Opinion
GEORGE, J.In this case, and in the companion case of Ford v. Gouin, post, page 339 [11 Cal.Rptr.2d 30, 834 P.2d 724], we face the question of the *300proper application of the “assumption of risk” doctrine in light of this court’s adoption of comparative fault principles in Li v. Yellow Cab Co. (1975) 13 Cal.3d 804 [119 Cal.Rptr. 858, 532 P.2d 1226, 78 A.L.R.3d 393], Although the Li decision itself addressed this issue, subsequent Court of Appeal decisions have differed in their interpretation of Li’s discussion of this point. We granted review to resolve the conflict among the Courts of Appeal.
I
We begin with a summary of the facts of this case, as set forth in the declarations and deposition transcripts submitted in support of and in opposition to defendant’s motion for summary judgment.
On January 25, 1987, the day of the 1987 Super Bowl football game, plaintiff Kendra Knight and defendant Michael Jewett, together with a number of other social acquaintances, attended a Super Bowl party at the home of a mutual friend. During half time of the Super Bowl, several guests decided to play an informal game of touch football on an adjoining dirt lot, . using a “peewee” football. Each team had four or five players and included both women and men; plaintiff and defendant were on opposing teams. No rules were explicitly discussed before the game.
Five to ten minutes into the game, defendant ran into plaintiff during a play. According to plaintiff, at that point she told defendant “not to play so rough or I was going to have to stop playing.” Her declaration stated that “[defendant] seemed to acknowledge my statement and left me with the impression that he would play less rough prospectively.” In his deposition, defendant recalled that plaintiff had asked him to “be careful,” but did not remember plaintiff saying that she would stop playing.
On the very next play, plaintiff sustained the injuries that gave rise to the present lawsuit. As defendant recalled the incident, his team was on defense on that play, and he jumped up in an attempt to intercept a pass. He touched the ball but did not catch it, and in coming down he collided with plaintiff, knocking her over. When he landed, he stepped backward onto plaintiff’s right hand, injuring her hand and little finger.
Both plaintiff and Andrea Starr, another participant in the game who was on the same team as plaintiff, recalled the incident differently from defendant. According to their declarations, at the time plaintiff was injured, Starr already had caught the pass. Defendant was running toward Starr, when he ran into plaintiff from behind, knocked her down, and stepped on her hand. Starr also stated that, after knocking plaintiff down, defendant continued *301running until he tagged Starr, “which tag was hard enough to cause me to lose my balance, resulting in a twisting or spraining of my ankle.”
The game ended with plaintiff’s injury, and plaintiff sought treatment shortly thereafter. After three operations failed to restore the movement in her little finger or to relieve the ongoing pain of the injury, plaintiff’s finger was amputated. Plaintiff then instituted the present proceeding, seeking damages from defendant on theories of negligence and assault and battery.
After filing an answer, defendant moved for summary judgment. Relying on the Court of Appeal decision in Ordway v. Superior Court (1988) 198 Cal.App.3d 98 [243 Cal.Rptr. 536], defendant maintained that “reasonable implied assumption of risk” continues to operate as a complete defense after Li v. Yellow Cab Co., supra, 13 Cal.3d 804 (hereafter Li), and that plaintiff’s action was barred under that doctrine. In this regard, defendant asserted that “[b]y participating in [the touch football game that resulted in her injury], plaintiff . . . impliedly agreed to reduce the duty of care owed to her by defendant ... to only a duty to avoid reckless or intentionally harmful conduct,” and that the undisputed facts established both that he did not intend to injure plaintiff and that the acts of defendant which resulted in plaintiff’s injury were not reckless. In support of his motion, defendant submitted his own declaration setting forth his version of the incident, as summarized above, and specifically stating that he did not intend to step on plaintiff’s hand or to injure her. Defendant also attached a copy of plaintiff’s deposition in which plaintiff acknowledged that she frequently watched professional football on television and thus was generally familiar with the risks associated with the sport of football, and in which she conceded that she had no reason to believe defendant had any intention of stepping on her hand or injuring her.
In opposing the summary judgment motion, plaintiff first noted that, in contrast to the Ordway decision, the Court of Appeal decision in Segoviano v. Housing Authority (1983) 143 Cal.App.3d 162 [191 Cal.Rptr. 578] specifically held that the doctrine of “reasonable implied assumption of risk” had been eliminated by the adoption of comparative fault principles, and thus under Segoviano the basic premise of defendant’s summary judgment motion was untenable and plaintiff was entitled to have the lawsuit proceed under comparative fault principles.
Furthermore, plaintiff maintained that even were the trial court inclined to follow the Ordway decision, there were numerous disputed material facts that precluded the granting of summary judgment in favor of defendant. First, plaintiff noted there was a clear dispute between defendant’s and *302plaintiff’s recollection of the specific facts of the play in which plaintiff was injured, and, in particular, of the details of defendant’s conduct that caused plaintiff’s injury. She claimed that under the facts as described by plaintiff and Starr, defendant’s conduct was at least reckless.
Second, plaintiff vigorously disputed defendant’s claim that, by participating in the game in question, she impliedly had agreed to reduce the duty of care, owed to her by defendant, to only a duty to avoid reckless or intentionally harmful conduct. Plaintiff maintained in her declaration that in view of the casual, social setting, the circumstance that women and men were joint participants in the game, and the rough dirt surface on which the game was played, she anticipated from the outset that it was the kind of “mock” football game in which there would be no forceful pushing or hard hitting or shoving. Plaintiff also asserted that the declarations and depositions of other players in the game, included in her opposition papers, demonstrated that the other participants, including defendant, shared her expectations and assumptions that the game was to be a “mellow” one and not a serious, competitive athletic event.1 Plaintiff claimed that there had been no injuries during touch football games in which she had participated on previous occasions, and that in view of the circumstances under which the game was played, “[t]he only type of injury which I reasonably anticipated would have been something in the nature of a bruise or bump.”
In addition, in further support of her claim that there was at least a factual dispute as to whether she impliedly had agreed to assume the risk of injury from the type of rough play defendant assertedly engaged in, plaintiff relied on the portion of her declaration in which she stated that (1) she specifically had told defendant, immediately prior to the play in question, that defendant was playing too rough and that she would not continue to play in the game if he was going to continue such conduct, and (2) defendant had given plaintiff the impression he would refrain from such conduct. Plaintiff maintained that her statement during the game established that a disputed factual issue existed as to whether she voluntarily had chosen to assume the risks of the type of conduct allegedly engaged in by defendant.
*303In his reply to plaintiff’s opposition, defendant acknowledged there were some factual details—“who ran where, when and how”—that were in dispute. He contended, however, that the material facts were not in dispute, stating those facts were “that plaintiff was injured in the context of playing touch football.”
After considering the parties’ submissions, the trial court granted defendant’s motion for summary judgment. On appeal, the Court of Appeal, recognizing the existing conflict in appellate court decisions with regard to the so-called “reasonable implied assumption of risk” doctrine, concluded that Ordway v. Superior Court, supra, 198 Cal.App.3d 98, rather than Segoviano v. Housing Authority, supra, 143 Cal.App.3d 162, should be followed, and further concluded that under the Ordway decision there were no disputed material facts to be determined. The Court of Appeal, holding that the trial court properly had granted summary judgment in favor of defendant, affirmed the judgment.
As noted, we granted review to resolve the conflict among Court of Appeal decisions as to the proper application of the assumption of risk doctrine in light of the adoption of comparative fault principles in Li, supra, 13 Cal.3d 804.
II
As every leading tort treatise has explained, the assumption of risk doctrine long has caused confusion both in definition and application, because the phrase “assumption of risk” traditionally has been used in a number of very different factual settings involving analytically distinct legal concepts. (See, e.g., Prosser & Keeton on Torts (5th ed. 1984) § 68, pp. 480-481; 4 Harper et al., The Law of Torts (2d ed. 1986) § 21.0, pp. 187-189; Schwartz, Comparative Negligence (2d ed. 1986) § 9.1, p. 154; 3 Speiser et al., The American Law of Torts (1986) §§ 12:46-12:47, pp. 636-640.) Indeed, almost a half-century ago, Justice Frankfurter described the term “assumption of risk” as a classic example of a felicitous phrase, “undiscriminatingly used to express different and sometimes contradictory ideas,” and whose uncritical use “bedevils the law.” (Tiller v. Atlantic Coast Line R. Co. (1943) 318 U.S. 54, 68 [87 L.Ed. 610, 618, 63 S.Ct. 444, 143 A.L.R. 967] (cone. opn. of Frankfurter, J.).)
In some settings—for example, most cases involving sports-related injuries—past assumption of risk decisions largely have been concerned with defining the contours of the legal duty that a given class of defendants—for example, owners of baseball stadiums or ice hockey rinks—owed to an *304injured plaintiff. (See, e.g., Quinn v. Recreation Park Assn. (1935) 3 Cal.2d 725, 729 [46 P.2d 144] [baseball stadium owner]; Shurman v. Fresno Ice Rink (1949) 91 Cal.App.2d 469, 474-477 [205 P.2d 77] [hockey rink owner].) In other settings, the assumption of risk terminology historically was applied to situations in which it was clear that the defendant had breached a legal duty of care to the plaintiff, and the inquiry focused on whether the plaintiff knowingly and voluntarily had chosen to encounter the specific risk of harm posed by the defendant’s breach of duty. (See, e.g., Vierra v. Fifth Avenue Rental Service (1963) 60 Cal.2d 266, 271 [32 Cal.Rptr. 193, 383 P.2d 777] [plaintiff hit in eye by flying piece of metal in area adjacent to drilling]; Prescott v. Ralphs Grocery Co. (1954) 42 Cal.2d 158, 161-162 [265 P.2d 904] [plaintiff injured on wet sidewalk on store premises].)
Prior to the adoption of comparative fault principles of liability, there often was no need to distinguish between the different categories of assumption of risk cases, because if a case fell into either category, the plaintiff’s recovery was totally barred. With the adoption of comparative fault, however, it became essential to differentiate between the distinct categories of cases that traditionally had been lumped together under the rubric of assumption of risk. This court’s seminal comparative fault decision in Li, supra, 13 Cal.3d 804, explicitly recognized the need for such differentiation, and attempted to explain which category of assumption of risk cases should be merged into the comparative fault system and which category should not. Accordingly, in considering the current viability of the assumption of risk doctrine in California, our analysis necessarily begins with the Li decision.
In Li, our court undertook a basic reexamination of the common law doctrine of contributory negligence. As Li noted, contributory negligence generally has been defined as “ ‘conduct on the part of the plaintiff which falls below the standard to which he should conform for his own protection, and which is a legally contributing cause cooperating with the negligence of the defendant in bringing about the plaintiff’s harm.’ ” (Li, supra, 13 Cal.3d at p. 809, quoting Rest.2d Torts, § 463.) Prior to Li, the common law rule was that “‘[e]xcept where the defendant has the last clear chance, the plaintiff’s contributory negligence bars recovery against a defendant whose negligent conduct would otherwise make him liable to the plaintiff for the harm sustained by him.’ ” (Li, supra, at pp. 809-810, italics added, quoting Rest.2d Torts, § 467.)
In Li, supra, 13 Cal.3d 804, we observed that “[i]t is unnecessary for us to catalogue the enormous amount of critical comment that has been directed over the years against the ‘all-or-nothing’ approach of the doctrine of contributory negligence. The essence of that criticism has been constant and *305clear: the doctrine is inequitable in its operation because it fails to distribute responsibility in proportion to fault .... The basic objection to the doctrine—grounded in the primal concept that in a system in which liability is based on fault, the extent of fault should govern the extent of liability— remains irresistible to reason and all intelligent notions of fairness.” (Id. at pp. 810-811, italics added.) After taking additional note of the untoward practical consequences of the doctrine in the litigation of cases and the increasing rejection of the doctrine in other jurisdictions, the Li court concluded that “[w]e are likewise persuaded that logic, practical experience, and fundamental justice counsel against the retention of the doctrine rendering contributory negligence a complete bar to recovery—and that it should be replaced in this state by a system under which liability for damage will be borne by those whose negligence caused it in direct proportion to their respective fault.” (Id. at pp. 812-813.)
After determining that the “all-or-nothing” contributory negligence doctrine should be replaced by a system of comparative negligence, the Li court went on to undertake a rather extensive discussion of the effect that the adoption of comparative negligence would have on a number of related tort doctrines, including the doctrines of last clear chance and assumption of risk. (Li, supra, 13 Cal.3d at pp. 823-826.)
Under the last clear chance doctrine, a defendant was rendered totally liable for an injury, even though the plaintiff’s contributory negligence had played a role in the accident, when the defendant had the “last clear chance” to avoid the accident. With regard to that doctrine, the Li decision, supra, 13 Cal.3d 804, observed: “Although several states which apply comparative negligence concepts retain the last clear chance doctrine [citation], the better reasoned position seems to be that when true comparative negligence is adopted, the need for last clear chance as a palliative of the hardships of the ‘all-or-nothing’ rule disappears and its retention results only in a windfall to the plaintiff in direct contravention of the principle of liability in proportion to fault. [Citations.]” (Id. at p. 824.) Accordingly, the court concluded that the doctrine should be “subsumed under the general process of assessing liability in proportion to fault.” (Id. at p. 826.)
With respect to the effect of the adoption of comparative negligence on the assumption of risk doctrine—the issue before us today—the Li decision, supra, 13 Cal.3d 804, stated as follows: “As for assumption of risk, we have recognized in this state that this defense overlaps that of contributory negligence to some extent and in fact is made up of at least two distinct defenses. ‘To simplify greatly, it has been observed . . . that in one kind of situation, to wit, where a plaintiff unreasonably undertakes to encounter a *306specific known risk imposed by a defendant’s negligence, plaintiff’s conduct, although he may encounter that risk in a prudent manner, is in reality a form of contributory negligence .... Other kinds of situations within the doctrine of assumption of risk are those, for example, where plaintiff is held to agree to relieve defendant of an obligation of reasonable conduct toward him. Such a situation would not involve contributory negligence, but rather a reduction of defendant’s duty of care.’ (Grey v. Fibreboard Paper Products Co. (1966) 65 Cal.2d 240, 245-246 [53 Cal.Rptr. 545,418 P.2d 153]; see also Fonseca v. County of Orange (1972) 28 Cal.App.3d 361, 368-369 [104 Cal.Rptr. 566]; see generally, 4 Witkin, Summary of Cal. Law [(8th ed. 1974)], Torts, § 723, pp. 3013-3014; 2 Harper & James, The Law of Torts [(1st ed. 1956)] §21.1, pp. 1162-1168; cf. Prosser, Torts [(4th ed. 1971)] § 68, pp. 439-441.) We think it clear that the adoption of a system of comparative negligence should entail the merger of the defense of assumption of risk into the general scheme of assessment of liability in proportion to fault in those particular cases in which the form of assumption of risk involved is no more than a variant of contributory negligence. (See generally, Schwartz, [Comparative Negligence (1st ed. 1974)] ch. 9, pp. 153-175.)” (Li. supra, 13 Cal.3d at pp. 824-825, original italics.)
As this passage indicates, the Li decision, supra, 13 Cal.3d 804, clearly contemplated that the assumption of risk doctrine was to be partially merged or subsumed into the comparative negligence scheme. Subsequent Court of Appeal decisions have disagreed, however, in interpreting Li, as to what category of assumption of risk cases would be merged into the comparative negligence scheme.
Á number of appellate decisions, focusing on the language in Li indicating that assumption of risk is in reality a form of contributory negligence “where a plaintiff unreasonably undertakes to encounter a specific known risk imposed by a defendant’s negligence” (13 Cal.3d at p. 824), have concluded that Li properly should be interpreted as drawing a distinction between those assumption of risk cases in which a plaintiff “unreasonably” encounters a known risk imposed by a defendant’s negligence and those assumption of risk cases in which a plaintiff “reasonably” encounters a known risk imposed by a defendant’s negligence. (See, e.g., Ordway v. Superior Court, supra, 198 Cal.App.3d 98, 103-105.) These decisions interpret Li as subsuming into the comparative fault scheme those cases in which the plaintiff acts unreasonably in encountering a specific known risk, but retaining the assumption of risk doctrine as a complete bar to recovery in those cases in which the plaintiff acts reasonably in encountering such a risk. Although aware of the apparent anomaly of a rule under which a plaintiff who acts reasonably is completely barred from recovery while a plaintiff who acts unreasonably *307only has his or her recovery reduced, these decisions nonetheless have concluded that this distinction and consequence were intended by the Li court.2
In our view, these decisions—regardless whether they reached the correct result on the facts at issue—have misinterpreted Li by suggesting that our decision contemplated less favorable legal treatment for a plaintiff who reasonably encounters a known risk than for a plaintiff who unreasonably encounters such a risk. Although the relevant passage in Li indicates that the assumption of risk doctrine would be merged into the comparative fault scheme in instances in which a plaintiff “ ‘unreasonably undertakes to encounter a specific known risk imposed by a defendant’s negligence’ ” (13 Cal.3d at p. 824), nothing in this passage suggests that the assumption of risk doctrine should survive as a total bar to the plaintiff’s recovery whenever a plaintiff acts reasonably in encountering such a risk. Instead, this portion of our opinion expressly contrasts the category of assumption of risk cases which “ ‘involve contributory negligence’ ” (and which therefore should be merged into the comparative fault scheme) with those assumption of risk *308cases which involve “ ‘a reduction of defendant’s duty of care.’ ” (Id. at p. 825.)
Indeed, particularly when the relevant passage in Li, supra, 13 Cal.3d at pages 824-825, is read as a whole and in conjunction with the authorities it cites, we believe it becomes clear that the distinction in assumption of risk cases to which the Li court referred in this passage was not a distinction between instances in which a plaintiff unreasonably encounters a known risk imposed by a defendant’s negligence and instances in which a plaintiff reasonably encounters such a risk. Rather, the distinction to which the Li court referred was between (1) those instances in which the assumption of risk doctrine embodies a legal conclusion that there is “no duty” on the part of the defendant to protect the plaintiff from a particular risk—the category of assumption of risk that the legal commentators generally refer to as “primary assumption of risk”—and (2) those instances in which the defendant does owe a duty of care to the plaintiff but the plaintiff knowingly encounters a risk of injury caused by the defendant’s breach of that duty— what most commentators have termed “secondary assumption of risk.”3 Properly interpreted, the relevant passage in Li provides that the category of assumption of risk cases that is not merged into the comparative negligence system and in which the plaintiff’s recovery continues to be completely barred involves those cases in which the defendant’s conduct did not breach a legal duty of care to the plaintiff, i.e., “primary assumption of risk” cases, whereas cases involving “secondary assumption of risk” properly are merged into the comprehensive comparative fault system adopted in Li.4
*309Although the difference between the “primary assumption of risk”/“secondary assumption of risk” nomenclature and the “reasonable implied assumption of risk”/“unreasonable implied assumption of risk” terminology embraced in many of the recent Court of Appeal decisions may appear at first blush to be only semantic, the significance extends beyond mere rhetoric. First, in “primary assumption of risk” cases—where the defendant owes no duty to protect the plaintiff from a particular risk of harm—a plaintiff who has suffered such harm is not entitled to recover from the defendant, whether the plaintiff’s conduct in undertaking the activity was reasonable or unreasonable. Second, in “secondary assumption of risk” cases—involving instances in which the defendant has breached the duty of care owed to the plaintiff—the defendant is not entitled to be entirely relieved of liability for an injury proximately caused by such breach, simply because the plaintiff’s conduct in encountering the risk of such an injury was reasonable rather than unreasonable. Third and finally, the question whether the defendant owed a legal duty to protect the plaintiff from a particular risk of harm does not turn on the reasonableness or unreasonableness of the plaintiff’s conduct, but rather on the nature of the activity or sport in which the defendant is engaged and the relationship of the defendant and the plaintiff to that activity or sport. For these reasons, use of the “reasonable implied assumption of risk”/“unreasonable implied assumption of risk” terminology, as a means of differentiating between the cases in which a plaintiff is barred from bringing an action and those in which he or she is not barred, is more misleading than helpful.5
*310Our reading of Li, supra, 13 Cal.3d 804, insofar as it draws a distinction between assumption of risk cases in which the defendant has not breached any legal duty to the plaintiff and those in which the defendant has breached a legal duty, is supported not only by the language of Li itself and the authorities it cites, but also, and perhaps most significantly, by the fundamental principle that led the Li court to replace the all-or-nothing contributory negligence defense with a comparative fault scheme. In “primary assumption of risk” cases, it is consistent with comparative fault principles totally to bar a plaintiff from pursuing a cause of action, because when the defendant has not breached a legal duty of care to the plaintiff, the defendant has not committed any conduct which would warrant the imposition of any liability whatsoever, and thus there is no occasion at all for invoking comparative fault principles. (See Prosser & Keeton on Torts, supra, § 68, at pp. 496-497.) By contrast, in the “secondary assumption of risk” context, the defendant has breached a duty of care owed to the plaintiff. When a risk of harm is created or imposed by a defendant’s breach of duty, and a plaintiff who chose to encounter the risk is injured, comparative fault principles preclude automatically placing all of the loss on the plaintiff, because the injury in such a case may have been caused by the combined effect of the defendant’s and the plaintiff’s culpable conduct. To retain assumption of risk as a complete defense in such a case would fly in the face of Li’s basic holding that when both parties are partially at fault for an injury, a rule which places all of the loss on one of the parties is inherently inequitable. (See id. at pp. 497-498.)
Thus, just as the court in Li reasoned it would be improper to retain the last clear chance doctrine as a means of imposing all liability on a defendant in cases in which the defendant is aware of the risk of harm created by the plaintiff’s negligence but fails to take the “last clear chance” to avoid the injury (Li, supra, 13 Cal.3d at p. 824), we believe the Li court similarly recognized that, in the assumption of risk context, it would be improper to *311impose all responsibility on a plaintiff who is aware of a risk of harm created by the defendant’s breach of duty but fails to avert the harm. In both instances, comparative fault principles call for a sharing of the burden of liability.
The dissenting opinion suggests, however, that, even when a defendant has breached its duty of care to the plaintiff, a plaintiff who reasonably has chosen to encounter a known risk of harm imposed by such a breach may be totally precluded from recovering any damages, without doing violence to comparative fault principles, on the theory that the plaintiff, by proceeding in the face of a known risk, has “impliedly consented” to any harm. (See dis. opn. by Kennard, J., post, pp. 331-333.) For a number of reasons, we conclude this contention does not withstand analysis.
First, the argument that a plaintiff who proceeds to encounter a known risk has “impliedly consented” to absolve a negligent defendant of liability for any ensuing harm logically would apply as much to a plaintiff who unreasonably has chosen to encounter a known risk, as to a plaintiff who reasonably has chosen to encounter such a risk. As we have seen, however, Li explicitly held that a plaintiff who “ ‘unreasonably undertakes to encounter a specific known risk imposed by a defendant’s negligence’ ” (Li, supra, 13 Cal.3d at p. 824) is not completely barred from recovery; instead, the recovery of such a plaintiff simply is reduced under comparative fault principles. Thus, the dissenting opinion’s implied consent argument is irreconcilable with Li itself.
Second, the implied consent rationale rests on a legal fiction that is untenable, at least as applied to conduct that represents a breach of the defendant’s duty of care to the plaintiff. It may be accurate to suggest that an individual who voluntarily engages in a potentially dangerous activity or sport “consents to” or “agrees to assume” the risks inherent in the activity or sport itself, such as the risks posed to a snow skier by moguls on a ski slope or the risks posed to a water skier by wind-whipped waves on a lake. But it is thoroughly unrealistic to suggest that, by engaging in a potentially dangerous activity or sport, an individual consents to (or agrees to excuse) a breach of duty by others that increases the risks inevitably posed by the activity or sport itself, even where the participating individual is aware of the possibility that such misconduct may occur.
A familiar example may help demonstrate this point. Although every driver of an automobile is aware that driving is a potentially hazardous activity and that inherent in the act of driving is the risk that he or she will be injured by the negligent driving of another, a person who voluntarily *312chooses to drive does not thereby “impliedly consent” to being injured by the negligence of another, nor has such a person “impliedly excused” others from performing their duty to use due care for the driver’s safety. Instead, the driver reasonably expects that if he or she is injured by another’s negligence, i.e., by the breach of the other person’s duty to use due care, the driver will be entitled to compensation for his or her injuries. Similarly, although a patient who undergoes elective surgery is aware that inherent in such an operation is the risk of injury in the event the surgeon is negligent, the patient, by voluntarily encountering such a risk, does not “impliedly consent” to negligently inflicted injury or “impliedly agree” to excuse the surgeon from a normal duty of care, but rather justifiably expects that the surgeon will be liable in the event of medical malpractice.
Thus, there is no merit to the dissenting opinion’s general claim that simply because a person is aware an activity involves a risk of harm that may arise from another’s negligence and voluntarily proceeds to participate in that activity despite such knowledge, that person should be barred from obtaining any recovery on the theory that he or she impliedly consented to the risk of harm. As we shall discuss in part III, legal liability for an injury which occurs during a sporting event is significantly affected by the assumption of risk doctrine, but only because the doctrine has been utilized in framing the duty of care owed by a defendant in the context of a sporting event, and not because the plaintiff in such a case has, in any realistic sense of the term, “consented” to relieve the defendant of liability.
Third, the dissenting opinion’s claim that the category of cases in which the assumption of risk doctrine operates to bar a plaintiff’s cause of action after Li properly should be gauged on the basis of an implied consent analysis, rather than on the duty analysis we have described above, is, in our view, untenable for another reason. In support of its implied consent theory, the dissenting opinion relies on a number of pre-Li cases, which arose in the “secondary assumption of risk” context, and which held that, in such a context, application of the assumption of risk doctrine was dependent on proof that the particular plaintiff subjectively knew, rather than simply should have known, of both the existence and magnitude of the specific risk of harm imposed by the defendant’s negligence. (See Vierra v. Fifth Avenue Rental Service, supra, 60 Cal.2d 266, 271-275; Prescott v. Ralphs Grocery Co., supra, 42 Cal.2d 158,161-162.) Consequently, as the dissenting opinion acknowledges, were its implied consent theory to govern application of the assumption of risk doctrine in the sports setting, the basic liability of a defendant who engages in a sport would depend on variable factors that the defendant frequently would have no way of ascertaining (for example, the particular plaintiff’s subjective knowledge and expectations), rather than on *313the nature of the sport itself. As a result, there would be drastic disparities in the manner in which the law would treat defendants who engaged in precisely the same conduct, based on the often unknown, subjective expectations of the particular plaintiff who happened to be injured by the defendant’s conduct.
Such an approach not only would be inconsistent with the principles of fairness underlying the Li decision, but also would be inimical to the fair and efficient administration of justice. If the application of the assumption of risk doctrine in a sports setting turned on the particular plaintiff’s subjective knowledge and awareness, summary judgment rarely would be available in such cases, for, as the present case reveals, it frequently will be easy to raise factual questions with regard to a particular plaintiff’s subjective expectations as to the existence and magnitude of the risks the plaintiff voluntarily chose to encounter. By contrast, the question of the existence and scope of a defendant’s duty of care is a legal question which depends on the nature of the sport or activity in question and on the parties’ general relationship to the activity, and is an issue to be decided by the court, rather than the jury. (See, e.g., 6 Witkin, Summary of Cal. Law, supra, Torts, § 748, pp. 83-86 and cases cited.) Thus, the question of assumption of risk is much more amenable to resolution by summary judgment under a duty analysis than under the dissenting opinion’s suggested implied consent approach.
An amicus curiae in the companion case has questioned, on a separate ground, the duty approach to the post-Li assumption of risk doctrine, suggesting that if a plaintiff’s action may go forward whenever a defendant’s breach of duty has played some role, however minor, in a plaintiff’s injury, a plaintiff who voluntarily engages in a highly dangerous sport—for example, skydiving or mountain climbing—will escape any responsibility for tiie injury so long as a jury finds that the plaintiff was not “unreasonable” in engaging in the sport. This argument rests on the premise that, under comparative fault principles, a jury may assign some portion of the responsibility for an injury to a plaintiff only if the jury finds that the plaintiff acted unreasonably, but not if the jury finds that the plaintiff knowingly and voluntarily, but reasonably, chose to engage in a dangerous activity. Amicus curiae contends that such a rule frequently would permit voluntary risk takers to avoid all responsibility for their own actions, and would impose an improper and undue burden on other participants.
Although we agree with the general thesis of amicus curiae’s argument that persons generally should bear personal responsibility for their own actions, the suggestion that a duty approach to the doctrine of assumption of risk is inconsistent with this thesis rests on a mistaken premise. Past *314California cases have made it clear that the “comparative fault” doctrine is a flexible, commonsense concept, under which a jury properly may consider and evaluate the relative responsibility of various parties for an injury (whether their responsibility for the injury rests on negligence, strict liability, or other theories of responsibility), in order to arrive at an “equitable apportionment or allocation of loss.” (See Daly v. General Motors Corp. (1978) 20 Cal.3d 725, 734-742 [144 Cal.Rptr. 380, 575 P.2d 1162]; Safeway Stores, Inc. v. Nest-Kart (1978) 21 Cal.3d 322, 328-332 [146 Cal.Rptr. 550, 579 P.2d 441]; Far West Financial Corp. v. D & S Co. (1988) 46 Cal.3d 796, 804, fn. 7 [251 Cal.Rptr. 202, 760 P.2d 399].)
Accordingly, contrary to amicus curiae’s assumption, we believe that under California’s comparative fault doctrine, a jury in a “secondary assumption of risk” case would be entitled to take into consideration a plaintiff’s voluntary action in choosing to engage in an unusually risky sport, whether or not the plaintiff’s decision to encounter the risk should be characterized as unreasonable, in determining whether the plaintiff properly should bear some share of responsibility for the injuries he or she suffered. (See, e.g., Kirk v. Washington State University (1987) 109 Wn.2d 448 [746 P.2d 285, 290-291]. See generally Schwartz, Comparative Negligence, supra, § 9.5, p. 180; Diamond, Assumption of Risk After Comparative Negligence: Integrating Contract Theory into Tort Doctrine (1991) 52 Ohio St. LJ. 717, 748-749.) Thus, in a case in which an injury has been caused by both a defendant’s breach of a legal duty to the plaintiff and the plaintiff’s voluntary decision to engage in an unusually risky sport, application of comparative fault principles will not operate to relieve either individual of responsibility for his or her actions, but rather will ensure that neither party will escape such responsibility.
It may be helpful at this point to summarize our general conclusions as to the current state of the doctrine of assumption of risk in light of the adoption of comparative fault principles in Li, supra, 13 Cal.3d 804, general conclusions that reflect the view of a majority of the justices of the court (i.e., the three justices who have signed this opinion and Justice Mosk (see cone, and dis. opn. by Mosk, J., post, p. 321)).6 In cases involving “primary assumption of risk”—where, by virtue of the nature of the activity and the parties’ *315relationship to the activity, the defendant owes no legal duty to protect the plaintiff from the particular risk of harm that caused the injury—the doctrine continues to operate as a complete bar to the plaintiff’s recovery. In cases involving “secondary assumption of risk”—where the defendant does owe a duty of care to the plaintiff, but the plaintiff proceeds to encounter a known risk imposed by the defendant’s breach of duty—the doctrine is merged into the comparative fault scheme, and the trier of fact, in apportioning the loss resulting from the injury, may consider the relative responsibility of the parties.
Accordingly, in determining the propriety of the trial court’s grant of summary judgment in favor of the defendant in this case, our inquiry does not turn on the reasonableness or unreasonableness of plaintiff’s conduct in choosing to subject herself to the risks of touch football or in continuing to participate in the game after she became aware of defendant’s allegedly rough play. Nor do we focus upon whether there is a factual dispute with regard to whether plaintiff subjectively knew of, and voluntarily chose to encounter, the risk of defendant’s conduct, or impliedly consented to relieve or excuse defendant from any duty of care to her. Instead, our resolution of this issue turns on whether, in light of the nature of the sporting activity in which defendant and plaintiff were engaged, defendant’s conduct breached a legal duty of care to plaintiff. We now turn to that question.
Ill
As a general rule, persons have a duty to use due care to avoid injury to others, and may be held liable if their careless conduct injures another person. (See Civ. Code, § 1714.) Thus, for example, a property owner ordinarily is required to use due care to eliminate dangerous conditions on his or her property. (See, e.g., Rowland, v. Christian (1968) 69 Cal.2d 108 [70 Cal.Rptr. 97, 443 P.2d 561, 32 A.L.R.3d 496].) In the sports setting, however, conditions or conduct that otherwise might be viewed as dangerous often are an integral part of the sport itself. Thus, although moguls on a ski run pose a risk of harm to skiers that might not exist were these configurations removed, the challenge and risks posed by the moguls are part of the sport of skiing, and a ski resort has no duty to eliminate them. (See generally Annot. (1987) 55 A.L.R.4th 632.) In this respect, the nature of a sport is highly relevant in defining the duty of care owed by the particular defendant.
Although defendants generally have no legal duty to eliminate (or protect a plaintiff against) risks inherent in the sport itself, it is well *316established that defendants generally do have a duty to use due care not to increase the risks to a participant over and above those inherent in the sport. Thus, although a ski resort has no duty to remove moguls from a ski run, it clearly does have a duty to use due care to maintain its towropes in a safe, working condition so as not to expose skiers to an increased risk of harm. The cases establish that the latter type of risk, posed by a ski resort’s negligence, clearly is not a risk (inherent in the sport) that is assumed by a participant. (See generally Annot. (1979) 95 A.L.R.3d 203.)
In some situations, however, the careless conduct of others is treated as an “inherent risk” of a sport, thus barring recovery by the plaintiff. For example, numerous cases recognize that in a game of baseball, a player generally cannot recover if he or she is hit and injured by a carelessly thrown ball (see, e.g., Mann v. Nutrilite, Inc. (1955) 136 Cal.App.2d 729, 734-735 [289 P.2d 282]), and that in a game of basketball, recovery is not permitted for an injury caused by a carelessly extended elbow (see, e.g., Thomas v. Barlow (1927) 5 N.J. Misc. 764 [138 A. 208]). The divergent results of the foregoing cases lead naturally to the question how courts are to determine when careless conduct of another properly should be considered an “inherent risk” of the sport that (as a matter of law) is assumed by the injured participant.
Contrary to the implied consent approach to the doctrine of assumption of risk, discussed above, the duty approach provides an answer which does not depend on the particular plaintiff’s subjective knowledge or appreciation of the potential risk. Even where the plaintiff, who falls while skiing over a mogul, is a total novice and lacks any knowledge of skiing whatsoever, the ski resort would not be liable for Ms or her injuries. (See Brown v. San Francisco Baseball Club (1950) 99 Cal.App.2d 484, 488-492 [222 P.2d 19] [baseball spectator’s alleged ignorance of the game did not warrant imposing liability on stadium owner for injury caused by a carelessly tMown ball].) And, on the other hand, even where the plaintiff actually is aware that a particular ski resort on occasion has been negligent in maintaimng its towropes, that knowledge would not preclude the skier from recovering if he or she were injured as a result of the resort’s repetition of such deficient conduct. In the latter context, although the plaintiff may have acted with knowledge of the potential negligence, he or she did not consent to such negligent conduct or agree to excuse the resort from liability in the event of such negligence.
Rather than being dependent on the knowledge or consent of the particular plaintiff, resolution of the question of the defendant’s liability in such cases turns on whether the defendant had a legal duty to avoid such conduct or to *317protect the plaintiff against a particular risk of harm. As already noted, the nature of a defendant’s duty in the sports context depends heavily on the nature of the sport itself. Additionally, the scope of the legal duty owed by a defendant frequently will also depend on the defendant’s role in, or relationship to, the sport.
The latter point is demonstrated by a review of one of the numerous cases involving an injury sustained by a spectator at a baseball game. In Ratcliff v. San Diego Baseball Club (1938) 27 Cal.App.2d 733 [81 P.2d 625], a baseball spectator was injured when, walking in the stands between home plate and first base during a game, she was hit by an accidentally thrown bat. She sued both the player who threw the bat and the baseball stadium owner. The jury returned a verdict in favor of the player, but found the stadium owner liable. On appeal, the Court of Appeal affirmed.
Had the Ratcliff court utilized an implied consent analysis, the court would have looked only to the knowledge of the particular plaintiff (the spectator) to determine whether the risk of being hit by an accidentally thrown bat was an inherent risk of the sport of baseball assumed by the plaintiff, and would have treated the plaintiff’s action against both defendants similarly with regard to such risk. The Ratcliff court did not analyze the case in that manner, however. Instead, the court implicitly recognized that two different potential duties were at issue—(1) the duty of the ballplayer to play the game without carelessly throwing his bat, and (2) the duty of the stadium owner to provide a reasonably safe stadium with regard to the relatively common (but particularly dangerous) hazard of a thrown bat. Because each defendant’s liability rested on a separate duty, there was no inconsistency in the jury verdict absolving the batter of liability but imposing liability on the stadium owner for its failure to provide the patron “protection from flying bats, at least in the area where the greatest danger exists and where such an occurrence is reasonably to be expected.” (Ratcliff v. San Diego Baseball Club, supra, 27 Cal.App.2d at p. 736.)
Other cases also have analyzed in a similar fashion the duty of the owner of a ballpark or ski resort, in the process defining the risks inherent in the sport not only by virtue of the nature of the sport itself, but also by reference to the steps the sponsoring business entity reasonably should be obligated to take in order to minimize the risks without altering the nature of the sport. (See, e.g., Quinn v. Recreation Park Assn., supra, 3 Cal.2d 725, 728-729 [discussing separately the potential liability of a player and a baseball stadium owner for injury to a spectator]; Shurman v. Fresno Ice Rink, supra, 91 Cal.App.2d 469, 474-477 [discussing duty owed by owner of ice hockey rink to spectators].)
*318Even a cursory review of the numerous sports injury cases reveals the diverse categories of defendants whose alleged misconduct may be at issue in such cases. Thus, for example, suits have been brought against owners of sports facilities such as baseball stadiums and ski resorts (see, e.g., Quinn v. Recreation Park Assn., supra, 3 Cal.2d 725; Danieley v. Goldmine Ski Associates, Inc. (1990) 218 Cal.App.3d 111 [266 Cal.Rptr. 749]), against manufacturers and reconditioners of sporting equipment (see, e.g., Holdsworth v. Nash Mfg., Inc. (1987) 161 Mich.App. 139 [409 N.W.2d 764]; Gentile v. MacGregor Mfg. Co. (1985) 201 N.J.Super. 612 [493 A.2d 647]), against sports instructors and coaches (see, e.g., Scroggs v. Coast Community College Dist. (1987) 193 Cal.App.3d 1399 [239 Cal.Rptr. 916]; Morris v. Union High School Dist. A (1931) 160 Wash. 121 [294 P. 998]), and against coparticipants (see, e.g., Tavernier v. Maes (1966) 242 Cal.App.2d 532 [51 Cal.Rptr. 575]), alleging that such persons, either by affirmative misconduct or by a failure to act, caused or contributed to the plaintiff’s injuries. These cases demonstrate that in the sports setting, as elsewhere, the nature of the applicable duty or standard of care frequently varies with the role of the defendant whose conduct is at issue in a given case.
In the present case, defendant was a participant in the touch football game in which plaintiff was engaged at the time of her injury, and thus the question before us involves the circumstances under which a participant in such a sport may be held liable for an injury sustained by another participant.
The overwhelming majority of the cases, both within and outside California, that have addressed the issue of coparticipant liability in such a sport, have concluded that it is improper to hold a sports participant liable to a coparticipant for ordinary careless conduct committed during the sport— for example, for an injury resulting from a carelessly thrown ball or bat during a baseball game—and that liability properly may be imposed on a participant only when he or she intentionally injures another player or engages in reckless conduct that is totally outside the range of the ordinary activity involved in the sport. (See, e.g., Gauvin v. Clark (1989) 404 Mass. 450 [537 N.E.2d 94, 96-97] and cases cited.)
In reaching the conclusion that a coparticipant’s duty of care should be limited in this fashion, the cases have explained that, in the heat of an active sporting event like baseball or football, a participant’s normal energetic conduct often includes accidentally careless behavior. The courts have concluded that vigorous participation in such sporting events likely would be chilled if legal liability were to be imposed on a participant on the basis of his or her ordinary careless conduct. The cases have recognized that, in such a sport, even when a participant’s conduct violates a rule of the game and *319may subject the violator to internal sanctions prescribed by the sport itself, imposition of legal liability for such conduct might well alter fundamentally the nature of the sport by deterring participants from vigorously engaging in activity that falls close to, but on the permissible side of, a prescribed rule.
A sampling of the cases that have dealt with the question of the potential tort liability of such sports participants is instructive. In Tavernier v. Maes, supra, 242 Cal.App.2d 532, for example, the Court of Appeal upheld a verdict denying recovery for an injury sustained by the plaintiff second baseman as an unintended consequence of the defendant baserunner’s hard slide into second base during a family picnic softball game. Similarly, in Gaspard v. Grain Dealers Mutual Insurance Company (La.Ct.App. 1961) 131 So.2d 831, the plaintiff baseball player was denied recovery when he was struck on the head by a bat which accidentally flew out of the hands of the defendant batter during a school game. (See also Gauvin v. Clark, supra, 404 Mass. 450 [537 N.E.2d 94, 96-97] [plaintiff hockey player injured when hit with hockey stick by opposing player; court held that defendant’s liability should be determined by whether he acted “with reckless disregard of safety”]; Marchetti v. Kalish (1990) 53 Ohio.St.3d 95 [559 N.E.2d 699, 703] [child injured while playing “kick the can”; “we join the weight of authority . . . and require that before a party may proceed with a cause of action involving injury resulting from recreational or sports activity, reckless or intentional conduct must exist”]; Kabella v. Bouschelle (1983) 100 N.M. 461 [672 P.2d 290, 294] [plaintiff injured in informal tackle football game; court held that “a cause of action for personal injuries between participants incurred during athletic competition must be predicated upon recklessness or intentional conduct, ‘not mere negligence’ ”]; Ross v. Clouser (Mo. 1982) 637 S.W.2d 11, 13-14 [plaintiff third baseman injured in collision with baserunner; court held that “a cause of action for personal injuries incurred during athletic competition must be predicated on recklessness, not mere negligence”]; Moe v. Steenberg (1966) 275 Minn. 448 [147 N.W.2d 587, 33 A.L.R.3d 311] [plaintiff ice skater denied recovery for injury incurred when another skater, who was skating backwards, accidentally tripped over her after she had fallen on the ice]; Thomas v. Barlow, supra, 5 N.J. Mise. 764 [138 A. 208] [recovery denied when appellate court concluded that plaintiff’s injury, incurred during a basketball game, resulted from an accidental contact with a member of the opposing team].)
By contrast, in Griggas v. Clauson (1955) 6 Ill.App.2d 412 [128 N.E.2d 363], the court upheld liability imposed on the defendant basketball player who, during a game, wantonly assaulted a player on the opposing team, apparently out of frustration with the progress of the game. And, in Bourque v. Duplechin (La.Ct.App. 1976) 331 So.2d 40, the court affirmed a judgment *320imposing liability for an injury incurred during a baseball game when the defendant baserunner, in an ostensible attempt to break up a double play, ran into the plaintiff second baseman at full speed, without sliding, after the second baseman had thrown the ball to first base and was standing four to five feet away from second base toward the pitcher’s mound; in upholding the judgment, the court stated that defendant “was under a duty to play softball in the ordinary fashion without unsportsmanlike conduct or wanton injury to his fellow players.” (Id. at p. 42.) (See also Averill v. Luttrell (1957) 44 Tenn.App. 56 [311 S.W.2d 812] [defendant baseball catcher properly held liable when, deliberately and without warning, he hit a batter in the head with his fist]; Hackbart v. Cincinnati Bengals, Inc. (10th Cir. 1979) 601 F.2d 516 [trial court erred in absolving defendant football player of liability when, acting out of anger and frustration, he struck a blow with his forearm to the back of the head of an opposing player, who was kneeling on the ground watching the end of a pass interception play]; Overall v. Kadella (1984) 138 Mich.App. 351 [361 N.W.2d 352] [hockey player permitted to recover when defendant player intentionally punched him in the face at the conclusion of the game].)
In our view, the reasoning of the foregoing cases is sound. Accordingly, we conclude that a participant in an active sport breaches a legal duty of care to other participants—i.e., engages in conduct that properly may subject him or her to financial liability—only if the participant intentionally injures another player or engages in conduct that is so reckless as to be totally outside the range of the ordinary activity involved in the sport.7
As applied to the present case, the foregoing legal principle clearly supports the trial court’s entry of summary judgment in favor of defendant. The declarations filed in support of and in opposition to the summary judgment motion establish that defendant was, at most, careless or negligent in knocking over plaintiff, stepping on her hand, and injuring her finger. Although plaintiff maintains that defendant’s rough play as described in her declaration and the declaration of Andrea Starr properly can be characterized as “reckless,” the conduct alleged in those declarations is not even closely comparable to the kind of conduct—conduct so reckless as to be totally *321outside the range of the ordinary activity involved in the sport—that is a prerequisite to the imposition of legal liability upon a participant in such a sport.
Therefore, we conclude that defendant’s conduct in the course of the touch football game did not breach any legal duty of care owed to plaintiff. Accordingly, this case falls within the primary assumption of risk doctrine, and thus the trial court properly granted summary judgment in favor of defendant. Because plaintiff’s action is barred under the primary assumption of risk doctrine, comparative fault principles do not come into play.
The judgment of the Court of Appeal, upholding the summary judgment entered by the trial court, is affirmed.
Lucas, C. J., and Arabian, J., concurred.
The portion of defendant’s deposition attached to plaintiff’s opposition included the following passage:
“Q: .... [F]rom your perspective—and I asked this same question of both of your friends yesterday—is the standard of care in which you were going to be dealing with people out there in the play field different, in your opinion, when you’re playing in that kind of a game, that is, the one that happened on that day versus if you’re out there playing in the exact same place and with a bunch of guys and no girls.
“A: Yeah, it would be different. Yes.
“Q: So, theoretically, you should be much more careful when the women are out there than if it was a bunch of guys?
“A: Right.”
In Ordway v. Superior Court, supra, 198 Cal.App.3d 98, the court suggested that the differentiation in the treatment accorded reasonable and unreasonable plaintiffs under an approach viewing “reasonable implied assumption of risk” as a complete bar to recovery was only “superficially anomalous” (id. at p. 104), and could be explained by reference to “the expectation of the defendant. He or she is permitted to ignore reasonably assumed risks and is not required to take extraordinary precautions with respect to them. The defendant must, however, anticipate that some risks will be unreasonably undertaken, and a failure to guard against these may result in liability.” (Id. at p. 105.)
Even when the matter is viewed from the defendant’s perspective, however, this suggested dichotomy is illogical and untenable. From the standpoint of a potential defendant, it is far more logical to require that the defendant take precautions with respect to risks that the defendant reasonably can foresee being undertaken, than it would be to impose liability only for risks that the defendant is less likely to anticipate will be encountered.
Ordway also attempted to explain the anomaly by reformulating the distinction between reasonable and unreasonable assumption of risk as one between plaintiffs who make a “knowing and intelligent” choice and those who act “negligently] or carelessfly]” (Ordway v. Superior Court, supra, 198 Cal.App.3d 98, 105), and the dissenting opinion cites this reformulated terminology with approval. (See dis. opn. by Kennard, J., post, p. 332.) The Li decision, however, specifically subsumed within comparative fault those assumption of risk cases in which a defendant “ ‘unreasonably undertakes to encounter a specific known risk’ ” (Li, supra, 13 Cal.Sd 804, 824, italics omitted and added), i.e., cases in which a defendant makes a knowing, but unreasonable, choice to undertake a risk. Indeed, in recasting the “unreasonable” assumption of risk category to include only those cases in which the plaintiff merely was careless and did not act with actual knowledge of the risk, Ordway inadvertently redefined the unreasonable assumption of risk category out of existence. The pre-Li decisions clearly held that where a plaintiff was injured as the result of a defendant’s breach of duty, the assumption of risk doctrine applied only to those instances in which the plaintiff actually knew of and appreciated the specific risk and nonetheless chose to encounter the risk. (See, e.g., Vierra v. Fifth Avenue Rental Service, supra, 60 Cal.2d 266, 271 [“Actual, and not merely constructive, knowledge of the danger is required.”].)
The introductory passage from the Harper and James treatise on The Law of Torts, that was cited with approval in Li, stated in this regard: “The term assumption of risk has led to no little confusion because it is used to refer to at least two different concepts, which largely overlap, have a common cultural background, and often produce the same legal result. But these concepts are nevertheless quite distinct rules involving slightly different policies and different conditions for their application. (1) In its primary sense the plaintiff’s assumption of a risk is only the counterpart of the defendant’s lack of duty to protect the plaintiff from that risk. In such a case plaintiff may not recover for his injury even though he was quite reasonable in encountering the risk that caused it. Volenti non fit injuria. (2) A plaintiff may also be said to assume a risk created by defendant’s breach of duty towards him, when he deliberately chooses to encounter that risk. In such a case, except possibly in master and servant cases, plaintiff will be barred from recovery only if he was unreasonable in encountering the risk under the circumstances. This is a form of contributory negligence. Hereafter we shall call this ‘assumption of risk in a secondary sense.’ ” (2 Harper & James, The Law of Torts (1st ed. 1956) § 21.1, p. 1162, fns. omitted, cited in Li, supra, 13 Cal.3d 804, 825.)
Although in the academic literature “express assumption of risk” often has been designated as a separate, contract-based species of assumption of risk distinct from both primary and secondary assumption of risk (see, e.g., Prosser & Keeton on Torts (5th ed. 1984) § 68, p. 496), cases involving express assumption of risk are concerned with instances in which, as the result of an express agreement, the defendant owes no duty to protect the plaintiff from an injury-causing risk. Thus in this respect express assumption of risk properly can be viewed as *309analogous to primary assumption of risk. One leading treatise describes express assumption of risk in the following terms: “In its most basic sense, assumption of risk means that the plaintiff, in advance, has given his express consent to relieve the defendant of an obligation of conduct toward him, and to take his chances of injury from a known risk arising from what the defendant is to do or leave undone .... The result is that the defendant is relieved of legal duty to the plaintiff; and being under no duty, he cannot be charged with negligence.” (Prosser & Keeton on Torts, supra, § 68, pp. 480-481, fn. omitted, second italics added.)
Since Li, California cases uniformly have recognized that so long as an express assumption of risk agreement does not violate public policy (see, e.g., Tunkl v. Regents of University of California (1963) 60 Cal.2d 92, 95-101 [32 Cal.Rptr. 33, 383 P.2d 441, 6 A.L.R.3d 693]), such an agreement operates to relieve the defendant of a legal duty to the plaintiff with respect to the risks encompassed by the agreement and, where applicable, to bar completely the plaintiff’s cause of action. (See, e.g., Madison v. Superior Court (1988) 203 Cal.App.3d 589, 597-602 [250 Cal.Rptr. 299], and cases cited.)
In addition to the sports setting, the primary assumption of risk doctrine also comes into play in the category of cases often described as involving the “firefighter’s rule.” (See Terhell v. American Commonwealth Associates (1985) 172 Cal.App.3d 434, 437 [218 Cal.Rptr. 256].) In its most classic form, the firefighter’s rule involves the question whether a person who negligently has started a fire is liable for an injury sustained by a firefighter who is summoned to fight the fire; the rule provides that the person who started the fire is not liable under such circumstances. (See, e.g., Walters v. Sloan (1977) 20 Cal.3d 199, 202 [142 Cal.Rptr. 152, 571 *310P.2d 609].) Although a number of theories have been cited to support this conclusion, the most persuasive explanation is that the party who negligently started the fire had no legal duty to protect the firefighter from the very danger that the firefighter is employed to confront. (See, e.g., Bakery. Superior Court (1982) 129 Cal.App.3d 710, 719-721 [181 Cal.Rptr. 311]; Nelson v. Hall (1985) 165 Cal.App.3d 709, 714 [211 Cal.Rptr. 668], See generally 6 Witkin, Summary of Cal. Law (9th ed. 1988) Torts, § 739, pp. 69-70 [discussing rule as one illustration of duty approach]; Anicet v. Gant (Fla.Dist.Ct.App. 1991) 580 So.2d 273, 276 [“a person specifically hired to encounter and combat particular dangers is owed no independent tort duty by those who have created those dangers . . . .”].) Because the defendant in such a case owes no duty to protect the firefighter from such risks, the firefighter has no cause of action even if the risk created by the fire was so great that a trier of fact could find it was unreasonable for the firefighter to choose to encounter the risk. This example again demonstrates that primary assumption of risk is not the same as “reasonable implied assumption of risk.”
Although Justice Mosk agrees that, in this context, a defendant’s liability should be analyzed under a duty analysis, he is of the view that the “primary” and “secondary" assumption of risk terminology is potentially confusing and would prefer entirely to eliminate the doctrine of implied assumption of risk as a bar to recovery and simply to apply comparative fault principles to determine liability. (See cone, and dis. opn. by Mosk, J., post, pp. 321-322.) Because the Li decision, supra, 13 Cal.3d 804, 824-825, indicated that the preexisting assumption of risk doctrine was to be only partially merged into the comparative fault system, the analysis set forth in the present opinion (distinguishing between primary and *315secondary assumption of risk) in our view more closely reflects the Li holding than does Justice Mosk’s proposal.
As suggested by the cases described in the text, the limited duty of care applicable to coparticipants has been applied in situations involving a wide variety of active sports, ranging from baseball to ice hockey and skating. Because the touch football game at issue in this case clearly falls within the rationale of this rule, we have no occasion to decide whether a comparable limited duty of care appropriately should be applied to other less active sports, such as archery or golf. We note that because of the special danger to others posed by the sport of hunting, past cases generally have found the ordinary duty of care to be applicable to hunting accidents. (See, e.g., Summers v. Tice (1948) 33 Cal.2d 80, 83 [199 P.2d 1, 5 A.L.R.2d 91].)