The parties accept that, in fact, James "died on or about 4 December 1986 in the Gibson Desert some 133 kilometres south of Balgo as a result of dehydration, exhaustion and hypothermia". Thus, the appellants learned of his death almost five months after it occurred. They were far away from James when he died.

34 I now turn to those facts about which there is a lack of clarity. The difficulties in question arise from par 8(c), par 8(d), par 8(e), par 9 and par 9A of the statement of claim, the parties having agreed that "the facts pleaded" in those paragraphs (amongst others) are to be assumed by the Court as true.

"[T]he [respondent] knew that if there was a breach of the duty of care owed to [James], that [sic] he may die in circumstances of particular distress to his parents having regard to his manner of death namely perishing in the desert."

The "duty of care owed to [James]" is not specifically identified in the statement of claim, save that in par 8(d) thereof, reference is made to "the duty of care owed by the [respondent] as the employer of [James] to [James]". I infer that the duty of care alleged is the duty of an employer owed to an employee to provide a safe system of work. It follows that the Court is required to assume that the respondent knew that if it breached that duty it owed to James, thereby causing James to die in the desert, the appellants would suffer "particular distress".

36 Paragraph 8(d) of the statement of claim alleges:

"[T]he [respondent] knew of the ongoing concern of the [appellants] in relation to the supervision of [James] and were on notice by reason of their [sic] earlier and ongoing inquiries that if there was a breach of the duty of care owed by the [respondent] as the employer of [James] to [James] that there was a foreseeable risk that the parents would suffer not only a grief reaction but in addition a reaction extending beyond grief to an entrenched psychiatric condition of the type which has since developed."

37 Plainly, par 8(d) alleges that the respondent knew that the appellants were concerned that it should supervise James. It is questionable, however, whether par 8(d) asserts that the respondent, in fact, knew that, if it committed a breach of the duty it owed James, there was a foreseeable risk that the appellants would sustain psychiatric injury. Importantly, the allegation is not made in par 8(d) that the respondent in fact knew that there was a foreseeable risk that, in the circumstances postulated, the appellants would sustain psychiatric injuries. The allegation is that the respondent was "on notice by reason of [the appellants'] earlier and ongoing inquiries" that there was a foreseeable risk as alleged. In other words, the allegation is that the respondent should have inferred from the inquiries that the appellants had made that there was such a foreseeable risk.

38 Whether a defendant in fact foresaw a particular risk, is a question of fact to be decided by reference to the usual criteria relating to credibility of witnesses. Whether a defendant, objectively, ought to have foreseen a particular risk, falls into a different category, and the criterion of a reasonable person has to be applied. In the context of a claim for psychiatric injury, the question of foreseeability "depends on what knowledge is to be attributed to the hypothetical reasonable man of the operation of cause and effect in psychiatric medicine": McLoughlin v O'Brian [1982] UKHL 3; [1983] 1 AC 410 at 432 per Lord Bridge. Moreover, in determining the question of foreseeability the test of an "ordinary person of normal fortitude" or "normal standards of susceptibility" is to be applied: Jaensch v Coffey [1984] HCA 52; (1984) 155 CLR 549 at 556 per Gibbs CJ, and at 568, 570 - 572 and 578 per Brennan J; see also Morgan v Tame [2000] A Tort Rep 63,870 at 63,872 - 63,874 per Spigelman CJ, at 63,887 - 63,888 per Mason P and at 63,891 per Handley JA. The resolution of these issues depend on more than findings of primary fact by the trial Judge and this underlies the observation by Mason P in Morgan v Tame (at 63,887) that "reasonable foreseeability is more than an issue of fact".

39 Against this background of the applicable legal principles, I have come to the conclusion that the allegation as to the respondent's knowledge of the foreseeable risk of psychiatric injury was not intended by the parties to be a fact that the Court was to assume to be true. The fact that the earlier inquiries were made is to be assumed as true, but not the allegation that in consequence the respondent ought to have known that there was a foreseeable risk of psychiatric injury should it breach its duty of care owed to James as his employer.

40 The answer to the question whether there was a reasonably foreseeable risk of psychiatric injury involves inferences being drawn from proved facts so as to determine foreseeability in an objective sense. This task is ordinarily a matter for judicial judgment. In the absence of unequivocal agreement by the parties as to the resolution of this question, I consider that it should be treated as having been left open for determination by the Court. I am reinforced in this conclusion by the fact that, at the trial, it was not assumed that the respondent had admitted that it ought to have known of the foreseeable risk of psychiatric injury. This was a live issue argued by the parties before Heenan J and adjudicated upon by his Honour in the appellants' favour. In my view, it should be treated as an open question on the appeal.

"[T]he [appellants] also allege that as parents of [James] they were within the range of persons who an employer owed a duty of care to and that breach of the employer's duty of care resulting in death to a young employee such as [James], would be likely to cause psychiatric injury to near relatives."

The comments I have made in regard to par 8(d) apply equally to this paragraph. In my opinion no facts are contained in par 8(e) which the Court is required to assume as being true.

"By reason of the relationship of proximity identified by the facts and matters referred in paragraph 8 above, the [respondent] owed a duty of care to the [appellants] as the parents of [James], that as his employer they would exercise reasonable care in the supervision of him and otherwise in the system of work that was put in place so as not to be in breach of their obligation to him in circumstances which could and did cause his death."

There are no facts alleged in par 9, which the Court is required to assume as being true.

"As a result of the matters referred to in paragraphs 2 - 7F hereof the [appellants] suffered injury."

This allegation contains within it several ambiguities, and is not a satisfactory basis for the preliminary issue as defined. Is the "injury" alleged in par 9A a "grief reaction", as pleaded in par 8(d), or a psychiatric injury, in the sense of a recognised psychiatric illness? Counsel for the appellants argued only for the latter, and I shall treat the allegation in that way. Is it alleged that the appellants suffered psychiatric injury on 6 December 1986 when they received the telephone call advising them that James was missing from his place of employment and was believed to have run away? There is no way of determining this, so for the purposes of these reasons I shall assume that that allegation is made. Then, is it alleged that the appellants suffered psychiatric injury at some time between 6 December 1986 and early January 1987 when they travelled to Halls Creek, spoke to various persons and were shown James' belongings including a hat covered in blood? If so, is it alleged that the psychiatric injury was caused by some particular incident or by the gradual accumulation of the relevant events? If the former, what precisely is the causative incident and when did it occur? I am afraid that there is no way in which these ambiguities can be resolved.

44 The position gets worse when regard is had to the allegations as to what occurred towards the end of April 1987. Is it alleged that the appellants suffered psychiatric injury at some time between 6 December 1986 and the end of April 1987? And again, is it alleged that the psychiatric injury was caused by some particular incident or by the gradual accumulation of the relevant events? There is particular force that attaches to the events towards the end of April 1987 as it was then that the appellants learned positively that James had died and Mr Annetts viewed the photograph of the remains. But it is not clear whether the appellants' claims are for psychiatric injuries caused by James' death and the circumstances under which they learned of it, or whether they claim that their psychiatric injuries pre-dated that knowledge.

45 Paragraph 9A alleges that the appellants suffered "injury" as a result of a broad range of matters, commencing with the telephone advice on 6 December 1986. Furthermore, par 9B of the statement of claim alleges that "it was at all material times reasonably foreseeable that in the event of injury or death befalling the child that the parents could thereby suffer psychiatric injury". On the other hand, par 8(c), par 8(e) and par 9 base the operative allegations relating to causation only on the death of James. Moreover, counsel for the appellants opened the appeal by saying that the appellants' claim was one "for damages for nervous shock suffered by the appellants arising from the death of their son ..." and did not refer to psychiatric injury being caused by events that pre-dated their knowledge of James' death. Counsel's entire argument was predicated upon the psychiatric injury being caused by the appellants learning of the death of James, not by what occurred earlier.

46 In the circumstances, I shall first assume that the appellants sustained psychiatric injury on 6 December 1986, caused by the shock of the advice that James was missing from his place of employment and was believed to have run away. Then, I shall assume, in the alternative, that they sustained psychiatric injury upon learning of James' death, the injury being caused by the shock of that advice coupled with what had transpired earlier. I think that these two scenarios cover all reasonable possibilities. Any variation would not have any significant legal consequence.

Policy issues that apply to claims for psychiatric harm

47 In Australia, the law relating to claims for damages for psychiatric injury based on negligence has not yet reached a state of certainty. Differences in judicial and academic opinion on various issues relating to the topic abound in this country and, indeed, throughout the common law world. Many learned articles have been produced and the relevant literature is extensive. It is the task, however, of an intermediate appellate court to approach the legal issues that arise conservatively, bearing in mind that, as this aspect of the law is still in the course of development, courts must act with caution, in company and not alone: Mount Isa Mines Ltd v Pusey [1970] HCA 60; (1970) 125 CLR 383 at 396 and 403; see also Sutherland Shire Council v Heyman [1985] HCA 41; (1985) 157 CLR 424 where Brennan J, in the course of commenting on the proposition that foreseeability of injury should be the sole criterion of a prima facie duty of care, said (at 481):

"It is preferable, in my view, that the law should develop novel categories of negligence incrementally and by analogy with established categories, rather than by a massive extension of a prima facie duty of care restrained only by indefinable 'considerations which ought to negative, or to reduce or limit the scope of the duty or the class of the person to whom it is owed.' "(See also at 497 - 498 per Deane J.)

48 There are particular reasons for the need to tread carefully in extending liability for psychiatric injury from those circumstances, which are by now firmly recognised by the law. These are the policy considerations that have always weighed heavily with courts from the time claims of this kind made their first appearance. It is important to bear these in mind. They were conveniently summarised by Lord Steyn in White v Chief Constable of South Yorkshire Police [1998] UKHL 45; [1999] 2 AC 455. His Lordship (at 493 - 494) described four distinctive features of psychiatric harm "which in combination may account for the differential treatment" between claims for damage of that kind and other claims based on negligence. First, there is the complexity and difficulties involved in drawing the line between acute grief and psychiatric harm. Secondly, there is the effect of the expansion of the availability of compensation on potential claimants who have witnessed gruesome events. His Lordship here had in mind the unconscious effect of the prospect of compensation on potential claimants. This is a concern of many judges in all common law countries, see for example Devji v Burnaby District (1999) 180 DLR (4th) 205 at 222 per McEachern CJ. Thirdly, his Lordship observed that the abolition or a relaxation of the special rules governing the recovery of damages for psychiatric harm would greatly increase the class of persons who can recover damages in tort. The fourth feature is related to the third. Lord Steyn observed (at 494):

"[T]he imposition of liability for pure psychiatric harm in a wide range of situations may result in a burden of liability on defendants which may be disproportionate to tortious conduct involving perhaps momentary lapses of concentration, eg in a motor car accident."

(See also Lord Hoffmann's discussion of the "control mechanisms" applicable in White (at 504 - 506) and the comments made by Mason P in Morgan v Tame (at 63,890) on the policy questions applicable.)

49 In Perre v Apand Pty Ltd [1999] HCA 36; (1999) 198 CLR 180 Gleeson CJ (at 192) observed that the consequences of a general rule that one person owes to another a duty to take care not to cause reasonably foreseeable financial harm "would be intolerable". There are, however, "circumstances in which the law recognises a duty of care such as will permit recovery of pure economic loss". The learned Chief Justice accepted the observation by Lord Oliver in Caparo Industries plc v Dickman [1990] UKHL 2; [1990] 2 AC 605 at 633 that a duty to avoid any reasonably foreseeable financial harm needs to be constrained by "some intelligible limits to keep the law of negligence within the bounds of common sense and practicality". In my opinion, the same sentiments apply to claims for psychiatric harm. Mental suffering of some kind is a common phenomenon in human intercourse and, once the law recognises the existence of a tortious duty for foreseeable mental harm caused by negligence, it is of fundamental importance that there be reasonable and defined limits confining liability thereunder.

50 The policy issues that arise involve a conflict between corrective justice and distributive justice: see the article by Mason P, "Fault, Causation and Liability" (2000) 19 Aust Bar Rev 201 (particularly at 203 - 205). Any dilution of what has been called the "control mechanisms" applicable to claims for mental harm could, at the stroke of the judicial pen, massively extend tortious liability, create a vast new category of litigants, and impose crushing burdens on negligent plaintiffs. Lord Steyn in White v Chief Constable of South Yorkshire Police mentions (at 495 - 496) the steady growth "in the more common place psychiatric injury proceedings based on the death, injury or imperilment of loved ones or fear of one's own safety". He observes that, nowadays, "it would be quite unrealistic to describe awards for psychiatric damage as modest". The potential number of victims of mental trauma brought about by negligence is infinite.

51 The warnings expressed by Rich J in Chester v Waverley Municipality [1939] HCA 25; (1939) 62 CLR 1 at 10 - 11, concerning the neglect of social consequences that may accompany unrestrained development of the law of tort, remain pertinent. Any alteration in legal principles that significantly enlarges potential liability for mental trauma will have profound changes for society. These would not only result in economic consequences involving, amongst other things, transfer of wealth on grounds not previously accepted. There would also be a major shift in the balance between plaintiffs and defendants and a marked impact upon the general administration of justice, involving litigants, lawyers, and courts. Health care and the practice of medicine would be affected. These are matters that fall appropriately within the province of legislative change. Indeed, some States (but not Western Australia) have enacted statutes that affect claims arising out of mental trauma. I do not think that it is for an intermediate appellate court to make changes of this kind.

The appropriateness of the preliminary issue

52 The question asked by the preliminary issue is whether the respondent owes a duty of care to the appellants "to exercise reasonable care and skill to avoid causing them psychiatric injury". This question assumes that the tort of negligence consists of three elements which can readily be separated, namely, a duty of care, a breach of that duty, and damage resulting to the plaintiff from that breach (which is not too remote in law). The arguments advanced by counsel assumed that the Court, in determining the preliminary issue, would be able to ignore the latter two elements entirely. These assumptions, however, are erroneous.

53 In Overseas Tank Ship (UK) Ltd v Mort's Dock & Engineering Co Ltd ("The Wagon Mound" [No 1]) [1961] UKPC 1; [1961] AC 388, Lord Simonds said (at 425):

"It is, no doubt proper when considering tortious liability for negligence to analyse its elements and to say that the plaintiff must prove a duty owed to him by the defendant, a breach of that duty by the defendant and consequent damage. But there can be no liability until the damage has been done. It is not the act but the consequences on which tortious liability is founded. ... It may, of course, become relevant to know what duty be owed to A, but the only liability that is in question is the liability for damage by fire. It is vain to isolate the liability from its context and to say that B is or is not liable, and then to ask for what damage he is liable. For his liability is in respect of that damage and no other. If, as admittedly it is, B's liability, (culpability) depends on the reasonable foreseeability of the consequent damage, how is that to be determined by the foreseeability of the damage which in fact happened - the damage in suit? And, if that damage is unforeseeable so as to displace liability at large, how can the liability be restored so as to make compensation payable?"

54 Similar views were expressed by Brennan J in Sutherland Shire Council v Heyman (at 487):

"I venture to repeat what I said in John Pfeiffer Pty Ltd v Canny [(1981) [1981] HCA 52; 148 CLR 218] at pp 241 - 242:

'His duty of care is a thing written on the wind unless damage is caused by the breach of that duty; there is no actionable negligence unless duty, breach and consequential damage coincide ... For the purposes of determining liability in a given case, each element can be defined only in terms of the others.' "

"The actual nature of the damage suffered is relevant to the existence and extent of any duty to avoid or prevent it."

(See also Caparo Industries plc v Dickman at 627 per Lord Bridge.)

"Foreseeability is relevant to the three different questions that may arise in an action for negligence - whether there was a duty of care; if so, whether the defendant was negligent; and whether the defendant was liable for the kind of damage that resulted from the negligence - and this sometimes tends to lead to confusion."

In Bourhill v Young [1942] UKHL 5; [1943] AC 92 Lord Russell noted that reasonable foreseeability is relevant not only to the existence of a duty of care but also to the question of remoteness of damage. In Morgan v Tame Mason P (at 63,885) pointed out that the issue of reasonable foreseeability, as an essential element of a duty of care and as a relevant factor in the inquiry into whether damages are too remote, "tend to merge": see also Rowe v McCartney [1976] 2 NSWLR 72 at 88 and Attia v British Gas plc [1987] EWCA Civ 8; [1988] 1 QB 304 at 319.

56 Remoteness of damage is also often relevant to questions of proximity. In Jaensch v Coffey Deane J (at 584 - 585) observed that causal proximity involved the "sense of the closeness or directness of the relationship between the particular act or cause of action and the injury sustained". In other words, causal proximity incorporates issues of remoteness of damage. That is not surprising, as "remoteness" is the antithesis of "proximity". Accordingly, questions of remoteness of damage are often caught up in the assessment of the existence of a duty of care.

57 Moreover, as Lord Oliver observed in Alcock v Chief Constable of South Yorkshire Police [1991] UKHL 5; [1992] 1 AC 310 at 412, the elements of reasonable foreseeability and proximity can coalesce. Additionally, in claims for psychiatric harm, the requirement that the harm be caused by shock, that is, a sudden assault on the senses (to which I refer in greater detail below), can have a bearing upon proximity, as well as causation and remoteness of damage.

58 In Dorset Yacht Co Ltd v Home Office [1970] UKHL 2; [1970] AC 1004 Lord Pearson (at 1052) said:

"The form of the order [directing argument of a preliminary question of law] assumes the familiar analysis of the tort of negligence into its three component elements, viz, the duty of care, the breach of that duty and the resulting damage. The analysis is logically correct and often convenient for the purposes of exposition, but it is only an analysis and should not eliminate consideration of the tort of negligence as a whole. It may be artificial and unhelpful to consider the question of the existence of a duty of care in isolation from the elements of breach of duty and damage."

59 In Caparo Industries plc v Dickman Lord Oliver said (at 633):

"Indeed it is difficult to resist the conclusion that what have been treated as three separate requirements are, at least in most cases, in fact merely facets of the same thing, for in some cases the degree of foreseeability is such that it is from that alone that the requisite proximity can be deduced, whilst in others the absence of that essential relationship can most rationally be attributed simply to the court's view that it would not be fair and reasonable to hold the defendant responsible."

Lord Oliver repeated these views in Murphy v Brentwood District Council [1990] 3 WLR 414 at 445 and a similar opinion was expressed by Lord Keith (at 425).

60 In Marc Rich & Co AG v Bishop Rock Marine Co Ltd [1995] UKHL 4; [1996] 1 AC 211 at 235 Lord Steyn approved the observation by Saville LJ that foreseeability, proximity and considerations of what is fair, just and reasonable "overlap with each other and are facets of the same thing" and, further, that the "so-called requirements of a duty of care are not to be treated as wholly separate and distinct requirements but rather as convenient and helpful approaches to the pragmatic question whether a duty should be imposed in any given case".

61 These authorities make it plain that in certain categories of cases of negligence, there are dangers in trying to make separate inquiries into duty of care, breach of duty and remoteness of damage (including causation). These dangers are inherent in the way the preliminary issue in this case is formulated.

62 In some instances, for example, in most motor vehicle accident cases, it is helpful and convenient to separate the three elements and to make separate inquiries into each. But in more complex situations, and particularly cases where damages are claimed for negligence causing psychiatric injury, there should be only one basic inquiry that incorporates an investigation into all the elements of negligence. In such a case those elements are dependent on each other and can only be defined by reference to each other.

63 I would adopt the approach of Samuels JA in Nader v Urban Transit Authority (NSW) (1985) 2 NSWLR 501, a case of psychiatric injury to a primary victim. His Honour said (at 506):

"I am inclined to agree with Professor Fleming "The Law of Torts" 6th ed 1983 at 191, that there is much to recommend the approach suggested by Denning LJ, as he then was, in Roe v Minister of Health [1954] EWCA Civ 7; [1954] 2 QB 66 at 85 which would propound one inquiry: 'Is the consequence fairly to be regarded as within the risk created by the negligence?' I think that cases of the present kind would be more fairly and readily determined by the application of such a principle."

64 Similar views were expressed in the Canadian case of Beecham v Hughes (1988) 52 DLR (4th) 625 by Lambert JA, a member of the Court of Appeal of British Columbia, who said (at 665 - 666):

"The questions of foreseeability, proximity, causation and remoteness are interlocked. There are not four answers to four questions, but one composite answer to one composite question ...

I do not consider that there is any one key which opens the door to a simple and straightforward answer ... There is no single question which solves the problem. There is no single limiting factor other than the composite answer, based on foreseeability, proximity, causation, and remoteness, to the composite question set out [above]."

65 I have expressed my views in some detail on these matters for two principal reasons. Firstly, despite the attempt in the formulation of the preliminary issue to quarantine the different elements of the tort from each other, I do not think it desirable or possible to do so. This was recognised some 30 years ago by Lord Pearson in Dorset Yacht Co Ltd v Home Office but the warning given therein seems to have gone unheeded. Secondly, further attempts at separation of the elements in this kind of case should be discouraged. Apart from the sterile arguments that tend to arise as to whether a particular facet of the claim falls under one element or another, there are fundamental difficulties in determining issues relating to overall liability on a piecemeal basis.

66 I would conclude on this aspect by repeating Lord Browne-Wilkinson's remarks in Barrett v Enfield London Borough Council [1999] 3 WLR 79 at 83 that:

"[I]n an area of the law which [is] uncertain and developing ... it is of great importance that such development be on the basis of the facts found at the trial not on hypothetical facts assumed (possibly wrongly) to be true ..."

His Lordship was referring to strike-out applications, but these sentiments apply equally to preliminary issues such as the present.

Reasonable foreseeability of psychiatric harm

67 There are two elements of the appellants' cause of action on which both counsel focussed to the exclusion of any others. One was foreseeability and the other proximity. I shall deal firstly with foreseeability.

68 There are two particular aspects of foreseeability that have to be considered. The one is the application of the "normal fortitude" test, to which I have already referred, and the other is the requirement that psychiatric harm be caused by a "sudden shock".

69 I shall first deal with the normal fortitude test. In Morgan v Tame (at 63,873) Spigelman CJ said:

"In Jaensch v Coffey Gibbs CJ at 556 expressly accepted a test of 'ordinary person of normal fortitude'. Brennan J provided a more extensive analysis of the test, to which he referred in terms of 'normal standards of susceptibility' at 568, 570 - 571, 572 and 578. I interpret Murphy J's reference at 557 to a 'normal person' to be supportive of the test. Deane J at 610 refers to the test with implicit approval.

The Full Court of the Federal Court and the Court of Appeal of the Supreme Court of Queensland applied this line of authority in, respectively, Wodrow v Commonwealth of Australia (1993) 45 FCR 52 at 72 - 73 and Mid-West Radio Ltd v Arnold (1999) EOC 92-970 at [28] to [29]. Although Windeyer J doubted the proposition and treated the issue as open in Mount Isa Mines Ltd v Pusey at 405 - 406, his Honour's obiter comments have not been adopted. (See also at 417 per Walsh J). ...

This Court should follow the majority reasoning of the High Court in Bunyan v Jordan [(1936-1937) [1937] HCA 5; 57 CLR 1 at 14, 16 - 17 and 18] and Jaensch v Coffey."

Mason P (at 63,887) observed that the "normal standard of susceptibility" test remained "part of the touchstone for liability for pure psychiatric illness". His Honour relied on the same authorities as Spigelman CJ. Handley JA agreed with Spigelman CJ and Mason P. Their Honours dealt with this test as part of the question whether the psychiatric injury suffered was reasonably foreseeable. I would follow the same approach.

70 In Morgan v Tame Spigelman CJ (at 63,876) said that the most frequently quoted authority in support of the proposition that liability for psychiatric injury caused by negligence is restricted to cases where the injury is induced by a sudden shock to the senses is that of Brennan J in Jaensch v Coffey (at 565 and 567). In that case, Brennan J said (at 565):

"A plaintiff may recover only if the psychiatric illness is the result of physical injury negligently inflicted on him by the defendant or if it is induced by 'shock'. Psychiatric illness caused in other ways attracts no damages, though it is reasonably foreseeable that psychiatric illness might be a consequence of the defendant's carelessness."

"I understand 'shock' in this context to mean a sudden sensory perception - that is, by seeing, hearing or touching - of a person, thing or event, which is so distressing that the perception of the phenomenon affronts or insults the plaintiff's mind and causes a recognizable psychiatric illness. A psychiatric illness induced by mere knowledge of a distressing fact is not compensable; perception by the plaintiff of a distressing phenomenon is essential. If mere knowledge of a distressing phenomenon sufficed, the bearers of sad tidings, able to foresee the depressing effect of what they have to impart, might be held liable as tortfeasors."

71 In Morgan v Tame, Spigelman CJ referred (at 63,877) to other authorities that required a sudden assault on the senses as being an essential element of the cause of action: These include Campbelltown City Council v Mackay (1989) 15 NSWLR 501 per McHugh JA, Chiaverini v Hockey [1993] A Tort Rep 62,254 at 62,255 per Handley JA,and at 62,257 per Sheller JA (with whom Meagher JA agreed), Spence v Percy [1992] 2 Qd R 299 at 313 - 314, 317 and 320 - 321 (where Shepherdson, Williams and de Jersey JJ accepted that there was a need for the plaintiff to prove that the psychiatric injury was caused by a "sudden sensory perception"), Pham v Lawson [1997] SASC 6086; (1997) 68 SASR 124 at 149 and 150 (where the South Australian Full Court accepted that it was necessary for the plaintiff to prove that the psychiatric injury was caused by a "sudden event"), and Davis v Scott [1998] SASC 7201; (1998) 71 SASR 361 at 379 - 381. I would also refer to Mount Isa Mines Ltd v Pusey where Windeyer J (at 407) said:

"I do not question decisions that nervous shock resulting simply from hearing distressing news does not sound in damages in the same way as does nervous shock from witnessing distressing events. If the sole cause of shock be what is told or read of some happening then I think it is correctly said that, unless there be an intention to cause a nervous shock, no action lies against either the bearer of the bad tidings or the person who caused the event of which they tell. There is no duty in law to break bad news gently or to do nothing which creates bad news."

"Although, in my opinion, there is no authoritative statement of the High Court which formally binds this Court to apply the requirement of a sudden assault on the senses, this Court should follow its own earlier decisions and that of other intermediate courts of appeal."

72 Mason P undertook a similar exercise. He referred (at 63,889) to the remarks of Brennan J in Jaensch v Coffey and pointed out that:

"Deane J (with whose reasons Gibbs CJ was in general agreement) also appeared to recognise the need in a nervous shock action for the psycho-neurosis or mental illness to be the result of a sudden shock suffered by the plaintiff (see at 602, 606)."

"In Campbelltown City Council v Mackay (1989) 15 NSWLR 501 McHugh JA (with whose reasons Samuels JA agreed) held (at 507 - 8) that the matter was foreclosed by the decision in Jaensch. Based on this view, he said in effect that a plaintiff in an action for pure psychiatric illness, as distinct from one when the illness stemmed from physical injury, damage to property or defamation, had to establish that the psychiatric illness is the result of a shock caused by the sudden perception of some phenomenon from which the defendant is responsible (see at 511). Kirby P reserved his position as to the correctness of the parties' concession to this effect (see at 503 - 4. See also Coates v Government Insurance Office [(1995) 36 NSWLR 1] at 8 - 11).

I accept that the views of McHugh JA and Samuels JA in Campbelltown [City Council] both represent the common law as declared by this Court and correctly interpret the reasoning of the majority of the High Court in Jaensch. See also Chiaverini v Hocking (1993) [A Tort Rep] 81-223; Buljabacic v Ah Lam, Court of Appeal; Supreme Court of New South Wales, unreported 3 September 1997."

73 Handley JA (at 63,891) agreed with both Spigelman CJ and Mason P and held that the action failed because the plaintiff did not suffer:

"a sudden affront or assault on her psyche from the perception of a horrifying event which was a necessary pre-condition for liability for psychiatric damage in a case such as this, as explained by Brennan J in Jaensch v Coffey (1994) 155 CLR 349, 565, 567; and subsequent decisions of this Court".

The law as stated in Morgan v Tame is consistent with the principles laid down by the House of Lords in White v Chief Constable of South Yorkshire Police and Alcock v Chief Constable of South Yorkshire Police. It is also in accord with the Canadian authorities of Rhodes v Canadian National Railway (1990) 75 DLR (4th) 248 and Devji v Burnaby District. In my opinion, this Court should follow the law so stated.

74 In Jaensch v Coffey, the element of harm being caused by sudden shock was explained by Brennan J as being part of the inquiry into foreseeability. His Honour observed (at 563):

"[It] is not necessary for a plaintiff to prove that a reasonable person in the defendant's position could foresee that any particular psychiatric illness might be caused by his conduct; it suffices that he could have foreseen that his conduct might cause some recognised psychiatric illness induced by shock ... It is not necessary that the precise events leading to the administration of the shock should be foreseeable. It is sufficient that shock and a psychiatric illness induced by it are reasonably foreseeable."

"The notion of psychiatric illness induced by shock is a compound, not a simple, idea. Its elements are, on the one hand, psychiatric illness and, on the other, shock which causes it. Liability in negligence for nervous shock depends upon the reasonable foreseeability of both elements and of the causal relationship between them." (See also at 572.)

75 While counsel for the appellants addressed the need for the appellants to have been injured by a "shock", counsel for the respondent did not. He submitted that the question of sudden shock had nothing to do with the existence of a duty of care. He argued that the question whether psychiatric injury stemmed from a sudden isolated assault on the sensory systems went to "the distinct issues of the nature of actionable damage, to causation, and to the aetiology of disorder". He said: "Those are matters entirely distinct from the duty of care." I do not accept these submissions. The passages from the judgment of Brennan J in Jaensch v Coffey to which I have referred establish that the question whether a plaintiff suffers psychiatric injury through shock forms part of the inquiry whether the psychiatric injury was reasonably foreseeable. Of course, the requirement of the injury being induced by shock may be part of the issues identified by counsel for the respondent, but that is not presently relevant.

76 In Morgan v Tame, Mason P said (at 63,886):

"Foreseeability involves the judge looking with hindsight at an event which has occurred and considering the perception of a hypothetical person placed generally in the position of the tortfeasor."

77 Accordingly, when the Court looks with hindsight at the events which occurred (in order to determine foreseeability), it must decide whether the psychiatric harm sustained by the appellants was in fact caused by shock.

The foreseeability of the psychiatric injury sustained by the appellants

78 Heenan J held that it was reasonably foreseeable that the appellants would suffer psychiatric injury if harm befell James as a consequence of the negligence of the respondent. His Honour relied on the following remarks of Lord Oliver in Alcock v Chief Constable of South Yorkshire Police (at 411).

"The traumatic effect on ... a mother on the death of her child is as readily foreseeable in a case where the circumstances are described to her by an eye witness at the inquest as it is in a case where she learns of it at a hospital immediately after the event. ...[Moreover,] the news brought by a policeman hours after the event may be as sudden and unexpected to the recipient as the occurrence of the event is to the spectator present at the scene."

This is consistent with the statement of Lander J (with whom Bollen J agreed) in Pham v Lawson (at 148):

"There is no doubt that it is foreseeable that if a tortfeasor injures or kills a child that child's mother might suffer injury in receiving the information of the injury or death."

79 It is to be observed, however, that Lander J made no reference to the normal fortitude test and nor did Lord Oliver in Alcock v Chief Constable of South Yorkshire Police. Further, as regards the observations of Lord Oliver to which Heenan J referred, his Lordship, I think, was not referring to the case where a mother is simply told of the fact of the death of her child. He was, in my view, referring to circumstances where the mother is given information of a frightening or horrifying kind, and the fright or horror she experiences, in so being told of the death of her child, causes psychiatric harm. This is apparent from the passage cited, where Lord Oliver (at 411) referred to the traumatic effect on a mother "in a case where the circumstances are described to her by an eye witness", in the context of his Lordship accepting that liability depended upon "shock" (that is, "a direct assault upon the mind or nervous system" (at 406)) constituted by a "sudden and direct visual impression on the plaintiff's mind of actually witnessing the event or its immediate aftermath" (at 416).

80 I now turn to the two scenarios (to which I have referred above) that arise for consideration from the facts to be assumed. In doing so, I bear in mind that, as Lord Oliver said in Alcock v Chief Constable of South Yorkshire Police (at 416):

"Grief, sorrow, deprivation and the necessity for caring for loved ones who have suffered injury or misfortune must, I think, be considered as ordinary and inevitable incidents of life which, regardless of individual susceptibilities, must be sustained without compensation."

81 The first scenario is that the appellants sustained psychiatric injury by the information they received during the telephone call on 6 December 1986 when Mr Annetts was informed that James was missing from his place of employment and was believed to have run away. As mentioned upon receipt of that advice Mr Annetts collapsed and Mrs Annetts continued the telephone conversation.

82 At that stage, neither of the appellants was told when James had run away and for how long he had been missing. That being so, applying the normal fortitude test, while it is quite possible for a reasonable person in the respondent's position to foresee that both the appellants would suffer deep anxiety and grief from the information conveyed, I do not think that it was reasonably foreseeable that they might suffer psychiatric injury.

83 The essential question, however, is whether (to paraphrase Brennan J in Jaensch v Coffey at 567) the respondent should have foreseen that the breach of duty on its part might result in a sudden sensory perception on the part of the appellants of a phenomenon so distressing that a recognisable psychiatric illness would be caused thereby. As mentioned, this determination has to take place in hindsight, and it is necessary, therefore, to determine whether in fact such an event occurred. I do not accept that the telephone call of 6 December 1986 constituted such a phenomenon. Rather, it gave rise only to "knowledge of a distressing fact", which is not compensable. In my opinion, on 6 December 1986 there was no perception by the appellants of a distressing phenomenon in the sense that this term was used by Brennan J. Accordingly, in my view, the appellants have failed to establish that the requisite element of foreseeability existed in regard to the first scenario.

84 The second scenario involves assuming that the appellants sustained psychiatric injury upon learning of James' death, the injury being caused by the receipt of that advice coupled with what had transpired earlier.

85 I have difficulty in accepting that it is reasonably foreseeable that a parent of normal fortitude might sustain psychiatric injury upon being informed of the death of a 16-year-old child. Undoubtedly, such a parent is likely to suffer deep grief and sadness, but as Lord Oliver noted in Alcock v Chief Constable of South Yorkshire Police (at 416) deprivation of a loved one through misfortune is an ordinary and inevitable incident of life. Psychiatric injury, however, is a different class of harm, and I am not persuaded that, according to the test applicable, it was reasonably foreseeable that the appellants might suffer a medically recognised mental illness upon the death of James (cf Chester v Waverley Municipality (at 10, 11, and 13-14)).

86 I recognise, however, that opinions may differ on this issue and I shall assume that foreseeability of psychiatric injury, in general terms, has been established. Nevertheless, the essential question for the purposes of determining whether the respondent owed the appellants a duty of care is different and more specific, as it must incorporate the element of harm caused by a sudden shock to the senses. The question in these terms, as I have indicated, is whether (tested in hindsight by reference to the relevant facts) the respondent should have foreseen that its breach of duty might result in a sudden sensory perception on the part of the appellants of an event so distressing that they would suffer a recognisable psychiatric illness.

87 After 6 December 1986, the appellants must gradually have come to realise that, in the harsh, desert-like area where James was missing, the probabilities were that he would have died. As the weeks and months went by, those probabilities must have strengthened. In the circumstances, the final confirmation of James' death that the appellants received at the end of April 1987, when Mr Annetts identified the remains from the photograph, cannot be regarded as a sudden sensory perception of a distressing phenomenon.

88 Accordingly, in regard to the second scenario (applying the test of foreseeability in hindsight), this element is not established.

89 In White v Chief Constable of South Yorkshire Police, Lord Goff (at 472) accepted that it was settled law that in a claim for psychiatric injury the necessary proximity of the relationship between the plaintiff and the defendant must be proved. To do this, the plaintiff must establish:

"(1) a close tie of love and affection to the immediate victim;

(2) closeness in time and space to the incident or its aftermath; and

(3) perception by sight or hearing, or its equivalent, of the event or its aftermath."

Lord Steyn expressed the same views (at 496), as did Lord Hoffmann (at 502).

90 In Australia, the concept of proximity has long been recognised as an important factor in determining liability for the tort of negligence. It has developed a particular meaning in regard to claims for psychiatric injury and, subject to some differences of opinion as to the content and extent of the second and third requirements referred to by Lord Goff, that meaning is along similar lines to that stated by his Lordship.

91 In Perre v Apand Pty Ltd, views were expressed by some of the members of the High Court that might be thought to diminish the significance of proximity. But, in my view, nothing was said in this respect that bears directly on the requirement of proximity in claims for psychiatric injury.

92 Gleeson CJ (at 194) pointed out, with apparent approval, that in Caparo Industries plc v Dickman:

" Lord Oliver emphasised (at 643) that, in this field of discourse [that is, claims for mere economic loss], the mere foreseeability of possible damage, without some further control, (which he summarised as 'proximity', after explaining what he meant by that term), would not be useful as the test of liability."

"It may well be that, at this stage, the notion of proximity can serve no purpose beyond signifying that it is necessary to identify a factor or factors of special significance in addition to the foreseeability of harm before the law will impose liability for the negligent infliction of economic loss."

McHugh J, while saying (at 209 - 210) that, "this Court no longer sees proximity as the unifying criterion of duties of care", nevertheless stated that neither the High Court nor the English Courts have entirely abandoned the use of proximity as a factor in determining duty. His Honour said (at 217):

"[I]n an area of law such as awarding damages for negligently inflicted economic loss, which is still developing and which has been recently cast adrift from any unifying principle, there is no alternative to a cautious development of the law on a case by case basis. ... Until a unifying principle again emerges, however, the best solution is to proceed incrementally from the established cases and principles."

Kirby J (at 275) said that in deciding whether a legal duty of care in negligence exists, the decision-maker must ask three questions, the second of which being:

"Does there exist between the alleged wrongdoer and such person a relationship characterised by the law as one of 'proximity' or 'neighbourhood'?"

Hayne J (at 301) said, "references to proximity" should not be discarded. Callinan J (at 326) made it plain that he was of the view that proximity remained a relevant factor.

93 In the circumstances, I consider that the remarks of Deane J in Jaensch v Coffey (to which reference is most often made in connection with proximity in claims for psychiatric injury) remain of fundamental importance. His Honour (at 584 - 585) referred to the concept of proximity as discussed by Lord Atkin in Donoghue v Stevenson [1932] AC 562 and observed that:

"[Proximity] involves the notion of nearness of closeness and embraces physical proximity (in the sense of space and time) between the person or property of the plaintiff and the person or property of the defendant, circumstantial proximity such as an overriding relationship of employer and employee or of a professional man and his client and causal proximity in the sense of the closeness or directness of the relationship between the particular act or cause of action and the injury sustained."

"[T]he requisite duty relationship will not, on the present state of the law, exist in a case where mere psychiatric injury results from subsequent contact, away from the scene of the accident and its aftermath, with a person suffering from the effects of the accident."

Deane J explained that the element of proximity was a requisite of liability and stressed that there has to be "causal proximity" between the act of carelessness and the resulting injury. Accordingly, where the psychiatric injury results from contact with more remote consequences - such as the subsequent effect of the accident upon an injured person - the necessary element of proximity will not be established (see, particularly, at 606 -608). Thus, as the respondent in Jaensch v Coffey sustained psychiatric injury as a result of the combined effect of what she saw at the hospital where her husband was in intensive care (that being part of "the aftermath" of the accident) and what she was told, the element of proximity was present. His Honour regarded the aftermath as including the journey by ambulance to hospital and the scene at the hospital itself.

94 Gibbs CJ (at 551) agreed generally with Deane J's reasons and Dawson J (at 612) accepted that for liability to be established it was not necessary for the plaintiff to observe the consequences of the defendant's breach of duty at the scene of the accident, as they "may be observed as part of the aftermath and the aftermath may extend to the journey by ambulance to a hospital and to the scene at the hospital itself".

95 In Spence v Percy, the Queensland Full Court followed the principles expressed by Deane J in Jaensch v Coffey. Shepherdson J adopted the causal proximity test. Williams J said (at 317 - 318):

"Gibbs CJ, Deane J and Dawson J each appeared to be of the view that one relevant question was whether the events which caused the nervous shock were part of the aftermath of the accident. No attempt was made to define the term, and it could hardly be considered as a term of art. It is true, as the learned trial judge here observed, that dictionary meanings indicate that it has no temporal connotation other than one of consequence. But that overlooks, in my respectful view, the context in which the term was used by the High Court; it was used to limit the extent of the tortfeasor's liability. In consequence it seems to me that the very use of the term 'aftermath' indicates that the proximity test will not be satisfied where there is a significant time lapse between the tortious act and the onset of the psychiatric illness."

"In Jaensch v Coffey Deane J referred (at 607) to a need to establish 'logical or causal proximity between the act of carelessness and the resulting injury', rather than mere 'physical proximity, in the sense of space and time, between the accident and its immediate aftermath on the one hand, and the injury on the other' (at 606). He said that that criterion for liability was better adapted to reflect 'notions of fairness and commonsense' than a requirement based 'merely upon mechanical considerations of geographical or temporal proximity' (at 607). Gibbs CJ may be taken to have agreed (at 551). Brennan J spoke of remoteness in proximity with relation to the question of reasonable foreseeability (at 570 - 571), and Dawson J expressed the limitation upon liability in terms of policy (at 612 - 613). "

"The features which, in combination, persuade me that the requisite degree of proximity is absent in this case, are that very substantial time difference; that many events intervened; that those intervening circumstances included grief and sorrow which, while obviously bearing upon the ultimate injury, were nevertheless not themselves compensable; that the ultimate injury was produced by an immediate cause (the shock) separate and distinct from (although causally related to) the original act of negligence; and that the death which led to the shock which produced the injury was outside the temporal and geographical aftermath of the accident, using the word 'aftermath' in its ordinary sense of the further events arising from the accident."

"It is clear, as Deane J said (at 606), that liability will not attach 'in a case where mere psychiatric injury results from subsequent contact, away from the scene of the accident and its aftermath, with a person suffering from the effects of the accident'."

99 In FAI General Insurance Co Ltd v Curtin (1997) A Tort Rep 64,479, the Queensland Court of Appeal again accepted that proximity was a necessary ingredient of a duty of care to avoid psychiatric injury caused by negligence. In Chiaverini v Hockey, the New South Wales Court of Appeal (Meagher, Handley and Sheller JJA) applied the test of proximity as explained by Deane J in Jaensch v Coffey. In Buljabasic v Ah Lam, unreported; C of A; SCt of NSW; No 40417196; 3 September 1997, the New South Wales Court of Appeal once more applied the principles expressed by Deane J in regard to the "aftermath"; thus where the injury was caused by matters that occurred a day or two after the accident, the plaintiff's claim failed. In Pham v Lawson, Lander J (with whom Cox J, on this issue, and Bollen J agreed), after referring to the various judgments in Jaensch v Coffey, said (at 144):

"[T]hose judgments establish that a duty of care will be owed by the tortfeasor to the spouse of an injured person where that spouse has suffered nervous shock and consequent psychiatric illness in circumstances where the spouse was not present at the time of the accident and did not attend the scene of the accident but was later told of the consequences of the accident in relation to her spouse and attended at the hospital and perceived for herself some of the consequences. ...

The existence of the duty of care becomes less likely as all of the matters which are important for existence become more remote. So that if the relative was not at the scene or does not attend at the scene then there is less likelihood of the determination of a duty of care as that person has less direct involvement in space and time and therefore less direct perception of the injuries suffered by the person for whom that relative cares.

It is a matter of degree. It is a matter of common sense when the stage is reached that a court must say that there can be no duty of care in a given case because the involvement of the person who suffered the nervous shock is not sufficiently close in terms of relationship, involvement or perception."

"It seems to me therefore that the respondent was only entitled to succeed before the learned trial judge if what she was told by the police officer, what she observed on the way to the hospital and what she observed at the hospital brought her into a class of persons to whom a duty of care was owed and were factors sufficient of themselves to cause her a nervous shock and in fact did so."

101 Finally, on this review of the authorities, I should note the position in three other Commonwealth countries. In Van Soest v Residual Health Management Unit [1999] NZCA 206; (2000) 1 NZLR 179, Blanchard J who delivered the judgment of the majority (Gault, Henry, Keith and Blanchard JJ) of the New Zealand Court of Appeal said (at 187):

"High authority establishes that it is not enough for a plaintiff to show that his or her mental suffering as a secondary victim is a reasonably foreseeability consequence of the defendant's negligence towards the primary victim. Three further matters need to be considered:

(a) The nature of the secondary victim's mental suffering;

(b) The physical proximity of the secondary victim to the primary victim's accident or misadventure; and

(c) The relational proximity (the closeness of the relationship) of the primary and secondary victims."

Thomas J dissented, and applied, in effect, a test of foreseeability alone.

102 A similar test to that suggested by Thomas J was applied by the Court of Appeal of South Africa in Barnard v Santam Bpk [1998] ZASCA 84; (1999) 1 SA 202, but this authority is of limited value as the law of negligence in South Africa is based on the Roman lex Aquilia, pursuant to which delictual liability for negligence is based substantially on the test of foreseeability, limited only by causation and matters of public policy (and the concept of proximity plays no role).

103 In Canada, the judgment of Deane J in Jaensch v Coffey was approved by the British Columbia Court of Appeal in Beecham v Hughes and, in deciding Devji v Burnaby District and Rhodes v Canadian National Railway, that Court relied on proximity, as understood by the House of Lords in White v Chief Constable of South Yorkshire Police and Alcock v Chief Constable of South Yorkshire Police.

104 On the basis of the material to which I have referred, the overwhelming weight of authority is to the effect that the appellants are required to show that they were in a relationship of proximity to the respondent as explained by Deane J in Jaensch v Coffey. This approach satisfies the policy considerations that I earlier identified, allows the law in this area to proceed casuistically, and enables the legal position to be established without undue uncertainty. I propose to follow it.

105 There is an aspect of proximity, however, that requires additional elaboration, namely, the question whether a plaintiff has directly to perceive the phenomenon or its aftermath, or whether it is sufficient if the plaintiff is informed of the consequences of the phenomenon.

106 Direct perception of the phenomenon or its aftermath is a settled requirement according to the law of England: White v Chief Constable of South Yorkshire Police, Alcock v Chief Constable of South Yorkshire Police. It is also a requirement upheld in Canada at the intermediate appellate level (Devji v Burnaby District). It is not a requirement in South Africa: Barnard v Santam Bpk.

107 The issue has not yet finally been resolved in Australia. I have referred to the remarks of Windeyer J in Mount Isa Mines Ltd v Pusey (at 407) and Brennan J in Jaensch v Coffey (at 566-567) to the effect that a psychiatric illness induced by mere knowledge of a distressing fact is not compensable and perception by the plaintiff of a distressing phenomenon is essential. In Jaensch v Coffey the need for direct perception of a distressing phenomenon was expressly left open by Gibbs CJ (at 555), Murphy J (at 556), and Deane J (at 608 - 609). It was also left open in Coates v Government Insurance Office (NSW) (1995) 36 NSWLR 1 by Gleeson CJ (at 5) and Clarke JA (at 23). Kirby P accepted that proximity was a necessary element of a cause of action for damage for psychiatric injury, but was not prepared to restrict recovery to perceptions derived from acts of physical proximity to the happening of the tort or to its immediate aftermath. There has been much academic support for his Honour's view, and in Pham v Lawson, Lander J, after making the remarks to which I have referred above, made obiter comments (at 148) agreeing with what Kirby P had said in Coates. Bollen J agreed with Lander J (at 125) and Cox J did not decide the point (at 125). In Morgan v Tame, Spigelman CJ (at 63,876) regarded the point as being open.

108 In Petrie v Dowling [1992] 1 Qd R 284, Kneipp J upheld a claim for damages for psychiatric injury based solely on what the plaintiff had been told. Counsel for the defendant did not base any argument on the fact that the plaintiff's shock was caused solely by something said to her and not by things seen or a combination of things seen and heard. In the circumstances I do not regard this case as persuasive authority.

109 In Knight v Pedersen & Ors [1999] NSWCA 333 at [24], the Court of Appeal referred to the observations by Kirby P in Coates v Government Insurance Office (NSW) (at 7 and 8) and said:

"[T]he present state of the common law in New South Wales supports the view that liability for nervous shock suffered by the child of a parent killed by neglect or default is not limited to cases where the parent was killed in the sight or hearing of the child."

I accept, however, the submission of counsel for the respondent that this statement refers only to the rule that liability will arise when the plaintiff suffers shock during the aftermath of the accident.

110 In my opinion, the views of Windeyer J Mount Isa Mines Ltd v Pusey and Brennan J in Jaensch v Coffey and those of the House of Lords in White and Alcock should be followed. Despite the criticisms of its restrictive consequences, the direct perception requirement satisfies the policy issues that have long troubled Judges at the highest level, establishes criteria for an appropriately close relationship of proximity, and has the undoubted merit of relative certainty.

111 On the basis of the direct perception requirement, the appellants have not established the requisite degree of proximity under either of the scenarios I have postulated. Apart from the occasion (in January 1987) when Mr Annetts saw a blood covered hat belonging to James, and when he identified James' remains (from a photograph seen some five months after his death), they did not directly perceive the consequences of the respondent's breach of duty. I do not consider the two instances I have mentioned as satisfying the requirement.

The element of proximity if direct perception is not an essential requisite

112 Despite the views I have expressed, however, I shall assume, in favour of the appellants, that direct perception of the phenomenon is not an absolute requirement. In making this assumption I shall nevertheless have regard to the fact that the appellants' direct perception of the relevant events was limited to the extent that I have stated. This aspect of the matter remains significantly relevant to "causal proximity".

113 Causal proximity includes the means by which the psychiatric injuries were caused and, in this sense, is akin to remoteness of damage. Proximity, as a whole, however, must embrace the whole gamut of factors that bear upon the relationship between the parties. These include the physical proximity of the appellants to the injury inducing events, the time between those events and the onset of the psychiatric injuries. Issues of foreseeability and reliance may also concern proximity. For example, the respondent's knowledge that the appellants were relying upon it to supervise James, and that they were concerned parents, are relevant in this regard.

114 I turn now to the facts of the case. On the first scenario to which I have referred, the relevant factors are that the information was conveyed to the appellants by telephone, there was no direct perception of the relevant event, and, while it is not known when James ran away, he died some two days before the telephone call was made. I take into account (to the degree explained hereafter) the respondent's knowledge about the appellants' concerns and their reliance upon it, to which I have previously adverted. I also have regard to the respondent's admission that it knew, if it breached the duty of care it owed to James to provide him with a safe system of work and if, as a result, James died in the desert, that the appellants would suffer "particular distress".

115 The delay in learning that James was missing was not due to any inaction on the part of the appellants. On the admitted and assumed facts, the respondent knew that he was missing on 3 December 1986 and had reason to suspect on that date that James was in grave danger of injury or death. The respondent, however, only informed the appellants on 6 December 1986 that James was missing. I accept that there would probably have been no difference in the appellants' reaction had they been told earlier that James was missing. In these circumstances, I do not think it appropriate to hold that there was a remoteness in time between the moment when James ran away and the moment when the appellants learned that he had done so.

116 The telephone call of 6 December 1986 was a direct result of the respondent's failure to supervise and care for James in circumstances where it knew that the appellants were relying upon it to do so. This is an important consideration in the appellants' favour.

117 But the telephone call involved merely a statement by the police officer that James had run away and was missing. On the admitted and assumed facts, nothing was said as to when he had run away and for how long he had been missing. Nothing was said about his condition. No facts were given that made the situation particularly horrifying. I realise that some of these matters concern the requirement of sudden shock, but they also bear on causal proximity. I would add that nothing in the admitted and assumed facts indicates that there were other circumstances, bearing upon the information conveyed by the telephone call, that were capable of reinforcing the element of causal proximity. Importantly, the appellants were not present to perceive the actual circumstances under which James was missing. The appellants were far away from the relevant events. There was a complete absence of geographic proximity.

118 I think it desirable that I should comment on the "reliance" placed by the appellants on the respondent to "supervise" James, and the respondent's knowledge that the appellants would suffer "particular distress" should it fail properly to do so, and thereby cause James to die in the desert. I accept that these facts assist the appellants to some relatively minor degree in establishing the requisite proximity, but I do not regard them as being otherwise of particular significance. On the admitted and assumed facts, the appellants did not alter their position or take or refrain from taking steps because of their reliance on the respondent. There was no contractual or other relationship known to the law between the respondent and the appellants. The reliance so placed by the appellants on the respondent was not akin to that placed by parents, say, on a nurse whom they employ to care for their child, where a direct relationship arises. The reliance was similar to that placed by all caring parents on employers of their adolescent children. In my view, such reliance is a relevant but incidental aspect of the proximity inquiry.

119 It is a matter of judgment whether, in the circumstances I have described, proximity was established. In exercising that judgment it is relevant to observe that no case has been cited to us (and many cases indeed have been drawn to our attention) in which, in relatively analogous circumstances, a plaintiff has succeeded. There are cases in which the proximity requirement was held not to have been met even though the facts supported a contrary finding to a far greater degree than the facts applicable under the first scenario. I shall mention but two of these, which may be said to be typical of the approach that has been adopted.

120 In Palmer v Tees Health Authority [1999] EWCA Civ 1533; [1999] Lloyd's Rep Med 351, a child was abducted, sexually assaulted and murdered by a psychiatric patient who was being treated by the respondent. The appellant mother of the child learned that the child was missing on the day she was abducted, but did not see the child's body until six days later when she viewed the body at the mortuary. The Court of Appeal held that proximity had not been established. Stuart-Smith LJ said (at 356):

"[O]nce rules are established, it is not open to the courts to extend the accepted principles of proximity simply because the facts of a given case are particularly horrifying or heart-rending."

121 In Devji v Burnaby District, the deceased was the daughter of two of the plaintiffs and the sister of two other plaintiffs. She was killed in a motor vehicle accident that occurred in the late evening of 30 December 1991. At about 11.00 pm that evening, the plaintiffs were notified of her death at their home. They went immediately to the hospital where the deceased was and waited there while the body was prepared for viewing. At about 12.30 am they viewed and identified the body. The Court of Appeal held that proximity had not been established.

122 In regard to the first scenario, I consider that, even on the assumed basis on which I have proceeded, the appellants have not established the requisite degree of proximity.

123 I do not think that the answer is any different in regard to the second scenario. Indeed, in the circumstances postulated thereunder, the proximity relationship is even more remote, having regard to the relatively long period of five months between the death of James and the confirmation that he had died, and the gradual causative effect of all the events that intervened. These matters detract substantially from the appellants' arguments in regard to both causal proximity and proximity in time.

124 In this respect, in addition to the cases to which I have referred relating to the first scenario, I would note, as a further example of the Court's approach to issues of this kind, that in Alcock v Chief Constable of South Yorkshire Police the plaintiffs visited a mortuary some nine hours after the disaster for the purposes of identifying the victims. Although the circumstances involved tragic and horrifying events, the delay was held sufficient to "take them outside the immediate aftermath of this disaster" (per Lord Jauncey at 424).

125 Heenan J held that the involvement of the appellants in and their perception of the events, which led to James' death, were remote and not sufficiently close to give rise to a duty of care owed to them by the respondent. I agree with his Honour's conclusions in regard to proximity.

126 The statement of claim alleges other grounds on which the claimed duty of care should be found to exist (that is, grounds that are different to those to which I have already referred). None of these other grounds was mentioned by counsel for the appellants in his written outline of submissions or in his oral argument. It is only necessary to refer to one of them. That is the argument that the respondent owed the appellants a duty of care, derived somehow or other (in a way not pleaded or otherwise explained) from the duty of care it owed James as his employer, such that, upon breach of the duty to James, the respondent would be liable to the appellants for psychiatric injury sustained by them.

127 There is a difficulty in principle in the proposition that, merely because the respondent owed James a duty of care as his employer, he owed the appellants a duty of care. Counsel for the appellants, understandably, made no attempt to address this question.

128 Even if it is accepted that the appellants relied on the respondent to fulfil its duty of care to James, it does not follow merely from this that the respondent owed a duty to the appellants not to cause them psychiatric harm. I have pointed out that the appellants did not alter their position or take or refrain from taking steps because of their reliance on the respondent, and there was no contractual or other relationship known to the law between the respondent and the appellants. The appellants' claim is governed by the ordinary principles of the law of tort and these contain restrictions on the recovery of compensation for such harm. These principles do not presently recognise that an employer owes the parents of an employee a duty of care not to cause them psychiatric injury, stemming merely from the existence of the duty of care owed by the employer to the employee to provide a safe system of work, together with the fact that the parents, to the employer's knowledge, rely on the employer (in the general sense I have explained) to comply with that duty. The employment relationship between the respondent and James does not, in itself, contain a sufficient basis for liability for the psychiatric harm sustained by the appellants.

129 In my view, the appellants have failed to establish the requisite elements of foreseeability and proximity. I would dismiss the appeal.