Concurrence Opinion

MERRITT, Circuit Judge,

concurring.

I concur fully in the result and in the reasoning of Judge Moore’s opinion for the Court in Sections I, II and IV. I concur in the result reached but not all of the language and reasoning of part III of Judge Moore’s opinion. I concur in the result reached in part III because I agree that the case should be remanded to the District Court for findings and conclusions on the Fifth Amendment argument raised by Woodard.

DAVID A. NELSON, Circuit Judge, concurring in part and dissenting in part.

Insofar as this court’s judgment affirms that of the district court, I concur. Insofar as the judgment of the district court is vacated and the case remanded for further proceedings, I respectfully dissent.

On the evening of June 20, 1990, a 19-year-old college student named Mani Akram was driving an Oldsmobile car down Harvard Avenue near East 175th Street in the city of Cleveland. The vehicle was equipped with custom wheels and a special stereo system.

Four young men who had decided to go “gaffling”1 in a ear stolen earlier in the evening spotted Mr. Akram’s vehicle and intentionally collided with it. Two of the gaf-flers got out of the stolen car, and one of them — identified by an eye witness as appellant Eugene Woodard — went up to the Oldsmobile and shot Mr. Akram in the chest with a handgun.

Mr. Akram, who subsequently died from the wound, was dumped on the ground and left there for dead. His car was driven away, stripped, and torched. The gafflers argued over the division of the spoils, and a witness testified at trial that he heard Mr. Woodard say “I should get the radio. I’m the one that shot the guy.” Other witnesses testified that they heard Mr. Woodard make similar statements, including, “I’m the one with the murder case,” and “if the nigger die, it’s me that shot him, so I should get the most.” See Woodard, 68 Ohio St.3d at 71, 623 N.E.2d at 76-77.

An Ohio jury found Mr. Woodard guilty of a number of crimes, including aggravated murder in the commission of a felony. The jury recommended imposition of the death penalty, and this recommendation was accepted by the trial court. Affirmances by a state court of appeals (see State v. Woodard, 1992 WL 84888 (Ohio App. 8th Dist. Apr. 23, 1992)) and the Supreme Court of Ohio followed. Mr. Woodard is now pursuing collateral remedies in the Ohio courts, as I understand it. His lawyers have said that if these efforts are unsuccessful, Woodard will seek habeas corpus relief in the federal courts.

Should these attempts to obtain collateral relief meet the same fate that the direct appeals did, Mr. Woodard plans to throw himself on the mercy of Ohio’s governor. Article III, § 11 of the Ohio Constitution, as amended effective January 1, 1996, empowers the governor “to grant reprieves, commutations, and pardons, for all crimes and offenses, except treason and cases of impeachment, upon such conditions as the Governor may think proper....” As noted in the report and recommendation issued by the magistrate judge in the present civil rights action, “the Governor enjoys almost unfettered discretion in determining whether to grant or deny clemency.” (Citation omitted.)

The manner of applying for pardons has long been subject to regulation by the Ohio General Assembly, and the constitutional amendment that became effective at the beginning of 1996 subjected the manner of ap*-389plying for commutations to such regulation as well. It is crystal clear, however, that the legislature may not prescribe substantive regulations limiting the Governor’s discretion to grant or withhold either type of clemency. “For example, the General Assembly could not ... enact a statute requiring the Governor to accept the recommendation of the [Ohio Adult Parole Authority] in the exercise of his clemency power.” State ex rel. Maurer v. Sheward, 71 Ohio St.3d 513, 520, 644 N.E.2d 369, 375 (1994).

Under Ohio Rev.Code § 2967.07, all applications for executive clemency must be made in writing to the Ohio Adult Parole Authority. Mr. Woodard has not yet submitted such an application. The parole authority (or “parole board,” as it sometimes calls itself) nonetheless scheduled a “clemency review” for Mr. Woodard to be held on September 16, 1994. In connection with the review Mr. Woodard was notified that he could have a pre-hearing interview with representatives of the parole board if he wished. This notification prompted the filing of the present civil rights suit, in which Mr. Woodard seeks a declaration that Ohio’s death penalty clemency procedure is unconstitutional.

There appears to have been nothing sinister in the parole board’s decision to initiate a clemency review before the filing of an application therefor. Under an Ohio Department of Rehabilitation and Correction regulation that became effective on July 21, 1994 — the “Death Penalty Clemency Procedure” challenged here by Mr. Woodard — the parole board must schedule a clemency review 45 days in advance of any scheduled execution date, unless that date has been stayed. The Ohio Supreme Court did not stay Mr. Woodard’s execution until after the 45-day clock had started running, so the parole board seems to have had no choice in the matter. On the day after the filing of the complaint in the present lawsuit, however,' the parties agreed to postpone the clemency review and not to reschedule it during the pendency of the suit without 30 days’ notice. The district court subsequently entered a consent order to that effect.

If and when the parole board has occasion to reschedule the clemency review, the challenged regulation will give Mr. Woodard the option of requesting a personal interview. Part IV C.2. of the regulation provides as follows:

“If the prisoner requests the personal interview, the parole board chair will appoint one or more board members or hearing officers to interview the prisoner at the confining institution. Only the prisoner and the person or persons assigned by the parole board chair shall be present at the interview.”

The regulation goes on to provide that any statement given by the prisoner at the interview shall be considered at the clemency hearing, along with all other relevant information that is available. (Such information can, of course, include affidavits, briefs, or other materials submitted by the prisoner’s counsel.)

It was the submission of the parole board, in its motion for judgment, that an executive clemency review is not an integral part of any prior judicial proceedings; that there is no constitutional right to counsel at a clemency interview conducted pursuant to the prisoner’s request; and that Ohio’s executive clemency procedure does not compel the prisoner to be a witness against himself in violation of the Fifth Amendment. Magistrate Judge Kemp and District Judge Kin-neary agreed with these submissions.' I agree with them too.

The Declaration of Independence declares it to be self-evident that all men are endowed by their Creator with unalienable rights to life and to liberty. These God-given rights are not absolute — if a Eugene Woodard takes the life of a Mani Akram in the commission of a felony, for example, the murderer obviously runs a risk that the state may take his life in return — but our Constitution prohibits a state from depriving any person of life or liberty without due process of law.

My colleagues on the panel discern two strands in the due process jurisprudence relevant to the case before us. The first strand involves life or liberty interests that would not exist at all if not created by state or *-388federal law in connection with the establishment of a procedure such as the one at issue here. The second strand involves “life” and “liberty” in the ordinary sense of those terms — i.e., life and liberty interests that are independent of any state procedure.

As to the first strand, Judge Moore’s lead opinion concludes (relying in part upon Connecticut Board of Pardons v. Dumschat, 452 U.S. 458, 101 S.Ct. 2460, 69 L.Ed.2d 158 (1981)) that Mr. Woodard has no constitutionally protected life or liberty interest in Ohio’s clemency procedure itself. This conclusion is clearly correct, in my view.

As to the second strand, the lead opinion suggests (with Judge Merritt fully concurring) that the State of Ohio has made its clemency procedure “an integral part of the ... system for finally adjudicating the guilt or innocence of a defendant.” See Griffin v. Illinois, 351 U.S. 12, 18, 76 S.Ct. 585, 590, 100 L.Ed. 891 (1956), as quoted in Evitts v. Lucey, 469 U.S. 387, 393, 105 S.Ct. 830, 834, 83 L.Ed.2d 821 (1985). The court concludes from this that Ohio’s clemency procedure must be subjected to the same type of due process scrutiny applicable to every other stage of the adjudicatory process. I have difficulty reconciling this conclusion with what the Supreme Court did in Dumschat.

The plaintiff in Dumschat was a convicted murderer serving a life sentence in the Connecticut prison system. He had repeatedly applied to the Connecticut Board of Pardons for commutation of his sentence, and the Board had rejected each application without explanation. Mr. Dumschat brought a civil rights suit requesting the federal courts to declare that the Board had violated his due process rights in denying his applications without giving any reasons for doing so. The Supreme Court refused to entertain this request, finding no constitutional liberty interest that could trigger due process protection.

It is true that the Court’s opinion in Dumschat — an opinion delivered by Chief Justice Burger — dealt largely with the question whether Connecticut itself had created a federally protected liberty interest. My colleagues’ taxonomy places this sort of question in the first strand of due process jurisprudence, not the second strand. But it cannot be supposed that the Dumschat Court simply overlooked the second strand. Every member of the Court obviously knew that, as Justice Stevens explicitly said in dissent, the liberty protected by the Constitution “is not merely ‘a statutory creation of the State.’ ” Dumschat, 452 U.S. at 469, 101 S.Ct. at 2466 (Stevens, J., dissenting), quoting Wolff v. McDonnell, 418 U.S. 539, 558, 94 S.Ct. 2963, 2976, 41 L.Ed.2d 935 (1974). Both strands of due process jurisprudence were unquestionably implicated by the facts presented in Dumschat.

Justice Stevens would have dealt with the second strand in much the same way that my colleagues have done here. As Justice Stevens saw it, the decision to grant or withhold commutation of the prisoner’s sentence in Dumschat constituted “a regular and critical component of the decisionmaking process employed by the State of Connecticut to determine the magnitude of its deprivation of the prisoner’s liberty.” Id. at 471, 101 S.Ct. at 2468 (footnote omitted). “In my opinion,” Justice Stevens continued, “the Due Process Clause applies to each step [including the step at which clemency is granted or withheld] and denies the State the power to act arbitrarily.” Id. (footnote omitted).

I can readily appreciate the logic of Justice Stevens’ dissent, just as I can appreciate the logic of my colleagues’ apparent wish that the dissent had been the majority opinion. But there are many cases where, as the old saying has it, “a page of history is worth a volume of logic” — and it is history and tradition, I take it, on which the majority of the Supreme Court relied in declining to follow the course urged by Justice Stevens in Dum-schat. “[P]ardon and commutation decisions,” as Chief Justice Burger said for the majority, “have not traditionally been the business of' courts; as such, they are rarely, if ever, appropriate subjects for judicial review.” Dumschat, 452 U.S. at 464, 101 S.Ct. at 2464, citing Meachum v. Fano, 427 U.S. 215, 225, 96 S.Ct. 2532, 2538-39, 49 L.Ed.2d 451 (1976) (footnote omitted).

Evitts v. Lucey, decided four years after Dumschat, did not purport to overrule that case — and Evitts did not involve anything *-387remotely resembling a pardon or commutation decision. The question presented in Ev-itts, rather, was whether the Due Process Clause guarantees a criminal defendant the effective assistance of counsel on his first appeal as of right. The Court answered this question in the affirmative, notwithstanding the absence of a constitutional requirement for any right of appeal at all. If a state elects to create an appellate procedure as an integral part of its adjudicatory process, the Evitts Court reasoned, the procedure must meet the demands of the Due Process Clause.

It may be true, as my colleagues suggest, that the same reasoning logically applies to any and all proceedings made available by the government beyond the first appeal as of right. But the Supreme Court has sometimes been dubious about the wisdom of pushing the logic of its decisions to the extreme. To do so can produce unintended consequences in the real world, for one thing,2 to say nothing of creating hopeless doctrinal conflicts. Evitts itself points out, citing Ross v. Moffitt, 417 U.S. 600, 94 S.Ct. 2437, 41 L.Ed.2d 341 (1974), that the constitutional right to counsel is limited to the first appeal as of right and does not extend as far as habeas corpus proceedings. Evitts, 469 U.S. at 394, 105 S.Ct. at 834-35. Yet habeas proceedings, I should have thought, form an integral part of the adjudicative system much more clearly than executive clemency proceedings do.

Acknowledging the “great distance” that lies between the trial and appeal stage, on the one hand, and the executive clemency stage on the other, my colleagues suggest that the due process protections mandated at the clemency stage “will necessarily be minimal, perhaps even barely perceptible....” But once the camel’s nose is under the tent, I suspect that it may prove difficult to keep the rest of the animal out. And if it really is true that clemency proceedings are to be constitutionalized only to a degree that is barely perceptible, perhaps this in itself casts doubt on the justification for constitutionaliz-ing such proceedings in the first place. No other federal court has done so, as far as I am aware — and generally speaking, at least, we have it on the highest authority that judicial review of clemency proceedings is simply not “appropriate.” Dumschat, 452 U.S. at 464, 101 S.Ct. at 2464.

Ill

Part III of the lead opinion, as to which Judge Merritt has concurred in result but not in all of the language and reasoning, addresses Mr. Woodard’s argument that Ohio’s clemency procedure presents death row inmates with a “Hobson’s choice” between participating in the procedure, on the one hand, and asserting a Fifth Amendment privilege against self-incrimination on the other hand. Noting that the argument “raises the specter of an unconstitutional condition,” and suggesting that there may be no “compelling state interest in requiring waiver of Fifth Amendment rights by clemency interviewees,” the opinion concludes that, absent any such state interest, “Woodard has raised a colorable unconstitutional conditions claim” that necessitates a remand of the case.

With respect, I disagree. Mr. Woodard has not raised a colorable unconstitutional conditions claim, in my view; the concerns he raises strike me as far too spectral, at this point, to warrant intervention by a federal court.

The Fifth Amendment, as applied to the states by the Fourteenth Amendment, says that “[no person] shall be compelled in any criminal case to be a witness against himself. ...” Mr. Woodard, who is represented by able counsel, is surely not unaware of this provision. And if Mr. Woodard should choose to exercise the option of speaking *-386privately to representatives of the board that will issue a recommendation on relief from the death sentence, I am aware of no reason at all to suppose that he could be “compelled” to say anything that might be offered against him in a different criminal case or in a new trial for the murder of which he currently stands convicted.

Ohio’s clemency procedures, as I read them, do not empower the parole board or its representatives to compel a prisoner who requests a personal interview to answer a single question — or, having answered some questions, to answer others that may be asked. On its face, the challenged regulation simply does not require a waiver of the Fifth Amendment right against self-incrimination; as far as the regulation discloses, a prisoner who requests a clemency interview remains free to say as much or as little at the interview as he wishes.3

It is conceivable, of course, that a prisoner who requests a personal interview could volunteer something that might tend to incriminate him. It is conceivable that evidence of what the prisoner volunteered could later be offered against him in an Ohio court. And it is conceivable that the Ohio court could rule such evidence admissible, just as it is conceivable that such a ruling could be upheld on appeal. But I think it would be “premature,” to borrow the adjective used by the United States Supreme Court in Baxter v. Palmigiano, 425 U.S. 308, 324, 96 S.Ct. 1551, 1560-61, 47 L.Ed.2d 810 (1976), for a federal court to instruct Ohio on how to deal with situations that may never materialize at all and the facts of which can only be imagined at this point.

What is it, precisely — besides writing a longer opinion — that the district court is supposed to do with this case on remand? We are told that the district court should “analyze Woodard’s specific arguments under Ev-itts ” and the principles outlined in Part II of this court’s opinion, but Part II itself would seem to have accomplished much of this task already. Perhaps the district court is to hold an evidentiary hearing to determine what percentage of prisoners granted clemency have taken advantage of the interview option. Dumschat, however, seems to suggest that this datum would be irrelevant. See 452 U.S. at 465, 101 S.Ct. at 2465. In any event, we are told that the district court should “determine more fully the nature of the APA’s clemency procedures,” consider whether the clemency interview procedure “might” impose an unconstitutional condition, and decide whether Mr. Woodard will “risk” self-incrimination if he participates in an interview. The assessment of such a risk ex ante will be a highly theoretical undertaking, it seems fair to say, and I seriously question its utility.

Mr. Woodard tells us in his complaint that he “does not contemplate that this action will have any effect on his state convictions and/or sentences (including his death sentence) _” (Emphasis in original.) If there is a theoretical risk of self-incrimination here, it is a risk relating to criminal proceedings that have not yet been initiated and may never be initiated. In this respect, it seems to me, the case at bar is precisely analogous to Baxter v. Palmigiano. And just as the relief granted by the lower courts in Baxter was held to be “premature” (at least insofar as it was not flatly inconsistent with Wolff v. McDonnell), I think it is premature for us to require the district court to consider granting relief against a risk of self-incrimination that is purely academic at this point.

I would affirm the judgment of the district court across the board.

. The term has been defined as “forcibly removing people from their cars to rob them." State v. Woodard, 68 Ohio St.3d 70, 71, 623 N.E.2d 75, 76 (1993), cert. denied, 512 U.S. 1246, 114 S.Ct. 2770, 129 L.Ed.2d 883 (1994).

. In Sandin v. Conner, — U.S. —, 115 S.Ct. 2293, 132 L.Ed.2d 418 (1995), for example, the Court pointed out that Hewitt v. Helms, 459 U.S. 460, 103 S.Ct. 864, 74 L.Ed.2d 675 (1983), had pushed the logic of Greenholtz v. Inmates of Nebraska Penal and Correctional Complex, 442 U.S. 1, 99 S.Ct. 2100, 60 L.Ed.2d 668 (1979), to a point that had "create[d] disincentives for States to codify prison management procedures in the interest of uniform treatment.” Sandin, — U.S. at—, 115 S.Ct. at 2299. Our disposition of the case at bar, similarly, could create disincentives for the State of Ohio to offer clemency petitioners an opportunity to have a personal interview with representatives of the parole board.

. Acknowledging that "we cannot be certain that the defendants would actually impose” a waiver requirement, the lead opinion suggests that for present purposes we must accept the allegations of the plaintiff's complaint as true. But the complaint merely pleads, in conclusory form, that "Mr. Woodard will be irreparably harmed in that he will be permanently deprived of his Fifth ... and Fourteenth Amendment rights against self-incrimination_" The pleading of a legal conclusion such as this does not foreclose us from making our own assessment of the law in light of all well-pleaded facts. The regulation containing the Ohio Death Penalty Clemency Procedure is properly before us, I believe, but speculation as to matters not addressed in the regulation can hardly qualify as a well-pleaded fact.

Lead Opinion

MOORE, J., delivered the opinion of the court. MERRITT, J. (p. 1194), delivered a separate concurring opinion. NELSON, J. (pp. 1194-98), delivered a separate opinion concurring in part and dissenting in part.

MOORE, Circuit Judge.

This is a § 1983 action involving Ohio’s death-penalty clemency procedures. Eugene Woodard, a death-row inmate, seeks injunc-tive relief against the Ohio Adult Parole Authority, arguing that its clemency procedures violate due process and various other constitutional rights. The district court entered a judgment on the pleadings against Woodard, finding his claims to be wholly without merit. Although the court’s analysis was not entirely incorrect, its discussion was incomplete and its conclusions, in significant respects, premature. As a result, we vacate the judgment in part and remand for further proceedings.

I. BACKGROUND

The underlying facts associated with Woodard’s aggravated murder conviction are not particularly germane to this appeal. Nevertheless, we pause to observe that Woodard apparently killed his victim during a carjacking by him and his accomplices in southeastern Cleveland. See State v. Woodard, 68 Ohio St.3d 70, 623 N.E.2d 75 (1993). Woodard was nineteen years old at the time. His conviction and death sentence were affirmed by the Ohio Court of Appeals and Ohio Supreme Court, see id., and the U.S. Supreme Court denied certiorari, Woodard v. Ohio, 512 U.S. 1246, 114 S.Ct. 2770, 129 L.Ed.2d 883 (1994).

Although the Ohio Supreme Court set October 7, 1994 as Woodard’s execution date, a stay was issued on August 24, 1994 so that Woodard could pursue state post-conviction relief. Because the execution was not stayed 45 days in advance, however, the Ohio Adult Parole Authority (APA) initiated the clemency process in his case, notifying his counsel on September 6 that a hearing would be held on September 16. In addition, the APA informed Woodard’s attorneys that if he de*-402sired a pre-hearing interview with the APA, such an interview would be available on September 9. Woodard’s counsel objected to the short notice given by the APA, pointing out that they had only recently been assigned to represent him. Counsel further sought assurances from the APA that any clemency procedures would not interfere with their representation of Woodard — namely, that counsel would be permitted to attend and participate in the scheduled interview and hearing. When the APA did not respond, Woodard filed this § 1983 action on September 14, 1994, moving additionally for a temporary restraining order against any clemency procedures in his case. According to Woodard, a premature review would serve as an obstacle to a later, more timely clemency review, because such later review could only be obtained by leave of the APA.

On September 16, a consent decree was filed, postponing the clemency proceedings indefinitely. Defendants then moved for a judgment on the pleadings, stating that Woodard could prove no set of facts to support any of his claims. The magistrate judge to whom the case had been referred agreed, noting that the Governor’s broad discretion in making clemency decisions meant that Woodard lacked any protected liberty interest in the clemency procedures which might be subject to due process requirements. The magistrate judge also found that Woodard did not have a substantive due process claim, that the clemency procedures did not violate Woodard’s privilege against self-inerimination, that Woodard did not have a right to counsel in the clemency proceedings, and that the Eighth Amendment prohibition against cruel and unusual punishments was not implicated in this case. The magistrate judge did not address Woodard’s additional equal protection and Ninth Amendment arguments. Despite objection, the district court adopted the magistrate judge’s report and recommendation in full and entered judgment against Woodard. This appeal followed.

II. PROCEDURAL DUE PROCESS

We start with what is undoubtedly the most complicated issue in this case: procedural due process. Close scrutiny of Woodard’s and amici’s briefs reveals that there are actually two kinds of procedural due process arguments to be extracted therefrom, although these arguments are hardly apparent at first glance: one relates to the deprivation of a life or liberty interest arising from the clemency procedures themselves; the other deals with the general deprivation of life or liberty that begins with the initial criminal prosecution and ends with the clemency decision. In the first type of argument, the inquiry focuses on whether a life or liberty interest exists in clemency and whether it has been deprived by the clemency procedures. In the second, the question is the degree to which the clemency procedures play a necessary role in the state’s overall scheme of prosecution and punishment. Given the opacity of these abstract formulations, it is understandable that the district court only recognized the first approach — in our view, the one less helpful to Woodard. As these concepts require further explanation, we set forth our analysis in some detail, beginning with basic doctrine.

A. Deprivation of a Life or Liberty Interest in the Clemency Procedures Themselves

1. Basic Doctrine

When a state acts to deprive someone of his or her interest in life, liberty, or property, that deprivation must take place in accordance with due process of law. See U.S. Const, amend. XIV, § 1. Three questions are thus instantly generated for any given situation: (1) is there a life, liberty, or property interest at stake here? (2) is there a deprivation of that protected interest? and (3) if so, what process is due? Sometimes, a state will do something disadvantageous to an individual by, for example, summarily terminating his employment or suddenly moving her from a prison to a mental hospital. The question whether a protected property or liberty interest exists in such instances of deprivation is the threshold inquiry, and according to prevailing doctrine, state law controls as to the existence of a property interest. See, e.g., Cleveland Bd. of Educ. v. Loudermill, 470 U.S. 532, 538-41, 105 S.Ct. *-4011487, 1491-93, 84 L.Ed.2d 494 (1985) (termination of public employment); Memphis Light, Gas & Water Div. v. Craft, 436 U.S. 1, 9-12, 98 S.Ct. 1554, 1560-62, 56 L.Ed.2d 30 (1978) (termination of utility service). As to whether there is a liberty interest at stake, however, the Supreme Court has made it clear that both state law and the Due Process Clause itself may create such an interest. See Sandin v. Conner, — U.S.—, —, 115 S.Ct. 2293, 2297-2302, 132 L.Ed.2d 418 (1995) (transfer from Hawaii prison to California prison); see also Rich-abb Fallon, Daniel Meltzeb & David Shapi-RO, Haet and WeohsleR’s The FedeRal Courts and the Federal System 557-62 (4th ed. 1996). According to the Court, a liberty interest may arise from federal law when the change in a prisoner’s situation is “not within the range of conditions of confinement to which a prison sentence subjects an individual” — that is, when it is “qualitatively different from the punishment characteristically suffered by a person convicted of crime.” Olim v. Wakinekona, 461 U.S. 238, 245, 103 S.Ct. 1741, 1745, 75 L.Ed.2d 813 (1983) (quoting Vitek v. Jones, 445 U.S. 480, 493, 100 S.Ct. 1254, 1264, 63 L.Ed.2d 552 (1980)). In addition, state procedures may create liberty interests that are deprived when a state actor deviates from these procedures. Sandin, — .U.S. at—, 115 S.Ct. at 2299-2301. Of course,- even though state law plays a role in determining the existence of property or liberty interests, the ultimate question of the degree of due process protection to be afforded under the Constitution remains a federal one.

Usually, disputes over “liberty” interests arise in the prison context, where some change in the circumstances of incarceration forms the basis for a due process claim. Although it provides the most common ground for complaint, the change does not always have to involve a transfer from one type of imprisonment to another. In Wolff v. McDonnell, 418 U.S. 539, 94 S.Ct. 2963, 41 L.Ed.2d 935 (1974), inmates challenged the prison authorities’ summary revocation of good time credits. In Connecticut Bd. of Pardons v. Dumschat, 452 U.S. 458, 101 S.Ct. 2460, 69 L.Ed.2d 158 (1981), plaintiffs sought a declaratory judgment that the Connecticut board’s commutation and pardon procedures did not comport with procedural due process.

2. The Existence of a Federally Created Life or Liberty Interest

In this case, Woodard contends that the Ohio clemency procedures are constitutionally defective under the foregoing framework. His primary difficulty, however, lies in showing the existence of a protected life or liberty interest. First, as far as the Due Process Clause itself is concerned, the Supreme Court has decisively rejected the argument that federal law can create a liberty interest in clemency. Dumschat, 452 U.S. at 464-65, 101 S.Ct. at 2464-65. According to the Court, “The ground for a constitutional claim, if any, must be found in statutes or other rules defining the obligations of the authority charged with exercising clemency.” Id. at 465, 101 S.Ct. at 2465.

Woodard tries to distinguish Dumschat by noting that the claimants in that ease were only non-capital inmates — that unlike them, he has a more powerful life interest at stake in the clemency proceedings. Although Woodard correctly points out that capital punishment is constitutionally unique and that its imposition generally requires extra procedural protections, he is unable to demonstrate how his death sentence can form a principled basis for departing from Dum-schat’s rationale. In Dumschat, the Court explained that the claimed liberty interest must derive from the state’s actions after a valid conviction and sentence. Id. at 463-65, 101 S.Ct. at 2463-65. In other words, the fact of an inmate’s incarceration does not itself give rise to a liberty interest, because it is assumed that the inmate’s “original” liberty interest — i.e., the one he possessed before he was arrested, charged, and convicted for his crime — has already been (or is still being) afforded due process: the right to trial, appeal, post-conviction remedies, and similar process. The liberty interest here, after incarceration, must come from actions taken by the state to change the prisoner’s present situation, and the question becomes whether procedural due process places any constraints on making that change. Such liber*-400ty-affecting change is easiest to see when prison authorities move an inmate from one prison to another, or from a prison to a mental hospital. It is far more difficult to identify when the parole authorities simply alter their mode of dealing with clemency applications. In the latter case, rather than a tangible, further restraint on one’s freedom, the state’s action constitutes only a change in a prisoner’s expectations concerning the likelihood of a pardon or commutation. As stated by the Court, a “felon’s expectation that a lawfully imposed sentence will be commuted or that he will be pardoned is no more substantial than an inmate’s expectation, for example, that he will not be transferred to another prison; it is simply a unilateral hope.” Dumschat, 452 U.S. at 465, 101 S.Ct. at 2465 (footnote omitted).

The same rationale applies to the life interest claimed by Woodard. Woodard’s “original” life interest — his right not to be sentenced to death — is already covered by independent process: his trial, appeal, habe-as petitions, and even (as will be discussed in Part II.B below) his clemency application. In order to argue that Ohio’s clemency procedure by itself impinges upon a protected life interest, however, he must point to a separate act that might constitute a deprivation of that interest — a separate, life-affecting change in his situation. In this ease, the change is no different from that in Dum-schat. It is just a change in expectation as to the possibility of having his sentence commuted.

Woodard also tries to distinguish Dum-schat by arguing that the asserted life and liberty interest here is not in being granted clemency, but in having “a meaningful clemency procedure.” Appellant’s Br. at 6. This argument makes the error of confusing the liberty or life interest with the process that is due. Woodard can only be entitled to a meaningful procedure, the process that is due, if he first shows the deprivation of a protected constitutional interest — life, liberty, or property. The protected interest cannot be the process itself. In Dumschat, the plaintiffs . also sought “meaningful” clemency procedures — including a statement of reasons for the parole board’s decision — but the Court made it plain that the Constitution gave them no right to such procedures until they could first demonstrate a cognizable liberty interest. Id. at 463-67, 101 S.Ct. at 2463-66.

3. The Existence of a State-Created Interest

Woodard next asserts that even if there is no federally created life or liberty interest in clemency, Ohio’s statutory scheme creates such an interest. Again, this claim is squarely met by contrary ease law. In addition, the Supreme Court’s recent decision in Sandin v. Conner, — U.S.—, 115 S.Ct. 2293, 132 L.Ed.2d 418 (1995), does not alter our view in this regard.

Pr e-Sandin doctrine was relatively straightforward. Under that framework, a state could create protected liberty interests “by placing substantive limitations on official discretion.” Olim v. Wakinekona, 461 U.S. 238, 249, 103 S.Ct. 1741, 1747, 75 L.Ed.2d 813 (1983). Thus, if state rules set forth mandatory procedures for prison transfers, and prison officials deviated from those procedures, there might be a deprivation of a state-created liberty interest. On the other hand, if state procedures essentially allowed “unfettered discretion” in making the ultimate decision, no constitutionally protected liberty interest would be created, even if some of the procedures leading up to the final decision might be couched in mandatory terms. Id. at 249-50, 103 S.Ct. at 1747-48; see also Sweeton v. Brown, 27 F.3d 1162, 1164-65 (6th Cir.1994) (en banc) (following Olim), cert. denied, — U.S.—, 115 S.Ct. 1118, 130 L.Ed.2d 1082 (1995). In all material respects, this case is indistinguishable from Olim. Although Woodard and the ACLU place great emphasis on the fact that the Governor of Ohio must wait for a recommendation from the APA before making a decision,1 the critical point, which even *-399Woodard acknowledges, is that the Governor “is not bound by the Ohio APA’s recommendation.” Appellant’s Br. at 23. Under Olim, it does not matter whether the APA must follow elaborate, mandatory procedures, and it does not matter whether the APA’s recommendation is itself discretionary or tightly regulated. The only thing that counts is that there are “no standards governing the [final decisionmaker’s] exercise of his discretion.” Olim, 461 U.S. at 243, 249-50, 103 S.Ct. at 1744, 1747-48. In Ohio, the final decision-maker’s discretion is entirely “unfettered.”

Sandin has now modified this analysis, although its ultimate effect on the pre-exist-ing framework remains uncertain. Faced with a situation involving mandatory prison regulations, the Supreme Court in Sandin decided to change course, holding that the mandatory/discretionary distinction which had been developed in prior decisions would no longer be controlling. According to the Court, the ease law delineating the mandatory/discretionary rubric “has produced at least two undesirable effects. First, it creates disincentives for States to codify prison management procedures in the interest of uniform treatment_ Second, the ... approach has led to the involvement of federal courts in the day-to-day management of prisons, often squandering judicial resources with little offsetting benefit to anyone.” Sandin, — U.S. at—, 115 S.Ct. at 2299. As a result, the Sandin Court formulated a new approach, which continued to “recognize that States may under certain circumstances create liberty interests which are protected by the Due Process Clause,” but cautioned that these interests would “be generally limited to freedom from restraint which, while not exceeding the sentence in such an unexpected manner as to give rise to protection by the Due Process Clause of its own force, ... nonetheless imposes atypical and significant hardship on the inmate in relation to the ordinary incidents of prison life.” Id. at —, 115 S.Ct. at 2300 (citations omitted). Stated differently, the prison authorities’ actions would need to “present a dramatic departure from the basic conditions of [the inmate’s] sentence.” Id. at—, 115 S.Ct. at 2301.

Notably absent from the Court’s opinion is any explanation as to how a state might actually create these liberty interests. As a result, it remains unclear whether the Court intended to eliminate the mandatory/discretionary distinction altogether, looking to alternate means of protecting “atypical and significant” hardships under state law, see Sandin, — U.S. at—, 115 S.Ct. at 2306 (Breyer, J., dissenting) (expressing uncertainty as to the majority’s intent), or whether the Court simply intended to set up a further requirement for forming state-law liberty interests, making mandatory language in state procedures a necessary but no longer sufficient condition in finding such interests. See id.; RICHARD Fallon, Damel Meltzer & David Shapiro, Hart and Weohsler’s The Federal Courts and the Federal System 560-61 (4th ed.1996). Under the latter interpretation, there would be no change to our analysis of Woodard’s situation under Olim, because the final decisionmaker’s discretion in Ohio remains uncabined. Under the former interpretation, we would no longer be able to rest our holding on Olim, but we would still conclude that Woodard is not entitled to relief. Even if state law can recognize and create liberty interests through means other than mandatory or discretionary language, it cannot be said that a change in a prisoner’s expectations regarding the likelihood of being granted clemency may be deemed an “atypical and significant hardship on the inmate in relation to the ordinary incidents of prison life,” Sandin, — U.S. at —, 115 S.Ct. at 2300, given that the Su*-398preme Court has already effectively held, in Dmnschat, that such an expectation is not “qualitatively different from the punishment characteristically suffered by a person convicted of crime,” Vitek v. Jones, 445 U.S. 480, 493, 100 S.Ct. 1254, 1264, 68 L.Ed.2d 552 (1980), and given that the Sandin Court did not even consider the harsh change in conditions presented in its own case — disciplinary, segregated confinement for 30 days — to impose an “atypical and significant hardship.” In short, we do not need to decide which of these interpretations of Sandin is the correct one, as Woodard is simply unable to prevail under either.

Because Woodard cannot demonstrate the existence of a protected life or liberty interest arising from the clemency procedures themselves, this first strand of procedural due process analysis is ultimately of little assistance to him. Also, because no such interest exists, it remains unnecessary for us to reach the question of what process might be due.

B. The Role of Clemency in the Overall Scheme of Adjudication

In our view, the district court’s focus on the foregoing strand of due process analysis, though analytically correct, as far as it goes, indicates that the court did not take account of the second strand — an approach that centers on the role of clemency in the entire punitive scheme. Under this approach, the district court should not have held that Woodard was entirely without a procedural due process claim. The court should have identified Woodard’s “original” life and liberty interests under Evitts v. Lucey, 469 U.S. 387, 105 S.Ct. 830, 83 L.Ed.2d 821 (1985), and it should then have proceeded to an inquiry of the degree of process that would be due.

As mentioned in Part II.A.2 above, Woodard has always had “original” liberty and life interests in not being imprisoned and in not being executed by the state. These interests could only be deprived with due process of law, which at a minimum would include the choice of having a trial conducted in accordance with the Sixth Amendment. See U.S. Const, amend. VI. The Constitution does not require a state also to provide a system of appeals, but if the state chooses to do so, the appeal, too, must comply with the basic requirements of due process. Evitts, 469 U.S. at 393, 105 S.Ct. at 834. As stated in Evitts: “[I]f a State has created appellate courts as ‘an integral part of the ... system for finally adjudicating the guilt or innocence of a defendant,’ the procedures used in deciding appeals must comport with the demands of the Due Process and Equal Protection Clauses of the Constitution.” Id. (quoting Griffin v. Illinois, 351 U.S. 12, 18, 76 S.Ct. 585, 590, 100 L.Ed. 891 (1956)). As this circuit has observed:

Implicit in this conclusion is the idea that the enactment of an appellate system necessarily “alter[s] the balance between the layers of adjudication in a constitutionally significant manner”_ That is, the introduction of a second tier of adjudication unavoidably affects the operation and significance of the first tier. For this reason, due process places constraints on an appeal even if it does not require an appeal in the first place. The appeal forms an “integral” and inextricable part of the procedures for determining whether a defendant should be deprived of his life, liberty, or property.

United States v. Smith, 94 F.3d 204, 207 (6th Cir.1996) (citation omitted), cert. denied, — U.S. —, 117 S.Ct. 997, 136 L.Ed.2d 877 (1997).

This reasoning logically applies to further proceedings made available by the government, as well: discretionary appeals, post-conviction proceedings, clemency procedures. The degree to which each component forms “an integral part” of the overall adjudicative ■ system determines the degree to which due process plays a role. Of course, the more distant the procedure from initial proceedings, the less stringent the requirements must be. For example, the speedy trial guarantee is well established under the Sixth Amendment, but courts' have only recently begun to recognize the right to a speedy criminal appeal under the Due Process Clause. See, e.g., Smith, 94 F.3d at 206-07. Similarly, a defendant has the right to counsel in the first appeal as of right, Douglas v. California, 372 U.S. 353, 83 S.Ct. *-397814, 9 L.Ed.2d 811 (1963), but no such right exists in post-conviction proceedings, even for capital defendants. Murray v. Giarratano, 492 U.S. 1, 10, 109 S.Ct. 2765, 2770-71, 106 L.Ed.2d 1 (1989).

Should proceedings as far down the line as death-penalty clemency procedures be analyzed under this overall framework? We believe so. Like appeals, clemency procedures are not required by the Due Process Clause, yet all thirty-eight states authorizing capital punishment make provisions for clemency. See Herrera v. Collins, 506 U.S. 390, 414, 113 S.Ct. 853, 867, 122 L.Ed.2d 203 (1993) (noting that “all 36 states” allowing capital punishment at the time had “constitutional or statutory provisions for clemency”); Kan. Const. art I, § 7, Kan. Stat. Ann. § 22-3701; N.Y. Const. art. IV, § 4, N.Y. Correct. Law §§ 261-66. Furthermore, as the Supreme Court itself emphasized in Herrera, “Clemency is deeply rooted in our Anglo-American tradition of law, and is the historic remedy for preventing miscarriages of justice where judicial process has been exhausted.” 506 U.S. at 411-12, 113 S.Ct. at 866. Indeed, the Court in Herrera explicitly relied on the existence of executive clemency in denying habeas relief to a petitioner claiming “actual innocence.” Id. at 411-17, 113 S.Ct. at 866-69. Given this clear recognition by the Court of the integral part played by clemency in every state’s death-penalty scheme, we find it only rational to apply Evitts ’s principle in equal measure to the clemency procedures. We acknowledge at the same time that the due process at the clemency stage will necessarily be minimal, perhaps even barely perceptible, because of the great distance from the truly fundamental process — the trial, the appeal — and the significant distance from other extremely important process — discretionary appeals and post-conviction petitions.2

This gives rise to the question whether an action under 42 U.S.C. § 1983 provides the appropriate vehicle for raising this type of due process claim. In our opinion, it does — more so than a habeas petition under 28 U.S.C. § 2254. In Preiser v. Rodriguez, 411 U.S. 475, 500, 93 S.Ct. 1827, 1841-42, 36 L.Ed.2d 439 (1973), the Supreme Court held that a state prisoner challenging the “fact or duration of his physical imprisonment” and seeking “immediate release or a speedier release” would be limited in his federal remedies to a writ of habeas corpus. Preiser is distinguishable from the instant case, however, because Woodard “is not claiming that he is entitled to clemency or that the [APA] should have granted his request. Instead, he is challenging only the process by which the [APA would consider] his request for clemency_” Otey v. Hopkins, 5 F.3d 1125, 1131 (8th Cir.1993), cert. denied, 512 U.S. 1246, 114 S.Ct. 2768, 129 L.Ed.2d 881 (1994). Indeed, the Court in Preiser itself suggested that a § 1983 action would be the proper approach if the prisoner were “attacking something other than the fact or length” of his sentence and “seeking something other than immediate or more speedy release — the traditional purpose of habeas corpus.” 411 U.S. at 494, 93 S.Ct. at 1838. Consequently, notwithstanding the interrelationship between successive stages of adjudication under Evitts, it is an injunction under § 1983 that provides the more suitable means of reforming the clemency procedures, where the prisoner is not directly laying claim to clemency itself, or any other form of instant relief from his sentence. As stated by the Eighth Circuit in Otéy, “This is not the stuff of redress which can be sought under § 2254.” 5 F.3d at 1131.

We come, then, to the ultimate question: how much process is due? It appears that a court must ultimately address this constitutional question by making a normative judgment about the functions of various procedures — e.g., what counts as a “trial”? what are the minimum requirements for an “appeal”? — something for which there is very little guidance by the time we arrive at clemency procedures. In Evitts, the Court used *-396the touchstone found in Douglas v. California, 372 U.S. at 358, 83 S.Ct. at 817, for the first appeal as of right: whether the appeal would be a “meaningless ritual” without the claimed procedural entitlement. 469 U.S. at 394, 105 S.Ct. at 834-35. In Evitts, the Court decided that an appeal without the effective assistance of counsel would be one such meaningless ritual. In United States v. Smith, this circuit employed the same principle in recognizing the right to a speedy criminal appeal, noting that excessive delay could also render an appeal “meaningless” for all practical purposes. 94 F.3d at 207-08.

In clemency, we are confronted with procedures that are frequently informal and almost always overwhelmingly discretionary. Once a state chooses to adopt a clemency process, what are the minimum characteristics to be expected? What goes into a baseline “death-penalty clemency procedure”? These are delicate questions, which courts— and especially federal courts — should approach with great caution and deference. Nevertheless, it is not too difficult to imagine extreme situations in which federal due process would be offended. For example, a procedure in which a governor or parole board merely pulled names out of a lottery bin or flipped coins to make clemency decisions would undoubtedly constitute a “meaningless ritual.” Similarly, a death-penalty “clemency hearing” occurring immediately after conviction but before commencement of a defendant’s sentencing proceeding would be almost equally without meaning.

In this case, Woodard argues that the APA’s current procedures provide inadequate notice and opportunity to be heard, for a number of reasons. We decline to address his specific due process arguments at this time, however, because we conclude that this case must in any event be remanded to the district court for additional proceedings on Woodard’s Fifth Amendment claim (discussed in Part III below). Because the district court did not consider this second strand of procedural due process analysis, holding instead that the Fourteenth Amendment was completely inapplicable to the state’s clemency proceedings, we believe that the district court should be given the opportunity to address this due process claim in the first instance. On remand, the district court should analyze Woodard’s specific arguments under Evitts and the principles outlined above, keeping in mind that the clemency procedure is not necessarily a formal, adversarial proceeding, and that the APA is generally free to decide what sources of information and evidence would best serve its informal investigation. Whether or not the district court ultimately finds merit in Woodard’s due process contentions, it is undeniably better equipped to be the first to evaluate them, as we are limited on this appeal to consideration of the parties’ bare pleadings.

III. THE FIFTH AMENDMENT PRIVILEGE AGAINST SELF-INCRIMINATION

Woodard claims that the APA’s clemency interview procedure burdens his privilege against self-incrimination because it provides for a private, uncounseled interview without a guarantee of immunity for the things he might say. The district court ruled that no such burden exists here because the interview is completely voluntary, an optional procedure that Woodard is free to accept or decline. In so holding, the court cited Simmons v. United States, 390 U.S. 377, 389-94, 88 S.Ct. 967, 973-77, 19 L.Ed.2d 1247 (1968), but Simmons does not provide any support for that conclusion. More important, the district court did not consider whether the interview procedure might impose an unconstitutional condition on one of Woodard’s fundamental rights.

A. The Doctrine of Unconstitutional Conditions

On appeal, the Ohio Death Penalty Task Force’s amicus brief argues that under Simmons, the APA should be required to modify its interview procedure so as not to violate the Fifth Amendment (applicable to the states via the Fourteenth Amendment). Simmons, however, involved a situation in which the defendant was forced to choose between pressing a Fourth Amendment claim and asserting his Fifth Amendment privilege against self-incrimination. According to the Supreme Court, this situation was *-395“intolerable” because an “undeniable tension” had been created between two constitutional rights. 390 U.S. at 394, 88 S.Ct. at 976-77. The instant case, on the other hand, presents no such tension — the only constitutional right at issue is Woodard’s Fifth Amendment privilege. Consequently, we are not persuaded that Simmons by itself provides a basis for reversing the district court’s judgment.

Instead, we are drawn to Woodard’s argument that the Ohio procedure calls for death-row inmates to make a “Hobson’s choice” between asserting the Fifth Amendment right and participating in the clemency review process.3 This raises the specter of an unconstitutional condition, something which the district court did not address, and something which we believe now requires us to remand the case.

As explained by a leading commentator, “The doctrine of unconstitutional conditions holds that government may not grant a benefit on the condition that the beneficiary surrender a constitutional right, even if the government may withhold that benefit altogether. It reflects the triumph of the view that government may not do indirectly what it may not do directly over the view that the greater power to deny a benefit includes the lesser power to impose a condition on its receipt.” Kathleen M. Sullivan, Unconstitutional Conditions, 102 Harv. L.Rev. 1413, 1415 (1989). In every respect, Ohio’s clemency interview procedure appears to fall within the doctrine’s parameters. In Woodard’s case, the “benefit” is the uncounseled clemency interview, and the “unconstitutional condition” is the requirement that he waive his Fifth Amendment right against self-incrimination.4 The clemency interview undoubtedly presents a serious matter for Woodard, whose life may literally depend upon it, yet in deciding between this interview and the risk of self-incrimination, Woodard has been offered a choice by the state that can hardly be deemed a constitutionally unencumbered one.

Once applied, the doctrine requires a court to employ strict scrutiny in analyzing the challenged condition. See Shapiro v. Thompson, 394 U.S. 618, 629-33, 89 S.Ct. 1322, 1328-31, 22 L.Ed.2d 600 (1969). At this stage, with little evidence before us, there does not appear to be any compelling state interest in requiring waiver of Fifth Amendment rights by clemency interviewees. Indeed, it would seem to be in the state’s own interest to encourage an open and frank interview, as far as this is possible, in order to help it decide whether mercy is ultimately appropriate. By the same token, a prisoner who still has ongoing post-conviction proceedings (such as Woodard) obviously has a measurable interest in avoiding self-incrimination. And there is always the possibility that a prisoner might be granted clemency but then be charged with different crimes that are revealed during the interview. As a result, unless a compelling reason can be brought forward which counsels against applying the doctrine, we would hold that Woodard has raised a colorable unconstitutional conditions claim regarding the interview procedure, and we would vacate the judgment of the district court.

B. Problems in Applying the Doctrine

A number of reasonable objections might be raised against finding an unconstitutional condition. Although these arguments are certainly worthy of careful consideration, in the end they do not dissuade us from favor*-394ing an application of the doctrine. We consider these arguments in turn.

1. The Clemency Interview As a “Benefit”

First, there is the idea that the only possible “benefit” to Woodard is clemency itself. Therefore, the interview procedure does not pose an unconstitutional condition because it is just an optional event — i.e., even if Woodard were given the type of interview he wanted, he still might not get clemency, in which case the interview will have been worthless. There are two possible responses to this.

One, the interview may indeed be a benefit because it dramatically increases Woodard’s chances of being granted clemency. Although we are not in a position to evaluate this hypothesis ourselves at this time, Woodard could conceivably produce evidence on remand that all, or the overwhelming majority, of the prisoners granted clemency were those who took advantage of the “optional” interview. In such event, the district court would be hard-pressed not to find a benefit in the interview.

Two, it is not entirely true that the interview is worthless without an actual grant of clemency. Giving Woodard the type of interview he desires — an open and frank discussion with no danger of self-incrimination— would give Woodard the sense that he had participated meaningfully in the process, even if he were ultimately denied clemency. On these terms, the interview procedure may also be deemed a benefit.

In short, the interview has both instrumental and intrinsic value. See LAURENCE H. Tmbe, American ConstitutioNAl Law 663-77 (1988). Whether we accept one or both of these grounds, the argument that the interview is not a “benefit” deserves, at the very least, a remand.

2. Applicability of the Doctrine in this Constitutional Context

A more difficult question is whether the unconstitutional conditions doctrine should apply in the context of the Fifth Amendment. In general, the doctrine has only consistently been applied to protect First Amendment rights. See, e.g., Board of County Comm’rs v. Umbehr, — U.S.—, 116 S.Ct. 2842, 135 L.Ed.2d 843 (1996) (free speech); FCC v. League of Women Voters, 468 U.S. 364, 104 S.Ct. 3106, 82 L.Ed.2d 278 (1984) (free speech); Sherbert v. Verner, 374 U.S. 398, 83 S.Ct. 1790, 10 L.Ed.2d 965 (1963) (Free Exercise Clause); Speiser v. Randall, 357 U.S. 513, 78 S.Ct. 1332, 2 L.Ed.2d 1460 (1958) (free speech); see also Dolan v. City of Tigard, 512 U.S. 374, 407-09, 114 S.Ct. 2309, 2328, 129 L.Ed.2d 304 (1994) (Stevens, J., dissenting) (doctrine “especially” protects freedom of speech). Nevertheless, the Supreme Court has also employed it sporadically in other areas. See Dolan, 512 U.S. at 385-87, 114 S.Ct. at 2317 (Takings Clause); Nolian v. California Coastal Comm’n, 483 U.S. 825, 107 S.Ct. 3141, 97 L.Ed.2d 677 (1987) (Takings Clause). Indeed, in Shapiro v. Thompson, 394 U.S. 618, 629-31, 89 S.Ct. 1322, 1328-30, 22 L.Ed.2d 600 (1969), the Court even held that conditioning welfare benefits on residency requirements impermissibly burdened the fundamental right of interstate travel. We are unable to discern any logical reason for holding the doctrine inapplicable to the Fifth Amendment here.

The doctrine can certainly be characterized as “unruly.” Richard A. Epstein, Unconstitutional Conditions, State Power, and the Limits of Consent, 102 Harv. L.Rev. 4, 13 (1988). As Professor Sullivan similarly observes, the Supreme Court’s recent jurisprudence seems “a minefield to be traversed gingerly.” Sullivan, 102 Harv. L.Rev. at 1415. However, it appears that much of the confusion surrounding the doctrine has focused on the “elusive distinction between withholding a subsidy and imposing a penalty.” Tribe, American Constitutional Law 781-84; see also Sullivan, 102 Harv. L.Rev. at 1439-42. In this area of confusion, the government may not penalize the exercise of a constitutional right, but the government retains the power to subsidize or encourage certain activities under its spending power. The trouble arises when by subsidizing one activity, the government is by negative implication penalizing another, and the “another” happens to involve a constitutional right. Compare FCC v. League of Women Voters, *-393468 U.S. 364, 399-401, 104 S.Ct. 3106, 3127-28, 82 L.Ed.2d 278 (1984) with Regan v. Taxation with Representation, 461 U.S. 540, 544-46, 103 S.Ct. 1997, 2000-01, 76 L.Ed.2d 129 (1983).

The same problem is not present in the instant case. We are not dealing with monetary subsidies or tax exemptions by the government. The government cannot be said here to have a legitimate purpose in limiting its “encouragement” of clemency interviews to those inmates who relinquish their Fifth Amendment rights. For that reason, this case might perhaps be described as a “purer,” and therefore easier, case of an unconstitutional condition — similar to the Supreme Court’s Takings Clause cases, where the benefits, zoning permits, are also non-monetary and (at least theoretically) unlimited. Indeed, in Dolan, the most recent Takings Clause decision addressing the issue, the Court describes the unconstitutional conditions doctrine as “well-settled.” 512 U.S. at 385-87, 114 S.Ct. at 2317. In light of this most recent pronouncement, we conclude that a degree of uncertainty in the doctrine is not a sufficient basis for resisting its application.

3. Application of the Doctrine in Light of Instances of Waiver of Rights

The final possible obstacle to application of the unconstitutional conditions doctrine lies in instances where criminal defendants have waived constitutional rights in exchange for “benefits,” without any resulting unconstitutional conditions problem. Two instances are particularly salient here: (1) the application of § 3E1.1 of the federal Sentencing Guidelines, the “acceptance of responsibility” guideline, and (2) the testimony of defendants at their own criminal trials.

First, the federal courts have continued to uphold the constitutionality of USSG § 3E1.1, see, e.g., United States v. Monsour, 893 F.2d 126, 129 (6th Cir.1990), which in many ways poses an unconstitutional condition like the one found in the instant case: the guideline gives defendants the “benefit” of a two-point or three-point decrease in offense level if they demonstrate “acceptance of responsibility” — that is, if they plead guilty and thereby give up their right to trial and privilege against self-incrimination. See United States Sentencing Commission, Guidelines Manual, § 3E1.1 (Nov.1995). How can we reconcile this acceptance of USSG § 3E1.1 with the notion that the clemency interview procedure may pose an unconstitutional condition?

Second, in criminal trials generally, defendants retain the option of testifying in their own behalf, but if they do so, they also subject themselves to cross-examination on the witness stand. To a certain extent, this situation could also be characterized in terms of an unconstitutional condition: in order to be granted the “benefit” of testifying in their defense, defendants must give up their Fifth Amendment privilege against self-incrimination. Again, the question arises whether the unconstitutional conditions doctrine must be rendered inapplicable in this context, and therefore in the similar, clemency-interview context, as well.

We should be careful to refrain from addressing issues not directly related to this case. Nevertheless, we note briefly that a reasonable argument could be made that both USSG § 3E1.1 and the trial testimony situation would actually withstand a direct unconstitutional conditions challenge, even as they impinge upon the right against self-incrimination. A court might find what appears to be an unconstitutional condition in each instance, apply strict scrutiny, but ultimately conclude that the condition is narrowly tailored to furthering a compelling government interest. With respect to the sentencing guideline, a court might identify a compelling interest in encouraging a defendant to admit responsibility, first for the sake of rehabilitating the defendant, and second to maintain administrative order in the overcrowded criminal justice system. With respect to a defendant’s testimony at trial, a court might conclude that the entire purpose of the defendant’s testimony is to serve the truth-seeking function of an adversarial, criminal trial — a purpose that would be subverted if only one side of the story were allowed to be presented. Whereas a trial is specifically designed to determine the guilt or innocence of the defendant — the very sub*-392ject matter for which the Fifth Amendment right is considered necessary — a clemency review is designed for something very different: to determine whether a prisoner is entitled to the state’s mercy — something which may or may not include consideration of the defendant’s culpability. For this reason, a defendant who testifies at her own trial may properly be deemed to have waived the right against self-incrimination, whereas no such assumption is warranted for a prisoner who chooses to speak to state officials regarding the possibility of being granted a pardon, commutation, or reprieve.

Furthermore, in the trial testimony instance, the argument could be made that no condition on any constitutional right even exists. Because the prosecution always bears the burden of proving its case, and because non-defendant witnesses are generally subject to the government’s subpoena, the fact that distinguishes a defendant from other individuals in a ease is his ability not to testify and not to be subject to questioning by the prosecution. In other words, the primary “benefit” to the defendant is the right to refuse testimony, not the ability to give it. Any benefit enjoyed by a particular defendant in speaking in her own defense is necessarily subordinate to the more important benefit of being able to remain silent while the government attempts to prove its ease. Again, that is one of the main protections the Fifth Amendment right was designed to accomplish.

Despite that protection, in Baxter v. Palmigiano, 425 U.S. 308, 96 S.Ct. 1551, 47 L.Ed.2d 810 (1976), the Supreme Court upheld the imposition of a choice between waiving the Fifth Amendment privilege to testify at a prison disciplinary hearing, and remaining silent and having an adverse inference drawn from that silence. Id. at 316-20, 96 S.Ct. at 1557-59. Though the Court did not analyze the issue in terms of a choice, but rather assessed the legality of the adverse inference, Baxter does appear on the surface to be analogous to this case. A prison disciplinary hearing, however, serves very different purposes and implicates very different interests than a clemency interview. In a prison disciplinary hearing, as in a criminal proceeding, the burden of proof is on the state; in the clemency process, however, the burden is entirely on the prisoner. A prison disciplinary hearing serves the purpose of determining whether a prisoner has violated prison rules, so that he or she may be punished, while a clemency interview serves the purpose of supporting a prisoner’s plea for mercy, so that he or she may be permitted to live or ‘receive leniency. The necessity of speaking openly to the state, therefore, is much greater in the clemency context than in Baxter, and more similar to Simmons. Baxter, therefore, is clearly distinguishable and does not affect our analysis of the clemency interview.

In the present context it is the prisoner who must demonstrate his eligibility for clemency, and so his refusal to engage in an interview can only in the most improbable of circumstances be considered a benefit. Moreover, there does not appear to be any persuasive reason why the APA should condition the benefit of an uncounseled clemency interview on the waiver of the prisoner’s Fifth Amendment right. Whereas the prisoner’s interest in trying to obtain clemency appears quite strong, the state’s interest in insisting on such a waiver appears comparatively weak. The possibility of damage to a state interest by applying the privilege seems remote, but the prisoner’s risk of self-incrimination is significant. Though' “an act for which a person has been convicted no longer tends to criminate, in the sense of the privilege,” VIII Wigmore on Evidence § 2279 (3d ed. 1940), a defendant who has not exhausted all post-conviction remedies still faces a risk of self-incrimination. If a defendant obtains a new trial, incriminating statements made during the appeal or post-conviction process, including those made during a clemency interview, might be admitted as evidence in that trial. Furthermore, an open-ended clemency interview could elicit incriminating statements regarding other acts unrelated to the prior conviction and could support new prosecutions for additional crimes. A coerced waiver of the privilege is particularly troubling here, where the state has arranged for the clemency review process to take place *-391before or during Woodard’s post-conviction proceedings in state court.5

Because this case is before the court on the pleadings, the facts have not yet been fully developed. We remand to the district court to determine more fully the nature of the APA’s clemency procedures. If it is indeed the case that Woodard must risk self-incrimination to participate in the “optional” clemency interview, the district court should apply the unconstitutional conditions doctrine.

If the court on remand does find that the clemency process imposes an unconstitutional condition, one possible resolution of the Hob-son’s choice is illustrated in United States v. Bounos, 693 F.2d 38 (7th Cir.1982). Writing for the court, Judge Posner held that individuals who were compelled to testify by a Hobson’s choice — between asserting their double jeopardy claims and retaining their Fifth Amendment rights — did not waive the privilege against self-incrimination. Id. at 40. In fact, the court held that the government’s position that they would waive the privilege by testifying in the double jeopardy hearing was “untenable in light of Simmons ....” Id. Although Seventh Circuit law did not recognize judicially created use immunity, “[t]he Fifth Amendment of its own force, without the superaddition of a judicial immunity order” would prevent the government from using the individuals’ statements against them in a subsequent prosecution. Id. Of course, such a holding in the case at hand would be premature at this point; the district court must first find the facts and determine whether an -unconstitutional condition exists.

IV. EQUAL PROTECTION AND REMAINING CLAIMS

Woodard raised an equal protection claim in his complaint, and he raises it here on appeal. Unfortunately, he failed to object when the magistrate judge did not address it in the report and recommendation to the district court. This failure is made more glaring by the fact that Woodard did object to the magistrate judge’s omission of the Ninth Amendment claim. As a consequence, it seems that Woodard has waived the claim on this appeal. See United States v. Walters, 638 F.2d 947, 949-60 (6th Cir.1981). Even if we were to address the claim on the merits, it is doubtful that we would find grounds for ruling in Woodard’s favor. First, capital inmates are not a suspect class, contrary to Woodard’s suggestion. Second, the State of Ohio could have a rational basis for treating capital and non-capital prisoners differently in clemency proceedings.

Woodard’s remaining substantive due process, Eighth, and Ninth Amendment claims are equally without merit. Woodard cannot show a fundamental right to clemency — as noted earlier, the Due Process Clause does not necessarily guarantee even a first appeal as of right. Similarly, as long as procedural due process is not offended, there can be no independent Eighth Amendment argument that the clemency procedures inflict “cruel and unusual punishment.” Finally, there is no precedent for attaching clemency rights to the Ninth Amendment.

V. CONCLUSION

The issues presented in this case are undoubtedly complicated and novel, but they also are not fully formed. As a consequence, because we conclude that Woodard has raised a Fifth Amendment argument warranting further scrutiny, we find it necessary to remand this case to the district court. On remand the district court should move beyond the pleadings stage and determine based on evidentiary materials whether Woodard must risk self-incrimination to participate in the clemency interview under the APA’s procedures; if so, the district court should apply the unconstitutional conditions doctrine as discussed in Part III. In addition, we disagree with the district court’s determination that procedural due process is irrelevant to an evaluation of the state’s death-penalty clemency procedures. On remand, *-390the district court should reconsider Woodard’s due process arguments in light of Ev-itts v. Lucey and related case law and in light of relevant evidentiary materials submitted by the parties. Finally, on all remaining arguments by Woodard, we affirm the district court.

AFFIRMED in part, VACATED in part, and REMANDED.

. In State ex rel. Maurer v. Sheward, 71 Ohio St.3d 513, 644 N.E.2d 369 (1994), the Ohio Supreme Court invalidated a pardon granted by Governor Richard Celeste, because Celeste had failed to wait for an APA recommendation as required by statute. The court simultaneously upheld eight statutorily "premature” commutations by Celeste, because the court concluded *-399that the Ohio Constitution only allowed the General Assembly to place procedural restrictions on the Governor’s pardon power, not the power to grant reprieves or commutations.

In 1995, the Ohio Constitution was amended by referendum, allowing the General Assembly to make the Governor follow additional procedures before granting commutations, as well as pardons. See Ohio Const, art. Ill, § 11.

Although the cumulative effect of these changes can only be to make the granting of clemency slightly more difficult (at least as a practical and political matter), it is clear that there is still no substantive restriction on the Governor’s discretion.

. We assume here that clemency procedures will normally fall last in line, as a final recourse for prisoners who have expended all possible effort in the courthouse, even though such procedures are certainly flexible enough to be pushed forward in line. For instance, state authorities in this case attempted to hold the clemency proceedings while Woodard’s state post-conviction petition was still in its initial stages.

. We are instructed by the Ninth Circuit that “[t]he phrase ‘Hobson’s choice’ is often used incorrectly. The phrase comes from Thomas Hobson, an English liveryman who required every customer to choose the horse nearest the door.... A Hobson's choice is thus an apparently free choice when there is no real alternative. ... Here, however, [plaintiff] technically was not confronted with a Hobson’s choice— rather, [plaintiff] was forced to choose between two harmful options.” Wang v. Reno, 81 F.3d 808, 813 n. 5 (9th Cir.1996) (citations omitted).

. Of course, at this juncture, we cannot be certain that the defendants would actually impose this condition. However, they have not denied that they would, and we must accept all of the plaintiff's allegations as true in reviewing this judgment on the pleadings. See Astor v. International Business Machines Corp., 7 F.3d 533, 538 (6th Cir. 1993).

. Though this case especially raises concerns because of the ongoing post-conviction proceedings, we note that we do not rest our holding on that fact — a prisoner who had exhausted his post-conviction remedies might still risk self-incrimination with regard to other acts, and therefore might be subjected to the same unconstitutional condition.