ISSN 1743-8713

:: Articles

:: Discussion

Last updated: 14/10/02

Oxford University Comparative Law Forum

Liability for Nonfeasance; a comparative study

by Jeroen Kortmann

(2001) Oxford U Comparative L Forum 1 at ouclf.iuscomp.org | How to cite this article | Discuss this article

Table of content

Headnote
Acknowledgements
I. Introduction
II. What is ‘nonfeasance’?
III. Theoretical justifications for distinguishing between feasance and nonfeasance
IV. Liability for nonfeasance in continental european private law
V. Liability for nonfeasance in English private law
VI. Comparative analysis and conclusion
Abbreviations
Bibliography
Table of Cases
Table of Legislation
Footnotes

Headnote

This article was submitted for the degree of Master of Studies in Michaelmas term 1999. The article includes few references to material that appeared after that date. Some of the foreign material that is referred to may be rather outdated. The author would like to emphasise that most of this article is part of a larger project, provisionally titled ‘Altruism in Private Law’ (to be submitted for the degree of Doctor of Philosophy), and should be regarded as ‘work in progress’. Any comments or suggestions would therefore be very much appreciated.

Acknowledgements

I owe a debt of gratitude first and foremost to my supervisor, Mr Robert Stevens, for the enthusiasm with which he guided me through the past year of research. I would also like to thank Dr Gerhard Dannemann and Dr David Ibbetson for agreeing to read drafts of sections of this thesis which touched upon their respective areas of specialisation. Their comments and suggestions were very helpful.

Dr Ibbetson allowed me to read draft-chapters of his forthcoming book, A Historical Introduction to the Law of Obligations, which proved to be indispensable as will become apparent from the frequent references to it in the fifth chapter of this thesis. I also had the pleasure of reading a draft of an article titled ‘Barrett, Negligence and Discretionary Powers’, by Professor Paul Craig and Mr Duncan Fairgrieve. I am very thankful to these authors for allowing me a ‘sneak-preview’.

I very much appreciate the effort of Nick Smith and Shannon McBriar in proof-reading drafts of this thesis.

Finally, I would like to take this opportunity to thank the many institutions that have so generously supported me: the European Commission through the TMR (Training and Mobility in Reasearch) Programme, the Radboudstichting for granting me a ‘stimuleringssubsidie’, the Art and Humanities Research Board for awarding me a studentship and finally Merton College, for awarding me a Domus Graduate Scholarship. Without their continued financial assistance I would not have been (and still would not be) able to afford my stay in Oxford.

I. Introduction

‘But he, willing to justify himself, said unto Jesus, And who is my neighbour?

And Jesus answering said, A certain man went down from Jerusalem to Jericho, and fell among thieves, which stripped him of his raiment, and wounded him, and departed, leaving him half dead.

And by chance there came down a certain priest that way: and when he saw him, he passed by on the other side.

And likewise a Levite, when he was at the place, came and looked on him, and passed by on the other side.

But a certain Samaritan, as he journeyed, came where he was: and when he saw him, he had compassion on him,

And went to him, and bound up his wounds, pouring in oil and wine, and set him on his own beast, and brought him to an inn, and took care of him.

And on the morrow when he departed, he took out two pence, and gave them to the host, and said unto him, Take care of him; and whatsoever thou spendest more, when I come again, I will repay thee.

Which now of these three, thinkest thou, was neighbour unto him that fell among thieves?

And he said, He that shewed mercy on him. Then said Jesus unto him, Go, and do thou likewise.’ (Luke, 10:29-37)¹

From a private law point of view, the parable of the Good Samaritan raises two questions:

1) Do we have a legal duty to give aid to our fellow human beings? In particular: can we be held liable for damages if we fail to do so? This question falls mainly in the area of tort-law.

2) If we do come to the rescue, like the Samaritan did, will we have any claim for the expenses that we incurred in doing so? Such claims could arise in tort, in unjust enrichment or in "negotiorum gestio".

This thesis will only deal with the first question: do we have a duty to give aid to our fellow human beings? Lord Atkin's speech in Donoghue v Stevenson² gives reason to believe that the answer is affirmative under English law:

The rule that you are to love your neighbour becomes in law, you must not injure your neighbour; (...) Who, then, in law is my neighbour? The answer seems to be - persons who are so closely affected by my act that I ought reasonably to have them in contemplation as being so affected when I am directing mind to the acts or omissions which are called in question.³

However, in Home Office v Dorset Yacht Co. Ltd⁴ Lord Diplock criticized Lord Atkin's test in Donoghue. In the view of Lord Diplock omissions give rise to no legal liability:

The very parable of the Good Samaritan which was evoked by Lord Atkin in Donoghue v Stevenson illustrates, in the conduct of the priest and of the Levite who passed by on the other side, an omission which was likely to have as its reasonable and probable consequence damage to the health of the victim of the thieves, but for which the priest and the Levite would have incurred no civil liability in English law.⁵

It has been confirmed in later judgments that English law does not in principle recognise an affirmative duty to act,⁶ and thus in general does not require us to give aid to our neigbours in need. In this respect English law differs from most legal systems in Continental Europe. The delictual provisions in the civil codes of for example Germany and France do not distinguish between acts and inaction.

In this thesis the theoretical justifications for distinguishing between acts (or ‘feasance') and inaction (or ‘nonfeasance') will be set out in chapter three. Chapter four will proceed with examining the law in two Continental European countries, Germany and France. In chapter five the position in English law will be subjected to a closer analysis. The final aim of this thesis is to investigate if and, if so, to what extent the position of English law on liability for nonfeasance is different from the position of its European counterparts. However, before we can start a more detailed investigation into ‘liability for nonfeasance', it needs to be decided what definition of ‘nonfeasance' will be used.

II. What is 'nonfeasance'?

When dealing with liability for nonfeasance, other terms that are generally used to indicate a lack of feasance are ‘inaction', ‘omission' or ‘failure to act'. There is a subtle difference between these terms. On the one hand we have ‘nonfeasance' and ‘inaction', which are both neutral terms.⁷ ‘Omission' and ‘failure to act' on the other hand, presuppose the existence of a duty, be it legal, moral or otherwise. To say that a person has ‘omitted' or ‘failed to act', is to say that he really should have done what he omitted or failed to do.⁸ In the following the terms ‘omission' and ‘failure to act' will therefore only be used in cases in which the existence of a duty to act is undebatable.

So what is nonfeasance? Answers like ‘the opposite of feasance' do not really help us. In truth, there are many situations in which it is very difficult to draw any logical distinction between feasance and nonfeasance.⁹ A cardriver who does not brake in time to avoid hitting a pedestrian can be said to have driven negligently. But can it not equally be argued that his negligence did not consist of an act, but rather of a failure to act? Fleming's answer is negative:

What superficially looks like non-feasance is often, upon correct analysis, a case of misfeasance. A motorist's failure to brake in time to prevent a collision is not an example of supine inaction: an omission is involved, but it is merely the element that makes his active conduct - driving - negligent.¹⁰

In other words: although the failure to brake, in isolation, is nonfeasance, it is part of a larger activity, driving, which has to be qualified as feasance. The problem with this view is that it helps you with answering one question - ‘are we dealing with feasance or nonfeasance?' - but at the same time gives rise to two new questions:

(i) ‘Is there a larger activity?', and

(ii) ‘Is the isolated nonfeasance part of this larger activity?'

The answer to the first question is often arbitrary. Fleming does not hesitate to regard ‘driving' as a larger activity. Yet, when railways fail to provide signals at a level crossing, he regards this as inaction.¹¹ Does this mean that ‘running a railway company' does not qualify as a larger activity? Similarly, ‘supervising your children' and ‘babysitting' are apparently not to be regarded as larger activities, for Fleming treats ‘losing control over children' as (isolated) inaction.¹²

If we have identified a larger activity, the second question will prove at least as high a hurdle as the first one. Why is it that we regard a motorist's failure to brake as part of the larger activity of driving? Let us consider the following example: a motorist drives his car down a country road. Suddenly he sees a severely wounded man ‘stripped of his raiment', beckoning for help. He decides not to brake. The man continues his journey and a little while later he is confronted with a pedestrian crossing the road. Again he fails to brake, and the pedestrian is killed in the collision.¹³ Now here we have two instances of failure to brake. Yet, Fleming would probably only regard the second failure to brake as being part of the larger activity ‘driving'. The most likely reason for doing so, is that the traffic rules command the motorist to stop in the second case, but not in the first. In other words: he regards only the second instance of inaction - not braking for a crossing pedestrian - as part of the larger activity - driving - because only in the second instance the motorist has a duty to act. Thus, to Fleming the mere existence of a duty to act can make the difference between inaction and action. For obvious reasons, Fleming's way of distinguishing between feasance and nonfeasance cannot be adopted in this thesis: by qualifying inaction as part of a larger act, he unwittingly obscures from our view the very duty to act that we are trying to investigate. Instead, we will try to look at nonfeasance itself, isolating it from the acts it may have been accompanied by. Thus, we will regard both ‘not stopping your car to help a man in need' and ‘not stopping your car to avoid collision' as instances of nonfeasance. This is of course not to say that similar liability will arise in both cases, because the existence of a duty to act is not at all certain in the first case, whereas it probably is in the second.¹⁴

The courts however, do not adopt one uniform definition of nonfeasance. As a result it sometimes proves difficult to establish in individual cases whether the court has made its decision on the presumption that the behaviour complained of is misfeasance or (also) nonfeasance. In those cases close attention should be paid to the court's reasoning. Most importantly, we must try to establish what particular part of the defendant's behaviour prompted the court to conclude that he had been negligent. E.g. was it the fact that he hit a cricketball over the fence or that he had failed to build a higher fence? Some of the judgments in the famous cricket-case of Bolton v Stone indeed give reason to believe that this case was concerned with liability for nonfeasance, rather than feasance.¹⁵

Thus, although a hard rule that is always able to make the proper distinction between feasance and nonfeasance is impossible to give,¹⁶ this thesis will apply the following two guidelines:

I) All action and inaction that is part of the defendant's behaviour will be regarded in isolation. No inaction is ‘part of an act' and vice versa.

II) Close attention will be paid to the court's reasoning; which part of the defendant's behaviour prompted the court to conclude that he had been negligent?

III. Theoretical justifications for distinguishing between feasance and nonfeasance

III:1. Introduction

By not recognizing a general affirmative duty to act, English law does not encourage members of the public to give aid to those in need of help. Quite conversely, English law in fact discourages giving aid, for once someone has started to give aid, he can be held liable for mistakes he makes in doing so. The advice Gregory gave in 1966 - ‘Don't ever "fease" unless you have to'¹⁷ - is still valid under modern English law. Nevertheless, there are apparently enough justifications for the House of Lords to uphold the distinction between feasance and nonfeasance, and to deny the existence of a general duty to act. In the recent case of Stovin v Wise,¹⁸ Lord Hoffmann explained:

There are sound reasons why omissions require different treatment from positive conduct. (...) One can put the matter in political, moral or economic terms. In political terms it is less of an invasion of an individual's freedom for the law to require him to consider the safety of others in his actions than to impose upon him a duty to rescue or protect. A moral version of this point may be called the "Why pick on me?" argument. A duty to prevent harm to others or to render assistance to a person in danger or distress may apply to a large and indeterminate class of people who happen to be able to do something. Why should one be held liable rather than the other? In economic terms, the efficient allocation of resources usually requires an activity should bear its own costs. If it benefits from being able to impose some of its costs on other people (what economists call ‘externalities') the market is distorted because the activity appears cheaper than it really is. So liability to pay compensation for loss caused by negligent conduct acts as a deterrent against increasing the cost of the activity to the community and reduces externalities. But there is no similar justification for requiring a person who is not doing anything to spend money on behalf of someone else. Except in special cases (such as marine salvage) English law does not reward someone who voluntarily confers a benefit on another. So there must be some special reason why he should have to put his hand in his pocket.

Each of the justifications given by Lord Hoffmann will be examined in the following, together with additional justifications that have been brought forward in academic writings.

III:2. A positive duty to act would be too much of an invasion of an individual's freedom

In academic debates a mere reference to the sanctity of freedom is often treated as irrefutable and conclusive to the outcome of the debate. We seem satisfied that invasion of an individual's freedom cannot be tolerated, and often regard an argument to that effect as self-evident and self-justifying, thus rendering additional reasons unnecessary.¹⁹ Yet, what an individual's freedom actually consists of, is a question that has hardly been answered by two ages or two countries alike.²⁰ Hence, Lord Hoffmann is not being very informative when he tells us that omissions require different treatment from positive conduct because ‘it is less of an invasion of an individual's freedom for the law to require him to consider the safety of others in his actions, than to impose upon him a duty to rescue or protect'.²¹ At the most we can deduce from this statement that Lord Hoffmann's understanding of ‘freedom' includes the right to refrain from acting. Regrettably, Lord Hoffmann fails to give us his definition of ‘freedom', nor does he tell us why one kind of ‘invasion of an individual's freedom', the imposition of liability for wrongful inaction, would be illegitimate where another, the existing liability for wrongful acts, apparently is legitimate.

Freedom or liberty (I shall regard both words as synonymous), is normally said to be the absence of interference by others, or indeed the state, with an individual's activity. If I am prevented by others from doing what I could otherwise do, I am to that degree unfree.²² This is not to say that interference is necessarily a bad thing. If we want to guarantee everyone a minimum degree of freedom, we will have to restrain everyone from interfering with each other's freedom. Thus, the freedom of one individual finds its limits where it collides with the freedom of others.²³ The difficulty lies in finding the proper standard of interference to apportion a fair share of freedom to each individual. Lawyers and philosophers alike have made numerous attempts to define such a standard of interference. This is not the time, nor the place for a detailed survey of the different standards that have been suggested through the centuries. For present purposes we will have to content ourselves with a summary of the philosophical views that are most often referred to by supporters and adversaries of the rule of no liability for nonfeasance.²⁴

Supporters of the no-liability rule most often refer to the works of Kant for confirmation of their views.²⁵ In his The Metaphysics of Morals, first published in 1797, Kant made a distinction between the doctrine of right on the one hand, and the doctrine of virtue on the other. Regarding the doctrine of right he wrote:

The concept of Right, insofar as it is related to an obligation corresponding to it (i.e., the moral concept of Right), has to do, first, only with the external and indeed practical relation of one person to another, insofar as their actions, as facts, can have (direct or indirect) influence on each other.²⁶ (Footnotes omitted)

Because the doctrine of right only deals with external actions, Kant is said to have been of the view that the law should not require any particular affirmative actions.²⁷ An individual's choice whether or not to act was an internal affair, which was governed by the doctrine of virtue, rather than of right. Accordingly, a duty to aid others was an ethical, not a legal duty:

(...) duties of benevolence, even though they are external duties (obligations to external actions), are still assigned to ethics because their lawgiving can be only internal.²⁸

Thus, the Kantian idea appears to be, as Howarth puts it, ‘that the state has no business telling people that they are bad people, that they are immoral or wicked'.²⁹ However, Kant himself may not have been as opposed to state interference as is often assumed. The Metaphysics of Morals included a section on the rights of a state, in which Kant wrote that a government was authorised to ‘constrain the wealthy to provide the means of sustenance to those who are unable to provide for even their most necessary natural needs', for example by imposing taxes to that effect.³⁰ One could argue that the imposition of taxes is different from the imposition of liability in that the former does not involve casting a judgment on an individual's character; after all, everyone has to pay taxes regardless of one's (internal) virtue. However, Kant's suggested solution for the maintaining of abandoned children does appear to involve a straightforward judgment on the taxpayers' virtue:

(...) the state has a right to charge the people with the duty of not knowingly letting them die, even though they are an unwelcome addition to the resources of the state. Whether this should be done by taxing elderly unmarried people of both sexes generally (by which I mean wealthy unmarried people), since they are in part to blame for there being abandoned children, in order to establish foundling homes, or whether it can be done rightly in another way (it would be hard to find another means for preventing this) is a problem which has not yet been solved (...)³¹ (Second emphasis added)

Adversaries of the no-liability rule tend to refer to the works of Bentham and Mill.³² Bentham appears to have agreed with Kant that ‘the rules of beneficence' generally belonged to the ‘jurisdiction of private ethics'.³³ However, in cases involving danger to others, he thought that the law ought to interfere:

(...) why should it not be made the duty of every man to save another from mischief, when it can be done without prejudicing himself, as well as to abstain from bringing it on him?³⁴

Unfortunately, Bentham did not proceed to reveal his arguments in favour of imposing such a duty to rescue.³⁵ Mill's views are more helpful in this respect. In his On Liberty,³⁶ Mill made an attempt to define the proper standard of state interference:

(...) the sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number, is self-protection. That the only purpose for which power can be rightfully exercised over any member of a civilised community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant. He cannot rightfully be compelled to do or forbear because it will be better for him to do so, because it will make him happier, because, in the opinions of others, to do so would be wise, or even right. (...) Over himself, over his own body and mind, the individual is sovereign.³⁷ (Footnotes omitted)

The proper standard according to Mill was what we now know as the ‘harm principle'; the community is only permitted to restrain the behaviour of its members, insofar as this behaviour causes harm to others.³⁸ Consequently we, as individuals, should in principle³⁹ be free to do what we want as long as we do not cause harm to others. From the very moment of its publication, On Liberty has been lauded by some,⁴⁰ while denounced by others.⁴¹ The main objection to the harm principle seems to be its apparent vagueness. If ‘harm to others' is to be the standard of legitimate interference, what must ‘harm' be understood to mean? In a chapter devoted to applications of his harm principle, Mill wrote that a man should be free to be drunk if he wants to, provided that he does not harm others in doing so.⁴² Many of Mill's contemporaries however, regarded the mere state of drunkenness as in itself harmful to others, because in any case it would set a bad example for the rest of the community.⁴³ Thus interpreted, application of Mill's harm principle could justify (state) interference with nearly every type of human behaviour. However, such interpretation does not do justice to Mill's theory. Mill used a common-sense concept of harm, which did not include mere offences,⁴⁴ nor the kind of negative effects that can be attributed to the free will and consent of the ‘harmed' person. Thus, one cannot stop a drunkard from being drunk or forbid an atheist preacher from leading his audience astray, because those who may be negatively affected are free to ignore the drunkard and the atheist priest and have themselves to blame if they do not.⁴⁵

Although Mill's critics reject the harm principle as the general standard of (state) interference, most of them accept that ‘harm to others' can be a sufficient reason (though not the only reason) to justify such interference. The reviewer of The Athenaeum predicted this in 1859:

We doubt if anyone, in modern times, will venture to dispute the principle. Some will take it as their guiding rule on principle: others will aver that, whatever other principles they may also allow to act, there is no proper case of application in which this principle does not also apply.⁴⁶

It has been argued that even Kantian ethics are not altogether inconsistent with the harm principle.⁴⁷ Thus, the predominant opinion appears to be that if a certain behaviour causes harm (in Mill's narrow sense of the word) to others, there is a prima facie case for state interference. We must then ask ourselves whether nonfeasance is or can be the kind of behaviour that causes harm to others. This question will be dealt with in more detail in section III:6 of this thesis. For the moment we will content ourselves with Mill's response:

A person may cause evil to others not only by his actions but by his inaction, and in either case he is justly accountable to them for the injury. The latter case, it is true, requires a much more cautious exercise of compulsion than the former. To make anyone answerable for doing evil to others is the rule; to make him answerable for not preventing evil is, comparatively speaking, the exception. Yet there are many cases clear enough and grave enough to justify that exception.⁴⁸

Accordingly if we assume, as Mill does, that nonfeasance can cause harm, there appears to be a prima facie case for interference. Nonetheless, there may be good reasons ‘secunda facie', ‘at second sight', to support Lord Hoffmann's argument that liability for nonfeasance would be too much of an invasion of an individual's freedom. Regrettably however, Lord Hoffmann does not advance such reasons.

One reason that has been suggested is that we may consider positive duties to be more burdensome than negative duties. This argument is easily met by Atiyah.⁴⁹ For although positive duties can indeed be quite burdensome, we deny liability even in those cases where there would hardly be any burden. Atiyah gives the example of a person who sees a neighbour's house on fire. ‘How burdensome would it be to require that person to telephone for the fire brigade?'⁵⁰ We should also remember that negative duties can sometimes be very burdensome. For although ‘do not act uncarefully' may seem to be an easy duty to perform, this will become more difficult the more we raise the standard of care. The behaviour of learner-drivers, for example, is measured against the standards of experienced drivers.⁵¹ Consequently, the learner driver will either have to drive as carefully as an experienced driver, which to him is nearly impossible, or not drive at all. And if anyone were to suggest that the latter option, ‘not driving at all', is hardly burdensome, he is mistaken. Where the use of a car may have been a luxury in the past, it is now to many people an indispensable tool of transportation.⁵²

Also, it should be noted that the burdensomeness of a duty is a factor that the courts can take into account at another stage of their judgement, namely when deciding whether the defendant was at fault. If the court finds the performance of a duty too burdensome, it can decide that in the specific case before the court, the defendant has behaved like a reasonable man. Take for example the case of Latimer v AEC Ltd.,⁵³ where the defendant was held not to be at fault for failing to prevent an accident that happened when an employee slipped on the wet factory floor after a flooding. The court was of the opinion that the only possible way of preventing the accident from occurring would have been to close the whole factory altogether until the floor was cleaned. But it regarded this remedy as too burdensome, and concluded that the defendant had behaved like a reasonable man. Hence, the burdensomeness of a duty does not have to prevent the imposition of liability in a whole category of cases, but can be taken into account when deciding in specific cases whether the behaviour complained of was reasonable.⁵⁴ Thus, if Lord Hoffmann's ‘freedom-argument' relies on the presumption that the imposition of liability for nonfeasance would be too burdensome, it is difficult to support his view.

III:2 (a) Lord Nicholls's and the ‘freedom of the will'

When discussing the notion of freedom, Samuel von Pufendorf - a 17th century writer of the Natural law school - used to emphasise the importance of the individual's ability to choose. The very reason for holding people responsible for their acts, he argued, is that they voluntarily choose to act the way they do:

And just as the chief reason for considering a man responsible for his own acts is that he undertook them of his own will, so we must especially observe that the freedom of the will is by all means to be asserted, at least in regard to the acts for which a man is commonly held to account before a human court.⁵⁵

In his speech in Stovin v Wise, Lord Nicholls appears to have adopted a notion similar to Pufendorf's ‘freedom of the will' to justify the feasance-nonfeasance distinction:

Liability for omissions gives rise to a problem not present with liability for careless acts. He who wishes to act must act carefully or not at all. (...) With liability for omissions, however, a person is not offered a choice. The law compels him to act when left to himself he might do nothing.⁵⁶

It must not be doubted that every duty that the law imposes - whether positive or negative - restricts to a certain extent our freedom to choose. Yet, Lord Nicholls argues, positive duties are more restrictive because they deny us the option to avoid being subjected to them. In general, negative duties do offer us this option. If we do not feel competent enough to meet a specific negative duty, we can avoid being subjected to it by not acting at all. Thus, if I feel I cannot live up to the standard of the reasonable driver, I can decide not to drive at all. However, if the law were to require a witness of an accident to act as the reasonable rescuer would do, he would not have an alternative option that would allow him to avoid being subjected to this duty to rescue.

Lord Nicholls's justification of the feasance-nonfeasance distinction would appear to be a powerful one. There are however some objections to his argument. Firstly, the law already imposes liability in some cases in which we do not have a choice in the sense here discussed. If I blamelessly become the source of a danger, the law imposes on me a duty to prevent this danger from culminating in harm.⁵⁷ Thus, if through no fault of my own I fall on the road (e.g. I was pushed by a stranger), I have a duty to stand up and clear the road in order to avoid endangering other road-users, even though I had no choice which would have allowed me to avoid being subjected to this duty. Therefore, if the law recognises Lord Nicholls's version of ‘freedom of the will', it does not consistently apply it.

Secondly, it is not entirely true to say that we do not have an option that allows us to avoid being subjected to affirmative duties. For in theory, we could lock ourselves in our houses in an attempt to avoid witnessing accidents or other occasions that could require our positive action. True, few people would regard this as a realistic option, but the same can be said in some cases involving negative duties. E.g. ‘not driving at all' in order to avoid being subjected to the relatively high standard of care required of cardrivers is to many of us not a realistic option either. Hence, the distinction between cases of feasance and of nonfeasance as it is drawn here, appears to be one of degree, not of principle. Imposing positive duties does not leave us without an option to avoid being subjected to these duties, but rather with less of an option than is commonly the case in respect of negative duties.

III:3. ‘Why pick on me?'

When caught red-handed with his pockets full of sweets, a child may try to avoid punishment by pointing at his bolder sibling who did the actual stealing. Parents are rarely impressed by this ‘why pick on me?' argument. After all, one child's fault is not erased by the fact that another child was equally or even more at fault. Though essentially still the same argument, the ‘why pick on me?' argument may seem more convincing in complex situations involving nonfeasance. Let us for example recall the (in)famous case of Kitty Genovese who was raped and murdered in broad daylight. Dozens of people watched from their windows and ignored her cries for help. Nobody called the police.⁵⁸ Why would we allow Kitty's family to ‘pick' on one of these people, by bringing a claim in negligence? After all, were the other spectators not equally to blame? And, is it not true that the rapist/murderer was the ‘real' culprit? As with the sweets-stealing siblings, the answer is fairly straightforward: the question is not whether others were equally or even more culpable, but whether this individual defendant was at fault.⁵⁹ In this respect there is no reason to differentiate between feasance and nonfeasance.⁶⁰ Thus, we allow a plaintiff to ‘pick' on one of a large group of potential defendants because they all individually failed to do what a reasonable man would have done.⁶¹

This is not to say that the presence of other people has no bearing whatsoever on the duty of an individual. Depending on the circumstances, a reasonable man may sometimes refrain from acting or limit his actions because there are other people who are in a better position to intervene than he is. The passer-by in the street who witnesses Kitty's ordeal may be required to run to a phonebox and call the police if he is the only one there. However, another conclusion may be reached if he notices that there are dozens of people witnessing the crime from their windows. Under such circumstances it can be argued that the passer-by may legitimately believe that his intervention is unnecessary, relying on one of the other witnesses to make the phonecall. Thus, the presence of other people - especially if they are in a better position to intervene - can be taken into account when we decide whether the defendant's behaviour was reasonable. However, if and when we decide that his behaviour was unreasonable, there is no (good) reason to deny a claim against him merely because someone else was equally or even more culpable. Thus, we allow a plaintiff to single out, or ‘pick on' the defendant of his choice. Any unfairness that might arise from this system, ought to be overcome by the Civil Liability (Contribution) Act 1978, which allows the defendant to bring an action against anyone else ‘liable in respect of the same damage' to recover a contribution from them.⁶²

III:4. Economic arguments

In the second half of this century, the law of torts has increasingly been subjected to economic analyses. Terms like ‘wealth maximization', ‘price deterrence', ‘internalisation of externalities' and ‘the cheapest cost avoider' have become commonplace in textbooks on the law of torts. This is not the right place to examine the various economic theories that have been developed.⁶³ However, the concept of ‘price deterrence' requires some attention, if only because Lord Hoffmann directly refers to this concept in his speech in Stovin v Wise. Price deterrence has been put forward as an economic justification for the imposition of tort liability on the person who causes loss to others. Essential to the concept of price deterrence is the presumption that if we hold a person liable for the losses his behaviour causes, he will eventually be unable to afford to continue or repeat his harmful behaviour. Take for example two factories that produce a similar type of car. Factory A causes no harm in doing so. Factory B however, uses a production method that has a negative effect on others, say, it causes pollution to the neighbouring land. If the social costs represented by the pollution are not transferred to factory B, there are ‘externalities'. If we ‘internalise' these externalities to the activity of factory B, the price of the cars of factory B will rise to a level reflecting the ‘true social costs' of the product. As a consequence, some consumers will buy the cars of factory A instead of B, and eventually factory B will go out of business and the pollution will stop.⁶⁴ Thus, imposition of tort liability removes externalities and consequently enhances efficiency. As is the case with the general argument of deterrence, the argument of price deterrence is open to considerable criticism.⁶⁵ Let us however presume that it has some merit in justifying the imposition of liability. Then, why is it that Lord Hoffmann alleges that this theory does not provide a ‘similar justification for requiring a person who is not doing anything to spend money on behalf of someone else'⁶⁶? If it is true that to accomplish efficiency the externalities caused by actions need to be internalised, then why is this not similarly true for the externalities caused by inaction? Inaction, like action, can cause externalities. Factory C may produce cars without causing any pollution or danger, but if it fails to provide accurate instruction manuals with the cars, this might very well lead to mishandling of the cars and consequently to damage, ‘externalities'. There is no (good) reason why the economic reasons to put factory B out of business do not equally apply to factory C. Apparently, Lord Hoffmann's assumes that inaction does not cause loss, and therefore does not cause externalities. From the example with the instruction manuals, we see that this point of view is not always easy to maintain. In section III:6 the question whether nonfeasance can cause loss will be subjected to a closer investigation.

A rather different economic argument is put forward by Honoré in a 1991 tribute to Atiyah:⁶⁷

The most plausible arguments for sanctioning acts more severely than omissions have so far been economic. One can refrain from a great many harmful acts, like poisoning, at the same time, while it is difficult or impossible to combine more than one helpful act, like providing food. The benefit of abstention is also widely distributed, in that if I abstain from doing harm no one is harmed, while the benefit of helpful acts is confined to the few who get the food I can provide.⁶⁸ (Footnotes omitted)

In other words: the performance of a negative duty involves a smaller effort than the performance of a affirmative duty, because one can ‘refrain from a great many harmful acts (...) at the same time'. Moreover, the performance of a negative duty has a more widespread effect, because ‘if I abstain from doing harm no one is harmed'. However, both submissions are debatable. As for the effort required, for the individual the performance of a negative duty often requires less effort than the performance of an affirmative duty. However, the performance of a negative duty is not always without effort. Sometimes one cannot perform a negative duty other than by taking affirmative action. When confronted with a traffic sign that tells you not to turn left on a T-junction, you will either have to stop your car or turn right. Moreover, for a negative duty to have its required effect, it needs to be performed by all the members of the public who are in the vicinity of the protected interest. The cumulative effort of these members of the public in performing the negative duty can very well exceed the individual effort that is required for the performance of the corresponding affirmative duty. An example can clarify: during a stormy night an unlucky pedestrian is knocked unconscious by a falling branch and falls down in the middle of a street. During the morning rush hour the unconscious man creates a considerable obstacle. To protect the man from being hurt, we can either impose a negative duty not to drive into the man or an affirmative duty to revive the man and/or remove him from the street. From the individual cardriver the negative duty requires less effort. It is easier to manoeuvre his car around the unconscious man than to have to stop his car, get out and help him. But to avoid the man being hurt, it is not just this cardriver who has to guide his car around the man. Every single cardriver who passes through this street will have to do the same. Conversely, it only takes one cardriver's effort to reach the same result by performing the affirmative duty. The cumulative effort of the cardrivers in performing the negative duty may in this case very well exceed the individual effort in performing the corresponding affirmative duty, reaching the same effect.

The example of the unconscious man also shows that the performance of a negative duty does not always have a more ‘widely distributed' benefit; it is only the unconscious man who benefits from the cardrivers' abstention from driving into him. Yet, affirmative action - e.g. taking the man off the street - will not only benefit the unconscious man, but also all the cardrivers who will be passing through the street; they will no longer have to manoeuvre their car around the unconscious man.

Even in cases where the benefit of the performance of a negative duty is more ‘widely distributed' than the benefit of the performance of an affirmative duty, this does not force us to the conclusion that imposition of the negative duty is necessarily more efficient. For the ‘output' of a positive act depends on more than just its ‘distribution'. The example given by Honoré provides a good illustration:⁶⁹ although everybody in my vicinity enjoys the benefit of my abstention to poison them, this is relatively only a very small benefit. For the fact that I abstain from poisoning a hungry man by no means protects him from being poisoned by somebody else. And even if we all abstain from poisoning this hungry man, he may still die of starvation. If I actively feed him, he will survive. Thus, the benefit of a positive act may often be less ‘widely distributed', but also larger than the benefit of inaction.

The most elaborate attempt to give an economic justification of the general denial of liability in cases of nonfeasance was made by the influential American legal economists Landes and Posner. In 1973, Posner had voiced the opinion that - at least in rescue cases - there are sound economic reasons for imposing affirmative duties.⁷⁰ In rescue cases, Posner argued, transaction costs are so high that we cannot rely on the free market principle for efficiency. We must therefore create an incentive for potential rescuers to come to the rescue when such intervention is efficient, by introducing liability for failure to rescue. When the costs of rescue are lower than its benefits, potential rescuers will feel compelled to come to the rescue in order to avoid liability. When the rescue costs are higher than the benefits, the ‘reasonable man standard' allows potential rescuers to abstain from intervention. By 1978 however, Posner's views had radically changed. After a detailed study into the economic aspects of altruism, carried out with Landes, Posner came to the conclusion that introduction of a positive duty to rescue would not necessarily promote efficiency.⁷¹ In individual cases, Posner and Landes argued, the view that the presence of a duty to rescue would promote efficiency, is broadly sound. As a general rule however, liability for failure to rescue would have negative side-effects that might lead to inefficiency. Landes and Posner had the following negative side-effects in mind:

1) Potential rescuers would avoid liability by ‘substituting away' from activities that give rise to rescue opportunities.⁷² E.g. good swimmers would avoid crowded beaches.⁷³ Hence, there would be less potential rescuers on the beach, and those remaining would be more likely to be called upon.

2) Even if imposition of a legal duty to rescue would lead to a rise in successful rescues, this would encourage potential victims to disregard their own safety. Moreover, the number of potential victims would increase as hazardous activities would become more attractive.⁷⁴

3) Under the present Common Law regime, rescuers enjoy public recognition as altruists. Under a regime of liability for failure to rescue, it would be impossible for a rescuer to prove that he was motivated by altruism. This could again discourage potential rescuers from going to places where there is a relatively high likelihood that they will be called upon (e.g. crowded beaches).⁷⁵

4) The costs of administering a liability rule are high. ‘Legal error costs in particular might be high because of the difficulty, in many settings (e.g., on a crowded beach), of identifying potential rescuers.'⁷⁶

It must be seriously doubted whether these predicted side-effects are likely to occur. The prediction of the first side-effect overlooks three important points. Firstly, we learn from experience that imposition of liability rarely dissuades people from indulging in activities that may give rise to such liability. Rather, they take more careful precautions or insure themselves against the increased risk of liability.⁷⁷ Secondly, there are so many activities that may give rise to rescue opportunities, that it is virtually impossible for potential rescuers to substitute away from them. Not only will they have to avoid crowded beaches, but also sporting grounds, industrial sites, busy roads and areas with a relatively high crime rate. Thirdly, Posner and Landes overlook the fact that society does not consist of categories of potential rescuers on the one hand, and potential victims on the other. If I go to the beach, am I a potential victim or a potential rescuer? ‘Both', is probably the right answer. Most of the people present on a beach are potential rescuers and potential victims at the same time. So even if we accept that the amount of potential rescuers would decrease after the imposition of liability, the amount of potential victims would decrease correspondingly.

The suggested second side-effect must be met with similar scepticism. People do not risk their own safety just because they know that other people will have a legal duty to rescue them if an accident occurs. In fact, there are clear indications that their own safety is one of the few considerations - if not the only consideration - that keep people from acting hazardously.⁷⁸ Moreover, most legal systems accept that by taking unreasonable risks, one may incur liability in negligence towards the person who comes to the rescue.⁷⁹ Thus, liability in negligence provides an additional deterrent against hazardous behaviour. The suggestion that more potential victims would indulge in hazardous activities is, again, highly speculative. But even if we accept that this is true, the amount of potential rescuers would rise correspondingly. For every extra potential victim that comes to the beach, is also a potential rescuer.

When Posner and Landes predict the third side-effect, that rescuers would no longer enjoy public recognition for their altruism, we must wonder what kind of liability rule they had in mind. Thus far, not a single proposal for a legal duty to rescue entails a standard of care that is higher than that of the ‘reasonable man'. A rule of liability for nonfeasance would at the most require a person to take affirmative action if it would be unreasonable not to do so. It would not require acts that are beyond the call of moral duty, acts of heroism or acts of sacrifice.⁸⁰ In other words: it would not require the kind of acts that deserve special praise and public recognition for being truly altruistic. Hence, the public recognition for jumping into a sea with treacherous tidal currents in order to rescue someone, for going into a burning house to save a cat and even for giving large donations to charities will remain unchanged, because these acts are, and would remain, ‘beyond the call of duty'.

As for the fourth and last side-effect, it is conceivable that imposition of liability in cases of nonfeasance would bring with it an increase in administrative costs, especially in cases of multiple potential defendants. The reason for this increase in costs is not, as Posner and Landes seem to suggest, that it would be difficult for the victim to identify a defendant. In cases with multiple potential defendants (e.g. the case of the crowded beach) it is relatively easy for the victim to identify a defendant, since there are so many of them. However, for the defendant bringing an action under the Civil Liability (Contribution) Act 1978, it would indeed be difficult to identify all the other potential defendants, as well as to establish the contribution that each of them would have to pay. Yet, this is not a side-effect that is particular to cases of nonfeasance. The same problem, and similar costs, occur in cases of feasance involving multiple potential defendants.⁸¹

We must therefore conclude that the economic arguments put forward by Lord Hoffmann, by Honoré and by Posner and Landes amount to very little. It is here submitted that in general the imposition of affirmative duties would promote efficiency as much⁸² (or as little) as the existing negative duties do. Thus, economic analysis does not offer a valid argument for a general denial of liability for nonfeasance.

III:5. The law does not reward the performance of affirmative duties

As was indicated in the introduction to this thesis, the law relating to ‘negotiorum gestio' falls outside of the scope of this thesis. Hence, Lord Hoffmann's statement that ‘English law does not reward someone who voluntarily confers a benefit to another'⁸³ will here remain unchallenged. However, even if we assume this statement to be correct, we need not accept it as an argument against the imposition of affirmative duties. After all, the performance of many negative duties is not rewarded by English law either. Moreover, the mere absence of a remedy in one area of the law does not necessarily provide sufficient justification for the absence of a remedy in another area of the law; it may be that both areas of the law need to be reconsidered.

III:6. Nonfeasance does not ‘cause' harm

Another argument that has been used to exclude liability for nonfeasance is that it is difficult to see how nonfeasance can be the cause of any loss.⁸⁴ In the case of positive action the defendant has made the position of the plaintiff worse. Yet, in the case of inaction the defendant has merely failed to confer a benefit to the plaintiff.⁸⁵

However, there are also instances of positive action in which the defendant merely prevents a benefit from being conferred to the plaintiff.⁸⁶ In these cases we do not seem to have any problem with regarding the defendant's behaviour as a ‘cause' of the plaintiff's loss. Let us presume that the Priest and the Levite did not walk by, but waited for the Samaritan and convinced him not to help the victim. It is hardly debatable that the actions of the Priest and the Levite would constitute a ‘cause' of the victim's loss. Yet is it not true that the Priest and the Levite merely prevented a benefit from coming the victim's way? The opposite view would be that ‘convincing somebody to abstain from helping a victim' actually makes the position of the victim worse. But if this is true, why should ‘convincing oneself to abstain from giving aid' be regarded as any different?⁸⁷

The problem that has often been felt with regarding nonfeasance as a ‘cause', originates in our understanding of the word ‘cause' as implying a positive interference, not a mere inaction. Historically, the notion of ‘movement' has been central to our concept of causation.⁸⁸ Honoré explains:

(...) the reason why movements feature so prominently in our assessment of responsibility is that we have a picture of a world as a matrix into which, by our movements and especially our manipulation of objects, we introduce changes. By moving their bodies people change the world, though they may not know or foresee exactly what changes will come about. It is inherent in this world-picture that positive acts count as interventions and attract at least a minimal responsibility for their outcomes. On the other hand, non-movement is prima facie not an intervention in the world, so that the agent is not responsible for its consequences.⁸⁹ (Footnotes omitted)

These days, the legal concept of causation is no longer completely governed by the notion of ‘movement'. If and when we find that the defendant had an affirmative duty to act, we have, at least in England, no special difficulty with regarding the defendant's omission as the ‘cause' of the loss.⁹⁰ Thus, a wounded man's refusal to see a doctor must be regarded as a cause of his death.⁹¹ Likewise, a borstal officer's failure to supervise some borstal trainees is treated as a cause of the damage that the escaped trainees inflict on someone's yacht.⁹² Honoré continues his explanation:

(...) superimposed on this picture of the world is a refinement of it, which treats disruptions of the normal course of events as similar to interventions which bring about change, though the change is not now a change in the existing state of affairs but rather in the normal or expected sequence of events. The regular course of things may take the form not only of recurrent natural events, but of regular human conduct such as clearing the street when the snow stops falling. If this regular or expected conduct fails to occur, the failure then counts not, as the first world-view would suggest, as a non-intervention but rather, on a wider view of what amounts to change, as a sort of intervention in the world.⁹³

Thus, if the local government suddenly stops cleaning the streets and collecting the garbage, there can be no doubt that the refusal to do so is a legal cause of the resulting chaos. However, the decision of the House of Lords in East Suffolk Rivers Catchment Board v Kent⁹⁴ has been the source of some confusion. Owing to a very high tide, a breach was made in a sea wall as a consequence of which the respondent's land was flooded. The appellants exercised their statutory powers to repair the wall but did the work so carelessly that the flooding continued for 178 days. It appeared from the evidence that by the exercise of reasonable skill the breach in the wall might have been repaired in 14 days. If Honoré's understanding of causation had been applied in this case, the appellants' failure to repair the wall properly would doubtlessly have been regarded as the cause of the continued flooding. Honoré would have regarded the 164 extra days of flooding as a ‘change in the normal or expected sequence of events'. However, the House of Lords by a majority rejected the respondent's claim for damage. Parts of the judgments in this case seem to suggest that the reason why the respondent's claim is rejected, is because his loss was not caused by the appellants' incompetence. Viscount Simon argued:

In the present case, the damage done by the flooding was not due to the exercise of the appellant's statutory powers at all. It was due to the forces of nature, which the defendants, albeit unskillfully, were endeavouring to counteract.⁹⁵

Similarly, Lord Thankerton's judgment read:

(...) I have felt throughout that the only real question in this appeal relates to causation, and the difficulty arises from the fact that, prior to the intervention of the appellants, a source of damage had been already created by the action of the flood waters (...) The point on which admittedly the success or failure of the appeal depends is whether the failure of the appellants to close the breach sooner by an efficient method can be held as the causa causans of the damage which accrued during these extra days, to the exclusion of the flood breach as a cause of the damage. However, it was in fact still the action of the water, rendered possible by the original breach, which caused the damage during these days, and failure to stop such action of the water cannot alter the fact that it is the water coming through the breach which causes the damage.⁹⁶

If these judgments were correct, failures to remove an already present source of damage would not constitute a legal cause. A wounded man's refusal to see a doctor would not be a cause of his death, the local governments failure to clean the streets would not constitute a cause of the resulting mess and a doctor's failure to diagnose poisoning would not be regarded as a cause of the patient's death.⁹⁷ These implications are unacceptable, and the conclusion that East Suffolk Rivers Catchment Board v Kent cannot be supported on the grounds put forward by Viscount Simon and Lord Thankerton is unavoidable. The question in this case was not whether the appellant's incompetence had caused the damage resulting from the 164 extra days of flooding. The causal connection was undebatable.⁹⁸ The real question was whether the appellant had an affirmative duty to abate the flooding or merely a negative duty not to make it worse.⁹⁹ The recent judgment of the Court of Appeal in Capital and Counties v Hampshire CC¹⁰⁰ must be understood in the same way. In one of three separate cases a firebrigade arrived at the scene of an explosion and satisfied themselves that all the fires had been extinguished. They left the scene without inspecting the plaintiff's premises. A fire later broke out, which severely damaged the plaintiff's premises. The reason why the plaintiff's claim was rejected is not that the firebrigade's failure to inspect the plaintiff's premises was not a cause of the damage. The reason is that a firebrigade, in the opinion of the majority of the Court of Appeal, merely had a duty not to cause additional damage by their action, not an affirmative duty to extinguish the fire.

The confusion between issues of ‘causation' on the one hand and issues of ‘duty' on the other is hardly surprising. By incorporating into our ‘picture of the world', as Honoré calls it, not only the ‘existing state of affairs' but also ‘the normal or expected sequence of events',¹⁰¹ we are asking for trouble. For what we ‘expect' of people is often - but not always¹⁰² - the same as their legal duty. Atiyah summarizes the issue most lucidly:

Many people feel instinctively, if irrationally, that one can only treat nonfeasance as a ‘cause' if one first starts by assuming that there was a duty to act. (...) the way we view the causation issue depends on whether we think that the defendant ought morally to have done something, and whether we think that this moral duty ought to be translated into a legal duty. This is a matter of legal policy and does not depend on any objective distinction between misfeasance and nonfeasance.¹⁰³ (Footnotes omitted)

III:7. Nonfeasance is less culpable

In his already quoted tribute to Atiyah, Honoré put forward a powerful defence of the ‘acts-and-omissions doctrine'.¹⁰⁴ In general, contends Honoré, omissions are less culpable than acts that bring about the same or similar outcome, other things being equal. Thus, hitting someone is worse than not preventing someone from being hit.¹⁰⁵ However, says Honoré, omissions that violate a distinct duty that we owe to other people or associations are usually on a level with positive acts which violate those duties.¹⁰⁶ For a nurse it is usually just as culpable to give the wrong medicine, as it is to give no medicine at all.¹⁰⁷ And even if there is a distinct duty, it is sometimes worse to act than to omit:

Take the problem of litter in the streets. If a person is employed by the local council to keep the streets clean, it can make no noticeable difference to the appearance of the streets whether he drops a wrapper in the street himself or fails to pick up a wrapper dropped by someone else. The outcome in each case is that there is one more wrapper in the street than there would otherwise have been, and the urban scene is to that extent less tidy. It is nevertheless probably worse for him to drop a wrapper than to fail to pick one up, since he is in that way adding to the very conditions he is employed to remedy.

Honoré's contentions are hardly debatable. All other things being equal, feasance is often regarded as being more culpable than nonfeasance, not in the least because many acts imply an additional inaction. When we say that ‘dropping a wrapper is worse than not picking someone else's wrapper up', what we mean to say is that ‘dropping a wrapper and not picking it up' is worse than ‘not picking up someone else's wrapper'. And if we say that - presuming that we have a duty to protect one another - hitting someone is worse than not preventing someone from being hit by a third person, this must be true, because by hitting we break two duties at once: the negative duty not to hit and the affirmative duty to protect.

Yet, accepting Honoré's contentions, do we also have to accept that, as he puts it, ‘the acts-and-omissions doctrine is broadly sound'? Surely just the fact that nonfeasance is usually less culpable than feasance - all other things being equal - cannot justify a general denial of liability for nonfeasance? The harmful acts of a rookie doctor are usually less culpable than those of an experienced doctor. Likewise, unintentionally harmful acts are usually less culpable than acts committed with the intent to cause harm. Yet neither against the rookie doctor, nor against the unintentional wrongdoer do we deny the victim a claim.

If all other things are not equal, the culpability of nonfeasance can very well exceed the culpability of feasance. If an inaction causes more harm than an action, the inaction is often seen as more culpable. There can be no doubt that leaving a man to drown where one could easily throw him a rope is more culpable than accidentally breaking someone's tea-pot. In rescue cases it may even happen that, although the act causes more harm than the inaction, the inaction is still regarded as more culpable than the action. Imagine that the Good Samaritan, trying to do his very best to lift the wounded man ‘on his beast', slipped and dropped the man, causing him to die earlier than he would have done if the Samaritan would have left him alone. The act of the Samaritan has caused more harm than the Priest's and the Levite's inaction. However, it is not unlikely that the behaviour of the Samaritan, who after all did his very best to save a fellow human being, will be regarded as the less culpable. Thus, if culpability is to be the test for imposing liability, there is no (good) reason to make a general distinction between cases of feasance and cases of nonfeasance.¹⁰⁸

IV. Liability for nonfeasance in continental european private law

IV:1. Historical background

Whether directly or indirectly, many provisions in Western European codifications of private law have been influenced by the corresponding provisions in the laws of the Roman Empire. The Romans, too, were confronted with cases of damage caused by nonfeasance. The relevant question under Roman law was whether the Lex Aquilia, the Roman statute establishing liability for killing slaves and cattle and for causing damage to goods, was applicable to cases of damage (or death) caused by nonfeasance. Ulpian treats a case in which a stoker-slave falls asleep at the furnace, as a consequence of which the house burns down.¹⁰⁹ If the slave has lit the fire in the furnace himself, there seems to be no doubt in Ulpian's mind that liability arises. The lighting of the fire, however inculpable in itself, has, to use English terminology, raised a duty of care to control the fire. More difficult is the case in which one man lights the fire, but another watches over it carelessly. Ulpian gives the following view, as recorded in emperor Justinian's Corpus Iuris Civilis, D. 9.2.27.9:

(...) but if one man lit the furnace and another watched it negligently, will the one who lit it be liable? For the one who watched it did nothing, while the one who lit it in the proper way was not at fault. What then are we to say? I think that an actio utilis lies against both the man who fell asleep at the furnace and the man who negligently watched it; nor can anyone say that the man who fell asleep was overtaken by a natural human weakness, because his duty was either to put out the fire or to make certain that it should not spread.¹¹⁰

In 19th century Germany, in parts of which the Corpus Iuris Civilis still had force of law, Ulpian's words were much discussed. Heavily relying on this text, Pernice came to the conclusion that the Romans did not draw a general distinction between feasance and nonfeasance; if all other requirements set in the Lex Aquilia were fulfilled, the careless inactor, like the careless actor, would incur liability:

The only question that is always asked is whether there is a substantial loss on the one hand and careless behaviour as its cause on the other hand. When a stone falls from a cart and damages an animal, this may have occurred because the stones were badly placed, or because the necessary plank was not pushed in its place. One is an action, the other an inaction. This however, does not make a difference; what matters is that it was bad placement, that the necessary plank was not in its place.¹¹¹

If Pernice was right, the Roman law position on liability for nonfeasance was very much in agreement with the position that the natural law movement adopted. The natural lawyers abandoned the distinction between feasance and nonfeasance for all practical purposes and subjected liability for both feasance and nonfeasance to the same requirements.¹¹² Pernice's opinion was, however, not the opinion of the majority of writers in the 19th century. Although his important contemporary Dernburg agreed with Pernice,¹¹³ the majority followed Windscheid, who thought that the Romans generally denied Aquilian liability in cases of nonfeasance.¹¹⁴ According to Windscheid, there was one exception:

(...) one will only be obliged by inaction, if action had become required due to an earlier or simultaneous act.¹¹⁵ (Footnotes omitted)

Thus, according to Windscheid, one only had a ‘duty to act' if one's earlier or simultaneous behaviour had created the need for intervention. Hence the liability of a man who carefully lights a fire, but does not control it properly and the liability of a doctor who performs a skillful operation but neglects the aftercare.¹¹⁶

With distinguished scholars like Pernice, Dernburg and Windscheid having such different views on the subject,¹¹⁷ it is hardly surprising that it proves difficult to establish the true nature of liability for nonfeasance under Roman law. Although one cannot be absolutely certain, it is now generally assumed that the Lex Aquilia initially did not apply to cases of nonfeasance.¹¹⁸ This assumption is based on Gaius' Institutes, in which it is written that there was only a direct action under the statute when the defendant had done the damage with his own body, corpore suo.¹¹⁹ When the damage was not a direct consequence of the wrongdoer's physical behaviour, the Lex Aquilia did not apply. Thus, it appears that there was initially no remedy against someone who persuaded a slave to go down a dangerous well where he was killed,¹²⁰ nor against him who killed a slave by pushing a third party into him.¹²¹ If the Lex Aquilia originally required that damage be done corpore suo, this requirement must have blocked liability in cases of nonfeasance. For the very notion of nonfeasance implies the absence of direct physical contact between wrongdoer and victim.¹²²

No matter what the original scope of the Lex Aquilia may have been, over the centuries it came to apply also to cases in which the damage was not done corpore suo. If the direct action on the Lex did not lie, one was granted a decretal action to the same effect;¹²³ an actio in factum or utilis.¹²⁴ It is likely that along the same lines decretal actions have been developed for certain instances of loss caused by nonfeasance. By the time of Emperor Justinian's reign (6th century A.D.) we can distinguish at least three types of cases in which liability for nonfeasance could arise:

1) The wrongdoer had created a risk and failed to control it; hence the liability of the doctor who neglected the aftercare,¹²⁵ of the stoker who fell asleep¹²⁶ and of the pruner who failed to warn for falling branches.¹²⁷

2) The wrongdoer had been trusted with the specific task of doing what he failed to do; hence the liability of the man who did not light the fire in the furnace, but was supposed to watch it and failed to do so¹²⁸ and of the sailors who failed to control their ship in order to prevent a collision.¹²⁹

3) The wrongdoer was in a special position of control over the person who directly inflicted the harm; hence the liability of the master who knew that his slave intended to wound or kill someone, but failed to prevent it.¹³⁰

It is worth noting that the Roman term ‘neglegentia', from which our ‘negligence' derives, originally appears to have denoted a (negligent) failure to perform a specific duty; nonfeasance. For active misconduct the term culpa was used.¹³¹

IV:2. The law in Germany

Since the introduction of the German Civil Code (BGB) in the year 1900, the Roman texts have lost force of law in Germany. The liability rules for the German equivalent of ‘negligence' can now be found in § 823 BGB:

§ 823 [Duty to compensate for damage]

(1) A person who, wilfully or negligently, unlawfully injures the life, body, health, freedom, property or other right of another is bound to compensate him for any damage arising therefrom.

(2) The same obligation is placed upon a person who infringes a statute intended for the protection of others. If, according to the provisions of the statute, an infringement of this is possible even without fault, the duty to make compensation arises only in the event of fault.¹³²

Thus, the German legislator did not appear to differentiate between liability for feasance and for nonfeasance.¹³³ However, when interpreting the requirement of ‘unlawfulness' in section (1) of § 823, the courts have taken a different approach towards actions on the one hand and inaction on the other.

If one of the protected interests of section (1) has been infringed through a positive act, this act is presumed to have been unlawful. The only way for the defendant to rebut this presumption is by invoking a special justification for his act (e.g. self-defence).¹³⁴

However, if one of the protected interests of section (1) has been infringed by inaction, this inaction is not automatically presumed to have been unlawful. Inaction is only regarded as unlawful if it violated a ‘duty to act' (Pflicht zum Handeln).¹³⁵

During the second half of this century the different approach taken by the courts towards cases of feasance and of nonfeasance has been heavily criticized by many influential authors.¹³⁶ According to these authors one should not presume positive acts to be unlawful, merely on the ground of their harmful result. As with inaction, positive acts should only be regarded as unlawful if they violate a pre-existing duty.¹³⁷ Thus, according to this view, there should no longer be a difference in approach towards cases of nonfeasance and cases of feasance: both actions and inaction should only be regarded as unlawful if a duty is violated. In theory this development could have serious implications for cases of nonfeasance.¹³⁸ In practice however, the courts have been - and are likely to stay - much more reluctant to recognise the existence of affirmative duties than of negative duties.¹³⁹

German law roughly distinguishes between three types of affirmative duties. Firstly, the defendant may, explicitly or implicitly, have undertaken a responsibility for the well-being of others. Thus, contractual parties have duties to act in good faith, regardless of the exact wording of the contract. In the absence of a contract, too, an undertaking may give rise to an affirmative duty. Thus, a neighbour who promises to look after a child or dog, can be held liable for not doing so, even though his promise may not have been contractually binding.¹⁴⁰

Because their occupations put them in a special position of responsibility, specific protective duties are imposed on nannies, doctors, the police, sportsteachers etc. Similar duties to ensure the safety of others, are imposed on those who have a family relationship or a relationship of love and affection with the person in need of protection. Thus, although the Levite may not be required to render assistance to everyone, he does have a duty to aid his close relatives, his spouse or his close friends.¹⁴¹

Where the first two types of affirmative duties arise out of the special responsibility the defendant has for the safety of the endangered person, a third type arises when the defendant is in some way connected with the danger itself. Examples are the duty of owners of dangerous premises, substances or animals to prevent others from being harmed. Under this heading one usually also includes the duty arising from a ‘previous act' (‘aus vorangegangenem Tun'); those who, no matter how inculpably, create a danger, have a duty to prevent the danger from harming others. Thus, like under Roman law, someone who lights a fire will be liable under German law if he fails to control the fire.¹⁴²

Most of the above affirmative duties have been expressly recognised by the courts through creation of the so-called ‘Verkehrspflichten' and ‘Verkehrssicherungspflichten'.¹⁴³ These terms are not easily translated. For present purposes it will suffice to say that Verkehrs(sicherungs)pflichten are legal duties, that require someone who, within the scope of his responsibility, establishes a source of potential danger or allows it to remain, to protect the interests and rights of others against such danger.¹⁴⁴ Very soon after the introduction of the BGB, the German courts started recognising the existence of Verkehrs(sicherungs)pflichten.¹⁴⁵ The following case was considered by the Reichsgericht in 1903:¹⁴⁶ On a cold winter evening the plaintiff had lost his balance when walking on slippery stone steps in his hometown. He brought a claim for damages against the local authorities, arguing that they ought to have gritted the steps. The Reichsgericht found for the plaintiff, establishing that the local authorities had a general duty to take care¹⁴⁷ that the steps were safe for the public to walk on. This duty, argued the Reichsgericht, arose from the local authorities' ‘power of disposition'¹⁴⁸ over the steps. Although the local authorities in this case were the legal owners of the steps, it was explicitly stated that this duty of care was not limited to legal owners:

The main focus does not lie on the right of ownership as such; those who possess other rights that entail power of disposition may have a similar duty to take care. The question is, whether the person concerned, by dint of his actual and legal relationship with the property, ought to take any kind of care towards third parties when disposing of this property, respectively when handling or using it in legal relations.¹⁴⁹

The amount of Verkehrs(sicherungs)pflichten that is recognised by the courts has increased immensely over the years. In the majority of cases they entail an affirmative duty to act, and arise either out of the special relationship the defendant has with a particular thing or piece of property or out of the defendant's special position of responsibility.¹⁵⁰ Examples are the duty of local authorities to ensure the safety of the roads,¹⁵¹ the duty of hosts to ensure the safety of their guests on their premises,¹⁵² the duty of care that road-users owe to other road-users¹⁵³ and the duty of those whose trade or profession may affect others, to use particular care and skill in exercising their trade or profession.¹⁵⁴ What standard of care the Verkehrs(sicherungs)pflichten impose, depends on the particular circumstances of each individual case. The most important factors that are taken into account are the likelihood of a particular risk to materialise, the serious nature of its consequences if the risk does materialise and the difficulty of preventing the risk from materialising.¹⁵⁵

Another possible source of affirmative duties are the statutory provisions. If a statutory duty is violated, as a result of which someone is harmed, the victim can bring a claim under section (2) of § 823BGB. This claim will only be successful if the statute was intended to protect others (if the statute is a so-called ‘Schutzgesetz'). Hence, if someone forgets to light his carlights at night, as required by § 17 of the German Traffic Regulations (Straßenverkehrsordnung), he will be liable to those who suffer harm as a consequence of his omission.¹⁵⁶ Of particular interest in this context is § 323c of the German Criminal Code (StGB), a provision originating from the days of the Nazi regime:¹⁵⁷

§ 323c [Failure to render assistance to those in need]

Whoever fails to render assistance in case of accident, common danger or emergency, although such assistance was needed and could have been expected from him under the circumstances, especially since he could have rendered it without placing himself in significant danger and without violating any important duties, shall be punished by up to one year's imprisonment or by fine.¹⁵⁸

After the collapse of the Nazi regime, the question arose whether § 323c StGB¹⁵⁹ was still in force. The Großer Senat für Strafsachen answered this question in the affirmative, stating that the duty to render assistance in emergencies is a ‘moral obligation that has existed from time immemorial' and adding that ‘aid to neighbours in need was, in particular, always an imperative command of Christian doctrine'.¹⁶⁰

Thus, German law continued to impose a criminal law duty to rescue on those who can do so without the risk of harming themselves. At first sight, this statutory provision would seem to be ‘intended for the protection of others', in the sense of § 823 (2) BGB. Indeed, it is likely that the German legislator in 1935 was of the opinion that the criminal duty of § 323c StGB would give rise to a private law duty under § 823 (2) BGB.¹⁶¹ However, legal opinion has changed and it is now the predominant view that § 323StGB was not intended to protect individual persons but only society as a whole.¹⁶² This legalistic argument does seem peculiar, especially considering the following words by Haager:

(...) It is sufficient, that a statutory provision mainly pursues another goal, but simultaneously seeks to protect certain individual interests. (...) Only statutory provisions that are exclusively intended to protect the order of society, the entire state as such, its external intactness and its internal constitution and administration, fall completely outside of the context of such ‘Schutzgesetz', like the criminal provisions on high treason and treason, on resistance against state authority, on breach of the constitution etc.¹⁶³

It is hard to imagine that § 323c StGB was exclusively intended to protect the ‘entire state as such' and not (at least) also certain individual interests. It is likely that there are other, more profound reasons that lie behind the denial of a civil duty to rescue under § 823 (2) BGB. In a judgment in 1988, the Oberlandesgericht of Frankfurt used an argument which strongly resembles the ‘Why pick on me?' argument:

The criminal provision (of § 323c StGB, JSK) is not a ‘Schutzgesetz' in the sense of § 823 (2) BGB, because there is no reason to hold the person who omits to render assistance liable in the same way as the person who inflicted the harm.¹⁶⁴ (Footnotes omitted)

Even in the absence of a Verkehrspflicht or a Schutzgesetz, liability for nonfeasance may arise if the defendant's refusal to act was intentional and contra bonos mores.¹⁶⁵ § 826BGB reads:

§ 826 [Intentional damage contra bonos mores]

A person who, in a manner contra bonos mores, intentionally causes damage to another, is bound to compensate him for the damage.¹⁶⁶

‘Intention' is taken to exist, if the defendant was aware of the harmful nature of his behaviour. It is not necessary that the motive of his behaviour was to cause harm.¹⁶⁷ However, not every intentional refusal to act is regarded to be contra bonos mores. Schäfer, for example, submits that an intentional refusal to render assistance to someone in need is not contra bonos mores, unless such refusal will lead to death, serious bodily harm or a dangerous road accident and unless rendering assistance is easy and can be done without disregarding one's own important interests.¹⁶⁸

There are relatively few cases involving nonfeasance in which successful claims have been brought under § 826 BGB. Contractual parties have been held liable for intentionally failing to disclose relevant information in the course of the negotiations.¹⁶⁹ Also, claims have been allowed against persons who noticed that their signature had been forged on an official document, but failed to give a warning to that effect.¹⁷⁰

Apart from the general provisions of §§ 823 and 826, the BGB imposes certain specific affirmative duties in the §§ 831 to 838. Thus, if one has failed to exercise the proper supervision or control, one will be liable for damage caused by one's employees, one's animals and the buildings or structures on one's land.

Finally, affirmative duties to act can arise from contractual and pre-contractual relationships. Where English law only imposes contractual duties if a contract has been concluded, German law recognises that contract-type duties can arise from pre-contractual relationships. Thus, a contract-type claim may lie in the event of carelessness of the defendant during the pre-contractual phase (‘culpa in contrahendo'), even if no contract is ever concluded between the parties.¹⁷¹ Also, German law recognises that contracts may have protective effects towards third parties (Schutzwirkung für Dritte).¹⁷² The famous ‘vegetable leaf case'¹⁷³ provides an example of both doctrines at work at the same time. The plaintiff went with her mother to the defendant's self-service store. While her mother stood at the till, the plaintiff slipped over a vegetable leaf. She sued the defendant for breach of his duty to provide safe access. The Bundesgerichtshof had no doubt that if the plaintiff's mother had been injured in the same way as her daughter, the defendant would have been liable for culpa in contrahendo.¹⁷⁴ But even the plaintiff, who did not have a pre-contractual relationship with the defendant, was allowed a claim on the basis of the Schutzwirkung of her mother's pre-contractual relationship with the defendant:

It accords with the long-standing case-law of this Senate in particular that in special circumstances even bystanders who do not themselves participate in a contract are included in the protection afforded by it (...)¹⁷⁵

Thus, German law imposes affirmative duties under a variety of headings: § 823BGB in connection with the Verkehrspflichten, § 823 (2)BGB and its Schutzgesetze, § 826BGB, §§ 831 to 838BGB and finally the law of contract. It must however be noted, that, in the absence of express statutory provisions, the courts have thus far only imposed affirmative duties, when the defendant was in some way connected with the endangered person or with the danger itself. If a case like the one of the Good Samaritan were to be considered now, it seems unlikely that the German courts would have held the priest or the Levite liable for the harm that was caused by their inaction.

IV:3. The law in France

Until the seventeenth century, the French law of delict had been a mixture of Roman law, canon law and local laws and customs.¹⁷⁶ There was no general rule of liability, neither was there a clear distinction between the compensatory and the punitive function of awarding damages. In 1689 Domat published his ‘Les loix civiles dans leur ordre naturel', a most influential work which would eventually provide the fundamentals of the modern French law of delictual liability. Central to Domat's understanding of delictual liability was the notion of faute. Domat distinguished three types of fautes: those that consist of a crime or offence, those that are a breach of agreements and those that are neither, ‘like when one thoughtlessly throws something out of a window and spoils someone's clothes, when animals that are not properly looked after cause damage, when one causes a fire through carelessness, or when a building under the threat of falling down is not repaired and collapses on another building and causes damage'.¹⁷⁷ Delictual liability is concerned with only the latter type of fault. When under a duty to act, Domat continued, a defendant could be held liable for a failure to prevent harm:

Those who, being able to prevent harm from occurring, when a certain duty committed them to prevent this harm, failed to do so, may be held liable depending on the circumstances. Thus, a master who sees and allows the infliction of harm by his servant, is held liable.¹⁷⁸

Domat's ideas were adopted, expanded and improved by Pothier, the most influential French scholar of the eighteenth century. By the end of that century, the French legislator decided to draft a liability rule that was meant to be universally and eternally applicable.¹⁷⁹ Domat's theories were clearly the source of inspiration of the articles 1382 and 1383 of the French Civil Code (C. civ.), which came into force in 1805:¹⁸⁰

1382 Any behaviour whatever of man which causes damage to another obliges him by whose fault (‘faute', JSK) it occurred to make reparation.¹⁸¹

1383 Each one is liable for the damage which he causes not only by his own behaviour but also by his negligence or imprudence.¹⁸²

When defining the notion of faute, the French usually distinguish between faute de commission (fault by commission) and faute d'abstention (fault by abstention).¹⁸³ With regard to the latter, a further distinction is made between abstentions dans l'action (abstentions within an action) and abstentions pures et simples (pure abstentions).¹⁸⁴ When we discussed a similar distinction made by Fleming - between ‘real' nonfeasance and nonfeasance that is part of a larger activity - in the second chapter of this thesis, we concluded that such a distinction is probably not very useful. The same view is adopted by Tunc and the brothers Mazeaud in their influential Traité Théorique et Pratique de la Responsabilité Civile.¹⁸⁵ However, the courts have continued to use the distinction.

In cases of abstentions dans l'action, the courts do not seem to have had any problems with imposing liability.¹⁸⁶ The so-called ‘Branly-case' provides what is probably the most well known example of an abstention dans l'action; Professor Turpain, the author of an article with the title ‘Historique de la T.S.F' (‘History of Wireless Telegraphy') had omitted to mention the name and the works of the plaintiff, Edouard Branly. The French Cour de Cassation decided:

That the court should have investigated whether Turpain, when writing a history of wireless telegraphy in which the name and the works of Branly were willfully omitted, had behaved like a prudent writer or historian, aware and conscious of the duty to be objective that rested upon him.¹⁸⁷

Thus, abstentions dans l'action are measured against the same standard as actions are:¹⁸⁸ i.e. the French equivalent of the ‘reasonable man standard'.¹⁸⁹ It is worth noting that the Cour de Cassation was apparently of the opinion that imposition of liability was not too much of an invasion of Professor Turpain's freedom of speech and expression.¹⁹⁰ On the same principle those who, however blamelessly, create a risk, have an obligation to take reasonable care to prevent this risk from materialising.¹⁹¹ Hence the liability of a railway company for failing to provide the appropriate warning signals at a level-crossing.¹⁹²

In the area of abstentions pures et simples, too, the Branly-case would play a vital role. Initially, the courts showed a reluctance, similar to that in Germany, to impose liability for ‘pure' nonfeasance. In a 1924 judgment, the Cour de Cassation held that:

(...) if everyone is liable for his negligence, an omission can only bring about liability if the person who is held liable had an obligation to do that which was omitted.¹⁹³

At the time, this judgment was interpreted as meaning that liability for nonfeasance would only be imposed if it violated an affirmative duty that had been created by statute or at least by custom.¹⁹⁴ Obviously, this requirement restricted the scope of liability for abstentions pures et simples.¹⁹⁵ The main reason for the reluctance of the courts is said to have been that one should be allowed at least the freedom not to act. Additionally, it was doubted whether a causal connection could exist between nonfeasance and harm.¹⁹⁶ In the 1940's, the recognition of affirmative duties by the courts was accelerated by a development in the area of criminal law. Unlike the Germans,¹⁹⁷ the French regard the commission of any criminal offence which causes harm to others as a ‘faute' for the purpose of articles 1382 and 1383 C. civ. No further inquiry into the intended scope of the criminal provision is necessary. This concept is referred to as the ‘unity of criminal and civil faults'.¹⁹⁸ Thus, when the legislator introduced a duty to act under criminal law, this was bound to have an immediate effect on the interpretation of ‘faute' in the law of negligence.¹⁹⁹ The first two paragraphs of article 63 of the French Criminal Code (C. pén.), after it was amended in 1945,²⁰⁰ read:

Article 63 Any person who, by his immediate action and without danger to himself or others, could have prevented either a felonious act or a misdemeanor against the person, willfully fails to do so, shall be punished by jailing for no less than one month nor more than three years and by fine from 12,000 to 500,000 francs, or either punishment, unless more severe punishments are provided by this Code or special law.

Any person who willfully fails to render or to obtain assistance to an endangered person when such was possible without danger to himself or others, shall be subject to like punishments.²⁰¹

The concept of ‘unity of criminal and civil faults' meant that the courts could translate the criminal duties imposed by article 63 C. pén. into private law duties. In 1947, the Tribunal Correctionnel d'Aix decided to award 25,000 francs in damages to the plaintiff, who had nearly drowned when he fell through ice into a deep canal. The defendant, who was the plaintiff's father-in-law, had walked away from the scene after refusing to assist a third person who tried to rescue the plaintiff by handing him an iron bar to which he might cling.²⁰²

In a later case, a doctor who was held liable by the Cour d'Appel of Paris for failing to assist a patient in need, argued that article 63 C. pén. was merely intended to protect the public interest, and should not give rise to private law claims. The Cour de Cassation dismissed the submission:

(...) if the provisions of article 63, 2nd paragraph of the Criminal Code concern the general interest of society and repress the disturbance caused by a delictual abstention to law and order, they also aim to protect private interests;²⁰³

One noteworthy consequence of the civil liability arising from article 63 C. pén., is that members of the medical profession are only allowed to strike in non-emergency cases.²⁰⁴

It should be mentioned that a new Criminal Code has come into force in 1994. The former article 63 C. pén. is now incorporated in a larger section, which increases the penalties imposed by article 63 C. pén. and introduces a further affirmative duty to assist in the case of a natural disaster, even when there is no immediate danger to any person.²⁰⁵

The decisive breakthrough towards a complete recognition of liability for nonfeasance came with the aforementioned Branly-case. Technically, the Branly-case was only concerned with an abstention dans l'action and not with abstentions pures et simples. The Cour de Cassation however, probably strongly influenced by the works of the brothers Mazeaud,²⁰⁶ did not draw such distinction:

(...) fault within the articles 1382 and 1383 C. civ. can consist of an abstention as well as a positive act; (...) an abstention gives rise to liability of the person responsible - even if there was no malice and intention to harm - when the act that was omitted should have been done either by virtue of a legal, statutory or contractual obligations, or by virtue of the demand for objective information within a profession, particularly when a historian is concerned.²⁰⁷

If we were to interpret this judgment to the letter, it would not cause a big change in the area of liability for nonfeasance.²⁰⁸ In the majority of the subsequent caselaw, however, the courts have proceeded on the presumption that the Branly-case represents a breach with the distinction between liability for feasance and for nonfeasance.²⁰⁹ At present, legal theory appears to have adopted a similar position. Whenever they think it appropriate the courts should be able to impose legal duties, whether affirmative or negative.²¹⁰

In practice however, the courts appear to adopt a rather cautious approach towards imposition of affirmative duties.²¹¹ In particular, the personal characteristics of the defendant are taken into account. Thus, the courts are reluctant to impose affirmative duties on children.²¹² Also, the courts are not likely to hold a defendant liable, who did not understand that his intervention was necessary or who was physically or mentally unable to render the required assistance.²¹³ If, on the other hand, the nonfeasance was malicious and harm was intended, caselaw and legal theory agree that liability should arise.²¹⁴

Finally, it should be noted that the Code civil contains express provisions imposing liability for damage caused by the persons for whom one is responsible,²¹⁵ by the things one has in one's control,²¹⁶ by the animals one owns,²¹⁷ as well as for damage caused by the collapse of one's buildings.²¹⁸

V. Liability for nonfeasance in English private law

V:1 Historical background

Central to the later medieval law of torts were the action of trespass and the action upon the case.²¹⁹ An action of trespass would almost certainly not have been available in cases of nonfeasance. For by analogy with the Roman requirement that damage be done corpore suo,²²⁰ the defendant's behaviour was only considered a ‘trespass' if it had directly harmed the plaintiff.²²¹ This requirement of ‘directness' is illustrated by Fortescue J in Reynolds v Clarke:

If a man throws a log into the highway, and in that act it hits me; I may maintain trespass, because it is an immediate wrong; but if as it lies there I tumble over it, and receive an injury, I must bring an action on the case; because it is only prejudicial in consequence, for which originally I could have no action at all.²²²

Because in cases of nonfeasance the harm is caused indirectly, the action of trespass would not have lain.

The action upon the case on the other hand, would typically have lain in cases of indirect or ‘consequential' loss. In 1762, Sir John Comyns, Chief Baron of the Exchequer, included in his Digest a title for the ‘Action upon the Case for Negligence'. ‘Negligence', rather unlike its wide modern connotations, was understood to mean ‘neglect' or ‘omission'.²²³ An action upon the case based on such an omission would have lain for ‘negligence in a man's trust' or ‘duty',²²⁴ for negligence in a man's office,²²⁵ for a man's ‘neglect in doing that, which by law he ought to do', for his ‘neglect to do that, which he has undertaken', for negligence in taking care of one's animals and for negligently keeping a fire.²²⁶ As we will see in the remaining sections of this chapter, many similar categories still feature in the modern law.

In the course of the nineteenth century the division between the action of trespass and the action upon the case was overshadowed by the development of a new tort: the tort of negligence. ‘Negligence' was no longer understood as merely referring to omissions, but came to cover cases of both acts and inaction.²²⁷ In Blyth v Birmingham Waterworks Company,²²⁸ Baron Alderson defined negligence as follows:

Negligence is the omission to do something which a reasonable man, guided upon those considerations which ordinarily regulate the conduct of human affairs, would do, or doing something which a prudent and reasonable man would not do.²²⁹

On the face of it, this definition does not indicate that there was any difference in the law's approach towards feasance and nonfeasance; apparently both types of behaviour were measured against the standard of the reasonable man. However, as Pollock was careful to point out, Baron Alderson's definition must have been based on the presumption that the party whose conduct was in question, was already under a ‘duty of taking care'.²³⁰ And it was at this ‘duty'-stage of the trial, that the judicial approach in cases of nonfeasance differed from that in cases of feasance. For acts and their results, Pollock explained, the actor was generally speaking held answerable. For mere omission he was not, unless he was under some specific duty.²³¹ An example of the judicial reluctance to impose affirmative duties can be found in Gantret v Egerton. The defendants owned land which was intersected by a canal and cuttings, and bridges ‘leading to certain docks of the defendants'.²³² Leon Gantret, who was lawfully passing over and using this land and these bridges, fell into one of the cuttings and drowned. The plaintiff claimed that the defendants had wrongfully and improperly kept and maintained their land, canal, cuttings and bridges, and had allowed them to continue to be in a condition which rendered them unsafe for persons lawfully passing along. Herschell, for the plaintiff, argued that the defendants would have to prove that they did not owe a duty of care towards the plaintiff. At this point, according to the Bar Reports, Willes J interrupted Herschell:

There is no duty to do anything, but there is a duty to abstain from doing anything that would injure.²³³

In the Law Reports, where this case appears under the name Gautret v Egerton, he is said to have added the following illustration:

No action will lie against a spiteful man who, seeing another running into a position of danger, merely omits the warning.²³⁴

Because the plaintiffs declaration did not reveal the breach of a specific affirmative duty, it was held that the plaintiff had no cause of action. As we will see in subsection V:6 (2), below, Gantret/Gautret v Egerton is no longer good law.²³⁵

V:2. The law in England at present

More than a hundred and thirty years have passed since Willes J gave his decision in Gantret/Gautret v Egerton. In that time it has been generally assumed that English law persisted in its reluctance to impose liability for nonfeasance. In the words of Markesinis: ‘it is a well-known fact that the common law, first for historical reasons and later on philosophical grounds, has taken a hostile view towards imposing tortious liability for pure omissions'.²³⁶ Nevertheless it proves surprisingly difficult to find binding authority for this no-liability rule.²³⁷

True, there have been occasional references to the rule in obiter dicta.²³⁸ Yet cases in which it is part of the ratio decidendi are scarce. One of the few exceptions, which for that reason features in almost every textbook as authority on the matter, is said to be Smith v Littlewoods, in which Lord Goff stated:

Why does the law not recognise a general duty of care to prevent others from suffering loss or damage caused by the deliberate wrongdoing of third parties? The fundamental reason is that the common law does not impose liability for what are called pure omissions.²³⁹

Although the no-liability rule was here clearly an important argument in Lord Goff's speech, it must be doubted whether Smith v Littlewoods provides binding authority. Firstly, Lord Goff was the only one of their Lordships to refer to the doctrine of no liability for nonfeasance.²⁴⁰ Secondly, Lord Goff himself went on to express his doubts about the no-liability rule, saying that it ‘may therefore require one day to be reconsidered'.²⁴¹

Encouraged by Lord Goff's latter statement Markesinis expressed in a 1989 article the hope and expectation that the no-liability rule would soon belong to the past.²⁴² Many other influential scholars too, have argued in favour of abandoning the rigid distinction between cases of feasance and of nonfeasance.²⁴³ However, in its recent decision in Stovin v Wise²⁴⁴ the House of Lords made clear that such development will not take place in the near future. In a powerful speech Lord Hoffmann not only confirmed the principle of no liability for nonfeasance but went on to defend it. The most relevant part of this speech was quoted in section III:1 of this thesis. Lord Goff and Lord Jauncey agreed with Lord Hoffmann. The minority of their Lordships, consisting of Lord Nicholls and Lord Slynn, also adhered to the general no-liability rule. After stating that the distinction between liability for acts and liability for omissions is ‘fundamentally sound',²⁴⁵ Lord Nicholls said:

The recognised legal position is that the bystander does not owe the drowning child or the heedless pedestrian a duty to take steps to save him. Something more is required than being a bystander. There must be some additional reason why it is fair and reasonable that one person should be regarded as his brother's keeper and have legal obligations in that regard. When this additional reason exists, there is said to be sufficient proximity. That is the customary label.²⁴⁶

On the facts of the case, Lord Nicholls and Lord Slynn were of the opinion that there were specific circumstances that gave rise to a duty of care.

Thus, Stovin v Wise provided new and ample authority for the no-liability rule. Only if we are under a specific duty to act, will we be held liable for our failure to act. The crucial question becomes therefore: ‘when do such specific duties arise?' Part of the answer was given by Lord Hoffmann in Stovin v Wise:

There may be a duty to act if one has undertaken to do so or induced a person to rely upon doing so. Or the ownership or occupation of land may give rise to a duty to take positive steps for the benefit of those who come upon the land and sometimes for the benefit of neighbours.²⁴⁷

In the following sections some specific circumstances that may give rise to affirmative duties will be assessed and investigated. It should at the outset be noted that the categories of circumstances that are distinguished in these sections sometimes overlap. As we will see, in many cases it is the combination of circumstances that has given rise to the imposition of an affirmative duty.

V:3. The defendant has created the risk

When the defendant has created a risk, however blamelessly, and then fails to control it, the courts tend to qualify the case as one of feasance, rather than nonfeasance. In chapter II) we criticised this tendency and discussed an example of a cardriver who failed to brake in time to avoid hitting a pedestrian. For the reasons given there, we decided to treat his failure to brake as an instance of ‘nonfeasance', rather than, as Fleming argues, ‘the element that makes his active conduct - driving - negligent'.²⁴⁸ Accordingly, we regarded the stoker who fell asleep at the furnace as an example of someone who would incur liability for nonfeasance in Roman law.²⁴⁹

Most English courts however, would regard both the cardriver's failure to brake and the stoker's failure to control the fire as active. Thus, by qualifying the defendant's behaviour as active rather than passive, the courts allow themselves to impose liability without having to address the issue of liability for nonfeasance. This technique is not unlike that of the French courts, when they distinguish ‘omissions dans l'action' from the ‘real' nonfeasance, the ‘omissions pures et simples'. It would perhaps be more realistic for both the French and the English courts to accept that the cardriver's and the stoker's conduct are instances of ‘real' nonfeasance, albeit instances that fall under an exception to the no-liability rule.²⁵⁰

There is no doubt that under English law, as under French and Roman law, creating a risk gives rise to an affirmative duty to take reasonable steps to prevent the risk from culminating in injury.²⁵¹ Authority for this proposition can be derived from a host of cases.²⁵² In Musgrove v Pandelis²⁵³ the defendant's servant started the engine of the defendant's car. For some unexplained reason the carburettor exploded and the petrol caught fire. If the servant had closed the petrol tap immediately the fire would have burnt out harmlessly. Although the servant could not be blamed for the explosion that ignited the petrol, he was held to have come under a duty to take reasonable care in preventing the fire from spreading. The defendant was held vicariously liable.

Although no express reference was made to the doctrine of liability for nonfeasance, we may conclude that creating a risk, no matter how blamelessly, gives rise to a duty to prevent the risk from culminating in h