(back to preceding text)

    57.  Secondly, the drafting history, examined in Saks at pp.401-3, indicates that it was deliberately decided to address separately and in separate terms liability for (a) persons, (b) goods and baggage and (c) delay. The original draft submitted to the second international conference in Warsaw in 1929 would, in contrast, have imposed liability "for damage sustained during carriage" in all these three cases, subject only to the carrier's ability to avoid liability under article 22 by proving it had taken reasonable measures to avoid liability. But the process of negotiation led to a differently conceived scheme, which identifies the necessary cause of damage, in relation to persons, as an "accident" (article 17) and, in relation to goods and baggage, as an "occurrence" (article 18), but speaks simply, in relation to delay, of "damage occasioned by delay in the carriage by air" (article 19). As the Supreme Court observed in Saks (p.403), "an 'accident' must mean something different than an 'occurrence' on the plane". Any accident may be regarded as an occurrence, but the reverse does not hold true.

    58.  Thirdly, the pre-condition to liability under article 17 is not simply that there should have been an accident; the damage sustained in the event of death, wounding or bodily injury must be shown to have been caused by an "accident" which "took place on board the aircraft or in the course of any of the operations of embarking or disembarking". For these reasons, courts throughout the world, when addressing claims under article 17, have had to engage in analysis, and application to varying factual situations, of the term of "accident".

    59.  Mr Scrivener does not challenge any of this. He expressly accepts the conclusion of the Supreme Court in Saks (at pp.400-5) that the death or wounding of or other bodily injury to a passenger cannot of itself be regarded as a relevant "accident" under article 17; and that the death, wounding or bodily injury "must be caused by an accident, and an accident must mean something different than an 'occurrence' on the plane" (p.403). Justice O'Connor, giving the Supreme Court's judgment, supported this conclusion by examination of the drafting history (pp.400-403), by reference to subsequent conduct of the parties and subsequent interpretations of the signatories (pp.400 and 403-4) and by reference to what she described as "the weight of precedent in foreign and American courts" (p.400 and 404-5). Again, no issue has been taken with her approach or conclusions. Further, Mr Scrivener expressly accepts the paraphrase of the term "accident" suggested in Saks (at pp. 400 and 405) - "an unexpected or unusual event or happening that is external to the passenger". But he underlines Justice O'Connor's immediately following words:

"This definition should be flexibly applied after assessment of all the circumstances surrounding a passenger's injuries"

    It is in this context that Mr Scrivener lays particular stress upon what he submits are relevant purposive considerations.

    60.  Up to a point, the parties are also on common ground regarding the object and purpose of article 17. Both sides accept that the Convention as amended seeks to strike a balance between competing interests in its articles governing liability towards passengers. It was introduced in the infancy of international air travel at a time when airlines were exposed to unrestricted liability under whatever law was regarded as applicable in whatever forum the plaintiff was able to and did chose. This open legal market situation inhibited financial investment in and growth of the airline industry: cf El Al Israel Airlines Ltd. v. Tseng 525 US 155 (1999), 170 and Zicherman v. Korean Airlines Co. Ltd. 516 US 217 (1996), 230. The Convention brought benefits to both sides. My noble and learned friend Lord Hope described these in Morris v. KLM Royal Dutch Airlines [2002] UKHL7; 2 AC 628, para. 66:

"66.  From the point of view of the passenger or the owner of baggage or cargo, the imposition of liability without proof of fault on the carrier and the nullification of provisions relieving him of liability or restricting the amount of his liability are very significant advantages. From the point of view of the carrier too however there are significant advantages in the system laid down by the Convention. A principal consequence of that system is the exposure of the carrier to liabilities without the freedom to contract out of them. But it defines those situations in which compensation is to be available, and it sets out the limits of liability and the conditions under which claims to establish liability, if disputed, are to be made. A balance has been struck between these competing interests, in the interests of certainty and uniformity."

    61.  Lord Hobhouse put the matter as follows at para. 150:

"150.  Thirdly, the code involves a division of risk. It strikes a balance. It is wrong to construe it as a code designed to advantage one interest or the other, the carrier or the customer. Like any code of this character, it contains familiar types of provision assisting one interest balanced against others assisting the other interest. For the passenger, a simple criterion of causation by an 'accident' is adopted but counterbalanced by strong provisions enabling the carrier to limit his liability. These are not like exemption clauses to be construed against one or other party. …"

    62.  It is now established that the remedy provided by article 17 is exclusive, in the sense that no other common or civil law remedy (e.g. for want of reasonably care or negligence) is available to a passenger who sustains bodily injury or dies in the course of international carriage by air, but has no claim against the carrier under article 17 (e.g. because the injury or death was not caused by an accident on board or in the course of embarkation or disembarkation). The House decided this in Sidhu v. British Airways plc [1997] AC 430, where passengers were captured by Iraqi forces in an air terminal in Kuwait where their aircraft had landed just before the Iraqi invasion of August 1990.

    63.  Both sides also reject an originalist approach to the application of article 17, or a limitation of the term "accident" by reference to the type of aviation disaster affecting a whole airplane that was probably all too familiar in 1929. It is, Lord Hobhouse said in Morris v. KLM Royal Dutch Airlines [2002] 2 AC 628, para. 149

"… mistaken to interpret a convention such as the Warsaw Convention, or the various amended versions of it, as if they were intended to be historical documents frozen in time. They are intended to provide an enduring uniform code which will govern contractual and, where relevant, delictual relationships not just for a finite time but for the future as the transactions to which they apply are entered into. The contracts into which the code is to be incorporated are to be made and will be performed at dates in the future, maybe long into the future. Notionally to relate them back to a supposed state of affairs existing in 1929 is not only wrong but wholly impractical. It leads to the complication and confusion to which I have already referred. It is also destructive of uniformity since when a convention has later been amended, logic would require that one starts the clock again and asks what was in the minds of the delegates at the later conference. It is not necessary in order to understand this point to have regard to the principles applicable to 'always speaking' constitutional documents. The principle is more simple. Words have a meaning which does not change but the application of those words to the decision of any question depends on the facts and circumstances of the case in which that question arises. It is the facts and circumstances of the cases that change, not the meaning of the contractual words."

It is thus fully open to the appellants to argue that the present claims involved "bodily injury" caused by an "accident" within article 7, although that article was originally agreed in an era when no-one conceived that air travel could involve the complex risk of DVT and such international flights as existed involved short refuelling hops which may have meant that it did not.

    64.  At this point, the respective submissions start to diverge. Mr Scrivener submits that the general conception underlying article 17 is that an airline should bear the risk of, and answer for, matters within its control, in respect of which it can take precautions and insure, and that a court should be reluctant to regard an airline as free from responsibility where a passenger has suffered on board an injury due to conditions on board over which the airline had complete control. This, he submits, is especially so, if and so far as any element of culpability could be attributed to the airline. Rather than concentrate on airline practice at the relevant time, he submits that the court should concentrate on what the ordinary passenger or member of the public would have expected of an airline which had, or ought to have had, the awareness of the risks of DVT which the claimants allege; otherwise, an airline could, by relying supinely on existing practice, avoid liability for known risks and would lack incentive to make changes; the flexible approach to the term "accident" suggested by the Supreme Court in Saks should enable and require the court to give effect to such considerations.

    65.  In contrast, Mr Robert Webb QC appearing with Mr Lawson for British Airways argues, in submissions that were adopted by Mr Lawson appearing for China Airlines, that the balance struck by the Warsaw Convention as amended by the Hague Protocol allows limited scope for deeper examination of any underlying object or purpose behind its wording or, indeed, for examination of the consequences of any particular interpretation. The word "flexibly" used by the Supreme Court does not, he submits, necessarily equate with "broadly". Mr Webb refers to my noble and learned friend Lord Steyn's words in Morris at para. 15, and submits that they are equally applicable to the interpretation of the term "accident":

"15.  The Warsaw Convention is an exclusive code of limited liability of carriers to passengers. On the other hand, it enables passengers to recover damages even though, in the absence of the Convention and the Act, they might have no cause of action which would entitle them to succeed: Swiss Bank Corp. v. Brink's MAT Ltd [1986] QB 853, 856G-H, per Bingham J (now Lord Bingham of Cornhill). It is therefore not necessarily right to approach the meaning of the phrase 'bodily injury' in article 17 of the Convention through the spectacles of full corrective justice."

    66.  The Supreme Court's admonition in Saks to apply its suggested definition of accident "flexibly" was illustrated with examples, in particular the interpretation by lower courts of the term "broadly enough to encompass torts committed by terrorists or fellow passengers" (p.405). While this indicates that "flexibly" can mean "broadly", it does not point to a general abandonment of the idea of an "unexpected or unusual event or happening that is external to the passenger". An assault by a terrorist or by another passenger can be regarded as falling within a broad conception of these words. It can, as far as the airline and the passenger are concerned, also be described, without very great difficulty, as an accident of the carriage. But the Supreme Court's definition was suggested with the clear purpose of delineating boundaries to the concept of "accident". The claimant's submission in Saks had been that "accident" embraced any "hazard of air travel", and the Court of Appeals, whose judgment the Supreme Court was reversing, had used a definition derived from another convention (that on International Civil Aviation of Dec. 7 1944), namely "an occurrence associated with the operation of an aircraft which takes place between the time any person boards an aircraft with the intention of flight and all such persons have disembarked". Such an approach would in one respect have introduced a limitation on the concept (to hazards or risk in some way associated with aviation travel) which is certainly not explicit if it exists in any degree in the Supreme Court's paraphrase (which I need not consider - cf Wallace v. Korean Air 214 F.3d 293 (2000); U.S. Court of Appeals, 2nd Circ. for discussion of the point). They would have opened up airlines to liability for bodily injury attributable to any such hazard or risk, whether or not involving anything untoward or unexpected or indeed internal in origin.

    67.  The practical significance of a determination whether or not an accident took place within the meaning of article 17 has altered with changes in the surrounding legal terrain. In Saks the plaintiff was pursuing a state negligence action, which the Supreme Court remitted to the Court of Appeals (pp.394 and 408). Since the Supreme Court's decision in El Al Israel Airlines Ltd. v. Tseng 525 US 155 and the House of Lord's decision in Sidhu v. British Airways plc [1997] AC 430, it has been clear that article 17 is an exclusive remedy in respect of death or bodily injury sustained during air carriage. Prior to those decisions, a claimant might well argue for a restricted interpretation of the term "accident", in order to pursue an unlimited claim for (say) negligence at common law, while carriers might respond by arguing for a wider interpretation (cf Fishman v. Delta Air Lines, Inc. 132 F.3d 138 (1998); U.S. Court of Appeals, 2d Circ. for an example). Since Tseng and Sidhu, these incentives have been reversed, with carriers benefiting by a complete exclusion of liability if they can negative the existence of any "accident" within the meaning of article 17. However, there are complicating cross-currents, capable of inducing different attitudes: under the Montreal Agreement of 1966, carriers for flights originating, terminating or stopping in the United States agreed voluntarily to forego any defence that they had taken all necessary measures (or that it was impossible to take them) under article 20; while the Montreal Convention of 1999 (set out in Schedule 1B to the 1961 Act and in force in the United Kingdom since June 2004) omits altogether any such defence (substituting only limited provisions for exoneration in the case of negligence or other wrongful act or omission of the passenger). At the same time, the Montreal Protocol underlines the requirement of an accident under article 17 which becomes:

"The carrier is liable for damage sustained in case of death or bodily injury of a passenger upon condition only that the accident which caused the death or bodily injury took place on board the aircraft or in the course of any of the operations of embarking or disembarking."

    These changes in the surrounding legal terrain may, as a matter of psychological reality, justify a degree of caution about any attempt to reconcile every past legal decision (cf words of Sutherland J in Quinn v. Canadian Airlines International Ltd. (1994) 18 O.R. (3d) 326, 331c-f). But there are enough decisions by different courts to mean that that is in any event an unpromising exercise. In my view, even if the application has been inconsistent, it is unconvincing to suggest that the essential meaning of the term "accident" in the relevant articles has in law changed.

    68.  In Sidhu, where the British courts recognised the exclusivity of article 17, my noble and learned friend Lord Hope said at p.453H-454A that "No doubt the domestic courts will try, as carefully as they may, to apply the wording of article 17 to the facts to enable the passenger to obtain a remedy under the Convention", before going on to point out that it was, in that case, conceded that no such remedy was available. A similar thought was expressed by Kirby J in Povey v. Qantas Airways Ltd. [2005] HCA 33 when he said at para. 137 that, although the travaux préparatoires

"show the compromises that were struck in adopting the text of the Warsaw Convention, they also indicate that that recovery of passengers of acceptable compensation for damage, consequent upon injury, as defined, was an objective of the Warsaw Convention";

    and also when he said, with reference to article 25, that

"[if] airlines with the requisite knowledge deliberately or recklessly withheld essential information from passengers (assuming that to be provable) the imposition of liability for consequent serious damage involving injury or death would not be entirely surprising, given the overall objects of the treaty".

    Nonetheless, Kirby J joined with the majority in concluding that the incidence of DVT in Povey had not been caused by any accident. In a similar vein Callinan J at para.205 questioned whether the concept of "accident" might not now be giving carriers by air "the benefit of an anachronistic approach to the perils of travel".

    69.  In an impressive review of authority and the factors in play in this difficult area it was suggested by Marrero DJ in Fulop v. Malev Hungarian Airlines 175 F.Supp.2d 651 (S.D.N.Y. 2001), 657 that, both before and after the decision in Saks, courts grappling with the application of article 17 had weighed various factors, focusing in particular on

"the relationship of the claimed accidents to (1) the normal operation of the aircraft or airline; (2) the knowledge and/or complicity of crew members in the events surrounding the alleged accident; (3) the acts of fellow passengers, whether intentional or not; (4) the acts of third persons who are not crew or passengers, e.g. hijackers and terrorists; (5) the location of the occurrence in the continuum of travel; (6) the role, condition and reaction of the complainant in connection with the occurrence at issue; and (7) the kind of risks inherent in air travel."

    He went on:

"Insofar as any decisive pattern may be discerned [in the authorities], the fulcrum of these considerations often rests on the extent to which the circumstances giving rise to the claimed accident fall within the causal purview or control of the carrier - or at least within its practical ability to influence -as an aspect of the operations of the aircraft or airline. The larger the role of the airline in the causal chain, and the greater the knowledge and involvement of its personnel and operations in bringing about the harmful event, the more likely it is that liability will be found.

Conversely, as the causal balance shifts towards acts and conditions that are independent of the knowledge or will of the carrier, or not associated with the operation of the aircraft or airline nor arising from risks characteristic of air travel, and instead are more unique to the passenger alleging injury, the lesser the claimant's probability."

Taking the paraphrase suggested in Saks which both sides accepted as a starting point, the issue before us can be broken down into three aspects, namely whether there was (a) an event, which was (b) unexpected or unusual and (c) external to the passenger. These could be viewed as involving, if only broadly, Marrero DJ's suggested factors (1), (2), (5), (6) and (7). But in relation to issue (a), the decision of the Court of the Appeal raises a sub-issue not explicitly identified by Marrero DJ, namely the extent, if any, to which inaction can constitute an event. Further, factor (2), crew "knowledge and/or complicity", arises in an expanded form in this case, having regard to the appellants' positive case that the airline or aircrew were at fault; the allegations of fault, relating to the condition of the aircraft and/or failure to warn of the risks of DVT, are advanced as relevant to all three issues, that is to show that (a) there was an event, (b) it was unexpected or unusual and (c) it was external to the appellant passengers. In this case, if they are sufficient to establish an event and that it was unexpected and unusual, then it would also be external. If there was no such unexpected or unusual event, the injury suffered by the appellants would flow from internal causes. So it is on issues (a) and (b) that it is necessary to focus.

    70.  In the Court of Appeal, [2003] EWCA Civ 1005; [2004] QB 234, Lord Phillips of Worth Matravers MR, in a judgment with which Judge and Kay LJJ agreed while adding some reasoning of their own, disposed of the appellants' claim on the basis that it involved no event, merely alleged inaction on the carriers' part in relation to a standard flight on a standard aircraft operating normally and without any undue incident. He said at para. 25:

"25.  A critical issue in this appeal is whether a failure to act, or an omission, can constitute an accident for the purposes of article 17. Often a failure to act results in an accident, or forms part of a series of acts or omissions which together constitute an accident. In such circumstances it may not be easy to distinguish between acts and omissions. I cannot see, however, how inaction itself can ever properly be described as an accident. It is not an event; it is a non-event. Inaction is the antithesis of an accident."

In deciding that the matrix disclosed no event, he said, first, that the permanent features of the aircraft, or the subjecting of the passengers to carriage in aircraft with these features was not capable of amounting to an "accident" within article 17 (para. 28), and, secondly, that the alleged failure to warn of the risk of DVT, or to advise on precautions which would avoid or minimise the risk, could also not be categorised as an accident: it was "simply something that did not happen - a non-event" (para. 29). He went on to say that "The question whether there took place on the flight events which were 'unexpected or unusual' was and is unreal, having regard to the judge's conclusion and my provisional conclusion that there was no relevant event" (para. 3l), adding only at paras. 37 and 38:

"37.  I am not persuaded that, when considering whether an action has the unusual, unexpected or untoward quality necessary to enable one to qualify it as an accident under the Air France v. Saks test, the questions of whether the actor has knowledge which makes it culpable to perform the action can never be relevant. I am, however, persuaded that it is simply not possible to apply to a state of affairs, or an omission to act, the test that is relevant to deciding whether an event is an accident.

38.  Looking at the position overall, which is the only realistic way of looking at it, the matrix, as augmented by the Further Information, discloses nothing which is capable of constituting an "accident" within the meaning of article 17."

    71.  Later, he addressed the decisions of the United States District Court and Court of Appeals in Husain v. Olympic Airways (2000) 116 F Supp 2d 1121 and 316 3d 829 (since affirmed by the Supreme Court by a majority at 540 US 644 (2004), 124 S.Ct. 1221 (2004)). The facts there were that Dr Hanson suffered from asthma and was particularly sensitive to second-hand smoke. He and his wife had requested non-smoking seats, but found that these were only three rows in front of the smoking section, which was not partitioned off. His wife's requests on three occasions (twice while the aircraft was on the ground and once in the air) that he be moved further away were refused by the flight attendant, who asserted on the second and third occasions that there were no spare seats, though she also said on the third occasion that they could ask other passengers to exchange seats, but would not be assisted by the crew. Dr Hanson walked to the front of the plane to get some fresh air, and there collapsed and died. The lower courts found that the airline's "failure to act was a 'blatant disregard of industry standards and airline policies", and that it satisfied the meaning of "accident". The Master of the Rolls did not question the result, but did question the reasoning, suggesting (para. 50) that:

"The refusal of the flight attendant to move Dr Hanson cannot properly be considered as mere inertia, or a non-event. It was a refusal to provide an alternative seat which formed part of a more complex incident, whereby Dr Hanson was exposed to smoke in circumstances that can properly be described as unusual or unexpected. … the direct cause of his death was the unnecessary exposure to the smoke. The refusal of the attendant to move him could be described as insistence that he remain seated in the area exposed to smoke. The exposure to smoke in these circumstances could … properly be described as an unusual or unexpected event."

    72.  The decision in Husain was affirmed in the Supreme Court, where the Master of the Rolls' words were quoted by the majority as indicating that their conclusion was not inconsistent with that of the Court of Appeal in the present case. However, the majority in the Supreme Court also rejected (in footnote 9 on p.1229) the Master of the Rolls' "precise reasoning" to the extent that it differed from their own, illustrating what they regarded as "the fallacy of the petitioner's position that an 'accident' cannot take the form of inaction" with the example (in footnote 10) of a failure in accordance with industry standards to divert a flight when a passenger otherwise faces imminent death as illustrating that

"the failure of an airline crew to take certain necessary vital steps could quite naturally and, in routine usage of language, be an 'event or happening'"

Mr Scrivener places particular reliance on these last footnote words, but they must be read in context. The essence of the majority's reasoning at pp.1229-1330 was that the combination of an expected and usual reaction and a categorical refusal or failure by the crew to react in that way on a particular flight could amount to an "accident".

    73.  The minority judges in Husain (Justices Scalia and O'Connor) took issue with the majority's conclusion that inaction could amount to an "accident". They agreed with the Master of the Rolls' analysis of the legal position, while questioning whether it was on the particular facts right to treat the flight attendant as having "insisted" on Dr Hanson remaining seated where he did (since she had in fact left it to Dr Hanson to try to find a seat elsewhere). They did not consider that anything unusual or unexpected had necessarily occurred at the time of the two requests prior to take-off, and they considered that, subject to one point, the flight attendant had been guilty of no more than inaction at the time of the third request. The one point which in their view required remission to the District Court related to the attendant's repeated representation that the plane was full. If that had led Dr Hanson or his wife not to look for another seat, and was both unusual and unexpected (as to which the minority evidently felt little difficulty) and a cause of his death, then it could be said to be an "accident" under article 17.

    74.  The Court of Appeal and the minority in Husain would thus draw a line between action and inaction, while the majority in Husain distinguished between on the one hand the unusual and unexpected and on the other hand the usual and expected (in each case whether involving action or inaction). But the difference between the two approaches may not be as great as might at first sight appear, for reasons indicated by passages from paras 25 and 50 of the Master of the Rolls's judgment quoted in paras. 70 and 71 above. These passages come close to saying that an event may in some circumstances exist, where there has been crew inaction in a context where action would normally be expected. A similar view was expressed by Kirby J in para. 172 of his separately reasoned judgment concurring in the result with the majority in Povey in the High Court of Australia. If that may be so, then the question whether there has been an event inter-relates with the question whether what happened was unexpected or unusual. I therefore turn to that further question.