. 4
( 18)


3.1 The nature of the duty of care
The concept of negligent conduct, which was discussed in chapter 2, together with
the notions of causation and remoteness of damage (which are discussed in ch. 5),
may be said to constitute the concept of fault as embodied in the tort of negligence.
But not all faulty conduct in this sense gives rise to legal liability. The tort of negli-
gence, it is sometimes said, cannot be committed ˜in the air™. A person will be liable
for negligent conduct only if that person owed the claimant a duty to take care. In
the famous case of Donoghue v Stevenson1 Lord Atkin enunciated the equally famous
˜neighbour principle™ according to which a duty of care is owed to persons whom you
ought reasonably to foresee as likely to be injured if you do not take reasonable care.
On the basis of this principle it was, for many years, said that the test of duty of care
was foreseeability. However, in the 1980s the House of Lords became dissatis¬ed with
this test, especially in relation to cases involving liability for economic loss; and in a
series of cases2 it developed a threefold test for the imposition of a duty of care: ¬rst,
was it foreseeable that the claimant might su¬er damage if the defendant did not take
reasonable care? Secondly, was there a su¬cient relationship of proximity between
the claimant and the defendant? And, thirdly, is it just and reasonable in all the cir-
cumstances of the case to impose a duty of care. The House of Lords has also shown
unwillingness, in some cases at least, to depart from well-established common law
rules denying a duty of care even if these three requirements are satis¬ed.
We have already examined the concept of foreseeability and come to the conclu-
sion that it signi¬es little more than that liability will be imposed if the court thinks
it fair that the defendant should bear responsibility. Some judges have been prepared
to admit that the notion of proximity is also just a means of giving e¬ect to (while
at the same time concealing) value judgments about the proper scope of liability for
negligently caused injury. This is obviously true of the third criterion of duty: justice
and reasonableness. Unwillingness to depart from old rules usually arises out of a
desire not to upset settled expectations especially in the business community.
These developments in the law relating to duty of care have mainly a¬ected lia-

1 [1932] AC 562.
2 The leading case is Caparo Industries plc v. Dickman [1990] 2 AC 605.

The scope of the tort of negligence 69

bility in the tort of negligence for purely economic loss, that is loss other than injury
to person or damage to tangible property, and economic loss consequential on such
injury or damage.3 In the typical case of personal injury (except ˜nervous shock™)
or damage to tangible property, foreseeability is, in practice if not in theory, the sole
criterion of the existence of a duty of care. Therefore, the threefold test of duty is
not of much importance to the subject matter of this book. It should also be noted
that whereas Lord Atkin seems to have put forward the neighbour principle as a
way of expanding the scope of liability for negligence, the duty of care concept is
most commonly used in modern cases as a means of justifying refusal to impose
liability for negligence.
This brief account of the law indicates that the main function of the concept of
duty of care is to de¬ne the boundaries of liability for damage caused by negligent
conduct by reference to what are commonly called ˜policy considerations™. So, for
example, for fairly obvious reasons, soldiers owe no duty of care to fellow soldiers
when engaging the enemy in battle; nor is the army under a duty to provide a ˜safe
system of work™ on the battle¬eld.4 Until 2000, barristers owed no duty of care to
their lay clients in the conduct of litigation in court, even if such conduct was neg-
ligent and caused foreseeable damage to the client, in order (it was said) to avoid
creating con¬‚icts between the barrister™s duties to the court and to the client.5 There
is no duty to take care not to cause a person economic loss by damaging tangible
property belonging to a third party because, it is said, the extent of liability such
loss may be ˜indeterminate™ and, perhaps, uninsurable. And here is a ¬nal example:
suppose a doctor negligently performs a sterilization operation with the result that
a woman conceives and bears a child. Whether healthy or disabled, the child can
recover no damages; and, at least if the child is healthy, the only remedy available is
an award of £15,000 for interference with the mother™s ˜reproductive autonomy™.6
In each of these cases, the denial of liability is based on value judgments about the
desirability of imposing liability in the type of case in question. In the last type of
case, courts have typically not justi¬ed the result by denying the existence of a duty
of care to the child, but simply by saying that to allow recovery would be undesir-
able for various reasons. To say that a person owes a duty of care in a particular situ-
ation means (and means only) that the person will be liable for causing damage by
negligence in that situation.7

3 This extremely complex topic is beyond the scope of this book. See generally P. Cane, Tort Law
and Economic Interests, 2nd edn (Oxford, 1996).
4 Mulcahy v. Ministry of Defence [1996] QB 737.
5 The rule was reversed in Arthur JS Hall & Co. v Simons [2002] 1 AC 615.
6 The leading case in this di¬cult area (for the moment, at least) is Rees v. Darlington Memorial
Hospital NHS Trust [2004] 1 AC 3009.
7 For criticism of this view see N.J. McBride, ˜Duties of Care “ Do They Exist?™ (2004) 24 OJLS 417.
I agree with McBride that negligence law is at least as concerned with telling people how they
ought to behave as with imposing liability for failure to behave as the law prescribes. In this
sense, to say that a person owes a duty of care means more than that they can be held liable for
negligence. But in practice, the only function of the duty-of-care concept in legal reasoning is to
de¬ne the scope of liability to pay damages for negligence.
70 Chapter 3

Use of the duty-of-care concept to create immunities from negligence liability has
been particularly controversial in relation to the liability of public authorities, such
as the police, and education and welfare agencies. Although the reasoning in such
cases tends to be very complex, what it boils down to is that a public authority will
be immune from liability for negligence in the performance of its statutory functions
(i.e. will owe no duty of care to persons injured by its negligence) unless the court
thinks that imposing such liability would be compatible with the terms of the rele-
vant statute and would not interfere unduly with the performance of those functions.
In Osman v UK 8 the European Court of Human Rights (ECtHR) held that this tech-
nique for denying liability was inconsistent with the right to a fair hearing under
Article 6 of the European Convention on Human Rights (ECHR). The ECtHR
objected to the fact that the English court decided the duty issue as a ˜preliminary
point of law™, without giving detailed consideration to all the facts of the case. The
House of Lords in Barrett v En¬eld LBC 9 was in¬‚uenced by this decision and held, in
e¬ect, that the issues of compatibility with the statute, and so on, should be resolved
on the basis of a careful consideration of all the relevant facts of individual cases, and
not by creating what were called ˜blanket immunities™ for particular functions (such
as taking a child into care or providing protection to potential victims of crime). The
ECtHR has since resiled from its approach in Osman,10 but it has left open the pos-
sibility that use of the duty-of-care technique might infringe the right to an e¬ective
remedy in a national court for breaches of the ECHR (under Article 13 of the ECHR)
in a case where the defendant™s allegedly negligent conduct also constituted a breach
of a Convention right.11 The argument is that denying liability on the basis of a ˜no-
duty™ immunity might preclude proper investigation of the claimant™s allegations and
hence deny the claimant an e¬ective remedy.
This chapter contains an examination of certain issues relevant to legal liability
for death and personal injury that are usually discussed by lawyers in terms of
whether a duty of care is owed. In other words, these are issues relevant to the scope
of legal liability for negligently in¬‚icted death and personal injury.

3.2 Speci¬c duty issues
3.2.1 Common situations in which duties of care have been imposed
In practice, the two most important areas of tort liability for death and personal in-
juries relate to road accidents and industrial accidents. Legal liability for negligence
resulting in road accidents has been recognized certainly since the seventeenth

8 (2000) 29 EHRR 245.
9 [2001] 2 AC 550.
10 TP & KM v. UK (2002) 34 EHRR 42; Z v. UK (2002) 34 EHRR 97; DP & JC v UK (2003) 36 EHRR
11 E.g. in McGlinchey v. UK (2003) 37 EHRR 41 conduct of prison authorities that amounted, in
e¬ect, to negligent failure to care adequately for a sick prisoner was held to constitute a breach of
the right not to be subjected to inhuman or degrading treatment under Article 3.
The scope of the tort of negligence 71

century and perhaps earlier. There has never been any doubt that those using the
highways are under a duty of care in so doing, and the legal position today is plain:
any person using the roads, whether as a motorist, pedestrian or cyclist, will be liable
if, by positive action,12 that person negligently causes physical injury to anybody else.
A lawyer would scarcely ever waste time in an ordinary road accident case by inquir-
ing whether the defendant owed a duty of care to the claimant. This would simply be
taken for granted. The general principle also holds good for positive, negligent action
resulting in industrial accidents.
Another important area of negligence liability is liability for defective products
(although the importance of the common law has been considerably reduced by
enactment of a regime of ˜strict™ liability in the Consumer Protection Act 1987,
which is discussed in ch. 4). The leading case is Donoghue v Stevenson, to which we
have already referred. In this case the claimant allegedly su¬ered gastroenteritis
and shock as a result of drinking a bottle of ginger-beer which was said to have
contained the remains of a decomposed snail. The bottle of ginger-beer had been
bought for the claimant by a friend in a caf©, but the claimant sued the manufac-
turer. The question at issue was whether, assuming that the presence of the snail
was due to lack of reasonable care on the part of the manufacturer, it would be
liable to the claimant. To us, it may seem astonishing that the answer could ever
have been in doubt, since there are several good arguments in favour of liability in
such circumstances “ the desire to compensate the claimant for injuries; the value
of providing an incentive for manufacturers of food and drink for public con-
sumption to take precautions against such events; and ¬nally the fact that the
manufacturer is better able than the consumer to bear the loss and distribute it by
making allowance for it in the price of its products. Nevertheless, despite all this,
liability was very much doubted at the time of the case, and it is generally agreed
that the majority in the House of Lords, in ¬nding for the claimant, made ˜new law™
by departing from precedents suggesting that there would be no liability on such
The real importance of Donoghue v Stevenson was that it decided that a claimant
could recover damages for negligence against a defendant even though there was
no contract between them (in other words, even though they were not ˜in privity of
contract™ with one another). Although liability for negligent acts was well estab-
lished, long before this case, in some areas (such as road accidents) even in the
absence of any contractual relationship, there is no doubt that the privity-of-
contract principle had become a severe limitation on the extension of the law of
negligence, and that it had been used by the courts in the nineteenth century to
restrict liability for negligent acts. Donoghue v Stevenson removed this restriction
on liability for negligence, and this began a movement towards general liability for
physical damage caused by positive, negligent conduct that has been going on ever

12 Concerning omissions, see 3.2.2.
72 Chapter 3

A signi¬cant area of law for our purposes is that relating to the liability of an
occupier of premises for personal injury and property damage su¬ered on the
premises by persons who come onto (or ˜visit™) them (˜occupiers™ liability™).
Loosely speaking, visitors are divided into two classes, namely lawful and unlaw-
ful visitors. As a result of the enactment of the Occupiers™ Liability Act 1957, the
liability of an occupier to lawful visitors has been very largely assimilated to that
of ordinary liability in negligence, so that occupiers now owe to their lawful visi-
tors a ˜common duty of care™, which is for all practical purposes indistinguishable
from the ordinary duty-of-care concept used in most common law actions for neg-
ligence. Nothing here need detain us because it is plain that occupiers always owe
their lawful visitors a duty to take care, and are therefore liable for causing them
physical injury by negligence. Indeed liability can be imposed on an occupier
either for positive negligence (˜misfeasance™) or negative negligence (˜nonfeasance™
or ˜omission™). Liability to unlawful visitors (or ˜trespassers™) is governed by the
Occupiers™ Liability Act 1984, which imposes on occupiers a duty to take such care
as is reasonable considering, in particular, that trespassers by de¬nition force their
presence on the occupier without the latter™s consent.
Another important statutory source of negligence liability is the Defective
Premises Act 1972, s. 4 of which imposes extensive liability on landlords for injury
caused, to persons coming onto rented premises, by failure to repair. A number of
other statutes have reversed common law rules denying the existence of a duty of
care in various circumstances. For example, the Animals Act 1971 largely (though
not quite entirely) removes an immunity from liability once enjoyed by owners of
animals which cause injury or damage as a result of straying on to a highway. The
Highways (Miscellaneous Provisions) Act 1961 imposes liability on highway author-
ities, for negligently failing to repair a highway; and the Law Reform (Husband and
Wife) Act 1962 enables husbands and wives to sue each other for negligence (with a
view to obtaining damages from insurance companies). The Congenital Disabilities
(Civil Liability) Act 1976 also eliminates any doubts about another possible no-duty
situation by making it clear that, in general, legal liability will exist for negligently
in¬‚icting injuries on an unborn child.

3.2.2 The distinction between acts and omissions
The paradigm instance of negligence liability arises where bodily injury or pro-
perty damage results from what we might call ˜positive conduct™ “ where, for
instance, two speeding cars collide injuring occupants and vehicles. Lawyers often
refer to positive conduct as ˜misfeasance™, and they contrast this with ˜nonfeasance™.
This contrast may also be expressed in terms of a distinction between ˜acts™ and
˜omissions™ (failures to act).13 It must be said at the outset that there are many

13 See T. Honor©, ˜Are Omissions Less Culpable?™ in P. Cane and J. Stapleton eds., Essays for Patrick
Atiyah (Oxford, 1991); H.L.A. Hart and T. Honor©, Causation in the Law, 2nd edn (Oxford, 1985);
M.S. Shapo, The Duty to Act (Austin, Texas and London, 1977).
The scope of the tort of negligence 73

situations in which it is impossible to draw any clear line between misfeasance and
nonfeasance. A solicitor instructed to draft a will allows it to be wrongly witnessed
so that it is invalid: this may be seen as misfeasance in preparing the will or as failure
to ensure that it was properly witnessed. A person digs a hole on their land and a
visitor falls into it: this may be seen as a¬rmative conduct in digging the hole or as
nonfeasance in failing to fence the hole or give a warning. A person turns right
across a line of tra¬c without signalling: this is either positive bad driving or a neg-
ative failure to signal.
More generally, whether failure to act is viewed as nonfeasance or misfeasance
depends largely on whether the failure is viewed in isolation or as part of a larger
activity. Nevertheless, despite the di¬culty of the distinction, the law recognizes
and acts on it. It is an important aspect of the di¬erence between tort and contract
liability: a person is often not bound to take positive action unless they have agreed
to do so, and have been paid for doing so; but people are in general bound to abstain
from causing damage by negligence whether or not they have agreed to do so, or
have been paid for doing so. Tort law embodies a general bias against imposing lia-
bility for nonfeasance. However, it is by no means the case that failure to take rea-
sonable steps to prevent another su¬ering injury or loss is never actionable in tort.
What lies behind the distinction between nonfeasance and misfeasance? How
can we regard absent-mindedly driving a motor vehicle through a red light as
something more reprehensible than walking by while a child is drowning in a few
feet of water? Yet failure to save a drowning stranger is a stock example of a clear
case of immunity from liability for negligence (and indeed in the criminal law).
Such questions are often posed in terms of a distinction between law and morality:
if tort law is based on some concept of moral fault, why does it embody quite a
sharp distinction between acts and omissions? There are at least two reasons why
we need to be a little wary of thinking about the issue in this way. First, although,
according to some views about morality, nonfeasance may be just as reprehensible
misfeasance, many people would give at least some moral weight to the distinction
between acts and omissions in deciding the right thing to do in various situations
and in assessing the behaviour of others. There is no single version of ˜morality™ that
can be easily contrasted with ˜the law™. Secondly, it is one thing to say that render-
ing assistance to someone in danger or distress is (morally) the right thing to do,
but quite another to say that a person who fails to do it should be (legally) obliged
to pay compensation for harm resulting from the failure to act. A good reason for
not turning every moral duty into a legal obligation lies in the nature of legal sanc-
tions and remedies compared with the sorts of disapproval with which breaches of
morality are often met. And a person who has reservations about imposing oblig-
ations in tort to compensate for harm resulting from nonfeasance would be even
less willing to use the criminal law to punish failure to act.
Nevertheless, it is often assumed that the law™s approach to nonfeasance is out
of step with morality; and so it is worth asking why the law of tort distinguishes
between misfeasance and nonfeasance. In seeking explanations it is necessary, ¬rst,
74 Chapter 3

to remember that in this book we are primarily discussing cases of physical damage
and injury. We will not consider cases of nonfeasance that cause only ¬nancial loss,
as where a person fails to warn another that they face a risk of su¬ering ¬nancial
loss, which the former knew about but the latter did not.14 The question to be
answered is how we can justify immunity from liability for nonfeasance causing
physical damage or injury.
The ¬rst possible consideration is that the imposition of duties to prevent harm
(by failure to act) is often more burdensome than the imposition of duties not to
cause harm (by acting). In its main spheres of practical operation the law of negli-
gence tends to prescribe not what we are to do, but only how we are to do things we
choose to do. Thus I am generally quite free to drive my car when and where I want
to on the roads; and it is not particularly onerous to be required to drive it carefully.
This obligation does not prevent me going where I want to, when I want to, though
it may force me to go a little more slowly than I might have chosen. And even when
the law does impose duties to prevent harm, they are frequently of a type that does
not involve much expenditure of time and e¬ort.
Requiring someone to render assistance to (or ˜rescue™) persons in danger may
not only be burdensome in time and e¬ort, but may also involve expenditure of
signi¬cant amounts of money, and involve signi¬cant risks to the person(s) provid-
ing assistance. If the law were to impose a general obligation to rescue, who would
pay the costs of so doing? And if there is an element of risk in rendering assistance,
and the risk eventuates, should the rescuer be able to recover compensation for this?
Suppose a person dies in the course of rescuing someone who is drowning: who is
going to maintain any dependants of the dead person? English law does not recog-
nize any general right to reward or even recompense for rescue. If someone (includ-
ing the rescued person) created the dangerous situation by negligence, the rescuer
may have an action against that person for costs incurred and any injuries su¬ered
in e¬ecting the rescue. It should be noted, however, that in practice, an action against
the rescued person would rarely be covered by liability insurance.
A second reason why it may be felt desirable to distinguish between misfeasance
and nonfeasance is that in the case of misfeasance, the defendant is normally self-
identi¬ed by the conduct that results in harm. On the other hand, a person accused
of nonfeasance is likely to feel, ˜Why pick on me? I didn™t do anything.™ This sort of
reaction may take one of two forms. First, the person may be asserting that they
are merely one of hundreds, and that it is unfair to pick on one individual while
bypassing all the others. If, for instance, a driver negligently knocks down and
injures a pedestrian, it is easy enough to justify fastening liability onto the driver for
misfeasance. But if the unfortunate pedestrian is left bleeding in the road and a
dozen, or a hundred, other motorists drive by without stopping to render assis-
tance, there would be no clear justi¬cation for imposing liability for nonfeasance

14 E.g. Banque Keyser Ullman v Skandia Insurance [1990] 1 QB 665; Reid v Rush & Tompkins [1990]
1 WLR 212.
The scope of the tort of negligence 75

on any one rather than another. In this type of situation a person is unlikely to deny
that they ought to have stopped to render assistance: the complaint is that so ought
many others.
The reaction may take a second form, which raises rather di¬erent issues. The
person accused of nonfeasance may be asserting that it was not up to them to do
something, and that the burden of taking the desired precautions really rested on
someone else “ perhaps on the person in danger or on a third party, but in any case
not them. For instance, suppose that a window cleaner sent by an employer to clean
the windows at a block of o¬ces is injured in a fall resulting from the use of defec-
tive belt-hooks. If the cleaner sues the employer, the complaint will probably be one
of nonfeasance; that is, the employee will be complaining that the employer ought
to have checked the belt-hooks or ought to have supplied safer means of cleaning
windows that did not depend on possibly unsafe belt-hooks. In this situation the
employer™s reply will probably be: ˜It was not my responsibility to check the belt-
hooks. The occupiers of the o¬ces should have seen that the belt-hooks were safe;
it was their responsibility, not mine.™ In some cases the employer might argue that
it was the claimant™s own responsibility to take the necessary precautions. For
instance, suppose an employee™s belongings are stolen at their place of work.15
Whose responsibility is it to take precautions against this possibility: the employee™s
or the employer™s?
A third possible ground for distinguishing between nonfeasance and misfeasance
is based on notions of causation. In some cases at least, we would hesitate to say that
a person guilty of nonfeasance caused injury or damage, and might prefer to say that
they failed to prevent injury or damage being caused by someone or something else.
Assuming that if assistance had been given to a person in danger, it would have pre-
vented the injury or damage occurring, there is, of course, a sense in which a person
who negligently failed to render such assistance ˜caused™ the injury or harm “ if they
had helped, the injury or harm would not have occurred. But there is a di¬erence
between saying that a person caused harm in this (˜counterfactual™) sense and saying
that they should be held responsible for the harm and, perhaps, liable to pay com-
pensation to the victim. According to the ˜causal™ argument for distinguishing
between acts and omissions, we might want to say that although harm would not
have occurred but for a person™s negligent nonfeasance, nevertheless the person did
not ˜really™ cause it, and so should not be held legally responsible for it.
As grounds for creating immunities from liability for nonfeasance, these argu-
ments are weak. The ¬rst point “ the possible burdensomeness of a¬rmative
obligations “ can easily be met. For one thing, the immunity from liability for non-
feasance does not apply only where it would be unduly burdensome to require
a¬rmative conduct. It applies even, for instance, where all that a person has to do is
shout a warning or make a telephone call. A person sees a stranger™s house on

15 Deyong v Shenburn [1946] KB 227 (held that the employer could not be held liable for the theft
of the employee™s belongings by a third party).
76 Chapter 3

¬re: how burdensome would it be to require that person to telephone for the ¬re
brigade? A sighted person can watch a blind stranger walk straight into a hole in
the road without liability in tort or any other branch of the law. What burden would
it be to require the sighted person to shout a warning? Even if a duty to render assis-
tance would in some cases involve a signi¬cant burden, there seems no reason why
the burdensomeness of the duty sought to be imposed should not be weighed
against the bene¬t of preventing harm. If the burden seems disproportionate to the
bene¬t, no duty need be imposed. Such an exercise would simply involve applying
to cases of nonfeasance the negligence formula used when determining liability for
misfeasance. No doubt it would be advisable to move cautiously here so that the
standard of what it is reasonable to expect by way of obligations to prevent harm is
not pitched too high.
The second argument, that there is di¬culty in identifying the person liable for
nonfeasance, and that to impose liability would often be to fasten onto the nearest
convenient defendant, can be rebutted by observing that the law does this even
where misfeasance is in issue. For example, bad road design contributes to many
motor vehicle accidents, but road authorities are rarely sued because the negligent
driver is a much more convenient target. The fact that the driver may be less or no
more culpable than the road designer is not of much importance given that the
driver will always be insured and will not pay the damages personally. So the ques-
tion is not whether there are others more or equally culpable, but whether this
defendant was personally at fault.
As for the third (causal) argument, it does often seem easier to justify holding
someone responsible for an outcome which that person has ˜caused™ by a¬rmative
conduct, and we may have doubts about treating nonfeasance as a cause at all.
Many people intuitively feel that nonfeasance can be treated as a ˜cause™ only if
there was a duty to act.16 Take the following example:17 suppose a passenger on a
small pleasure boat falls overboard, through their own carelessness, into ice-cold
waters, and eventually drowns. We might well want to say that a fellow passen-
ger who did nothing to help the drowning person could not be held in any way
responsible for, and did not ˜cause™, the death by failing to jump in and attempt a
rescue, because they were under no duty to take such action. But we might feel
di¬erently if the owner of the boat failed to attempt to manoeuvre it into a posi-
tion where a lifeline could be thrown to the drowning person. The owner could
more easily be said to be a cause of the death because under the circumstances the
owner surely had a duty to take advantage of having control over the vessel to help
the passenger.
However, this causal argument is open to a strong objection. At ¬rst sight, to say
that the defendant™s nonfeasance did not cause the claimant™s loss seems to provide

16 For a review of the literature on this question see A.M. Honor©, International Encyclopedia of
Comparative Law vol. XI, (1971), ch. 7, ss. 25“8.
17 See Horsley v MacLaren (The Ogopogo) [1971] 2 Lloyd™s Reports 210 on which this example is
The scope of the tort of negligence 77

a sort of objective justi¬cation for not imposing liability. But the way we view the
causation issue depends on whether we think that the defendant ought to have
done something to help; and if so, whether we think that breach of this duty ought
to be translated into a legal liability to pay compensation. These are matters of judg-
ment that cannot be resolved in any ˜objective™ way. Of course, the question of
whether the harm would have been prevented if the defendant had taken action “
in other words, whether the defendant™s failure to act caused the harm in a coun-
terfactual sense “ can, in theory at least, be answered ˜objectively™ or scienti¬cally,
on the basis of facts alone and without having to make value judgments. But the
question of whether the defendant™s nonfeasance ˜really™ caused the harm cannot
be answered in this way because in this richer sense of the word ˜cause™, the language
of causation provides a way of expressing a judgment about the proper limits of
responsibility and liability for negligent failure to act.
While each of these three arguments can be used to explain and rationalize
certain cases in which it does not seem fair to impose legal liability for nonfeasance,
none justi¬es a sharp distinction between misfeasance “ which may attract legal
liability for resulting harm “ and nonfeasance “ which will not. The distinction
between acts and omissions is an important one, but it does not mark the bound-
ary between liability and no liability. In fact, there are various situations in which
tort liability for nonfeasance can arise, and these can be conveniently grouped
under several headings. Undertakings
A person who contracts to do something may incur liability for not doing it. There
is also a somewhat hazy and undeveloped part of the law dealing with voluntary
(i.e. non-contractual) undertakings. Suppose a motorist comes upon a car accident
in a remote area and tells a person badly injured in the accident that she will call
an ambulance from the nearest settlement, but then fails to do so. This would seem
a strong case for liability especially if, in reliance on the undertaking, the victim
declined help from someone else. Or suppose an altruistic citizen who regularly
frequents an isolated beach and is trained as a lifeguard, o¬ers her services as a
voluntary lifeguard to those using the beach. This might justify imposition of lia-
bility if the self-appointed lifeguard made no attempt to rescue someone who relied
on the o¬er in using the beach. Not unrelated are cases in which a person carries
on an especially dangerous activity or creates a physically dangerous situation: here
the law might well impose a duty to take reasonable steps to obviate or to warn of
the danger.18 For example, a local authority obliterated road markings when the
road was resurfaced but then failed to repaint the road or to warn of the resulting
dangerous situation; the council was held liable for injuries su¬ered by a motor-
ist as a result.19 Again, for example, a motor manufacturer could probably be held

18 E.g. Arnold v Teno (1978) 83 DLR (3d) 609.
19 Bird v Pearce [1979] RTR 369.
78 Chapter 3

liable for failure to recall vehicles discovered to su¬er from a dangerous defect (this
is frequently done voluntarily).
Even contractual undertakings are relevant to the problem of liability for non-
feasance in tort, because the person injured may not be the other contracting party.20
For example: A contracts with B that A will clear the snow o¬ B™s doorstep during
winter snowfalls. One day A fails to do this and C, a visitor to B™s house, slips on the
snow and is injured. Is A liable to C in tort for failing to take reasonable care? A has
not created the danger, and although A was under a duty towards B to remove it, can
C rely on that duty? In this kind of case there is much to be said for imposing lia-
bility, and a court would probably do so.21 Because A owes a contractual obligation
(to B) to clear the snow, there is less reason for reluctance to impose tort liability on
A in favour of C. Furthermore, in many cases of this kind the undertaking leads
other people to rely on it, thus creating dangers which would not have arisen
without it. For example, a person takes a car to a garage to have the brakes repaired.
The garage omits to do so but the owner (reasonably) thinks that the repair has been
done. Here the car owner is induced by reliance on the garage to continue driving
the car, thereby creating dangers to third parties on the road. By their undertaking
and failure to carry it out, the garage has made a positive contribution to the danger. Duties of physical protection
There are various situations in which the law is prepared to impose duties, to
prevent harm, on people who are in a particularly good position to protect or
rescue others from physical dangers and who, it might be thought, should o¬er
such protection because of their relationship with the person in danger. So, for
example, employers owe their employees legal duties to provide safe tools, a safe
workplace and safe working systems; and the basis of such duties is that the indi-
vidual employee typically has little control over working conditions. Again, doctors
and hospitals may be held liable for failure to provide treatment,22 or for failure to
warn patients of risks associated with particular treatments, or for failure to protect
a known ˜psychiatric and suicide risk™ who seriously injures himself by jumping out
of a window.23 More generally, professionals, by virtue of their status as profes-
sionals, owe duties to their clients and, to a lesser extent, to third parties, to take
positive steps, in the exercise of their professional skills, to protect them from
injury, loss or damage and to warn them of impending danger.

20 Also because, in general, clients can sue their professional advisers in either contract or tort:
Henderson v Merrett Syndicates Ltd [1995] 2 AC 145. Not all contractual undertakings are action-
able in tort, but a contractual undertaking (whether express or implied) to take all reasonable care
and (positively) to do everything reasonably necessary to protect the client from injury, damage
or loss, certainly is.
21 But not if the loss su¬ered was purely economic. For other examples see A.J.E. Ja¬ey, The Duty of
Care (Aldershot, 1992), 52“3.
22 Barnett v Chelsea and Kensington Hospital Management Committee [1969] 1 QB 428 (but the claim
failed on the issue of causation).
23 See Selfe v Ilford & District Hospital Management Committee, The Times, 26 November 1970.
The scope of the tort of negligence 79

Adults in charge of children may be required to take positive steps to protect or
help them; and prison authorities owe a duty of protection to prisoners.24 Similarly,
a person could probably be held liable for failing to call medical assistance for an
occupant or guest in their house who became helpless through disease or accident.
There seems little doubt that occupiers of business premises, such as hotels, restau-
rants, shops and even o¬ces, would be held liable if they failed to take reasonable
steps to summon medical help in an emergency arising from sudden illness to a
visitor to the premises. In Canada it has been held that a person in charge of a vessel
owes a duty to assist a person who falls overboard;25 that a hotel proprietor may owe
a duty of reasonable protection to an intoxicated person turned out of the hotel and
who is subsequently run down by a car;26 and that vehicle owners owe a duty of
protection to unlicensed and uninstructed persons whom they allow to use their
vehicles. In another case the Supreme Court of Canada held that a ski resort oper-
ator owed a duty to take reasonable steps to discourage an intoxicated patron from
taking part in a dangerous competition run by the operator.27
On the other hand, English courts have held that a taxi driver owes no duty of
care to an intoxicated passenger once the passenger has left the taxi;28 and that the
¬re service and other emergency services are not liable simply for failing to prevent
injury or damage, even with negligence.29 Liability will arise only if they positively
make matters worse. In one case, a local authority was sued for failure to exercise
its power to make a road safer by removing an obstruction to sight.30 The principle
underlying the decision that the authority was not liable is that public authorities
can only be held liable in tort for failing to perform their public functions, in such
a way as to prevent harm occurring, if their conduct was so unreasonable that no
reasonable public body could have considered it appropriate.
The basic issue that arises in all of these cases and situations is the extent to which
people should take responsibility for their own safety rather than expecting others
to protect or rescue them from physical danger “ or, at least, rather than expecting
to be able to recover damages for harm resulting from another™s failure to protect
or rescue. To say that the law is sometimes justi¬ed in imposing duties of physical
protection is probably banal and uncontroversial. Moreover, in imposing duties of

24 Kirkham v. Chief Constable of Manchester [1990] 2 QB 283. See also Metropolitan Police
Commissioner v Reeves [2000] 1 AC 360.
25 Horsley v MacLaren (The Ogopogo) [1972] 2 Lloyd™s Reports 210.
26 For the position in Australia see Cole v South Tweed Heads Rugby Football Club Ltd (2004) 217
CLR 469 (commercial supplier of alcohol owes no duty to intoxicated customer). Concerning
the liability of hosts to persons injured by intoxicated guests see J. Horder, ˜Tort and the Road
to Temperance: A Di¬erent Kind of O¬ensive against the Drinking Driver™ (1988) 51 Modern
LR 735.
27 Crocker v Sundance Northwest Resorts Ltd [1988] 1 SCR 1186 (in which the two previous exam-
ples also are cited).
28 Gri¬ths v Brown, The Times, 23 October 1998.
29 Capital & Counties Plc v Hampshire CC [1997] QB 1004; OLL Ltd v Secretary of State for Transport
[1997] 3 All ER 897; but contrast Kent v Gri¬ths [2001] QB 36 (ambulance service).
30 Stovin v Wise [1996] AC 923.
80 Chapter 3

protection, the law seems to re¬‚ect commonly held ethical views by taking account
of the nature and closeness of the relationship between the parties, and of the iden-
tity of the person who created the dangerous situation “ the claimant, the defendant
or a third party. The di¬cult and contentious issue is where the line should be drawn
between protection of self and protection of others. For example, consider a case in
which a naval airman drank large amounts of alcohol and, as a result, choked to
death on his own vomit. While his employer, the Ministry of Defence, was held not
liable for allowing him to get himself drunk, it was held liable for failing to look after
him even though his colleagues knew that he was incapable of taking care of
himself.31 This decision has been described as ˜o¬ensive to normal ideas of
justice™32 “ although the judges who held in the claimant™s favour would obviously
not accept this description of their decision.
There are two important points to bear in mind here. First, while courts may
sometimes refuse to impose liability for failure to prevent harm on the basis of a
rather abstract statement to the e¬ect that ˜the defendant did not owe the claimant
a duty of protection™, decisions whether or not to impose such liability are more
often based on a detailed consideration of whether, in the particular circum-
stances of the case, the defendant ought to have taken steps to protect the
claimant. The court will treat the issue in the case as being whether, on the facts,
the defendant was at fault and acted negligently, rather than whether, in the
abstract, the defendant owed the claimant a duty of protection. So, for instance,
disagreement about the case of the airman, considered in the previous paragraph,
centred not on whether, in the abstract, an employer owes a duty of protection to
its employees “ clearly it should and does. The real issue was whether, in the cir-
cumstances of the case, the employer acted negligently in not protecting the
employee “ in lawyer™s jargon, whether the employer breached its duty of care to
protect the employee. The second point to note is that even if a court holds that
the defendant owed the claimant a duty of protection, and negligently failed to
protect the claimant from harm, that may not be the end of the matter. For
instance, in the case of the airman, the court held that although the defendant
ought to have taken steps to protect the claimant from his own fecklessness, the
airmen was also partly responsible for what happened. In fact, the court appor-
tioned the responsibility two-thirds to the claimant and only one-third to the
defendant. Both of these points illustrate the distinction between the issue of duty
of care and the issue of fault (considered in ch. 2). Duty of care is concerned with
whether, in principle, the defendant can be held liable: the scope of liability for
negligence, as it was put earlier. This leaves open the issue of fault “ whether and
to what extent a defendant who owes a duty of care breached that duty and ought
to be held responsible (˜at fault™) for what happened.

31 Barrett v Ministry of Defence [1995] 1 WLR 1217. See also Jebson v Ministry of Defence [2000]
1 WLR 2055.
32 P.S. Atiyah, The Damages Lottery (Oxford, 1997), 40“1.
The scope of the tort of negligence 81

Not only may people disagree about when the law should impose duties of phys-
ical protection, but views about this may also change over time. A good illustration
is provided by cases where a person su¬ers injury as a result of diving into shallow
water and claims that the defendant ought to have erected a sign warning of the
danger. In 1993 the High Court of Australia imposed liability in such a case,33 and
the decision can be seen as part of a general trend in the twentieth century of expan-
sion of the scope of tort liability for personal injury. In the last few years, however,
a reaction has set in, and increasing emphasis is now being put in self-reliance and
˜personal responsibility™. This change of attitude is re¬‚ected in decisions of the
courts in personal injury cases.34 In a recent English case of a young man rendered
tetraplegic as a result of diving into shallow water, the House of Lords, by majority,
refused to impose liability on the defendant for failing to take steps that would have
prevented the tragedy.35 In a section of his judgment headed ˜Free will™ Lord
Ho¬mann stressed that the claimant had acted ˜freely and voluntarily™, and that the
law should not expect occupiers of land ˜paternalistically™ to prevent visitors from
undertaking ˜inherently risky activities™ on their land in order to protect them from
harming themselves. It is worth noting, however, that the Court of Appeal had
decided the case in the claimant™s favour. Duties to control the conduct of others
A person who has the power to control another may be liable for failure to exercise
it. So, for example, parents and school authorities are under a duty to control young
children, and prison authorities are under a duty to control inmates.36 If a child, for
instance, is given a gun by a third party, and is known by the parent to have a gun,
the parent becomes responsible for seeing that the child is old enough and sensible
enough to be allowed to have the gun; for instructing the child in how to use it
safely, and so on. If the parent does nothing at all, and if he or she is shown to have
been negligent to do nothing, the parent will be liable for injuries caused by the
child.37 Similarly, an employer is under a duty to control employees, although this
is not of much practical importance because an employer is vicariously liable for
the negligence of employees whether the employer was negligent or not; but there
are a few cases in which the employer may be liable for personal nonfeasance
though not vicariously liable because, for instance, the servant was not acting in the
course of employment.38
A person who sees someone being beaten up in the street may walk on without
assisting, but a hotel proprietor who saw someone in danger of being attacked by

33 Nagle v Rottnest Island Authority (1993) 177 CLR 423.
34 E.g. Wyong Shire Council v Vairy [2004] NSWCA 247; a¬rmed Vairy v Wyong Shire Council [2005]
HCA 62. But contrast Mulligan v Co¬s Harbour City Council [2005] HCA 63.
35 Tomlinson v Congleton BC [2004] 1 AC 46.
36 Home O¬ce v Dorset Yacht Co. Ltd [1970] AC 1004.
37 Newton v Edgerley [1959] 1 WLR 1031.
38 Hudson v Ridge Manufacturing Co. [1957] 2 QB 348.
82 Chapter 3

a guest in the hotel might well be liable for failing to take some reasonable steps to
control the attacker. Those in charge of public transport vehicles or vessels would
also have some duty to control passengers. The same might also apply to private
vehicles; for example, a car driver might be held liable to an injured cyclist if the
driver sat and watched a passenger negligently opening the o¬side door of the car
in the way of approaching tra¬c, at least if the driver was in a position to stop the
door being opened.39
On the other hand, the police cannot normally be held liable for failure to
prevent crime.40 More generally, regulatory bodies whose function it is to monitor
and control potentially dangerous activities are typically not liable for failure to
exercise their regulatory powers. The cases which ¬rst established this principle
dealt with ¬nancial loss “ where, for example, authorities responsible for regulat-
ing banks are sued by depositors and investors who lose their money when the bank
collapses.41 But the principle has also been applied where, for instance, social
workers fail to prevent child abuse;42 and it has even been hinted that a local author-
ity might not be held liable if occupants of a house were injured as a result of bad
and illegal construction work which the authority negligently failed to detect when
it inspected the house for compliance with building regulations.43 On the other
hand, liability may be imposed where, because of personal dealings between the
defendant and the claimant, it would be unreasonable for the defendant not to act
for the claimant™s bene¬t. But the mere fact that a public body has powers to control
the conduct of others, which it could exercise for the bene¬t of members of the
public, does not open it to tort liability for negligent failure to do so. One reason
for this approach is that the courts do not want public funds to be used to com-
pensate individuals when they could be used for the bene¬t of the public as a whole
or sections of it. Another is a fear that if public authorities are vulnerable to tort lia-
bility, they may be led to act ˜defensively™ in performing their functions; that is, in a
way designed to avoid potential tort liability regardless of whether such action is in
the wider public interest.
It should be noted, however, that in these decisions, English courts demonstrated
a much greater unwillingness to impose liability than courts in other major common
law jurisdictions, such as Australia, Canada and New Zealand, have displayed. It
should also be recalled (as noted in 3.1) that the ˜no-duty™ technique used in these
cases was held by the ECtHR to be incompatible with Article 6 of the ECHR.
Subsequently, the House of Lords signalled that it would modify its approach to
take more account of the facts of individual cases. Increased willingness to impose

39 See e.g. Brown v Roberts [1965] 1 QB 1 where the claim was rejected on the ground that the owner
of the car was not in fact negligent in failing to prevent the passenger opening the door.
40 Hill v Chief Constable of West Yorkshire [1989] AC 53.
41 Yuen Kun Yeu v Attorney-General of Hong Kong [1988] AC 175.
42 X v Bedfordshire County Council [1995] 2 AC 633; but see now Barrett v En¬eld LBC [2001]
2 AC 550.
43 Murphy v Brentwood District Council [1991] 1 AC 398, 463 (Lord Keith of Kinkel).
The scope of the tort of negligence 83

liability in such cases can, perhaps, also be detected. For its part, the European Court
has drawn back somewhat from its disapproval of the duty-of-care technique by
holding that it does not infringe Article 6. However, it has also held that the technique
may be incompatible with Article 13 of the ECHR. It remains unclear what impact
this jurisprudence of the ECtHR will have on the use of the duty-of-care technique
negatively to limit liability for negligence rather than positively, to provide a frame-
work or justi¬cation for the imposition of liability. Control over property
Another ground on which liability for nonfeasance may be imposed is that the defen-
dant was in control of some property from which, or by means of which, the damage
was done. For example, the duties of occupiers of land to their lawful visitors44
require them to take positive steps to ensure that visitors are safe either by removing
dangers from the premises or by warning of such dangers.45 Another example of this
kind of liability is to be found in the case of Goldman v Hargrave 46 in which a tree on
the defendant™s land was struck by lightning and caught ¬re. The defendant took
some, but (it was found) negligent and ine¬ectual, steps to put the ¬re out, and it
spread to the claimant™s property, causing damage there. The (successful) complaint
against the defendant was simply that he had failed to take reasonable steps to put
out the ¬re. He had not started it, nor even created any conditions on his land which
could be said to have contributed to the risk of the ¬re: it was simply a natural
hazard.47 The argument in favour of such liability is that the ownership of land
should entail responsibilities as well as rights.
An important argument against such liability, however, is that if people can be
held liable for careless attempts to avert danger, they might be discouraged from
helping in the ¬rst place. The law is keen not to discourage altruism and so, for
example, it allows rescuers to recover compensation for injury or loss su¬ered in
the process of rendering assistance from any person negligently responsible for cre-
ating the dangerous situation which prompted the ˜rescuer™ to act.48 In Goldman v
Hargrave the Privy Council countered this argument by saying that the defendant,
as a landowner, would have been liable even if he had done nothing, provided
failure to do anything at all would have been negligent in the circumstances. But
suppose a person decides to render assistance to the victim of a road accident and
unfortunately does so negligently, thus making matters worse. In the absence of
special circumstances, if the helper had done nothing there would, probably, have

44 It is unclear to what extent the duty of positive action extends to unlawful visitors.
45 This may include a duty to control other visitors to the premises: Hosie v Arbroath Football Club
Ltd 1987 SLT 122.
46 [1967] 1 AC 645. See also Leakey v National Trust [1980] QB 485 (Trust held liable when a natural
mound of earth subsided on to neighbouring land even though the neighbour had been given
permission to enter the land and shore up the mound).
47 Related are cases in which damage is caused to a neighbour by the activities of a third party on the
defendant™s land. For a discussion see Ja¬ey, The Duty of Care, 72“6.
48 E.g. Chadwick v. British Railways Board [1967] 1 WLR 912.
84 Chapter 3

been no liability. Should the helper, then, be open to liability for careless interven-
tion? Perhaps liability should be possible if it was clearly unreasonable of the helper
to render aid personally rather than, for example, to call a doctor or ambulance.
Once we have reached this point, however, there seems little reason why we should
not go one step further and allow of the possibility of liability for failure to do any-
thing at all, at least if the burden of doing something to help would not be very
great. In other words, provided the standard of care is pitched low enough, there
may seem little objection to imposing a duty to rescue.
In an attempt to meet this last point, the Privy Council held, contrary to the
normal rule (namely that the personal capabilities of the defendant are irrelevant),
that in deciding what were ˜reasonable™ steps which an occupier of land must take to
prevent the land being a source of danger to others, account had to be taken of the
resources and capacities of the particular occupier: what is reasonable for the indi-
vidual landowner may not be reasonable for the large company. This may be a sound
approach in some cases “ our road accident example, for instance. But in the great
majority of cases, the cost facing an occupier of land is not really the cost of taking
precautions against the land being a danger to neighbours, but the cost of insuring
against this; and it is perfectly reasonable to regard this insurance cost as an essen-
tial part of the cost of using land which the landowner must pay, or vacate the land.
For instance, it may be doubted whether a poor cricket club which fails to erect a
fence around its ground to protect passers-by from being hit by stray cricket balls
should be any better o¬ than a wealthy cricket club. The real e¬ect of imposing lia-
bility in such a case would not be to force clubs to build fences but to force them to
buy insurance; and a club which could not a¬ord the insurance premium should not
be in any di¬erent position from a club which could not a¬ord to pay rates.
In the particular circumstances of Goldman v Hargrave, the Privy Council may
have been wise in restricting the extent of liability. It is very doubtful whether any
good purpose is served by extending liability in tort for damage caused by ¬re, at
all events in this country; the position in Australia may be a¬ected by special con-
siderations, such as the great danger of bush ¬res and the importance of providing
every inducement to landowners to take precautions against ¬re. But the technique
chosen to restrict liability “ that is by personalizing the standard of care “ is not
desirable. In the wider context of liability for personal injury arising from the use
of land it would be even more unsatisfactory to use a personalized test of standard
of care. Admittedly, the personalized standard of care for nonfeasance may make it
more likely that liability for nonfeasance will be expanded in the future. But not
only is the distinction thus drawn between misfeasance and nonfeasance of doubt-
ful value; also, to consider the wealth of the defendant without regard to whether
they are or ought to be insured, is di¬cult to justify.

3.3 Nervous shock
There is one other limitation on the scope of liability for negligence which is dealt
with in terms of duty of care principles and which requires discussion here. This
The scope of the tort of negligence 85

concerns liability for what has traditionally been called ˜nervous shock™. This term
is often now objected to as having no obvious meaning, and phrases such as ˜mental
injury™ or ˜psychiatric damage™ are often put in its place. But these do not capture
the full range of situations covered by the older label, and so it is used here as an
umbrella term. Nervous shock is injury caused by the impact on the mind, through
the senses, of external events. Injury caused by the impact on the mind of external
events, which is recognized by law, is of three types: physical injury “ a pregnant
woman may su¬er a miscarriage or a person may su¬er a heart attack or a stroke;
psychological injury such as hysteria, neurosis, depression or any other recogni-
zed psychiatric illness; and psychosomatic e¬ects of psychiatric illnesses, such as
The history of the law concerning tort liability for nervous shock in the twenti-
eth century was one of gradual expansion of the grounds of recovery as both
knowledge of the brain and the mind, and sympathy for those a¬„icted by mental
disturbance, have increased. There is still, however, a bias in the law against allow-
ing recovery for nervous shock. Several arguments have traditionally been put
forward to justify this approach. One is that mental injury that has no bodily symp-
toms, or only psychosomatic symptoms, is relatively di¬cult to prove; and, more-
over, people vary more widely in their susceptibility to mental upset than in their
susceptibility to physical injury. The law attempts to deal with this problem by
being prepared to compensate for mental injury that is not accompanied by bodily
injury to its su¬erer only if it amounts to some ˜recognizable psychiatric illness™.49
Thus, expert medical evidence will normally be necessary to establish that a person
has su¬ered nervous shock in this sense. Mere grief, anguish, fear, unhappiness,
humiliation, outrage and so on, however distressing they may be, can (with one
exception)50 attract compensation only if they are the result of bodily injury to the
person su¬ering any of these feelings; or, perhaps, of damage to their property51 or
to ¬nancial interests which are protected by law.
Even though a judge may, with the aid of expert evidence, not have too much
di¬culty in distinguishing a real psychiatric illness from less serious mental
disturbance, this may not justify drawing a sharp line between the two, because at
least some psychiatric disorders (e.g. depression) are just extreme versions of
commonplace emotional states. Moreover, it is di¬cult to assess how big a
problem the need to draw this line creates in cases which do not go to court, but
are settled out of court simply on the basis of written medical reports and with-
out the bene¬t of cross-examination of expert witnesses. It may well be that in
practice, much more will turn, in settled cases, on the e¬ect of the symptoms
on the claimant™s lifestyle (e.g. are they con¬ned to bed, unable to work, and so
on), rather than on whether the symptoms amount to a recognized psychiatric

49 See Hinz v Berry [1970] 2 QB 40.
50 Damages for bereavement: 3.4.
51 Attia v British Gas [1988] QB 304.
86 Chapter 3

illness,52 especially in cases where the claim is accompanied by claims for bodily
injury by other members of the claimant™s family. Furthermore, it is probably the
case in practice that damages for mental disturbance are most often paid to a
family member who su¬ers it as a result of injuries to other family members, and
that such damages are only relatively rarely paid to persons who are not related to
the physically injured person. This is signi¬cant because it is well recognized that
the closer the relationship between the injured person and the person who su¬ers
mental disturbance, the more serious the mental disturbance is likely to be.
Another reason for the restrictive approach to mental distress is the so-called
˜¬‚oodgates argument™: if recovery for mental distress (like recovery for bodily injury)
were allowed simply on the basis that it was foreseeable, there might well be a ¬‚ood
of claims which would clog up the court system and divert too many of society™s
resources into compensating the victims of nervous shock at the expense of the many
who presently receive little or no compensation even for physical injuries su¬ered as
a result of negligent conduct. The force of the ¬‚oodgates argument is disputed by
judges and commentators even in cases where it is relevant to what happened. On the
other hand, given the large number of serious accidents each year, and the fact that
a person may su¬er mental distress even if they are in no personal physical danger, it
might be expected that many people would su¬er some sort of mental distress as a
result of witnessing harrowing events. But it must also be remembered that the
narrow de¬nition of nervous shock would probably rule out very many, if not most,
of such cases.
Having noted the traditional bias in the law against recovery for nervous shock,
it must now be said that the law divides victims of nervous shock into two groups “
primary victims and secondary victims; and that the bias against recovery for
nervous shock now really only applies to secondary victims. For instance, a person
who su¬ers physical injuries as a result of another™s tort (say, in a car accident) may
also recover damages for nervous shock resulting from their physical injuries.
Moreover, they may, in addition, recover damages for other ˜lesser™ forms of mental
distress such as pain and su¬ering, awareness of a shortened expectation of life, dis-
comfort and inconvenience arising from con¬nement to bed or hospital or wheel-
chair. The extent to which damages for mental injuries may be awarded where
physical injury has also been su¬ered has never been treated as raising a problem
involving the duty of care, but merely as involving a problem in the assessment of
damages; and this is dealt with fully in chapter 6. Another type of primary victim
is a person who su¬ers nervous shock as a result of being tortiously exposed to
a risk of physical injury but who actually su¬ers no physical injury “ for instance,
a passenger in a car which is involved in a road accident who escapes physically

52 In 1991 more than £1 million was paid by insurers under the terms of an out-of-court settlement
to ¬fty-nine families of children who had, it was alleged, been negligently and falsely diagnosed
as having been subjected to sexual abuse. The amounts awarded apparently included sums for
mental distress short of nervous shock. On the current state of the law, it is by no means clear that
the employers of the allegedly negligent doctors were liable for such mental injury.
The scope of the tort of negligence 87

unscathed. Such a person may recover damages for their nervous shock (but not
for lesser forms of mental distress) even if it was an abnormal or extreme reaction
to what happened, simply because there was a risk that they would su¬er physical
injury.53 A variant of this type of case is where people su¬er ˜fear for the future™. In
1998, damages were awarded to people who, as a result of having been treated with
human growth hormone as children, su¬ered ˜deteriorating psychiatric health™ as a
result of ˜rational fears™ of one day succumbing to a ghastly lingering death from
CJD.54 A third type of primary victim is a person who, for instance, su¬ers psychi-
atric injury as a result of being exposed to excessively stressful or dangerous
working conditions by their employer. Provided their mental injury was not an
abnormal or extreme reaction to the situation they were in, they may recover
damages for it.55
It is in this area, perhaps, that there is greatest pressure to expand the boundaries
of tort liability by recognising new types of mental harm as appropriate subjects for
compensation. Concepts of illness (especially mental illness) are, to some extent at
least, socially constructed; and although the law uses the category of ˜(medically)
recognisable psychiatric illness™ as a device to control the expansion of liability, the
courts themselves can play a part in causing particular sets of symptoms to be char-
acterised in this way. For this reason, judicial activity in this area is a prime target
for those who think that tort law has gone too far in protecting people from life™s
The typical secondary victim of nervous shock is a person who witnesses an
accident in which someone known to them is killed or injured, and then sues the
person responsible for the injuries or death.56 Liability for nervous shock su¬ered
by secondary victims is hedged about with limitations. First, the secondary vic-
tim™s mental injury must have been the result of su¬ering a ˜shock™ in the colloquial
sense.57 For instance, a secondary victim could not sue in respect of psychiatric
illness resulting from having cared over a long period of time for an injured per-
son. This limitation seems to be an illogical result of calling the injury in such cases
˜nervous shock™. It is probably one of the factors that led to judicial recognition of the

53 Page v Smith [1996] AC 155. The reasoning was that in this day and age, no distinction should be
drawn between bodily and mental injury. So if a person could have recovered damages if they had
su¬ered bodily injury as a result of their exposure to risk, they should be allowed to recover for
nervous shock su¬ered as a result of exposure to the same risk. A person who did not face a risk
of physical injury but reasonably feared that they did could recover for resulting nervous shock
provided it was not an unusual or extreme reaction.
54 For the legal justi¬cation for these awards see Creutzfeldt-Jakob Disease Litigation; Group B
Plainti¬s v Medical Research Council (1997) 41 BMLR 157.
55 Hatton v Sutherland [2002] 2 All ER 1. Strictly, this type of victim may not be ˜primary™ because
it was held in Frost v Chief Constable of South Yorkshire [1999] 2 AC 455 that by de¬nition,
a primary victim is one who was subjected to a risk of physical injury.
56 Recovery may also be allowed where, as a result of another™s tort, a person su¬ers shock arising
from reasonable fear for the safety of a third party, even if the fear was, in fact, groundless.
57 What was called a ˜sudden sensory perception™ by Brennan J in Jaensch v Co¬ey (1983“4) 155 CLR
549. This limitation was expressly adopted by Lord Ackner in Alcock v Chief Constable of South
Yorkshire [1992] 1 AC 310.
88 Chapter 3

condition called ˜post-traumatic stress disorder™ (PTSD).58 Secondly, the secondary
victim™s mental condition must not have been an abnormal or extreme reaction to
the incidents in question. Liability will arise only if a person of ˜reasonable fortitude™
would have su¬ered shock. This limitation is based on the perception that di¬erent
people™s susceptibility to mental injury varies much more than their susceptibility
to bodily injury. Thirdly, the secondary victim must (as a general rule) have been in
a relationship of ˜love and a¬ection™ with the person injured or killed. Bizarrely, it
has been held that some relationships (such as parent and child) are assumed to be
relationships of love and a¬ection, but that others (such as aunt and nephew) will
only qualify as such if the claimant can prove that there were close ties of love and
a¬ection between them and the person injured or killed. How can we justify a rule
that requires mentally traumatized people to go to court and prove that they have
strong feelings of love and a¬ection towards another? For many years, it was thought
that there was an exception to this rule that allowed rescuers to recover for nervous
shock even if they were not in a relationship of love and a¬ection with the victims.59
However, it now seems that a ˜rescuer™ who su¬ers nervous shock will recover only
if they were subjected to a risk of physical injury (i.e. only if they were a ˜primary
Fourthly, a ¬rm line was traditionally drawn between secondary victims who
su¬er shock merely as a result of being told of events and those who actually witness
the events or their aftermath; the former were allowed to recover. Leaving aside the
question of whether this distinction has any sound scienti¬c basis, the advent of
simultaneous broadcasting of sporting and other events has put severe strain on the
law. In the Hillsborough stadium case, some of the claimants claimed damages for
shock su¬ered as a result of seeing the terrible events on television. The House of
Lords held that the television pictures in this case were not su¬ciently equivalent
to being in the stadium itself to warrant recovery, although the judges did not rule
out that a media broadcast might be detailed and graphic enough to give rise to a
claim. There is no precise de¬nition of ˜aftermath™, but it is said to require a fairly
high degree of temporal and physical proximity to the incident. The longer the gap
of time between the accident and the witnessing of its consequences, the less likely
is it that recovery for nervous shock will be allowed. For example, one judge in the
Hillsborough football stadium case61 said that shock su¬ered as a result of seeing
the corpse of a dead relative in a morgue 8 hours after the accident which caused
the death would not attract compensation.

58 For an illuminating discussion of this condition see I. Freckleton,˜Post-Traumatic Stress Disorder:
A Challenge for Public and Private Health Law™ (1985) 5 J. of Law and Medicine 252.
59 Chadwick v British Railways Board [1967] 1 WLR 912. Other exceptions have been allowed in cases
where the secondary victim is a fellow worker of the person injured or killed (Dooley v Cammell
Laird [1951] 1 Lloyd™s Reports 271; Mt Isa Mines Ltd v Pusey (1970) 125 CLR 383) provided they
were involved as an actor in the dangerous incident: Hunter v British Coal Corporation [1998] 2
All ER 97.
60 Frost v Chief Constable of South Yorkshire [1999] 2 AC 455.
61 Lord Ackner in Alcock v Chief Constable of South Yorkshire [1992] 1 AC 310, 405.
The scope of the tort of negligence 89

These limitations on liability to secondary victims of nervous shock (and,
indeed, the distinction between primary and secondary victims of shock) involve
the drawing of gruesome, invidious and often di¬cult distinctions that do the law
no credit. In fact, many people ¬nd the legal regime in this area unsatisfactory and
even repugnant. In the late 1990s the Law Commission reviewed the law relating to
nervous shock.62 It made many recommendations, but three stand out as being
important. One was that the shock requirement should be abolished.63 Another was
that the distinction between primary and secondary victims should be abandoned.
This distinction has been much criticised; but the problem with the Commission™s
proposal is that their third main recommendation assumes the continued existence
of a distinction between people who su¬er nervous shock as a result of another
being killed, injured or imperilled, and people who su¬er nervous shock in other
circumstances. This is essentially a distinction between primary and secondary
victims. The Commission™s third main recommendation is for the abolition of the
requirement that a secondary victim of nervous shock may recover damages only
if they witnessed the accident or its aftermath personally; and that the only require-
ment should be that there were close ties of love and a¬ection between the victim
of nervous shock and the person killed, injured or imperilled. The Commission
further proposed that there should be a ¬xed list of relationships which are deemed
to involve close ties of love and a¬ection, but that people in relationships not on
this list should be allowed to prove that they did nevertheless have close ties of love
and a¬ection to the person killed, injured or imperilled. While the Commission™s
proposals would, if adopted, rid the law of some of its more objectionable and
complex features, this third recommendation in particular retains some of the
more unedifying aspects of the law in this area.

3.4 Family claims
As noted above (3.3 n. 50), there is one exception to the rule that mere grief, anguish
or unhappiness cannot attract compensation unless it is a consequence of some
other actionable injury. This exception is usually treated as part of the law concern-
ing assessment of damages rather than duty of care,64 but for ease of comparison, it
will be dealt with here. Under the Fatal Accidents Act 197665 an award of a ¬xed sum
of £10,000 (called damages for bereavement) may be made to a husband or wife or
a civil partner66 in respect of the death of his or her spouse or civil partner, or to

62 Law Com No. 249, Liability for Psychiatric Illness (1998).
63 The High Court of Australia has abandoned this requirement: Tame v. New South Wales (2002)
211 CLR 317. The court has not adopted the distinction between primary and secondary victims.
Australian law on this topic has been contrasted with ˜the possibly over-re¬ned state™ of English
law: R.P. Balkin and J.L.R. Davis, Law of Torts, 3rd edn (Sydney, 2004), 253.
64 Partly because the death need not have been caused by negligence but can have been caused by
any ˜wrongful act, neglect or default™. But negligence is the most common trigger of claims under
the Act.
65 Section 1A inserted by the Administration of Justice Act 1982.
66 As a result of the Civil Partnership Act 2004.
90 Chapter 3

parents in respect of the death of an unmarried minor child. These damages are
meant as ˜solace™ (or ˜solatium™) for the grief caused by the spouse™s or child™s death
(so they cannot be recovered by the estate of a deceased spouse or parent). This head
of damages also constitutes an exception to the principle laid down by the courts
that damages under the Fatal Accidents Act 1976 (which, loosely, allows members of
the close family of a deceased person to recover damages in respect of that person™s
death, if it was wrongfully caused) are meant to compensate only for ¬nancial loss
“ basically, loss of ¬nancial support formerly provided by the deceased. Damages for
bereavement were designed to replace damages for loss of expectation of life
(awarded to the deceased™s estate), which were abolished in 1982.67 There is a similar
provision for damages for ˜loss of society™ in Scotland under the Damages (Scotland)
Act 1976 but, unlike in England, there is no statutory limit to the award, and the class
of eligible claimants is de¬ned more widely.
Naturally the death of a close relative in an accident must give rise to sympathy
for the survivors, but damages for bereavement are nevertheless highly objection-
able. There are two main objections to all awards by way of solatium. One is that
the motives of relatives in seeking such awards may be questionable. Much more
importantly, it seems arbitrary to select the death of a close relative as the criterion
for paying what is still to many people a substantial sum of money. It must be
remembered that the relatives of a person who is very severely injured (but not
killed) in an accident may well su¬er much greater mental su¬ering than the rela-
tives of someone who is killed. For one thing, the su¬ering is continuous and may
be prolonged in such cases for many years. Even if the victim™s injuries were the
result of negligent conduct, the su¬ering of the relatives would not be recoverable
as nervous shock either. It does not seem right that, when nothing is awarded in
such a case, damages should nevertheless be awarded for the death of a child.
In addition, the fact that the sum to be awarded is ¬xed by statute means that the
same sum would be awarded in a very wide variety of situations, for example, to a
mother for the death of a newly born child; to parents of an older child irrespective
of whether the child was a comfort or a trial to its parents; and to a spouse or civil
partner irrespective of the age, state of health of the other spouse or civil partner,
and regardless of whether the spouses or civil partners were the best of friends or
had been separated for years and were not on speaking terms. Furthermore,
damages for bereavement, unlike damages for ¬nancial loss resulting from
another™s death, are only recoverable by a spouse to a legal marriage or a party to a
civil partnership, and not by parties who cohabit without marrying or entering a
civil partnership. Apart from all this, there is a further fundamental point that
damages by way of solatium ought to be a very low priority in any legal system
which still denies adequate compensation for loss of income to so many of those
injured in accidents or crippled by disabling illness.

67 By s. 1 of the Administration of Justice Act 1982. For the history of these provisions see the fourth
edition of this book, 76“7.
The scope of the tort of negligence 91

Besides creating an exception to the rule about liability for mental harm, the
Fatal Accidents Act 1976 (the original version of which, known as Lord Campbell™s
Act, was enacted in 1846) also creates an exception to a basic principle of tort law
that damages may not be recovered for ¬nancial loss arising out of harm to another
person or another person™s property. Because this book is primarily concerned with
compensation for personal injury and death, we will not discuss this basic princi-
ple in any detail. The point to make is that in order to prevent the principle being
swamped by the exception, the Fatal Accidents Act 1976 has always contained a list
of classes of persons who are entitled to make a claim under the Act. The current
list of eligible claimants covers a wide range of de iure and de facto family relation-
ships. In 1999 the Law Commission recommended that the Act be amended to
allow any person who was formerly being ˜maintained™ by the deceased to make a
claim, regardless of the nature of their relationship with the deceased; but this rec-
ommendation has not been acted upon.

Departures from the fault principle

4.1 Fault liability and strict liability

The fault principle, as embodied in the concept of negligence, is not the only basis
of legal liability for personal injuries and death, although it is, in practice at least,
by far the most important. In this chapter we will consider modi¬cations to and
departures from the fault principle. Such modi¬cations and departures are often
said to impose ˜strict liability™ as opposed to fault liability. Whereas fault liability
is based on a judgment that a person should have behaved di¬erently (for
instance, by taking certain precautions), strict liability does not involve any judg-
ment that the person should have behaved di¬erently. Putting the same point
another way, fault liability is liability for the way a person behaved whereas strict
liability is liability for consequences of a person™s conduct. Strict liability has often
been thought to be morally unjusti¬able, even if it has its uses as a legal device “
how can it be fair to hold someone liable for the consequences of behaving in a
perfectly acceptable way? How can we justify responsibility in the absence of cul-
pability? The best answer to this question appears to be that even in morality (as
opposed to law) we sometimes accept responsibility and hold others responsible
for things that were not our, or their, fault. For example, if a young child acciden-
tally breaks a neighbour™s window while playing ball, its parents might well feel
that they ought (morally) to accept responsibility for the broken window and pay
to have it replaced, even if they took all reasonable care in supervising their child.
Indeed, this example shows that morality might impose strict liability in situa-
tions where the law would not “ the parents would not be legally liable for the
damage done by the child in such a case. So it may be fair to hold someone liable
for the consequences of their conduct even if that conduct was not faulty.
However, just saying this does not tell us under what circumstances strict liabil-
ity can be fair. The phrase ˜liability without fault™ merely eliminates fault as a nec-
essary condition of liability; it does not put anything else in its place. Thus strict
liability is not one possible alternative to liability for fault, but a collection of such
alternatives. The phrase ˜liability for fault™ tells us that liability ought to be placed
on a faulty party (although it does not tell us on which, if there are more than

Departures from the fault principle 93

one). But the term ˜strict liability™ implies no criterion for deciding on whom
liability should rest.1
This point is often not appreciated because it is often taken for granted that strict
liability is based on the concept of ˜legal causation™.2 For example, it is often assumed
that if strict liability were extended to road accidents, a motorist would be held
liable under such a regime if, for example, he or she ˜caused™ an accident by collid-
ing with a pedestrian, even without fault. Similarly, a gas undertaking might be
thought to ˜cause™ accidents arising through leaks from their pipes, and strict lia-
bility would simply make the gas undertaking liable for such accidents. But few if
any existing forms of strict liability are based on legal causation, at least if we give
the word ˜cause™ its most common meaning. For example, a zoo-keeper whose lion
escapes despite all due care is strictly liable for the damage it does, but the zoo
owner would not be said to have ˜caused™ the death of someone killed by the lion.
What the zoo owner has done is to create a risky situation by keeping a lion in cap-
tivity. Again, as a general rule, employers are strictly (vicariously) liable for torts
committed by their employees that injure third parties, even though the employer
would not be said to have ˜caused™ the injuries. In any event, as we shall see later, the
notion of causation is a problematic basis for liability because there are consider-
able di¬culties in formulating principles of causation and in justifying legal liabil-
ity on the basis of such principles.
This is not to say that the concept of ˜cause™ may not in many cases identify the
party who, as a matter of sound policy, ought to be made liable “ as in the cases of
the road accident or the gas leak mentioned above. But a possible criterion of lia-
bility which would cover both of these cases and that of the zoo-keeper, would be
to ask which party could more easily bear and distribute the losses caused by the
accident, by insurance or other means. Clearly, for example, not only could the
driver more easily insure against the risk than the pedestrian, but also it is much
easier to enforce compulsory insurance against motorists than it would be against
pedestrians. This line of reasoning, however, has important practical implications.
It is true, for example, that a scheme imposing strict liability for road accidents
caused without fault3 would entitle more personal injury victims to claim com-
pensation, but only if there was another motorist involved who could be held
strictly liable. However, many road accidents involve only one driver, for example,
where a car veers o¬ the road and collides with a tree or perhaps another (station-
ary) car. Motorists as a group are just as able, via insurance, to bear and distri-
bute the costs of such accidents as those of accidents involving more than one
driver. It is arguments such as these that have led many reformers towards ˜no-fault™

1 It does not even tell us that liability ought to be placed on someone who is not at fault. Strict lia-
bility is not liability in the absence of fault but liability regardless of the presence or absence of fault.
2 For a detailed explanation of this notion see 5.2.
3 For a limited proposal of this nature see Compensation for the Injured (1971), a report of a com-
mittee of the Society of Conservative Lawyers, and Professor Atiyah™s criticisms in ˜Compensation
for the Injured™ (1971) 34 Modern LR 432.
94 Chapter 4

compensation schemes under which entitlement to compensation depends on
being injured and not on being able to ¬nd someone to sue.
This is not to say that the search for the best ˜risk bearer™ is the only possible
reason for extending strict liability. Another may be that placing the risk on one
person rather than another would reduce accidents. Full examination of all these
objectives is deferred to chapter 17 of this book. All that needs to be stressed at this
stage is that extending ˜strict liability™ is not a positive programme for reform on its
own. We also need to decide on whom strict liability is to be placed or, in other
words, what the criterion of liability will be, if it is not to be fault.

4.2 ˜Procedural™ devices
In cases where liability is based on fault, it is the injured party who normally ˜bears
the burden of proving™ that the injurer was at fault. This rule about burden of proof
is generally considered to be a corollary of the negative part of the fault principle,
namely no liability without (proof of) fault. One way of making it more di¬cult
for a person allegedly guilty of faulty conduct to escape liability for that conduct is
to require that person to prove that their conduct was not faulty. This device is
referred to as ˜shifting the burden of proof ™. It may be done directly. For example,
the EC Commission once issued a draft Directive on liability for services; under the
Directive the basis of liability was to be fault, but the burden of proof on the issue
of fault was to rest on the provider of the services and not on the person who claims
to have been caused injury or loss by the service-provider. The same e¬ect may be
achieved more indirectly by the application of a principle referred to by the Latin
tag ˜res ipsa loquitur™ (literally: ˜the thing speaks for itself ™). This principle applies in
cases where harm has been caused by a thing or a process which was under the
exclusive control of an identi¬ed person; and where the harm-causing incident was
of a type which would not, in the ordinary course of things, happen without neg-
ligence on the part of that person.4 In such cases, the harm-doer runs a real risk of
being held liable unless they can at least give a plausible explanation of how the
harm-causing incident might have occurred without negligence on their part.5
Because it is, in practice, often much more di¬cult to establish absence of negli-
gence than to prove negligence, it is sometimes said that application of res ipsa
loquitur may e¬ectively impose liability without (proof of) fault.
The res ipsa loquitur principle is particularly important in product liability
cases6 in which the user of a product has su¬ered injury as a result of what is often
called a ˜manufacturing defect™ in the product caused by some malfunction in the
manufacturing process. If the question is whether the malfunction was the result

4 See generally P.S. Atiyah, ˜Res Ipsa Loquitur in England and Australia™ (1972) 35 Modern LR 337.
5 Ng Chun Pui v. Lee Chuen Tat [1988] RTR 298. The principle has been abolished in Canada:
Fontaine v. Insurance Corporation of British Columbia (1998) 156 DLR 577.
6 It is rarely of much use in a road accident case; but see Widdowson v. Newgate Meat Corporation
[1998] PIQR P138.
Departures from the fault principle 95

of negligence on the part of the manufacturer, it will usually be up to the manu-
facturer to prove that it was not, and it will not be for the injured person to prove
that it was. The doctrine gains additional power when used in conjunction with
the doctrine of vicarious liability.7 Suppose a patient goes into hospital for an
operation to cure sti¬ness in a ¬nger, but that something goes wrong and the
person leaves hospital with ¬ve sti¬ ¬ngers. On such facts8 a court would proba-
bly be prepared to apply the doctrine of res ipsa loquitur and place on the hospital
the burden of proof on the question of whether the accident was the result of neg-
ligence. Assume that all the medical sta¬ involved in the person™s treatment were
employees of the hospital. On that assumption, the hospital would be vicariously
liable for tortious conduct of any of those medical employees. Even if the injured
person could not say which particular member of the medical sta¬ was negligent,
the hospital might be held liable for the harm unless the hospital could prove that
the harm was not the result of fault on the part of any of these people “ which
would usually be very di¬cult indeed.9
Although the e¬ect of shifting the burden of proof on the issue of fault may be
to impose strict liability, in theory it does not alter the basis of the liability, which
remains that the harm-doer should have behaved di¬erently. Therefore, in seeking
a justi¬cation for shifting the burden of proof we are not looking for a justi¬cation
of strict liability but rather a justi¬cation for relieving the injured person of the
normal burden of proof. So far as res ipsa loquitur is concerned, the justi¬cation is
that where a harm is caused by a thing or process under someone™s exclusive
control, that person is in a much better position than the injured party to know, or
to ¬nd out, how it happened. In the case of the proposed services Directive, placing
the burden on the defendant was essentially a consumer protection measure. In
general, disparity of knowledge or resources relevant to resolving the issue of fault
is the basic justi¬cation for imposing the burden of proof on the issue of fault on
the defendant.

4.3 Breach of statutory duty
Whether an action for damages lies for breach of a statutory duty depends in theory
on whether Parliament intended to confer a civil remedy when it created the duty.
But this is pure theory, because it is only in very recent times that Parliament has ever
paused to consider whether it wishes to confer such a remedy. In practice, the action
for breach of statutory duty is almost entirely con¬ned to industrial accidents.
Factory legislation and mines legislation have long been held to confer a right of
action for breach. This dates back to the last years of the nineteenth century when the
¬rst Workmen™s Compensation Act was passed, and the whole question of industrial

7 4.7.
8 The doctrine is unlikely to be very useful in less straightforward cases: M.A. Jones, ˜Res Ipsa
Loquitur in Medical Negligence Actions: Enough Said™ (1998) 14 Professional Negligence 174.
9 See Cassidy v. Ministry of Health [1951] 2 KB 343.
96 Chapter 4

safety was a prominent subject of discussion.10 There was little con¬dence at that
time that safety legislation was being adequately enforced, or could adequately be
enforced by the government inspectors appointed under the Factories Act; and it may
well be that these factors in¬‚uenced the courts in their decision to impose civil lia-
bility for breach of duties of this nature.
Attempts to extend the action for breach of statutory duty to other situations
have almost invariably been rebu¬ed. In particular, in 1923 the Court of Appeal
refused to allow an action for breach of statutory duty for breach of Ministry of
Transport regulations relating to the construction and use of motor vehicles.11 If a
motorist takes reasonable care to maintain a vehicle “ for example, by having it
regularly serviced by a reputable garage “ and (e. g.) the vehicle™s brakes suddenly
fail, the motorist will not be guilty of negligence, though they may well be guilty of
an o¬ence under statutory regulations. The court refused to impose strict liability
for breach of statutory duty on the ostensible ground that Parliament did not
˜intend™ to confer a civil remedy. Perhaps the court was in¬‚uenced “ consciously or
unconsciously “ by the fact that in 1923 it was still not compulsory to insure against
third party liability, and it may have shrunk from imposing a form of liability
without fault on individual motorists, who might not have had the resources to
meet a judgment for damages. Had this issue arisen after compulsory insurance was
introduced in 1930, the result might have been di¬erent.
Liability for breach of a statutory duty can often be imposed even in the absence
of proof of fault on the part of the party in breach. But this is by no means always
so in practice, and sometimes not even so in theory. Much depends on the wording
of the statutory provision imposing the duty. Some prescribe a result to be attained,
the most famous and important being s. 14 of the Factories Act 1961 which declares
that ˜every dangerous part of any machinery . . . shall be securely fenced™. In such a
case it is no defence to plead that all reasonable care was taken to fence the machin-
ery; or that the machine would be unusable if securely fenced. Other statutory duties
may be stated in terms that do not di¬er greatly from the usual de¬nitions of the
standard of care required by the law of negligence, though the actual requirements
of due care will usually be speci¬ed in much more detail than at common law. On
the other hand, for example, a statute may simply require precautions to be taken
˜so far as is reasonably practicable™, or words to that e¬ect; although since the burden
of proof on the issue of practicable precautions rests on the defendant, such a pro-
vision may, in e¬ect, impose liability without proof of fault.12
Courts tend to interpret even detailed provisions of industrial legislation in the
light of common law notions of fault. For example, ˜dangerous machinery™ may be
held to mean ˜machinery which is capable of causing injury if not carefully oper-
ated™. Hence a requirement to fence dangerous machinery, though not expressed in

10 Groves v. Wimborne [1898] 2 QB 402; and 13.1.1.
11 Phillips v. Britannia Hygienic Laundry [1923] 2 KB 832; see also Tan Chye Choo v. Chong Kew Moi
[1970] 1 All ER 266.
12 Nimmo v. Alexander Cowan & Sons Ltd [1968] AC 107; Larner v. British Steel Plc [1993] ICR 551.
Departures from the fault principle 97

terms of due care at all, may not in fact impose a burden signi¬cantly more onerous
than that of taking the due care.13 Obviously, if machinery is only dangerous when
injury from its use can be foreseen, it would not be a breach of the statute to omit
to fence machinery which was not a foreseeable source of danger “ and this is not
very di¬erent from the ordinary requirements of the law of negligence. For another
thing, contributory negligence and principles of causation and remoteness of
damage14 remain as limitations on liability for breach of statutory duty.
Moreover, many of the requirements imposed by statutes or regulations are in
fact no more than what reasonable care would require if only as much was known
about accident causation by the reasonable person (or the courts) as by the appro-
priate government department. Industrial legislation imposing this or that require-
ment may not always appear to require only what is reasonable care: it may appear
unnecessarily solicitous or ˜fussy™. This is often because the court or the reasonable
person does not know how many accidents are caused by the omission to take the
required precautions. Statutory requirements may be drawn up as a direct response
to a serious accident rate in this or that area of industry.15 Sometimes the main
purpose of detailed legislation is to give the employer greater guidance as to what
is required in the way of reasonable care; clearly in this event, the object is not to
impose strict liability.
Much of the time spent by appellate courts in deciding on the proper interpre-
tation of detailed provisions in industrial safety legislation is a waste, because the
search for the correct interpretation assumes that small di¬erences in wording
between di¬erent provisions were intended by the legislator to re¬‚ect important
considerations of policy, which they rarely, if ever, do.16 More often they re¬‚ect
either poor draftsmanship or a desire to cover every possible contingency.17 Even
if courts were to look for policy considerations to guide their decisions as to
whether compensation ought to be given in particular cases, it would not be easy
to ¬nd a rational legislative approach. The problem lies principally in the fact that
the primary justi¬cation for strict liability for industrial accidents was to a large
extent removed by the introduction of the industrial injuries system in 1948.18 The
main justi¬cation for strict liability for industrial accidents is (in e¬ect) that this
is a form of insurance for the bene¬t of the worker at the expense of the employer,
rather than that the employer is in some way at fault. An excellent case can be made
for saying that workers should be insured against industrial accidents, and that
this should be wholly or partly paid for by employers; but this is precisely the ratio-
nale of the industrial injuries system. It is di¬cult enough to justify the continued

13 See e.g. Leversley v. Thomas Firth [1953] 1 WLR 1206, 1210.
14 See 5.2.
15 See Annual Report of the Chief Inspector of Factories (Cmnd 3745, 1967), 23.
16 See the comments of Lord Reid in Nimmo v. Alexander Cowan & Sons [1968] AC 107.
17 See e.g. Robens Committee Report, para. 29: ˜the attempt to cover contingency after contingency
has resulted in a degree of elaboration, detail and complexity that deters even the most determined
18 See 12.4.
98 Chapter 4

existence of liability based on fault in industrial accidents despite the existence of
the industrial injuries system; it is almost impossible to justify the continued exist-
ence of strict liability.19 This is why, in 1946, the Monckton Committee on
Alternative Remedies recommended (in e¬ect) that, with the enactment of the
Industrial Injuries Act of 1946, liability for breach of statutory duty should cease
to be ˜strict™.20 This recommendation was never implemented by Parliament, and
in consequence, the courts have had to approach the problem of interpreting
industrial legislation against the background of an indefensible policy decision.
Small wonder, then, that the courts have failed to evolve any consistent approach
to the problem based on a clear and intelligible policy. A statement by Lord
Diplock in Haigh v. Ireland 21 suggests that the courts now appreciate the true situ-
ation more clearly. In this case Lord Diplock said that the courts must resist the
temptation to stretch the interpretation of industrial legislation in order to ensure
the compensation of injured workers. Compensation without fault, as he pointed
out, is available under the industrial injuries system to all workers injured in the
course of their employment. Statutory provisions should, therefore, be interpreted
without any bias in favour of injured workers.
A fresh start in this area was made in the Health and Safety at Work etc. Act 1974.
This Act imposes on employers various general duties “ such as a duty to ensure the
health, safety and welfare of employees. These general duties are all quali¬ed by the
phrase ˜so far as is reasonably practicable™, making them, in theory, duties of reason-
able care; but in practice, since the burden of proof of practicability rests on the
defendant, the e¬ect may be to impose liability without (proof of) fault. The Act
expressly states (s. 47(1)) that the general duties do not give rise to civil liability for
damages; and they are enforceable only by criminal prosecution. By contrast, s. 47(2)
of the Act provides that breaches of health and safety regulations are actionable in
damages unless the contrary is expressly stated “ a rare occurrence.22 As health and
safety regulations that existed at the time the 1974 Act was passed are gradually
replaced by regulations made under that Act, the importance of the older case-law
dealing with the availability of an action for damages, will diminish in this area.
In relation to industrial injuries, the action for breach of statutory duty would
appear to be of great importance if judged solely by the number of statutory duties
imposed on employers and by the number of reported cases. We have seen that a
very large proportion of all litigation is personal injury litigation, and that about
25% of this is industrial in origin (much arising out of breaches of statutory
duties). Nevertheless, in practice the great bulk of employees su¬ering injury still
appear to have no legal cause of action; at any rate they do not in fact make tort

19 This was one of the reasons why the Pearson Commission (Pearson Report, vol. 1, paras. 918“22)
rejected a suggestion that the burden of proof in actions for breach of industrial safety regulations
should be reversed.
20 Cmnd 6860; see further 15.4.5.
21 [1974] 1 WLR 43, 54“5.
22 An example is Management of Health and Safety at Work Regulations 1999, reg. 22(1).
Departures from the fault principle 99

claims for damages.23 It is also noticeable that although there are many ˜strict duties™
under industrial legislation, and very few under road safety legislation, a much
larger proportion of road accident victims than industrial accident victims actually
succeeds in recovering some tort damages.

4.4 Contractual duties
The law of contract is primarily concerned with ¬nancial losses and not with phys-


. 4
( 18)