Since the early 1990s the American Law Institute has been in the process of adopting the Restatement of the Law (Third) of Torts [“Restatement (Third)”]. When this piecemeal process is eventually complete, we will have a thoughtful and reasonably up-to-date version of existing tort doctrine in the United States.1
Of course, so long as it remains a matter of state law, there can be no single law of torts
in the United States. Technically, we have more than fifty different tort law regimes, each with its own idiosyncrasies. Nevertheless, most of the states have reasonably similar rules and doctrines governing most issues, and those similarities are what the Restatement (Third) seeks to capture.
In any event, a Restatement, by its nature, is predominately backward-looking to what is, rather than forward-looking to what should be. In my view, American tort doctrine should be structured in ways that differ in important respects from the way that the Restatement (Third) will present them.
In hopes of finding some support for an alternative structure for tort doctrine, I attended the National Judicial Institute’s Civil Law Torts Seminar and the Conference of Canadian Tort Law Professors in May 2002 in Montreal to learn first-hand about Canadian tort law. I also came to help celebrate the career of Justice Allen M. Linden, an outstanding judge and Canada’s most distinguished scholar of tort law.
What I initially learned in Montreal was that, because Canada does not have a separate common law of torts for each common law province, the Supreme Court of Canada has enormous authority to establish a common tort regime for all (or most) of the nation – a power that no American court enjoys. Moreover, I came to appreciate that the Supreme Court of Canada has used that power in bold and imaginative ways.
I next learned that Canadian provinces do have the authority to over-ride the national common law with individually tailored statutory provisions that are applicable to a single province. Thinking in terms of American politics, I had expected to find that various provinces had exercised their autonomy in ways that had resulted in very different tort law rules from place to place in Canada. Certainly in the United States in recent years our state legislatures have routinely changed their common law tort rules by statute, and on some issues, such as comparative fault, joint and several liability, and the liability of alcohol servers, those changes have been quite different from place to place, just as the underlying American common law differs significantly from state to state on some issues.
It is true that Canadian provinces have taken dramatically different positions as to how personal injury resulting from auto accidents should be treated, and hence the rules governing compensation for auto injuries do differ substantially across Canada.2 But apart from that, I was unable to uncover any other important statutorily-enacted tort law differences among the common law provinces. The failure of either plaintiff or defendant interests in Canada to obtain legal advantages through the legislative process may be a result of Canada’s parliamentary form of government, or perhaps of Canadian tradition, or perhaps it merely reflects a widespread deference to the wisdom of the Canadian courts. But for whatever reason, this legislative inaction means, among other things, that Canadian tort law remains more dominated by judicially-created common law principles than is true for a large number of American states.
What I further learned, of course, is that alongside a national common law of torts, Canada has a civil law of torts as well, applicable to Quebec. I did not come close to mastering the nuances of the interaction between the civil law and the common law and the ways that Quebec judges and the Supreme Court of Canada deal with each other so as to reach what I understand to be largely similar outcomes to most typical tort problems, even if achieved through very different analytical processes. But what I did manage to learn a little bit about was the very different doctrinal structures of the common law and civil law systems.
I came away from the Montreal experience appreciating that on the common law side, perhaps not surprisingly, the Canadian courts still tend to look to England and somewhat to Australia, but rather rarely to the United States, and that on the civil law side, not at all surprisingly, the French tradition is still very strong. Nonetheless, I now recognize that Canada has carved out its own distinct common law and that Quebec has its own unique Code governing civil liability.
Most importantly for me, I have concluded that there are very attractive features of both systems of the Canadian law that could be combined to create a new tort law doctrine that would be better than either existing regime. When I say “better” I mean that, as compared with tort law today, whether American or Canadian of either stripe, the restructured tort doctrine would be simpler and thereby easier to teach, understand and apply to new problems; it would employ distinctive and more suitable doctrinal headings to deal with distinctive matters of fairness and policy; and it would more appropriately allocate burdens of persuasion to plaintiffs and defendants.
To be sure, Canadian common law, in my view, has doctrinal shortcomings – shortcomings, as I will explain, that are frequently very similar to the shortcomings I see in American tort law. So too, Canadian civil law, in my opinion, contains doctrinal shortcomings, discussed below, that also characterize the Continental legal system. Hence, in setting out my new way to structure tort doctrine, I will emphasize aspects of Canadian tort law from both systems that I would embrace and reject.
In a recent article honoring Professor Gary Schwartz, who, alas, died while at work on a portion of the Restatement (Third), I set out in some detail what I term “Visions of a Restatement (Fourth) of Torts”.3 Those visions also present the reformulated tort doctrine that I favor, but set entirely in the American context. In this article, I present those visions in a way that draws on what I see as the best of the Canadian common law and civil law systems.
II. Drawing on Canadian Tort Law’s Strengths and Avoiding Its Weaknesses
1. Two Bases for Tort Liability, Not Many Torts
In my judgment, a substantial disadvantage of the common law, not only in Canada but also the United States, the United Kingdom and Australia, is that there are too many separate torts. This is especially true for the many so-called intentional torts such as battery, false imprisonment, intentional infliction of emotional distress, defamation, and the like. All these different torts have resulted in the proliferation of separate, but needless, doctrinal labels that only create complications. These different doctrines make it difficult for courts and scholars to recognize that certain issues are common to the different torts, issues that presumptively ought to be resolved in the same way, but which instead are sometimes blithely treated inconsistently. Further, these many different doctrines also get in the way of making what otherwise would be helpful analogies between different sorts of injuries.
Instead of having many torts, I believe that all of tort law should be structured around two simple headings – first, the core and basic principle of fault-based liability, and second, the rather exceptional principle of strict liability (that is, liability imposed in the absence of fault).
Here the Canadian civil law is decidedly on the right track. In article 1457 the Civil Code of Quebec4 contains a basic and sweeping provision governing civil liability. Although this provision does not precisely state that when one is injured as a result of the fault of another, the latter is liable for the harm caused, that is the thrust of article 1457 and that is how it has been interpreted. Specifically, article 1457, in its English version, provides: “Every person has a duty to abide by the rules of conduct which lie upon him, according to the circumstances, usage or law, so as not to cause injury to another. Where he is endowed with reason and fails in this duty, he is responsible for any injury he causes to another person and is liable to reparation for the injury, whether it be bodily, moral or material in nature.” The basic point, for my purposes, is that the “rules of conduct” imposed on people by “circumstances, usage or law” essentially create the obligation to act reasonably under the circumstances: the same obligation that is imposed on people by the negligence or fault principle of the common law.
Hence, I would begin my restructured tort doctrine with an even more transparent version of the portion of article 1457 just quoted, making clear that an actor is generally liable to those injured by the actor’s fault.
Article 1457 itself goes on to provide that “[h]e is also liable, in certain cases, to reparation for injury caused to another by the act or fault of another person or by the act of things in his custody.” This rather obscure provision is later elaborated in articles 1459 through 1464, governing the fault of others, and in articles 1465 through 1469, governing the so-called acts of things. These provisions are unfortunate, in my view, because they mix together those circumstances in which one is liable without fault, those in which there is a rebuttable presumption of fault, and those in which, duplicating article 1457, it is made clear that proof of fault is required. So, for example, there is strict liability for the torts of an employee or agent (covered in article 1463) and for injuries caused by one’s animals (covered in article 1466), a rebuttable presumption of fault by parents and others with parental authority for the injuries caused by the children in their care (covered in articles 1459 and 1460), and liability only upon proof of fault of someone gratuitously supervising a child (also covered in article 1460). So, too, with respect to various things, sometimes liability is imposed unless the defendant can prove he or she is not at fault (article 1465 concerning so-called “autonomous acts” of things) and sometimes liability appears to be strict (as in articles 1467, 1468 and 1469 concerning “movable” and “immovable” property), although even this is vague because the definition of “defect” used in these articles does not make clear whether and, if so, when there can be a defect in the absence of fault.
Rather than the Civil Code’s use of these specialized provisions in articles 1459 through 1469, I would separate out the mere burden-shifting cases from the true strict liability cases. The former, as elaborated below, would become a sub-part to the basic fault principle. The latter would be grouped under a second and special heading governing liability without fault.
In the United States the three important instances of true strict liability today are: (1) liability for harms caused by so-called “abnormally dangerous activities”, which is the eventual American outcome of the doctrinal rule that began with the nineteenth-century English case of Rylands v. Fletcher;5 (2) liability for harms caused by so-called “manufacturing defects” in products liability cases;6 and (3) the vicarious liability of employers, and others similarly situated, for the torts of their employees.7 Less important instances of strict liability in American law concern certain injuries caused by animals in one’s care8 and harm reasonably caused to another in certain circumstances of so-called “private necessity”, such as when, because of an unexpectedly fierce storm, a ship owner is forced to tie up the ship at a private dock, but then, albeit without negligence, the ship causes damage to the dock.9 Canadian common law is even less receptive to strict liability. Like English common law, Canada appears to have eviscerated the Rylands rule.10 Canadian common law continues to follow the leading English case of Donoghue v. Stevenson,11 thereby still requiring proof of fault even in cases of injuries arising from manufacturing defects, and because of the power of Donoghue, Canadian common law, like English common law, also appears not to impose strict liability on those who harm others in situations of “necessity”.12 Hence, if Canadian common law were to embrace the new simplified approach to tort doctrine that I propose, there would be far fewer matters covered under the strict liability heading than there would be in American law.
2. The Plaintiff’s Burden: Proof of Breach and Factual Causation
As I see it, then, it should normally be the injured plaintiff’s job to show either that the defendant unreasonably failed to prevent the injury and thereby was at fault or that the injury is one that attracts liability without fault. That is, the plaintiff should first have to demonstrate that his or her case is properly covered by at least one of the two basic doctrinal principles covering tort liability. By doing so, the plaintiff would carry the burden of showing a breach of the relevant standard. Notice that, for fault-based liability, this actually requires a showing of how the defendant should have acted, not merely a showing of how the defendant actually behaved.
Although I will not go into detail here, this new approach would mean, for example, that plaintiffs would not be asked to prove the various elements of a common law battery or false imprisonment, but rather they would have to show that they were unreasonably struck or unreasonably detained by the defendant, thereby bringing the case under the fault heading. This is what plaintiffs now must do under the civil law regime.
The plaintiff should also have to prove what is perhaps best termed “factual causation”. For fault-based claims, this means showing that, if the defendant had acted reasonably, the plaintiff would not have been injured. This is the so-called “but for” test of causation, which, it is perhaps worth noting, actually requires answering a hypothetical question: what would have happened had the defendant acted carefully instead of the way he or she did act?
3. Some Special Exceptions Concerning Proof of Fault or Proof of Causation
A special exception to the normal rule concerning proof of fault should be available if, as a matter of policy, it is thought desirable in certain situations to shift the burden from the plaintiff to the defendant. This might be true, for example, where the defendant clearly has much better access to the evidence, such as instances of exploding soda bottles that one suspects were carelessly charged at the bottling plant,13 or perhaps in instances of claims against parents for injuries caused by their children (as now provided in the Civil Code, noted earlier). However, the rules governing discovery in the jurisdiction might well make shifting the burden of proof unnecessary in these sorts of cases. Moreover, in some situations this type of case may have already been classified in the strict liability category, thereby rendering proof of fault irrelevant, as is the case with manufacturing-defect product injuries in the United States.
In other situations, the basic rules governing the proof of factual cause might also properly be altered a matter of public policy. One such instance concerns so-called “over-determined causation”. In such settings, each of two at-fault parties might convince us that the plaintiff would have been injured anyway had either of them acted carefully. However, allowing both defendants to escape liability by pointing the finger at the other means that the plaintiff, who would not have been injured in the absence of fault, would recover nothing. A classic example of this sort of case involves two independently acting motorcyclists who simultaneously and carelessly roar their engines loudly enough that each noise is sufficient to frighten the plaintiff’s nearby horse into tossing him or her out of the saddle and onto the ground, thereby causing injury. Clearly, in such settings, fairness should preclude each careless defendant from taking advantage of the simultaneous coincidence of another carelessly-acting defendant.
A second instance that perhaps calls for relaxing the rules about proof of factual causation concerns multiple, but independent, defendants who all acted unreasonably, but for understandable reasons the victims of those actions cannot determine which of those actors actually caused their harm. A paradigm example concerns a drug that the plaintiffs took that caused harm, where the precise manufacturer of the various pills in question is indeterminate because many made identical products, records are understandably unavailable, and so on. In such settings, perhaps some sort of market share liability would much more fairly allow plaintiffs to recover and wrongful injurers to pay approximately what they should owe, even if that means relaxing the ordinary rules of one-to-one matching of causation.14 A different example of this problem arises when two vehicles carelessly run into each other, whirl around, and somehow together strike an innocent pedestrian. Again, because of proof impossibilities, fairness may well require that the two defendants be held either fully liable or at least liable for half of the harm done.
The Civil Code addresses some of these sorts of causation problems in article 1480, although that provision’s precise language restricts the sharing of liability to those who have “jointly taken part in a wrongful act”. I have primarily in mind here those settings in which the parties would not ordinarily be considered to be joint actors. For those matters, the civil law judges, like their common law counterparts, must seek the proper result without clear statutory guidance.
4. Clarified Roles for the Separate Principles of Causation
The Civil Code clearly provides separate treatment of factual causation and the problems dealt with by the common law doctrine of “proximate cause”. Indeed, whereas factual causation is covered in article 1457, there are at least two distinct provisions in the CivilCode that deal with “proximate cause” problems. Article 1607 provides that plaintiffs are entitled to damages for an injury that is “an immediate and direct consequence” of the defendant’s breach of his or her legal obligation. Furthermore, article 1470, which is grouped along with other “defences”, provides that one can escape liability by showing that the injury in question resulted from a “superior force”, which is, in turn, defined as “an unforeseeable and irresistible event”.
Canadian common law, like American common law, uses the word “cause” in connection with the separate concepts of factual causation and proximate cause. This has tended to create confusion. In the United States, for example, the notion of “substantial factor” that was initially meant to serve as a different way of expressing what is a “proximate cause” has now unfortunately migrated and has come to be used in connection with problems that are instead about “cause-in-fact”.15 The Supreme Court of Canada’s embrace of the phrase “materially contributed” may be laden with the same shortcomings.16 I would clearly differentiate these two doctrinal matters. The “cause-in-fact” or factual causation requirement would be an integral part of the basic provisions governing fault liability and strict liability, with the rules about the burden of proof and the special exceptions to the “but for” test handled as discussed above. The “proximate cause” problem would be re-characterized as the problem of injuries that are “outside of the scope of the risk taken” and would become one of the “defences” in my restructured tort doctrine.
To illustrate, assume that the plaintiff has shown that the defendant either was at fault or acted in a way that generally gives rise to strict liability, and assume that the plaintiff has also shown the necessary factual causal link between the defendant’s act and the plaintiff’s injury. Ordinarily, fairness would require that the defendant compensate the plaintiff for his or her loss. Nevertheless, sometimes the defendant can convince us that the injury that this plaintiff actually incurred is not fairly part of the risk, or series of risks, that the defendant took that gave rise to actionable conduct. When the defendant can so convince us, I believe that it is unfair to hold him or her liable.
I should emphasize that this doctrine rests, not on various policy considerations of the sort relevant to the duty issue,17 but rather simply on considerations of fairness in the specific case. Defendants are, in effect, saying that, because of the unexpected way in which the injury took place, someone else or some force of nature is where the finger of responsibility should be pointed and that it would be unfair to instead point it at them.
In my view, the sorts of cases that make some injuries not fairly part of the risk taken by the defendant can generally be grouped under three headings: (1) where the type of harm suffered by the plaintiff was not foreseeable, even though one or more other types of harm were; (2) where the plaintiff was not foreseeably put at risk by the defendant’s conduct, even though one or more others were; and (3) where the intervening action of a third party was both unforeseeable and of a sort that is fairly understood to eclipse the responsibility of the defendant. Cases readily falling under the second and third of these headings can sometimes also be understood to fall under the first
The first heading is nicely illustrated by a plank dropped into a hold of a ship which risked striking a person or property but which in no way was expected to start the fire it did, as happened in the famous English case of In re Polemis.18 Of course, in Polemis the defendant was held liable for the fire, but that case was later over-ruled, essentially on the ground that the fire was outside the scope of the risk the defendant took when dropping the plank: fire was not the type of risk that the defendant foreseeably created. A more complicated example under this same heading is illustrated by the Privy Council’s decision in the Australian case of Overseas Tankship U.K. Ltd. v. Morts Dock and Engineering Co., Ltd. (The Wagon Mound) (No. 1),19 which overturned Polemis. There, the defendants negligently spilled some furnace oil that they should have foreseen would have mucked up the harbor, but which (surprisingly) they could not have foreseen would have caught fire, but which caught fire nonetheless. Because the fire was outside the scope of the risk taken, that is, not the type of risk they foreseeably ran, the defendants were not liable.
My second heading is illustrated by the famous New York case of Palsgraf v.Long Island Railroad Co.20 There, railway guards carelessly pushed a passenger onto a train, thereby endangering him and the package he was carrying. In no way could they have anticipated, however, that this package carried explosives that would be set off by their actions, causing the overturning of some scales well down the platform and thereby injuring the plaintiff. Although Justice Cardozo phrased it very differently, the better way to understand the result in Palsfgraf, in which the railway was held not liable, is that harm to the plaintiff was outside the scope of the risk the careless guards took because she was not foreseeably put at risk by the guards.
My third heading is illustrated by a California case21 in which defective wood was carelessly supplied by the defendant to a buyer for the purpose of building a platform. When the defect in the wood was discovered by the buyer, rather than seeking replacement wood, the buyer amazingly and unexpectedly went ahead and built the platform anyway. The plaintiff later was injured when the platform predictably collapsed. Here, although the defendant would have been liable had the buyer built the platform unaware of the condition of the wood, the defendant was relieved of liability. The buyer’s decision to build the platform with knowledge of the defective wood was seen to be outside the scope of the risk taken. This is an instance of an unforeseeable intervening action by a third party that eclipsed the responsibility of the supplier of the wood.
It is by no means always evident whether a particular case fits under any of these three headings. In practice, effective advocacy can be crucial here. Moreover, because under my restructured tort doctrine it would be up to the defendant to carry the burden of showing that the plaintiff’s injury is not fairly part of the risk taken, I want to re-emphasize that this doctrine would serve as a defence, in line with the thinking behind article 1470 of the Civil Code. By contrast, at common law, proving “proximate cause” is traditionally viewed as part of the plaintiff’s burden.
5. Clarified Role for the “No Duty” Principle
In the common law, it is widely appreciated that one is not liable if one has no legal duty with respect to the injury that occurred. Hence, in common law jurisdictions there is considerable case law exploring when a defendant does or does not owe a legal duty. This separate treatment is appropriate because the “no duty” issue is quite different from the issues of “no proximate cause” and “no factual causation.” Hence, duty should have a distinct doctrinal heading. Canadian civil law, however, appears to have no specific provision that focuses on the “no duty” issue. Instead, under the Civil Code judges seem to employ a combination of the proximate cause requirement set out in article 1607, the factual causation requirement set out in article 1457, and other provisions to deal with problems that common law systems handle with the “no duty” principle. This is undesirably confusing.
As compared with the United States, Canadian common law has been especially attentive to the duty issue. The Supreme Court of Canada has wrestled long and hard with this doctrinal question, including in its recent attempt in Cooper v. Hobart,22 both to reflect and to distinguish similar British common law developments arising from the leading case of Anns v. Merton London Borough Council.23 Alas, although the Supreme Court of Canada has offered many helpful insights, I believe that the current state of Canadian common law doctrine on this question is undesirably confused and complex. Similarly, though, little comfort may be taken from the American experience. In California, for example, although the California Supreme Court has also wrestled with the duty issue, it too has advanced confusing and undesirable considerations as relevant to making the duty determination.24 For example, both courts talk about “foreseeability” as a relevant criterion for the duty determination. This is wrong, in my view. If the defendant has not run any foreseeable risk, then there can be no breach – that is, the defendant cannot be found to be at fault. It is not a matter of duty. So, too, foreseeability is relevant to the determination of whether the injury to the plaintiff is fairly part of the risk that the defendant took. But this too is not a matter of duty.
The Supreme Court of Canada also talks about “proximity” as relevant to the duty issue.25 While proximity may be relevant to the question of whether the harm is fairly part of the risk taken, once again, it is not, in my view, relevant to the duty issue. On the other hand, the court is exactly right, in my view, when it self-consciously acknowledges that duty is question of social policy.26 Other commentators have struggled to determine precisely what change, if any, the Supreme Court of Canada was intending to bring about, or perhaps inadvertently brought about, in Cooper, which re-formulated the previously-embraced two-part Anns test. That is not my concern here. Rather, the difficulty I see is that the court has not been as thematic as it might have been in setting out precisely what policy considerations a court should appropriately take into account in denying recovery to an otherwise deserving plaintiff. That is, what sort of social policy arguments justify a finding of “no duty”?
I have explored this question, and although I do not purport at present to have a complete list of these policy arguments, I can present eight different reasons that I believe can, if truly found applicable to the situation in question, justify a “no duty” conclusion. I hope my presentation of these policy arguments makes it clear that these reasons for “no duty” are very different from the individual fairness considerations that might lead one to conclude, in a particular case, that the harm caused is outside the scope of the risk taken. To make these policy arguments more concrete, I generally provide a familiar example from tort law to illustrate each policy reason offered, although if one focused on each of these examples, one might conclude that more than one of the policy reasons applies.
Seven such reasons are: (1) to allow recovery will so flood the courts with small injury claims (such as, perhaps, minor emotional distress claims) and thus create a backlog that will inevitably deny justice to more seriously injured plaintiffs; (2) to allow recovery will deny the public essential public services (as in, perhaps, certain public utility cases or certain pure economic loss cases) so as to produce a socially worse outcome than that which comes from denying recovery to otherwise deserving plaintiffs; (3) a well-functioning parallel private justice system applies to the setting in question (as in, perhaps, professional sports injury cases) in which the public goals of victim compensation, safety promotion and punishment are already well served; (4) there is a different judicial setting (such as, perhaps, divorce court with respect to emotional misconduct during the marriage) in which the dispute between the parties is better resolved; (5) in this sort of case the judiciary is unlikely to be able reach a just result (for example, because it is incompetent to determine the truth or it legitimately fears collusion as to the evidence, as in, perhaps, certain intra-family disputes); (6) to allow the plaintiff a remedy violates the principle of separation of powers and thereby intrudes on legislative or administrative policy-making, so that the plaintiff’s relief, if any, must lie elsewhere (as in certain public agency defendant cases); and (7) to allow recovery will generate perverse behavioral responses (as in, perhaps, recreational injuries cases) that will be more harmful to society than the harm caused by denying recovery to otherwise deserving plaintiffs.
One last “no duty” policy argument (8) is perhaps more controversial but, in my view, it is also appropriately included in the list. I have in mind situations in which there is a trumping social value of great importance at stake that should win out despite the plaintiff showing that, in this particular case, the defendant’s conduct seems unreasonable. This might also be thought of in terms of making the “breach” determination at the “wholesale” level, whereas normally it is made at the “retail” level. In other words, normally we focus on the risk a defendant imposed on a plaintiff and then decide whether that risk was reasonable or unreasonable to run. In some circumstances, however, it might be concluded that, if we broaden our focus well beyond this plaintiff, the risk was reasonable, even if it might not have seemed so with a narrower focus.
This is an especially important argument in the United States and in Canadian provinces that use juries, because of the concern that juries might focus too narrowly on the facts of the specific case and be inadequately attentive to the wider policy issues at stake. Examples in this category might include a “no duty” rule applicable to television broadcaster defendants whose programs depict violence that is then “copied” by criminals in their attacks on victims, because of the trumping value of free speech; or a “no duty” rule applicable to strangers who fail to rescue others, if the true reason for such a rule is the trumping value of the liberty to remain uninvolved in the lives of others.
I believe that each of the eight reasons I have set forth for why a tort duty might be denied is highly relevant to a largely judicially-created system of private recovery.27 Moreover, it seems to me that in most instances courts are well-equipped to decide whether the reason or reasons asserted for “no duty” actually apply to the particular injury setting in question. Yet even when these are difficult issues for courts, I believe that judges have an obligation to do their best to avoid having the system they run bring about more social harm than good.
On the other hand, I am not asserting that the policy arguments necessarily convincingly apply to the examples I provide by way of illustration. I recognize that people can quite reasonably believe that, even if the policy arguments I offer are convincing in the abstract, they are simply not persuasively applicable to many of the settings in which they have traditionally been applied. Indeed, to guard against too quickly embracing defendants’ claims about the socially undesirable outcomes of imposing a tort duty in the sort of case that is before the court, I should further emphasize that I would make the “no duty” issue one which falls to the defendant to raise and prove. After all, refusing recovery to an otherwise deserving plaintiff is presumptively a substantial denial of justice, and so there must generally be a convincingly strong reason for doing so.
It would be better if tort doctrine were to start, as the Civil Code does, with the basic principle of liability for fault. This would be in lieu of the series of separate torts featured in the Canadian and American common law. The fault principle would be followed by a second, exceptional principle for the special instances of liability without fault.
Tort liability should be precluded, however, on three grounds. One is the failure of the plaintiff to prove factual causation, and unlike in both the Canadian common law and civil law, the “cause” label should be applied exclusively to this ground. A second ground for denying recovery is a convincing showing that the defendant should not owe a legal duty to the plaintiff, and this ground should draw on Canadian common law’s open recognition that duty is a policy issue. A third ground for denying recovery is a convincing showing that the injury suffered by the plaintiff is not fairly part of the risk taken by the defendant, and this ground should rest on principles of fairness found in both the common and civil law traditions.
Of course, these few principles do not completely describe a reformulated structure of tort doctrine. For example, issues of shared responsibility must also be addressed, both as between plaintiffs and defendants in cases where a plaintiff is at fault, and as among defendants in multiple-defendant settings. Attention must also be given to the type of damages that are to be awarded to plaintiffs, and to their quantification. Of course, under the fault principle the idea of what constitutes unreasonable behavior in various settings could be elaborated in ways already done in the common law.
For now, the basic point is that one may draw on the best features of Canadian common law and civil law to build a new vision of tort doctrine, not only for Canada but also for her neighboring jurisdictions to the south. This new vision of tort doctrine might permit Justice Linden to shorten his treatise on Canadian tort law, even if it also might require him to reorganize his leading torts casebook.
* Agnes Roddy Robb Professor of Law, University of California (Berkeley). McGill University Law Professor Shauna Van Praagh provided especially helpful comments on drafts of this article.
1 So far the Institute has adopted provisions on product liability, in American Law Institute, Restatement of the Law (Third) of Torts: Product Liability (1998), and on the apportionment of liability, in American Law Institute, Restatement of the Law (Third) of Torts: Apportionment of Liability (2000), and it is in the process of adopting provisions concerning basic principles governing physical injury and property damage.
2 Not too long ago I wrote enviously of the Quebec auto no-fault scheme: Sugarman, “Quebec’s Comprehensive Auto No-Fault Scheme and the Failure of Any of the United States to Follow” (1998), 39 C. de D. 303.
3 Sugarman, “Re-Thinking Tort Doctrine: Visions of a Restatement (Fourth) of Torts” (2002), 50 U.C.L.A. L.Rev. (forthcoming).
4 S.Q. 1991, c. 64 [“Civil Code”].
5 (1868) L.R. 3 H.L. 330 [“Rylands”].
6 See section 402A of American Law Institute, Restatement of the Law (Second) of Torts (1965) and sections 1 and 2 of American Law Institute, Restatement of the Law (Third) of Torts: Products Liability (1998).
7 See generally American Law Institute, Restatement of the Law (Second) of Agency (1958).
8 Restatement of the Law Torts: Liability for Physical Harm (Basic Principles) Tentative Draft No. 1, sections 21 and 22.
9 Vincent v. Lake Erie Transportation Co., 109 Minn. 456 (1910). See generally section 197 of American Law Institute, Restatement of the Law (Second) of Torts (1965).
10 See for example Cambridge Water Co. v. Eastern Counties Leather plc,  2 A.C. 264 (H.L.) and Tock v. St. John’s Metropolitan Area Board,  2 S.C.R. 1181. See generally Linden, “Whatever Happened to Rylands v. Fletcher?” in Klar (ed.), Studies in Canadian Tort Law (1977).
11  A.C. 562 (H.L.) [“Donoghue”].
12 See Munn and Co. v. “The Sir John Crosbie”,  1 Ex. C.R. 94 and Cope v. Sharpe (No. 2),  1 K.B. 496 (C.A.).
13 See for example Escola v. Coca Cola Bottling Co. of Fresno, 24 Cal.2d 453 (1944).
14 See for example Hymowitz v. Eli Lilly & Co., 73 N.Y.2d 487 (1989).
15 See for example Mitchell v. Gonzales, 819 P.2d 872 (Cal. 1991).
16 See Walker Estate v. York-Finch General Hospital,  1 S.C.R. 647.
17 See “Clarified Role for the ‘No Duty’ Principle”, infra.
18  3 K.B. 560 [“Polemis”].
19  A.C. 388 (P.C.).
20 248 N.Y. 339 (1928) [“Palsgraf”].
21 Stultz v. Benson Lumber Co., 6 Cal.2d 688 (1936).
24 For a leading California case discussing the factors relevant to finding a duty see Rowland v. Christian, 443 P.2d 561 (Cal. 1968).
25 See Cooper, supra, note 22, at paras. 30-36.
26 See id., at paras. 25, 30 and 37.
27 For another effort to set out several categories of “no duty” cases see Stapleton, “Duty of Care Factors: a Selection from the Judicial Menus” in Cane and Stapleton (eds.), The Law of Obligations: Essays inCelebration of John Fleming (1998).