Heinous, Atrocious & Cruel - Apprendi, Indeterminate Sentencing & the Meaning of Punishment, Columbia Law Review, 2009
Download original document:
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
COLUMBIA LAW REVIEW VOL. 109 JUNE 2009 NO.5 ARTICLES HEINOUS, ATROCIOUS, AND CRUEL: APPRENDI, INDETERMINATE SENTENCING, AND THE MEANING OF PUNISHMENT W David Ball* Under Apprendi v. New Jersey, any fact that increases an offender's maximum punishment must be found by a jury beyond a reasonable doubt. The Apprendi literature has focused on the allocation of power between judge and jury, ignoring entirely the role of the parole board in indeterminate sentences-that is, sentences which terminate in discretionary parole release. In an indeterminate sentence, a judge makes a pronouncement about the length of the prescriptive sentence to be imposed, but the parole board decides the actual sentence that is, in fact, imposed. In this Article, I explore the Apprendi ramifications of indeterminate sentencing. In states where a prisoner is presumptively entitled to parole release, the denial of pa1'Ole increases the maximum punishment without a jury finding the relevant facts beyond a reasonable doubt. In California specifically, the parole board can transform parole eligible offenses into parole ineligible offenses based on its own findings of fact about the crime, even when these findings contradict the jury's. A purely mechanical reading of Apprendi would require jury findings offact for all components of the parole release decision. I argue for a reading of the Apprendi rule more consonant with fustice Stevens's Apprendi opinion itself, where the jury's power comes from its role as the retributive conscience of the community. Because indeterminate sentences combine -rt'tributive and rehabilitative components, they delinertte where-and, more * Research Fellow, Stanford CrirninaIJustice Center, Stanford Law School. I've been extremely blessed with supportive, curious, and inspiring tcachers and colleagues at Stanford and Santa Clara law schools. I'm particularly grateful to my co-clerk Meredith Osborn, to my colleagues in the Stanford Fellows program (especially Andy eoan) for their rigorous and spirited review of an early draft of this Article, to Stephanie Wildman and the Santa Clara Center for Social Justice and Public Service for giving me the time and financial support to finish a large chunk of this Article, to my wife for her critical help in shaping the argument and the Article, and to my young sons for the joy that sustains me. As inadequate a token of my gratitude as this may be,] would like to dedicate this Article to JudgeJohn T. Noonan,Jr., and Professor Robert Weisberg. ] hope you both, in some ways, find this superfluous-that you know how much I value you as mentors, teachers, and friends. Thank you for all the opportunities you've given me, and consider this a down payment on a career that will owe so much to both of you. Know that I will endeavor to pass on to my students and colleagues your rare combination of humanity, integrity, and intellect. 893 COLUMBIA LAW REVIEW 894 [Vol. 109:893 importantly, why-the Apprendi jury right applies to some facts and not to others. loxploring Apprendi in this context restores needed coherence to the doctrine, illustrating larger issues about the punitive and rehabilitative aspects of sentencing as well as the judicial and executive limits of punishment. INTRODUCTION ., .....-~~-.-:-~:~-:~-. . . . . . . . . . . . . . . . . . . . . . . • . . . . . . : :.. ' - --sg5-·---~~---~~- I. UNSUITABLE LIFE SENTENCES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 906 A. Claritying Terms: Indeterminate Sentences Are Sentences with Discretionary Parole Release. . . . . . . . . . B. California Uses Both Determinate and Indeterminate Sentences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. California's Parole Release Process ... . . . . . . . . . . . . . . . . 1. Parole Suitability Basics. . . . . . . . . . . . . . . . . . . . . . . . . . 2. The Use of Commitment Offense Facts in Suitability Determinations. . . . . . . . . . . . . . . . . . . . . . . . 3. Deferential Review............................... D. The Result: Prisoners Serve Much Longer than Their Enumerated Term of Years. 11. WHICH BODY FINDS WHICH FACTS? . . . . . . . . . . . . . . . . .. . . . . . . A. Parole Discretion and Unsuitable Life Sentences. . . . . . 1. Parole Boards Are Given Authority to Determine a Wide Range of Facts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Parole Board Factfinding and Unsuitable Life Sentences.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Apprendi'sJury Requirement.......................... 1. The Jury Finds Retributive Facts, the Judge Finds Public Safety Ones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Apprendi Factfinding and Unsuitable Life Sentences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Reconciling Apprendi and Parole Due Process: Indeterminate Sentences Serve Split Purposes 1. Who Should Find Mixed Retributive/Public Safety Facts?........................................... 2. Relabeling Desert as Public Safety. . . . . . . . . . . . . . . . 111. SENTENCE LIMITATIONS . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . A. Parole Cases and the Expectancy of Release. . . . . . . . . . 1. The Liberty Interest in Parole. . . .. . . . . . . . . . . . . . . . 2. The Liberty Interest in Parole Applied to Unsuitable Life Sentences. . . . . . . . . . . . . . . . . . . . . . . . B. Apprendi, the Statutory Maximum, and the Enumerated Term................................................ 1. The Meaning of the Functional Statutory Maximum Sentence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. The Functional Statutory Maximum of an Indeterminate Sentence is the Enumerated Term. 906 909 911 912 915 916 918 920 920 921 921 923 923 932 935 939 942 943 944 944 948 950 950 952 2009] 3. IV. . _ 895 HEINOUS, ATROCIOUS, AND CRUEL Avoiding Apprendi through Changing Parole; or, Why Does This Even Matter? . 955 THE STANDARD OF PROOF AND THE DEFINITION OF AN 957 OFFENSE . A. Parole's Standard of Proof . 958 _ . _.._ L_.R<lrol.~.:>ui\.;l!>ili!ris Reviewed for "Some .. _"'" Evidence." . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 958 2. Applying "Some Evidence" to Unsuitable Life Sentences . 959 B. Apprendi's Requirement of Proof Beyond a Reasonable Doubt . 961 I. Apprendi Defines Crimes Functionally . 962 2. Apprendi Functionalism and Unsuitable Life Sentences: Are Parole Board Unsuitability Facts Elemen ts of Crimes? . 966 V. POLICY CONSIDERATIONS . CONCLUSION . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 968 971 INTRODUCTION Apprendi v. New Jersey threw contemporary sentencing into disarray when it held that "any fact that increases the penalty for a Clime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt."l But while Apprendi's impact on state and federal practices has been pronounced, the cases that follow it are generally seen to contribute little to a more profound understanding of the jUrisprudence of punishment. 2 Instead, Apprendi is seen as a for· mal rule, applying only to a process that ends when a judge in a court· room pronounces a sentence. 3 1. 530 U.S. 466,490 (2000). 2. Justice O'Connor decried Apprendi's rule as a "meaningless formalism" with "several plausible interpretations of the constitutional principle on which the Court's decision rests." Id. al539-40 (O'Connor,]., dissenting). In Blakely v. Washington, the next major case in the line, she reiterated that it was "difficult for [her] to discern what principle besides doctrinaire fonnalism acnlally motivates today's decision." 542 U.S. 296, 321 (2004) (O'Connor, j., dissenting). Academics have also described Blakely as "a destructive rule in search of a sound principle." See Douglas A. Bennan, Conceptualizing Blakely, 17 Fed. Sent'g Rep. 89, 89 (2004) [hereinafler Berman, Conceptualizing Blakely] (explaining views of other commentators). 3. The Apprendi literature focuses only on decisions a judge makes about the prescriptive sentence to be imposed, not on the actual sentence that is, in fact, imposed. Laura Appleman's recent article is the sole exception. See Laura I. Appleman, Retributive Justice and Hidden Senlencing, 68 Ohio St. LJ. 1307 (2007). Appleman tentatively concludes that Blakely does not affect the workings of parole proceedings, id. at 1372-73, but suggests that an expansive reading of Blakely might aJfect parole, id. at 1373-76. I discuss Appleman's article in greater detail infra notes 175-] 78 and accompanying text. Several scholars have raised the parole issue in papers that focus on other aspect.s of sentencing, however. See, e.g., Kevin R. Reitz, The New Sentencing Conundrum: Policy and Constitutional Law at Cross-Purposes, 105 Colum. L. Rev. 1082, ]094 (2005) (asking 896 COLUMBIA LAW REVIEW [Vol. 109:893 I challenge those views in this Article. Apprendi is not a formal doctrine but a functional one that looks past the formal taxonomy of a criminal statute to the way it operates. Apprendi's protections are not limited to temporal slices of the judicial process but should apply any time the jury power is infringed, including during the substantial percentage of ____________. ~~.0..<.:;m~entenc'Cs.._!.ha~ terElin~.'C...~I1_ a p<l!:0le board's discret!ona~release decision.' Indeed, exploring the Apprendi right throughout the entire sentencing process helps clarify the nature and scope of that right. 5 I conclude that Apprendi is primarily concerned not with time served but with punishment, defined as the stigmatic deprivation of liberty. The jury is central not because of institutional or formal concerns, but because it expresses the moral judgments ineluctably tied to the retributive principles at the heart of criminal law. * * * The Apprendi revolution began when Charles Apprendi pleaded guilty to three weapons charges, two of which were each punishable by a five- to ten-year sentence. 6 The judge sentenced Apprendi to an "extended term" of twelve years on one of the charges, however, based on a finding by a preponderance of the evidence that he had acted with racial bias, a fact triggering a hate crime sentence enhancement. 7 The U.S. Supreme Court vacated the sentence, holding that a jury must find beyond a reasonable doubt any fact that increases the penalty a defendant faces beyond the "statutory maximum."8 As I discuss in greater detail in Parts II and IV, the reasons for the Supreme Court's finding have to do with Apprendi's view that punishment involves both the stigma of wrongdoing and the restriction of liberty: Only the jury may make such a stigmatic finding, and this finding must be made conclusively, beyond a reasonable doubt. After Apprendi, the Supreme Court was chiefly concerned with figuring out the circumstances under which the new rule would apply. The Court first decided in Ring v. Arizona that Apprendi's holding barred a whether Blakely applies "[i][ state law presumes that prisoners will be released [into parole] at a certain time in the absence of adverse findingsn);]on Wool & Don Stemen, Aggravated Sentencing: Blakely v. Washington; Practical Implications for State Sentencing Systems, 17 Fed. Sent'g Rep. 60, 68 0.27 (2004) (explaining use of parole in New Jersey indctenninate sentencing) . 4. I use the phrase "discretionary release" to postpone discussion of a key issue to which I return infra Part I.A: the way in which unclear usage of the term "indeterminate sentencing" has muddled the Apprendi doctrine. For an exposition of this Article's terminology, see infra notes 67-68 and accompanying text. 5. I will argue that Apprendi is not a "sentencing" case at all: It is a case about the way facts must relate to criminal punishment. See infra text accompanying notcs 224-226. 6. The irony is not lost on the author that the case reestablishing the jury's central role in American sentencing did not, itsclf, involve a jury. 7. 530 U.s. at 468-69. 471. 8. Id. at 490. _ 2009] HEINOUS, ATROCIOUS, AND CRUEL 897 judge from finding aggravating facts justifying the death penalty'" It then turned to the kind of cases that have since come to dominate the Appnmdi line: those which tease out the meaning of the statutory maximum. Here the issue is not so much what Apprendi "means" as when it applies. It is this focus which has given Apprendi its formalist reputation: The cases turn on technical readings of sentencing guidelines and statutes and deal _·_-----liweWiilitlie-goals -imd interests ofthe-jufy-iignL"-Tne two serriTnalcases---··------·--·-----are Blakely v. Washington lo and United States v. Booker. ll Blakely held that the statutory maximum for a given crime is not necessarily the maximum specified in the statute (what I call the taxonomic statutory maximum), but the sentence a judge is bound to impose based on the facts in the jury's conviction or the offender's plea (what I call the functional statutory maximum). Ralph Blakely's kidnapping of his wife was a felony with a maximum term of 120 months. Washington State's binding guidelines specified a range of forty-nine to fifty-three months for the kidnapping. Blakely was sentenced to ninety months, however, because the judge found an aggravating fact-that Blakely had acted with "deliberate cmelty. "12 Even though fifty-three months and ninety months are both less than 120 months-the maximum in the statute-the Supreme Court nevertheless vacated the sentence, holding that "the relevant 'statutory maximum' is not the maximum sentence ajudge may impose after finding additional facts, but the maximum he may impose without any additional findings." 13 Booker used Blakely's definition of the (functional) statutory maximum I4 but is most notable for its remedy: making the Federal Sentencing Guidelines advisory.1s Advisory guidelines pose no Apfn-endi problem; higher sentences no longer require certain findings of fact, so nonjuries may find them. The State of California subsequently followed tlle Supreme Court's lead: Mter Cunningham v. California invalidated California's Determinate Sentencing Law (DSL) on Apprendi grounds,16 the California state legislature made its sentencing guidelines advisory as well. l7 It is the technical simplicity of this Apprendi compliance move, and the cynicism it engenders, that has contributed to the view that Apprendi means very little. 9. 536 U.S. 584, 609 (2002) ("Because [the] enumerated aggravating factors operate as 'the functional equivalent of an element of a greater offense,' the Sixth Amendment requires that they be found by ajury." (quoting Ap/mmdi, 530 U.S. at 494 n.19)). 10. 542 U.S. 296 (2004). 11. 543 U.S. 220 (2005). 12. Blakely, 542 U.S. at 298-300. 13. Id. at 303-04. 14. Booker, 543 U.S. at 226-27 (Stevens,.!-, opinion of the Court) ("We hold that ... the Sixth Amendment as construed in Blakely docs apply to the Sentencing Guidelines."). 15. Id. at 245 (Breyer,]., opinion of the Court). 16. 549 U.S. 270, 293 (2007). 17. See infra note 87 and accompanying text. 898 .._.. ~._. COLUMBIA LAW REVIEW [Vol. 109:893 The end result is that Apprendi's rule applies only to binding systems, and only to facts relating to the commitment offense-that is, the crime for which the inmate was charged, sentenced, and committed to prison. Apprendi does not require a jury to find prior offenses,18 nor does it require a jury to find facts justifying mandatory minimum sentences. 19 . Non..':-",f tlJ.':. cases in th~~endi lin..':.. has dealt with indetermin~te ._. sentences,20 however, and none has dealt with any finder of fact beyond the judge. As a practical matter, because Blakely and Booker invalidated such huge swaths of sentencing practice, they, not Apprendi, became the go-to cases for judges and lawyers who needed to apply Apprendi to a given case. Ironically, then, Apprendi's contribution to our understanding of sentencing got lost even as its holding became more legally significant. All that seems to matter after Blakely-and, to an even greater extent, after Booker-is an almost mathematical focus on the structure and phrasing of sentencing guidelines themselves. 21 This focus on statutory maxima has left an enormous hole in the jurisprudence: We now have a rule that is as clear as it is uncompelling. None of the cases has explained fully, for example, why the shift from binding guidelines to advisory ones serves to insulate nonjury factfinding from an Apprendi challenge. We know for certain that Apprendi does not apply in advisory situations, but it is hard to explain why dlat rule is as it is. This Article is, in part, my attempt to draw out a rationale behind Apprendi's rule, a rationale implicit in Apprendi and its predecessors: The jury must find facts leading to punishment, since punishment conveys a morally stigmatizing judgment about the commitment offense. The problem is that the moral dimension has not presented itself clearly in any of the post-Apprendi cases. But what if the nature of given facts, or the way they were put to use, were crucial in determining whether the jury had to decide those facts beyond a reasonable doubt? Such is the case with sentences terminating in discretionary parole release. 18. Apprendi left in place Almendarez-Torres v. United Stales, which rejected an argument that recidivism must be treated as an element of the offense because it increases the potential maximum sentence. 523 U.S. 224, 246--47 (199S). 19. Harris v. United States, 536 U.S. 545, 565 (2002) ("[N]othing in this history suggests that it is impermissible for judges to find facts that give rise to a mandatory minimum sentence below the maximum penalty for the crime committed," (internal quotation marks omitted». 20. For a discussion of the confused usages of "determinate" and "indetenninate" in Supreme Court sentencing cases, and more workable definitions of my o\"'n, see infra Part l.A. 21. I also note that the literature-and Supreme Court jurisprudence-has disproportionately focused on the Federal Sentencing Guidelines, even though the overwhelming majority of sentencing and incarceration takes place in the fifty SLales. This Article focuses on the states; I will discuss federal sentencing cases only in passing. ._. . __. __ 2009] HEINOUS, ATROCIOUS, AND CRUEL 899 Sentences with discretionary parole release serve two purposes: punishment and the preservation of public safety.22 An offender serves an enumerated term of years as punishment and is then kept in prison until he is safe to be released. In these sentences, the distinction between incarceration and the meaning of that incarceration is drawn sharply: The enumerated term punishes and the rest of the time serves public safety. "---------------·--Explonng-'ihecontoursof ApprenJJ.-as- appfied to these sentences"is the--"----------~ --""perfect opportunity to see why this distinction is desirable from both a policy and a doctrinal view. If Apprendi is simply mechanical, and applies to retributive and non retributive incarceration, then the jury must find all facts justifying a denial of parole and the parole board is rendered moot. Locating the Apprendi right in the jury's retributive role, however, narrows its scope: The jUlY need only find facts expressing moral stigma. Facts that bear primarily23 on an offender's threat to society and need for incapacitation or rehabilitation, however, can be found by bodies besides the jury, including the parole board. The view that Apprendi is a mechanical rule applying only during the judicial pronouncement of the sentence subverts the very jury power that Apprendi established. In this Article, I focus on one specific example of this larger phenomenon: California's practice of imposing what I call unsuitable life sentences, where the same offense which justified a parole eligible sentence at trial is itself the reason why the parole board 24 later finds a prisoner unsuitable for parole. 25 I focus on California for several reasons: The state has a large number of prisoners serving sentences subject to discretionary parole release; state courts have grappled with the unsuitable life sentence problem in a series of cases; and the parole regulations literally duplicate the elements of parole ineligible crimes. California parole release is also presumptive, meaning that in order to extend a term of incarceration, a parole board must find a reason not to release a prisoner. It is this feature that allows a parole board finding of 22. For the purposes of this Article, I will lump all of the consequentialist purposes of punishment (including incapacitation) under the rubric of rehabilitation and public safety. The important distinction here is between retributive and nonretributive purposes of punishment. ] do not discuss the impacts my argument has on deterrence, in part because of my skepticism that criminal penalties can deter second degree crimes of passion and in part because the goals of deterrence can be more readily achieved through changes to sentencing laws, not changes to parole release practices. 23. There are "mixed" facts that bear on both punishment and rehabilitation. I will discuss these infra Part II.C.l. The commitment offense primarily underlies the penalty portion of a sentence, however, and inasmuch as it relates to rehabilitation, the fact'i are often so prejudicial as to overwhelm any rehabilitative steps-positive or negative-thar an offender has taken since entering prison. Sec infra Part lG2. 24. The name of the parole board is officially the California Board of Parole Hearings.. Because this was recently changed from the Board of Pl'ison Temls, see Cat. Penal Code § 5075 (West Supp. 2008), I will use the more generic term parole board. 25. Life sentences are only unsuitable where denial is based on the commitment offense. Garden variety parole denials, where prisoners are denied parole for poor institutional behavior, for example, are not the subject of this Article. 900 ................................ COLUMBIA LAW REVIEW [Vol. 109:893 fact to increase the statutory maximum punishment to which the prisoner is subjected. 26 Of the 173,000 prisoners in California prisons,27 approximately 27,000 are serving life sentences with the possibility of parole. 28 Of these, around 10,000 have already served their enumerated term of years and are eligible to be deemed suitable for parole. 29 ..Mth.O.1!gh..1.h.i.LArJLc!!:.LQ.clJ~eL9.11J:;jjJif.QI.ni.!h.. thGJactic.e..Qf d~~ng.. __._. parole on the basis of the commitment offense is by no means limited to one state. 30 While it is difficult to characterize the variety of sentencing systems in the fifty states and the District of Columbia, the majority of states still impose sentences terminating in discretionary parole release," and the "vast majority ... have retained some form of discretionary parole release and postrelease supervision."32 Nationwide, discretionary pa26. Sec discussion infra Part IItB. Whether this argument extends to other states' systems, then, depends on the structure of a state's parole statutes and regulations. While I have not conducted a conclusive fifty-state survey, other states' parole s)'stems indicate that the argument might apply there as well. See infra notes 30-42 and accompanying text. 27. Data Analysis Unit, Cal. Dep't of Corr. & Rehab., Prison Census Data as of June 30, 2007, al tbl.l (2007), available at http://www.cdcr.ca.gov/Reports.Researcb/ OffendcT_Infonnation_Services_Branch/Annual/Census/Censusd0706.pdf (on file with the Columbia Law Review). 28. Adam Liptak, To More Inmates, Life Tenn Means Dying Behind Bars, N.Y. Times, Oct. 2, 2005, at AI. Some of these inmates might be serving time for nonmurder convictions prior to the passage of the DSL in 1977, but they are outside this Article's scope. For more on the history of the DSL, see infra text accompanying note 85-87. 29. In re Criscione, No. 71614, slip op. at 9 (CaL Super. Ct. Aug. 30, 2007), available at http://www.bayareanewsgroup.com/multimedia/mn/news/criscione_complainc091~~ 07.pdf (on file with the Columbia Law Review). 30. In 2001, fourteen states reported releasing fewer than ten lifers, and eight states reported releasing fewer than two dozen each. Adam Liptak, Serving Life, with No Chance of Redemption, N.Y. Times, Oct. 5, 2005, at Al. 31. See Michael TonI)', Reconsidering 1ndetenninate and Structured Sentencing, Sent'g & Corrections (U,S. Dep't ofJustice, Washington D.C.), Sept. 1999, at 1, available at http://www.ncjrs.gov/pdffilesl/nij/175722.pdf (on file with the ColumiJia Law Reoiew) [hereinafter Tonry, Reconsidering] ("The numbers are imprecise because systems differ so greatly that reasonable people can disagree over which label best characterizes a particular system."). Indeed, the latest federal review of state sentencing systems labels California a "determinate" system and New York an "indetenninate" system, even though both states employ sentences with and without discretionary parole relea'ie. See Bureau of Justice Assistance, 1996 National Sunrey of State Sentencing Structures 3 (1998), available at http://www.ncjrs,gov/pdffiles/169270.pdf (on file "ith the ColumiJia Law Reoiew) [hereinafter 1996 National Survey] (describing classification of state sentencing systems by guideline structure). For a further discussion ofthc tenns determinate and indetenninate, see infra Part LA, 32. 1996 National Survey, supra note 31, at xi; see also id, at 15 exhibit 1-8 (detailing type of postrelease supenrision available in twelve states that abolished discretionary parole), Joan Petersilia counts sixteen states where parole boards have full release pmvcrs, nineteen where parole boards have limited release powers, and fifteen (plus the federal system) where discretionary parole release has been abolished. Joan Petersilia, ""Vhen Prisoners Come Home 66-67 tb1.3.1 (2003) [hereinafter Petersilia, When Prisoners Come Home]. Bolstering the point that categorization is difficult, however, California is listed among the states with no discretionary parole. Id. _ .. 2009] . ~_~~~_~ HEINOUS, ATROCIOUS, AND CR~L 901 role release has decreased as a percentage of released prisoners." But as of 2003, approximately ten percent of prisoners were serving life sentences (three-quarters of them with the possibility of parole),34 at a cost of billions of dollars a year. 35 The number of these "lifers" has nearly doubled since 1992. 36 Additionally, several other states have language ._~~i.~i~~."~()ricl~J1ti~al-=8..~()J!1.~J~~~<JE~in. the ~.'l!if()~J1ia~lJ~cler.sta.t.':'~_. triggering an increase in penalty, and many states also consider the commitment offense in determining parole release, either explicitly39 or implicitly, through rules about the evidence parole boards are required to review. 40 New York currently faces a lawsuit from parole eligible prisoners charging that the state parole board has an "unwritten policy of rejecting parole in most of the cases solely because of the severity of the crime."4! Whether Apprendi might apply to these systems the way it does to California's is beyond the scope of this Article, because the question 33. Timothy A. Hughes, Doris James Wilson & Allen J. Beck, U.S. Dep't of Justice, Bureau ofJustice Statistics Special Report: Trends in State Parole, 1990-2000, at 1 (200]), available at http://ojp.usdoj.gov/bjs/pub/pdfltspOO.pdf (on file with the Columbia Law &vi>fW). 3'1. Marc Mauer, Ryan S. King & Malcolm C. Young, The Sentencing Project, The Meaning of "Life"'; Long Prison Sentences in Context 9 (2004), available at http://www. sentcncingproject.org/Adrnin/Documents/Publications/inc_meaningoflife.pdf (on file with the Columbia Law Rez1iew). 35. Id. at 25. 36. Id. at ll. 37. See, e.g., Ariz. Rev. Stat. Ann. § 13-702(C) (5) (2001) ("especially heinous, cruel, or depraved manner"); Colo. Rev. Stat. Ann. § 18-1.3-1201 (5) (j) (West 2004) (same); Conn. Gen. Stat. Ann. § 53a46a(i) (West 2007) (same); N.H. Rev. Stat. Ann. § 630:5(VII) (h) (2007) (same); NJ Stat. Ann. § 2C:44-I(a)(l) (West 2005 & Supp. 2008) (same); Wash. Rev. Code Ann. § 13.40.150(3) (i) (ii) (West 2004) (same). 38. See, e.g., Ala. Code § 13A-549 (LexisNexis 2005 & Supp. 2007) ("especially heinous, atrocious, or cruel"); D.C. Code Ann. § 22-2104.01 (b) (4) (LexisNexis 2001 & Supp.2008) (same); Fla. Stat. Ann. § 921.0016(3) (b) (West 2006) (same); Haw. Rev. Stat. Ann. § 706-657 (LexisNexis 2008) (same); Idaho Code Ann. § 19-2515(9)(e) (2004 & Supp. 2008) (same); Kan. Stat. Ann. § 214636(1) (2007) (same); La. Code Crim. Proc. Ann. art. 905.4 (A) (7) (2008) (same); Miss. Code Ann. § 99-19-101 (5) (h) (1972) (same); Neb. Rev. Stat. § 29-2523(1) (d) (1995) (same); N.C. Gen. Stat. § 15A-1340.16(d)(7) (2007) (same); Okla. Stat. Ann. tit. 21, § 701.12(4) (West 2002) (same); Tenn. Code Ann. § 39-1~204(i)(5) (2006) (same); Utah Code Ann. § 76-5-202(1)(r) (2003 & Supp. 2008) (same). 39. See, e.g., Ga. Code Ann. § 42-9-40(a) (1997) ("severity of current offense"); Haw. Rev. Stat. Ann. § 706-669(8) (LexisNexis 2007) ("nature and degree of the offense"); Ind. Code Ann. § ll-I3-3-3(h)(I) (LexisNexis 2003 & Supp. 2007) ("nature and circumstances of the crime"); Mont. Code Ann. § 46-23-202(1) (2007) ("circumstances of the offense"); N.M. Stat. § 31-2J-10(A)(2)(a) (1978 & Supp. 2008) (same); N.Y. Exec. Law § 259i(l)(a)(i) (McKinney 2005 & Supp. 2008) ("seriousness of the offense"); R.I. Gen. Laws § 13-8-23(3) (2002) ("circumstances surrounding his or her offense"). 40. See, e.g., Mo. Ann. Stat. § 217.690(7) (West 2004 & Supp. 2008) (including victim participation in parole hearings); N.H. Rev. Stat. Ann. § 651-A:8 (2007) (allowing attorney general to present evidence at hearings); Utah Code Ann. § 77-27-9 (2003 & Supp. 2008) (including victim participation in parole hearings). 41. Sam Roberts. Violent Felons Move Forward with Lawsuit over Their Rights to Parole, N.Y. Times, Dec. 30, 2007, Metro Section, at 21. A lawyer for the inmates argues ._.. . __._..__.. 902 COLUMBIA LAW REVIEW [Vol. 109:893 depends in part on the degree to which parole regulations are binding, whether facts about the commitment offense are sufficient in themselves to deny parole, and whether an inmate has a statutorily created liberty interest in the granting of parole 42 * * * California has four penalties for murder. Two include the possibility of discretionary parole release: second degree murder (fifteen years to life) and first degree murder (twenty-five years to life).43 Offenders serving either of these sentences are to be granted parole 44 after fifteen or twenty-five years, respectively, unless the parole board finds them unsuitable for parole by "some evidence."45 The other two murder sentences do not include the possibility of parole and are to be imposed if the jury finds certain statutorily enumerated special circumstances: the aptly named life without the possibility of parole (also known as LWOPP) and the death penalty.46 One special circumstance in California is that a murder was "especially heinous, atrocious, or cruel, manifesting exceptional depravity."47 If a murderer is sentenced to a parole eligible sentence, it necessarily means the jury did not find the murder heinous, atrocious, or cruel beyond a reasonable doubt, since that is a special circumstance mandating a sentence of LWOPP or death. According to the California parole statute, the parole board must examine the prisoner's commitment offense when it analyzes his potential threat to puhlic safety.48 One of the commitment offense factors determining unsuitability for parole is whether the commitment offense was heinous, atrocious, or cruel. 49 A parole board can deny parole indefithat "a large number of people who have exemplary prison records [are] being denied parole just because of the nature of their crime." Jd. 42. For example, Florida repeats some language about the heinousness of the crime in its murder statute and its parole release regulations. Sec Fla. Admin. Code Ann. r. 2321.010(5)(a)(I)(d) (1998) (listing "bru~,1 or heinous behavior" in offense as example of aggravation justifying beyond-matrix sentence). But New York does not permit denials of parole based solely on the commitment offense. See Edward R. Hammock & James F. Seelandt, New York's Sentencing and Parole Law: An Unanticipated and Unacceptable Distortion of the Parole Boards' Discretion, 13 St. John's J. Legal Comment. 527, 537 (1999) (ciLing Maye v. Russi, N.Y. LJ., Feb. 5, 1996, at 28). 43. Cal. Penal Code § 190.2(a) (West 2008). 44. Id. § 3041 (b) (West 2000 & Supp. 2008) (requiring board La ",eL a relea,e date unless it determines that ... public safety requires a more lengthy period of incarceration" (emphasis added». 45. E.g.. Superintendent v. Hill, 472 U.S. 445, 455-56 (1985) ("[T]he requirements of due process are satisfied if some evidence supports the decision by the plison disciplinal)' board ...."). For a detailed discussion of the Hill standard, see infra notes 3'10-3·15 and accompanying text. 46. Cal. Penal Code § 190.2(a) (West 2008). 47. 1d. § 190.2(a)(14). 48. Id. § 3041(b) (West 2000 & Supp. 2008). 49. Cal. Code Regs. tit. 15, § 2402(c) (I) (2005). 2009] HEINOUS, ATROCIOUS, AND CRUEL 903 nitely on its own finding that the murder was heinous, atrocious, or crue!.SD So we see the problem. DUling the trial, Apprendi requires the jury to find the fact that the crime is so heinous, atrocious, or cruel that the offender does not deserve parole. sl Ajudge cannot make this finding if a _~ __ ~~ ~~~_._ ju!Y..2id n()!:".~ ..!?.':'! ..g,~.j>a2:.()le_~1J()<lrt!£("n.ci~!lJ'.paE()I(:'Xo.Ll~(:,itl()lIS murders, even if the jury did not reach the issue or the jury explicitly found otherwise. 53 In practice, the existence of the heinous, atrocious, or cruel catchall means that almost no first and second degree murderers are ever released. A recent case from a California state court reviewed almost 2,700 parole denials and determined that the parole board found the commitment offense heinous, atrocious, or cruel in each and every case. S4 Statewide, only about one percent of parole eligible prisoners earns release each year. 55 The dysfunction in the California parole system is more than just an administrative problem, however. It raises some deeper questions at the heart of criminal procedure. Who gets to decide what when it comes to punishment? Why is the parole board deciding about the wrongfulness of crimes? Why is the parole board, and not the jury or judge, deciding anything about the commitment offense at all? The contours of these policy problems map precisely onto Apprendi's constitutional concernsso not only does an exploration of Apprendi via parole teach us about Apprendi, exploring parole via Apprendi teaches us about parole. * * * State and federal judges have cast about for legal rationales permitting them to overturn unsuitable life sentences. The problem is that parole is currently governed by an underdeveloped and confusing series of due process cases that is badly in need of review. Almost every limitation on parole board decisionmaking depends on a state's parole statute. Due process itself does not bar a parole board from considering any facts the 50. [n re Dannenberg, [04 P.3d 783, 802-03 (Cal. 2005) (finding sufficient evidence for Board's determination that defendant's crime was "especially callous and cruel"). 51. Apprendi v. New Jersey, 530 U.S. 466, 490 (2000) ("[A]ny fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to ajury, and proved beyond a reasonable doubt."). 52. For example, ajudge cannot impose a statutory sentence enhancement for use of a firearm unless the fact was '"alleged in the accusatory pleading and either admitted by the defendam in open court or found to be true by the trier of fact." Cal. Penal Code § 12022.53UJ· 53. See discussion infra Part I.C.3. 54. In re Criscione, No. 716[ 4, slip op. at I (Cal. Super. Ct. Aug. 30, 2007), availahle at http://wwvv.bayareanewsgroup.comimuitimedia/mn/news/criscione_complaint~0913 07.pdf (on rile with (he Colmnhia Law Review). 55. Sasha Abrarnsky, Barred for Life, SF Wkly., Aug. 15,2007, at 16 (citing California defense attorney). 904 . .._._~_~ COLUMBIA LAW REVIEW [Vol. 109:893 board deems relevant to its decisions, including the commitment offense. 56 Nor must the parole board find facts beyond a reasonable doubt: "Some evidence" is enough. 57 A prisoner has no inherent liberty interest in parole release,58 but state parole statutes can create that interest. 59 (Some recent Supreme Court prison administration cases, however, have rej ec!ed s.~tut()Ii.IL<:I"~<lte_tl_!i?~r-tLiIlter~s.ts._<lIltl.I<:l(}..l<(}_nlr.t()~leil~~e r-cIl! nature and weight of the deprivation involved.60) None of the due process cases limit~ the parole board's authority to its core institutional competence-measuring an offender's threat to public safety-and parole boards can therefore find facts that sound in both retribution and dangerousness. Cases using due process analysis to overturn unsuitability determinations are, accordingly, rare,61 but a few recent cases have so held. The Ninth Circuit found that a prisoner's liberty interest in parole ripens if she has been denied parole based on an unchanging factor. 62 The California Supreme Court has also recently held that the state parole statute requires a finding that the commitment offense is some evidence of present dangerousness-not merely that it provides some evidence of an unsuitability factor. 63 While these cases might yield the desired result, an Apprendi-based limit would be more clearly grounded in the Constitution, would more closely align with justifications for sentencing, and would allow the court to address the problem of unsuitable life sentences directly. Courts have declined to use this solution, however.64 Perhaps this is due to the Apprendi line's formalist reputation, as judges are loath to extend Apprendi beyond the "formal" context of judicial sentencing. 65 Another barrier could be the misapprehension of how Apprendi and Blakely idiosyncratically define the statutory maximum punishment: Sentences with 56. See discussion infra Part II.A. L 57. See discussion infra Part N.A.l. 58. Greenholtz v. Inmates of Neb. Penal & Corr. Complex, 442 U.S. 1, 7 (1979) ("There is no constitutional or inherent right of a convicted person to be conditionally released before the expiration of a valid sentence."); see also infra notes 262-267 and accompanying text. 59. See discussion infra Part IILA.I. 60. See discussion infra Part IILA.I. 61. For a discussion of recent cases overturning parole board decisions on these grounds, sec infra notes 161-163 and accompanying text. 62. Ha}ward v. Marshall, 512 F.3d 536, 546-47 (9th Cir. 2008) (finding Governor violates defendants' due process rights by relying on "stale and static factor" when denying parole). The U.S. Supreme Court has not dealt with the unsuitable life sentence issue. 63. In re Lawrence, 190 P.3d 535, 553 (Cal. 2008) ("[T]he relevant inquiry is whether some evidence supports the decision . .. that the inmate constitutes a current threat to public safety, and not merely ... the existence of certain factual findings."). 64. Sec discussion infra note 309 and accompanying text. 65. Sec infra text accompanying notes 224-226 (questioning whether Apprendi is really a ··sentencing" decision at all). .0___0_ 2009] Hr-iNOUS, ATROCIOUS, AND CRUEL 905 presumptive parole release have statutory maxima of the enumerated tenn of years, not life. 55 Ultimately, Apprendi must be understood to bar a parole board from encroaching on the jury's power, but its application should be limited only to those facts that justify continued punishment. A mechanical ap.~.. ~.~~~~~.~.~.~~ ...~.l'Ji~atiQI!2L4PP:r:":,,di!.()1l~t~1<lt.f<liJ~.to(ji~tingllisl1])et.':"~.e.I1 .. r~tri])lltiveand . rehabilitative incarceration, would require a jury to find all facts justifying a finding of parole unsuitability, including those related to public safety. This would upset well-established principles of deference in areas in which the parole board is most competent, and it would fail to advance Apprendi's core interests. But because sentences terminating in discretionary parole release combine retributive and rehabilitative components, they provide the ultimate test case for determining where-and, more importantly, why-the Apprendi right applies. Exploring Apprendi in the context of these sentences restores needed coherence to the Apprendi right, and it provides a clear example of why different parts of a single sentence, and different facts justifying the imposition of that sentence, do not all serve the same purpose. Different rights and interests attach to each. * * * This Article is divided into five parts. In Part I, I define the "problem" of the unsuitable life sentence: the denial of parole based on facts about the commitment offense. The parole board's finding of unsuitability on the basis of the commitment offense is the "text" I then examine under the Apprendi and parole due process lines of cases. In Part II, I take a deeper look at the jUrisprudence of punishment as a means of explaining why a jUlY needs to find certain facts and why a parole board can find others. Part III explores whether an eligible prisoner can be deemed "punished" when the parole board finds her unsuitable for parole. Part IV examines what, if anything, justifies the different standards of proofjudges, juries, and parole boards use to find the same facts. Part V discusses some of the policy considerations arising from the application of Apprendi to parole. Woven throughout these issues are potential separation of powers and due process concerns, and questions about whether the purpose of criminal penalties is retributive or rehabilitative. I conclude that while Apprendi must apply to parole board suitability hearings, it should not apply to all facts in those hearings. The Apprendi and due process doctrines can best be reconciled by looking to the purpose of each part of a sentence that tenninates in discretionary parole 66. I will also argue that Apprendi applies under a fannal (or taxonomic) approach as welL The legislature has distinguished parole ineligible murders from lesser, parole eligible ones, based, in part, on whether the murders were heinous, atrocious, or cruel. The parole board is thus both functionally increasing punishment and violating the legislative taxonomy of degrees of murder. See discussion infra Part IlI.B.2. COLUMBIA LAW REVIEW 906 ~. .~_ ..~ [Vol. 109:893 release. The enumerated term of years an inmate serves before his first parole hearing-a term which depends on jury findings about the commitment offense-is punitive, and the parole board may not extend this punishment. The second "to years" portion governing release is rehabili, tative, and Apprendi should not bar the parole board's ability to make ~~_flJ:l!,iiJ:l,gu~l~~~J:l!_t..(Uhis~ai!J:l:~_g!~..!h:.cl~l!J:leatiJ:1g~!!l.!: .. role of !11_e..l'.a_I()\.!:_~~___~. ~_~__ ~_ board as a rehabilitative/public safety body has one final benefit: restoring discretionary parole release as a useful tool in sentencing. If a parole board is properly constrained to find facts about potential threats to public safety, it is an extremely effective part of the sentencing process, one uniquely situated to determine dangerousness at the point of release. Removing punishment from the parole board's jurisdiction in the way Apprendi suggests would give states all of the benefits of discretionary parole release as a tool while reducing the more arbitrary and capricious aspects of its administration. I. UNSUITABLE LIFE SENTENCES Indeterminate sentences are sentences terminating in discretionary parole release. California, which uses a mixture of determinate and indeterminate sentencing, grants release to indeterminately sentenced prisoners after they are found suitable for parole. In practice, however, California's parole boards rarely find prisoners suitable, as "some evidence" exists to find almost every commitment offense unsuitably heinous, atrocious, or cruel, even when juries have failed to make this finding beyond a reasonable doubt. The result is that an offense that justifies a parole eligible sentence is found, by itself, sufficient to render a prisoner unsuitable for parole, transforming an indeterminate sentence into what I call an unsuitable life sentence. This Part explores California's sentencing and parole release process to map the contours of the unsuitable life sentence problem. A. Clarifying Terms: Indeterminate Sentences Are Sentences with Discretionary Parole Release The time an inmate spends in prison is potentially subject to two types of discretion: the judge's discretion at the point a sentence is imposed, and the parole board's discretion at the point of release. Not all systems have both kinds of discretion. Since the Apprendi line limits some types of discretion but not others, I use this section to define and distinguish the commonly used (and commonly confused) sentencing terms "determinate" and "indeterminate." ClarifYing these terms helps explain why and when Apprendi governs. In this Article, I use the term "indeterminate" to refer to sentences subject to parole board discretion at the point of release, and "determinate" as its antonym, referring to sentences 2009] HEINOUS, ATROCIOUS, AND CRUEL 907 where a prisoner must be released after a fixed term-no more, no less. 67 I use "advisory" to refer to systems that allow judicial discretion at the point a sentence is pronounced, and "binding" as its antonym, referring to a system where ajudge must impose a presumptive sentence upon conviction. Discretionary systems can be further characterized as guided/ ..__ .~.~s!J:l1ct,,:~':(UEE.£'tdll1.~,,:gg!:s.t,,_c1. ral1.g"s()fter_IIis.()fyears)_()r_llngllidc:<iL______________ unstructured (making no suggestions about penalties within the statutory range).68 Clarifying this terminology immediately resolves some of the muddier parts of the Apprendi doctrine, as the Supreme Court has often conflated these ly£es of discretion,69 using "indeterminate" to mean "advisory" and "determinate" to mean "binding" (i.e., determinative of the outcome). In Blakely, for example, Justice Scalia wrote that "[i]ndeterminate sentencing ... increases judicial discretion" and that "indeterminate schemes involve judicial factfinding."7o Similarly, Justice A1ito, dissenting in Cunningham, equated "fully discretionary sentencing" with "indeterminate sentencing ranges," and concluded that (at the 67. These definitions follow Steven Chanenson's succinct summary: "Indetcrnlinate systems use discretionary parole release while determinate systems do not." Steven 1.. Chanenson, The Next Era of Sentencing Reform, 54 Emory LJ. 377, 382-83 (2005) [hereinafter Chanenson, Next Era]. Chanenson's definitions, in turn, track those used by the Bureau of Justice Assistance: Determinate sentencing involves "a fixed leOll" and indetenninatc sentencing involves "an administrative agency, generally a parole board" with "the authority to releao;.e an offender." 1996 National SUIvey, supra note 31, at 1-2. Because the federal system has used only determinate sentences since 1984, when Congress eliminated discretionary parole release, I do not discuss federal cases beyond Booker. See Sentencing Reform Act of 1984, Pub. L. No. 98-473, § 3624, 98 Star. 1987, 2008-2009 (codified as amended at 18 U.S.C. § 3624 (2006» (indicating that prisoner is to be released "on the date of the expiration of his tenn of imprisonment" less only time credited for good behavior). It is confusing that federal sentencing ranges sometimes include the phrase "X years to life," but the phrase indicates a range of determinate terms ajudge may impose-up to life in prison-not an indeterminate sentence itself. In other words, a federally sentenced prisoner will never get an indeterminate, fifteen-years-to-life sentence. He might get any number of determinate sentences within that range, however: fifteen years, life, or some enumerated term in between. 68. Chanenson, Next Era, supra note 67, at 383. The post-Booker Federal Sentencing Guidelines are a determinate, advisory, structured system. They suggest, but do not require, determinate sentences of years for particular commitment offense facrs. 69. The Supreme Court is in good company here. The 1996 National Survey of Scale Sentencing Structures observed "a lack of consensus regarding the meaning of commonly used terms" including determinate sentencing, indetenninate sentencing, and the two types of sentencing guidelines (advisory and presumptive). 1996 National Survey, supra note 31, at 1. Almost ten years later, Steven Chanenson noted that the labels "indeterminate" and "determinate" were still frequently applied "imprecisely or improperly, leading to confusion . .. [that] has become more acute in the immediate aftermath of Blakely." Chanenson, Next Era, supra note 67, at 381-82. 70. Blakely v. Washington, 542 U.S. 296, 309 (2004). 908 COLUMBIA LAW REVIEW [Vol. 109:893 judge's discretion) these ranges yield a "precise sentence" at the time of sentencing?' There are some indications that Apprendi might not govern systems with discretionary parole release, however. Justice Kennedy stated in his Cunningham dissent that an indeterminate sentencing system, which re...... _~~~.~_.J?os~.!L<:\i§.<;retioIl2cryI'"I..,<lSe.:.i!'..~.()!iliJdicialagencY..1()_S!".L'!...t:<,,!ea§~.ga.t~ . for convicted felons," would seem to be "untouched by Apprendi.""" Justice Scalia observed in Blakely that "indeterminate schemes involve judicial factfinding, in that a judge (like a parole board) may implicitly rule on those facts he deems important to the exercise of his sentencing discretion" without violating the Sixth Amendment. 73 While at first blush these statements would seem to forestall the argument I raise in this Article,74 a closer examination reveals that the Justices have conflated two different types of discretion. Indeterminate sentences do not necessarily involve judicial faetfinding. For example, a system can grant a judge discretion to impose an indeterminate sentence, but it could also bind her to a particular indeterminate sentence for a given set oEfacts. 75 Similarly, regulations might bind parole boards in their consideration of release (as in California), or parole boards might have no guidelines or purely advisory guidelines. Justice Scalia's statement further assumes that parole must be "early release" to whieh a prisoner is not otherwise entitled,76 even though the Supreme Court has itself twice found that a parole scheme can create an expectancy of parole protected by the Fourteenth Amendment Due Process Clausen Determining 71. Cunningham v. California, 549 U.S. 270, 298 n.1 (2007) (Ali to, j., dissenting) (internal quotation marks omitted). 72. Id. at 295-96 (Kennedy, j., dissenting). Although Justice Kcnncdy "asked a number of questions about indeterminate sentencing, parole and parole boards, parole commissions, and parole guidelines" during ora] argument in Booker, Robert Weisberg, Excerpts from "The Future of American Sentencing: A National Roundtable on Blakely," 2 Ohio St.]. Crim. L. 619, 633-34 (2005), his interest appears only in these two sentences in Cunningham. The Supreme Court has offered no more guidance on exactly why and how Apprendi could be reconciled with indeterminate sentences. 73. Blakely, 542 U.S. at 309. 74. See, e.g., Appleman, supra note 3, at 1372 (noting that this excerpt from Blakely "seems to specifically exempt parole"). 75. Note also that an indeterminate sentence need not include the possibility of life in prison. An indetcmlinate sentence could lake the form of "five to ten years," with five years the enumerated term and "to ten" denoting the period during which the prisoner could be released into parole. ]n a "five to ten" sentence, an inmate could be released five years into his sentence, but not sooner, or ten years into his sentence, but not later. 76. Blake(v, 542 U.S. at 309 ("But the facts do not pertain to whether the defendant has a legal Tight La a lesser sentence-and that makes all the difference insofar as judicial impingement upon the traditional role of the jury is concerned."). 77. See Bd. of Pardons v. Allen, 482 U.S. 369, 377 n.8 (1987) (rejecting idea that "parole is ... a matter of grace" and finding that while states are not required to establish parole systems, and may place conditions on parole release, it is otherwise a "right"); Greenholtz v. Inmates of Neb. Penal & Corr. Complex, 442 U.S. I, 12 (1979) ("[T]he expectancy of release provided in [Nebraska's] statute is entitled to some measure of 2009] HEINOUS, ATROCIOUS, AND CRUEL 909 whether and how Apprendi applies thus requires us to look to the features and operations of a given parole system.?8 California's indetenninate sentences do, in fact, bind the parole board: The board must find a prisoner suitable for release unless it finds facts justifying continued incarceration 7 " The board is bound to release .....................~ Pr;~£:':l~.t: <t!:t.~.~.l1is . pr~sllmptive sel1.t~I1.<:~.=!I1.~ ~.Il.ll.I11.~E<l!e.~ terl11 of years-unless it finds specific facts justifying a longer sentence. California's parole system therefore exhibits all of the features of the sentencing regime Blakelv found unconstitutional so B. California Uses Both Determinate and Indeterminate Sentences California's sentencing system is largely detenninative, but it contains a hodgepodge of indeterminate practices left over from an earlier era. The existence of discretionary parole release in a largely determinate system represents a transitional moment in sentencing policy. Analyzing Apprendi's application to discretionary parole release is at the center of one of Apprendi's primary issues: what criminal penalties mean. This section provides an overview of California sentencing, focusing on this combination of retributive and rehabilitative practices. At the beginning of the 1970s, every state, the federal government, and the District of Columbia used indeterminate sentencing schemes S1 California courts did not detennine the length of imprisonment; instead, prisoners were sentenced under the Indeterminate Sentence Law to the range of years prescribed by statute-which could be as vague as "one constitutional protection."). The Allen Court found that, while the state has no duty to establish a parole system, 482 U.S. at 377 n.8, parole statutes can and do creatc duties and rights, id. at 378 n.9 ("This Court ... [has] recognized the relevance of regulations to a determination of whether a certain scheme gives rise to a liberty interest . .. [T] he Montana statute ... obligate[s] the Board to consider certain information in making its parole-release decision." (internal citations omitted)). But see Ohio Adult Parole Auth. v. Woodard, 523 U.S. 272, 280 (1998) (citing Greenholtz in clemency case for proposition that "[t]he individual's interest in release or commutation 'is indistinguishable from the initial resistance to being confined,' and that interest has already been extinguished by the conviction and sentence" (internal citations omitted)). 78. This, too, is part of the theme that Apprendi is functional, not [annal. See infra Parts III, IV. 79. That is, even though the application of facts to parole standards is, in some sense, discretionary (i.e., parole boards can decide whether a given fact meets a standard), the decision to release is governed by binding regulations. The last time the Supreme Court issued a significant opinion about parole, in Allen, 482 U.S. 369, it made the same distinction. Allen distinguished between the "two entirely distinct uses of the tcon discretion," one meaning "not bound by standards set by ... authority," the other meaning to "use judgment in applying the standards set ... by authority." Id. at 375 (internal quotation marks omitted). In the former instance, no statute or regulation binds the judge or parole board, and a prisoner thus has no statutorily created protected liberty interest in release. In the latter instance, the stanue or regulation specifies the contours of the liberty interest. See discussion infra notes 264-273 and accompanying text. 80, See supra text accompanying notes 12-13. 81. Michael Tonry, Reconsidering, supra note 31, at 1. 910 COLUJi,1BJA LAW REVIEW [Vol. 109:893 year to life"-and the parole board "determined the amount of time a felon would ultimately spend in prison."82 Indeterminate sentencing faced growing criticism in the 1970s both nationally and in Califo rnia 83 Some critics argued that there was evidence that parole reduced recidivism; others focused on the arbitrariof release decisions"~4I!!J~I71g<lli[()E!!i<l.'sQSLr eplaced. nate sentences with fixed terms for most crimes,s5 declaring explicitly that "the purpose of imprisonment for crime is punishment."86 California's determinate sentencing scheme punished offenders with "tenus proportionate to the seriousness of the offense with provision for uniformity in the sentences of offenders committing the same offense under similar circumstances. "87 82. Cunningham v. California, 549 U.S. 270. 276-77 (2007). 83. See, e.g., Kate Stith & Jose Cabranes. Fear of Judging 29-36 (1998) (discussing revolt against discretionary sentencing); Jeremy Travis, But They All Come Back 17-18 (2005) ("Reliance on the exercise of discretion by judges, corrections administrators, parole boards, and parole officers was criticized as arbitrary, racially discriminatory, and fundamentally unfair."). 84. See, e.g., Petersilia. When Prisoners Come Home, supra notc 32, at 65 (explaining that those seeking to deemphasize rehabilitation sought in particular to abolish indetenninate sentencing and discretionary parole release); David Rothman, Conscience and Convenience: The Asylum and Its Alternatives in Progressive America 159 (1980) (entitling chapter on parole, "A Game of Chance"). 85. 1976 Cal. Stat. 5140 (codified as amended at Cal. Penal Code § 1170(a)). invalidated by Cunningham, 549 U.S. at 274; see infra note 87 (noting change in DSL to advisory after Cunningham decision). For the historical background of indeterminate sentencing, see Petersilia, 'When Prisoners Come Horne, supra note 32, at 55-68. 86. 1976 Cal. Stat. 5140. The name of the Department of Corrections was, however, changed to the Department of Corrections and Rehabilitation in an attempt to signal a shift in the department's priorities in the opposite direction. 2005 Cal. Legis. Servo 18 (West); Sara B. Miller, California Prison Boom Ends, Signaling a Shift in Priorities, Christian Sci. Monitor, June 20. 2005. al3; see also Cal. Penal Code § 1170(a)(2) (stating, in section immediately following declaration that punishment is the purpose of imprisonment, that "[t}he Legislature encourages the development of policies and programs designed to educate and rehabilitate nonviolent felony offenders"). 87. 1976 Cal. Stat. 5140. The DSL regime subjected most offenses to three possible fixed-term sentences-a lower, middle, and upper term-rather than a broad, open-ended prison term. See Cunningham, 549 U.S. at 277. A guilty verdict or plea resulted in the imposition of a presumptive middle-tenn sentence unless the judge found facts in aggravation (justifying an upper-tenn sentence) or in mitigation Uustifying a lower-term sentence). Id. (citing Cal. Penal Code § 1170(b) (West Supp. 2006)). In 2007. however, Cunningham v. California applied Apprendi to the DSL and declared it unconstitutional. Id. at 274, Because the DSL gave judges "authority to find the facts that expose a defendant La an elevated 'upper term' sentence," facts that were "neither inherent in the jury's verdict nor embraced by the defendant's plea" and which needed only to be "established by a preponderance of the evidence, not beyond a reasonable doubt," it violated the Appre11di right to trial by jury. Id. Among the facts in aggravation needing to be found by a JUT)' were "facts relating to the crime," including those "disclosing a high degree of cruelty. viciousness, or callousness." Id. at 278 & n.7. To cure the constjtutional infirmities identified in Cunningham, California has since made the DSL wholly a.dvisory. See Cal. Penal Code § 1170(b) (West Supp. 2008) ("When ajudgment of imprisonment is to be imposed and the statute specifies three possible terms, the choice of the appropriate term 2009] HEINOUS, ATROCIOUS, AND CRUEL 911 The DSL did not completely overhaul the system, however, and what remains mixes and matches determinate and indeterminate sentencing practices with retributive and rehabilitative goals. For example, California continues to impose indeterminate sentences for crimes that seem particularly reprehensible: sex offenses,s8 homicides,s9 and "three stti k':'~::':'£feIls,:,s:!l~~lli~~ilityf,:,,:.p,,r,:,l,:,re Ieas.':' is~ ':'':'<tI.ll."t.':'<.I ~''cs':'<.I ':'Il "Il.. offender's threat to public safety, yet the means by which this threat is analyzed uses moral criteria like heinous, atrocious, and cruel. Language about term uniformity in the parole statute arguably mirrors the proportionality concerns of the DSL,91 yet consideration of the individual offense requires the parole board to "eschew term uniformity, based simply on similar punishment for similar Climes, in the interest of public safety in the particular case."92 Finally, California has retained post-release parole supervision for all prisoners, including those serving determinate sentences, even though the discretionary parole release decision is made only for indeterminately sentenced prisoners."' C. California's Parole Release Process This section reviews California's parole release process. California law sets out a two~stage procedure for determining when prisoners serving indeterminate sentences can be released. First, a panel of parole board members decides whether a prisoner is suitable for parole. 94 If a shall rest within the sound discretion of the court"); 2007 Cal. Legis. Servo 1 (West) ("It is the intent of the Legislature in enacting this provision to respond to the decision of the United States Supreme Court in Cunningham v. California."). Cunningham addressed only the DSL-it did not apply Apprendi to California's indeterminate sentences. 88. See, e.g., Cal. Penal Code § 667.51 (West 1999 & Supp. 2008) (prior lewd act with a child); id. § 667.61 (sex offenses): id. § 667.71 (habitual sex offender). 89. Id. § 190 (West 2008). 90. [d. § 667(e)(2) (A) (West 1999 & Supp. 2008). Support for the "three strikes" initiative among the California public was linked to a "general support for harsher punishment of rule breakers," even though the third strike is punished with an indeterminate, twenty-five-years-to-life sentence. See Tom R. Tyler & Robert J. Boeckmann, Three Strikes and You Are Out, but Why? The Psychology of Public Support for Punishing Rule Breakers, 31 Law & Soc'y Rev. 237, 250 (1997). 91. In re Dannenberg, f04 P.3d 783, 791 (Cal. 2005) (explaining that purpose of punishment is "best served by terms proportionate (0 the seriousness of the offense" (quoting Cal. Penal Code § 1f70(a)(I) (West 2004»). 92. Id. at 794. 93. Cal. Penal Code § 3000(b) (I) (Wes' 2000 & Supp. 2008). Most prisoners on parole served determinate sentences. Pctersilia, When Prisoners Come Home, supra note 32, at 59. Because parole is automatically granted for determinate sentences, I do not discuss it in this Article. I note, however, that the system of mandatory parole shows signs of changing: The state is testing reforms that would give certain parolees "earned discharge" from parole supervision after six months, focusing resources on the most crimeprone parolees. See Joan Petersilia, Op-Ed., Parole, the Right Way, L.A. Times, Oct. 8, 2007, at Al5 [hereinafter Pctcrsilia. Parole]. 94. See Cal. Penal Code § 5076.1 (C) ("The board may meet and transact business in panels. Each panel shall consist of two or more persons . ... "). 912 COLUMBIA LAW REVIEW [Vol. 109:893 prisoner is found suitable, the panel proceeds to the second stage (release date calculation), when it determines how much time the inmate must serve before release. 95 In both stages, the panel makes its decision based on facts about the commitment offense that were removed from the judge's purview in the Apprendi line. Practically all panels designate ................. .............................................t;h".e,•. . c,o':mmitment offense as "heinous, atrocious, or cruel" up()Iltl1eJil1c1~ ing of "some evidence," transforming indeterminate sentences into the LWOPP-equivalent unsuitable life sentence."6 1. Parole Suitability Basics. - One year before a California prisoner's minimum eligible parole release date (MPRD),97 he has his first parole suitability hearing 98 The issue before the panel during a suitability hearing is whether the prisoner "is suitable for parole, not when he should be released."99 A prisoner unsuitable for parole is one who "will pose an unreasonable risk of danger to society if released from prison."roo If the prisoner is found unsuitable for release, the panel "must provide a definitive written statement of its reasons for denying parole"ror and must have "some evidence" to support its decision Io2 The prisoner's next parole 95. In re Elkins, 50 Cal. Rptr. 3d 503, 520 ( Ct. App. 2006) (citing Cal. Code Regs. tit. 15, §§ 2402-2403 (2005)). "[T]he Board shall first determine suitability and shall set a base term (thus establishing a parole release date) if the prisoner is deemed suitable for parole." Dannenberg, 104 P.3d at 800. The existence of these two stages has "long been noted in the case law." Id. at 792. 96. Irons v. Warden of Cal. State Prison-Solano (Irons 1), 358 F. Supp. 2d 936, 942 (E,D. Cal. 2005). When a parole board finds a crime heinous, atrocious, or cruel that a jury did not, it "transforms an offense for which California law provides eligibility for parole into a de facto life imprisonment without the possibility of parole. . . . The circumstances of the crimes will always be what they were, and petitioner's motive for committing them will always be trivial." Id. at 947. 97. The MPRD is calculated by taking the minimum term of the sentence and subtracting credits for "good behavior and participation." See Cal. Penal Code §§ 2930-2935 (West 2000) (detailing accumulation of and reduction in such credits). The precise fonnula depends on a number of factors-different crimes yield good time at different rates, or require a defendant to serve a given percentage of the sentence. See generally Steven Fama et al., California State Prisoners Handbook 172-74 (3d ed. 2001) (summarizing different good time formulas). The calculation of good time is sufficiently complicated that thousands of California inmates have had their release dates miscalculated. See Julia Reynolds, Suit Says Inmates Released Late, Monterey County Herald, Dec. 12, 2007, at B1. In an interesting twist, the prison guards' union, not the affected inmates, brought the suit against the corrections system, alleging that understaffing prevented the guards from calculating release dates accurately. Merely serving the minimum term of an indeterminate sentence does not, of course, entitle a prisoner to parole. Dannenberg, 104 P.3d at 786. 98. See Cal. Penal Code § 3041 (a) (West 2000 & Supp 2008) (detailing procedure of parole suitability hearing). 99. Sass v. Cal. Bd. of Prison Tenns, 461 F.3d 1I23, 1132 (9th Cir. 2006) (Reinhardt. J., dissenting). 100. Cal. Code Regs. tit. 15, § 2402(a). 101. In re Rosenkrantz, 59 P.3d 174, 203 (Cal. 2002). 102. Id. at 183. For a discussion of how the "some evidence" requirement makes judicial review practically impossible, see infra Part N.A.2. 2009] HEINOUS, ATROCIOUS, AND CRUEL 913 suitability hearing can be scheduled anywhere from one to five years later, a term left to the parole board's discretion. 103 The determination of parole suitability takes place at a healing which follows the same format as a trial: There is a finding of fact, and a decision is made on the basis of that finding of fact. The issue of which facts can.be found, by what body, according to which standard of proof, and with what discretion to impose a result, all have clear analogies in the sentencing process Apprendi governs. 104 Consider, as an example, Sandra Lawrence's parole hearing. I05 Lawrence, her attorney, and a Deputy District Attorney from Los Angeles were present. 106 Lawrence was found essentially "competent" to be present at the hearing. 10? The details of the committnent offense-her state of mind, the weapons she used, details about the murder-occupied much more of the hearing than discussion of her conduct in prison or post-release plans. lOS After Lawrence concluded her testimony, the District Attorney recommended that the panel find her unsuitable for parole. 109 Lawrence's attorney responded with her own closing arguments,1I0 and Lawrence then spoke on her own behalf. l l l After a recess, the panel returned with a sentencing decision-that Lawrence was suita- 103. Cal. Code Regs. tit. 15, § 2268 (2008). Delays of more than one year must be justified by written findings. Id. 104. My equation of parole board hearings and trials/sentencing is not novel. David Rothman describes parole board hearings a<; "in effect retrying, in capsule fashion, the original offense." Rothman, supra notc 84, at 166. 105. In rc Life Term Parole Consideration I-fearing of Sandra Lawrence, CDC No. W19366 (Cal. Inst. for Women Aug. 25, 2005) (on file with the Columbia Law Review). 106. Id. at 1-2. 107. The parole panel made sure she did not have any disabilities that needed to be accommodated, and Lawrence testified that she had not participated in any programs for inmates with mental health issues. Id. at 2-5. 108. The testimonial portion of the hearing transcript ran approximately seventy-one pages, id. at 10-81, of which a total of roughly forty pages was devoted to the commitment offense, see id. at 11-36 (initial testimony about the murder); id. at 40-41 (influence of drugs/alcohol at time of the murder); id. at 42-48 (Governor's prior reversal, focusing on commitment offense, read into record); id. at 64-67 (follow up questions about potato peeler used in murder); id. at 72-74 (District Attorney's questions about her stale of mind at time of murder); id at 75 (Lawrence's attorney asks about commitment offense). The testimony devoted to everything else ran approximately thirty pages. See id. at 36-38 (juvenile record and early social history); id. at 38-40 (history with alcohol and drugs); id. at 49-53 (custodial behavior, including discipline); id. at 53-60 (letters supporting her release); id. at 60-63 (psychiatric report); id. at 68-71 (post-release employment plans); id. at 75-79 (summary of how she rehabilitated in prison); id. at 79-81 (recovery from alcohol). 109. Id. at 81-83. The District Attorney focused almost entirely on the murder as well. Id. 11 O. Id. at 83-88. 111. Id. at 88-89. 914 COLUMBIA LAW REVIEW [Vol. 109:893 ble for parole l12-then calculated a release date, determining the remaining time she would serve on her sentence 1l3 When calculating a release date, the panel looks to characteristics of the commitment offense. 1I4 The panel refers to a matrix to set a "base term" of incarceration, then adds enhancements for factors such as possession ()f a firea nn during the offense, 11 5 .The nlatrices bear a striking resemblance to the Federal Sentencing Guidelines Booker invalidated. Each matrix establishes three term lengths for a given offense, depending on how the crime was committed, the relationship between the offender and the victim, and the injury to the victim. 1I6 If there is no matrix for a given offense, the board establishes a base term by comparing the crime to offenses of similar gravity and magnitude. II? As a practical matter, when a prisoner is eventually found suitable for parole, he is almost always long past even the highest range of the release-date guidelines. 1I8 Therefore, I do not spend much time discussing the Apprendi ramifications of release-date calculations, even though they, too, use commitment offense facts to determine time served in prison. If a parole panel finds a prisoner suitable for parole, the parole board as a whole can review and remand a panel's suitability decision within 120 days. 11 9 The governor can also reverse suitability findings 112. Id. at 90. 113. Id. at 93-94. Her total period of confinement was calculated at 130 months, starting in 1983. Id. at 94. The hearing took place in August 2005. Even assuming that the trial concluded on the very last day of 1983, Lawrence still would have served at least 261 months-more than double the sentence time the parole board eventually calculated. 114. In re Dannenberg, 104 P.3d 783, 800 (Cal. 2005) ("[T]he Board shall first determine suitability and shall set a base term (thus establishing a parole release date) if the prisoner is deemed suitable for parole."); In re Elkins, 50 Cal. Rptr. 3d 503. 520 (CL. App.2006) (citing Cal. Code Regs. tit. 15, §§ 2402-2403 (2003». This process is meant to ensure that "uniformity in sentencing is taken into account." In re Scott (Scott!), 15 Cal. Rptr. 3d 32. 42 n.6 (Ct. App. 2004). 115. Cal. Code Regs. tit. 15, § 2285 (2003). The release date can be years into the future. Elkins, 50 Cal. Rptr. 3d at 520. In no case maya prisoner serving life be released without having served the greater of seven calendar years or a "minimum period of confinement under a life sentence before eligibility for parole." Cal. Penal Code § 3046 (West 2000 & Supp. 2008). In addition, no prisoner may be released on parole within sixty days of his suitability hearing. Id. § 3042(b). 116. See Fama et al., supra note 97, at 181. 117. Id. 118. See infra Part J.D. 119. Cal. Penal Code § 3041 (b). Parole can also be rescinded between a successful suitability hearing and the actual date of release. Cal. Code Regs. tit. 15, § 2450 (2000). (The date can also be moved up, pursuant to id. § 2269 (2003).) Behaviors leading to rescission include serious disciplinary issues (assault, escape, distribution of intoxicants) and psychiatric deterioration. Id. § 2451 (a) (2000). The board can also initiate rescission proceedings if "fundamental errors" resulted in the "improvident granting of a parole date," or if "[a]ny new information ... indicates parole should not occur." ld. § 2451(c). As in the initial parole hearing, the evidentiary standard for the board's ability to rescind parole is the very low "some evidence" threshold that is derived from Superintendent v. Hill, 472 U.S. 445 (1985). For a discussion of the Hill standard, see infra Part IVA 2009] HEINOUS, ATROCIOUS, AND CRUEL 915 within thirty days if the commitment offense was murder. J20 Gubernatorial reversals of parole grants must be made on the same basis as parole board decisions themselves. 121 2. The Use of Commitment Offense Facts in Suitability Determinations. Although the panel is to consider "[a]1I relevant, reliable information available." in determining suitability,122 the statute governing parole release focuses exclusively on the offender's commitment offense and criminal history.123 One of the circumstances "tending to show unsuitability" for release is the nature of the commitment offense: Crimes committed "in an especially heinous, atrocious, or cruel manner" tend to indicate unsuitability for release. 124 "[A] conviction for second degree murder" is not supposed to "automatically render one unsuitable" for parole by virtue of the fact that the offender committed the crime. 125 The Califomia regulations enumerate specific facts indicating that a murder was heinous, atrocious, or cruel, such as a murder with multiple victims (including those assaulted or injured), crimes "carried out in a dispassionate and calculated manner," 120. Cal. Penal Code § 3011.2 (West 2000). 121. In re Elkins, 50 Cal. Rptr. 3d 503, 512-13 (Ct. App. 2006). 122. Cal. Code Regs. tit. 15, § 2402 (2005). These factors include the circumstances of the prisoner's social history; past and present mental state; past criminal history, including involvement in other criminal misconduct which is reliably documented; the base and other commitment offenses, including behavior before, during and after the crime; past and present attitude toward the crime; any conditions of treatment or control, including the use of special conditions under which the prisoner may safely be released to the community; and any other information which bears on the prisoner's suitability for relca')c. Id. The cited regulations refer specifically (Q "murders committed on or after November 8, 1978," id. § 2400, but they are substantially the same for other prisoners serving indeterminate life sentences, see, e,g., id. ("The suitability criteria are the same for [those convicted of murdel-s on or before NO\!cmber 7, 1978] ... ."). Whether the governing rules are statutory or regulatory "lacks constitutional significance." United States v. Booker, 543 U.s. 220, 237 (2005) (Stevens, j., opinion of the Court). 123. The statute provides: The panel or the board, sitting en bane, shall set a release date unless it determines that the gravity of the currcnt convicted offense or offenses, or thc timing and gravity of currcnt or past convicted offense or offenses, is such that consideration of the public safety requires a more lengthy period of incarceration for this individual, and that a parole date, therefore, cannot be fixed at this meeting. Cal. Penal Code § 3011 (b) (West 2000 & Supp. 2008). 124. Cal. Code Regs. tit. 15, § 2402(c) (1). Other factors include a previous history of violence, an unstable social history. sadistic sexual offenses, a history of mental problems, and institutional misconduct. Id. § 2402(c). 125. In re Scott (ScottI), 15 Cal. Rptr. 3d 32, 45 (Ct. App. 2004) (quoting In re Smith, 7 Cal. Rptr. 3d 655, 673 (Ct. App. 2003); see also In re Dannenberg, 104 P.3d 783,802 (Cal. 2005) ("[T]he violence or viciousness of the inmate's crime must be more than minimally necessary to convict him of the offense for which he is confined."). 916 COLUMBIA LAW REVIEW [Vol. 109:893 mutilation of victims, "callous disregard for human suffering," and a trivial motive for killing. 126 In practice, however, these enumerated factors provide no obstacle to deeming almost all murders heinous, atrocious, or cruel. In re Criscione, a recent case from a California state court, reviewed thousands of parole denials and found that "in. one hundred percent of 2690 randomly chosen cases, the [Parole] Board found the commitment offense to be 'especial1y heinous, atrocious, or cruel', a factor tending to show unsuitability under Title 15 § 2402(c)(1)."127 The court emphasized t11at parole boards found both the absence of and existence of the same facts to be indicators of unsuitability: For example, if the inmate's actions result in an instant death the Board finds that it was done in a "dispassionate and calculated manner, such as an execution-style murder." At the same time the Board finds that a murder not resulting in near instant death shows a "callous disregard for human suffering" without any further analysis or articulation of facts which justiry that conclusion. If a knife or blunt object was used, the victim was "abused, defiled, or mutilated." If a gun was used the murder was performed in a "dispassionate and calculated manner, such as an execution-style murder." If bare hands were used to extinguish another human life then the crime is "particularly heinous and atrocious."128 3. Deferential Review. - Prisoners found unsuitable can only challenge the board's finding t11rough state habeas corpus review: California does not al10w direct appeal of parole board decisions. 129 In habeas, the prisoner bears the burden of proof on any disputed issues of fact. 130 Deference to the parole board's "broad discretion over parole suitability decisions" means that "courts should refrain from reweighing the evidence [during state habeas], and should be reluctant to direct a particular re126. Cal. Code Regs. tit. 15, § 2402(c)(I). 127. In re Criscione, No. 71614, slip op. all (Cal. Super. Ct. Aug. 30, 2007), available at http://bayareanewsgroup.com/multimedia/mn/news/criscionc_complainc091307. pdf (on file with the Columbia Law Review). 128. [d. at 13. The opinion continued: Similarly, if several acts, spanning some amount of time, were necessary ror the murder the Board may deny parole because the inmate had "opportunities Lo stop" but did not. However jfthe murder was accomplished quickly parole will be denied because it was done in a dispassionate and calculated manner and the victim never had a chance to defend themselves or flee. If the crime OCCUlTed in public, or with other people in the vicinity, it has been said that the inmate "showed a callous disregard" or "lack of respect" for the "community." However if the crime occurs when the victim is found alone it could be said that the inmate's actions were aggravated because the victim was isolated and more vulnerable. Id. at 13-14. 129. Redd v. McGrath, 343 F.3d 1077, 1082 (9th Cir. 2003) ("[Sltale habeas re,oew is the first and only opportunity the California state courts have to hear a prisoner's constitutional claims."). 130. In re Rosenkrantz, 59 P.3d 174, 217 (Cal. 2002). 2009] HEINOUS, ATROCIOUS, AND CRUEL 917 sult."13] Even a valid habeas claim cannot result in a grant of parole: California state courts may only "requir[e] the Board to conduct another parole suitability hearing."] 32 The result is that the parole board findings of fact are upheld on a standard lower than the preponderance of evidence standard at issue in the Apprendi line, even when the board finds facts the jury explicitly failed to find. In fact, a "parole authority may credit evidence suggesting the inmate committed a greater degree of the offense than his or her conviction evidences."133 Murderers sentenced to indeterminate life sentences are, incredibly, not "entitled to earlier release simply because their convictions are in the second rather than the first degree."!34 Ajury acquittal on a given grounds meant only that the jury found a reasonable doubtnot that it lacked "some evidence" of a given fact. ]35 In Dannenberg, for example, the California Supreme Court found that the facts of the offense "pelmitted an inference" by the parole board that Dannenberg drowned his wife,I36 even though "how this happened is unclear.'·!37 In Rosenkrantz, the California Supreme Court found that even if "the jury, for whatever reason, did not find beyond a reasonable doubt" that a crime was premeditated, that does not "preclude this court from detelmining that some evidence" supports a finding at the parole stage that it was premeditated. I3s In re Scott (Scott II) found the prosecuting district attorney using the parole suitability hearing as a second opportunity to prove premeditation. Scott was offered a pretrial plea for manslaughter (a crime that did not require premeditation), and was acquitted at trial of first degree murder (a crime that did require premeditation).!39 The district attorney's office nevertheless argued that Scott's crime was premeditated, explain131. In re Ramirez, ]]4 Cal. Rptr. 2d 381, 398 (CI. App. 2001), overruled on olher grounds by In re Dannenberg, 104 P.3d 783 (Cal. 2005). 132. Id.; see also In re Irons, No. 121937 A, 1 (Cal. Super. CI. Jan. 30, 2002) (order denying petition of habeas corpus) (on file with the Columbia Law Review) ("The jurisdiction of the courts in this area is limited to dctcnnining whether procedural requirements have been met. If the Board's action was not arbitrary, the separation of powers doctrine precludes review in [sic] the merits." (emphasis added», 133. Dannenberg, 104 P.3d al 803 n.15 (citing Rosenkrantz, 59 P.3d at 219). 134. 1d. 135. &senkrantz, 59 P.3d at 219. 135. Dannenberg, 104 P.3d at 802. 137. Id. at 785. 138. Rosenkrantz, 59 P.3d at 219. Thus, a state need not bother with jury proof beyond a reasonable doubt, as long as the proceeding is not called sentencing and the decision is not made by the presiding judge. The state can effectively stack the deck even further: The prisoner must contest the finding of fact, but contesting the facts is likely to weigh against his suitability for parole. In re Scott (Scali II), 34 Cal. Rptr. 3d 905. 924 (Ct. App. 2005) (explaining prisoner's effort to show his crime is not heinous "will be seen as unwillingness to accept responsibility and therefore evidence of unsuitability"). He must then challenge the finding at habeas, where he bears the burden of proof. Rosenkrantz, 59 P.3d at 217. 139. Scali II, 34 Cal. Rptr. 3d al 913. 918 COLUMBIA LAW REVIEW [Vol. 109:893 ing that his pretrial offer "had nothing at all to do with .. what we're talking about today" at the parole hearing, but only with "what we felt the evidence would be able to prove."140 In Brodsky v. Kane, Federal District Judge Charles Breyer upheld the parole denial of a defendant who pleaded guilty to second degree murdeL I41 Even though the defendant was "being detained on the basis of first-degree facts, such as the fact that his crime involved a murder-forhire and was 'carried out in a dispassionate and calculated manner,' "142 Judge Breyer nonetheless found that "the State of California has unbridled discretion in determining the standard by which to grant parole, if at all."143 D. The Result: Prisoners Serve Much Longer than Their Enumerated Term. of Years The widespread imposition of unsuitable life sentences means that few prisoners serving indeterminate sentences are ultimately released. From 1999 to 2002, the California parole board conducted approximately 12,000 suitability hearings. IH Just 140 prisoners were found suitable for release. 145 Of those 140 determinations, 127 were reversed by thenGovernor Gray Davis and eleven remanded, leaving a total of two prisoners released into parole-both of them "battered women who killed their abusers."146 Davis's predecessor released sixty-eight prisoners in eight years, and his successor, Arnold Schwarzenegger, has released 170 prisoners out of the 771 found suitable. 147 One current estimate is that three percent oflifers eligible for parole are found suitable each year and given 140. Id. (alteration in original) (internal quotation marks omitted). The board granted Scott's parole in Scott II, only to be overturned by the Governor on the sole basis that his crime was heinous, atrocious, or cruel. Id. at 914 (quoting Governor that "[t]he gravity of the murder ... alone is a sufficient basis on which to conclude that his release from plison at this time would pose an unreasonable public safety risk" (first alteration in original) (internal quotation marks omitted». ]n Scott I, however, the board had looked at the same facts-including all the facts about the plea offer and the jury's verdict-and found him unsuitable for parole on the basis of his commitment offense. See In re Scott (ScottI), 15 Cal. Rptr. 3d 32, 43-45 (Ct. App. 2004). Scott's habeas petition was granted in Scoll II, and he was subsequently released. 34 Cal. Rptr. 3d at 927. 141. No. C 06-02288 CRB, 2007 U.S. Dist. LEXIS 56401, at *1-*2 (N.D. Cal. July 24, 2007). 142. Id. at *31. 143. Id. at *32. 144. Editorial, Da,is Keeps Door Shut, S.F. Chron., Sept. 29, 2002, at D4. 145. Id. 146. Id. But the court in In re Rosenkrantz rejected the contention that then-Governor Davis had a "no parole" policy, in part because of Davis's statements that second degree murderers "should serve at least a life sentence in prison." 59 P.3d 174, 223 (Cal. 2002) (internal quotation marks omitted). By the end of his term, Davis overturned all but six parole suitability tindings for prisoners serving murder sentences. Abramsky, supra note 55. 147. Editorial, Home for Christmas, S.F. Chron., Dec. 24, 2007, at B4; see also Bob Egelko, Court Overrules Parole Veto, S.F. ehron., Nov. 28, 2007, at B2 (stating that 2009] HEINOUS, ATROCIOUS, AND CRUEL 919 release dates by the board, but only one percent survive full board and gubernatorial review on the way to release.'48 These figures cover only prisoners who have had hearings, however; the parole board currently has a substantial backlog of parole hearings and is subject to court orders resulting from a case filed by parole eligible inmates. l4" California data from 2006 show that second degree and first degree murderers served roughly the same time in prison, even though a first degree sentence is twenty"five years to life and a second degree sentence is fifteen years to life. 15o The median time served for paroled second degree murderers was twenty-three years, one month. The median time served for paroled first degree murderers was twenty-six years, three months.!S! * * * In short, prisoners serving unsuitable life sentences are initially given parole eligible sentences based on their commitment offense, but are later denied parole on the basis of that same offense. There is almost no way to challenge parole board findings of fact given the deferential standard of review, and existing cases governing due process rights in parole are incoherent and ineffective. Apprendi, however, allocates the domain of factfinding about the crime to the jury, discusses limitations on the Governor Schwarzenegger has overturned parole board determination about three quarters of the time). 148. Abramsky, supra note 55. 149. Michael Rothfeld, State Parole Board Gets a Grilling, L.A. Times, July 6, 2008, at B1 (counting backlog of ],400 parole heal;ngs). Part of the problem is high turnover in commissioners. See id. 150. See Data Analysis Unit, Cal. Dep't of Corr. & Rehab., Time Served on Prison Sentence 2 tbl.1 (2007), available at http://www.cdcr.ca.gov/Reports_Research/ Olfender_Information_Services_Branch/AnnuaI/TlME6/TIME6d2006.pdf (on file with the Columbia Law Review). 151. Id. The data are perhaps skewed by the small sample size. Only three first degree and twenty-seven second degree murderers ,,",'ere granted parole. Given the thousands of prisoners senring first and second degree tenTIS, however, it is perhaps safe to assume that prisoners yet to be released are serving even longer. AI; one California state judge noted: [l]t should be self evident that after an inmate has sCIlled the equivalent of 25 years, whether his actions were more than minimally necessary for a second degree conviction . . . is no longer the appropriate question. [The board's] position, that inmates who were only convicted of second degree may forever be denied parole based on some modicum of evidence that their acts rose to the level of a first, without acknowledging the fact that they have already served the time for a first, should be seen as so ridiculous that simply to state it is to refute it. In re Weider, 52 Cal. Rptr. 3d 147, 155-56 (Ct. App. 2006) (alteration in original) (internal quotation marks omitted). This is, ultimately, why 1 do not spend much time discussing the Apprendi ramifications of release-date calculations, even though they, too, apply commitment offense faclS to a matrix of sentence lengths. A'i a practical matter, when a prisoner is eventually found suitable for parole, he is almost always long past even the highest range of the release-date guidelines. COLUMBIA LAW REFlEW 920 [Vol. 109:893 penalty that can be imposed for a given verdict, and requires heightened standards of proof for facts increasing punishment. Might it provide a barrier to unsuitable life sentences? 11. WI-JICH BODY FINDS WHICH FACTS? I look at the unsuitable life sentence problem in three parts: limitations on which body gets to decide which facts (Part 11); limitations on the penalty that can be imposed (Part III); and limitations on the standard of proof used to justifY those penalties (Part IV). I examine each of these from two perspectives-the parole due process line of cases l52 and the Apprendi line J53-and then attempt to synthesize the conclusions of both lines. Removing commitment offense facts from the parole board's purview vindicates Appendi's key interests and preserves parole board discretion over areas in which it is the more competent and accurate factfinder. Parole due process cases traditionally grant almost complete deference to the parole board to decide any fact about the prisoner or her offense. Discretionary release determinations are seen as more art than science: Judicial oversight neither preserves rights nor improves outcomes. Apprendi, however, gives no deference to nonjury bodies when a given fact increases punishment. Any fact increasing punishment beyond the statutot]' maximum must be found by the jury alone. Reconciling these two lines of cases, then, turns on the meaning of punishment. A mechanical/formal version of Apprendi, one in which punishment merely means an increase in prison time, would present an "all or nodling" simation where the jury supplanted the parole board: The jury would have to find any fact used in a parole suitability determination, including those relating to post-trial institutional behavior. A version of Appendi grounded in the meaning of punishment, howeverwhere Appendi only applies to retributive incarceration, not rehabilitative incarceration-would mean the jury need only find commitment offense facts. This would preserve parole board discretion to determine rehabilitative factors in a way that is implicit, and sometimes explicit, in Apprendi and its predecessors. This split, in tum, maps onto the two parts of a California indeterminate sentence: The enumerated term of years governed by the jury is punitive, and the rest of the term governed by the parole board measures rehabilitation. A. Parole Discretion and Unsuitable Life Sentences The Supreme Court has given parole boards immense latitude to make suitability determinations: Parole requires only "some orderly pro152. See supra text accompanying notes 56-63. 153. See supra text accompanying notes 6-21. 2009] HEINOUS, ATROCIOUS, AND CRUEL 921 cess, however informal."!54 The essence of the Supreme Court's approach is that the inherent subjectivity of parole determinations means that judicial second-guessing yields no benefit. This section explores the ways in which parole due process does not constrain the parole board's authority to find commitment offense facts, enabling it to impose unsuitable life sentences. 1. Parole Boards Are Given Authority to Determine a Wide Range of Facts. - There are two parts to any parole suitability determination: factfinding and judgment. In the first stage, "[ t] he function of legal process ... in the realm of factfinding, is to minimize the risk of erroneous decisions."!55 When it comes to the parole board's judgment, however, due process does not require the parole board to "specify the particular 'evidence' ... on which it rests the discretionary determination that an inmate is not ready for conditional release."!56 This is because suitability decisions involve "a predictive judgment" about "what is best both for the individual inmate and for the community,"!5? an "equity-type judgment" synthesizing facts, personal observations, and the experience of the decisionmaker.'58 The parole board's "decision is much like a sentencing judge's choice ... to grant or deny probation following a judgment of guilt ...."159 In this view, then, there is no constitutional floor for parole-the board can consider whatever it wants. The board's decisionmaking is cabined only by parole statutes and regulations, which, in California, encourage the board to consider the commitment offense. 2. Parole Board Factfinding and Unsuitahle Life Sentences. - The parole line of cases poses only a minimal barrier to unsuitable life sentences. As long as a parole board follows the law and provides "some evidence" supporting its decision, courts may not intervene. In states like California, 154. Morrissey v. Brewer, 408 U.s. 471, 482 (1972). Morrissey concerns parole revocation, but, if anything, the due process requirements are even more minimal for prospective parole release, See Greenholtz v. Inmates of Neb. Penal & COlT. Complex, 442 U.S. 1, 9 (1979) ("There is a crucial distinction between being deprived of a liberty one has, as in parole, and being denied a conditional liberty that onc desires,"). I will therefore cite Mumney for the general principles involved in parole administration, at least insofar as they are comistcnt with the two leading cases about parole, Greenhollz and Board of Pardons v. Allen. 482 U.S. 369 (1987). The method used for establishing the liberty interest in parole is quite different among these cases, however, and will be discussed in greater detail infra Part lILA. I. 155. Gmmhollz. 442 U.S. at 13; see also Morrissey, 408 U.S. at 479-80 (explaining that first step in parole revocation is a "wholly retrospective factual question," where accuracy is at a premium. and the second step is more "predictive and discretionary"). Note that the separation of factfinding from disposition bears a striking resemblance to the Apprendi line's separation of the jury's power to find facts from the judge's power to make dispositional decisions based on those facts. 156. Greenhullz, 442 U.S. at ]5. However, the board could "communicate the reason for its denial as a guide to the inmate for his future behavior." ld. 157. Id. at 8. 158. Allen. 482 U.S. at 374-75 (internal citations omilted). 159. Greenholtz, 442 U.S. at 16. But see infra Part Il.B.l.b for a discussion of how this discretion might itself depend on rehabilitative concerns. 922 COLUMBIA LAW REVIEW [Vol. 109:893 with parole statutes that encourage the board to consider the commitment offense in determining suitability, any potential due process barriers all but disappear.] 60 While parole cases permit unsuitable life sentences, courts are nevertheless grappling with the sense that there is something wrong with them. Courts have used due process reasoning to overturn commitment offense parole denials at least twenty-eight times since late 2005. 161 In the main, these cases have held that the evidence supporting denial must not be merely "that a particular factor or factors indicating unsuitability exist, but that a prisoner's release will unreasonably endanger public safety."162 "It violates a prisoner's right to due process when the Board or Governor attaches significance to evidence that forewarns no danger to the public."163 The California Supreme Court recently considered "the question of the extent that the [Parole] Board and the Governor should consider an inmate's current dangerousness in making a parole suitability determination, and at what point, if ever, the gravity of the commitment offense and prior criminality are insufficient to deny parole when an inmate otherwise appears rehabilitated."164 In re Lawrence held that a commitment offense must show not only "some evidence" of a given statutory factor (such as "heinousness"), but also some evidence of the ultimate conclusion-that the inmate is presently dangerous. 165 The dissent argued, 160. For a discussion of California's parole statute, see supra notcs 43-50 and accompanying text. Practically speaking, the consideration of the commitment offense makes a parole board finding of suitability almost impossible. What difference does it make that a prisoner has done well in, say, anger management class, when that is measured against his taking another person's life? No CPA in a prisoner's GED cla'is is going to be high enough to compensate for her killing someone. The facts in mitigation thus have almost no chance to tip the balance towards release. 161. Michael Rothfeld, Is This Paroled Killer Still a 111reat?, L.A. Times,July J 3,2008, at AI. One inmate attorney characterizes this number as "an extremely high reversal rate," adding "[i]t was so completely unfair, the courts finally had to do something. The governor can basically resentence these inmates to life without possibility of parole." Id. (internal quotation marks omitted). 162. Hayward v. Marshall, 512 F.3d 536, 543 (9th Cir. 2008). 163. In re Tripp, 58 Cal. Rptr. 3d 64, 68 (Ct. App. 2007). 164. In re Viray, 75 Cal. Rptr. 3d 190, 195 (Ct. App. 2008) (citing seven cases then pending review before California Supreme Court). 165. "In some cases" where evidence of rehabilitation is overwhelming and the only evidence of unsuitability is the commitment offense, the commitment offense "does not provide 'some evidence' inevitably supporting the ultimate decision that the inmate remains a threat to public safety." In re Lawrence, 190 P,3d 535, 539 (Cal. 2008), The court concluded: [A] lthough the Board and the Governor may rely upon the aggravated circumstances of the commitment offense as a basis for a decision denying parole, the aggravated nature of lhe crime does not in and of itself provide some evidence of current dangerousness to the public unless the record also . . . indicates that the implications regarding the prisoner's dangerousness that derive from his or her commission of the commitment offense remain probative to the statutory determination of a continuing threat to public safety. 2009] HEINOUS, ATROCIOUS, AND CRUEL 923 with some reason, that the court was substituting its own judgment for the parole board's, and that there were significant separation of powers issues at stake. 166 Cases overturning unsuitability findings do not dispute whether a parole board can use the commitment offense to justiry parole denial, however, only whether this justification comports with the language of the parole statute. The jurisprudential tools available in the Apprendi line provide a more cogent and direct means of reining in unsuitable life sentences. B. Apprendi '5 Jury Requirement In this section, I explain the rationale behind Apprendi's jury requirement by looking at the institutional competencies of the judge and jury. This analysis also helps to explain the contours of Apprendi's rule-for example, why a judge may find proof of prior offenses to increase an offender's sentence. Using this institutional competence analysis, I then map Apprendi's rule onto indeterminate sentencing, concluding that the parole board may not find facts related to the commitment offense when making its suitability determination: The commitment offense is largely a matter of retribution, and any public safety implications are better handled by the judge. 1. The Jury Finds Retributive Facts, the Judge Finds Public Safety Ones. In this subsection, I explain Apprendi's division of labor between the judge and the jury by analyzing the rehabilitative and retributive components of confinement. Apprendi isolates two necessary dimensions of punishment: the stigma of conviction, which comes from the wrongfulness of the conduct, and the loss of liberty, which is the consequence of the transgressive behavior. While punishment is necessarily stigmatic, confinement need not be: An offender can be confined in order to incapacitate and rehabilitate him. Thus, the reasons why confinement is imposed-that is, whether or not confinement is punitive-helps determine which facts a jury must find. In Part n.B.l.a, I explain that juries must find facts leading to punishment because they serve as the moral representatives of the community. Retribution is not a legal issue or a criminological issue, but a moral one best left to the conscience of the community.16? In Part n.B.l.b, I Id. at 555. 166. Id. at 566 (Chin, j., dissenting) ("[I]f a factor is properly part of the [parole suitability] evaluation ... and [its] existence ... is supported by some evidence, ... the electorate entrusted [the ultimate conclusion regarding parole suitability] to the Governor's discretion, not the courts . . . ." (citation and internal quotation marks omitted» . 167. When considering the division of labor between judge and jury, one need not consider the diflerence between detenninate and indeterminate sentences. These sentences differ only at the point of release-not when ajudge pronounces the sentenceand in neither ca.se is the release decision made by either the judge or the jury. In a 924 COLUMBIA LAW REVIEW [Vol. 109:893 explain why public safety evaluations require competence of a different sort. Nonjury repeat players, such as parole boards and judges, are more likely to know which consequences applied to which criminals are most effective in neutralizing their threat to public safety. Public safety evaluations are not punitive, so juries need not be involved. I use this understanding of institutional competencies to explain the exceptions to Apprendi's jury factfinding requirement in Part n.B.I.e. a. TheJury as Moral Agent. - Understanding Justice Stevens's view of juries is the key to understanding his majority opinion in Apprendi. Punishment involves not just the deprivation ofliberty but the imposition of moral stigma, and the jury, the conscience of the community, is uniquely suited to make moral judgments. Stevens's views were developed in a series of death penalty cases where the defendant argued that judicial factfindingjustifying a death sentence violated the Sixth Amendment.!68 (Some of these cases also explicitly examined the "heinous, atrocious, or cruel" standard used in the death penalty statutes at issue.!6g) Often writing in dissent,Justice Stevens argued that the moral questions underlying punishment meant that the jury had to find facts. In Spaziano v. Florida, for example, the Court held that a judge'S imposition of the death penalty over a jury recommendation of life imprisonment does not violate the Sixth Amendment, noting that capital sentencing is not "like a trial in respects significant to the Sixth Amendment's guarantee of a jury trial," because it involves questions of punishment, not questions of guilt or innocence 170 Dissenting, Justice Stevens argued that juries alone should make the decision about death, because they are "best able to express tile conscience of the community on the ultimate question oflife or death."!7! Because capital punishment is "an expression of society's moral outrage,"!72 the "life-or-death decision ... depends upon its link to community values for its moral and constitutional legitimacy," just as the jury trial right at the guilt phase ties the deprivation of liberty to community values.l7 3 "UJ urics more accudetcr-minate system, there is no release decision to be made; in an indeterminate system, the release decision is made by the parole board. For a more detailed discussion of determinate and indctcnninate sentences, sec supra Part LA. 168. Rory Little has argued that the "triumphant architect" in the Apprendi line is not justice Scalia, but Justice Stevens, based on how his view of the jUl)"s role developed in these early cases. Weisberg, supra note 72, at 630. 169. See Walton v. Arizona, 497 U.s. 639, 646 (1990), overmJcd by Ring v. Arizona. 536 U.S. 584 (2002) (upholding death sentence given out under "heinous, atrocious, or depraved" statute); Clemons v. Mississippi, 494 U.S. 738, 742 (1990) (announcing that death sentence may be imposed on grounds that murder was committed for pecuniary gain and was "heinous, atrocious, or cruel"); Maynard v. Carhvright 486 U.S. 356, 359-60 (1988) (holding Oklahoma's "heinous, atrocious, or cruel" aggravating death penalty factor unconstitutionally vague as applied). 170. 468 U.S. 447, 458-59 (1984). 171. Id. at 470 (Stevens,]., dissenting) (internal quotation marks omitted). 172. Id. at 480 (internal quotation marks omitted). 173. Id. at 483. 2009] HEINOUS, ATROCIOUS, AND CRUEL 925 rately reflect the conscience of the community than can a single judger, which] is the central reason that the jury right has been recognized at the guilt stage in our jurisprudence."!74 Although she does not examine Spaziano and other Apprendi precedents, Laura Appleman has also located the Apprendi line's theory of sentencing in a limited "expressive retributive theory of punishment for sentencing."!75 Part of the purpose of punishment is to "reinforce belief' in the community's values.!76 "[W]hen the jury determines sentencing facts ... [an offender's] community and his peers have pronounced his blameworthiness ...."!77 The jury expresses "the community's condemnation of the act."!78 With the jury's moral role as background, it is easier to understand Apprendi's definition of punishment. That is, while Apprendi spoke generally of whether the fact at issue would increase the penalty!79 or punishment!SO [or a crime, it did not equate penalties and punishments with prison sentences. Punishment involves something more: not just the loss of liberty, but also the stigma of the offense.!S! Different punishments attach to different degrees of "criminal culpability."!82 174. Id. at 487. Justice Stevens concluded: Juries-comprised as they are of a fair cross section of the community-are more representative institutions than is the judiciary; they reflect more accurately the composition and experiences of the community as a whole, and inevitably make decisions based on community values more reliably, than can that segment of the community that is selected for senrice on the bench. Id. at 486-87 (internal citation omitted). 175. Appleman, supra note 3, at 1326. Appleman's theory owes a great deal [0 Jean Hampton's work, which is based on the idea that there are "two [onns of damage to value effected by wrongful actions": what the action does and what the action means. Jean Hampton, Correcting Harms Versus Righting Wrongs: The Goal of Retribution, 39 UCLA L. Rev. 1659, 1672, 1680 (1992) (noting that while a woman stealing a book from a library sends a message that she considers herself superior to others, "her actions actually make it the case" that she gets superior treatment). 176. Hampton, supra note 175, at 1678. This approach is not without its limitations. In discussing Durkheim's very similar dramaturgical theory of the law (that punishment is "directed primarily at the community and not at the offender" and works "indirectly through affirmation and reiteration of basic norms"), Michael Tonry observes that "[c]riminaI law and punishment ... are not the primary means of socialization into right values. That function belongs to primary social organizations such as the family, the church, the workplace, the community, and to kinship and friendship networks." Michael Tonry, Thinking About Crime: Sense and Sensibility in American Penal Culture 99 (2004). 177. Appleman, supra note 3, at 1332. 178. Id. at 1338. The judge is a less effective instrument of punishment because "the offender may very well attribute his punishment to the State and shrug off the desired feelings of responsibility or awareness of his wrongdoing.'" Id. at 1332. 179. E.g., Apprendi v. New Jersey, 530 U.S. 466, 472, 473. 476, 482, 483. 486, 490, 495, 497 (2000). 180. E.g., id. at 476, 478, 480-88, 491, 495. 181. ld. at 484. 182. Id. at 485. Justice Stevens noted that "[t]he degree of criminal culpability the legislature chooses to associate with particular, factually distinct conduct has significant 926 COLUMBIA LAW REVIEW [Vol. 109:893 If a defendant faces punishment beyond that provided by statute when an offense is committed under certain circumstances but not others, it is obvious that both the loss of liberty and the stigma attaching to the offense are heightened; it necessarily follows that the defendant should not-at the moment the State is put to proof of those circumstances-be deprived of protections that have, until that point, unquestionably attached. 's3 Stigma is important in its own right, even when criminal penalties are not involved. In re Winship, a noncriminal, indeterminate sentencing case from 1970 184 that lies at the heart of Apprendi's holding about the burden of proof,ls5 held that the adjudication of a juvenile as a delinquent-which generated the stigma of having done something wrong'8fi -could be punitive even though the result of this decision was that Winship was sent to reform school. 18 ? b, TheJudge as Rehabilitation Expert. -Judicial discretion to find facts at sentencing is unrelated to moral or retributive judgments. A judge's factfinding power derives instead from her role as an experienced and astute analyst of what sentence might best preserve the public safety and rehabilitate an individual offender. This discretion is rooted historically in an era where both indeterminate sentences and the goals of rehabilitation were dominant. With the resurgence of determinate sentences and punitive goals, the rules about judicial discretion have parted ways with their underlying rationales. In the middle of the twentieth century, when judicial discretion was at its peak, crime was seen as a "moral disease," and the judge was an expert who prescribed treatment for the offender. 188 Judges got broad factfinding discretion in order to "tailor dispositions to dle treatment needs of individual offenders and the public safety risks they posed."189 implications both for a defendant's vcry liberty, and for the heightened stigma associated with an offense the legislature has selected as worthy of greater punishment." ld. at 495. 183. Id. at 484. Moreover, both the "absolute years behind bars ... and ... the more severe stigma attached" are constitutionally significant. Id. at 495. "Prosecution subjects the criminal defendant both to the possibility that he may lose his liberty upon conviction and . .. the certainty that he would be stigmatized by the conviction." Id. at 484 (emphasis added) (omission in original) (internal quotation marks omitted) (citing In re Winship, 397 U.S. 358, 363 (970)). Justice Stevens, in his dissent in McMillan v. Pennsylvania, offered this variation on the Apprendi/Winship rule: "[T]he Due Process Clause requires ... [the State] to prove any component of the prohibited transaction that gives rise to both a special stigma and a special punishment beyond a reasonable doubt." 477 U.S. 79, 95 (1985) (Stevens,].. dissenting). 184. Winship. 397 U.S. at 358-59. Winship, adjudged a juvenile delinquent, faced an indeterminate sentence, subject to annual renewal. Id. at 360. 185. See infra Part IV. ]86. Winship, 397 U.S. at 366-67 (discussing the stigmatic harms of delinquency adjudication) . 187. Id. at 359. 188. Nancy Gertner, What Has Hams Wrought, 15 Fed. Sent'g Rep. 83, 84 (2002-2003). 189. Tomy, Reconsidering, supra note 31, at 3. 2009] HEINOUS, ATROCIOUS, AND CRUEL 927 In this judicial environment, "[n]o one challenged judges' sentencing procedures as somehow undermining the Sixth Amendment's right to a jury trial precisely because the judge and jury had 'specialized roles.' "190 Williams v. New l'ork1 91 presents the classic example of how the therapeutic model of criminal sanctions was essential to unfettered judicial factfinding. Williams was convicted of first degree murder. 192 Although the jury recommended a sentence of life in prison, the judge sentenced Williams to death, based in part on reports of thirty burglaries for which Williams had not been convicted.!9' In upholding Williams's death sentence, the Supreme Court focused on the statutory discretion given New York judges,194 explaining that in the "prevalent modern philosophy of penology ... the punishment should fit the offender and not merely the crime."195 In order to treat the offender, the judge needed unfettered access to facts-including uncharged facts-that would permit him to make his diagnosis. 196 Beginning in the mid-1970s, "many of the rationales and practices of indeterminate sentencing began to be challenged"197 as determinate, re190. Gertner, supra note 188, at 8,1 ("However flawed a judge's decision might be, it was not the case that he or she was usurping the jury's role."). 191. 337 U.S. 241 (1949). 192. Id. at 242. 193. Id. at 242-44. Williams confessed to some of the burglaries, but evidence of others came from witnesses Williams did not have the opportunity to cross-examine. Id. The judge also referred to the probation report's indication that Williams possessed "a morbid sexuality." Id. at 244. 194. Id. at 245. 195. Id. at 247 (internal citations omitted). 196. See Berman, Conceptualizing Blakely, supra notc 2, at 94. As Bennan explains: [S]ince sentencing was long conceived-at least form all)', if not in actuality-as an enterprise designed to help "cure" the sick defendant, the idea of significant procedural rights at sentencing almost did not make sense. Just as patients are nol thought to need "procedural rights" when being treated by a doctor, defendants were not thought to need procedural rights when being sentenced by a judge. But, of course, it has been nearly a quarter century since the rehabilitative model of sentencing has held sway, and yet until Apprendi and Blakely came along, our sentencing structures still relied without much question on Ia..x procedures for proving the truth of facts that could lead to extended sentences. Id. The type of discretion in Williams was closely related to, if not determined by, the purpose of sentencing. "Indeterminate sentences" that involved discretionary release had "to a large extent taken the place of the old rigidly fixed punishments," indicating that "[r]etribution is no longer the dominant objective of the criminal Jaw. Reformation and rehabilitation of offenders have become important goals of criminal jurisprudence." Williams, 337 U.S. at 248. Of course, it seems strange to talk about rehabilitation as a goal in a case where the judge sentenced the defendant to a sentence without any possibility of reform-death-but this could be seen as a diagnosis that the offender was a hopeless case. 197. Tonry, Reconsidering, supra note 31, at 5. 928 COLUMBIA LAW REVIEW [Vol. 109:893 tributive sentencing systems became increasingly popular. 19S States' decisions to implement binding, determinate sentencing schemes "appropriately raised Sixth Amendment concerns."!99 The Supreme Court's sentencing doctrine, however, has not always acknowledged this shift from rehabilitation to retribution or its implications on precedent. 200 Justices freely mix and match cases from either side of the rehabilitation/ retribution line, drawing conclusions about judicial power unmoored from their rationales. This failure to address the goals of determinate and indeterminate regimes has resulted in an "inherently schizophrenic"20! body of sentencing law. For example, Williams has survived to some extent, even though its foundation-the consensus that reformation and rehabilitation are the goals of sentencing-has not. While the post-Apprendi ca~e Ring v. Arizona effectively reversed Williams to require jury findings of aggravating factors justifying capital punishment,202 Williams's view ofjudicial discretion in subcapital sentencing has remained untouched. In an advisory system, a judge may consider any fact, including uncharged or acquitted conduct. 2ll3 Supreme Court Justices have also ignored Williams's grounding of judicial discretion in rehabilitation. Justice Scalia, for example, distinguished Blakely from Williams solely on the basis of whether the sentencing system was binding or advisory.204 But the source of the discre- 198. See supra text accompanying notes 83-84. Although Williams had announced confidently that the theory of punishment underlying detenninate, retributive sentencing "no longer prevails,'" Williams, 337 U.S. at 247, the theory became the focal point of reformers who focused on, inter alia, systemic racial bias, ineffective treatment, and undue leniency in indeterminate systems. Tonry, Reconsidering, supra nole 31, at 5. 199. Gertner, supra note 188, at 84. Indeed, showing that there is nothing new under the sun, Williams contained a strong dissent by Justice Murphy in which he set forth a retributive theory of punishmentjustifying jury factfinding, Murphy identified the jury ';as the representative of the community" whose "voice is that of the society against which the crime was committed." Williams, 337 U,S, at 253 (Murphy, j., dissenting). Although a judge might have "statutory authority" to override the jury, he "should hesitate indeed to increase the severity of such a community expression," Id. 200. See Douglas A. Berman & Stephanos Bibas, Making Sentencing Sensible, 4 Ohio St. J. erim. L. 37, 48-49 (2006) (arguing Supreme Court declined to recast sentencing doctrine to reflect modern reforms), 201. Reitz, supra notc 3, at 1096. 202. See 536 U.S. 584, 609 (2002). 203. Reitz, supra note 3, at 1095 (discussing considerations in Williams of "thirty burglaries that were never charged nor proven, plus additional uncharged offenses"); see also id. at 1083 n,3 (citing Supreme Court holding that "acquittal on gun possession charge does not preclude consideration ... at sentencing"). 204. Blakely v. Washington, 542 U.S. 296, 305 (2004) ("Williams involved an indeterminate-sentencing [that is, advisory] regime that allowed a judge (but did not compel him) to rely on facts outside the trial record in determining whether to sentence a defendant to death."). For definitions of the terms I use in this Article, see supra text accompanying notes 67-68. 2009] HEINOUS, ATROCIOUS, AND CRUEL 929 tion given the judge in Williams was tied to his rehabilitative expertisehe was not given unfettered discretion to punish. 205 c. Could Retribution Explain Apprendi's Exceptions? - While the Apprendi line has not explicitly grounded its rule in a theory of punishment, mapping the jury power to retribution helps explain why the jury must find certain facts and why a judge can find others. This approach might also provide a way to resolve some of the anomalies in the Apprendi line, such as why Apprendi does not govern advisory sentencing regimes or a judge's finding of prior convictions, even when such findings trigger more severe sentences. 206 One of Apprendi's enduring mysteries is why, if jury factfinding is so important, that right is only implicated in binding, presumptive systems. 207 One way of explaining this phenomenon is to read binding, determinate sentencing schemes as clear legislative statements that a sentence serves only retributive goals. A "one size fits all" sentence punishes the offense without regard to the threat an individual offender poses to public safety. Because binding schemes punish the crime, not the offender, a jury is in its bailiwick when it finds facts justifying punishment. An advisory system, in contrast, can be read to implicitly endorse public safety considerations. The judge is permitted to find facts because, as in Williams, she must tailor her sentence to an individual offender's history and risk of recidivism. Any deviation from a normal sentence is not necessarily a statement about punishment, but about the greater (or lesser) need for rehabilitation and incapacitation. A jury is less suited to make this judgment than a judge, and therefore need not find all facts bearing on this judgment. 208 205. The Supreme Court has not addressed whether l1'"illiams's discretion is a necessary function of rehabilitative goals, however. An advisol)', indeterminate system could conceivably promote retribution by giving a judge more discretion [0 punish, but this is not the reasoning behind Williams, nor is it typically the policy justification for establishing such systems. In fact, proponents of determinate regimes often point to indetemlinate systems' perceived "softness" on crime. Sec generally Tyler & Boeckmann, supra notc 90, at 250 (finding support for "three strikes" rules stems from "general support for harsher punishment of rule breakers"). 206. My claim here is not that the Justices conceived of their reasons this way, but that their actions can be explained this way. 207. See Reitz, supra note 3, at 1101 ("[T]he Court's solicitude to juries as privileged decisionmakers is absent at the extremes, and engages only the middle ground of presumptive rules."). 208. Conversely, the departure could indicate a judge's taste for punishmcnt, as noted supra note 205 and accompanying text, or policy disagreements with, for example, the crack/powder cocaine sentencing disparity. See Kimbrough v. United States, 128 S. Ct. 558, 570 (2007) (noting that courts may vary from Guidelines ranges on policy considerations alone). 930 COLUMBIA LAW REVIEW [Vol. 109:893 The retribution/public safety division might also explain why the Apprendi line caIVes out an exception for prior convictions. 209 Even though prior convictions are facts that can operate to increase punishment by adding prison time, Apprendi does not require a jury to find them. The retribution/rehabilitation model could explain this puzzling omission in one of two ways.210 Since repeat offenders are more likely to commit future crimes, prior convictions might be seen as public safety factors about the offender, not the offense. 211 Alternatively, because the prior convictions themselves have already comported with due process, their use in punishment does not violate due process. 2'2 The retribution/rehabilitation model also maps onto Douglas Berman's distinction between facts about the offense and facts about the offendcr. 213 Justice Kennedy's dissent in Cunningham v. California en209. For example, Almendarez-Torres v. United States, 523 U.S. 224, 226-27 (1998), which pennitted a judge to find prior offenses and use them to impose a higher sentence in a binding sentencing scheme, survived the Apprendi decision. 210. The Apprendi court saw no need to explain the survival of Almendarez-Torres, however. See Apprendi v. New Jersey, 530 U.S. 466, 489-90 (2000) ("Even though it is arguable that Almendarez-T017es was incorrectly decided .. , Apprendi does not contest the decision's validity and we need not revisit it for purposes of our decision today . . . ." (internal footnote omitted». 211. Apprendi distinguished Almendarez-Torres along offense/offender lines: "Whereas recidivism 'does not relate to the commission of the offense' itself," the statute in Apprendi went "precisely to what happened in the 'commission of the offense.'" Id. at 496 (quoting Almendarez-Torres, 523 U.S. at 244). Almendarez-Torres rejected this interpretation, however, stating that recidivism "goes to the punishment." 523 U.S. at 244 (internal quoLation marks omitted); see also United States v. Rodriquez, 128 S. Ct. 1783, 1789 (2008) (explaining that recidivism bears on both seriousness of offense and public safety). 212. Almendarez-Torres admitted his priors "pursuant to proceedings with substantial procedural safeguards." Apprendi, 530 U.S. at 488. Ultimately, "[b]oth the certainty that procedural safeguards attached to any 'fact' of prior conviction, and the reality that Almendarez-Torres did not challenge the accuracy of that 'fact' in his case, mitigated the due process and Sixth Amendment concerns." Id. The increased penalty AlmendarczTorres faced depended only on the fact of his conviction-not the facts implicit or explicit in that conviction. That is, the way in which Almendarez-Torres committed his offenses did not matter; in A.pprendi, however, the defendant pleaded guilty to two counts of second degree possession of a fireann for an unlawful purpose but not to the racial animus under which he was eventually sentenced. Id. at 470-71. Exonerated and uncharged behavior does, however, still operate to increase sentences, and while this practice is outside the scope of my Article, a public safety/ punishment (or offense/offender) distinction between Williams and Apprnuti might help explain the different treatment. For more on the use of uncharged or acquitted conduct in the federal sys.tem, see generally Amy Baron-Evans & Jennifer Niles Coffin, Nat'l Fed. Defender Sentencing Res. Council, Deconstructing the Relevant Conduct Guideline: Challenging the Use of Uncharged and Acquitted Offenses in Sentencing (2008), available at http://www.fd.org/pdUib/relevant%20conduct2.pdf (on file \vith the Columbia Law Review). 213. Douglas A. Bennan, Distinguishing Offense Conduct and Offender Characteristics in Modern Sentencing Reforms, 58 Stan. L. Rev. 277, 277-80 (2005) [hereinafter Bennan, Distinguishing] (providing background about development of offense/offender distinction in sentencing), 2009] HEINOUS, ATROCIOUS, AND CRUEL 931 dorsed Berman's distinction, arguing that courts should "distinguish between sentencing enhancements based on the nature of the offense, where the Apprendi principle would apply, and sentencing enhancements based on the nature of the offender, where it would not."214 The retribution/rehabilitation line could explain this modified rule as follows: Punishment for a given offense takes account of the facts involved, such as use of a weapon, quantity of drugs, or deliberate cruelty. The jury has to find these facts because it is determining how wrong the particular offense was, and therefore how much punishment should be given. In contrast, a judge finds facts about the offender, such as prior record or other threats to public safety, because she is deciding how long to treat (or quarantine) a particular individual. 215 The Supreme Court had an opportunity to distinguish these two kinds of facts in Cunningham. Cunningham was a sexual predator sentenced to serve an aggravated, upper-level prison term on the judge's finding of certain facts. Among the aggravating facts the judge found in that case were facts "relating to the crime" and facts "relating to the defendant" (more specifically, "violent conduct indicating a serious danger to society").216 The majority opinion did not distinguish between the two kinds of facts, however, and instead applied Apprendi mechanically to require the jury to find all facts in aggravation irrespective of what kind they were. 217 The limitations to the Cunningham approach are laid bare when one considers an Apprendized parole system. If all facts increasing punishment must be found by a jury irrespective of their relationship to retribution, then the jury would completely supplant the parole board. A jury would have to find any fact justitying unsuitability, including those related to public safety-an area where the parole board has a clear advantage in expertise. Grounding APPrendi in a theory of retribution serves as a kind 214. 549 U.S. 270, 295 (2007) (Kennedy,]., dissenting). 215. Bennan, Distinguishing, supra notc 213, at 287 ("When the law ties punishment consequences to aspects of a person's past and character, .. the state is not defining what conduct it believes merits criminal sanction, but rather is instructing judges how to view and assess an offender's personal history at sentencing."). 216. Cunningham, 549 U.S. at 276 n.l (internal quotation marks omitted), 217. Id. at 294. This holding support'; the Court's view in Apprendi and Blakely that the jury is a "bulwark" against "oppression and tyranny on the part of rulers," Apprendi, 530 U.S. at 477, and a "fundamental reservation of power in our constitutional structure:' Blakely v. Washington, 542 U.S. 296, 306 (2004) (':Just as suffrage ensures the people's ultimate control in the legislative and executive branches, jury trial is meant to ensure their control in the judiciary."). I should note that the bulwark theory lacks detail, however. If the jury is the bulwark, it is unclear why it should apply only to binding systems, or why it does not apply to civil commitments. The purpose of confinement does nothing to change its effect on a defendant's liberty, only the meaning of that deprivation. Cunningham might ultimately reflect a skepticism about legislative labels, or an acknowledgement of the difficulty in drawing lines between commitment offense facts that are retributive and those that relate to public safety. For a more in-depth discussion of both points, see infra Parts II.C.l, II.C.2. 932 COLUMBIA LAW REVIEW [Vol. 109:893 of compromise. It stands between the present system, where the parole board can encroach on the jury right at will, and an alternative where the parole board has no power. It preserves the institutional competency of both the jury and the parole board. 2. Apprendi Factfinding and Unsuitable Life Sentences. - The Sixth Amendment "by its terms is not a limitation on judicial power, but a reservation of jury power."2IB This power is infringed any time the jury is supplanted-whether the judge, the parole board, or any other body is doing the supplanting. 219 As a precursor to Apprendi explained, the Sixth Amendment dichotomy is between jury and nonjury,220 not jury and judge. Thus, the parole board is equally forbidden to do what ajudge is forbidden to do. In this Part, I first dispense with the criticism that the Apprendi protections end once a judge has pronounced the sentence. I then examine unsuitable life sentences against the backdrop of the jury's role as the retributive body. a. Textual Arguments Limiting the Sixth Amendment to a Procedure. Steven Chanenson (among others) has argued that "structured parole release" is "outside the ambit of the Sixth Amendment. "221 The text of the Sixth Amendment refers to the accused and to "criminal prosecutions"; parole decisions involve convicted prisoners whose prosecutions have long since concluded. 222 218. Blnkely. 542 U.S. al 308. 219. The Court has argued that, historically, the linkage of facts La punishment and "[t]he defendant's ability to predict with certainty the judgment from the face of the felony" underlay the Sixth Amendment. See A.pprendi, 530 U.S. at 477-83. This \~ew has been contested vigorously. See, c.g., Blake~)', 542 U.S. at 329 (Breyer, j., dissenting) (accepting "that, classically speaking. the difference between a traditional sentencing factor and an clement of a greater offense often comes down to a legislative choice about which label to affix:' but declining to conclude from this "that the Sixth Amendment always requires identical treatment of the two scenarios"'); APlffendi, 530 U.S, at 525-29 (O'Connor,]., dissenting) ("None ofthe history contained in the Court's opinion requires the rule it ultimately adopt'i."); Weisberg, supra note 72, at 629 (quoting Rory Little's assertion that "[ilt's undeniable that in Apprendi the Court was wrong about its history"). The Court in Apprendi also argues that the individual's right to ajury is a procedural protection to make society apportion punishment ex ante, through the legislature. See Apprendi, 530 U.S. at 480 ("Where a statute annexes a higher degree of punishment to a , , . felony, if committed under particulal- circumstances, an indictment for the offence ... must expressly charge it to have been committed under those circumstances, and must state the circumstances with certainty and precision." (internal quotation marks omitted». 220. Jones v. Uni'ed States. 526 U.S. 227, 243-44 (1999). The Cour' wrote: If a potential penalty might rise from 15 years to life on a nonjury determination, the jury's role would correspondingly shrink from the significance usually carried by detenninations of guilt to the relative importance of low-level gate keeping; in some cases, a jury finding of fact necessary for a maximum IS-year sentence would merely open the door to a judicial finding sufficient for life imprisonment. 1d. 221. Chanenson, Next Era, supra note 67, at 448. 222. Id. ("[T]he Sixth Amendment applies solely to 'criminal prosecutions' and provides rights only to 'the accused.' No such event and no such person are involved in 2009] HEINOUS, ATROCIOUS, AND CRUEL 933 The Apprendi right clearly attaches during ajudge's pronouncement of the sentence, however, and at this stage the jury has already convicted the offender and has been dismissed. If Chanenson were correct, then no sentencing procedure would come under Sixth Amendment protection. Moreover, an offender has often surrendered her liberty pending sentencing, and in some instances (depending on the state statute and the judge's discretion) might even receive sentence credit for the jail time she serves before the verdict or between the verdict and sentencing. Whether or not she was convicted before or after constructively beginning her sentence cannot possibly be dispositive of whether her Sixth Amendment rights apply.223 Chanenson's objection raises a larger question, however: Is Apprendi a "sentencing case" at all?224 Calling Apprendi a sentencing case reifies the idea that sentencing is somehow an entirely separate event from criminal conviction (and, a fortiori, that parole is further removed still). This view enables us to compartmentalize Apprendi's overarching concerns into ones dealing with discrete processes. 225 But saying that Apprendi applies at sentencing and not, say, at guilt, begs the questions "guilty of what?" and "sentencing for what?"226 Apprendi's central concerns are both more nuanced and expansive than that, requiring us to ask how the determinations at one temporal stage affect the outcomes in another. Apprendi, ultimately, tells us how facts must be tied to punishment. VVhen something is detennined matters less than what is detennined, who decides, how sure they are, and the punishment that results. Viewed this way, the unsuitable life sentence problem is not, in fact, parole board the parole release process. The criminal prosecution is over. The accused is now a convicted prisoner." (quoting U.S. Const. amend. Vl». 223. Chanenson's citations in support of his position come from cases which predate Apprendi and one district court case from Utah. See, e.g., id. at 449 (citing POI. Bd. ofPrab. & Parole v. Scott, 524 U.S. 357. 354 (1998); Ohio Adult Pamle Auth. v. Woodard, 523 U.S. 272, 284 (1998); Canz v. Bensinger, 480 F.2d 88, 89 (7th Gir. 1973); United States v. Gmxford, 324 F. Supp. 2d 1255, 1254 (D. Utah 2004». Ghanenson's opinion also depends on the view that parole release is "early" and therefore can be analyzed as a downward departure from the maximum sentence oflife. He specifically discusses a particular policy proposal of his, Indetenninate Structured Sentencing, in which parole release is an act of mitigation from a judicially imposed maximum sentence. See id. at 447. 224. I am indebted Lo Jeffrey Fisher for this insight. 225. Arguably, this is one way to interpret the Supreme Court's recent decision, Oregon v. lee, which held that a judge may find fact':i which justify the imposition of consecutive sentences. No. 07-901, 2009 U.S. LEX1S 582, at *8 (U.S. Jan. 14, 2009). Because the jury historically has "played no role in the decision to impose sentences consecutively or concurrently," it was held not to implicate "the core concerns that prompted . .. A.pprendi." ld. at *15-*16. But perhaps the case was simpler than the Court let on-the issue of consecutive sentences had been raised and dismissed in Apprendi itself. See infra text accompanying note 319. 226. "The decisive movement in the conjuring tlick has been made, and it was the very one that we thought quite innocent." Ludwig Wittgcnstein, Philosophical Investigations § 308 (G.E.M. Anscombe trans., Blackwell Publ'g 3d ed. 2001) (1953). 934 COLUMBIA LAW REVIEW [Vol. 109:893 resentencing so much as the parole board retrying an offender for an aggravated, parole ineligible offense fifteen or twenty years later. b. Unsuitable Life Sentences Explicitly Contradict the Jury s Findings. Unsuitable life sentences occupy the heartland of Apprendi's retributive concerns. Just as the judge in Apprendi made findings about the racial animus underlying Charles Apprendi's act, the parole board finds commitment offense facts about the heinous, atrocious, or cruel nature of a murder. These facts are not implicit in the jury's verdict: 227 In fact, a jury cannot have found a crime heinous, atrocious, or cruel and failed to find a defendant guilty of the parole ineligible crime of first degree murder with special circumstances. 228 If society is punishing conduct, the jury, as society's representative, must find facts about that conduct,229 especially those facts which involve moral concerns. AsJustice Breyer has explained, "[w]ords like 'especially heinous,' 'cruel,' or 'depraved' ... require reference to community-based standards, standards that incorporate values [that a] jury is better equipped than a judge to identify and to apply accurately."230 These judgments are particularly retributive because the parole board considers them in the suitability stage of the proceeding. "Unsuitability for parole" is both semantically retributive (connoting unworthiness) and methodologically distinct from calculating a release date on the basis of public safety. Most importantly, if the crime itself is so "heinous, atrocious, or cruel" as to make the offender unsuitable for parole, a jury can make that finding at trial. No intervening factors justifY denial of parole, particularly given that "heinousness" is an element of a greater 227. Cf. United States v. Booker, 543 U.S. 220. 231 (2005) (Stevens,.J.. opinion of the Court) ("[T]he procedural error in Ring's case might have been harmless because the necessary finding was implicit in the jury's guilty verdict ...." (citing Ring v. Arizona, 536 U.S. 584. 609 n.7 (2002»)). 228. See discussion of parole eligible and parole ineligible murder sentences supra text accompanying notcs 43-47. 229. See Bennan, Conceptualizing Blakely, supra note 2, at 90 ("[I]fand when the law provides that certain offense conduct will have certain punishment consequences, the law has triggered the jury trial right ...."). 230. Schriro v. Summerlin, 542 U.S. 348, 361-62 (2004) (Breyer,.J., dissenting). Schriro held that Ring "announced a new procedural rule that does not apply retroactively to cases already final on direct review." Id, at 358; see also Ring, 536 U.S. at 615 (Breyer,.J., concurring in the judgment) ("In respect to retribution, jurors possess an important comparative advantage over judges."), This leaves open a question of whether an inmate's expression of remorse is a factual, morally blameworthy issue, The issue is beyond the scope of this Article, and deals with another dimension of institutional competence: temporality, Parole boards might best consider remorse because there is no real way to measure it at the point a sentence is imposed. For a fuller discussion of remorse in the parole context, see generally Daniel S. Medwed, The Innocent Prisoner's Dilemma: Consequences of Failing to Admit Guilt at Parole Hearings, 93 Iowa L. Rev, 491 (2008) (discussing origins and purposes of parole, normative implications of relying on admissions of guilt and expressions of remorse during parole hearings, and recommending reforms), 2009] HEINOUS, ATROCIOUS, AND CRUEL 935 offense punishable by a parole ineligible sentence. 231 The only things that have changed are that the parole board is a body less representative of the public at large,232 the standard of proof is lower, and the record has grown cold in the fifteen years or more since trial. The parole board is well situated to analyze rehabilitation, however, since it has access to contemporaty psychiatric and behavioral reports, and because the offender's state of mind is examined immediately prior to release. These are facts that neither the jury nor the judge can know at the time of sentencing, facts that "either will not be discovered, or are not discoverable, prior to trial. "233 These contemporary data enable the parole board to measure rehabilitation better than either the judge or the jury could at the time of sentencing. We get better at measuring facts about an offender's public safety risk with the passage of time; we do not, however, get better at measuring the blameworthiness of her crime. C. Reconciling Apprendi and Parole Due Process: Indeterminate Sentences Serve Split Purposes Under a mechanical application of Apprendi, such as that used in Cunningham, the jury must find all facts, including those related to public safety. Under this reading, the rule applies completely or not at all. In other words, an Apprendized parole board would have no jurisdiction to find any facts. This Article argues for an alternate understanding of Apprendi, one where the jury must find only those facts relating to retribution. In this section, I explain how to separate the retributive and rehabilitative portions of California's indeterminate sentences, leaving room for an intermediate solution that reconciles parole due process cases and Apprendi. In Part II.Gr, I address the question of who should find mixed factsthose that relate to both retribution and rehabilitation. In Part ILC.2, I discuss how state legislatures might be limited in recharacterizing the purposes of parole to avoid application of Apprendi at the parole stage. 231. In In re Rosenkrantz, for example, the Governor reversed parole on the ba.'iis of the prisoner's commitment offense and went so far as to say that RosenkranlZ "should be grateful that he was not convicted aftirst degree murder," 59 P,3d 174, 189 (Cal. 2002). The Governor then rehashed the evidence presented at trial about premeditation, a theory that the jury rejected in finding him guilty of second degree murder. Jd. at 215-16; see also In re Elkins, 50 Cal. Rptr. 3d 503, 509-10 (Ct. App. 2006) (upholding Govemor's denial on basis of "atrocious [ness]" of murder, as the "gravity, .. alone [was] sufficient" to deny parole). There was little discussion of future threats to public safety, and much discussion of the crime itself. 232. For retribution to be effective, the offender has to associate his punishment with the opinion of the community. Parole boards, which are typically composed of political appointees of the governor, are not more representative of community values than ajury, As for their expressive component, parole boards operate in prisons, where there is little attention paid to their decisions, and where these decisions come years after the offending conduct. 233. Blakely v. Washington, 542 U.S. 296, 319 (2004) (O'Connor,,/., dissenting). 936 COLUMBIA LAW REVIEW [Vol. 109:893 The Supreme Court has indicated that all time served in prison, including that after parole denial, is punitive."34 But using the duration of a sentence as the sole measure of the degree of punishment has perverse results. Parole is, itself, a form of incarceration, albeit a constructive one where the prisoner serves time in the community; time served on parole satisfies time served on a sentence. 235 Under a purely durational reading, there is no difference between the penalties for first degree murder with special circumstances, first degree murder, and second degree murder: 236 All are life sentences of equal dura234. I argue that the extra time after a finding of unsuitability should be seen as rehabilitative incarceration, provided the parole board does not find commitment offense facts. The Supreme Court has not found that to be the case, however. Beginning with California Department of Corrections v. Morales, 514 U.S. 499 (] 995), the Supreme Court considered a series of challenges to changed parole regulations under the Ex Post Facto Clause. These regulations changed the frequency of subsequent parole hearings. Id. at 501. The prisoners filed suit, claiming that this increased their punishment and was therefore forbidden under the Ex Post Facto Clause. Id. at 508. Though their claims were rejected, the reason for the rejection was not that extra time in prison was not punishment: That was assumed. Instead, the Court found that these changes to the parole system did not materially decrease the chance that they would have been found suitable for parole. Id. at 508-09. For a more recent view, see Justice Scalia's dissent in Oregon TJ. Ice, which argues that total time senred is all that matters, not the purpose the incarceration serves or the stigma involved. No. 07-901, 2009 U.S. LEXIS 582, at *26 (U.S. Jan. 14. 2009) (Scalia,]., dissenting) (arguing that prisoner's "date of release from prison" is "the single consequence most important to convicted noncapital defendants" and that "[t]wo consecutive 10-year sentences are in most circumstances a more severe punishment than any number of concurrent IO-year sentences"). If this is the case, however, parole boards punish and in any parole system with presumptive release, Apprendi requires ajury to find all facts. If the fact of confinement, not its meaning, is key, it is unclear why Apprendi does not also govern civil commitment proceedings. For more on this subject, see W. David Ball, Civil, Criminal, or Mary Jane: Stigma, Legislative Labels, and the Civil Case at the Heart of Criminal Procedure (Apr. 24, 2009) (unpublished manuscript, on file with the Columbia Law Review). 235. See. e.g., Morrissey v. Brewer, 408 U.S. 471, 474-75 (1972) (citing Eighth Circuit's reasoning that parole is only "a correctional device authorizing service of sentence outside the penitentiary"); see also Samson v. California, 547 U.S. 843, 847 (2006) (finding California parolees have no expectation of privacy and no Fourth Amendment protection against suspicionless searches since they are constructively in custody of state prison system). 236. A parolee is still in the legal custody of the California Department of Corrections and Rehabilitation (CDCR). Cal. Penal Code § 3056 (West 2000): see also Samson, 547 U.S. at 851 ("A California inmate may senre his parole period either in physical custody, or ... out of physical custody and subject to certain conditions. Under the latter option, an inmate-turned-parolee remains in the legal custody of the California Department of Corrections through the remainder of his term ...." (internal citation omitted)). Most prisoners have an enumerated term of parole, but "[i]n the case of any inmate sentenced under Section 1168 for any offense of first or second degree murder with a maximum term of life imprisonment, the period of parole, if parole is granted, shall be the remainder of the inmate's life." Cal. Penal Code § 3000.1(a). However, the functional maximum parole tenn is five years for second degree murder and seven years for first degree murder. The California Penal Code directs the parole board to discharge people from parole after that 2009] HEINOUS, ATROCIOUS, AND CRUEL 937 tion. 237 Dispositional issues or stigmatic issues must, therefore, be considered a~ important and distinguishing dimensions of a sentence. Otherwise, there is simply one sentence for murder, no matter what the degree: life. One could argue that the problem with applying Apprendi to indeterminate sentences is that in indeterminate sentences, nonretributive concerns dominate. As in Williams, there is no need for a jury to find facts where the system serves nonretributive goals. Perhaps purely indeterminate sentences serve this purpose,238 but California's indeterminate sentences contain definite terms of punishment: the enumerated terms of years. Judges in California have no discretion over how many years a prisoner must serve before becoming eligible for parole. This term is assuredly a retributive component, assessed for the offense, and thus governed by Apprendi. Because California's indeterminate sentences have both retributive and rehabilitative components-a lengthy term of years followed by discretionary parole release-they provide a unique opportunity to isolate the meaning of incarceration from its duration. With determinate sentences, it is difficult to isolate stigmatic punishments from durational ones, since in most instances-including every case in the Apprendi linethe state punishes more stigmatized crimes with longer sentences. In a certain sense, a determinate sentence is nothing but punishment. 239 With an indeterminate sentence, however, stigma and duration are separable. Prisoners convicted of different degrees of murder might end up serving similar amounts of time, but the degree of the murder assigns a different amount of stigma to each. A second degree murderer is less blameworthy than a first degree murderer. If he serves a longer sentence, it is because he is more dangerous, not because his crime is worse. Because indeterminate sentences more cleanly separate the issues of period unless the board makes a "good cause" determination to retain them. Id. § 3000.1 (b). 237. This line of reasoning also conflicts with existing cases about the liberty interest in parole. The type of lime is a necessary dimension of a person'5 liberty interest; otherwise, there would be no difference between the due process rights in parole revocation and prospective parole release. See infra Part III.A.I; see also supra notc 154. 238. Prior California indeterminate sentences were as vague as "one year to life." The enumerated tenn was merely long enough to make the crime a felony, at which point the state could incarcerate the offender in a state prison. See Cunningham v. California, 549 U.S. 270, 276-77 (2007) (explaining that for sixty years under California's indeterminate sentencing regime, courts imposed open-ended tenns and parole board "determined the amount of time a felon would ultimately spend in prison"'). I do not claim that these old indeterminate sentences were necessarily hybl;ds, with only the enumerated term of a single year denoting the punishment. However, modern enumerated terms, with degrees of offense paired to longer enumer:.Hed terms, fit the "split purpose" analysis. For a more exact definition of "split purpose," see infra text accompanying note 240. 239. Of course, one could also argue that what makes a crime more stigmatized is the length of its sentence; we measure society's degree of disapproval of a crime by the way in which we punish it. 938 COLUMBIA LAW REVIEW [Vol. 109:893 blameworthiness and time served, they help isolate why Apprendi requires jury findings in some instances and permits nonjury findings in others. The two parts of an indeterminate sentence ultimately serve what I call "split purposes." The enumerated term of years exhausts the punishment for the crime and the rest of the sentence measures rehabilitation. This split purpose lays the foundation for how the parole due process cases and Apprendi should be reconciled. I propose that Apprendi be reserved only for the retributive part of the sentence and that retribution for the commitment offense should end with the enumerated term of years. The parole board, in turn, should retain its deference when it measures rehabilitation, but only for that purpose-not for moral facts like "heinousness." A split purpose interpretation would preserve the statutory construction of the California homicide statute. That is, first degree murder and second degree murder would still be punished as different crimes, with different levels of stigma, even though a prisoner convicted of second degree murder could end up serving the same amount of time (or more) as a prisoner convicted of first degree murder. 240 Allowing a parole board, rather than a jury, to make rehabilitation-related decisions under the split purpose analysis also accords with traditional deference to prison officials to deal with in-prison issues and corrections decisions. Expert, repeat players can still access all the information necessary to come up with the most accurate "treatment plan," as long as their decision is about public safety risk, not stigmatic or punitive issues that relate back to the original crime."" This split provides an ideal test case for distinguishing 240. Courts have heretofore dealt with this problem by unfailingly reading parole release as "carly." The bulk of this Article is dedicated to showing why this is theoretically problematic. It is also incorrect as a matter of statutory interpretation since in many states, including California, a prisoner is presumptively released absent a specific finding of fact to the contrary. In fact, in both Greenholtz v. Inmates of Neb. Penal & Corr. Complex, 442 U.S. 1 (1979), and Bd. of Pardons v. Allen, 482 U.S. 359 (1987), the Supreme Court found that the state parole statute granted parole release presumptively. See infra Part lILA. Calculating a release date-the second stage of the parole process-would also require jury findings for certain calculations that increased sentence time. The release date matrix for second degree murder, for example, includes characteristics about the murder victim (the victim was an "accomplice or othenvise implicated in a criminal act" with the offender (minimum), the victim had a personal relationship with the offender (medium), or there was no personal relationship (maximum)) and the level of the offender's participation in the murder (indirect/accomplice liability (minimum), victim initiated struggle (medium), death the result of "severe trauma inflicted with deadly intensity" (maximum». Cal. Code Regs. tit. 15, § 2403(c) (2005). Because the matrices are binding, ajury would have to find each of these facts beyond a reasonable doubt before the board could calculate a sentence based on them. If all time served is punitive, then the calculation of the release date is also subject to Apprendi scrutiny. 241. A criminal conviction gives the parole board the discretion to make decisions about everything attendant to the execution of the sentence. It cannot give the board discretion to extend the conviction to an aggravated form of the commitment offense, however, any more than "ajudge could sentence a man for committing murder even if the jury convicted him only of illegally possessing the firearm used to commit it." Blakely v. 2009] HEINOUS, ATROCIOUS, AND CRUEL 939 facts about an offender from facts about an offense 242 in a manner which is coextensive with case law that understands incarceration as distinct from punishment. 243 Finally, my proposal solves a practical problem: Applying Apprendi to every fact that a parole board considers-that is, reading it mechanically, separate from the meaning of punishmentwould be an administrative nightmare. The jury would have to find facts about psychiatric evaluations, a prisoner's post-release employment plans, and social history. Adopting the retribution/rehabilitation distinction limits the Apprendi line's control to "the central sphere of [its] concern, while reducing the collateral, widespread harm to the criminal justice system and the corrections process."244 It does leave two questions, however. First, what should be done with facts which are indicia both of reprehensibility and threats to public safety? Second, what would prevent the parole board from finding retributive facts under the guise of public safety? I address each in tum. 1. Who Should Find Mixed Retributive/Public Safety Facts? - The commitment offense is not only probative of moral culpability; it might also provide some evidence of a threat to public safety.245 Murderers are blameworthy because they have murdered, but they are also dangerous Washington, 542 U.S. 296, 306 (2004). Thus, if a conviction authorizing the parole board's discretion failed to find the crime heinous, atrocious, or cruel, the discretion it grants does not include the power to find the crime heinous, atrocious, or crueL 242. It is unclear, however, that the current regulations governing the calculation of a release date would survive even under this theory. To the extent they rely on commitment offense facts, they would have to be found by the jury beyond a reasonable doubt. I would argue that guidelines, as opposed to actuarial tables, are inherently retributive. Guidelines express a judgment about the time one should get for a given set of facts. They can be distinguished from actuarial tables, which measure the probable public safety risk associated with a given set of facts. 243. See Kansas v. Hendricks, 521 U.S. 346, 363 (1997) ("'[T]he mere fact that a person is detained does not inexorably lead to the conclusion that the government has imposed punishment.'" (quoting United States v. Salerno, 481 U.S. 739, 746 (1987»)); see also Soling v. Young, 531 U.S. 250, 272-73 (2001) (Thomas,]., concurring) (explaining that confinement is not dispositive factor in making sanction criminal); Hendricks, 521 U.S. at 363 ("If detention for the purpose of protecting the community from harm necessarily constituted punishment, then all involuntary civil commitments would have to be considered punishment. But we have never so held."). A state's power to confine someone to protect the public safety under its parens patriae power is well established. Civil commitments require the state to show cause of future dangerousness every few years, but a state can confine someone indefinitely upon this showing. Civil commitments are specifically nonpenological, of course, and the laws governing them are not located in the penal code. See, for example, California's Sexually Violent Predators Act, which seeks to protect against the "sexually violent criminal behavior" of a person "who has been convicted of a sexually violent offense . . . and who has a diagnosed mental disorder that makes the person a danger to the health and safety of others." Cal. Welf. & Inst. Code § 6600(a) (West 1998 & Supp. 2008). 244. Cunningham v. California, 549 U.S. 270,295 (2007) (Kennedy,]., dissenting). 245. See Berman, Conceptualizing Blakely, supra note 2, at 91 (explaining that certain facts "involve a mix of offense conduct and offender characteristics"). 940 COLUMBIA LAW REVIEW [Vol. 109:893 because they have murdered. Despite Justice Kennedy's claim that "in most instances ... the problem of categories would not be difficult,"246 the commitment offense is irreducibly both a particle and a wave. This Article did not create the mixed fact problem, nor will it attempt to solve it definitively. The problem has been addressed in other areas of the law, however, and I will look to one particularly salient example-the use of criminal priors in sexually violent predator (SVP) lawsas a means of sketching out some of the terrain. In SVP proceedings, prior crimes are considered as evidence of future dangerousness, not as justification for punishment. 247 There is some concern, however, that despite the taxonomically civil label of these procedures, they are functionally serving to repunish the offender. 248 We might be able to borrow some of the tools from this line of jurisprudence to move toward a substantive definition of punishment. 249 But while I acknowledge the difficulty (or impossibility) of determining whether a fact is, in the abstract, more probative of iniquity or dangerousness, the concern in this Article is much more limited: Even if the commitment offense is a mixed fact, the issue is whether the parole board should be the body to consider its effects on dangerousness. Here 1 think the answer is no. First, the California parole board has, in fact, turned into a retributive body, so the danger of punishment overwhelming public safety is real. Second, the parole board's relative expertise does not extend to the commitment offense. If a nonjury body needs to consider a commitment offense's public safety ramifications, a judge is in a much better position than the parole board to do so. The judge is present at 246. Cunningham, 549 U.S. at 297 (Kennedy,]., dissenting). 247. See, e.g., Hendricks, 521 U.S. at 361-62 (finding Kansas statute only sees past behavior as evidence of mental abnonnality or future dangerousness). 248. See, e.g., id. at 373-74 (Breyer,]., dissenting) ("[C]ertain ... special features of the Act convince me that it was not simply an effort to commit Hendricks civilly, but rather an effort to inflict further punishment upon him."); Allen v. Illinois, 478 U.S. 364, 376-77 (1986) (Stevens,j., dissenting) ("Neither the word 'civil' nor the unsettling term applied by the State-'sexually dangerous pcrson'-should be penniltcd to obscure our analysis.") . 249. Kennedy v. Mendoza-Martinez, 372 U.S. 144, 168-69 (1963), for example, established a widely used seven factor test to determine whether a law is criminal or civil: Whether the sanction involves an affinnative disability or restraint, whether it has historically been regarded as a punishment[,] whether it comes into play only on a finding of scienter, whether its operation will promote the traditional aims of punishment-retribution and deterrence, whether the behavior to which it applies is already a crime, whether an alternative purpose to which it may rationally be connected is assignable for it, and whether it appears excessive in relation to the alternative purpose assigned .. Id. (internal footnotes omitted). Most of these factors do not apply to a law that is expressly criminal; however, the notion of proportionality (factor seven) would be particularly useful in situations where second degree murderers serve the equivalent of a first degree sentence. See supra text accompanying note 150. Under this test, the sentence length might be found to be disproportionate, hence punitive, even if the commitment offense facts were nominally used as indicia of public safety. 2009] HEINOUS, ATROCIOUS, AND CRUEL 941 trial and makes his or her decisions based on a fresh record. The parole board reads a cold record, and the hearing provides only a new opportunity, not a new basis, to reconsider the commitment offense. Given the lax standards of review given to parole board decisions, the potential for the mixed fact exception to swallow the rule is, literally, unbounded by judicial oversight. If a parole board can deny parole on the basis of the commitment offense alone, there is currently no way to cure abuses through the judicial system. Judicial sentences, meanwhile, can always be reviewed by higher courtS. 250 Finally, because a commitment offense's effect on public safety would not be dispositive in most cases, the problem is merely an interesting possibility, not an actual obstacle. Dangerous prisoners give fresh evidence of their threats through disciplinary infractions, failure to participate in rehabilitative programming, and poor psychiatric diagnoses. A parole board can cite these factors in finding a prisoner unsuitable. The only case in which the commitment offense is dispositive of dangerousness, therefore, is when a prisoner has no institutional infractions, and the parole board literally has no other evidence of unsuitability. No actuarial tool measuring the risk of recidivism "attempts to predict future dangerousness by virtue of the commitment offense alone. "251 250. One additional argument might be made-that the legislature, not the judiciary, should make the call about who has jurisdiction over mixed facts. Under this viewpoint, the legislature has already decided to give mixed facts to the parole board. The indeterminate sentence is authorized precisely because the legislature is sufficiently concerned about public safety not to run the risk of releasing a dangerous prisoner after a determinate sentence. So if a parole board has discretion to consider public safety, why deny it the discretion to consider the public safety effects of a murder? Perhaps there should be no limitation on the factors the parole board can consider, as long as it is considering public safety and not retribution. See John Monahan, AJurisprudence of Risk Assessment: Forecasting Harm Among Prisoners, Predators, and Patients, 92 Va. L. Rev. 391, 395 (2006) (discussing how, in civil commitment context, where "a legal determination about future conduct in which blameworthiness for past conduct plays no part ... the use of violence risk factors ... is jurisprudentially unconstrained except for classifications subject to strict scrutiny" under the Equal Protection Clause). This line of argument would seem to be foreclosed by the Apprendi line, but Oregon v. Ice resurrected Patterson v. New York, citing it to support a state legislature's sovereign interest in structuring its criminal justice system. Oregon v. Ice, No. 07-901, 2009 U.S. LEXIS 582, at *20 (U.S. Jan. 14, 2009) (citing Patterson v. New York, 432 U.S. 197,201 (1977»; see also infra note 383 (discussing Patterson-era case law). 251. Brief for Stanford Criminal Justice Center as Amicus Curiae Supporting Petitioner Sandra Davis Lawrence at 14-15, In re Lawrence, 190 P.3d 535 (Cal. 2008) (No. 5154018) (on file with the Columbia Law Review) (emphasis omitted) (discussing "vociferous" academic debate about merits of different models that exist, with this conclusion). As an empirical matter, recidivism for lifers is much lower than the rate for the prison population as a whole: Only about a fifth oflifers are rearrested, compared with two-thirds of all released prisoners. Mauer et aI., supra note 34, at 24. Those paroled lifers who are rearrested are no more likely to be rearrested for a violent offense than are those offenders who were initially incarcerated for drug or property offenses. Id. (Of course, some might argue that the system is working because of these low recidivism numbers, and only the 942 COLUMBIA LAW REVIEW [Vol. 109:893 2. Relabeling Desert as Public Safety. - Another loophole might be for the parole board to cite public safety and really mean desert. For example, it could remove the "heinous, atrocious, or cruel" language it uses to describe the commiunent offense. So even if the parole board were restricted to public safety considerations, it could simply expand the factors that bear on public safety, and describe them in morally neutral terms. California already uses this method of recharacterization in the second stage of the parole process, during the calculation of a release date for a suitable prisoner. The release date matrix for second degree murder, for example, includes characteristics about the murder victim and the level of the offender's participation in the murder. 252 Clearly these facts themselves are not used for stigmatic or expressively retributive purposes, but because the matrices are binding,253 they would pose an Apprendi issue under a purely mechanical reading that does not consider the offense/offender distinction. Apprendi's "functional" analysis, which works well in other areas of the doctrine, is of little use here. Function in terms of what operates to constitute a crime is easy.254 Function in terms of meaning is more difficult. It is hard to lie about what you're doing, but much easier-at least for lawyers-to lie about why you're doing it. Given the deference to parole board decisions, it would be difficult to scrutinize the "real" meaning behind either a parole regulation or a parole board decision, although, safe are being released. The number of prisoners released in California is vanishingly small, however, supra text accompanying notcs 144-151, so by this logic, California could guarantee a recidivism rate of zero by transforming all murder sentences into life without the possibility of parole.) Commitment offense facts are also much less probative than fact'i about more recent behavior,such as programming, discipline, and psychiatric reports. The commitment offense may indicate a petitioner's instability, cruelty, impulsiveness, violent tendencies and the like. However, after fifteen or so years in the caldron of prison life, not exactly an ideal therapeutic environment to say the least, and after repeated demonstrations that despite the recognized hardships of prison, this petitioner does not possess those attributes, the predictive ability of the circumstances of the crime is near zero. Irons v. Warden of Cal. State Prison-Solano (fronsl). 358 F. Supp. 2d 936, 947 0.2 (E.O. Cal. 2005). The Ninth Circuit recently endorsed this proposition in Ha)'\v<\rd v. Marshall, 512 F.3d 536, 545 (9th Cir. 2008). The circuit court concluded that Hayward's educacion and conduct in prison meant that his "commitment offense, which occurred twenty-five years ago, cannot demonstrate that Hayward's release will pose an imminent danger to public safety." [d. at 546. 252. Cal. Code Regs. tit. 15, § 2403(c) (2005). For details of specific characteristics included in the matrix, see supra note 240. 253. Id. § 2403(a) ("The panel shall impose the middle ba'ic term renected in the matrix unless the panel finds circumstances in aggravation or mitigation."). 254. See infra Part IV (discussing how under AppTendi's functional definition of "crime," what is important is what statutes actually do, not what they purport to do). 2009] HEINOUS, ATROCIOUS, AND CRUEL 943 arguably, that is exactly what the California Supreme Court tried to do in In re Lawrence. 255 But the stated reason for a given action has been deemed important in other areas of the law, particularly when the stated reason is stigmatic. For example, in Brown v. Board ofEducation, the Supreme Court held that school segregation violated the Equal Protection Clause. 256 Brown was not self-enforcing, however, and the principle it established required concerted efforts in both the political and judicial realms; even though the principle was clear, it was difficult for courts to figure out which practices constituted clandestine or de facto segregation. 257 Ultimately, a determined government can always find new ways to achieve a foreclosed result, but judicial restrictions on the means it can use to do so are nevertheless important: How we do something communicates important messages about what we mean by doing something. 258 Restricting parole findings to public safety means the debate is about who is safe to be let out-not who deserves to be let out. A Supreme Court holding that parole cannot be retributive will not cure all the ills of a system, but it can frame the debate in a way that helps us confront the real issues at hand. III. SENTENCE LtMtTATIONS What, if anything, limits a parole board's power to find a prisoner unsuitable for parole? Can a parole board find a prisoner unsuitable indefinitely? While Part II argues that the relative competencies of criminal justice institutions should determine the kinds offact~ they can consider, this Part explores the limits of what they can do with those facts. The issue, once again, turns on the definition of punishment. While Apprendi applies to retributive facts throughout both the imposition and the execution of the sentence, Apprendi places no limits on a parole board's consideration of nonretributive facts. Under a split purpose interpretation of indeterminate sentencing, the jury verdict limits the state's power to punish to the enumerated term of years, while it allows a parole board to detain prisoners indefinitely on the basis of public safety. Part IILA examines extant parole due process jUrisprudence in depth, arguing that it places no real limits on parole board decisionmak255. 190 P.3d 535 (Cal. 2008). Lawrence essentially reads the plain text of the California parole statute and requires something more than simply finding factors that indicate dangerousness. Id. at 546-49. That is to say. the parole board has to rcally mean that the particular factors it cites in this particular case indicate that this particular prisoner is dangerous. It is a tcst of sincerity, It changes nothing about the statute, only what the parole board is to emphasize when complying with it. 256. 347 U.S. 483, 495 (1954). 257. See generally Charles J. Ogletree, Jr., All Deliberate Speed: Reflections on the First Half Century of Brown v. Board oj Education 124-34 (2004). 258. See infra note 386 and accompanying text (discussing how transparency in operation of criminal laws is crucial interest underlying Apprendi); infra text accompanying note 332 (explaining why changing parole regulations is theoretically possible, but not necessarily probable or even logical). 944 COLUMBIA LAW REVIEW [Vol. 109:893 ing. Part IlI.B then shows how applying Apprendi along retributive/rehabilitative lines would effectively restore parole as a rehabilitative sentencing tool. A. Parole Cases and the Expectancy of Release Supreme Court parole cases place no limit on the amount of time a prisoner can serve after his parole eligibility date, analyzing the issue via a messy, underdefined series of due process cases. According to this line of cases, a parole board is free to deny parole for whatever reason, on whatever facts, for however long. Part liLA. I examines the limited liberty interest a prisoner may have in parole. Part IlI.A.2 discusses this liberty interest in the context of unsuitable life sentences. 1. The Liberty Interest in Parole. - The Supreme Court has generally held that the liberty interest in parole is greatest when an offender has already been released into the community: The state may not revoke parole and return an offender to prison without some minimal process. 259 The Court has also generally held that there is no inherent liberty interest in an incarcerated prisoner's prospective release into parole. 26o Ultimately, the liberty interest in parole suitability depends on the parole statute-if the statute grants an expectation of parole, then some process is due. 261 The reason that this description only applies generally is that the Supreme Court has vacillated on the method it uses to evaluate the liberty interest in parole release. Morrissey v. Brewer, a 1972 parole revocation case, looked to the nature and weight of the liberty interest involved to determine what, if any, due process protections applied: 262 Only a "grievous loss," "one within the contemplation of the 'liberty or property' language of the Fourteenth Amendment," is protected. 263 Seven years later, the Supreme Court abandoned Morrissey's nature and weight methodology in Greenholtz, looking instead to the statutes involved. Greenholtz analyzed the liberty interest in prospective parole release by looking to 259. Morrissey v. Brewer, 408 U.S. 471, 488-89 (1972). 260. Greenholtz v. Inmates of Neb. Penal & Corr. Complex, 442 U.S. 1, 7 (1979) ("There is no constitutional or inherent light of a convicted person to be conditionally released before the expiration of a valid sentence ... [because] the conviction, with all its procedural safeguards, has extinguished that liberty right."), Absent a statutory right, the possibility of parole is but "a mere hope that the benefit will be obtained." Id. at 11. 261. Id. at 12 ("We can accept respondents' view that the expectancy of the release provided in this statute is entitled to some measure of constitutional protection."). 262. Both offenders admitted the parole violations of which they were accused, but alleged that they had been denied due process because their paroles were revoked without hearings, Morrissey, 408 U.S. at 473-74, a striking parallel to the petitioner in Apprendi, who pleaded guilty to the charges but successfully appealed on the basis of the due process jury trial rigbt, Apprendi v. New Jersey, 530 U.S. 466, 470-71 (2000). 263. l\Jorrisey. 408 U.S. at 481 (internal citations omitted). 2009] HEINOUS, ATROCIOUS, AND CRUEL 945 whether the statute created a "protectible expectation of parole"264 that "is entitled to some measure of constitutional protection."265 In 1987, Bom'd of Pardons v. Allen again looked to a state parole statute to find a protected liberty interest,266 rejecting the argument that the discretion vested in parole board officials outweighed the statutorily created interest in release. 267 By the early 1990s, then, Morrissey's nature and weight methodology was left for dead. 268 In 1995, however, the Supreme Court revived Morrissey's method in Sandin v. Conner. 269 The Court criticized Greenholtz's holding that statutes and regulations can create liberty interests,270 complaining that it was forced to "wrestle[ ] with the language of intricate, often rather routine prison guidelines to determine whether mandatory language and substantive predicates created an enforceable expectation."271 Greenholtz also encouraged "prisoners to comb regula264. Greenholtz, 442 U.s. at 11. Even if a statutory expectation exists, however, there need not be "repeated, adversary hearings in order to continue the confinement." Id. at 14. 265. Id. at 12. The Supreme Court held that the Nebraska parole statute granted a protected expectancy of parole release through "the structure of the provision together with the usc of the word 'shall' , ..." Id. at 11-12. It found no inherent liberty interest in parole, however. Id. at 7. 266. 482 U.S. 369, 376-77 (1987). 267. The district court found that the Montana parole statute contained language mandating release, but concluded that the board's discretion was "too broad to provide a prisoner with a liberty interest in parole release." Id. at 372. The Ninth Circuit reversed: While the statute vested "great discretion" in the parole board, the board could not deny parole "once it determines that harm is not probable." Id. (internal quotation marks omitted). The Supreme Court affirmed. Id. at 381. 268. See, e.g., Thomas J. Bamonte, The Viability of Morrissey v. Brewer and the Due Process Rights of Parolees and Other Conditional Releasees, 18 S. Ill. U. LJ. 121, 124 (1993) ("Morrissey's apparent viability is deceptive ... [because] the role played by parole in the correctional system has changed substantially over the twenty years since Morrissey was decided."). 269. 515 U.S. 472 (1995). Prisoner Conner had been sent to administrative segregation ("the hole") for a disciplinary infraction and had not been allowed to present witnesses at his disciplinary hearing. Id. at 475-76. Sandin is one in a line of prison due process cases, such as Turner v. Safley, where the Supreme Court was willing to defer substantially to prison officials and their legitimate penological interests, even where they impinged on core constitutional rights. See, e.g., Turner v. Safley, 482 U.S. 78,81 (1987) (upholding restrictions on inmate correspondence against First Amendment challenge, but striking down restrictions on inmate marriages). But see Johnson v. California, 543 U.S. 499, 509 (2005) (applying strict scrutiny to California's practice of racially segregating inmates during prison intake). For more on the revival of Morrissey, see generally Sharif A. Jacob, Note, The Rebirth of Morrissey: Towards a Coherent Theory of Due Process for Prisoners and Parolees, 57 Hastings LJ. 1213, 1226 (2006). 270. Sandin, 515 U.S. at 479-81. It is unclear, however, how parole statutes or regulations that create a presumption of parole release are different from sentencing statutes or regulations that create presumptive (or functional) statutory maxima. See infra Part IV.B.!. 271. Sandin, 515 U.S. at 480-81. The Prison Litigation Reform Act (PLRA) requires inmates to exhaust all their administrative appeals before filing suit in court. 42 U.S.c. 946 COLUMBIA LAW REVIEW [Vol. 109:893 tions in search of mandatory language."272 The Court therefore "retum[ed] to due process principles" and held that the states could generally create only those interests related to freedom from restraint which, while not so unexpected "as to give rise to protection by the Due Process Clause of its own force ... nonetheless impose [d] atypical and significant hardship on the inmate in relation to the ordinary incidents of prison life. "273 The Supreme Court used Sandin to find an inherent, protected liberty interest-one which involved eligibility for parole-in Wilkinson v. Austin. 274 The decision to place an offender into a "supermax" facility, which rendered otherwise eligible prisoners ineligible for parole, constituted an "atypical and significant hardship" and thus required procedural protection. 275 While there was no "liberty interest in avoiding transfer to § 1997e(a) (2000). Prison regulations are thus stacked against the prisoner-they afford no protection, but they create all sorts of hurdles for prisoners filing suit. Prisons have an incentive to write Byzantine regulations as a fonn of protection against lawsuits. None of the regulations will bind them negatively, and jf a prisoner fails to observe any of the niceties therein, he or she will be barred from filing suit to reach the substance of his or her complaint. 272. Sandin, 515 U.S. at 481. Indeed, Justice Ginsburg, in dissent, noted this "practical anomaly" which applies equally to sentencing laws: "[A] State that scarcely attempts to control the behavior of its prison guards may, for that very laxity, escape constitutional accountability; a State that tightly cabins the discretion of its prison workers may, for that attentiveness, become vulnerable to constitutional claims." Id. at 490 (Ginsburg,]., dissenting). (Kevin Reitz has noted a similar problem with the Apprendi line. See supra note 207.) Justice Ginsburg thought the prison officials' behavior was itself a deprivation ofliberty under the Due Process Clause. Sandin, 515 U.S. at 491 (Ginsburg,]., dissenting) . 273. Sandin, 515 U.S. at 483-84 (majority opinion). Note that there are potentially two separate due process claims: one where due process arises out of its own force (presumably a practice or incident that shocks the conscience or is arbitrary and capricious), and one that imposes an atypical and significant hardship. Because the Hawaii state prison's action did not "present a dramatic departure from the ba.'iic conditions of Conner's indetenninate sentence," the state did not create a liberty interest in its decision. 1d. at 485-86. Ironically (or incoherently), the Sandin Court argued that a prisoner would not necessarily serve more time for discipline infractions because he would have "procedural protection at his parole hearing in order to explain the circumstances behind his misconduct record." Id. at 487. This protection came, stupefyingly, from the state's administrative rules-the rules that Sandin said created no liberty interest. Id. Thus, regulations which do not protect inmates at disciplinary hearings are of no consequence because other regulations do protect them during parole hearings-even though due process guarantees the presence of neither. 274. 545 U.S. 209, 224 (2004). 275. Id. at 223. An inmate could be sent to the supermax prison if he were convicted of certain crimes, or upon certain behaviors in prison. Id. at 216. This decision did not depend on any independent findings of fact about the commitment offense, however. Id. at 216-17. 2009] HEINOUS, ATROCIOUS, AND CRUEL 947 more adverse conditions of confinement,"276 inmates' ineligibility for parole during confinement in the supermax facility, coupled with the duration of time a prisoner could spend there, "impose[d] an atypical and significant hardship within the correctional context."277 By this reasoning, eligibility for prospective parole release is sufficiently weighty to create a protectible liberty interest. It is unclear whether a statute can still create a protectable liberty interest in parole. 278 Wilkinson stated that "Sandin abrogated Greenholtz's ... [statutory] methodology."279 But Sandin itself stated that abandoning statutory methodology "[did] not technically require [it] to overrule any holding of this Court. "280 The Ninth Circuit has taken its own approach in analyzing the due process limits of parole denials, holding that unsuitability findings based on unchanging factors can violate due process. Here the unchanging factors include not only the commitment offense but also the inmate's prior record and his social history before incarceration. In Biggs v. Terhune, the Ninth Circuit denied the habeas petition of a prisoner challenging his unsuitable life sentence but suggested in dictum that "continued reliance in the future on an unchanging factor" could ultimately "result in a due process violation."28l The Ninth Circuit cited the Biggs dicta in two cases without applying it 282 before holding that a state's reliance on the commitment offense, a "stale and static factor," violated a prisoner's due process rights."8' 276. [d. at 221. These conditions involved "extreme isolation" in a seven-by-fourteen foot cell where the light was always on and the architecture prevented conversation or communication with other inmates. [d. at 214. 277. Id. at 221. The Supreme Court ultimately held that the process the state employed was sufficient to prevent erroneous placement, however, and found no due process violation. Id. at 225. 278. The Supreme Court has yet to apply Sandin (or reapply Aforrisscy) in a case explicitly about the denial of parole. Wilkinson, however, turned to these cases to assess the constitutionality of classifications affecting a prisoner's eligibility for parole. 279. n'ilkinson, 5·15 U.S. at 229. 280. Sandin v, Conner, 515 U,S, 472, 483 n.5 (1995), The decision "only abandon[cd] an approach that in practlce is difficult to administer and which produces anomalous results." ld. 281. 334 F,3d 910, 917 (9th CiL 2003), 282, See Irons v, Carey (Irons /1), 505 F,3d 846, 853-54 (9th CiL 2007) (finding no due process violation in part because Irons's suitability determination was made before expiration of minimum tenn of years); Sass v. Cal. Bd. of Prison Terms, 461 F.3d 1123, 1129 (9th Cir. 2006) (declining to follow Biggs dicta because of Antiterrorism and Effective Death Penalty Act of 1996 requirement that habeas be granted only where state decisions upholding parole denial are "contrary to ... or involve an unreasonable application of, clearly established Federal law as determined by the Supreme Court"). 283. Ha}Ward v, Marshall, 512 F,3d 536, 546-47 (9th CiL 2008), In graming Hayward's habeas petition, the Ninth Circuit focused primarily on whether there was "some evidence" supporting the Governor's reversal of the parole board's decision to grant parole. Id. at 544. After concluding that "no evidence" supported the Governor's decision, the panel independently held that the Governor's reliance on an unchanging factor violated due process. Id. at 544, 546-47 ("[T]he unchanging factor of the gravity of 948 COLUMBIA LAW REVIEW [Vol. 109:893 Biggs did not address Morrissey, Greenholtz, or Sandin, and is not good law outside the Ninth Circuit. Nevertheless, it may illustrate one way in which courts might construe an inherent liberty interest in parole release. In Wilkinson, where the state's interest in controlling dangerous, undisciplined inmates was paramount,284 the Supreme Court identified some interest in ongoing suitability hearings. Conducting repeated parole hearings and denying each time on the same facts might therefore constitute an atypical and significant hardship. All of this analysis is speculative, though; post-Biggs cases do not explain the meaning behind the Biggs dictum or situate it within the existing due process jurisprudence, but instead focus on whether the number of denials has caused the due process interest to ripen. 285 2. The Liberty Interest in Parole Applied to Unsuitable Life Sentences. Given the uncertainty in the parole due process line, it is unclear whether a California prisoner has a liberty interest in parole release, and, if so, whether an unsuitable life sentence deprives him of this interest without due process of law. If Greenholtz governs, we look to the California parole statute to see if it creates an expectation of parole. If Morrissey or Sandin governs, we look to the ordinary incidents of punishment-which, under Wilkinson, might require some regular opportunity for parole release. 286 Under the Greenholtz statutory method of analysis, courts look to see whether the parole statute creates a protected liberty interest in paHa)"'Vard's commitment offense had no predictive value regarding his suitability for parole [and i]n the circumstances of this case, the Governor \~olated Hayward's due process rights by relying on that stale and static factor in reversing his parole grant."}, At the same time, the court limited its holding in a footnole that explained that "certain conviction [sic] offenses may be so 'heinous, atrocious, or cruel' that a prisoner's due process rights might not be violated ifhe or she were denied parole solely on ... [that] basis." Id, at 547 n.l0. 284. Sec Wilkinson, 545 U.S. at 227 ("Prison security, imperiled by the brutal reality of prison gangs, provides the backdrop of the State's interest."); see also id. at 229 ("Prolonged confinement in Supermax may be the State's only option for the control of some inmates .... "). 285. Note, however, that Sass tried to tie the Biggs dictum to the arbitrariness concerns of the Due Process Clause. Sass, 461 F.3d at 1129. It is arbitrary if the crime that has rendered onc unsuitable for parole suddenly poses no obstacle. The magistrate in Irons v. Warden oj California State Prison-Solano (Irons /) also argued that, in Irons's unsuitable life sentence, the parole board would either make the same finding about the commitment offense each year or make a different decision on the same facts-an argument which clearly raises "arbitrary and capricious" due process concerns. 358 F. Supp. 2d 936, 947 (E.D. Cal. 2005); see also In re Criscione, No. 71614, slip op. at 20 (Cal. Super. Ct. Aug. 30, 2007). available at http://www.bayareanewsgroup.com/multimedia/ rnn/ncws/Cliscione_complainc091307.pdf (on file with the Columbia Law Review) ("The fact that Title 15, § 2402, has been invoked in every case, but then sometime later not invoked, tends to show either completely arbitrary and capricious behavior or that unwritten standards are what really determine outcomes."). 286. Or at least a parole hearing. Then again, if we are really to look at the nature and weight of the right, depriving someone of a parole hearing that is essentially an empty fonnal gesture seems unlikely to generate an interest of sufficient weight that Wilkinson would deem it wonhy of protection. 2009] HEINOUS, ATROCIOUS, AND CRUEL 949 role. 287 Even under this method, however, it is unclear whether the mandatory language in the California statute creates such an interest. While the California parole statute is nearly identical to the statutes Greenholtz and Allen identified as creating a protectible liberty interest in parole,2BB the California Supreme Court in Dannenberg held that the "statutory concern for public safety in the individual case trumps any expectancy the indeterminate life inmate may have."es9 According to the Ninth Circuit, however, Dannenberg did "not explicitly or implicitly hold that there is no constitutionally protected liberty interest in parole."29o Instead, it simply "addressed the nan-ow question whether the Board must engage in a comparative proportionality analysis in setting parole dates ... before determining whether an inmate is suitable for parole."291 Thus, California might have created a protectable liberty interest by statute, in which case due process might limit the practice of unsuitable life sentences. Under the Morrissey/Sandin "nature and weight" method, we look to see if unsuitable life sentences impose an atypical and significant hardship on prisoners. 292 Wilkinson held that classification affecting parole eligibility can impose such a hardship. A court might therefore look at whether California classifies parole eligible prisoners differently from parole ineligible ones as evidence of what might be an atypical and significant hardship. In California, prisoners who serve parole ineligible sentences are automatically classified in maximum security prisons, while 287. See supra notcs 264-265 and accompanying text. 288. Compare Cal. Penal Code § 3041 (b) (West 2000 & Supp. 2008) (explaining that the board "shall set a release dale unless it determines ... that the consideration of the public safety requires a more lengthy period of incarceration")' with Sci. of Pardons v. Allen, 482 U.S. 369, 376-77 (1987) ("'SUbject to the following provisions, the board shall release [a prisoner] on parole ... [but] only when the board believes that he is able and willing to fulfill the obligations of a law-abiding citizen.''' (quoting Mont. Code Ann. § 4623-201 (1985»), and Greenholtzv.lnmates of Neb. Penal & Corr. Complex, 442 U.S. I, II (979) (explaining that state law requires that parole board "'shall order'" prisoner's relcase unless it chooses to defer parole for one of four specific reasons (quoting Neb. Rev. Stat. § 83-1,1l4(2) (1976»). 289. In re Dannenberg, 104 P.3d 783, 795 (Cal. 2005). 290. Compare Hayward v. Marshall, 512 F.3d 536, 542 (9th Cil. 2008) (reaffirming California statute's creation of a liberty interest), and Sass, 461 F.3d at 1128, with Dannenberg, 104 P.3d at 794 (describing second degree murder sentences as "imprisonment/oT life, subject to the possibility of sooner release on parole"). Three years before Dannenberg, the Ninth Circuit had held that the California statute created an expectation of release. MeQuillion v. Duncan, 306 F.3d 895, 902 (9th Cil. 2002). This back-and-forth between the fcderal and state court\ is one of many problems with current due process doctrine. See Jacob, supra note 269, at 1228-29 (highlighting problems that lower courts have had understanding interaction of federal and statc law in due process detenninations) . 291. Sass, 461 F.3d at 1127-28. But even Dannenberg's reading, if controlling,justifies indefinite detention on the basis of public safely, not desert. See Dannenberg, 104 P.3d at 785-86 (framing question in public safety terms). 292. See supra notes 262-263, 268-273 and accompanying text. 950 COLUMBIA LAW REVIEW [Vol. 109:893 parole eligible, indetenninately sentenced prisoners are not so classified. 29s This might provide some evidence that Wilkinson applies. Then again, the amount of process required under Wilkinson is minimal. B. Apprendi, the Statutory Maximum, and the Enumerated Term The fonnalist critique of Apprendi focuses in part on Blakely and Booker's analysis of the statutory maximum sentence,Z94 but while the Blakely/Booker analysis is highly technical, it is functional, not formal. This section sets out the two different definitions of the statutory maximum, what I call the "functional statutory maximum" (the maximum for Apprendi purposes) and the "taxonomic statutory maximum" (the maximum laid out in a given criminal statute). The functional statutory maximum is the maximum punishment that can be imposed based on facts the jury has found. In a binding guidelines system, the functional statutory maximum is the presumptive sentence for a given offense. If a judge imposes a longer sentence without a jury finding the necessary facts, the resulting sentence violates the Apprendi right, even if its term is less than the taxonomic statutory maximum. According to the California parole statute, parole eligible prisoners shall be released at the end of the enumerated tenn unless the parole board finds facts justitying unsuitability for parole. That is, release is presumptive after the enumerated tenn. Thus, the functional statutory maximum sentence in a fifteen-years-to-life sentence is not life, but fifteen years. Apprendi's protections apply to any facts that increase punishment beyond that fifteen-year threshold. I. The Meaning of the Functional Statutory Maximum Sentence. - Blakely and Booker clarified that punishment, for Apprendi purposes, is defined functionally. In deciding whether an offender faces an increased maximum penalty, courts must look past legislative labels to analyze how a statute operates. In a binding system, the presumptive sentence is the statutory maximum, even where it is less than a taxonomic maximum sentence. In Blakely, the Supreme Court defined the statutory maximum as the presumptive sentence for a given set of facts, not the maximum sentence defined in the statute under which the offender is charged. Ralph Blakely kidnapped his estranged wife and pleaded guilty to second degree (nonaggravated) kidnapping, a class B felony.29s Washington State's 293. See generally Fama et al., supra notc 97, at 98, 100 (noting prisoners sentenced to death or LWOPP automatically classified as Level N). Viewed cynically, however, denial of parole is, in fact, quite an ordinary incident of punishment, since almost no one serving a life sentence is granted parole. See supra Part J.D. Thus, there are either gross due process violations occurring throughout the state, or the ordinary incidents tcst has some kind of evolving standard of (in)decency built into it. That is, a practice that is widespread cannot be significant and atypical-just as a practice that most states engage in cannot be cruel and unusual. 294. See supra note 2. 295. Blakely v. Washington, 542 U.S. 296, 298-99 (2004). 2009] HEINOUS, ATROCIOUS, AND CRUEL 951 sentencing statute specified a maximum penalty of 120 months for class B felonies. 2g6 The state's binding sentencing guidelines, however, provided a presumptive sentence for nonaggravated kidnapping of forty-nine to fifty-three months. 297 At sentencing, the judge made his own finding of fact that Blakely "had acted with 'deliberate cruelty,'" a statutorily enumerated aggravating factor that allowed the judge to impose a sentence of ninety months. 2gB Although both the presumptive sentence (fiftythree months) and the aggravated sentence (ninety months) were below the maximum term of 120 months allowed by statute, the Supreme Court nevertheless vacated the sentence, holding that the defendant was "sentenced to prison for more than three years beyond what the law allowed ... on the basis of a disputed finding that he had acted with 'deliberate cruelty.' "299 This judgment turns on Blakely's functional definition of the statutory maximum, which is different from the maximum the statute's text authorizes (the taxonomic statutory maximum). Washington's sentencing guidelines were binding; the judge could only sentence Blakely to a maximum of fifty-three months on the facts to which he pleaded guilty. The judge was not permitted to impose a ninety-month sentence unless a jury had found, or Blakely admitted, deliberate cruelty. The functional statutory maximum in Blakely's case, then, was fifty-three months. "When a judge inflicts punishment that the jury's verdict alone does not allow, the jury has not found all the facts 'which the law makes essential to the punishment' and the judge exceeds his proper authority."300 Pleading guilty to the presumptive crime is not the same as pleading guilty to the aggravated version of that crime, even if the taxonomic maximum sentence for the crime the legislature calls kidnapping is greater than either. Booker used Blakely'S functional analysis in holding the Federal Sentencing Guidelines unconstitutional. 3D] A jury found Freddie Booker guilty of possession with intent to distribute at least fifty grams of crack, an offense for which the binding guidelines sentence was at most twentyone years, ten months. 302 At sentencing, however, the judge found by a preponderance of the evidence that Booker "possessed an additional 566 grams of crack and that he was guilty of obstructing justice," exposing him to a sentencing range of thirty years to Iife. 303 Both sentences were less than the taxonomic maximum in the drug trafficking statute, which "prescribes a minimum sentence of 10 years in prison and a maximum 296. 297. 298. 299. 300. 301. Court). 302. 303. Id. at 303. Id. at 299. Id. at 300 (quoting Wash. Rev. Code Ann. § 9.94A.390(2)(h)(iii) (2000)). Id. at 313. Id. at 304 (quoting 1 j. Bishop, Criminal Procedure § 87. at 55 (2d ed. 1872). United States v. Booker, 543 U.S. 220, 232 (2005) (Stevens, j., opinion of the Id. at 227. Id. 952 COLUMBIA LAW REVIEW [Vol. 109:893 sentence of life for that offense."304 But because the guidelines were binding, and because they required the finding of certain aggravating facts before an increased guidelines sentence could be imposed, "as in Blakely, 'the jury's verdict alone does not authorize the sentence. The judge acquires that authority only upon finding some additional fact.' "305 The maximum punishment authorized by the jury's verdict was the presumptive sentence under the guidelines: twenty-one years, ten months. This was, in turn, the functional statutory maximum for Apprendi purposes,306 because the guidelines "are mandatory and binding on all judges."30? Again, the jury right attached to facts that would increase a binding guidelines sentence, even though the guidelines sentence was less than the taxonomic statutory maximum. 30s 2. The Functional Statutory Maximum of an Indeterminate Sentence Is the Enumerated Term. - Confusion over Apprendi's definition of the (functional) statutory maximum means that Apprendi-based parole challenges get rejected out of hand. Judges fail to look at the operation of the parole statute, note simply that the indeterminate sentence includes the phrase "to life," and conclude that a parole board cannot increase the maximum punishment to which a prisoner is subjected. 309 The 304. Id. (citing 21 U.S.C. § 841(b)(I) (A) (iii) (2000». 305. Id. at 235 (citing Blakely, 542 U.S. at 305) ("There is no relevant distinction between the sentence imposed pursuant to the Washington statutes in Blakely and the sentences imposed pursuant to the Federal Sentencing Guidelines in these cases."). 306. Id. 307. Id. at 233. 308. Critics have derided Blakely's definition as "not tied to colloquial meaning, legislative intent, or traditional usage of the tenn." Reitz, supra notc 3, at 1093; see also id. at 1090-92 (noting that overwhelming majority of lower courts before Blakely held that Apprendi did not apply to guidelines sentences below statutory maximum). 309. See, e.g.• In re Dannenberg, 104 P.3d 783, 804 (Cal. 2005) ("[T]he indeterminate sentence is in legal effect a sentence for the maximum term, subject only to the ameliorative power of the parole authority to set a lesser term." (internal quotation marks and citations omitted». A number of trial court" in the Ninth Circuit have rejected the invitation to apply Apprendi to parole suitability. See, e.g., Hawks v. Kane, No. C 0502853 JSW, 2006 U.S. Dist. LEXIS 86335, at *15 n.5 (N.D. Cal. Nov. 21, 2006) (noting defendant's "maximum sentence is life," and that "[b]y determining whether or not to grant parole, the Board is not increasing his penalty beyond the statutory maximum"); McCauley v. Brown, No. C 05-1817 SI, 2006 U.S. Dist. LEXIS 79040. at *23 (N.D. Cal. Oct. 23, 2006) ("Apprendi [is not] implicated by a parole board's decision whether to release him before th[e] maximum sentence has been served."); Bibbs v. Kane, No. C 05-04024 ISW, 2006 U.S. Dist. LEXIS 73659, at *14 (N.D. Cal. Sept. 27, 2006) ("By detennining whether or not to grant parole, the Board is not increasing his penalty beyond the statutory maximum."); Brumett v. Kane, No. C 04-05423 JSW, 2006 U.S. Dist. LEXIS 73661, at *16 (N.D. Cal. Sept. 27, 2006) ("By determining whether or not to grant parole, the Board is not increasing his penalty beyond the statutory maximum. Apprendi and Blakely are inapposite to Petitioner's challenge to the Board's parole determination."); cf. Diaz v. Curry, No. C 06-0586 CRB. 2007 U.S. Disl. LEXIS 53166, at *23-*24 n.2 (N.D. Cal.July 10, 2007) ("Even assuming [the Apprendi line is] pertinent in the parole context, the BPT [Board of Prison Terms] ... merely characterized the commitment offense based on facts already admitted by the defendant."). A case from the First Circuit also rejected an 2009] HEINOUS, ATROCIOUS, AND CRUEL 953 California Supreme Court has itself characterized the maximum of an indeterminate sentence as life, "subject to the possibility of sooner release on parole."310 But the statutory maximum punishment for Apprendi purposes is the functional statutory maximum, not the taxonomic statutory maximum. Under the California parole statute, release into parole is presumptive. 3JI A parole board has to provide reasons for denying suitability (and extending a sentence) after a prisoner has served his enumerated term of years. 312 This evidence takes the form of "aggravating facts beyond the minimum elements of that offense."313 Absent any threat to public safety, the board "shall set a release date."314 Thus, the functional statutory maximum in a fifteen-years-to-life sentence is fifteen years. 315 Only if the parole board were preauthorized to deny parole without finding any additional facts would the functional statutory maximum be life. This interpretation runs counter to some of the language in Justice Scalia's Apprendi concurrence, where he states that a prisoner getting released into parole has gotten "less than" a full sentence. 3J6 His language suggests that the maximum sentence in an indeterminate X-years-to-life sentence would be life, and dlat the "tenderhearted" act of being granted parole is not something to which a prisoner has a right: Will there be disparities? Of course. But the criminal will never get more punishment than he bargained for when he did the .1pprendi application to parole, but there the offender was convicted of first degree murder "by extreme atrocity or cruelty" by the jury. Obershaw v. Lanman, 453 F.3d 56, 57 (1st Cir. 2006). 310. Dannenbe7g, 104 P.3d at 794. 311. Cal. Penal Code § 3041 (a) (West 2008) (Slating that parole board "shall normally set a parole release date" during parolee's first hearing). This is the essential difference bcnveen a due process analysis and Apprendi statutory maximum analysis. Due process analysis says a prisoner has no right to release and the state need not give it to him. Apprendi analysis, on the other hand, starts from the opposite end: The state is entitled only to the punishment authorized by the jury and cannot subsequently increase it. Thus, the maximum penalty in each is different, because the initial allocation is different (the Coase Theorem be damned). 312. See Danne7lbe7g, 104 P.3d at 803 n.16; see also supra notes 101-102 and accompanying text. Note that suitability hearings generally take place a year before the enumerated term is up (less good time credits), but for simplicity, I have used round numbers. 313. Danne7lberg, 104 P.3d at 803 n.16. 314. Cal. Penal Code § 3041 (b) (West 2000 & Supp. 2008). 315. Indeed, a functional analysis of determinate sentences might engender Apprendi protections for disciplinary infractions leading to the loss of good time credits. Given that good time credits are presumptively given in most states, the presumptive release date of a detenninate sentence is actually the tenn of years minus good time credits. Disciplinary infractions leading to the revocation of good time credits increase the functional statutory maximum sentence, even though the taxonomic "pronounced" sentence remains unchanged. 316. See Apprendi v. New Jersey, 530 U.S. 466, 498 (2000) (Scalia, J., concurring) (arguing that felon should "thank the mercy of a tenderhearted parole commission if he is let out inordinately early"). 954 COLUMBIA LAW REVIEW [Vol. 109:893 crime, and his guilt of the crime (and hence the length of the sentence to which he is exposed) will be determined beyond a reasonable doubt by the unanimous vote of 12 of his fellow citizens. 317 In addition to making assumptions about the way a state's parole system operates,318 Justice Scalia's argument rests on an assumption that the key interest Apprendi vindicates is notice. If this is true, then, indeed, a criminal already knows he could do life for murder. But Blakely had constructive knowledge that he could do 120 months-the taxonomic maximum-for kidnapping, and his ninety-month sentence was vacated. The judge could have sentenced Charles Apprendi to consecutive sentences based on his guilty plea alone and put him in prison for a total of twenty years, but instead he found aggravating facts and sentenced Apprendi to twelve. 31g The taxonomic upper limit of the sentencing range, in other words, has never been the same as the functional statutory maximum for Apprendi purposes, nor has notice been Apprendi's sole concern. 320 Instead, in each of the above cases, the Supreme Court has made clear that a defendant faces only the presumptive sentence upon a verdict (or plea) of guilty, because only the jury can find facts which punish him more. 321 Under the current statutory regime in California, commitment offense parole denials operate as sentence enhancements for "heinousness." The parole board tacks on incarceration for stigmatic reasons re317. Id. 318. The parole system described in Justice Scalia's statement would have no expectation of parole, no standards for parole suitability, and no requirements that facts beyond the minimum for conviction be found. This does not describe California's parole system-the board is required to have reasons for denying parole, it must issue a written statement to the inmate setting forth the reasons or reasons for denial, Cal. Penal Code § 3041.5(b) (2) (West 2000), and these reasons must go beyond the minimum facts necessary for conviction. In other words, California's indeterminate sentences contain within them binding, presumptive sentences, and a defendant does not have notice that he will selVe more than the enumerated term based on his commitment offense. That is sufficient to take parole suitability out of the realm of a random, tenderhearted action and into "Apprendi-Iand." Ring v. Arizona. 536 U.S. 584. 613 (2002) (Scalia,.J., concurring). 319. Apprendi, 530 U.S. at 470-71. 320. See, e.g., infra note 391 and accompanying text (explaining that degree of culpability in murder statute is among concerns Apprendi vindicates). 321. 1 note also that Justice Scalia's interpretation conflates stigmatic incarceration (punishment) with incapacitative or rehabilitative punishment (public safety), two issues Apprendi was at pains to distinguish. See supra Part ILB.l. Confinement is not sufficient to define punishment (e.g., civil commitment), and arguably not necessary either (e.g., shaming punishments such as wearing an "I am a drunk driver" sign in public). See, e.g., United States v. Gementera. 379 F.3d 596, 598, 606 (9th Cir. 2004) (upholding supervised release condition requiring convicted mail thief to wear signboard stating, "I stole mail. This is my punishment," as part of "a comprehensive set of provisions that expose the defendant to social disapprobation"). An offender might, indeed, be kept in prison for life under an indeterminate sentence, but only because he poses a threat to public safety, not because he is being punished for life. Nevertheless, Justice Scalia has recently maintained that time served is all that matters, and that Apprendi's protections might cover punishments that are less stigmatic but result in more time served. See supra note 234. 2009] HEINOUS, ATROCIOUS, AND CRUEL 955 lated to the commission of the crime. After Apprendi, however, a judge cannot impose a two-year sentence enhancement for the use of a firearm during the course of a murder without a jury finding that fact 322-even though a fifteen-years-to-life sentence and a seventeen-years-to-life sentence both have a taxonomic maximum term oflife. 323 Similarly, ajudge cannot change a fifteen-years-to-life second degree murder sentence to a sixteen-years-to-life sentence based on her own finding of "heinousness," not only because there is no statutory authorization for the enhancement, but also because a sixteen-years-to-life sentence is a more stigmatic punishment than one of fifteen years to life. Only a jury can make the factual findings justifying these enhancements-no nonjury body, whether judge or parole board, may.324 3. Avoiding Apprendi Through Changing Parole; ar, VVhy Does This Even Matter? - Does reconciling the due process parole line with the Apprendi line depend on the exact statutes at play in California? To what extent, if any, can a state avoid Apprendi scrutiny by changing its parole statutes and guidelines-say, by removing the presumption of parole release, removing the requirement that there be any reasons for unsuitability, and removing any enumerated unsuitability factors? As Justice O'Connor argued in her Apprendi dissent, if New Jersey could change its statute to make "precisely the same differences in punishment turn on precisely the same facts, and can remove the assessment of those facts from the jury ... it is impossible to say that the Fifth, Sixth, and Fourteenth Amendments require the Court's rule."325 322. See Cunningham v. California. 549 U.S. 270, 280 (2007) ("[S]tatutory enhancements must be charged in the indictment, and the underlying facts must be proved to the jury beyond a reasonable doubt."), 323. Cal. Penal Code § 1170.1(e) (West 2004) ("All enhancements shall be alleged in the accusatory pleading and either admitted by the defendant in open court or found to be true by the trier of fact."). 324. Of course, unsuitable life sentences are not instantly converted to first degree punishment; it takes a series of denials to get there. To argue that Apprendi is not implicated only because the second degree conviction gets transformed into a first degree punishment incrementally-rather than all at once-values fonn over function, an approach Apprendi rejects. 325. Apprendi v. New Jersey, 530 U.S. 466,543 (2000) (O'Connor,]., dissenting); sec also Bd. of Pardons v. Allen, 482 U.S. 369, 381-85 (1987) (O'Connor,]., dissenting) (arguing that broadly granted discretion "belies any reasonable claim" to entitlement to parole). Justice Scalia joined Justice O'Connor's Allen dissent, in which she argued that the Montana parole statute did not give rise to a protected liberty interest because the state had not acted "to limit meaningfully the discretion of the decisionmakers." Id. at 382. A statute limited a parole board's discretion only if it constrained the relevant decisionmakers with "particularized standards or criteria." ld. (internal citations and quotation marks omitted). If a statute is so "broadly framed" as to "essentially leave the decision whether or not to grant release on parole to the discretion of the Board, .. the statute simply fails to create a legitimate entitlement to release," Id. at 384. In other words, penalties and privileges need to be tightly bound to individual facts, which means that the real difference between parole release and sentencing is the absence of particular terms associated with particular facts. (It follows from this that, since the Nebraska statute 956 COLUMBIA LAW REVIEW [Vol. 109:893 Kevin Reitz has observed that there are two ways of responding to the Apprendi jury right: the "approach" or "B1akelyization" option; and the "avoidance" or "Bookerization" option. 326 Polities choosing to approach Apprendi, for example, require jury determination of aggravating sentencing factors. 327 Avoidance is a far more popular approach, however, exploiting a "mile wide" hole in the Apprendi line of Sixth Amendment jurisprudence. s28 As in Booker, an avoiding polity need only make binding guidelines advisory. S29 After making the guidelines advisory, the taxonomic and functional statutory maxima are the same, so a jury need not find any facts a judge relies on. Without the mandatory provisions, the guidelines "fall outside the scope of Apprendi's requirement."33o In other words, a federal judge can and does find facts in imposing sentences, but only because no particular fact is required. 33 1 But while a state might choose to enact any number of statutes and regulations, it is unclear why it would choose to employ a completely ungoverned parole system. States issue parole standards and regulations because they serve purposes other than compliance with court decisions. If a state really wanted the ability to punish offenders more severely, it could simply eliminate indeterminate sentences and move towards determinate sentences of incredible duration-say, ninety years per offense. It would not need to take the long way there, increasing punishment through parole. So while it is possible a state could make a punitive, wholly discretionary version of parole, it would be much easier for states simply to dispense with indeterminate sentencing entirely. In other words, a state could certainly make these changes, but it is a mistake to think of them as somehow a trivial or obvious response. 332 in Greenholtz also failed to offer particularized standards, Greenholtz was an "aberration and should be reexamined and limited strictly to its facts." Id. at 385.) 326. Reitz, supra note 3, at 1108-09, 1114. 327. Id. at 1109. 328. Id. at 1113. 329. United States v. Booker, 543 U.S. 220, 259 (2005) (Breyer,]., opinion of the Court). 330. Id. 331. Exactly how advisory the federal guidelines are is a matter that the Supreme Court is tcasing out in a series of cases. Rita v. United States held that a circuit may establish a presumption of reasonableness to a within~guideJines sentence. 127 S. Ct. 2456, 2462 (2007). Gall v. United Slales held that circuit courts must review all sentences under a deferential abuse of discretion standard "whether inside, just outside, or significantly outside the Guidelines range." 128 S. Ct. 586, 591 (2007); see also Kimbrough v. United States, 128 S. Ct. 558, 564 (2007) ("[U]nder Booker, the cocaine Guidelines ... are advisory only, and ... the Court of Appeals erred in holding the crack/powder disparity effectively mandatory."): cf. Spears v. United States, 129 S. Ct. 840, 843 (2009) (allowing district courts to adopt and apply "replacement ratios" instead of the 100:1 crack/powder cocaine ratio from Guidelines). 332. Nor, frankly, is the prospect of a legislative or regulatory reaction to a court ruling strictly a problem with Apprendi: Governments are perpetually adjusting laws and regulations in response to judicial opinions. I would also note that part of the problem is that the Eighth Amendment has become a practical nullity, applying only to prison 2009] HEINOUS, ATROCIOUS, AND CRUEL 957 There might also be some due process limits to a state's choices after all. Current cases protect the imposition of the sentence (Apprendi) but not its result, and they protect the result of parole release (Morrissey) but not its imposition. 333 Ultimately, it seems absurd to hold that due process obtains during the sentencing process, which governs the prospective duration of restrained liberty, but that it is no restraint on the actual length of time those in prison will spend there. Nor does it make sense to hold that a parolee enjoys due process protections once she has been conditionally released, but that she has no protections during the hearing that determines whether or when she will be granted that conditional release. 334 An offender cannot have inherent due process interests in both of these stages but not in parole suitability itself''' IV. THE STANDARD OF PROOF AND THE DEFINITION OF AN OFFENSE Having discussed which bodies may find facts in Part Il, and what kinds of punishment they may impose on the basis of those facts in Part Ill, I now arrive at Part IV and the final part of the analysis: what standard of proof governs factfinding. The primary conflict between the Apprendi line and the due process line is, again, that the former is functional, while the latter remains entirely taxonomic. Apprendi, contrary to its reputation, relies on a functional definition of "crime."336 Apprendi looks to what statutes do, not what they say they do, in order to determine whether a given fact is an element of a crime. Under the existing cases governing parole board review, however, courts do not need to look at conditions and the death penalty. If there were limits to what a legislature could decide about punishment-that is, if the Eighth Amendment were read alongside the prohibition on bills of attainder as a kind of cQuntennajoritalian limit on punishment-then any "problem" Apprendi created could be solved by resorting to the Eighth Amendment. 333. Perhaps prison is the common factor here, a singularity from which no due process can enter or exit. "What, then, happens if an incarcerated prisoner faces a new charge and get') a new trial for a different crime? The pTisoner's due process rights are surely not lessened at his new trial just because he is already in prison-but how do we then draw the line? 334. Greenhollz held that "parole ulease and parole rl!vocalion are quite different," without explaining why the rights attaching to each are different. Greenholtz v. Inmates of Neb. Penal & Corr. Complex. 442 U.S. 1. 9 (1979). Parolees can certainly form the "enduring attachments of nonnal life," and unreleased inmates are "subject to all of the necessary restraints that inhere in a prison." ld. (internal quotation marks omitted). But why pull up the ladder once someone has been gralucd release, without acknowledging that the way she was released was through the very process-a suitability hearing-to which no due process rights attach? 335. Wilkinson v. Austin, 545 U.S. 209 (2005), might support this proposal. It is obvious that anyone sentenced to prison-as respondents were in Wilkinson, id. at 218has had his liberty restrained. But Wilkinson held that parole eligibility cannot be limited "atypical[ly] and significant[ly]." Id. at 223. (Or is that arbitrarily and capriciously? See supra notes 273, 285.) 336. In fact, A.pprendi's standard of proof holding is itself a reaction against the formalist practice of stashing the clements of real crimes in sentence enhancements. See discussion infra Part N.B. 958 COLUMBIA LAW REVIEW [Vol. 109:893 what the parole board is doing. They need only see that a parole board is doing it. Once the parole board is involved, it needs only to find facts by "some evidence." A. Parole's Standard of Proof Parole board findings of fact are essentially unconstrained. Part 1V.A.I explains that the parole board need only find "some evidence" of the facts justitying a finding of unsuitability. California state courts reviewing parole habeas petitions may not independently review these facts, and federal courts are doubly constrained by "some evidence" review and deference due to state courts under the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA).337 Only if there is no evidence of unsuitability will a court overturn a parole board. Part 1V.A.2 shows how the "some evidence" standard is used in unsuitable life sentence cases to render review practically impossible. 1. Parole Suitability Is Reviewed for "Sante Evidence. " - While state and federal cases have held that the parole board's suitability decisions must be supported by "some evidence,"33B the Supreme Court has never applied the "some evidence" test to a parole suitability case. 339 The "some evidence" standard comes from the prison discipline case Superintendent v. Hill, which involved the revocation of good time credits for an inprison disciplinary infraction. 340 The policy concerns unique to prison discipline weigh heavily in favor ofjudicial restraint, especially "the legitimate institutional needs of assuring the safety of inmates and prisoners, avoiding burdensome administrative requirements that might be susceptible to manipulation, and preserving the disciplinary process as a means 337. 28 U.S.C. § 2254 (2006). 338. See. e.g., Sass v. Cal. Bd. of Prison Terms, 461 F.3d 1123, 1128-29 (9th Cir. 2006) (finding requisite "some evidence" to support parole board determination); d. In re Rosenkrantz, 59 P.3d 174, 211-12 (Cal. 2002) (finding Governor's parole decision subject to same cOl1slitutional limitations as those of parole board, including requirement that "the faclUal basis of such a decision is supported by some evidence in the record that was before the Board"). 339. Because the Supreme Court has not spoken directly to this question, Ninth Circuit judges have argued about the exact contours of the "some evidence" standard, and whether it applies in the parole context at all. One dissenter has argued that a parole board determination must be both "supported by 'some evidence' and not 'otherwise arbitrary.'" Sass. 461 F.3d at 1133 (Reinhardt,]., dissenting) (emphasis omitted). On the other end of the spectrum is an argument that looking for "some evidence" to support denial of parole gives faD lillie deference to state parole boards, not too much. Irons v. Carey. 506 F.3d 951. 956 (9th Cir. 2007) (Kleinfeld,]., dissenting from denial of rehearing en banc) ("States are entitled to deny parole and require prisoners to serve their full sentences less 'good time,' even without 'some evidence' beyond the crimes for which the sentences were imposed." (emphasis added)); see also Boss v. Quarterman, No. 07-50448, 2008 WL 5194600. at *2 (5th Cir. Dec. 12,2008) (holding Superintendent v. Hill. 472 U.S. 445 (1985). does not govern parole release review). 340. 472 U.S. at 447. I-lill's credits were revoked even though no onc saw the assault and the victim provided written testimony that Hill and the other charged inmates had not caused his injury. Id. at 447-48. 2009] HEINOUS, ATROCIOUS, AND CRUEL 959 of rehabilitation."34l While an inmate has a "strong interest in assuring that the loss of good time credits is not imposed arbitrarily" because "the loss of such credits threatens his prospective freedom from confinement by extending the length of imprisonment,"342 this liberty interest has to be "accommodated in the distinctive setting of a prison, where disciplinary proceedings 'take place in a closed, tightly controlled environment peopled by those who have chosen to violate the criminal law and who have been lawfully incarcerated for doing so.' "343 Thus, "a modicum of evidence" supporting the revocation of good time credits is sufficient to "prevent arbitrary deprivations without threatening institutional interests or imposing undue administrative burdens."344 This standard does not "require examination of the entire record, independent assessment of the credibility of witnesses, or weighing of the evidence. Instead, the relevant question is whether there is any evidence in the record that could support the conclusion reached by the disciplinary board."3·" 2. Applying "Some }<,vidence" to Unsuitable Life Sentences. - The policy considerations driving the due process analysis of prison disciplinary proceedings map poorly, if at all, onto a parole board's finding of commitment offense unsuitability. None of the institutional concerns is the same. First, unsuitable life sentences have nothing to do with in-prison discipline. The only prisoners for whom the commitment offense is dispositive are those who have already demonstrated a pattern of good behavior. Their parole is denied despite good behavior, not because of it. 346 Second, the essence of prison disciplinary proceedings is that they "take place in a highly charged atmosphere" where "prison administrators must often act swiftly on the basis of evidence that might be insufficient in less exigent circumstances."347 There is no exigency in a parole hearing, particularly if the facts before the parole board concern the commitment offense. The parole board has notice of impending suitability hearings for more than a decade (and sometimes more than two decades). There is no recent event requiring a rapid response and no "highly charged atmosphere"34s arising from in-prison violence. Third, whereas facts in a disciplinary hearing come from "a closed, tightly controlled environment peopled by those who have chosen to violate the criminal law and who have been lawfully incarcerated for doing SO,"349 commitment offense facts come from court transcripts. Hill's concerns about witness reliability and intimidation in prison do not apply to 341. Id. at 454-55. 342. Id. at 454 (internal citations omitted). For an exploration of why the revocation of good time credits might nevertheless present an Apprendi problem, see supra note 315. 343. Hill, 472 U.S. at 454 (internal citations omitted). 344. Jd. at 455. 345. Jd. at 455-56. 346. See text accompanying note 251. 347. Hill, 472 U.S. at 456. 348. Id. 349. Id. at 454 (internal citations and quotation marks omitted). 960 COLUMBIA LAW REVIEW [Vol. 109:893 criminal trials, and the trial has procedural guarantees (such as crossexamination) that improve the accuracy of such testimony.35o Ironically, it is these procedural protections that underlie judicial deference to parole board decisions in the first place. 35 ! Finally, unsuitable life sentences have nothing to do with the safety of inmates, add no administrative burdens, and, in fact, undercut the prison discipline process by demonstrating to inmates that good behavior is necessary but not sufficient to gain release. The attenuated standards by which parole decisions are reviewed exacerbates the weakness of the standard of proof.352 While the prisoners denied good time credits in Hill were granted judicial review of prison disciplinary proceedings as a statutory right,353 California grants no such right. The only state judicial review in California is habeas review, which itself applies the "some evidence" standard and precludes any independent weighing of the evidence. 354 Federal courts, in tum, also review for "some evidence,"355 and are further constrained in their ability to review facts by AEDPA. 356 This leads to the very real possibility that there is no 350. A parole board also has a more tenuous relationship to the facts than the judge, who has at least observed the trial (or the plea allocution) and seen the evidence herself. 351. See supra notc 260. 352. The power to review parole decisions is crucial in making sure that discretion is exercised sensibly. Steven L. Chanenson, Guidance from Above and Beyond, 58 Stan L. Rev. 175, 178, 188-89 (2005) ("[A]ppellate courts playa vital role in [limiting discretion] as part of the proper functioning of a sensible, guided sentencing system."); see also Berman & Bibas, supra notc 200, at 43-44 (noting that review is onc mechanism by which discretion can be channeled). 353. Hill, 472 U.S. at 459-60 (Stevens,.J., dissenting) ("Massachusetts' law, wholly apart from the Federal Constitution, provides judicial review for the correction of errors 'in proceedings which ... are not otherwise reviewable by motion or appeal. '" (internal citations omitted)). 354. See supra notes 129-132 and accompanying text. 355. See supra note 338. 356. AEDPA limits the review of habeas petitions "on behalf of a person in custody pursuant to the judgment of a State court." 28 U.S.C. § 2254(d) (2006). Federal courts often interpret this phrase to apply to parole board decisions, since the prisoner is in custody pursuant to his conviction for his commitment offense. See, e.g., Sass v. Cal. Ed. of Prison Terms, 461 F.3d 1123, 1126-27 (9th Cir. 2006). Under AEDPA, federal courts may only grant relief "with respect to any claim that was adjudicated on the merits in State court proceedings" if the state court decision was "contrary to, or involved unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States" or if the decision was "based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding." 28 U.S.C. § 2254(d). In a future paper, I will argue that none of the comity and finality concerns underlying AEDPA is in operation here. See generally, e.g., Williams v. Taylor, 529 U.S. 362, 381 (2000) (citing Teague v. Laue, 489 U.S. 288, 301 (1989» (discussing habeas concerns developed in case law and codified by AEDPA). The parole board of the state of California is not a coequal body with a federal court, a parole board hearing docs not have the procedural safeguards or factfinding accuracy of a state trial court, and state habeas review makes no findings on the merits. A state court has plenary jurisdiction to reach constitutional issues; a parole board does not. For more on this subject, sec Nancy J. King & Suzanna Sherry, Habeas Corpus and State Sentencing Reform: A Story of Unintended 2009] HEINOUS, ATROCIOUS, AND CRUEL 961 meaningful judicial review of unsuitable life sentences. 357 Thus, not only is the jury's verdict not binding on the parole process itself, there is no real way to redress the parole board's subversion of the jury finding via the courts. The "some evidence" standard's application to parole should be reconsidered independently of the Apprendi argument raised in this Article. While it makes no sense to give a parole board's finding of commitment offense facts any deference, a parole board might, indeed, deserve deference for the facts it finds about other suitability factors besides the commitment offense, such as punishing in-prison misconduct (as in Hill) or providing a forward-looking assessment of a prisoner's public safety threat at the point of release. The parole board can and should find facts about in-prison discipline and the prisoner's current psychiatric and behavioral profile. These facts need not be found beyond a reasonable doubt, for the institutional interests identified in Hill. Limiting a parole board's discretion, by removing certain facts from its consideration, might ultimately be the best way to save it. B. Apprendi's Requirement of Proof Beyond a Reasonable Doubt Apprendi requires both that the jury find facts and that it find those facts beyond a reasonable doubt. Since both of these parts of the rule involve the jury and touch on similar interests,358 they are often conDated,359 but the requirement that facts be proven Consequences, 58 Duke LJ. I, 19 (2008) (stating that nearly one in five noncapital habeas petitions filed in federal district courts challenged "the constitutionality of a decision by state corrections or parole officials regarding the administration of the prisoner's sentence" (citing Nancy J. King et al., Final Technical Report: Habeas Litigation in U.S. District Courts 26 (2007), available at http://ssrn.com/abstract=999389 (on file with the Cotumbia Law Review))). 357. State courts have, however, granted relief on the basis that there was not "some evidence" supporting a finding that a murder was heinous, atrocious, or cruel. See, e.g., In re Elkins, 50 Cal. Rptr. 3d 503, 523 (Ct. App. 2006) ("The Governor's decision reversing the Board's grant of parole on the basis of the tacts of the offense lacks 'some evidence' that granting parole posed 'an unreasonable risk of danger to society.'''). 358. Both touch on due process: The procedure due a convicted offender involves bolh a jury finding and a finding beyond a reasonable doubt. Both requirements also touch on separation of powers questions: Can the state usurp the jury's power as the final bulwark against the deprivation of liberty, or, alternatively, can the judiciary impinge upon the legislative power to structure criminal laws? 359. See, e.g., Apprendi v. New Jersey, 530 U.S. 466, 476-77 (2000). In Appendi, the Court found that: At stake in this case are constitutional protections of surpassing importance: the proscription of any deprivation of liberty without "due process of law," and the guarantee that "[i]n all criminal prosecutions, the accused shall enjoy the right to a speedy and public trial, by an impartial jury." Taken together, these rights indisputably entitle a criminal defendant to "a jury determination that [he] is guilty of every element of the crime with which he is charged, beyond a reasonable doubL" 962 COLUMBIA LAW REVIEW [Vol. 109:893 beyond a reasonable doubt is a distinct part of the Apprendi rule. 350 Apprendi's holding about the standard of proof once again uses a functional method to define crimes. Crimes are made up of facts justifying punishment, whether or not the legislature labels these facts as elements, sentencing enhancements, or anything else. AIl of these facts must be proven beyond a reasonable doubt because of the stigma and liberty deprivation at stake, and because crimes have historically required this standard of proof.3 5l 1. Apprendi Defines Crimes Functionally. - The beyond a reasonable doubt standard of proof is "a profound judgment about the way in which law should be enforced and justice administered."352 The standard of proof for a given fact depends on whether that fact is an element of a crime: The state has borne the burden of "a higher degree of persuasion in criminal cases" since "ancient times," even though "its crystallization into the fonnula 'beyond a reasonable doubt' seems to have occurred as late as 1798."353 The real issue here is what makes a fact an element. Is an element an element because the legislature calls it an element, or is an element an element because it is the functional constituent of a crime? Again, I call the statutory approach taxonomic and the operational approach functiona],354 There are potential separation of powers issues at stake in both approaches: Taxonomist~ argue that the judiciary infringes on the legislature's power to define crimes by ignoring statutory text, while functionalists argue that the jury has the right to find facts about crimes, and the legislature may not "manipulate the prosecutor's burden of proof'35s by putting elements elsewhere in the code. Apprendi represents the triumph of the functional approach, ushering in an era where operant facts-"whether the statute calls them eleId. (alteration in original) (internal citations omitted). But, of course, affirmative defenses and other facts a jury finds are not found beyond a reasonable doubt. 360. Justice Stevens wrote in Apprendi that: "[I]t is unconstitutional for a legislature to remove from the jury the assessment of facts that increase the prescribed range of penalties to which a climinal defendant is exposed. It is equally clear that such facts must be established by proof beyond a reasonable doubt." Id. at 489 (alteration in original) (inlernal quotation marks omitted) (quoting Jones v. United States, 526 U.S. 227, 252-53 (1999) (Stevens,]., concurring)). 361. Of course, this history is contested, and the beyond a reasonable doubt standard applied only to taxonomic crimes, not functional ones. Then again, "[a]ny possible distinction between an 'element' of a felony offense and a 'sentencing factor' was unknown to the practice or criminal indictment, trial by jury, and judgment by court as it existed during the years surrounding OUf Nation's founding." [d. at 478 (intcITIal citations omitted). 362. [d. (internal quotation marks and citation omitted), 363. [d. (citing In re Winship, 397 U.S. 358, 361 (1970». 361. This division is similar to the methodological disputes over the liberty interest in parole-that is, whether the statute itself creates the interest, or whether there is an inhercm liberty interest in release. See discussion supra Part IILA.I. 365. Apprendi, 530 U.S. at 475. 2009] HEINOUS, ATROCIOUS, AND CRUEL 963 ments of the offense, sentencing factors, or Mary Jane-must be found ... beyond a reasonable doubt."366 Under Apprendi, a jury must find facts that increase punishment, no matter in what section of a state's code they might be found: "[M] erely because the state legislature" deems a given fact a sentence enhancement does not mean the fact "is not an essential element of the offense."367 "[T]he relevant inquiry is one not of form, but of effect."368 Apprendi eliminated the distinction between elements of an offense and sentencing factors, "repeatedly" instructing "that the [legislature's] characterization of a fact or circumstance as an 'element' or a 'sentencing factor' is not determinative of the question 'who decides,' judge or jUly."369 Apprendi sought to restrain state legislatures from defining elements of Climes in the offense statute and authorizing punishments in sentence enhancements. 37o In Apprendi, for example, the New Jersey legislature constructed a statutory scheme that "allow[ed] ajury to convict a defendant of a second-degree offense ... [and] after a subsequent and separate proceeding ... allow[ed] ajudge to impose punishment identical to that New Jersey provides for crimes of the first degree."371 This both circumvented the jury power and allowed the state to punish offenders for crimes it did not prove beyond a reasonable doubt. Because the fact of Charles Apprendi's racial animus enabled "an increase beyond the maximum authorized statutory sentence, it [was] the functional equivalent of an element of a greater offense than the one covered by the jury's guilty verdict."372 Whether Apprendi treated the sentencing enhancement disease with an even more formalist cure is a matter of continuing debate. 366. Ring v. Arizona, 536 U.S. 584, 610 (2002) (Scalia, J., concurring). These facts have to be related to punishment-that is, stigmatic loss of liberty. As the AptJTendi Court noted: [f a defendant faces punishment beyond that provided by statute when an offense is committed under certain circumstances but not others, it is obvious that both the loss of liberty and the stigma attaching to the oJTcnse are heighu'ned; it necessarily follows that the defendant should not-at the mornen t the State is put to proof of those circumstances-be deprived of protections that have, until that point, unquestionably attached. 530 U.S. at 484. 367. Apprendi, 530 U.S. at 495. 368. Id. at 494. 369. Ring, 536 U.s. at 604-05 (citing Appmuli, 530 U.S. at ·192, ·194, 495, 501). Of course, this passage conflates the standard of proof with the body that decides, assuming that a judge necessarily uses a preponderance standard of proof and a jury necessarily uses a beyond a reasonable doubt standard. 370. See United States v. Booker, 543 U.S. 220, 236 (2005) (Stevens,.J., opinion of the Court) ("As the enhancements became greater, the jury's finding of the underlying crime became less significant. "). 371. Apprendi, 530 U.S. at 491. 372. Id. at 494 0.19. COLUMBIA LAW REVIEW 964 [Vol. 109:893 Apprendi's standard of proof holding relies primarily on In re Winship,373 which itself takes a functionalist approach. 374 Although Apprendi (and earlier cases) applied Winship to criminal cases, Winship himself was a juvenile,375 charged not with a crime, but rather with "an act which would constitute a crime if committed by an adult."376 Winship did not face a prison term but was sentenced indeterminately to a training school for an eighteen-month term "subject to annual extensions ... until his 18th birthday."377 The New York Court of Appeals used a taxonomic approach in upholding the preponderance of evidence standard used to deem him a delinquent: Because a delinquency adjudication was not (taxonomically) a conviction, delinquency status was not (taxonomically) a crime, and the proceedings were not (taxonomically) Climinal,378 the beyond a reasonable doubt standard of proof did not apply. The Supreme Court reversed,379 looking beyond "the civil label-ofconvenience" to the functional effect of the proceeding. 38o Despite the rehabilitative interests of the juvenile system,381 the state could not subject "the child to the stigma of a finding that he violated a criminal law and to the possibility of institutional confinement on proof insufficient to convict him were he an adult."382 The Winship Court thus functionally analogized Winship's behavior-and his juvenile charge-to a crime, and held that Winship's due process rights were those that attached to (taxonomic) criminal proceedings. 373. '" [T] he Due Process Clause protects the accused against conviction except upon proof beyond a reasonable doubt of every fact necessary to consdtute the crime with which he is charged.''' Id. at 477 (alteration in original) (quoting In re Winship, 397 U.S. 358, 364 (1970». 374. 397 U.S. at 365-66 ("[C]ivillabels and good intentions do not (hemselves obviate the need for criminal due process safeguards in juvenile court" ...."). Winship used a functional analysis to consider the civil juvenile proceeding criminal; Apprendi used another functional step to apply Winship's holding to sentence enhancements. 375. Id. at 360. 376. Id. at 359. Winship stole $112, an act, which "if done by an adult, would constitute the crime or crimes of Larceny." Id. at 360. 377. Id. at 360. 378. Id. at 365. 379. [d. at 361. 380. Id. at 365-66 (internal quotation marks omitted) (explaining that criminal due process safeguards may be necessary in juvenile courts, even if adjudication is called civil, where child faces loss of liberty and the stigma of being adjudicated a delinquent). 381. The New York court had "attempted to justify the preponderance standard on the related ground that juvenile proceedings are designed 'not to punish, but to save the child.'" Id. (internal citations omitted). This illustrates, again, that the meaning of punishment is implicit in the Apprendi line. The Kansas Supreme Court recently held that juveniles have a right to trial by jury, since juvenile sentences are no longer rehabilitative. See In re L.M., 186 P.3d 164, 170 (Kan. 2008). 382. Winship, 397 U.S. at 367. For a discussion of the use of similar language in Apprendi, see supra notes 179-183 and accompanying text. 2009] HEINOUS, ATROCIOUS, AND CRUEL 965 The Supreme Court teased out Winship's boundaries in a series of cases beginning with Mullaney v. Wilbur'83 and ending with Jones v. United States, the case that first formulated Apprendi's standard of proof holding. 384 By the time Apprendi adopted jones's formulation of the standard of proof the Court had completely assimilated a rule from a civil case with an indeterminate sentence (Winship) into a line of criminal cases with determinate sentences. Apprendi is often accused of formalism because its rule can be circumvented by making binding guidelines advisory. Although accusations of formalism against Apprendi are now commonplace, Apprendi itself was a reaction to a different kind of formalism.'85 Without a functional constraint on the legislature's power to define crimes, a state could, conceivably, cut and paste all elements of climes into the sentencing enhance383. 121 U.S. 68,1 (1975). Mullaney supported the functionalist approach, holding that the state cannot "circumvent the protections of Winship merely by 'redefin [ing] the elements that constitute different crimes, characterizing them as factors that bear solely on the extent of punishment. '" Apprendi v. New Jersey, 530 U.S. 466, 485 (2000) (alteration in original) (quoting Mullaney, 421 U.S. at 698). Winship's due process protections were ultimately "concerned with substance, rather than, . ,fonnalism." Mullaney, 421 U.S. at 699. Patterson v. New York subsequently limited Mullaney to instances where the legislature created a presumption against the defendant, upholding a New York murder statute functionally very similar to the Maine law at issue in Mullaney. 432 U.S. 197, 214-15 (1977). Uustice Stevens, a later proponent of functionalism-and Apprendi's authorjoined the Paller-son majority in a 5-3 decision. Id. at 197.) McMillan v. Pennsylvania, 477 U.S. 79 (1986), was the taxonomists' high-water mark. upholding a "sentencing factor" that imposed mandatory minimum sentences for offenders found, by a preponderance of the evidence, to have '''visibly possessed a firearm' in the course of committing one of the specified felonies." Apprendi, 530 U.S. at 486 (quoting AfcMillan, 477 U.S. at 81-82). Pennsylvania could have chosen to include gun possession as an element of the specified felonies, but did not. The Court cited Pallerson for the proposition that "we should hesitate to conclude that due process bars the State from pursuing its chosen course in the area of defining crimes and prescribing penalties." McMillan, 477 U.S. at 86. This time, however,Justice Stevens joined the dissent in a 5-4 decision. Id. at 80. 381. 526 U.S. 227, 213 n.6 (1999) ("[A]ny fact (other than prior convictiou) that increases the maximum penalty for a crime must be charged in an indictment, submitted to a jUly, and proven beyond a reasonable doubt."). But because the prior cases "suggest[cdJ rather than establish [ed] this principle," the Jones Court did not so hold. Id. Instead, after noting the tension in the cases following Winship, id. at 240-43, the majority chose, as a means of avoiding "serious constitutional questions," id. at 251, to read the single federal carjacking statute as three separate offenses, id. at 229. Jones also referred to the factfinding issues in the capital cases Walton v. Arizona, 497 U.S. 639 (1990), overruled by Ring v. Arizona, 536 U.S. 581 (2002), and Spaziano v. Florida, 168 U.S. 117 (1981), as raising issues "too significant to be decided without being squarely faced." 526 U.S. at 251. For a more in-depth discussion of Spaziano, especially Justice Stevens's dissent, sec supra notes ]70-]74 and accompanying text. 385. Pallerson, which embraced wholly the legislature's taxonomic power to define crimes, was itself criticized as "indefensibly formalistic" and "a rather simplistic lesson in statutory draftsmanship," language that today would be associated with AttfJrendi's critics. Pattersim, 132 U.S. at 221 (Powell, J, dissenting). 966 COLUMBIA LAW REVIEW [Vol. 109:893 ment section of its code. Only the "formal" finding of guilt for the stripped-down crime would require proof beyond a reasonable doubt; all subsequent findings of fact about enhancements-the very facts that determine an offender's ultimate punishment-would require only a preponderance of evidence. Whether this kind of sentence enhancement formalism is more dangerous than Apprendi's binding guidelines formalism depends on how one models the political pressures that legislatures are subject to. The majority in Apprendi cited Justice Powell's Patterson dissent on this key point, arguing that element-to-enhancement relabeling hides true criminal punishments in a way that renders legislators less accountable: Our rule ensures that a State is obliged "to make its choices concerning the substantive content of its criminal laws with full awareness of the consequences, unable to mask substantive policy choices" of exposing all who are convicted to the maximum sentence it provides. So exposed, "the political check on potentially harsh legislative action is then more likely to operate."3S6 2. Apprendi Functionalism and Unsuitable Life Sentences: Are Parole Board Unsuitability Facts Elements of Crimes? - Under the functional approach, commitment offense facts used to deny parole fall under Apprendi's ambit-they are "elements" of "crimes," operating to increase punishment, even though the legislature has put parole suitability in a different part of the code. Apprendi emphasizes that elements of a crime are not those facts a legislature labels as elements; elements are facts which functionally deprive one of liberty through increased punishment. "If a State makes an increase in a defendant's authorized punishment contingent on the finding of a fact, that fact-no matter how the State labels it-must be found by ajury beyond a reasonable doubt."387 Under the Apprendi analysis presented in this Article, parole board findings of fact increase the punishment administered to an offender just as surely as a judge finding a fact about racial animus, aggravated kidnapping, or any sentencing enhancement does, and it does so without a state making clear choices about the substance of its criminal laws. The only argument that Apprendi does not cover parole board resentencing is one extrinsic to the Apprendi line itself: that rights at sentencing extend only to the judicial pronouncement of the prospective sentence, not the administration of the actual sentence itself. This, too, is inconsistent with Apprendi's 386. Appendi, 530 U.S. at 490-91 n.16 (quoting Patterson, 432 U.S. at 228-29 n.13 (Powell,]., dissenting)); see also Patterson, 432 U.S. at 223 (Powell,]., dissenting) (asserting that allowing legislature "to shift, virtually at will, the burden of persuasion, .. so long as it is careful not to mention the nonexistence of that factor in the statutory language that defines the crime," weakens the political pressures necessary to prevent abuse). 387. Ring, 536 U.S. at 602. 2009] HEINOUS, ATROCIOUS, AND CRUA'L 967 functional approach, since it suggests that in this one area, labels do, in fact, matter more than functions. 3ss These limits are the heart of the interests the Apprendi line protects. In Blakely, Justice Scalia highlighted their effects on notice. Without the protections of a jury and a high standard of proof, a defendant, with no warning in either his indictment or plea, would routinely see his maximum potential sentence balloon from as little as five years to as much as life imprisonment, based not on facts proved to his peers beyond a reasonable doubt, but on facts extracted after trial from a report compiled by a probation officer who the judge thinks more likely got it right than got it wrong. 389 Second degree murderers see their punishments balloon under worse circumstances-the parole board is operating at a greater remove from the trial than the probation officer, and the parole board need only find facts with a modicum of evidence, not the relative stricture of "more likely right than wrong."390 Notice is not the only, or even the primary, interest Apprendi protects, however, and parole board resentencing infringes on these interests as well. Apprendi is ultimately concerned with proportionality of punishment-not just what punishment an offender expects, but what punishment she deserves. Part of the reason the Court scrutinized the NewJersey sentencing law in Apprendi was that the jury's verdict did not serve to limit the punishment the judge could impose. This meant that the distinct degrees of offense the legislature codified in the statute collapsed into a single, more serious crime. The "scheme ... allows a jury to convict a defendant of a second-degree offense ... [and] after a subsequent and separate proceeding, it then allows a judge to impose punishment identical to that New Jersey provides for crimes of the first degree . . . ."391 Preserving the degrees of an offense is important not just because it maintains the legislature's power to structure its criminal statutes, but because, in some absolute sense, the degrees in the statute correspond to the moral opprobrium attached to particular crimes. Parole board resentencing eradicates any proportionality of punishment. Not only do second degree and first degree murderers serve similar amount, of time for different degrees of an offense, they do so because a parole board-not a jury-finds their Cl;mes similarly iniquitous (i.e., heinous, atrocious, or cruel) . 388. For a stronger argument that Apprendi is not confined to labels like sentencing phase and guilt phase, but is, instead, just concerned with crime and punishment, see supra text accompanying notes 224-226. 389. Blakely v. Washington, 542 U.S. 296, 311-12 (200·'1). 390. In Mullaney, Justice Powell described as "intolerable" the idea that "a defendant can be given a life sentence when the evidence indicates that it is as likely as not that he deserves a significantly lesser sentence." Mullaney v. Wilbur, 421 U.S. 684, 703 (1975). 391. "4pprendi, 530 U.S. at 491. 968 COLUMBIA LAW REVIEW [Vol. 109:893 Parole board factfinding thus runs afoul not only of the functional approach, but of the taxonomic approach as well. The California homicide statute defines killing in a heinous, atrocious, or cruel manner as one of the taxonomic elements of first degree murder with special circumstances, punishable by LWOPP or death. When the parole board finds identical facts in a parole eligible murderer's suitability hearing and denies parole on that basis alone, the board has thwarted the structure of the murder statute. 392 Parole board findings of fact thus have something for both sides to abhor. Under either a taxonomic or functional standard, the parole board cannot find heinous, atrocious, or cruel facts about the commitment offense without violating both the letter and the spirit of the law. Heinous, atrocious, or cruel murders are first degree offenses with special circumstances punishable by parole ineligible sentences. A second degree murder is necessarily one that is not heinous, atrocious, or cruel and cannot, of its own accord, foreclose parole. But an unsuitable life sentence is like Jack Nicholson's chicken salad sandwich in Five Easy Pieces: 393 fifteen years to life with the possibility of parole, hold the possibility of parole. V. POLICY CONSIDERATIONS If commitment offenses were no longer a factor in suitability determinations, the focus of the parole system could shift to inmates' behavior in prison. Parole boards arc both authorized and ideally situated to assess an offender's threat to public safety by looking at his disciplinary record, participation in rehabilitative programs, and mental health. 394 Indeed, evaluating these factors is closer to the heart of parole, since it both aids in the administration of prison populations (motivating prisoners to be well-behaved) and promotes public safety (preparing inmates for their return to society and evaluating whether they are ready to be released). The parole board has the institutional competence to make these assessments and is, crucially, examining behavior after the judicial imposition of the sentence, facts which were obviously unavailable to the jury (or the judge) at trial. Eliminating this retrospection in the evaluation of suitability might reorient the proceedings away from the crime and toward the inmate's readiness to reenter society. This orientation would emphasize public safety over desert and create incentives for prisoners to participate in rehabilitative programming, moving the entire system toward release and reintegration. Foreclosing the possibility of parole, regardless of how many positive steps an inmate takes while incarcerated, removes a prisoner's incentive both to behave and to treat underlying criminogenic risk 392. This is so even if ineligibility for parole is impermanent (albeit indefinite). 393. Five Easy Pieces (Columbia Pictures 1970). 394. Cal. Code Regs. tit. 15. § 2402(c) (2005). 2009] HEINOUS, ATROCIOUS, AND CRUEL 969 factors, ultimately damaging both prison administration and public safety. Even though a prisoner serving an indeterminate sentence might never be released, a climate of disorder and violence impedes the progress of the thousands of other inmates who, thanks to determinate sentencing, will be released, no matter what kind of threat they pose to public safety. A focus on release would also reframe the political discussion about parole by placing an outer limit on the punishment a given crime deserves, one that is immune to parole board reconsideration. A state would have to acknowledge that an offender's debt to society cannot perpetually be reassessed; after a while, an offender's debt to society has, in fact, been paid. Alternatively, if an offender's societal debt is to be reassessed, it should be the jury who does it, working in concert with the judiciary and legislature, not a body of political appointees, in prison, whose decision will almost never be scrutinized or stringently reviewed. 395 Shifting the focus toward public safety means that we can start to have discussions about who is ready to get out, not who deserves to get out. States would also have an opportunity to improve how they assess an offender's threat to public safety. California currently screens paroleesmostly those released from determinate sentences into mandatory parole-for twenty-six crime-related factors and assigns them a risk level based on their potential for recidivism. 396 The screening mechanisms already employed by the state (and nearly 500 other correctional agen395. Onc other option might be to establish a reentry jury for parole suitability. These juries would hear expert testimony from corrections officials and psychiatrists about an offender's threat to public safety, and the jury could decide whether an eligible inmate was, in fact, rcady for parole relca5c. Involving citizens in release decisions would acknowledge that reentry requires not only an offender willing to rcenter the community, but a community willing to accept him. The jUl)" as the voice of the community, would find all relevant fact'i and pass judgment. This time, however, its sentence would be a positive one announcing that an offender was ready to return. Community acknowledgement that an offender's debt has been paid would complete the retributive circle. Such a jury would not be limited in its ability to find all relevant facts, including ones about the commitment offense. This proposal embodies what Kevin Reitz calls the "approach" method to Apprendi compliance (e.g., Kansas's establishing separate jUlies to find all facts triggering sentence enhancements). Reitz, supra note 3, at llOS; see also supra note 308 (disclLssing critique of Blake~'V). Many issues would need to be addressed to flesh out the reentry jury proposal, including what administrative and budgetary resources it would use and whether it would replace traditional parole board determinations entirely or in part. The loss of parole board "expertise" itself might be minimal since parole board members are chosen without any requirements that they possess any particular expertise, and many of them face political pressure to find prisoners unsuitable. Juries drawn from the population at large might benefit from not being repeat players. Mter all, criminal (and tort) verdicts pilloried in public always make sense to the jurors themselves. Of course, ajury has no institutional obligations to release anyone, and California might find its prisons even more crowded. Given that so few parole eligible prisoners are released now, however, it is difficult to see how much more infrequent parole release could get. 396. Petersilia, Parole, supra note 93. 970 COLUMBIA LAW REVIEW [Vol. 109:893 cies)397 could also be used to determine when a lifer is ready to be released. Arguably, such a practice would provide more reliable indicia of an offender's threat to public safety than the parole board's judgment. 39B But perhaps the problem is not with parole itself, but with its implementation, and a lack of Apprendi analysis is not the cure for what ails the system. As Douglas Berman and Stephanos Bibas have observed, "formalistic constitutional doctrines alone cannot effectively promote jury involvement or prompt sound legislative responses."399 Removing the commitment offense from the parole board's consideration will not solve all the problems with parole board administration. States do not require particular qualifications to serve on parole boards. 40o Parole boards are notoriously arbitrary, even when not considering commitment offenses" ol Parole boards also face considerable political pressures to keep inmates in prison"02 If an Apprendi analysis results in greater discretion during suitability determinations (e.g., by making binding regulations ad397. Id. 398. See Ian Ayres, Super Crunchers 118 (2007). Ayres focuses on the dangers that arise when decisionmakers override statistical indicators of risk. Id. at 123. This is not at issue in California, however, where the parole board ignores statistical indicators of safety in deciding that practically everyone is too "dangerous" to release. See In re Dannenberg, 104 P.3d 783. 811 (Cal. 2005) (Moreno, j.. dissenting) (finding "no indication that the Board exhibits particular expertise regarding which prisoners constitute a threat to public safety or are othenvise suitable for parole," since it ignores the uncontradicted findings of psychiatric experts); see also Monahan, supra note 250, at 405-13 (arguing actuarial tools are generally superior to individuals employing a clinical approach). For statistics on parole release rates in California, see supra text accompanying notes ]44-]5l. 399. Berman & Bibas, supra note 200, at 38. 400. Rothman, supra note 84, at ]62; see also Mauer ct a1., supra note 34, at 30 (finding that "nvo-thirds of states maintain[ ] no standards for professional qualifications" for parole board staft). 401. Abramsky, supra note 55. A press secretary at the Department of Corrections and Rehabilitation described a process that almost defines arbitrary and capricious, describing the relevant factors in determining suitability as a mixture of instinct, body language, "[w]hat they say" and "[w]hat you hope they say but they don't. You can't presume that because one commissioner granted a date, another commissioner hearing the same case will do that. It's not a mathematical formula." Id. Treating similar cases dissimilarly is one of the hallmarks of capriciousness. At the CDCR, however, it's a policy. 402. Exact data on this subject is hard to come by, but at least one parole board member, appointed in part because her father and stepmother were murdered and she was supposed to bring a victims' rights perspective to the parole board, said that she was relieved of her post because she granted "too many" parole releases-twdve cases out of three hundred. Julia Reynolds, Parole Board Members Feel Pressure, Monterey County Herald, Oct. 9, 2007, at A]; see also Mauer et al, supra notc 34, at 30 ("The politicization of parole . . . renders the prospect of a rational and empirical consideration of each individual's application for parole unlikely, with future electoral concerns more likely to guide decisions."). Perhaps the fact that the parole board feels political pressures means that the system is getting the feedback it needs. I might argue, however, that the pressure of victims' groups represents a market failure/tragedy of the commons situation (that is, while the population at large has a greater stake in the administration of justice than victims' groups, the groups concentrate their power in a readily identifiable, politically powerful way), but that is outside the scope of this Article. 2009] HEINOUS, ATROCIOUS, AND CRUEL 971 visory), this arbitrariness is likely to increase. 403 But at least the parole board will be doing a job which it was designed to do, not redoingpoorly-a job the jury can handle just fine. CONCLUSION The Apprendi right limits the government's power to punish. Current policy allows the parole board to second-guess the jury, deeming a crime sufficiently heinous, atrocious, or cruel to deny suitability for parole even when a jury did not. Because tile "some evidence" standard is so deferential, a prisoner cannot effectively revisit this assessment in court. Even if a prisoner in California has no right to a finding of suitability, he has the right to have the limits of his conviction mean something. The parole board should, therefore, not consider the commitment offense in determining a prisoner's suitability for parole. Admittedly, second degree murder cases present difficult questions. The murders are often brutal and committed for trivial reasons. 404 Apprendi himself was an unsympathetic criminal (an admitted racist), as was Blakely (a wife beater). The unsavory nature of the crime is a heavy thumb on the scale, but this impulse is what the rule of law guards against. The parole board was not created to pass judgment on whether the crime was reprehensible-society, acting through the jury, has already expressed its opinion with a lengthy prison sentence. The California parole regulations "contemplate that an inmate may be deemed suitable for release even though his offense demonstrated 'exceptionally callous disregard for human suffering. "'405 And "all second degree murders by definition involve some callousness-i.e., lack of emotion or sympatlly, emotional insensitivity, indifference to the feelings and suffering of others."406 The punishment has already been inflicted; the parole board's role is only to assess whether the prisoner can be returned safely to society. California, through its legislature, has said that fifteen years is a sufficient penalty to serve for second degree murder. A jury may find facts justifying another punishment, or the legislature can change the punishment for a given set of behaviors. Society or the jury must make its decision up front, however, rather than sneaking retribution in through the 403. For a discussion of why capricious parole would have no policy appeal and is therefore more of a theoretical possibility than one likely to be enacted, see supra Part III.B.3. 404. See. e.g., Irons v. Carey (Irons II), 505 F.3d 846, 849 (9th CiL 2007) (relaying factual background of murder, where potential parolee shot housematc twelve times, stabbed him twice in the back, wrapped him in a sleeping bag for ten days, and finally dumped his body in the Pacific Ocean, all because he thought his housematc was dealing drugs and stealing from the home). 405. In re Scott (Scoll I), 15 Cal. Rptr. 3d 32, ·t6 n.1l (CI. App. 2004) (quoting Cal. Code Regs. til. 15, § 2402(c)(I)(D) (2003)). 406. In re Smith, 7 Cal. Rptr. 3d 655, 673 (CI. App. 2003). 972 COLUMBIA LAW REVIEW [Vol. 109:893 back door. If the inmate deserves a greater punishment than that currently prescribed in the second degree murder statute, society needs to establish that greater punishment through open deliberative processes, "'with full awareness of the consequences, unable to mask substantive policy choices' of exposing all who are convicted to the maximum sentence it provides."40? Unsuitable life sentences do none of these things. They do not measure rehabilitation, and they do not voice community disapproval. All they provide is some weak incapacitation effect, without any political checks and balances. 408 Although "[n]o ideal, error-free way to make parole-release decisions has been developed,"409 the role of the parole board is, properly, to make an assessment of the current state of the criminal, not a past judgment of his crime. That judgment is rightly, and constitutionally, left to the jury. The facts about a commitment offense underlying an unsuitable life sentence were known at the time of trial. The jury could have found these crimes heinous enough to warrant a sentence of life without the possibility of parole, but did not. The parole board may not revisit this judgment without violating a prisoner's Sixth Amendment and due process rights under Apprendi v. New Jersey. 407. Apprendi v. New Jersey, 530 U.S. 466, 490-91 n.l6 (2000) (quoting Patterson v. New York, 432 U.S. 197, 228-29 n.13 (1977) (Powell, j., dissenting)). 408. The incapacitation effect is weak because an offender serving an unsuitable life sentence displays no indications that his release will threaten public safety other than the facts of his commitment offense-which, after more than a decade, are not strongly probative of a threat. Prisoners who pose more obvious threats to public safetycontemporary violence, mental instability, persistent antisocial attitudes-would continue to be incapacitated (but not punished) under my analysis, since these are indicia of public safety threats well within the heartland of parole expertise. 409. Greenholtz v. Inmates of Neb. Penal & Corr. Complex, 442 U.S. I, 13 (1979).