I dissent. To hold that the operator of this underground garage, with its alleged hiding places, missing lights, broken security cameras, absence of supervision, and other evidence of neglect, lacked any duty as a matter of law to maintain the premises in a fit and safe condition for business so as not to attract crime—and implicitly that any garage operator lacks such a duty—defies logic. And the result is demonstrably unjust.
I
Sharon P. alleged that she was sexually assaulted in an underground parking garage. She claimed in essence that the garage’s dilapidation attracted her assailant. The trial court ruled that plaintiff was not owed a duty of care and entered summary judgment for defendants. It reasoned that the crime was not sufficiently foreseeable to give rise to a duty of care.
A
To recover for the consequences of another’s purportedly wrongful action, the victim must show that the tortfeasor owed a duty of care, that it breached *1203its duty, that the breach proximately caused the harm, and that the victim is entitled to money damages as a result. (Artiglio v. Corning, Inc. (1998) 18 Cal.4th 604, 614 [76 Cal.Rptr.2d 479, 957 P.2d 1313].) This is the bedrock of negligence law. Of these elements, ordinarily only duty is a question of law to be resolved by a court.1 Thus, we routinely say that the existence of a duty is a legal question. (Kentucky Fried Chicken of Cal., Inc. v. Superior Court (1997) 14 Cal.4th 814, 819 [59 Cal.Rptr.2d 756, 927 P.2d 1260] (Kentucky Fried Chicken).) The requirement of duty gives courts the ability to limit liability for reasons of social policy (e.g., Hoff v. Vacaville Unified School Dist. (1998) 19 Cal.4th 925, 933 [80 Cal.Rptr.2d 811, 968 P.2d 522]), lest the theoretically infinite reach of tort liability paralyze society with a rule that any action eventually leading to harm, no matter how remotely, is actionable. Thus, an actor has no legal duty to avoid unforeseeable harm. (Ann M. v. Pacific Plaza Shopping Center (1993) 6 Cal.4th 666, 678 [25 Cal.Rptr.2d 137, 863 P.2d 207] (Ann M.).)
In this case, however, the majority wrongly decide that defendants had no duty to maintain the premises so as not to guard against attacks on people therein. That is utterly contrary to settled legal principles. Civil Code section 1714 provides in relevant part: “Every one is responsible . . . for an injury occasioned to another by his want of ordinary care or skill in the management of his property . . . .” Thus, “California law requires landowners to maintain land in their possession and control in a reasonably safe condition.” (Ann M., supra, 6 Cal.4th 666, 674.) This obligation “includes a ‘duty to take affirmative action to control [i.e., guard against] the wrongful acts of third persons which threaten invitees where the occupant has reasonable cause to anticipate [i.e., foresee] such acts and the probability of injury resulting therefrom.’ ” (Kentucky Fried Chicken, supra, 14 Cal.4th 814, 819.) In turn, “ ‘foreseeability . . . includes whatever is likely enough in the setting of modem life that a reasonably thoughtful [person] would take account of it in guiding practical conduct.’ . . . Moreover, it is settled that what is required to be foreseeable is the general character of the event or harm . . . [and] not its precise nature or manner of occurrence.” (Bigbee v. Pacific Tel. & Tel. Co. (1983) 34 Cal.3d 49, 57-58 [192 Cal.Rptr. 857, 665 P.2d 947].)
As pertinent to a case such as this, in which a victim sues the possessor of land for a crime committed against her by a third party, “duty ... is *1204determined by a balancing of ‘foreseeability’ of the criminal acts against the ‘burdensomeness, vagueness, and efficacy’ of the proposed security measures.” (Ann M., supra, 6 Cal.4th 666, 679.) Applying Ann M.’s balancing test, one observes that a land possessor’s neglect of its property may increase the foreseeability of third party criminal conduct, and thus create a duty to ameliorate any dilapidation or deterioration that invites violent crime. (Isaacs v. Huntington Memorial Hospital (1985) 38 Cal.3d 112, 129 [211 Cal.Rptr. 356, 695 P.2d 653]; cf. Ann M., supra, 6 Cal.4th 666, 677-679 & fns. 6 & 7.)2 Foreseeability increases, as does the burden imposed on the land possessor to guard against third parties’ criminal acts, to the extent that those three factors were present and would notify a reasonable land possessor of the potential that its invitees could be injured.
This case comes before us on summary judgment. As is well known, a summary judgment motion is properly granted if the moving party meets the burden of showing that there are no triable issues of fact and that it is entitled to judgment as a matter of law. (Artiglio v. Corning, Inc., supra, 18 Cal.4th 604, 612.) Thus if, but only if, the land possessor meets the burden of showing that, as a matter of law, the character of the property and the history of crime on the property and in the vicinity, viewed in the aggregate, would not give notice to a reasonable land possessor of the potential for harm to its invitees, is the summary judgment motion properly granted.
The foregoing burden, of course, is and should be difficult to meet. Summary judgment is not “a substitute for a full trial.” (Hayman v. Block (1986) 176 Cal.App.3d 629, 639 [222 Cal.Rptr. 293].) “The purpose of the summary judgment procedure is not to try the issues but merely to discover, through the medium of affidavits, whether there are issues to be tried and whether the parties possess evidence which demands the analysis of trial.” (Colvin v. City of Gardena (1992) 11 Cal.App.4th 1270, 1275 [15 Cal.Rptr.2d 234].) Especially in a factually complex case such as this, the parties cannot be expected to present all of their evidence in the form of affidavits.
Applying the principles outlined above to the alleged facts of this case, it is clear that granting the summary judgment motion was improper. It was for *1205a jury to evaluate the conflicting evidence and determine whether, under all of the circumstances, defendants knew of or should have anticipated the possibility of criminal conduct in their parking garage, and failed to guard against it or even invited it by letting the premises deteriorate.
But the notion that there are limits to the limits on tort liability that courts may impose is absent from the majority’s analysis. They impose an iron rule of no potential liability despite plaintiff’s strong evidence of neglected and unsupervised property. They err.3
B
The majority rest their decision on the principle of unforeseeability. But their reasoning is difficult to understand, and I believe they confuse the two different ways in which foreseeability is applied. To be sure, “[t]he problem is complex, and has bedeviled many.” (Brewer v. Teano (1995) 40 Cal.App.4th 1024, 1030 [47 Cal.Rptr.2d 348].) The majority may now count themselves among them.
The majority contend that foreseeability, when used as a component of determining a defendant’s duty, is a matter of law for the court to decide. That is correct, but it is a rule of limited application. All it means is that, as a general matter, if the type of harm alleged is too remote a consequence of the type of misconduct alleged, the defendant is not liable. (Ballard v. Uribe (1986) 41 Cal.3d 564, 573, fn. 6 [224 Cal.Rptr. 664, 715 P.2d 624].) It is much the same as saying that the conduct, as a matter of law, was not the *1206proximate cause of the harm. But there is no question of such remoteness here. It is for the jury to decide whether, under the facts of this case, the harm to Sharon P. was a foreseeable consequence of defendants’ conduct.
The majority appear not to perceive the distinction, but they need look no further than Ballard v. Uribe, supra, 41 Cal.3d 564. Brewer v. Teano, supra, 40 Cal.App.4th 1024, usefully summarized Ballard’s holding. Ballard “acknowledged the confusion that had arisen over the respective roles played by the court and jury in dealing with the foreseeability concept in tort doctrine. In some contexts it is a question of fact for the jury, and in others a ‘part of the calculus to which a court looks in defining the boundaries of “duty.” ’ ‘ “[Djuty” is not an immutable fact of nature, “ ‘but only an expression of the sum total of those considerations of policy which lead the law to say that the particular plaintiff is entitled to protection.’ ” ’ The foreseeability of a particular kind of harm is significant in the duty calculus, ‘but a court’s task—in determining “duty”—is not to decide whether a particular plaintiff’s injury was reasonably foreseeable in light of a particular defendant’s conduct, but rather to evaluate more generally whether the category of negligent conduct at issue is sufficiently likely to result in the kind of harm experienced that liability may appropriately be imposed on the negligent party.’ By contrast, the jury considers foreseeability in two more focused, fact-specific settings; likelihood or foreseeability of injury in deciding whether defendant’s conduct was negligent in the first place, and whether the negligence was a proximate or legal cause of plaintiff’s injury.” (Id. at p. 1030.)
C
Today’s decision illustrates the impossibility of evaluating complex factual matters in the pages of the appellate reports. If the question of legal duty must not be “left to the jury,” lest “the court. . . abdicate] decision in favor of men who do not know the law” (Prosser, Proximate Cause in California (1950) 38 Cal. L.Rev. 369, 423), so the questions of proximate cause and foreseeability in the fact-specific context of this case must not be left to a reviewing court that, in the absence of a full trial, cannot and does not know all of the relevant facts.
II
I also disagree with the majority’s unsupported comments about crime and public safety.
Apparently addressing the question of proximate cause, the majority announce that “it is questionable whether plaintiff’s proposed measures *1207would have been effective to protect against the type of violent assault that occurred here. The record, for instance, contains no evidence that the security camera at issue was even aimed toward the area of the parking garage where plaintiff was attacked. Moreover, surveillance cameras do not deter all crime and criminals do not confine their activities to locations that are untidy or unkempt.” (Maj. opn., ante, at p. 1196.)
Of course surveillance cameras do not deter all crime. One question to be resolved at trial is the extent to which they would have on defendants’ premises. And of course criminals do not confine their conduct entirely to neglected, dank, dark, gloomy and unmonitored underground parking structures. Crime may occur in unlikely places: in 1982 an intruder awakened Queen Elizabeth II in her bedroom at Buckingham Palace. The question, ultimately, is whether the environment in which plaintiff was assaulted invited crime in a way that entitles her to damages. These are questions of fact for a jury to resolve.
Casual theorizing and straw-figure demolition pervade the majority’s discussion. The majority speculate that “surveillance cameras may be ineffectual to protect against crime unless there are employees who are available to continuously monitor video transmissions and respond effectively when suspicious or criminal behavior is observed.” (Maj. opn., ante, at p. 1196.) They also, with notable lack of persuasiveness, question the efficacy of good lighting. (Id. at pp. 1196-1197.) But at a trial, there might be testimony on such questions. That trial will not occur now.
One item of the majority’s social commentary is almost certainly incorrect. They justify their decision to excuse defendants in part by announcing that “violent crime [is] ‘endemic in today’s society.’ ” (Maj. opn., ante, at p. 1194.) The strange implication is that ever higher crime rates justify a new legal rule. But empirical data belie the majority’s observation: violent crime has declined to the levels of decades ago. “National Crime Victimization Survey [1998] violent crime rates ... are the lowest recorded since the survey’s inception in 1973.”4 And “the 1991 peak in the homicide rate was lower than the 1980 crest and even close to par with trends during the Great *1208Depression. Moreover, since 1991, the murder rate has regressed to levels not seen since the 1960s.”5
Moreover, the majority’s crime prevalance analysis is illogical. If crime is endemic in a particular area, then the duty to protect against it is heightened, not diminished. The relevant question in this regard is whether defendants’ garage was located in an area sufficiently prone to certain crimes so as to make a potential attack on an invitee foreseeable. If so, that fact, regardless of statistics, would call for defendants to provide more protection to their invitees, not less. To repeat, “ ‘foreseeability . . . includes whatever is likely enough in the setting of modem life that a reasonably thoughtful [person] would take account of it in guiding practical conduct.’ ” (Bigbee v. Pacific Tel. & Tel. Co., supra, 34 Cal.3d 49, 57.)
Ill
The Court of Appeal declared that this case presented the question whether underground “commercial parking garages can be fairly characterized as inherently dangerous” for attracting “the criminal acts of third persons.” It held that they could be so characterized.
In reversing the trial court’s grant of summary judgment, the Court of Appeal found two factors significant. The first was narrow and, though it led that court to a legal conclusion, was fact specific. Sharon P. alleged that defendants had neglected the property and it was located in a high-crime area. “[T]aking into account the physical conditions presented at this particular location as well as its recent history of criminal activity, including at least seven serious felonies (robberies) at the adjacent bank premises during the two-year period preceding the assault on plaintiff, we hold, as a matter of law, that a high degree of foreseeability existed that patrons of the defendants’ commercial parking garage might become victims of third person criminal assaults, such as robberies, shootings, rapes, or some other form of physical aggression. We necessarily further hold that, given all of such circumstances, specific evidence of prior similar criminal misconduct is not required in order for the defendants to have a duty to provide reasonable preventative measures which, depending on the total circumstances, might or might not include security guards.” (Italics omitted.)
The second factor in the Court of Appeal’s analysis was purely legal: it held that as a matter of law underground parking garages, by their nature, *1209invite crime. “ ‘[I]n its very operation of a parking structure, defendant may be said to have created “an especial temptation and opportunity for criminal misconduct,” thus increasing the foreseeability of the attack. [Citation.] In making this observation, we note the unique nature of a parking complex, which invites acts of theft and vandalism. In such structures, numerous tempting targets (car stereos, car contents, the cars themselves) are displayed for the thief; high walls, low ceilings and the absence of the cars’ owners allow the thief or vandal to work in privacy and give him time to complete his task. Such circumstances increase the likelihood of criminal misconduct. In addition, the deserted, labyrinthine nature of these structures, especially at night, makes them likely places for robbers and rapists to lie in wait. Robbery, rape, and violent consequences to anyone who interrupts these crimes, may thus also be foreseeable.’ ” The Court of Appeal was quoting Gomez v. Ticor (1983) 145 Cal.App.3d 622, 628 [193 Cal.Rptr. 600], invoking language that we also quoted approvingly in Isaacs v. Huntington Memorial Hospital, supra, 38 Cal.3d 112, 128.
These conclusions were, in my view, erroneous. The first conclusion was incorrect for the same reason that the majority err: the Court of Appeal decided as a matter of law a question of foreseeability that was for the jury, but decided it to the contrary. The second conclusion, that subsurface parking garages are inherently dangerous, was also erroneous. In fact some such parking garages, because they provide the security their location requires, are not dangerous. But the Court of Appeal cannot be faulted for having addressed the issue. It was inevitable that a reviewing court would do so after we queried an aside in Ann M., supra, 6 Cal.4th 666, “whether some types of commercial property are so inherently dangerous that, even in the absence of prior similar incidents, providing security guards will fall within the scope of a landowner’s duty of care.” (Id. at p. 680, fn. 8.) We suggested that parking garages were inherently dangerous, observing that Ann M. had “offered no evidence to show that, like a parking garage or an all-night convenience store, a retail store located in a shopping center creates ‘ “an especial temptation and opportunity for criminal conduct.” ’ ” (Ibid.) The Court of Appeal noted, “Although the [Ann M.] court phrased the question only in terms of a duty to provide security guards, the basic issue it left unresolved is whether certain types of commercial premises may be so inherently dangerous as to present a reasonable foreseeability that crimes against customers or tenants will be committed by third persons if the landowner does not take reasonable preventative measures which may or may not include security guards.”
But the majority have responded to the questionable aspects of the Court of Appeal majority’s analysis with their own errors. I disagree with the reasoning both of the majority on this court and that on the Court of Appeal.
*1210IV
I would affirm the judgment of the Court of Appeal, but on the grounds stated herein.
Whether defendants breached their duty is a question of fact for the jury to decide. (Mexicali Rose v. Superior Court (1992) 1 Cal.4th 617, 633 [4 Cal.Rptr.2d 145, 822 P.2d 1292].) So are the questions whether defendants’ conduct proximately caused plaintiff’s injuries (Hoyem v. Manhattan Beach City Sch. Dist. (1978) 22 Cal.3d 508, 520 [150 Cal.Rptr. 1, 585 P.2d 851]), and the amount of compensation owed for them (Torres v. City of Los Angeles (1962) 58 Cal.2d 35, 53 [22 Cal.Rptr. 866, 372 P.2d 906]).
In Isaacs we held that whether a reasonable land possessor would have foreseen the possibility of criminal conduct depends on the following factors, among others: “the nature, condition and location of the defendant’s premises.” (Isaacs v. Huntington Memorial Hospital, supra, 38 Cal.3d 112, 129.) Ann M. expressed doubt about a totality of the circumstances test contained in Isaacs, but did not overrule Isaacs in this regard; indeed, Ann M. noted that under circumstances similar to those Sharon P. alleges here, a plaintiff might establish a high degree of foreseeability. (Ann M., supra, 6 Cal.4th 666, 679, fn. 7.)
Nothing in Ann M., supra, 6 Cal.4th 666, is to the contrary. The question presented in Ann M. was narrow: whether a shopping center had a duty, despite a lack of knowledge of prior incidents of violent crime, to provide security guards to protect a tenant’s employee—or as stated in Ann M.’s own language, whether it “had reasonable cause to anticipate that criminal conduct such as rape would occur in the shopping center premises unless it provided security patrols in the common areas.” (Id. at p. 676.) We concluded that because of the high cost of security guards and the difficulty of knowing how many might be needed, “a high degree of foreseeability is required in order to find that the scope of a landlord’s duty of care includes the hiring of security guards.” (Id. at p. 679.) The “requisite degree of foreseeability rarely, if ever, can be proven in the absence of prior similar incidents of violent crime on the landowner’s premises.” (Ibid.)
But that is the extent of Ann M. As the Court of Appeal herein explained, “Ann M. did not totally rewrite Isaacs on the issue of prior similar incidents, but only addressed that issue vis-a-vis the claimed necessity of a specific preventative measure: security guards.” (Italics deleted.) What Ann M. does require is what Isaacs v. Huntington Memorial Hospital, supra, 38 Cal.3d 112, required: as stated, “a balancing of ‘foreseeability’ of the criminal acts against the ‘burdensomeness, vagueness, and efficacy’ of the proposed security measures.” (Ann M., supra, 6 Cal.4th 666, 679; cf. Isaacs, supra, 38 Cal.3d 112, 125, 131.) A court engages in that balancing test in determining the scope of the land possessor’s duty. (Ann M., supra, 6 Cal.4th at pp. 678-679.)
Bureau of Justice Statistics, United States Department of Justice, Criminal Victimization 1998 (Aug. 25, 1999) page 1 <http://www.ojp.usdoj.gov/bjs/abstract/cv98.htm> (as of September 21, 1999).
Bureau of Justice Statistics, Sourcebook of Criminal Justice Statistics (1998) Section 3, Nature and Distribution of Known Offenses <http://www.albany.edu/sourcebook/1995/ sec3intro.html> (as of September 21, 1999).