Eighty years on, Keith Patten traces the legacy of Donoghue v Stevenson
Was there ever a snail in the ginger beer bottle? We will never know, because the most famous case in the common law world, which is 80 years old this month, went all the way to the House of Lords on a preliminary point of law and was never tried on its facts. That also leaves open the perhaps even greater evidential problem of whether Mrs Donoghue could have established a causative link between the presence of the snail and the injury for which she claimed.
Much has been claimed for Donoghue v Stevenson  AC 562,  All ER Rep 1, but one thing that it did not do was to invent the law of negligence. The idea of liability for carelessly inflicted harm, separate from contract, arising where, but only where, there was a pre-existing duty of care was developing throughout the 19th century, and was well established long before Donoghue and her friend ventured into the Wellmeadow Café in Paisley on a daytrip from Glasgow. There had even been earlier attempts to generalise the duty concept, most notably by Brett MR in Heaven v Pender (1883) 11 QBD 503, [1881-5] All ER Rep 35, but what had emerged in the run up to Donoghue was a rather narrower cause of action. Duties had been held to exist in certain specific situations, for example between an occupier and his lawful visitors, or between bailor and bailee. Most, but not all, of these duty situations arose in circumstances where there was some background contractual or quasi-contractual relationship. Almost all arose in situations where there was at least some pre-existing relationship between the parties.
Donoghue v Stevenson in the House of Lords
The editor of the law reports who wrote the headnote to the case report does not appear to have thought that anything all that exciting had happened. He limited himself to recording that the House of Lords had decided that a manufacturer of goods owed a duty of care to their ultimate consumer in circumstances where there was no opportunity for intermediate inspection of those goods. And, indeed, strictly speaking that is probably all the ratio of the decision was. In that sense, it could be seen as just a further example of the situation-by-situation approach to duty of care, extended to a new situation.
But the judges in the House of Lords seem to have been alive to the fact that something more was going on, albeit that the expressions of these developments, including the much-quoted neighbour principle, were obiter. Indeed, it was a development which was a hotly contested battleground, because the outcome of Donoghue in the House of Lords was a 3 to 2 majority decision. A slight change in the composition of the panel hearing the case, and no law student might ever have heard of snails or ginger beer.
The principle protagonists on each side were Lord Atkin (of whom more shortly) and Lord Buckmaster.
Indeed, Lord Buckmaster, leading the line for the traditionalist minority (where he was joined by Lord Tomlin, he of “Tomlin order” fame), was positively dismissive of the claimant’s arguments. He seems to have taken a “thus far and no further” attitude to negligence law, painting an apocalyptic vision of the consequences of allowing duties of care to stray too far away from contractual relationships. His concern was the “floodgates” concern which has moved many judges since. It is interesting, when reading more recent floodgates arguments from the judiciary, to reflect on the fact that Lord Buckmaster’s expectation of cataclysmic consequences of allowing Donoghue to succeed never materialised. Floodgates arguments so rarely stand on the basis of anything more than their own self-justification. Empirical evidence does not tend to impinge.
The real significance of the case undoubtedly lies in Lord Atkin’s neighbour principle. What he was (probably) seeking to do was to move away from the situation-by-situation approach to duty of care and establish a more general approach that would apply across all situations. Such an approach could, if successful, unify the cases that had gone before and, perhaps more importantly, offer some element of predictivity in relation to cases yet to come. The terms of the neighbour principle are too familiar to restate, but it does contain within it the two key concepts which would come to dominate 20th century negligence law. These are foreseeability (which Lord Atkin mentions expressly) and proximity (a term he does not use but which is implicit in his definition of “who in law is my neighbour”).
Donoghue has never been overruled and remains part of English law. But if Lord Atkin’s aim was to provide a useable general approach to duty, then his success can be seen as no more than partial. One of the main attractions of the neighbour principle (its pleasing simplicity) is also its key weakness. Because, if the purpose was indeed to unify the court’s approach to duty, then it failed even to explain the cases that had gone before, let alone offer a workable approach to predicting the outcome of cases that would come. Take, for example, cases of pure economic loss. At the time of Donoghue these claims were, for all practical purposes, excluded entirely from the law of negligence. There simply was not a duty of care (in virtually all circumstances) not to cause pure economic loss and yet it cannot be said that such loss is necessarily any less foreseeable than physical damage, nor that there are not many situations where the parties to such a claim are not in a relationship of proximity or neighbourhood. To put it another way, many pure economic loss cases would seem to fit very neatly within the neighbour principle.
So it would appear that the neighbour principle entirely failed to capture something else which was going on in the law of negligence. This “something else”, unstated in Donoghue and indeed for several decades afterwards, was that rather slippery and problematic concept generally referred to as policy. Policy is a dangerous word because it has no intrinsically agreed meaning. Instead, it encompasses a whole bundle of ideas, some contradictory to each other, which determine the courts to impose, or, more often, not to impose, liability in negligence, independently of foreseeability and proximity. Foreseeability and proximity of themselves are simply not enough and never have been.
So, even the extension of liability which came about in the 1960s and 70s, (often in the field of economic loss but also in the field of third party liability) recognised that Donoghue, taken entirely at face value, would expand negligence law too far. So, in Dorset Yacht v Home Office  AC 1004,  2 All ER 294, when Lord Reid expressed the view that the time had come to apply Donoghue unless there were good reasons not to, even this was a recognition that, at times, there would be good reasons. And when, in Anns v Merton London Borough Council  AC 728,  2 All ER 492, the House of Lords sought to re-formulate the neighbour principle for those expansionist times, Lord Wilberforce recognised the need for an override mechanism. This, he said, was the question of whether there was some other substantial reason which might militate against imposing a duty of care, policy considerations by another name.
Anns, seen as the apotheosis of the expansionist era, fared rather less well than Donoghue, being effectively overruled by the House of Lords in Murphy v Brentwood District Council  1 AC 398,  2 All ER 908 and replaced as the test for duty of care in English law by Caparo Industries v Dickman  2 AC 605,  1 All ER 568.
The approach of the House of Lords to duty in Caparo is an interesting one. The reverence for Donoghue v Stevenson is such that the speeches do not seek to criticise it. But what they do is to subtly undermine the Atkinian project. If it is indeed the case that Lord Atkin was seeking to establish a general test for the existence of a duty (and there are commentators who deny that was ever his intention), then Caparo casts doubt not merely on the detail but on the entire aim. As Lord Oliver puts it, “it has to be recognised that to search for any single formula which will serve as a general test of liability is to pursue a will-o’-the-wisp”. The modern approach is to look to incremental development from earlier decisions, a step-by-step approach which seems to be in sharp contrast to the aim of a general principle. Because, if there is a general principle then how can it possibly matter whether the facts of a case are no more than a short step from the facts of a previous one?
Donoghue may not have invented the tort of negligence, but it freed the cause of action from the shackles that had inhibited its growth. It did this by opening up the field for duty of care to spread into new situations. While the neighbour principle is the most famous element of the case, it is the quotation from Lord MacMillan, another of the majority, that “the categories of negligence are never closed” which perhaps better caught the mood of the new departure that Donoghue marked. But, as has been recognised in more recent years, the simplicity of its approach to duty is, in reality, too simple. It fails to capture all the complicated variations of fact that negligence law has had to encounter. In many ways the modern approach is to deny the existence of any single approach to duty that works in all circumstances. There are different approaches for different contexts, something Lord Steyn has called “a mosaic” (MacFarlane v Tayside Health Board  2 AC 59,  4 All ER 961). In cases involving physical injury caused by a positive act, the neighbour principle still works pretty well, because those are cases where foreseeability and proximity will (in most cases) get you home. Once we stray outside these spheres into liability for non-physical injuries, or for omissions, or for the conduct of third parties, other considerations beyond foreseeability and proximity begin to acquire greater significance. In those cases, Donoghue stands as a foundation stone upon which much more has been built.
At the risk of extending the metaphor too far, the builders of negligence law have not all been working to the same plans over the years. In the years of expansion (exemplified by Anns), judges seemed to view negligence as an instrument of social progress, and themselves as crusaders for justice. It seemed that there were few wrongs which negligence could not be used to put right. This corresponded with a time when the state (in its widest sense) tended to be viewed positively as a device for advancement. Whether for good or for ill, that collectivist approach has clearly fallen significantly out of fashion, and it is no surprise that the law of negligence has come to reflect that change. The consequence was that the tendency to reach for negligence as the primary instrument of loss distribution has diminished. Indeed, the whole project that would suggest that it is a positive thing to distribute losses to those most able to bear them has been cast into doubt. The idea that losses should be left to lie where they fall has become much more prevalent.
Negligence did not become less expansionist because of the decision in Caparo. Rather, the decision in Caparo came about because of changes in social attitudes which led the judges to consider that the expansion had gone too far. Law is, therefore, a reflection of change more than it is a driver. From our current vantage point, it is hard to see a return to those expansionist times of the 60s and 70s, but attitudes that change can change back.
It is fruitless to ask what Lord Atkin would have made of how negligence has turned out. He was, in many ways, an Edwardian gentleman for whom much of the 21st century would make little sense. He may even have been surprised that this generation felt that the 80th anniversary of his judgment was worthy of note at all. But the enduring legacy of Donoghue v Stevenson does not rest only on the fact that it is a good story. It may be true that, had the same developments occurred in a more prosaic case, its fame may be less, but its impact as a foundational basis for modern negligence law would still have remained. History has not given us all that many influential snails. But Mrs Donoghue’s snail can proudly claim that title. Real or not, it has slithered its way to worldwide fame.