OPINION OF THE COURT

While attending a summer program administered by defendants on their premises, the infant plaintiff, Luke Anthony Trupia, rode and ultimately fell from a bannister, injuring himself seriously. The complaint seeks to recover principally upon a theory of negligent supervision; it alleges that at the time of the *394accident Luke, then not yet 12 years of age, had been left wholly unsupervised. This appeal arises from defendants’ motion, granted by Supreme Court, but subsequently denied in the order we now review, to amend their answer to allege assumption of risk; defendants propose to seek dismissal of the action upon the ground that Luke may be deemed to have consented in advance to the risks involved in sliding down a bannister, among them falling from the railing, something which, evidently, had happened to him before.

The Appellate Division denied the sought amendment upon the ground that, under its cases and those of its First Department counterpart, the assumption of risk doctrine is not generally applicable in negligence actions to nullify a defendant’s duty, but is appropriately interposed only to shield a defendant from exposure to liability arising from risks inhering in athletic and recreational activities (62 AD3d 67 [3d Dept 2009]). The Court did, however, note that the Second and Fourth Departments had permitted broader use of the doctrine, and presumably granted defendants leave to appeal from its unanimous decision so that the interdepartmental inconsistency over the applicability of the doctrine might be resolved. We now answer the question consequently certified to us by the Appellate Division—whether it erred “in reversing, on the law, the order of the Supreme Court by denying defendants’ motion for leave to amend their answer to include the affirmative defense of primary assumption of risk?”—in the negative.

In 1975, following this Court’s decision in Dole v Dow Chem. Co. (30 NY2d 143 [1972]), where we held, in part, that the “[r]ight to apportionment of liability ... as among parties involved together in causing damage by negligence, should rest on relative responsibility and ... be determined on the facts” (id. at 153), the Legislature abolished contributory negligence and assumption of risk as absolute defenses and provided instead that

“[i]n any action to recover damages for personal injury, injury to property, or wrongful death, the culpable conduct attributable to the claimant or to the decedent, including contributory negligence or assumption of risk, shall not bar recovery, but the amount of damages otherwise recoverable shall be diminished in the proportion which the culpable conduct attributable to the claimant or decedent bears to the culpable conduct which caused the damages” (CPLR 1411 [emphasis added]).

*395Nonetheless, assumption of risk has survived as a bar to recovery. The theory upon which its retention has been explained and upon which it has been harmonized with the now dominant doctrine of comparative causation is that, by freely assuming a known risk, a plaintiff commensurately negates any duty on the part of the defendant to safeguard him or her from the risk (see Turcotte v Fell, 68 NY2d 432, 438-439 [1986]). The doctrine, then, is thought of as limiting duty through consent— indeed, it has been described a “principle of no duty” rather than an absolute defense based upon a plaintiff’s culpable conduct (Morgan v State of New York, 90 NY2d 471, 485 [1997] [emphasis omitted]; accord Turcotte v Fell, 68 NY2d at 438)— and, as thus conceptualized can, at least in theory, coexist with the comparative causation regimen. The reality, however, is that the effect of the doctrine’s application is often not different from that which would have obtained by resort to the complete defenses purportedly abandoned with the advent of comparative causation—culpable conduct on the part of a defendant causally related to a plaintiffs harm is rendered nonactionable by reason of culpable conduct on the plaintiff’s part that does not entirely account for the complained-of harm. While it may be theoretically satisfying to view such conduct by a plaintiff as signifying consent, in most contexts this is a highly artificial construct and all that is actually involved is a result-oriented application of a complete bar to recovery. Such a renaissance of contributory negligence replete with all its common-law potency is precisely what the comparative negligence statute was enacted to avoid.

The doctrine of assumption of risk does not, and cannot, sit comfortably with comparative causation. In the end, its retention is most persuasively justified not on the ground of doctrinal or practical compatibility, but simply for its utility in “facilitat[ing] free and vigorous participation in athletic activities” (Benitez v New York City Bd. of Educ., 73 NY2d 650, 657 [1989]; Morgan, 90 NY2d at 484; see Turcotte, 68 NY2d at 439). We have recognized that athletic and recreative activities possess enormous social value, even while they involve significantly heightened risks, and have employed the notion that these risks may be voluntarily assumed to preserve these beneficial pursuits as against the prohibitive liability to which they would otherwise give rise. We have not applied the doctrine outside of this limited context and it is clear that its application must be closely circumscribed if it is not seriously to undermine and displace the principles of comparative causation (see Arbegast v Board of *396Educ. of S. New Berlin Cent. School, 65 NY2d 161, 168 [1985]) that the Legislature has deemed applicable to “any action to recover damages for personal injury, injury to property, or wrongful death” (CPLR 1411 [emphasis added]).

No suitably compelling policy justification has been advanced to permit an assertion of assumption of risk in the present circumstances. The injury-producing activity here at issue, referred to by the parties as “horseplay,” is not one that recommends itself as worthy of protection, particularly not in its “free and vigorous” incarnation, and there is, moreover, no nexus between the activity and defendants’ auspices, except perhaps negligence. This is, in short, not a case in which the defendant solely by reason of having sponsored or otherwise supported some risk-laden but socially valuable voluntary activity has been called to account in damages.*

Allowing the defense here would have particularly unfortunate consequences. Little would remain of an educational institution’s obligation adequately to supervise the children in its charge (see Mirand v City of New York, 84 NY2d 44, 49 [1994]) if school children could generally be deemed to have consented in advance to risks of their misconduct. Children often act impulsively or without good judgment—that is part of being a child; they do not thereby consent to assume the consequently arising dangers, and it would not be a prudent rule of law that would broadly permit the conclusion that they had done so. If the infant plaintiff’s harm is attributable in some measure to his own conduct, and not to negligence on defendants’ part, that would be appropriately taken account of within a comparative fault allocation; it is not a predicate upon which an assumption of risk should be permitted to be applied.

We do not hold that children may never assume the risks of activities, such as athletics, in which they freely and knowingly engage, either in or out of school—only that the inference of such an assumption as a ground for exculpation may not be made in their case, or for that matter where adults are concerned, except in the context of pursuits both unusually risky and beneficial that the defendant has in some nonculpable way enabled.

*397Accordingly, the order of the Appellate Division should be affirmed, with costs, and the certified question answered in the negative.

This seems to me an extremely easy case. Assumption of risk cannot possibly be a defense here, because it is absurd to say that a 12-year-old boy “assumed the risk” that his teachers would fail to supervise him. That is a risk a great many children would happily assume, but they are not allowed to assume it for the same reason that the duty to supervise exists in the first place: Children are not mature, and it is for adults, not children, to decide how much supervision they need.

The majority makes this point, which is enough to dispose of the case, near the end of its opinion (majority op at 396 [“Little would remain of an educational institution’s obligation adequately to supervise the children in its charge ... if school children could generally be deemed to have consented in advance to risks of their misconduct”]). The rest of the majority opinion is, in my view, an extended dictum, which seems to say that the assumption of risk defense is largely if not entirely limited to cases involving “athletic and recreative activities” (majority op at 395).

The majority’s dictum invites a number of questions that the majority makes no attempt to answer. Most obvious among them: What exactly is “athletic or recreative” activity? Indeed, why was Luke Trupia’s chosen activity—sliding down a banister—not “recreative”? He was obviously doing it for fun. The majority says that “athletic and recreative activities possess enormous social value” (majority op at 395)—a value that presumably does not inhere in banister sliding. But why exactly is sliding down a banister (supposing it to be done by an adult with a taste for such amusement) of less “social value” than sliding down a ski slope or bobsled run? And if the latter activities are more socially valuable than the former, why is the banister slider, who chose the less desirable form of amusement, in a better position to recover damages than the skier or bobsledder?

Assumption of risk in tort law is a hard idea to understand, and I do not imply that the majority’s understanding of it is necessarily wrong. There may be perfectly good answers to the questions I have asked, and to the many others that could be asked about this subject. But I think it is a mistake to make *398sweeping pronouncements in a case that does not require it, while ignoring the questions those sweeping pronouncements raise.

Judges Ciparick, Graffeo and Jones concur with Chief Judge Lippman; Judge Smith concurs in result in a separate opinion in which Judges Read and Pigott concur.

Order affirmed, etc.