dissenting.
I do not believe that the plain meaning of “prison conditions” under § 1997e(a), or the definition in § 3626(g)(2), includes the random, violent assault alleged in this case.1 As that is the only question we need to resolve today, I will not reach the issue whether a categorical exemption exists for claims of excessive force, except to note that my analysis is consistent with opinions that conclude that a categorical exemption does exist. See Nussle v. Willette, 224 F.3d 95, 99-106 (2d Cir.2000), cert, granted, Porter v. Nussle, — U.S. —, 121 S.Ct. 2213, 150 L.Ed.2d 207 (2001); Booth v. Churner, C.O., 206 F.3d 289, 300-03 (3d Cir.2000) (Noonan, J., concurring and dissenting). Nor do I believe that a reasonable consideration of the context, object, and policy of § 1997e and the Prison Litigation Reform Act (“PLRA”) as a whole warrants concluding otherwise. I therefore respectfully dissent.
When the language of a statute is plain and unambiguous, we apply the plain meaning (period). Connecticut Nat’l Bank v. Germain, 503 U.S. 249, 253-54, 112 S.Ct. 1146, 117 L.Ed.2d 391 (1992); United States v. Hayward, 6 F.3d 1241, 1245 (7th Cir.1993). In ascertaining the plain meaning of the language, however, we should not read with blinders on-context matters. Robinson v. Shell Oil Co., 519 U.S. 337, 340-41, 117 S.Ct. 843, 136 L.Ed.2d 808 (1997). But that latter principle has limits; invoking “context,” we may not redraft language Congress enacted simply because the language may fail to achieve Congress’s entire purpose as we perceive it. See Dunn v. Commodity Futures Trading Comm’n, 519 U.S. 465, 470, 117 S.Ct. 913, 137 L.Ed.2d 93 (1997) (“[A]bsent any indication that doing so would frustrate Congress’s clear intention or yield patent absurdity, our obligation is to apply the statute as Congress wrote it.” (alteration in original) (internal quotation marks omitted)); see also Smith v. United States, 508 U.S. 223, 247 n. 4, 113 S.Ct. 2050, 124 L.Ed.2d 138 (1993) (Scalia, J., dissenting) (“Stretching language in order to write a more effective statute than Congress devised is not an exercise we should indulge in.”); Pavelic & LeFlore v. Marvel Entm’t Group, 493 U.S. 120, 126, 110 S.Ct. 456, 107 L.Ed.2d 438 (1989) (“Our task is to apply the text, not to improve upon it.”); United States v. Locke, 471 U.S. 84, 95, 105 S.Ct. 1785, 85 L.Ed.2d 64 (1985) (“[T]he fact that Congress might have acted with greater clarity or foresight does not give courts a carte blanche to redraft statutes in an effort to achieve that which Congress is perceived to have failed to do.”). In this case, the majority appears to be fixing the language of § 1997e(a), more than truly interpreting it according to its plain meaning, context considered.
To the language of the statute, then. The statute itself does not define “prison conditions.” If only as a preliminary, working definition, Webster’s Dictionary defines “conditions” as “attendant circumstances: existing state of affairs.” Web*454ster’s Third New International Dictionary-473 (1986). As examples, Webster’s lists “living conditions,” “playing conditions,” and “adverse weather conditions.” Id. Conditions, as the definition suggests, does not refer to random events. Accord Nussle, 224 F.3d at 101. For instance, a rotted tree that happens to fall on an unsuspecting golfer while playing on a golf course is not a “playing condition”2 of that course. Furthermore, when the random event is an action (i.e., conduct by a person), the description seems to evoke more discord than sense. For instance, and closer to the facts of our case, one would not say that a teacher’s sexual abuse of a student is part of “elementary school conditions.”
In common parlance, conditions-eircum-stances or states, under our definition-largely refer to the physical environment or surroundings in which something is situated. When applied to conduct specifically, the term implies, and seemingly requires, that the conduct occur with regularity, meaning that it is common and usual-which, when applied to conduct of officials part of an institution, suggests a policy or routine practice in the institution. As an example, in a high crime area a robbery (though not always frequent) is a “condition,” because it is common and usual in such an area. Or, taking the first of the two earlier examples, if rotted trees are widespread on a golf course and frequently fall (even if on the same hapless golfer), then falling trees could be said to be a “condition,” because common and usual to that course.3 But it is strange, if not strained, to refer to a random, violent assault by prison officials as conditions, or as a condition.4
But I do not understand my colleagues to quibble with the plain meaning of the statute we are asked to interpret. Instead, the majority borrows the definition of prison conditions contained in 18 U.S.C. § 3626(g)(2) — another provision of the PLRA passed along with § 1997e(a)— which defines a “civil action with respect to prison conditions” as “the conditions of confinement or the effects of actions by government officials on the lives of persons confined in prison.” I must, however, take issue with the conclusion that the random, violent assault in this case fits within that definition.
Apparently the majority concedes that excessive force is not included within the first clause of the definition. Ante, at 449; see also, e.g., Booth v. Churner, C.O., 206 F.3d 289, 294 (3d Cir.2000) ajfd. on other grounds, Booth v. C.O. Churner, 531 U.S. 956, 121 S.Ct. 1819, 149 L.Ed.2d 958 (2001). Indeed, excessive force fits no better in that definition (conditions of confinement) than the language of § 1997e(a) (prison conditions) for essentially the same reasons, which is hardly a surprise because the language is virtually the same. The majority concludes that excessive force is included within the second clause. Ante, at 449. But the language of the second *455clause addresses only claims relating to the “effects of actions.”
Claims of excessive force are not claims relating to effects; they are claims relating to actions. See Hudson v. McMillian, 503 U.S. 1, 7-11, 112- S.Ct. 995, 117 L.Ed.2d 156 (1992). Hudson recognizes this difference, excluding excessive force claims from the “extreme deprivation” requirement applicable to conditions of confinement claims under the Eighth Amendment. Id. For claims of excessive force, the action itself violates the Eighth Amendment. Id. at 9, 112 S.Ct. 995 (“When prison officials maliciously and sadistically use force to cause harm, contemporary standards of decency always are violated. This is true whether or not significant injury is evident.” (citations omitted)). But for conditions-of-confinement-type claims, the extreme deprivation — ie., the effects — creates an Eighth Amendment violation, not the actions themselves. See Rhodes v. Chapman, 452 U.S. 337, 345-47, 101 S.Ct. 2392, 69 L.Ed.2d 59 (1981). Therefore, the second clause definition, addressing claims relating to the “effects of actions,” cannot necessarily be said to include excessive force.
So it cannot be the language of the statute that is driving the majority to conclude that claims of excessive force are included within “prison conditions” under § 1997e(a). Rather, the majority relies on the context, object, and policy of the statute to conclude that claims of excessive force should be included within the exhaustion requirement of § 1997e(a). Ante, at 449-52. But equally persuasive arguments have been advanced by others that they should not be included. See Nussle, 224 F.3d at 103-06; Booth, 206 F.3d at 301-02 (Noonan, J., concurring and dissenting). We are simply not in the business of deciding what statutes should or should not say, but deciding what they in fact do say. And, this statute certainly does not say that the random, violent assault involved in this case must be exhausted in administrative proceedings.5
It may be that Congress really did want to include this kind of claim in § 1997e(a)’s “prison conditions” exhaustion requirement. The problem is that Congress did not write language in this statute to accomplish that objective. In the end, I am not convinced that we have a justifiable reason to strain the language Congress wrote to accomplish an intent we can only speculate exists. I do not believe that the plain meaning of prison conditions under § 1997e(a) includes the random, violent assault involved in this case, and therefore I respectfully dissent.
. Smith alleges that several prison guards entered his cell, ordered him to walk out, and immediately after he complied, beat him in the face, buttocks, and groin. According to Smith, the guards handcuffed him, beat him while cuffed, and took him to prison showers where they continued to hit, stomp, and jab him with batons. Smith was next taken outside to a special housing unit where the guards slammed his head into a metal plate, stripped him naked, and beat him repeatedly over the span of an unspecified length of time. The guards finally locked him in a segregation cell, naked and bleeding, releasing him the next day and denying him medical treatment until approximately one week later.
. Note, whether one uses the singular "condition” or the plural "conditions,” the incongruity remains the same.
. One would not say that falling trees are a "playing condition,” however, because dodging falling trees has nothing to do with the game of golf.
. Worse yet, it seems unreasonable to refer to such conduct as prison conditions. Accord Booth, 206 F.3d at 301 (Noonan, J., concurring and dissenting) ("That [the plaintiffs] alleged blow took place in a prison does not make it 'prison conditions.' ”); cf. Farmer v. Brennan, 511 U.S. 825, 834, 114 S.Ct. 1970, 128 L.Ed.2d 811 (1994) (“Being violently assaulted in prison is simply not 'part of the penalty that criminal offenders pay for their offenses against society.’ ”).
. I express no opinion whether the context, object, and policy of § 3626 (a very different statute) warrants including in that section claims of excessive force, or the random, violent assault in this case.