|
Torts - Negligence - Foreseeability/Remoteness (2). Hemmings v. Padmore
In Hemmings v. Padmore (SCC, 2026) the Supreme Court of Canada (in a closely-split decisions) dismissed two medical malpractice appeals (by the plaintiff H and defendant J) from the Ontario Court of Appeal, which had before dismissed actions against defendant P but held against defendant J.
Here the court considers the two elements of causation law in negligence - ie. "the factual cause of the plaintiff’s injuries and whether those injuries are too remote":C. The Methodology for Determining Factual Causation and Remoteness of Damage in Negligence
[42] To succeed in a claim in negligence, the plaintiff must establish that the defendant owed the plaintiff a duty of care, that the defendant breached that duty, that the plaintiff suffered damage, and that the damage was “caused, in fact and in law, by the defendant’s breach” (Deloitte & Touche v. Livent Inc. (Receiver of), 2017 SCC 63, [2017] 2 S.C.R. 855, at para. 77; see also Mustapha v. Culligan of Canada Ltd., 2008 SCC 27, [2008] 2 S.C.R. 114, at para. 3). Causation so described entails two distinct inquiries (Nelson (City) v. Marchi, 2021 SCC 41, [2021] 3 S.C.R. 55, at para. 96). First, the court must determine whether the defendant’s conduct was the factual cause of the plaintiff’s injuries (“factual causation”). Second, the court must determine whether the plaintiff’s injuries are too remote from the defendant’s conduct to assign liability (“remoteness of damage”).
[43] It is essential that factual causation and remoteness of damage remain analytically distinct and be treated separately. Only factual causation can accurately be characterized as addressing a strictly causal question. Remoteness, by contrast, is a normative inquiry into whether there is a sufficient connection between the defendant’s wrongful conduct and the plaintiff’s injuries such that the defendant should be liable. The process of determining this scope of liability “demands delicate value judgments and the drawing of fine lines. It is more a question of law and policy than fact” (A. M. Linden et al., Canadian Tort Law (13th ed. 2025), at §6.01). In reading the case law, one needs to be mindful that remoteness of damage has been referred to as “legal causation”, “proximate cause”, “proximity” or “cause-in-law”. These terms all refer to the same element of the negligence analysis and, for clarity, we will refer only to remoteness.
[44] We turn now to the principles relevant to each of the factual causation and remoteness inquiries.
(1) Factual Causation
[45] Factual causation requires the plaintiff to prove on a balance of probabilities that the defendant’s conduct caused their injury. In Clements v. Clements, 2012 SCC 32, [2012] 2 S.C.R. 181, this Court affirmed that “[t]he test for showing causation is the ‘but for’ test. The plaintiff must show on a balance of probabilities that ‘but for’ the defendant’s negligent act, the injury would not have occurred” (para. 8; see also Resurfice Corp. v. Hanke, 2007 SCC 7, [2007] 1 S.C.R. 333, at paras. 21-22; Marchi, at para. 96; Snell v. Farrell, 1990 CanLII 70 (SCC), [1990] 2 S.C.R. 311, at p. 330).
[46] The “but for” test requires that the defendant’s conduct was a necessary cause of the plaintiff’s injury, but not necessarily the sole cause (Clements, at para. 43; Snell, at p. 321). Where the plaintiff’s injuries are the result of the negligent acts of multiple actors, each of which is a necessary cause of the injury, then those actors are jointly and severally liable. The judge or jury may then apportion liability according to the degree of fault of each defendant subject to any contributory negligence legislation (Clements, at para. 12). Exceptionally, “where it is impossible to determine which of a number of negligent acts by multiple actors in fact caused the injury, but it is established that one or more of them did in fact cause it”, a plaintiff may recover on the basis of “material contribution to risk of injury” without showing “but for” causation (para. 13).
[47] Factual causation need not be proven with scientific precision (Snell, at p. 328). The plaintiff is not required to provide an opinion from a medical or scientific expert definitively supporting its theory of causation to discharge its burden of proof (pp. 328 and 330-31; Benhaim v. St-Germain, 2016 SCC 48, [2016] 2 S.C.R. 352, at paras. 47-50). “[T]he law requires proof of causation only on a balance of probabilities, whereas scientific or medical experts often require a higher degree of certainty before drawing conclusions on causation” (Benhaim, at para. 47, citing Snell, at p. 330). Whether factual causation is made out in given circumstances is “not algorithmically reducible” (R. Brown, “Known Unknowns in Cause-in-Fact” (2011), 39 Advocates’ Q. 37, at p. 64). Therefore, the “but for” test should be applied in a robust, common-sense fashion (Clements, at para. 9). This means that a court must consider “all the evidence which is before it, factual, statistical and that which the judge is entitled to presume” (Laferrière v. Lawson, 1991 CanLII 87 (SCC), [1991] 1 S.C.R. 541, at p. 606 (emphasis in original)).
[48] A defendant may rebut an inference of causation by introducing evidence demonstrating that the injury would have occurred regardless of their negligent conduct (Snell, at p. 336; Clements, at para. 11). However, even where the defendant argues or calls evidence that the plaintiff’s injuries would have happened with or without the defendant’s negligence, the burden remains with the plaintiff to prove factual causation (Clements, at para. 11; Snell, at p. 330). A finding of factual causation necessarily entails the exclusion of the counterfactual where the plaintiff’s injuries occur regardless of the defendant’s breach of the standard of care (Snell, at p. 327; Clements, at para. 21). In other words, if the plaintiff’s injury would not have occurred but for the defendant’s conduct that fell below the standard of care, it must also be true that, if the defendant’s conduct had met the standard of care, the plaintiff’s injury would not have occurred.
[49] In many cases, a common-sense inference of factual causation may be drawn from the evidence “without difficulty” (Clements, at para. 10). This includes, for example, cases with a single plaintiff and defendant, or cases where the defendant’s conduct is the sole physical cause of the full extent of the plaintiff’s injuries. In these circumstances, the trial judge may not need to explicitly explore and reject the counterfactual in their reasons. It can be inferred from a finding of factual causation that the trial judge did so.
[50] However, greater difficulty may arise in considering the counterfactual in some cases, particularly where there are multiple defendants, a significant amount of time has elapsed between the defendant’s negligent conduct and the plaintiff’s injury, each party has advanced competing theories of causation supported by the evidence, or expert evidence about causation is equivocal or contradictory. Medico-legal causation in particular can raise these kinds of complexities given the potential for multiple “interdependent variables that are rarely grounded in the decision-maker’s experience” (N. Hooper, “The Phenomenology of Medico-Legal Causation” (2017), 40 Dal. L.J. 579, at p. 590). In these circumstances, further analysis may be required to explain why, considering the evidence as a whole, the plaintiff has proven on a balance of probabilities that “the injury would not have occurred without the defendant’s negligence” (Clements, at para. 8).
[51] In these more complex circumstances, the court may need to engage in an explicit counterfactual analysis of whether, if a given defendant’s conduct had not fallen below the applicable standard of care, the harm to the plaintiff would nonetheless have occurred. In considering the counterfactual, the court determines the factual matrix leading to the plaintiff’s injuries, then omits the defendant’s negligent conduct and asks whether the plaintiff would still have suffered the same injury (A. Botterell, “Factual Causation”, in E. Chamberlain and S. G. A. Pitel, eds., Fridman’s The Law of Torts in Canada (4th ed. 2020), 503, at p. 506; Sacks v. Ross, 2017 ONCA 773, 417 D.L.R. (4th) 387, at para. 46). If after considering the counterfactual, the court concludes that “the damage would have happened just the same, fault or no fault, then the fault is not [the factual] cause of the damage” (Cork v. Kirby Maclean, Ltd., [1952] 2 All E.R. 402 (C.A.), at p. 407, per Denning L.J.).
(2) Remoteness of Damage
[52] Remoteness has remained a difficult concept to grasp throughout its history in Canadian jurisprudence and in the earlier English jurisprudence on which it is based. To clarify the principles of remoteness, it is useful to begin by examining its historical development leading to the doctrinal uncertainty that emerged during the 20th century and, to some extent, persists today. After doing so, we provide a summary of the principles relevant to the remoteness inquiry.
(a) Historical Development
[53] From the emergence of negligence as a basis of liability, courts have struggled with the scope of liability that should attach to a defendant’s negligent conduct (see Saadati v. Moorhead, 2017 SCC 28, [2017] 1 S.C.R. 543, at para. 18; P. H. Winfield, “The History of Negligence in the Law of Torts” (1926), 42 L.Q. Rev. 184; see also J. G. Fleming, “The Passing of Polemis” (1961), 39 Can. Bar Rev. 489, at p. 490). For what consequences should a defendant in breach of the standard of care be liable? In the cases from the late 19th and early 20th centuries, a conflict emerged between an approach that assigned liability based on direct causation (see Smith v. London and South Western Railway Co. (1870), L.R. 6 C.P. 14), and one focussed on foreseeability of harm (see, e.g., Greenland v. Chaplin (1850), 5 Ex. 243, 155 E.R. 104, at p. 106; Rigby v. Hewitt (1850), 5 Ex. 240, 155 E.R. 103, at p. 104; Victorian Railways Commissioners v. Coultas (1888), 13 App. Cas. 222 (P.C.); Lynch v. Knight (1861), 9 H.L.C. 577, 11 E.R. 854). The former would see a defendant liable for any consequence in the causal chain flowing from the defendant’s conduct, while the latter limits liability based on the “natural and probable” consequences of an act, viewed from the perspective of a reasonable person.
[54] The Court of Appeal of England and Wales provided a definitive statement on the preferred approach in Re Polemis and Furness, Withy & Co., [1921] 3 K.B. 560. There, the court considered and rejected the notion that a defendant should be liable only for those harms that could be reasonably foreseen at the time of their wrongdoing. Instead, it held that a defendant is liable for any harm that is directly caused by their wrongful conduct. In effect, this meant that defendants were liable for any harm for which their conduct was the factual cause. Re Polemis was subsequently followed in Thurogood v. Van Den Berghs & Jurgens Ld., [1951] 2 K.B. 537 (C.A.), at p. 555, and in Canada, the rule in Re Polemis was understood to represent settled law regarding the remoteness of damage in negligence (see, e.g., Toronto Railway Co. v. Grinsted (1895), 1895 CanLII 13 (SCC), 24 S.C.R. 570, at pp. 573-74; Winnipeg Electric Rway. Co. v. Canadian Northern Rway. Co. (1919), 1919 CanLII 529 (SCC), 59 S.C.R. 352, at pp. 367 and 369, per Duff and Anglin JJ.; The King v. Canadian Pacific Ry. Co., 1947 CanLII 34 (SCC), [1947] S.C.R. 185, at p. 196, per Rand J.; Booth v. City of St. Catharines, 1948 CanLII 10 (SCC), [1948] S.C.R. 564, at pp. 583-84, per Estey J.; see also Harding v. Edwards & Tatisich, 1929 CanLII 394 (ON SCAD), [1929] 4 D.L.R. 598 (Ont. S.C. (App. Div.)), per Middleton J.A.).
[55] It was not long, however, before the House of Lords began to question the coherence of the directness test and to find ways to reintroduce the concept of reasonable foreseeability into the remoteness analysis. By the 1940s, the directness test had been called into question by three House of Lords decisions (see Donoghue v. Stevenson, 1932 CanLII 536 (FOREP), [1932] A.C. 562; Bourhill v. Young, [1943] A.C. 92; Woods v. Duncan, [1946] A.C. 401; see also D. J. Payne, “The ‘Direct’ Consequences of a Negligent Act” (1952), 5 Curr. Legal Probs. 189, at pp. 195-202).
[56] In Donoghue, the House of Lords recognized the tort of negligence, and in so doing introduced reasonable foreseeability into the duty of care analysis. This gave rise to a tension between the use of the Re Polemis test for remoteness and reasonable foreseeability at the duty of care stage. On one hand, a negligent defendant would only owe a duty of care to those plaintiffs to whom they could reasonably foresee causing injury. On the other hand, once a duty of care was owed to that plaintiff, the defendant would be liable under Re Polemis for any injury that was directly caused by their conduct, no matter how far removed it may be from the negligent conduct (see P. H. Winfield, A Text-Book of The Law of Tort (5th ed. 1950), at p. 72).
[57] In Bourhill, Lord Macmillan, recognizing this tension, commented that while the law as set out by the Court of Appeal in Re Polemis appeared to suggest that a defendant is liable for “everything that directly follows” from a breach of the duty of care, “[w]hether the law there laid down is consonant with the law of England it will be for this House to pronounce when the occasion arises” (p. 106). Until that occasion arrived, the English courts took to resolving cases that raised questions of remoteness in the duty of care analysis, as in Bourhill. By doing so, courts arrived at results driven by common sense about what types of injury should be compensable “by the device of denying that any duty of care is owed in respect of damage which is too remote by the foresight test” (Payne, at p. 202).
[58] In 1961, the Judicial Committee of the Privy Council finally addressed Re Polemis and the role of remoteness in the negligence analysis with the decision of Overseas Tankship (U.K.) Ltd. v. Morts Dock & Engineering Co. (The Wagon Mound), [1961] A.C. 388 (“The Wagon Mound (No. 1)”). The defendant owned the eponymous ship, which took on fuel oil while docked in Sydney Harbour. Through the carelessness of the ship’s engineers, the oil leaked out of the ship and drifted on the surface of the water to a nearby wharf where two other ships were under repair. Workers on the wharf, despite seeing the oil, continued welding operations for the repairs after being advised it was safe to do so. The welding activity eventually set the fuel oil alight, causing extensive damage to the wharf and docked ships. The wharf owners sued the owners of the Wagon Mound in negligence and nuisance.
[59] The Privy Council was seized with the issue of whether the damage that resulted from the fire was too remote to hold the defendants liable. Both the trial judge and a full panel of the New South Wales Supreme Court, applying the rule in Re Polemis, found that the defendants were liable, notwithstanding a concession that the fire was not reasonably foreseeable. The Privy Council overruled Re Polemis. Viscount Simonds, in his speech, famously stated:It is a principle of civil liability, subject only to qualifications which have no present relevance, that a man must be considered to be responsible for the probable consequences of his act. To demand more of him is too harsh a rule, to demand less is to ignore that civilised order requires the observance of a minimum standard of behaviour. [pp. 422-23] [60] Given the rarity with which fuel oil was reasonably understood to ignite, the Privy Council concluded that the owners of the Wagon Mound could not be held liable for the fire damage to the wharf that resulted from the overflow of oil (see Fleming, at pp. 498-99). Liability was no longer determined by the characterization of consequences as direct or indirect, rather, consonant with the duty of care analysis, remoteness now turned on the reasonable foreseeability of those consequences.
[61] Shortly after The Wagon Mound (No. 1), Canadian courts began considering and, in some cases adopting, the reasonable foreseeability approach to remoteness (see Foster v. Registrar of Motor Vehicles for Province of Ontario, 1961 CanLII 144 (ON CA), [1961] O.R. 551 (C.A.); Oke v. Weide Transport Ltd. (1963), 1963 CanLII 574 (MB CA), 41 D.L.R. (2d) 53 (Man. C.A.); McKenzie v. Hyde (1967), 1967 CanLII 516 (MB QB), 64 D.L.R. (2d) 362 (Man. Q.B.), aff’d (1967), 1967 CanLII 618 (MB CA), 66 D.L.R. (2d) 655, per Dickson J. (as he then was); see also Buchanan v. Oulton (1965), 1965 CanLII 118 (NB CA), 51 D.L.R. (2d) 383 (N.B.C.A.), questioning the applicability of The Wagon Mound (No. 1) in New Brunswick).
[62] However, the decision did not resolve all questions related to the remoteness inquiry: there remained the question of the extent to which a particular sequence of events arising from the defendant’s conduct must be foreseeable to warrant assigning liability. In Hughes v. Lord Advocate, [1963] A.C. 837, the House of Lords held that a defendant need not foresee the exact way in which an accident occurs to fix liability, as long as the type of harm that occurred was reasonably foreseeable. Hughes represented something of a retreat from reasonable foreseeability in The Wagon Mound (No. 1) towards an approach closer to the directness test established in Re Polemis (see Linden et al., at §6.04).
[63] The Privy Council would further refine the test for remoteness of damage in Overseas Tankship (U.K.) Ltd. v. Miller Steamship Co. Pty., [1967] A.C. 617 (“The Wagon Mound (No. 2)”). The Privy Council again considered the same fuel spill that caused damage in The Wagon Mound (No. 1). This time, however, the plaintiffs were the owners of the two vessels docked at the wharf that were damaged by the fire. They were better positioned to advance an argument of reasonable foreseeability of damage without risking a finding of contributory negligence against them (Linden et al., at §6.05).
[64] The Privy Council held that the defendants were liable to the owners of the two docked vessels. In so doing, the Privy Council took the view that even an event that may be difficult to foresee can ground liability so long as there is a “possibility” or “real risk” of damage of the kind suffered by the plaintiff. Lord Reid clarified the concept of “real risk” in his seminal speech:If a real risk is one which would occur to the mind of a reasonable man in the position of the defendan[t] . . . and which he would not brush aside as far-fetched, and if the criterion is to be what [the] reasonable man would have done in the circumstances, then surely he would not neglect such a risk if action to eliminate it presented no difficulty, involved no disadvantage, and required no expense. [pp. 643-44] [65] Many saw The Wagon Mound (No. 2) as a further retreat from reasonable foreseeability and a move back towards the test set out in Re Polemis. If mere possibility is sufficient to attach liability to the defendant’s wrongful conduct, this would seem to eclipse the notion of reasonable foreseeability and represent a return to the idea of directness (see Botterell, “Remoteness”, in Chamberlain and Pitel, Fridman’s The Law of Torts in Canada, 529, at p. 535; J. C. Smith, “The Limits of Tort Liability in Canada: Remoteness, Foreseeability and Proximate Cause”, in A. M. Linden, ed., Studies In Canadian Tort Law (1968), 88; Linden et al., at §6.07). Some commentators argued that the lack of clarity regarding whether possibility or reasonable foreseeability was required to assign liability left a judge with a “choice of tests” between The Wagon Mound (No. 1) and (No. 2), depending on whether they “wanted” to hold for the defendant or the plaintiff, revealing the policy laden nature of the remoteness inquiry (C. Coval. J. C. Smith and J. Rush, “‘Out of the Maze’: Towards a ‘Clear Understanding’ of the Test for Remoteness of Damages in Negligence” (1983), 61 Can. Bar Rev. 559, at p. 566).
[66] This Court adopted the reasoning in The Wagon Mound (No. 2) in R. v. Côté, 1974 CanLII 31 (SCC), [1976] 1 S.C.R. 595. There, Justice Dickson (as he then was) built upon his reasoning as a judge of the Manitoba Court of Appeal in School Division of Assiniboine South, No. 3 v. Greater Winnipeg Gas Co., 1971 CanLII 959 (MB CA), [1971] 4 W.W.R. 746. In Assiniboine, the court concluded that a defendant “need not envisage ‘the precise concatenation of circumstances which led up to the accident’” to be held liable for the damage that resulted (p. 751). Rather, all that was required was that “the event giving rise to the damage is not regarded as ‘impossible’, . . . even [if] it ‘very rarely happened . . . only in very exceptional circumstances’” (ibid.). In Côté, Dickson J. adopted the language from Assiniboine, stating that “it is enough to fix liability if one can foresee in a general way the class or character of injury which occurred” (p. 604). Subsequently, this Court reaffirmed the applicability of The Wagon Mound (No. 2) in Bow Valley Husky (Bermuda) Ltd. v. Saint John Shipbuilding Ltd., 1997 CanLII 307 (SCC), [1997] 3 S.C.R. 1210.
[67] More recently, in Mustapha, remoteness was in issue in the case of a plaintiff who claimed to have suffered emotional distress from seeing a fly in a bottle of water. McLachlin C.J. restated the principles of remoteness. She held that the inquiry focuses on whether the harm suffered by the plaintiff is too unrelated to the wrongful conduct to fairly hold the defendant liable (para. 12). However, commenting on The Wagon Mound (No. 2), McLachlin C.J. emphasized that an approach distinguishing merely possible consequences from probable consequences was misleading, because “[a]ny harm which has actually occurred is ‘possible’” (para. 13). Accordingly, she held, “possibility alone does not provide a meaningful standard for the application of reasonable foreseeability” (ibid.). Rather, the test requires a “real risk” which would “occur [in] the mind of a reasonable man in the position of the defendan[t] . . . and which he would not brush aside as far-fetched” (ibid.).
[68] Subsequent decisions of this Court have affirmed that remoteness is an inquiry into whether the injury suffered by the plaintiff was a reasonably foreseeable consequence of the defendant’s negligent conduct (Livent, at para. 77; Saadati, at para. 20; Marchi, at para. 97). However, the Canadian version of the reasonable foreseeability test is narrower than what was stated in The Wagon Mound (No. 2): the boundary of reasonable foreseeability is not mere possibility, but some real risk that would not be dismissed by a reasonable person as far-fetched (Linden et al., at §6.07; see also Milliken v. Rowe, 2012 BCCA 490, 39 B.C.L.R. (5th) 283, at para. 25; St Isidore Co-op Ltd. v. AG Growth International Inc., 2020 ABCA 447, 16 Alta. L.R. (7th) 54, at paras. 29-30).
(b) The Principles Relevant to Remoteness of Damage
[69] The following principles relevant to the remoteness inquiry can be drawn from the jurisprudence.
[70] First, a negligent defendant can only be liable for those consequences that they could have reasonably foreseen at the time of their breach of the standard of care and from their vantage point (Mustapha, at para. 13; The Wagon Mound (No. 2), at p. 643). Professional defendants are assessed based on the skill and knowledge of the reasonable practitioner in their position at the time (see ter Neuzen v. Korn, 1995 CanLII 72 (SCC), [1995] 3 S.C.R. 674).
[71] Second, for the plaintiff’s injury to be reasonably foreseeable, there must have been a real risk or probability, one greater than a mere possibility, that that type of injury could occur (Livent, at para. 77). The risk must be sufficient in that a reasonable person in the defendant’s position would not brush it off as far-fetched and would have taken reasonable precautions to prevent it (Mustapha, at para. 13; The Wagon Mound (No. 2), at pp. 643-44).
[72] The defendant need not have foreseen the precise concatenation or sequence of events leading to the plaintiff’s injuries (Côté, at p. 604; Bow Valley Husky, at para. 76). Rather, the court must ask whether the defendant’s conduct gave rise to a real risk of the harm ultimately suffered by the plaintiff (Mustapha, at para. 13). Put another way, “[t]he type of accident or incident needs to have been foreseeable” but its particulars, as in the physical sequence of cause and effect that finally brought about the plaintiff’s injuries, do not (K. Cooper-Stephenson and E. Adjin-Tettey, Personal Injury Damages in Canada (3rd ed. 2018), at p. 1185).
[73] Academic commentators have opined that the language of real risk raised in The Wagon Mound (No. 2) and adopted in Mustapha has done little to resolve the doctrinal uncertainty around the scope of liability for negligence (see Linden et al., at §§6.06-6.07; L. N. Klar and C. S. G. Jefferies, Tort Law (7th ed. 2023), at pp. 604-5; see also R. W. M. Dias, “Remoteness of Liability and Legal Policy” (1962), 20 Cambridge L.J. 178, at pp. 189-90; Smith, at pp. 89 and 114; J. P. S. McLaren, “Negligence and Remoteness — The Aftermath of Wagon Mound” (1967), 32 Sask. L. Rev. 45, at p. 47). Some of these commentators argue that remoteness is a doctrine based firmly on notions of fairness and public policy, and that courts should acknowledge this policy rationale and do away with the artificial language of reasonable foreseeability. In some cases, they go further and advocate for a return to the directness test (see J. Murphy, “Reassessing Remoteness of Damage in Tort” (2025), 45 Oxford J. Legal Stud. 896).
[74] This is not a novel view. Indeed, in the landmark New York Court of Appeals decision of Palsgraf v. Long Island Railroad Co., 248 N.Y. 339 (1928), Justice Andrews, in dissent, opined that it is “because of convenience, of public policy, of a rough sense of justice, the law arbitrarily declines to trace a series of events beyond a certain point” (p. 352). He explained that it is to this end that courts deploy the notions of proximity and remoteness. Canadian courts have at times indicated similar views. For example, in Kienzle v. Stringer (1982), 1981 CanLII 1851 (ON CA), 35 O.R. (2d) 85 (C.A.), Zuber J.A. commented that reasonable foreseeability and related terms “necessarily include more policy than fact as courts attempt to find some fair measure of compensation to be paid to those who suffer damages by those who cause them” (p. 90).
[75] In our view, this appeal provides an opportunity to clarify that the inquiry into “real risk” is not merely a public policy exercise grounded in the judge’s view of what is fair in the abstract. Determining what is or is not a “real risk” of harm requires the court to address whether that risk is one that a reasonable person ought to have foreseen, and therefore taken reasonable precautions to prevent. This is what it means to determine whether the harm suffered by the defendant is “too unrelated to the wrongful conduct to hold the defendant fairly liable” (Linden et al., at §6.01; Mustapha, at para. 12 (emphasis added)). To reduce the “real risk” analysis to a matter of fairness and public policy, as some suggest, would introduce unacceptable uncertainty into tort law (see Coval, Smith and Rush, at pp. 565-66). It would allow the remoteness inquiry to come untethered from its purpose and place within the broader law of negligence.
[76] Instead, whether there was a “real risk” of the injury such that it was reasonably foreseeable is an inquiry that must be appreciated in light of the corrective justice orientation of all tort law (see Ahluwalia v. Ahluwalia, 2026 SCC 16, at para. 78; Atlantic Lottery Corp. Inc. v. Babstock, 2020 SCC 19, [2020] 2 S.C.R. 420, at para. 34; Clements, at para. 7). As this Court put it in Clements, corrective justice “assigns liability when the plaintiff and defendant are linked in a correlative relationship of doer and sufferer of the same harm” (para. 7, citing E. J. Weinrib, The Idea of Private Law (1995), at p. 156).
[77] From the perspective of corrective justice, compensation for private wrongdoing is not meant to punish the wrongdoer, but to remedy harm that results from the defendant’s unreasonable creation of the risk of that harm (Mustapha, at para. 7; Rankin (Rankin’s Garage & Sales) v. J.J., 2018 SCC 19, [2018] 1 S.C.R. 587, at para. 63; Babstock, at para. 33; see also Marchi, at para. 84). In this way, remoteness may be said to concern the scope or ambit of the risk created by a defendant’s negligent conduct. It follows that a defendant can only be liable for the plaintiff’s injuries if the probability that those injuries would occur was something the defendant ought to have considered and taken precautions to prevent, having regard to the degree of harm and the costs associated with preventing the risk of that harm (see The Wagon Mound (No. 2), at p. 718).
[78] Corrective justice offers a means to conceptualize the “why” behind the remoteness inquiry in the abstract and thereby guide its application. However, remoteness is ultimately a context-specific inquiry that must necessarily be guided by previous decisions taken by courts on concrete facts. This accumulated wisdom has given shape to the tort of negligence and must continue to inform its incremental development in new fact scenarios (see Ahluwalia, at paras. 69 and 72, per Kasirer J., at para. 285, per Karakatsanis J., concurring, and at para. 301, per Jamal J., dissenting, but not on this point; see also Friedmann Equity Developments Inc. v. Final Note Ltd., 2000 SCC 34, [2000] 1 S.C.R. 842, at para. 42).
[79] It is worth noting that the notion of real risk is deployed to ensure liability does not extend beyond its corrective justice foundation not just in the remoteness context, but also in establishing the duty of care. In the latter context, this Court has recognized that more than a mere possibility is required to establish reasonable foreseeability (Rankin, at paras. 45-46; see also Childs v. Desormeaux, 2006 SCC 18, [2006] 1 S.C.R. 643, at para. 29).
[80] While reasonable foreseeability is a concept relevant both for establishing a duty of care and determining whether the plaintiff’s injuries are too remote, it plays a different role in each inquiry. The duty of care analysis is aimed at establishing a sufficient link between the defendant’s wrongful conduct and the “type of injury” (Livent, at para. 78 (emphasis deleted); Marchi, at para. 97; see also Weinrib, at p. 158). This begins by looking to established categories of relationship where a duty of care has been recognized. Where there is no applicable established category, a novel duty of care may be recognized by applying the Anns/Cooper test (so named for the decision of the House of Lords in which the test was first set out, Anns v. Merton London Borough Council, [1978] A.C. 728, and the decision of this Court clarifying that test in the context of Canadian tort law, Cooper v. Hobart, 2001 SCC 79, [2001] 3 S.C.R. 537).
[81] The Anns/Cooper test requires that in order to establish a duty of care the plaintiff and defendant must be proximate, or in a sufficiently “close and direct” relationship to one another (Cooper, at para. 32). In some cases, factors like expectations, representations, reliance and property or other legal interests between the parties can support a finding of proximity (para. 34; Rankin, at para. 23). In many cases where there is no relationship between the parties, the question will turn on whether “the risk of the type of damage that occurred was reasonably foreseeable to the class of plaintiff that was damaged” (Rankin, at para. 24 (emphasis deleted), quoting A. M. Linden and B. Feldthusen, Canadian Tort Law (10th ed. 2015), at p. 322; see also Childs, at para. 31). Thus, at the duty of care stage, reasonable foreseeability is focussed on the type of harm and the class of plaintiff.
[82] In assessing risk for the purposes of the reasonable foreseeability analysis, it is important to take into account not just the likelihood of the plaintiff’s injury, but also its severity. This Court has observed, in different contexts, that the magnitude of risk is a product of both the gravity of the feared harm and its probability (see Sherman Estate v. Donovan, 2021 SCC 25, [2021] 2 S.C.R. 75, at para. 82; R. v. Mabior, 2012 SCC 47, [2012] 2 S.C.R. 584, at para. 86). Here the principle is no different. A reasonable person foreseeing the potential of their actions to cause harm would be expected to devote relatively more attention to those potential harms that are graver, given that the animating principle is that one ought to be “careful of the safety of his neighbour” (The Wagon Mound (No. 2), at p. 643). That said, reasonable foreseeability must also factor in the probability of the event actually happening, since many actions carry a very small but non-zero chance of catastrophic harm.
[83] Reasonable foreseeability plays a related but distinct role in the remoteness analysis. In the remoteness inquiry, the question of reasonable foreseeability is focussed on whether “the actual injury suffered by the plaintiff” was reasonably foreseeable to the defendant (Livent, at para. 78 (emphasis deleted), citing L. N. Klar and C. S. G. Jefferies, Tort Law (6th ed. 2017), at p. 565; Marchi, at para. 97). In this way, remoteness acts to ensure there is a sufficient connection between the defendant’s wrongful conduct and the plaintiff’s particular injury.
[84] In sum, the scope of liability cannot be determined categorically as the Hemmings and others before this Court suggest, based on general policy considerations that attach to a profession or type of activity. Rather, the extent of liability must be determined by reference to the reasonable person in the circumstances of a given defendant. That inquiry is rooted in corrective justice and guided by the accumulated wisdom reflected in the case law as applied to the specific circumstances of each case.
[85] Accordingly, while the characterization of remoteness as a question concerned with the scope of risk does not conflict with what we have set out above, we prefer to rely on the orthodox framing of the remoteness analysis in terms of reasonable foreseeability that has been developed in Canadian and English jurisprudence over the last century and a half.
|
