concurring:
I agree with the majority that this case is not moot; hence, I join in part II of its opinion. I also agree that 29 U.S.C. § 794 (1984) does not afford Ms. Clark the relief she seeks. See Maj. Op. 84-85 n. 3 (29 U.S.C. § 794). I believe, however, that the majority’s eleventh amendment analysis, Maj. Op. at 83-84, is incorrect, and that the eleventh amendment prevents Ms. Clark from receiving compensation for alleged violations of her constitutional rights occurring during her 29-year confinement at Laurelton Center.1 I would, nevertheless, grant Ms. Clark the relief she seeks on a different ground: not as compensation for historical constitutional deprivations, but on account of her present constitutional right to treatment. I therefore concur in the judgment affirming the district court’s order.
In part I of this concurrence, I shall explain my difference with the majority’s eleventh amendment analysis. In part II, I shall identify Ms. Clark’s right to treatment, describe its extent and limits, and explain its relevance in this case.
I.
The majority holds that Ms. Clark is entitled to relief now because of violations of her rights to procedural due process over the past 29 years, Maj.Op. at 86, and violations of her substantive due process rights since 1976, Maj.Op. at 87. I believe that this analysis, which predicates relief on historical violations of Ms. Clark’s constitutional rights, is barred by the eleventh amendment.
A. The Prospective-Retrospective Distinction in the Eleventh Amendment Jurisprudence
Although there is currently much scholarly debate about the origin and meaning of the eleventh amendment,2 the Supreme *88Court has spoken clearly on the matter. Ever since Ex parte Young, 209 U.S. 123, 28 S.Ct. 441, 52 L.Ed. 714 (1908), the rule has been that the eleventh amendment does not permit federal courts to entertain suits against states in which the complainants seek compensation for historical violations of federal rights, but that the federal courts do have the power to protect against states3 continuing violations of federal rights. See, e.g., Green v. Mansour, — U.S. —, —, 106 S.Ct. 423, 425-26, 88 L.Ed.2d 371 (1985); Pennhurst State School & Hospital v. Halderman, 465 U.S. 89, 102-03, 104 S.Ct. 900, 909, 79 L.Ed.2d 67 (1984); Quern v. Jordan, 440 U.S. 332, 337, 99 S.Ct. 1139, 1143, 59 L.Ed.2d 358 (1979); Edelman v. Jordan, 415 U.S. 651, 668, 94 S.Ct. 1347, 1358, 39 L.Ed.2d 662 (1974). The Supreme Court has summarized its interpretation by saying that prospective relief is permitted whereas retrospective relief is not. See Quern v. Jordan, 440 U.S. at 337, 99 S.Ct. at 1143.
The distinction between prospective and retrospective relief “will not in many instances be that between night and day,” Edelman v. Jordan, 415 U.S. at 667, 94 S.Ct. at 1357, and the Court has thus enunciated a standard for determining whether relief is prospective or retrospective: it is retrospective, and hence barred by the eleventh amendment, if it is “measurable in terms of a monetary loss resulting from a past breach of a legal duty on the part of the defendant state officials.” Id. at 668, 94 S.Ct. at 1358. Application of this standard to the majority’s solution in the instant case raises two questions: (1) is the relief sought — placement in the community living arrangement (CLA) — “measurable in terms of a monetary loss,”? and (2) does it (according to the majority’s theory of the case) “resultQ from a past breach of a legal duty” by the state?
The answer to both questions is “yes.” I do not see Ms. Clark’s loss of liberty, albeit tragic and outrageous, as different in kind from any of the other losses of liberty for which courts regularly recompense victims. In her demand for relief in this case, Ms. Clark seeks two million dollars to remedy the alleged violations of her substantive and procedural due process rights. Ms. Clark’s desired remedy is thus “measurable in terms of a monetary loss.” As to the second question, the majority clearly holds that Ms. Clark must be placed in the CLA because her due process rights have, for the past 29 years, been violated by the officials at Laurelton. See Maj.Op. at 83. There is therefore no doubt that the remedy is on account of a past breach of the state’s legal duty to Ms. Clark. Because the answer to both of these questions is in the affirmative, the majority’s theory of the case is thus foreclosed by the eleventh amendment.
This conclusion comports with the Eighth Circuit’s decision in Miener v. Missouri, 673 F.2d 969 (8th Cir.), cert. denied, 459 U.S. 909, 103 S.Ct. 215, 74 L.Ed.2d 171 (1982), in which the plaintiff, a sufferer of “serious learning disabilities and behaviorial disorders,” id. at 972, requested compensatory educational services from the state defendants on the ground that she had been unconstitutionally denied a free education on account of her condition. The court held the plaintiff’s suit barred by the eleventh amendment:
An award of tuition reimbursement would clearly be barred as an award of *89damages for past breach of legal duty under this test. Riley v. Ambach, 508 F.Supp. 1222, 1248 (E.D.N.Y.1980). We view the request for compensatory services as practically indistinguishable from a request for such reimbursement. Compensatory services, like the award of a money judgment, would be measurable against past educational deprivation. The expenditure of state monies to provide compensatory services would not, in other words, ensure “compliance in the future with a substantive federal question determination.” Edelman v. Jordan, supra, 415 U.S. at 668, 94 S.Ct. at 1358 (emphasis added). We conclude that the eleventh amendment bars the award of such compensatory relief as appellant has requested. [Citations omitted].
Miener v. Missouri, 673 F.2d at 982. See also Max M. v. Thompson, 566 F.Supp. 1330, 1336-37 (N.D.Ill.1983) (relying on Miener to deny claim for compensatory educational services).4 The Ninth Circuit has recently adopted the Miener analysis in Alexopulos v. Riles, 784 F.2d 1408, 1412 (9th Cir.1986). .1 believe that the Miener analysis is correct and fully applicable in this case.
B. The Meaning of Milliken II
The majority relies on Milliken v. Bradley, 433 U.S. 267, 97 S.Ct. 2749, 53 L.Ed.2d 745 (1977) (Milliken II) in finding that the Eleventh Amendment is not a bar to Ms. Clark’s suit. Because Milliken II is critical to the majority’s position, and because the lessons to be drawn from it are not entirely clear, it is necessary to review that decision at some length.
In Milliken II, the district court found that the City of Detroit had practiced de jure segregation for many years. The district court therefore ordered a busing program and a remedial education program, funded by the state defendants, for those children who had attended the worst schools in the system. The Supreme Court upheld the plan against an eleventh amendment challenge on the grounds that the remedial plan was necessary to “ ‘eliminate from the public schools all vestiges of state-imposed segregation.’ ” Milliken II, 433 U.S. at 290, 97 S.Ct. at 2762 (quoting Swann v. Charlotte-Mecklenburg Board of Education, 402 U.S. 1, 15, 91 S.Ct. 1267, 1275, 28 L.Ed.2d 554 (1971)). The Court went on:
[B]y the nature of the antecedent violation, which on this record caused significant deficiencies in communication skills — reading and speaking — the victims of Detroit’s de jure segregated system will continue to experience the effects of segregation until such future time as the remedial programs can help dissipate the continuing effects of past misconduct. Reading and speech deficiencies cannot be eliminated by judicial fiat; they will require time, patience, and skills of specially trained teachers. That the programs are also “compensatory” in nature does not change the fact that they are part of a plan that operates prospectively to bring about the delayed benefits of a unitary school system.
Milliken II, 433 U.S. at 290, 97 S.Ct. at 2762.
After reviewing Milliken II and quoting extensively from it, the majority asserts that Milliken II stands for the proposition that a federal court may order states to fund programs to remedy historical violations of federal rights so long as the measures are necessary to undo those harmful effects of the historical violations that continue into the future. Maj.Op. at 84. The majority thus reads Milliken II as imposing what I will call a “continuing effects” exception on the traditional prospective-retrospective eleventh amendment dichotomy. The continuing effects exception works as follows: regardless of when the constitutional violation at issue occurs, *90the state is liable for any effects of the violation that are felt after the time of the institution of the suit. Applying the continuing effects exception to the instant case, the majority has no difficulty concluding that Ms. Clark’s placement in a CLA was intended to remedy the continuing effects of her improper detention at Laurel-ton, ie., to prepare her to live in the general community with no, or minimal, assistance.5
Although a literal reading of Milliken II may support the continuing effects exception, I believe, for two reasons, that the exception is an unwarranted extension of the eleventh amendment jurisprudence. First, the exception would effectively obliterate the prospective-retrospective distinction: almost all wrongdoing has continuing effects, and a continuing effects exception would thus virtually emasculate the current, prevailing interpretation of the eleventh amendment. Virtually anyone who has a colorable argument of wrongdoing by the state will assert consequential damages; even if the claims eventually fail, the plaintiffs will have forced the state to defend itself in court — an often expensive and risky undertaking. There is no indication in Milliken II itself that the Court intended to cut back so substantially on the longstanding prospective-retrospective distinction, and, in the several Supreme Court cases following Milliken II that involved the prospective-retrospective distinction, the distinction has been continually reasserted without extensive discussion of Mil-liken II or any mention of a continuing effects exception, see, e.g., Green v. Mansour, supra; Pennhurst State Hospital v. Haldeman, supra; Quern v. Jordan, supra. Without some stronger indication from the Supreme Court, I cannot believe that Milliken II cuts so broad a swath as the majority perceives in the Eleventh Amendment jurisprudence.
Second, the exception is completely at odds with the reasoning underlying the prospective-retrospective distinction, which is that it is a greater affront to the sovereignty of the states to force them to pay for prior illegal acts (the illegality of which the states were presumably unaware) than to enjoin them from doing certain acts in the future. See Scott v. Plante, 691 F.2d 634, 637 (3d Cir.1982) (Gibbons, J.) (“Obviously the problem of hindsight interference with decisions made by hard-pressed professional staff members of state mental institutions is a more serious one than that of assisting them in directing prospective injunctive relief against appropriate state officials.”). The majority ignores this point, however, for its continuing effects exception imposes potentially significant costs for states’ prior acts. The incursion on state sovereignty is the same whether the state pays for continuing effects or past effects, for in either case the state is forced to pay for its prior acts.
The recent case of Green v. Mansour, supra, supports my position. In Green, recipients of federal funds (AFDC) sued the state official responsible for the administration of the funds, alleging violations of various federal statutes. Before the suit was decided, Congress amended the relevant federal statutes, and it was undisputed that the program was properly run from then on. Plaintiffs nevertheless sought relief in the form of (1) a declaratory judgment stating that the official’s conduct had violated federal law and (2) a notice sent by the state to all members of the class advising them that there were state administrative procedures available to determine whether they were eligible for past benefits.
The Supreme Court held that the requested relief was barred by the eleventh amendment. The Court began its analysis by restating and strongly reaffirming the prospective-retrospective distinction. Green v. Mansour, 106 S.Ct. at 426. Then, the Court made clear that as a consequence *91of the distinction the eleventh amendment barred all but the prevention of ongoing violations of federal laws: “Because ‘notice relief is not the type of remedy designed to prevent ongoing violations of federal law, the Eleventh Amendment limitation on the Art. Ill power of federal courts prevents them from ordering it as an independent form of relief.” Id. at 427. The Court’s reliance on such well-entrenched eleventh amendment doctrine in a case where no more was sought by way of arguably retrospective relief than an order that the state provide notice is suggestive of the weakness of the majority’s position here, for, as explained above, the majority’s position severely limits and undermines the conceptual underpinnings of the doctrine.
In addition, what the Green Court did not say may be as important as what it did. Cf. A. Doyle, The Silver Blaze (dog that did not bark provides key to case), in The Complete Sherlock Holmes (1938). The Green dissent criticized the Court’s reliance on the prospective-retrospective distinction in the eleventh amendment context, and in so doing adopted a position similar to the majority’s in this case. The Green dissent stated that “[t]he distinction is hardly so neat as the majority implies,” 106 S.Ct. at 432 n. * (Marshall, J., dissenting), and then supported its charge by quoting portions of Milliken II relied upon by the majority in this case. Compare id. with Maj.Op. at 83-84. The Green dissent may thus fairly be read as at least exploring a continuing effects exception derived from Milliken II. The fact that this line of argument was rioted by the Green dissent and not answered by the majority implies that the majority was aware of it, but 'unpersuaded.
We are left, then, with an anomaly in the Supreme Court’s eleventh amendment jurisprudence. The majority deals with the anomaly by ignoring it. The majority simply focuses on Milliken II and does not consider it in context with the Supreme Court’s other eleventh amendment cases. By refusing even to acknowledge the tension between Milliken II and those other cases, and by blindly following Milliken II, the majority reaches a holding that effectively overrules, or at least severely limits, the rest of the cases.
As I have indicated above, I think that this takes Milliken II much too far. When one decision is so completely at odds with a whole body of law as Milliken II is with the rest of the eleventh amendment cases, the most appropriate reading is to limit that case to its facts perhaps more severely than we would otherwise be inclined to do. This interpretive rule is derived from our role as inferior judges in the federal system: we cannot overturn the decisions of the Supreme Court, but must always attempt to harmonize them. When the harmony is on the whole strong and clear, as I believe it is in the eleventh amendment context, we should make sure that any discordant notes are muted.
The limitation must be principled, not arbitrary: it must draw its essence from the historical context in which the case arose. In the case of Milliken II, the historical context is rich and significant. Milliken II was part of a long line of Supreme Court cases dealing with racial desegregation, and long before Milliken II the Supreme Court had made clear that racial segregation was a unique blight on our nation that would be combatted by extraordinary measures. See, e.g., Swann v. Charlotte-Mecklenburg Board of Education, 402 U.S. 1, 91 S.Ct. 1267, 28 L.Ed.2d 554 (1971); Brown v. Bd. of Ed. of Topeka, 347 U.S. 483, 74 S.Ct. 686, 98 L.Ed. 873 (1954). The pronouncements and decisions that the Court made during the fight against school segregation should not automatically be applied to other situations that may not share the same historical factors that went into the Court’s desegregation decisions. The Court was aware when it was deciding the school desegregation cases that the history of Blacks in this country was unique, and subsequent courts cannot ignore that fact in interpreting those cases.
In light of Milliken IPs variance with the rest of the eleventh amendment jurispru*92dence, and in light of the striking historical factors underlying Milliken II, I believe that Milliken II should be read as adopting a continuing effects exception applicable to racial desegregation cases only, or, even more narrowly, school desegregation cases. I admit that no such limitation is to be found on the face of Milliken II, but I believe that my reading is faithful to the spirit of the opinion and the context in which it was written. This is not to say, of course, that every judicial decision concerning racial matters cannot have precedential effect beyond the area of race, but merely, as stated above, that when one decision is completely at odds with a whole body of law, as Milliken II is with the rest of the eleventh amendment cases, a limited but principled reading of the case is to be preferred.
Given this reading of Milliken II, it is clear that the continuing effects exception does not apply to this case. The main body of the eleventh amendment jurisprudence controls, and, for the reasons stated above, mandates that Ms. Clark cannot predicate her claim for relief on the historical mistreatment that she suffered while at Lau-relton.
II.
Because I believe that the eleventh amendment bars Ms. Clark from receiving the injunctive relief she seeks as a remedy for any historical violations of her due process rights, if Ms. Clark is entitled to such relief, it must be because she has a present constitutional entitlement to it, independent of any mistreatment she suffered in the past. I believe that the existence vel non of such a constitutional entitlement, and the extent of such an entitlement if it exists, are the central questions on this appeal.
A. Historical Foundation for the Right to Habilitation
The right to habilitation6 of the involuntarily 7 civilly committed has been the subject of numerous opinions and scholarly articles.8 Although the cases and articles *93have not coalesced into a unanimous position, a consensus has developed among the lower courts that the involuntarily civilly committed do have a right to habilitation. There are two principal theories concerning that right, the quid pro quo theory and the parens patriae theory.9 I shall describe them both here.
1. The Quid Pro Quo Theory — At the outset, it is important to distinguish the two powers pursuant to which a state may involuntarily civilly commit a person. The state may commit someone to involuntary civil confinement using its police power or its power of parens patriae. See Herman, Barriers to Providing Effective Treatment: A Critique of Revisions in Procedural, Substantive, and Dispositional Criteria in Involuntary Civil Commitment, 39 Vand.L.Rev. 83, 85 (1986). Under the police power, the state may confine people who are dangerous to others. The parens patriae power, by contrast, allows the state to confine people who, although not dangerous to others, are in need of care or treatment. The quid pro quo theory applies to any involuntarily civilly committed person; the parens patriae theory applies, as its name would suggest, only when the state involuntarily civilly commits someone pursuant to its parens patriae power.
The premise of the quid pro quo theory is that any involuntary commitment, civil or criminal, entails a “massive curtailment of liberty.” Humphrey v. Cady, 405 U.S. 504, 509, 92 S.Ct. 1048, 1052, 31 L.Ed.2d 394 (1972). Although the infringement on liberty is the same in criminal and civil commitment, the circumstances surrounding the two situations are quite different. When people are incarcerated for criminal acts, the curtailment of liberty is not unconstitutional because (1) the curtailment is attended by extensive procedural safeguards mandated by the Constitution, and (2) the curtailment is for a specific offense for which the incarcerated person was responsible; there is thus some sense in which he or she deserves, or has consented to, the curtailment of his or her liberty.
By contrast, those who are civilly committed have usually not been protected by the panoply of procedural safeguards found in the criminal system, and they have not committed any acts that could justify their confinement. Civil involuntary commitment is therefore constitution*94ally problematic, and to justify it the state must give the civilly committed persons something in exchange for their loss of liberty. As explained by Judge Alvin B. Rubin in Gary W. v. State of Louisiana, 437 F.Supp. 1209, 1216 (E.D.La.1976), aff'd on other grounds, 601 F.2d 240 (5th Cir.1979):
Long-term detention of an individual is ordinarily a denial of due process except when he has been proved, in a proceeding subject to the rigorous constitutional limitations of the due process clause and the Bill of Rights, to have committed a specific act defined as an offense against the state, and for which incarceration is permitted for a fixed term only. If an individual, adult or child, healthy or ill, is confined by the government for some reason other than his commission of a criminal offense, the state must provide some benefit to the individual in return for the deprivation of his liberty.
(emphasis added).
According to the quid pro quo theory, due process dictates that the benefit to which the involuntarily civilly committed are entitled is habilitation to enable them to leave their commitment. Just as one confined for civil contempt of court must “have the keys to the jailhouse in his pocket,” so those involuntarily civilly committed must be provided with the means to end their commitment, otherwise civil commitment would be equivalent to placement in “a penitentiary where one could be held indefinitely for no convicted offense.” Ragsdale v. Overholser, 281 F.2d 943, 950 (D.C.Cir.1960) (Fahy, J., concurring). Habilitation is the constitutionally required quid pro quo for civil confinement, hence the name of the theory. See Note, 72 Geo.L.J. at 1790-93 (identifying and discussing the quid pro quo theory of right to treatment).10
2. The Parens Patriae Theory — The derivation of the right under the parens patriae theory is quite straightforward. Simply put, the argument is that it violates the tenets of fundamental fairness embodied in the due process clause for the state to deprive a person of her or his liberty for the stated purpose of training that person, and then to fail even to attempt to give training. As eloquently explained by Judge Frank M. Johnson, Jr.: “To deprive any citizen of his or her liberty upon the altruistic theory that the confinement is for humane therapeutic reasons and then fail to provide adequate treatment violates the very fundamentals of due process.” Wyatt v. Stickney, 325 F.Supp. at 785.
Judge John Minor Wisdom made the same point in the first decision by a Court of Appeals to adopt the parens patriae theory:
[Where] the rationale for confinement is the “parens patriae” rationale that the patient is in need of treatment, the due process clause requires that minimally adequate treatment be in fact provided. This in turn requires that, at least for the nondangerous patient, constitutionally minimum standards of treatment be established and enforced.
Donaldson v. O'Connor, 493 F.2d at 521. See also Woe v. Cuomo, 729 F.2d 96, 105 (2d Cir.) (“If the justification for commitment rests, even in part, upon the need for care and treatment, ... then a State which commits must also treat.”), cert. denied, — U.S. —, 105 S.Ct. 339, 83 L.Ed.2d 274 (1984); Johnson v. Solomon, 484 F.Supp. 278, 300 (D.Md.1979) (“The right to treatment, then, is but the corresponding obligation of the state in light of its right to exercise its parens patriae power over its citizens.”) (foonote omitted); Welsch v. Li-*95kins, 373 F.Supp. at 499 (“civil commitment for reasons of mental retardation [must] be accompanied by minimally adequate treatment designed to give each committed person ‘a realistic opportunity to be cured or to improve his or her mental condition’ Wyatt v. Stickney, supra, 325 F.Supp. at 784.”).
Although the Supreme Court has not had occasion specifically to address this theory of a state’s obligations to those it commits pursuant to its parens patriae power, in Jackson v. Indiana, 406 U.S. 715, 738, 92 S.Ct. 1845, 1858, 32 L.Ed.2d 435 (1972), the Court enunciated a rule that gives strong support to the theory. In striking down a state law that permitted the state to confine indefinitely a mentally deficient deaf mute adjudged incompetent to stand trial, the Court said: “At the least, due process requires that the nature and duration of commitment bear some reasonable relation to the purpose for which the individual is committed.” This logic supports the par-ens patriae theory, for, as Judge Wisdom explained, “[i]f the ‘purpose’ of commitment is treatment, and treatment is not provided, then the ‘nature’ of the commitment bears no ‘reasonable relation’ to its ‘purpose,’ and the constitutional rule of Jackson is violated.” Donaldson, 493 F.2d at 521. See also Youngberg v. Romeo, 457 U.S. 307, 325-26, 102 S.Ct. 2452, 2463, 73 L.Ed.2d 28 (1982) (Blackmun, J., concurring).
3. Ms. Clark’s Right to Habilitation —Ms. Clark was involuntarily civilly committed, and the quid pro quo theory therefore applies to her case. In addition, as there is no indication in the record that Ms. Clark was, at the time of her initial confinement or at any time thereafter, a threat or danger to anyone in society, the state must have been acting at all times under its parens patriae authority in detaining her. The parens patriae theory is therefore also applicable. I thus conclude that, at all times when she was or will be11 involuntarily committed, Ms. Clark had and will continue to have a constitutional right to habilitation under both the quid pro quo and the parens patriae theories.
B. The Extent of the Right to Habilitation
Like all constitutional rights, the right to habilitation of involuntarily civilly committed persons is not absolute. The courts must therefore determine how strong a right it is. I shall not propose a comprehensive answer to that question here, for I believe that the answer will come about only as a result of case-by-case analysis and development. Rather, I shall propose and defend a limited right to treatment, based on Justice Blackmun’s concurring opinion in Youngberg v. Romeo, 457 U.S. 307, 325, 102 S.Ct. 2452, 2463, 73 L.Ed.2d 28 (1982), that is sufficient to decide the case at hand.
1. Justice Blackmun’s Concurrence: The Non-Deterioration Principle —Before considering Justice Blackmun’s concurring opinion in Youngberg v. Romeo, supra, it would be best to summarize the majority’s opinion in that case. In Youngberg, the Supreme Court held that a severely retarded, involuntarily civilly committed man had a due process right to such training or habilitation as was required “in light of [his] liberty interests in safety and freedom from unreasonable restraints.” 457 U.S. at 322, 102 S.Ct. at 2461. This is “such training as an appropriate professional would consider reasonable to ensure [the patient’s] safety and to facilitate his ability to function free from bodily restraints.” Id. at 324, 102 S.Ct. at 2462. Youngberg dealt exclusively with training related to physical restraints; the Court expressly stated that it was neither considering nor ruling on any broader right to habilitation. Id. at 316 & n. 19, 102 S.Ct. at 2458 & n. 19; id. at 318 & n. 23, 102 S.Ct. at 2459 & n. 23.
Justice Blackmun, in a concurrence joined by Justices Brennan and O’Connor, *96agreed that no broader right to habitation was before the Court. Id. at 326-27, 102 S.Ct. at 2463-64 (Blackmun, J., concurring). He suggested, however, that an involuntarily civilly committed person might have a due process right to such training as is necessary to preserve his or her self-care skills, i.e., to prevent them from deteriorating during the person’s commitment.
If a person could demonstrate that he entered a state institution with minimal self-care skills, but lost those skills after commitment because of the State’s unreasonable refusal to provide him training, then, it seems to me, he has alleged a loss of liberty quite distinct from — and as serious as — the loss of safety and freedom from unreasonable restraints. For many mentally retarded people, the difference between the capacity to do things for themselves within an institution and total dependence on the institution for all of their needs is as much liberty as they ever will know.
Id. at 327, 102 S.Ct. at 2464 (Blackmun, J., concurring). I shall refer to this as the non-deterioration principle. Although Justice Blackmun merely suggested the non-deterioration principle and did not definitively endorse it, id. at 329, 102 S.Ct. at 2465 (Blackmun, J., concurring), other courts have endorsed the principle. See Society for Good Will to Retarded Children v. Cuomo, 737 F.2d 1239, 1250 (2d Cir.1984); Lelsz v. Kavanaugh, 629 F.Supp. 1487 (N.D.Tex.1986); Association for Retarded Citizens of North Dakota v. Olsen, 561 F.Supp. at 487.
I, too, would endorse the non-deterioration principle. Liberty is more than merely the absence of physical confinement; as the Supreme Court’s privacy cases make clear, the right of liberty is a right to personal autonomy, see, e.g., Carey v. Population Servs. Int’l, 431 U.S. 678, 693 n. 15, 97 S.Ct. 2010, 2020, n. 15, 52 L.Ed.2d 675 (1977); Whalen v. Roe, 429 U.S. 589, 599-600 & n. 26, 97 S.Ct. 869, 876-77 & n. 26, 51 L.Ed.2d 64 (1977); Roe v. Wade, 410 U.S. 113, 153, 93 S.Ct. 705, 726, 35 L.Ed.2d 147 (1973). See also Garvey, Freedom and Choice in Constitutional Law, 94 Harv.L. Rev. 1756, 1760 (1981). The mentally disabled cannot have meaningful autonomy, and hence meaningful liberty, without basic skills; therefore, for the state to allow disabled persons’ skills to deteriorate is as sure a denial of their liberty as is their confinement to an institution. It is indeed a sad fact that for many mentally disabled people self-care skills are “as much liberty as they ever will know.” Youngberg v. Romeo, 457 U.S. at 327, 102 S.Ct. at 2464 (Blackmun, J., concurring). The right to habilitation would be a nullity if that right did not extend even to prevent the deterioration of people’s skills.
2. The Modified Non-Deterioration Principle —I am concerned by a potential limitation raised by one aspect of the non-deterioration principle, and therefore I wish to set forth my understanding of the contours of the principle. Consider a patient who is committed at an early age, who remains in confinement for a significant period of time, but whose self-care skills might have improved even if she had never been committed. (This example is not merely hypothetical; as will be seen below, it pertains to the case before us.) Under a restrictive interpretation of the non-deterioration principle, the state would be bound merely to assure that that person’s skills did not deteriorate below the level at which they were when she or he entered the state institution. It seems to me, however, that the state should be required to provide such a person with at least such training as would match the improvement that she or he would have experienced if never committed. The reasoning behind this position is analogous to the reasoning supporting the non-deterioration principle: by committing people and preventing them from developing self-care skills, the state is effectively depriving them of the opportunity to develop and exercise their autonomy. This is a deprivation of their due process right to liberty. I thus endorse a modified non-deterioration principle: involuntarily civilly committed persons have a right to treatment sufficient to develop their self-care skills to at least the level at which they would be if the persons had not been institutionalized.
*973. Application of the Modified Non-Deterioration Principle —As the majority notes, the district court found that “[b]ut for the then well-intentioned intervention of the state, Ms. Clark would most likely have lived an average life,” 613 F.Supp. at 707 (quoted in Maj.Op. at 83). That finding was not clearly erroneous. The majority also points to the stipulation that according to the very experts at the Laurelton Center, the barriers to Ms. Clark’s entering society without supervision arose primarily on account of her long institutionalization. Maj.Op. at 83. The conclusion to be drawn from these observations is that if Ms. Clark had not been institutionalized, she would have developed into a functioning member of society who would need no, or minimal, supervision. Because her institutionalization has led to her debilitation, the modified non-deterioration principle requires the state to provide Ms. Clark, who remains involuntarily committed, albeit to a CLA, see infra, with training and treatment sufficient to put her in the position in which she would have been had she never been institutionalized. Because it appears from the numerous, un-contradicted doctors’ reports in the record that placement in a CLA is the only or best way for Ms. Clark to receive the training that she is due, I believe that she has a right to placement in the CLA, until she has acquired the skills she would have had had she never been institutionalized.12
Because this is such a complicated area of law, I wish to make clear three questions I am not addressing in this opinion. First, I do not decide whether the right to habilitation survives the release of the involuntarily committed. This issue is irrelevant to Ms. Clark, for although the CLA is far less restrictive than Laurelton Center, it appears from the record that Ms. Clark is still involuntarily committed even while at the CLA. The district court’s order compelling the state to fund Ms. Clark’s CLA placement made no change in her status as an involuntarily civilly committed person. The record reveals that Ms. Clark cannot leave the CLA permanently without permission, and that she is not free to enjoy all lawful activities. Ms. Clark has therefore not been released. It is not clear, however, whether persons have rights to habilitation once they are released. The reason they might not is seen from the justification for the right to habilitation: because the right to habilitation is derived from the liberty infringement caused by involuntary confinement, once a person is no longer in the state’s custody, he or she may have no more right to habilitation. Furthermore, the eleventh amendment would prevent him or her from suing in federal court to vindicate the right that he or she had before release, for that would be a suit for vindication of an historical violation of rights. See supra part I.
*98Despite its apparent logic, this position may work a very unfair result. Were it to prevail, the state could release someone to whom it owed a grave obligation, and then hide behind its eleventh amendment shield to avoid that obligation. One solution may be to hold that although the right to habili-tation derives from the involuntarily committed person’s deprivation of liberty, it survives the termination of that deprivation. Thus, for example, one might argue that one has the right conferred by the modified non-deterioration principle until that right is satisfied, regardless of where the right-holder may be. I need not consider here the ultimate success of such an argument.13
Second, in enunciating the modified non-deterioration principle, I leave open the question whether the involuntarily civilly committed have a right to even greater habilitation than the modified non-deterioration principle would allow. It might be argued, for example, under the theories discussed here or some other theory of habilitation, that the involuntarily civilly committed have a right to as much habilitation as their capacity will allow. Alternatively, one might argue that all involuntarily civilly committed persons are entitled to treatment up to an absolute level, regardless of their status upon entrance to the institution. Because Ms. Clark was only slightly retarded, I need not consider these questions in her case.
Third, and integrally related to the second question, is the question of cost. How much should the state be forced to pay on behalf of each person’s habilitation? Must the state pay exorbitant costs, even if they are greatly in excess of the value of any improvement? Such a requirement might create a significant burden on state treasuries. Once again, I need not address this issue, for one of the perversities of this case is the district court’s non-clearly erroneous finding that it would have cost the state less to put Ms. Clark in a CLA than to keep her at Laurelton.
III.
I believe that the eleventh amendment prevents Ms. Clark from receiving compensation for any alleged violations of her constitutional rights during her confinement at Laurelton Center, and that the majority therefore erred in finding to the contrary. I further believe that the due process clause affords Ms. Clark a right, for as long as she is involuntarily committed, to at least as much treatment as is necessary to assure that her self-care skills develop as well as they would have if she had never been committed. This right mandates her placement in the CLA. I therefore concur in the judgment of the court.
. Because of my position on the eleventh amendment, I do not reach the merits of Ms. Clark’s allegations of unconstitutional deprivations while at Laurelton, discussed by the majority in part III.B. of its opinion.
. See, e.g., Fletcher, A Historical Interpretation of the Eleventh Amendment: A Narrow Construction of an Affirmative Grant of Jurisdiction Rather than a Prohibition Against Jurisdiction, 35 Stan.L.Rev. 1033 (1983); Gibbons, The Elev-*88etith Amendment and State Sovereign Immunity: A Reinterpretation, 83 Colum.L.Rev. 1889 (1983); C. Jacobs, The Eleventh Amendment and Sovereign Immunity (1972); Field, The Eleventh Amendment and Other Immunity Doctrines, 126 U.Pa.L.Rev. 515, 1203 (1978); Nowak, The Scope of Congressional Power to Create Causes of Action Against State Governments and the History of the Eleventh and Fourteenth Amendments, 75 Colum.L.Rev. 1413 (1975); Orth, The Interpretation of the Eleventh Amendment, 1798-1908: A Case Study of ludicial Power, 1983 U.Ill.L.Rev. 423; Shapiro, Wrong Turns: The Eleventh Amendment and the Pennhurst Case, 98 Harv.L.Rev. 61 (1984); Engdahl, Immunity and Accountability for Positive Governmental Wrongs, 44 U.Colo.L.Rev. 1 (1972).
. Young insisted that the complainant sue the responsible state official rather than the state itself. We are unconcerned with that legal fiction here.
. The district court later reversed itself on this point, explaining that it was constrained by dictum in Timms v. Metro. Sch. Dist. of Wabash County Ind., 722 F.2d 1310 (7th Cir.1983), which had been decided after the first Max M. decision. See Max M. v. Thompson, 585 F.Supp. 317, 324 (N.D.Ill.1984).
. Timms v. Metro. Sch. Dist. of Wabash County, Ind., 722 F.2d 1310, 1315-16 (7th Cir.1983) raised, but did not unequivocally endorse or rely upon, an argument similar to the majority’s in the case of a severely handicapped young woman seeking educational services. The Timms court decided the case on entirely separate grounds. See id. at 1316.
. Technically, one refers to "training” or "habili-tation" of the mentally retarded, and "treatment" of the mentally ill. See Youngberg v. Romeo, 457 U.S. 307, 309 n. 1, 102 S.Ct. 2452, 2454 n. 1, 73 L.Ed.2d 28 (1982); Note, Beyond Youngberg: Protecting the Fundamental Rights of the Mentally Retarded, 51 Fordham L.Rev. 1064, 1074 n. 53 (1983). Although the terms are frequently used interchangeably, see, e.g., Youngberg, I respect the distinction and will limit my discussion here to the right to habilitation.
. Ms. Clark was involuntarily committed, and I therefore limit this discussion to the right of the involuntarily committed. At least one court has stated that whether a person was voluntarily or involuntarily committed may influence his or her right to treatment. Doe v. Public Health Trust of Dade County, 696 F.2d 901, 903 (11th Cir.1983) (per curiam). Others, however, noting the frequent significant mental deficiencies of the voluntarily civilly committed, have questioned whether their consent to confinement was informed, and therefore whether there is any basis for treating them differently from those involuntarily civilly committed. See Association for Retarded Citizens of North Dakota v. Olson, 561 F.Supp. 473, 484 (D.N.D.1982); Note, The Constitutional Right to Treatment in Light of Youngberg v. Romeo, 72 Geo.L.J. 1785, 1791 (1984).
. See, e.g., Society for Good Will to Retarded Children v. Cuomo, 737 F.2d 1239 (2d Cir.1984); Phillips v. Thompson, 715 F.2d 365 (7th Cir.1983); Doe v. Public Health Trust of Dade County, 696 F.2d 901 (11th Cir.1983) (per curiam ); Scott v. Plante, 691 F.2d 634 (3d Cir.1982); Donaldson v. O’Connor, 493 F.2d 507 (5th Cir.1974), vacated and remanded, 422 U.S. 563, 95 S.Ct. 2486, 45 L.Ed.2d 396 (1975); Association of Retarded Citizens of North Dakota v. Olson, 561 F.Supp. 473 (D.N.D.1982); Evans v. Washington, 459 F.Supp. 483 (D.D.C.1978); Gary v. State of Louisiana, 437 F.Supp. 1209 (E.D.La.1976); Welsch v. Likins, 373 F.Supp. 487 (D.Minn.1974); Stachulak v. Coughlin, 364 F.Supp. 686, 687 (S.D.Ill.1973); Davy v. Sullivan, 354 F.Supp. 1320, 1324-30 (M.D.Ala.1973) (per curiam); Wyatt v. Stickney, 325 F.Supp. 781 (M.D.Ala.1971), aff’d, remanded, and reversed in part sub. nom. Wyatt v. Aderhalt, 503 F.2d 1305 (5th Cir.1974); Slovenko, The Past and Present of the Right to Treatment: A Slogan Gone Astray, 9 J. Psych. & Law 263 (1981); Spece, Preserving the Right to Treatment: A Critical Assessment and Constructive Development of Constitutional Right to Treatment Theories, 20 Ariz.L.Rev. 1 (1978); Mason & Menolascino, The Right to Treatment for Mentally Retarded Citizens: An Evolving Legal and Scientific Interface, 10 Creighton L.Rev. 124 (1976); Birnbaum, The Right to Treatment, 46 A.B.AJ. 499 (1960); Note, 72 Geo.L.J. 1785; Note, 51 Fordham *93L.Rev. 1064; Note, Developments in the Law— Civil Commitment of the Mentally III, 87 Harv.L. Rev. 1190 (1974). See generally 31 Stan.L.Rev. 541-829 (1979). (symposium on mental retardation and the law).
. There are two other theories of the right to habilitation that I do not consider here. First, the least restrictive alternative theory, relied upon by the court in Rone v. Fireman, 473 F.Supp. 92, 125 (N.D.Ohio 1979) and endorsed in Spece, Justifying Invigorated Scrutiny and the Least Restrictive Alternative As a Superior Form of Intermediate Review: Civil Commitment and the Right to Treatment As a Case Study, 21 Ariz.L.Rev. 1049 (1979), requires treatment on the grounds that the involuntarily civilly committed have a right to live in the least restrictive setting possible and treatment is necessary to effectuate that goal. This argument has been effectively foreclosed by this court’s decision in Rennie v. Klein, 720 F.2d 266 (1983) (en banc) in which six judges endorsed the view that Youngberg v. Romeo, 457 U.S. 307, 102 S.Ct. 2452, 73 L.Ed.2d 28 (1982), made least alternative analysis inapplicable to the involuntarily civilly committed. I discuss Youngberg briefly infra at 95.
Second, the incarceration due to status theory holds that, unless they are afforded treatment, the involuntarily civilly committed would be indistinguishable from criminals incarcerated for no crime; mental illness or impairment would be effectively status crimes, which violate the eighth amendment, see Robinson v. California, 370 U.S. 660, 666-67, 82 S.Ct. 1417, 1420, 8 L.Ed.2d 758 (1962). See Welsch v. Likins, 373 F.Supp. at 496 (endorsing this theory); Martarella v. Kelley, 349 F.Supp. 575, 599 (S.D.N.Y.1972) (same). The two problems with this theory are (1) it is doubtful that the eighth amendment has any force outside of the criminal context, see Ingraham v. Wright, 430 U.S. 651, 664-71, 97 S.Ct. 1401, 1408-12, 51 L.Ed.2d 711 (1977); Youngberg v. Romeo, 644 F.2d 147, 156 (3d Cir.1980) (en ban), vacated and remanded, 457 U.S. 307, 102 S.Ct. 2452, 73 L.Ed.2d 28 (1982); and (2) even if the eighth amendment does extend beyond the criminal context, it is doubtful that failure to treat is "punishment” proscribed by that amendment. See Garvey, Freedom and Choice in Constitutional Law, 94 Harv.L.Rev. 1756, 1788-89 n. 140 (1981). .
. See also Donaldson v. O'Connor, 493 F.2d at 522:
[WJhen the ... limitations on the government’s power to detain [that constrain the government in the criminal context] are absent, there must be a quid pro quo extended by the government to justify confinement. And the quid pro quo most commonly recognized is the provision of rehabilitative treatment, or, where rehabilitation is impossible, minimally adequate habilitation and care, beyond the subsistence level custodial care that would be provided in a penitentiary.
(footnotes omitted).
. As I explain below, even though she is currently at the CLA, Ms. Clark is still involuntarily committed. Because it is the mere fact of involuntary civil commitment that creates the right to treatment, Ms. Clark has a right to habilitation whether she is at Laurelton or at a CLA.
. My conclusion is at odds with that of the Seventh Circuit in Phillips v. Thompson, 715 F.2d 365, 367 (1983), which held that involuntarily civilly committed mentally disabled people had no due process right to live in a community program, like the CLA at issue here. (Although only voluntarily civilly committed persons were involved in Phillips, the court analyzed the case as if .they had the same rights as involuntarily committed persons.) The Phillips court based its conclusion on an incorrect reading of Youngberg v. Romeo according to which Youngberg meant that involuntarily committed people had a right to treatment only insofar as the treatment was necessary to support their freedom from physical constraints. See 715 F.2d 367-68. As I pointed out, supra at 95, this position is simply incorrect; Youngberg expressly stated that it was not concerned with the question of a general due process right to treatment. 457 U.S. at 316 & n. 19, 102 S.Ct. at 2458 & n. 19; id at 318 & n. 23, 102 S.Ct. at 2459 & n. 23. Also worthy of note is Society for Good Will to Retarded Children v. Cuomo, 737 F.2d 1239 (2d Cir.1984), which adopted Justice Black-mun’s nondeterioration principle, see supra at 96, but vacated that part of the district court’s order requiring the community placement of certain patients. 737 F.2d at 1251. That portion of Society for Good Will is distinguishable from this case, for the Society for Good Will court expressly found that "[institution] residents are not unduly restrained by residing at [the institution].” Id at 1251. Here, by contrast, the uncontradicted evidence is that Ms. Clark cannot reach the level she would have reached had she never been institutionalized— that is, she cannot enjoy her full right to liberty — without placement in the CLA. That crucial difference distinguishes Society for Good Will from this case. To the extent that Society for Good Will is not distinguishable, I think it wrongly decided.
. I also leave open whether the continuing right would obtain pursuant to both the parens patriae and quid pro quo theories. Whereas the continuing right to habilitation is clearly compatible with the parens patriae theory, it would appear that by its very nature the quid pro quo theory supports a right to habilitation only during the time of actual civil involuntary commitment.