Brown v. United States, 365 F. Supp. 328 (E.D. Pa. 1973)
September 28, 1973
UNITED STATES of America et al., Defendants.
United States District Court, E. D. Pennsylvania.
*329 *330 *331 Allen D. Black, Harold E. Kohn, P. A., Philadelphia, Pa., for plaintiffs.
Richard R. Galli, Asst. U. S. Atty., Philadelphia, Pa., for defendants.
OPINION AND ORDER
NEWCOMER, District Judge.
Plaintiff Richard E. Brown enlisted in the United States Marine Corps on July 22, 1966, for a term of four years. On February 7, 1969, plaintiff Brown was sentenced to punitive discharge, confinement, and forfeiture of pay and allowances by a special court-martial convened by the Commanding Officer, Transient Facility, Camp Smedley D. Butler, Camp Hansen, Okinawa. This special court-martial was ostensibly convened pursuant to the authority of 10 U.S.C. § 823(a) (7) which grants special courts-martial convening authority to "the commanding officer or officer in charge of any . . . command when empowered by the Secretary concerned."
Plaintiff David L. Taylor enlisted in the Marine Corps on July 28, 1967, for a term of four years. On October 25, 1968, he too was sentenced, after trial and conviction by a court-martial also convened ostensibly pursuant to § 823(a) (7) by the Commanding Officer, Student Company, Schools Battalion, Marine Corps Base, Camp Pendleton, California. On September 10, 1969, he was sentenced by a second similar tribunal after a trial on a different alleged offense.
On May 28, 1970, the Court of Military Appeals held in U. S. v. Greenwell, 19 USCMA 460, 42 CMR 62, that the procedure used by the Secretary of the Navy to empower commanders to convene special courts-martial pursuant to 10 U.S.C. § 823(a) (7) constituted an illegal delegation of power, and that the courts convened pursuant to such a delegation were without authority.
Plaintiff Brown and plaintiff Taylor have sued, on behalf of themselves and all others tried by such courts-martial, for relief in the nature of mandamus to direct the Secretary of the Navy and the other defendants to correct the pertinent records to reflect the invalidity of such convictions, that is, to expunge any clerical reflections of such a court-martial, and also to recover the back pay and monetary value of the benefits of which persons were deprived as a result of such convictions. Both sides have moved for summary judgment.
The first issue presented by this case is the propriety of such relief in the nature of mandamus. This Court has general authority under 28 U.S.C. § 1361 to entertain an action in the nature of an action for an original writ of mandamus. It is clear that in appropriate circumstances this Court may mandamus any of the defendants in this case, including the Secretary of the Navy. Decatur v. Paulding, 14 (U.S.) Pet. 497, 10 L. Ed. 559 (1840). Plaintiffs claim that the courts-martial which tried them were without authority of any kind, that this has in principle been determined by the Court of Military Appeals, and that the Secretary of the Navy and the other named defendants have wrongly refused to give retroactive effect to the decision of the Court of Military Appeals. The Court agrees with plaintiffs that, in general, the Navy has no right to maintain records of wholly void court-martial convictions as if they were valid, and to *332 continue bearing clerical and bureaucratic witness to their validity if they are not. To correct such a record is a purely ministerial act, which is a proper subject for a writ of mandamus. The fact that plaintiffs' right of relief and defendants' ministerial duties are disputed by defendants does not, as defendants argue, render the rights of plaintiffs "unclear" so as to necessitate dismissal of the action, or destroy this Court's jurisdiction. The time at which plaintiffs must establish to the Court the clear nature of their entitlement is at the point of final decision, not at the filing of the suit. United States ex rel. Joy v. Resor, 342 F. Supp. 70 (D.Vt. 1972). Plaintiffs have stated in their complaint facts sufficient to make out an arguable right to mandamus. Such a pleading vests this Court, pursuant to 28 U.S.C. 1361, with the jurisdiction to determine the facts, hear the arguments, and decide the case on the merits.
The defendants have argued that this Court is without authority to entertain a collateral attack on a court-martial conviction by way of mandamus. The history of the limits on collateral review of court-martial convictions is a somewhat tangled web. The military court system has traditionally been viewed as being as wholly distinct from the federal civil courts as the courts of the state are. The opinions on its exact status have ranged, in recent years, from those of Mr. Justice Vinson and Mr. Justice Minton in Burns v. Wilson, 346 U.S. 137, 73 S. Ct. 1045, 97 L. Ed. 1508 (1953), granting the military system an aura of sovereignty even superior to that of a state, to those of Mr. Justice Douglas in his concurring opinion in Parisi v. Davidson, 405 U.S. 34, 92 S. Ct. 815, 31 L. Ed. 2d 17, (1972), that the whole military "is simply another administrative agency, insofar as judicial review is concerned." Id. at 51, 92 S. Ct. at 825.
Still it must be admitted that the problems of administering justice within the military setting are special problems, and their solution has been given to Congress in the first instance by the Constitution. Congress has seen fit to create a system of military justice wholly apart from the civil system. But it has long been recognized that the Constitution does not, a priori, prevent the subjects of courts-martial from resorting to the civil courts for the vindication of certain rights. The most important avenue of resort is, of course, the writ of habeas corpus, specifically protected by the Constitution, traditionally capable of reaching issues concerning military jurisdiction, and protected from legislative encroachment as a matter of course. See Gusik v. Schilder, 340 U.S. 128, 71 S. Ct. 149, 95 L. Ed. 146 (1950). More will be said on the shifting scope of habeas corpus presently. But now it should be noted that habeas corpus was not the sole method of collateral attack available in the early years of the Republic. In Wise v. Withers, 3 (U.S.) Cranch 331, 2 L. Ed. 457 (1806), the Supreme Court held that the decision of a court-martial that a certain justice of the peace in the District of Columbia was not an officer of the United States, and was therefore a proper subject for Militia duty, was not binding in an action for trespass di bonis asportatis against the members of the court and the officer whom they sent to seize goods to pay a fine for failure to report which had been levied by the court-martial. The argument was made by counsel for the defendant that the decisions of a court-martial were "final and conclusive like those of an ecclesiastical court . . .", but the Supreme Court, *333 through Marshall, C. J., held that in any action the Court may review the jurisdiction of a court-martial, and if it has acted without any jurisdiction, the Court may so decide and enter judgment accordingly. The Supreme Court found that "the court and the officer are all trespassers", id. at 337. While it is true that this case involved the federal militia, and not the regular armed forces, there is no legal reason for distinguishing the cases, as the power of Congress over the federal militia was as great as over the Army. See also, Houston v. Moore, 5 (U.S.) Wheat 1, 5 L. Ed. 19 (1820) (Review of state militia court-martial jurisdiction through trespass); Martin v. Mott, 12 (U.S.) Wheat 19, 6 L. Ed. 537 (1827) (Review of federal court-martial jurisdiction over state militia man called to federal service, through replevin).
The above cases dealt with challenges to the jurisdiction of courts-martial over persons who alleged that they were not really in service when tried at all. In Dynes v. Hoover, 20 (U.S.) How. 65, 15 L. Ed. 838 (1857), the Supreme Court reviewed an assault and battery-false imprisonment action by a Navy seaman against his jailors. He had been sentenced to imprisonment by a court-martial, and though the Court affirmed the actions of the court-martial in Dynes' particular case, it held that such an action would lie if the court-martial had been without either subject matter or personal jurisdiction as had been alleged, even though Dynes was clearly in service at the time. It is interesting to note that Dynes' sentence had apparently been served prior to the bringing of the action, so that habeas corpus would not have been available, but this was not seen by the Court as any impediment to collateral attack on the court-martial conviction.
In Ex parte Reed, 100 U.S. 13, 25 L. Ed. 538 (1879), the Supreme Court affirmed, surprisingly for the first time, the availability of habeas corpus to a serviceman imprisoned by a court-martial lacking subject matter or personal jurisdiction, although such jurisdiction was found to be present in Reed's particular case. Just three years later, in Ex parte Mason, 105 U.S. 696, 26 L. Ed. 1213 (1882), some members of the Court doubted the power of the federal courts to issue habeas corpus in such a situation, though all agreed that it could only be issued on the basis of a lack of jurisdiction in the court-martial in question, which jurisdiction was found to exist in Mason's particular case as well.
Keyes v. United States, 109 U.S. 336, 3 S. Ct. 202, 27 L. Ed. 954 (1883), is the next important case, dealing as it does with the first review by the Supreme Court of what has become a fairly common method of collateral attack on the validity of a court-martial, the suit for back pay. There the Court recognized the validity of such a suit "[w]here there is no law authorizing the court-martial, or where the statutory conditions as to the constitution or jurisdiction of the court are not observed, [or] there is no tribunal authorized by law to render the judgment." id. at 340, 3 S. Ct. at 204. All these are ancient antecedents of the traditional concept of jurisdiction.
In Wales v. Whitney, 114 U.S. 564, 5 S. Ct. 1050, 29 L. Ed. 277 (1884), the Court's previous doubts as to the availability of habeas corpus seem to have vanished. In holding that actual custody was necessary for the issuance of habeas corpus, the court said
"In thus deciding, we are not leaving the appellant without remedy if his counsel are right in believing the court-martial has no jurisdiction of the offense of which he is charged. He can make that objection to the court before trial. He can make it before judgment after the facts are all before the court. He can make it before the reviewing tribunal."
"If that court finds him guilty, and imposes imprisonment as a part of a sentence, he can then have a writ to relieve him of that imprisonment. If he should be deprived of office, he can *334 sue for his pay and have the question of the jurisdiction of the court which made such an order inquired into in that suit. If his pay is stopped, in whole or in part, he can do the same thing. In all these modes he can have relief if the court is without jurisdiction . . ." Id. at 575, 5 S. Ct. at 1055.
Smith v. Whitney, 116 U.S. 167, 6 S. Ct. 570, 29 L. Ed. 601 (1885) involved the issue of whether a federal court could issue a writ of prohibition to forbid trial by a court-martial which lacked jurisdiction. The court specifically reserved judgment on the propriety of such a writ, finding that the court-martial in question had jurisdiction.
The availability of habeas corpus was again affirmed in principle in United States v. Grimley, 137 U.S. 147, 11 S. Ct. 54, 34 L. Ed. 636 (1890), Johnson v. Sayre, 158 U.S. 109, 15 S. Ct. 773, 39 L. Ed. 914 (1895), and Carter v. McClaughry, 183 U.S. 365, 22 S. Ct. 181, 46 L. Ed. 236 (1902), and the propriety of a suit for back pay was affirmed in principle in Swaim v. United States, 165 U.S. 553, 17 S. Ct. 448, 41 L. Ed. 823 (1897). But McClaughry v. Deming, 186 U.S. 49, 22 S. Ct. 786, 46 L. Ed. 1049 (1902) was in many ways the landmark in this line of cases, for it was in this case that the Supreme Court found a lack of court-martial jurisdiction for the first time, reversing the Eighth Circuit and ordering habeas corpus. Here is the affirmance that what the court had held in principle it would apply in fact in the appropriate case.
The question of the propriety of mandamus as a remedy against an improper court-martial proceeding does not appear to have ever arisen in the Supreme Court. This is not surprising, since prior to the passage of 28 U.S.C. § 1361 by Congress in 1962, it was held that the federal courts were without authority to entertain actions for mandamus as an original writ. See McIntire v. Wood, 7 (U.S.) Cranch 504, 3 L. Ed. 420 (1813), and Marshall v. Crotty, 185 F.2d 622 (1 Cir., 1950). The exception to this rule involving the courts of the District of Columbia, see Kendall v. United States, 12 (U.S.) Pet. 524, 9 L. Ed. 1181 (1838), does not appear to have been either widely known or widely utilized. However, it is interesting to note that in the two related cases of United States ex rel. Creary v. Weeks, 259 U.S. 336, 42 S. Ct. 509, 66 L. Ed. 973 (1922) and United States ex rel. French v. Weeks, 259 U.S. 326, 42 S. Ct. 505, 66 L. Ed. 965 (1922), which dealt with actions by the Army Classification Board, the Supreme Court cited the court-martial line of cases, habeas corpus, trespass, et al, for the proposition that mandamus could lie only to test the jurisdiction of a military tribunal (and by necessary inference, would lie then). It is true that mandamus has never been directly approved as an avenue of collateral attack upon the validity of a court martial conviction. Yet a review of the early cases impels the conclusion that, from the beginnings of our legal system, the absolute lack of the traditional accoutrements of jurisdiction could be litigated by whatever means was most handy; trespass d. b. a., replevin, assault and battery, false imprisonment, habeas corpus, prohibition, suit for back pay. Never has the Supreme Court said that any means of raising the issue was improper; rather it has always decided the issue. The same course was adopted as late as United States v. Augenblick, 393 U.S. 348, 89 S. Ct. 528, 21 L. Ed. 2d 537 (1969), in the context of a suit for back pay. It is true that habeas corpus has become the primary way to test the legal sufficiency of court-martial proceedings. It has, in the course of time, expanded beyond what Lord Mansfield would have recognized as questions of jurisdiction. See generally the discussion of the scope of habeas corpus in the context of court-martial judgments in Levy v. Parker and Resor, 478 F.2d 772 (3rd Cir., 1973). And the suit for back pay has filled most needs when the custody requirements of habeas corpus were not met. Yet this Court believes that, *335 at least in a case concerned with traditional concepts of jurisdiction, mandamus will lie to test the jurisdiction of a court-martial. Doubtless a District Court should pursue such a course only in the most exceptional circumstances, but for certain reasons of procedure and judicial economy, this case appears to present such circumstances.
The writ of mandamus has come into recent use to compel military boards for the correction of records established under 10 U.S.C. § 1552, to correct records of invalid court-martial proceedings after the refusal of such a board to do so. Ashe v. McNamara, 355 F.2d 277 (1st Cir., 1965); Smith v. McNamara, 395 F.2d 896 (9th Cir., 1968); Haines v. United States, 453 F.2d 233 (3rd Cir., 1970); Lima v. Secretary of the Army, 314 F. Supp. 337 (E.D.Pa., 1970); and Angle v. Laird, 429 F.2d 892 (10th Cir., 1970). See also Ragoni v. United States, 424 F.2d 261 (3rd Cir., 1970). The progenitor of this approach is the Ashe case. The opinion in Ashe is quite narrowly written. It is sometimes read as reflecting a general suspicion that review of proceedings by Boards of Records-Correction is a narrow exception to 10 U.S.C. § 876, which provides that court-martial sentences are "final and conclusive" and binding on, inter alia, all courts of the United States. This was the analysis of the court in Parrish v. Seamans, 343 F. Supp. 1087 (D.S.C., 1972). The essence of this argument is that 10 U.S.C. § 1553 is an exception to § 876, and that the right of judicial review of § 1553 proceedings found in the legislative history of that section is the only exception to § 876 other than habeas corpus. This Court can agree with the first part of this proposition but does not agree with the last.
This Court sees no impediment to the use of mandamus directly, and not merely on the board for the correction of records, arising out of 10 U.S.C. § 876. The provisions of § 876 embodied a concept of legal finality, and must be read to encompass the normal collateral exceptions to such finality. One of those exceptions is and has always been that complete lack of jurisdiction in the narrow sense may be raised in any available proceeding. The Supreme Court has held that this language in § 876 was not intended to preclude habeas corpus. Gusik v. Schilder, 340 U.S. 128, 71 S. Ct. 149, 95 L. Ed. 146 (1951). It is true that habeas corpus is a constitutionally protected remedy. However, the Supreme Court has had the opportunity to hold that a suit for back pay attacking a court-martial conviction was prohibited by § 876 and has chosen not to do so. Augenblick, supra. The suit for back pay attack has always been regarded as unaffected by § 876 in the Court of Claims, the forum which, until the 1964 amendment to the Tucker Act, 28 U.S.C. § 1346, had exclusive jurisdiction over such suits, and the forum in which the issue has generally been raised since then. See the Court of Claims opinion in Augenblick, 377 F.2d 586, 180 Ct.Cl. 131 (1967) and the cases cited therein. It is the opinion of this Court that these cases are correct, and further that if the constitutionally unprotected suit for back pay is not rendered unavailable by § 876, neither is jurisdiction to entertain direct mandamus in an appropriate case.
The argument against reading § 876 as a bar to either direct mandamus or a suit for back pay was stated succinctly by the Court of Claims in Augenblick as follows:
"There is no adequate reason for looking to habeas corpus alone, or for thinking that Congress limited its exception from `finality' to that specific proceeding. Liberty is of course important, but so are a man's career, his livelihood, his rights as a veteran, his status as a convicted criminal, and his reputation. To deny collateral attack *336 to one not in confinementthe consequence of saying that habeas corpus is the only remedywould be to deny the possibility of review by a constitutional court, and ultimately by the Supreme Court, of the constitutional claims of servicemen like plaintiff who have not been sentenced to jail or who have been released. [377 F.2d 586, 592 (1967).]"
This language was relied upon by the D. C. Circuit in Kauffman v. Secretary of the Air Force, 135 U.S.App.D.C. 1, 415 F.2d 991 (1969) to support its conclusion that habeas corpus is not the exclusive exception to § 876. Both the Kauffman case and the earlier case to the same effect of Gallagher v. Quinn, 124 U.S.App.D.C. 172, 363 F.2d 301 (1966) involved "declaratory judgment" and "mandatory injunction" requests whose jurisdictional basis can only have been 28 U.S.C. § 1361.
The Third Circuit has recognized without disapproval in Ragoni, supra, that "several courts have held that Article 76 (§ 876) was not intended to bar any collateral attack on court martial decisions", citing Kauffman, supra, and the Court of Claims opinion in Augenblick, supra. Further, Kauffman was cited by the Third Circuit with general approval in Levy v. Parker and Resor, supra. It is interesting to note that in Smith v. McNamara, supra, the 10th Circuit, in the context of a case involving mandamus to a board for the correction of records, did not in its opinion follow the narrow language of Ashe, but the broad ideas of Augenblick and Kauffman.
The Ashe decision may be read as holding no more than that jurisdiction exists to mandamus the board for the correction of records, but not that jurisdiction does not exist to mandamus the Secretary or other official directly in an appropriate case. This Court finds nothing in Haines or Ragoni to indicate that the Third Circuit has specifically adopted the view that mandamus to the board for the correction of records is the only action for which this Court has jurisdiction under § 1361. The existence of the Board may raise questions under the exhaustion doctrine, but that is a separate issue. To hold that mandamus to the Board is the only action over which this Court has jurisdiction, exhaustion aside, would unnecessarily tie the hands of the Federal Courts and is not a construction of the statute which should be adopted absent the clearest expression of congressional purpose. Such mandamus is only a thinly veiled use of the mandamus power to examine court-martial proceedings in any event. As will be discussed more fully below, in some cases it forces those Boards to make preliminary decisions on legal issues outside their competence, and makes the outcome of the litigation dependent to some extent on the skill of the parties in creating a record on a legal issue before a tribunal with no expertise on that issue.
We need not go so far as the courts have gone in Kauffman and Gallagher in this case, and hold that all "constitutional" issues may be raised by direct mandamus. We deal here only with jurisdiction in the narrow traditional sense. Such a case does not raise the problem of relitigation of factual issues that seemed to bother the Third Circuit in Ragoni. At any rate, such review of jurisdiction has traditionally been available in whatever proceeding was most convenient, as we have seen, and if Congress had wished to eliminate this by 10 U.S.C. § 876, and thereby put a punitively discharged or unincarcerated person convicted by an illegally convened authority on such a radically different footing than a person "in custody", the Court feels it would have said so more clearly and directly. This Court has jurisdiction over this case under 28 U.S.C. § 1361.
The defendants have also challenged this court's jurisdiction to entertain the claim of the named plaintiffs for back pay under the Tucker Act, 28 U.S.C. § 1346. The 1964 amendment to *337 the Tucker Act, gave this Court concurrent jurisdiction with the Court of Claims on this type of action in any civil action in which less than $10,000 was at issue. As we have seen, all of the relevant precedents in the Court of Claims favor the Court's jurisdiction in this circumstance under the Tucker Act. The possibility of such an action was recognized by the Supreme Court without adverse comment in Augenblick, supra, at 351 of 393 U.S., 89 S. Ct. 528. The above discussion makes clear that the Court is confident that its jurisdiction under the Tucker Act is not affected by 10 U.S.C. § 876. There is no doubt that, since the named plaintiffs make no claim for a sum greater than $10,000, this Court has jurisdiction over their cases. Perry v. United States, 308 F. Supp. 245 (D.Colo., 1970), aff'd. 442 F.2d 353 (10th Cir., 1971). Defendants have argued that the class nature of the complaint in this action makes it a suit for a sum in excess of $10,000. This is not a challenge to the Court's jurisdiction to hear the claims of the named plaintiffs, but to the Court's jurisdiction to hear a class action pursuant to the power conferred by § 1346. The defendants' position appears to be incorrect. See footnote 5 infra. The Court need not decide this issue, however, as the Court does not intend to accord class action treatment to the Tucker Act claims. This case requires class action treatment, binding on all members of the class without the chance to opt out of the class, to lay it to rest once and for all. It is probable that the class contains persons whose claims are, at least arguably under the precedents of the Court of Claims, in excess of $10,000. Under the Tucker Act this Court has no jurisdiction to hear their cases. They might come before the Court if they would waive any excess, but the Court cannot, consistent with due process of law, make waiver by fiat, and attempting to get such waiver would require notification of the class which would at best slow this case down for a long time. Even if the class action were to proceed under F.R.Civ.P. 23(b) (1) or (b) (2), there is no jurisdiction to adjudicate and bind those who make claims in excess of $10,000 under the Tucker Act. The case would not be decided with the finality it demands. However, there are no such jurisdictional problems with the mandamus action. This can be adjudicated fully under Rule 23(b) (2), as discussed more fully infra, and the entire class bound whatever the ultimate result. This is why the Court has gone to such pains to vindicate its mandamus jurisdiction in this most appropriate case.
This is probably the best point to raise the problems posed by the case of Carter v. Seamans, 411 F.2d 767 (5th Cir. 1969). Carter v. Seamans dealt with an action by a discharged Air Force Captain seeking to mandamus the Secretary of the Air Force to correct records pursuant to 10 U.S.C. § 1552. Plaintiff sought to be reinstated in the Air Force, promoted to the rank of Full Colonel, and have all his back pay and allowances paid from the date of his allegedly illegal discharge. The claim if paid in full would have approximated $135,000. The Court concluded in that case that the suit was most accurately characterized as a suit for back pay in the guise of a mandamus action. The Court correctly saw that to allow Carter's litigation as framed would virtually eliminate the $10,000 limitation on the jurisdiction of the District Courts under 28 U.S.C. § 1346. The Court dismissed the action and sent Mr. Carter to the Court of Claims, holding that "in cases of the type now under consideration the monetary limitation imposed on its jurisdiction by section 1346 must also be held to apply to a mandamus proceeding brought pursuant to section 1361." id. at 775.
This holding is, properly understood, an excellent one, but the case at bar is not a case of the type then under consideration. *338 First, as the Carter court noted, Carter was seeking various disputed and unliquidated sums in damages in his action. Although mandamus relief can reach money wrongfully held, the sum generally must be a sum certain or easily capable of reduction to a sum certain. Where, as in Carter, litigation is potentially as much over how much is to be paid as whether there is a duty to pay, it is the kind of case which should be handled as a contract-analogy case, not a mandamus action. Second, Carter, like most such plaintiffs, was suing for himself on the basis of facts unique to himself. There was no class consideration necessary, so that the absence of class procedures in the Court of Claims was of no import.
In the case at bar, even if we were to hold that unliquidated sums could be returned by mandamus, the cases of the named plaintiffs would not present any § 1346 problem, as in Carter, because, as noted, they have waived any claims in excess of $10,000. Further, the case at bar requires class treatment not available by the procedures of the Court of Claims.
This Court does not believe that it would be proper to mandamus the payment of unliquidated or disputed sums in the nature of damages. However, this Court does believe that it can mandamus the return of fines or forfeitures levied directly within the four corners of an invalid judgment by a courtmartial without jurisdiction. There is no discretion to retain the latter and no reasonable basis to dispute their amount. They are sums certain or readily capable of being made certain. The cases distinguished properly in Carter support this position here. Miguel v. McCarl, 291 U.S. 442, 54 S. Ct. 465, 78 L. Ed. 901 (1934); Clackamas County, Oregon v. McKay, 94 U.S.App.D.C. 108, 219 F.2d 479 (1954). The $10,000 limitation of § 1346 is not applicable to such sums, and would pose no problem to class treatment of the mandamus action even if it were not virtually certain that no direct fines or forfeitures totalled $10,000 for any one member of the class.
There is a distinct difference between returning a fine and litigating all the arguable elements of damage in a back pay case. It is true that the outcome of the mandamus class action may establish liability by res judicata in future back pay cases in the District Courts or in the Court of Claims. That result is merely the outcome of the common factual basis of the distinct claims. The scope of these actions is not necessarily coterminus in all contexts. When they are in a given case, this does not defeat the jurisdiction of this Court to decide a case properly before it. Besides, the limitation which Congress has placed upon this Court in § 1346 does not really go so much to who should determine liability as who should have the power to determine the proper elements and magnitude of the award. Any members of the class who might seek to claim more than $10,000 in damages aside from the return of fines or forfeitures will still have to seek their money in the Court of Claims. Nothing in § 1346 or Carter v. Seamans does, or should be read to, limit the availability of class treatment in a case such as this.
The next question concerns the applicability of the doctrines of exhaustion of remedies and abstention on comity *339 grounds to the case at bar. These doctrines are not properly viewed as jurisdictional in nature except when required by statute. Rather, they are doctrines of judicial self-discipline. This is not to say that they are questions on which a trial court has absolute discretion, or even that the decision of a trial court must be "clearly erroneous" before an appellate court will substitute its judgment for that of the trial court. Doctrines of judicial self-discipline occupy a special position in the legal system, and the supervisory authority of appellate courts gives them special latitude in such matters. These doctrines are not doctrines of substantive law nor fixed doctrines of jurisdiction or procedure. They do not concern themselves with the rights of the parties so much as the relationship of the courts with other governmental bodies which have spheres of responsibility which must not be disregarded as a matter of course. They are flexible, and dependent on the particular facts of the case. Their general rules have numerous exceptions, and exact precedents are rare. Yet the doctrines are very important to keeping the role of the courts in proper symbiotic balance with the other agencies of legal decision in society. Therefore, when an issue of exhaustion or abstention is raised, it must be examined with special care.
The differences between the exhaustion doctrines and the abstention doctrines have not always been recognized. See Parisi v. Davidson, 405 U.S. 34, 92 S. Ct. 815, 31 L. Ed. 2d 17, especially footnote 6. When a Court chooses not to exercise jurisdiction in favor of another tribunal which can make a binding judgment on an issue not subject to any sort of review by the first court, it abstains. Thus, in a typical situation where a U. S. District Court faces a Constitutional issue based upon the construction of a state statute which could be construed to avoid the issue, the Court may abstain in favor of a state court, awaiting a binding construction of the statute. However, when a court chooses not to exercise jurisdiction in favor of another tribunal, knowing that the decision of that tribunal may later be reviewed in one way or another in an appropriate proceeding between the same parties in the first court, the court is requiring exhaustion of other available remedies.
The first alternative remedy pressed by the defendants as available presents an interesting exercise in classification. Defendants say that plaintiffs may apply for writs of coram nobis in the Court of Military Appeals. If they did, and if they lost, would that decision be binding on this Court in proceedings such as these as res judicata? The answer appears to be no. The issue in this case is the availability of collateral relief from an erroneous exercise of jurisdiction within the military justice system. This opinion has spoken at length of the propriety of such relief by the civil courts. For the Court of Military Appeals to affirm the existence of jurisdiction does no more than add one more decision by the military justice system in its own favor. Coram nobis proceedings are not binding in proper collateral procedures any more than the original decision is binding. Thus denial of coram nobis in a state court does not foreclose federal habeas corpus. The decision of the Court of Military Appeals would probably be correct and it would be entitled to great respect and consideration, but it would not constitute res judicata if any collateral review of courtmartial jurisdiction is ever proper. The other two alternative remedies pressed, review by the Judge Advocate General under 10 U.S.C. § 869, and resort to the Board for the Correction of Naval Records, previously discussed, are clearly *340 administrative and reviewable here, so that we deal exclusively with the question of whether to require exhaustion of other available remedies. This Court does not believe that requiring resort to any of the proposed remedies is required by the policies of the exhaustion doctrine or is in the best interests of justice in this case.
As to the named plaintiffs in this case, they still have a viable cause of action for back pay pending in front of this Court, and application for a writ of coram nobis from the Court of Military Appeals could not give full relief, since that Court cannot grant the kind of relief that can be granted in a back pay suit through a writ of coram nobis. U. S. v. Snyder, 18 USCMA 480 (1969). This fact alone disposes of coram nobis as a potential alternative remedy to be exhausted.
It should be noted, however, that the availability of coram nobis in the Court of Military Appeals is by no means clear, either to the named plaintiffs or to any members of the class. A comparison of the decisions in U. S. v. Bevilacqua, 18 USCMA 10 (1968); U. S. v. Snyder, supra, and Osborne v. Bowman, 20 USCMA 385 (1971), make the availability of coram nobis in the Court of Military Appeals in the context of this case at best speculative. In the Osborne case, that Court held that its writ jurisdiction extended only to the prevention of events which would "tend . . . to defeat this Court's possible future [appellate] jurisdiction" or to "prevent the rendition of any relief shown to be necessary during the ordinary course of appellate review." 20 USCMA at 387. It is questionable that a writ of coram nobis to inquire into the jurisdiction of a court-martial long since beyond appeal to the Court would qualify within the Court's narrow definition of its writ power as tending to preserve the Court's future appellate jurisdiction. Plaintiffs cannot properly be required to exhaust a remedy which may not exist. Parisi v. Davidson, supra, quoting Noyd v. Bond, 395 U.S. at 698, note 11, 89 S. Ct. 1876.
Further, this case no longer concerns any issues in which the special expertise of the Court of Military Appeals would be of value. It has spoken on the issues of jurisdictional defects. The issue of retroactivity, insofar as it has not spoken on that issue in the recent case of U. S. v. Ferry, 22 USCMA 339, 46 CMR 339 (1973), requires no special expertise, difficult though the question may be. The Supreme Court has always been sensitive to the friction which might arise between the civilian courts and the military courts if the civilian courts prejudged issues within the special sphere of competence of the military courts before the military courts had had an opportunity to speak upon those issues. But where the issues involved in the case were not within such a special sphere, the court has been much less ready to find that comity between the two court systems requires exhaustion through the military courts first. Compare Noyd v. Bond, 395 U.S. 683, 89 S. Ct. 1876, 23 L. Ed. 2d 631 with Toth v. Quarles, 350 U.S. 11, 76 S. Ct. 1, 100 L. Ed. 8 (1955) and the other cases cited in footnote 8 of Noyd v. Bond at page 696, 89 S. Ct. 1876. This case in its present posture falls into the latter category. Further, insofar as the proper consideration in deciding whether to require exhaustion of military court remedies is the prevention of friction between the military court system and the civilian court system, there appears little likelihood of friction between the Court of Military Appeals and the civilian courts over our failure to send it the pretty package of a potential 30,000 or so applications for coram nobis.
The second "alternative remedy" proffered by the defendants is a resort to the Judge Advocate General pursuant to 10 U.S.C. § 869. The attorneys *341 for the defendants have admitted in their brief, quite candidly, that such resort is not available to the named plaintiffs by the very terms of the statute, since they received punitive discharges approved through the supervisory authority level of review and their cases were therefore necessarily reviewed by the United States Navy Court of Military Review pursuant to 10 U.S.C. § 866. So much for the availability of that proffered remedy. It should also be pointed out that review under Article 69 could not give named plaintiffs full relief on their back pay claims, and is futile in the face of the Judge Advocate General's published position that the decisions in Greenwell and Ortiz are not retroactive.
The Board for correction of Naval Records was established pursuant to 10 U.S.C. § 1552, providing for the creation of civilian boards from the executive part of a military department to correct military records to remove error or injustice. The main reason why Congress created this Board had nothing to do with reviewing the actions of courts-martial. Military records are mostly the product of administrative action. A person's records may be inaccurate from clerical error or from error in the administration of all sorts of non-criminal regulations. The effect on the victim of such an error can be both great and unjust. It may effect, inter alia, pay, assignment, promotion or rating for one in the military, and, inter alia, employability, veterans benefits and pension qualifications for one already discharged. Prior to the passage of § 1552, there was no efficient mechanism to correct such errors, and after World War II, Congress itself was flooded with private bills to work such corrections. In response to this situation, Congress passed § 1552. Ashe v. McNamara, supra.
While the main purpose of § 1552 boards was to correct administrative errors, the Congressional grant of power was broad enough arguably to allow the boards to correct, "errors in records" resulting from improper court-martial action. This would read § 1552 as an exception to § 876 finality. This, of course, is one of the holdings of Ashe v. McNamara. Even since the Ashe decision, the Secretary of the Navy and the Navy Board have not adopted the position that § 1552 contains a complete exception to § 876. The position of the Secretary and the Board are not without merit. The kind of expertise developed by the Board in its run-of-the-mill case is of no help in deciding the legal issues raised by an attack on the validity of a court-martial. Lack of jurisdiction and Constitutional error are technical legal issues. The Board is not a Court and is not necessarily made up of lawyers. If the Ashe decision was based merely on a reading of the statute and a conviction that Congress intended the Board to correct any error resulting in injustice, so be it. There are occasionally court-martial errors so glaring and clear that even a board of laymen will perceive them, and perhaps Congress did intend for the Board to then correct such an error. The Ashe decision, as has been noted, may also have reflected a suspicion by the Court that reading § 1552 as an exception to § 876 was necessary to the jurisdiction of the Court, a position consistent with the First Circuit's later ruling in Davies v. Clifford, 393 F.2d 496 (1968), but this position has already been dealt with at length above, and this Court has reached the opposite conclusion.
The question now before the Court is whether resort to the Board should be required under the exhaustion doctrine in this case. The Board has the authority to recommend to the Secretary relief which is much closer to full relief than that which could be given by the other agencies previously discussed. However, the relief is nowhere near full relief. As noted in the recent book Justice and the Military by Homer E. Moyer at § 6-225, the balance which has been struck by the boards for the correction of records between § 1552 and § 876 is such that ". . . military correction boards will review sentences imposed *342 by courts-martial, but no longer will the boards pass judgment on the conviction itself." A completely successful petitioner, Moyer notes "may have all the punitive consequences of a court-martial removed from his record, but the correction board will refuse, on jurisdictional grounds, to expunge the underlying conviction itself." This limitation on relief available in addition to other reasons discussed later in this section effectively eliminates military correction boards as an essential step in exhausting military remedies prior to collateral attacks on court-martial convictions in federal court."
Further, even in the area of its greatest responsibility and expertise, revealing administrative decisions, resort to the board is not always required before coming to court. Compare McCurdy v. Zuckert, 359 F.2d 491 (5th Cir., 1966) with Ogden v. Zuckert, 111 U.S.App.D.C. 398, 298 F.2d 312 (1961). More importantly, the Supreme Court has recognized that in many cases to require application to the board is not appropriate and not required by the exhaustion concept. See the order vacating the Ninth Circuit's decision requiring such exhaustion in Craycroft v. Ferrall, 397 U.S. 335, 90 S. Ct. 1152, 25 L. Ed. 2d 351 (1970) and the explanation of that action in Parisi v. Davidson, 405 U.S. 34 at 38 footnote 3, 92 S. Ct. 815. There appears even less reason to require exhaustion when the pertinent issue is a very technical legal issue quite outside of the special competence of the Board. The Court is aware that the D. C. Circuit in Owings v. Secretary of the Air Force required application to the Board under the exhaustion principles without opinion. See 145 U.S.App.D.C. 76, 447 F.2d 1245, 1249, footnote 8. But the Court is also well aware of the admonition of McKart v. United States, 395 U.S. 185, 89 S. Ct. 1657, 23 L. Ed. 2d 194 (1969) that the exhaustion doctrine "is applied in a number of different situations and is, like most judicial doctrines, subject to numerous exceptions. Application of the doctrine to specific cases requires an understanding of its purposes and of the particular administrative scheme involved." id. at 193, 89 S. Ct. at 1662. Whatever the merits of requiring resort to the Board may be in a case involving collateral attack upon a court-martial not based on jurisdictional grounds, and having relevant elements arguably falling within the special expertise of the board, the Court feels that such resort is not required in the context of this case.
The Court is aware that it could have based its decision not to require exhaustion of presently available remedies on the proposition that such exhaustion is never required when the basis for the collateral attack on a court-martial is "jurisdiction." See for example, Autry v. Wiley, 440 F.2d 799 (1st Cir., 1971); Gorko v. Commanding Officer, 314 F.2d 858 (10th Cir., 1963); and the many cases collected in Moyer, Justice and the Military, §§ 6-250 to 6-255. However, convinced as this Court may be that this case involves only a most narrow question of traditional jurisdiction, the scope of the term "jurisdiction" used in the cases, and the weakness underlying the policy that exhaustion is not required in cases attacking "jurisdiction" in some of its aspects, has justified the demonstration that exhaustion of remedies should not be required in any event.
One further policy factor militates against requiring exhaustion in this case, even if such exhaustion might arguably be otherwise required. None of the alternative remedies pressed by the defendants contains any mechanism for the class handling of the dispute, and some form of class treatment is without doubt both the most convenient and the most just way to dispose of the issues raised in this case.
We turn now to the substantive consideration of the case at bar. 10 U.S.C. § 823 reads:
(a) Special courts-martial may be convened by
(1) any person who may convene a general court-martial;
*343 (2) the commanding officer of a district, garrison, fort, camp, station, Air Force base, auxiliary air field, or other place where members of the Army or the Air Force are on duty;
(3) the commanding officer of a brigade, regiment, detached battalion, or corresponding unit of the Army;
(4) the commanding officer of a wing, group, or separate squadron of the Air Force;
(5) the commanding officer of any naval or Coast Guard vessel, shipyard, base, or station; the commanding officer of any Marine brigade, regiment, detached battalion, or corresponding unit; the commanding officer of any Marine barracks, wing, group, separate squadron, station, base, auxiliary air field, or other place where members of the Marine Corps are on duty;
(6) the commanding officer of any separate or detached command or group of detached units of any of the armed forces placed under a single commander for this purpose; or
(7) the commanding officer or officer in charge of any other command when empowered by the Secretary concerned.
(b) If any such officer is an accuser, the court shall be convened by superior competent authority, and may in any case be convened by such authority if considered advisable by him. Aug. 10, 1956, c. 1041, 70A Stat. 44.
In 1963, the Court of Military Appeals decided U. S. v. Ortiz, 15 USCMA 505, 36 CMR 3. In that case it held that 10 U.S.C. § 823(a) (6) was not a direct grant of power to convene special courts-martial to any separate or detached military unit of less than roughly battalion size. In 1970, the Court of Military Appeals decided U. S. v. Greenwell, 19 USCMA 460, 42 CMR 62. In that case, the Court of Military Appeals decided that 10 U.S.C. § 823(a) (7) did not authorize the Secretary of the Navy to make a categorical or bloc grant of power to convene special courts-martial, and delegate the power to determine which commands qualified for category membership to flag or field officers. The Court of Military Appeals held that to "empower" a Commander, the Secretary must act personally re that Commander under § 823(a) (7). In 1973, the Court of Military Appeals decided U. S. v. Ferry, 22 USCMA 339, 46 CMR 339, holding that the jurisdictional nature of the Greenwell decision mandated its retroactivity, at least as far as the power of that Court to forbid current use of past convictions within the military justice system might go.
Plaintiffs argue that this Court is well-nigh bound to follow those decisions, and to enter judgment for plaintiffs almost in vindication of the authority of the Court of Military Appeals. Defendants argue that this Court is not so bound.
If the decision in the above cases had gone against plaintiffs, they would argue, and quite rightly, that this is a proper collateral proceeding and that the Court is not bound. One should not invoke merely a hollow doctrine of mutuality to decide that the Court is not bound in the present circumstance either however.
It is true that this Court owes a duty of obedience to the highest court of a state on issues of state law. The argument is made that the Court of Military Appeals is the State Supreme Court of Military Law, and must be followed. There is a certain appeal to the argument, but the Court of Military Appeals is not a State Supreme Court in one important way it cannot give civil effect to its own judgments, for it does not have jurisdiction over any civil court. It has absolute supervisory and decisional authority over the military criminal courts, but when litigants seek to give its decisions civil effect they must resort to the federal civil court. What then *344 should be the posture of such a court? Only one reported decision has considered the problem. In Robb v. United States, 456 F.2d 768 (1972), the Court of Claims wrote:
"A further compelling reason for following Averette is that Congress seems to have endowed the rulings of the Court of Military Appeals with absolute finality where they favor the serviceman (or other court-martial defendant). Article 76 of the Uniform Code of Military Justice, 10 U.S.C. § 876 (1970), provides that military review of court-martial convictions (including decisions of the Court of Military Appeals) shall be `final and conclusive' and `binding upon all . . . courts . . . of the United States.' If the accused has been convicted, there is an implied exception in his behalf for relief through habeas corpus (United States v. Augenblick, 393 U.S. 348, 349-350, 89 S. Ct. 528, 21 L. Ed. 2d 537 (1969)), and in the view of some Courts of Appeals (footnote omitted) and of this court (footnote omitted), also by way of declaratory judgment or other form of civil action, provided that the judicial review is confined in scope to that comparable to habeas corpus scrutiny. But there appears to be no such review available to the Government if the Court of Military Appeals decides against the Government and in favor of the court-martial defendant. Article 76 makes a final military determination `binding upon all departments . . . agencies, and officers of the United States,' and the implied exceptions are for the benefit of the accused individual so that he can seek a ruling by an Article III court on his constitutional or jurisdictional claims. The Government, it would seem, has no such right to ask a civilian court to overturn a determination by the Court of Military Appeals in favor of the accused. There was no way, for example, in which the Government could have Averette's conviction reinstated by an Article III court after it had been quashed by the Court of Military Appeals. Congress can, of course, restrict the Government to a decision by the military tribunals, and Article 76 appears to have done just that. We do not suggest that we are technically bound by the Averette ruling, as we might be if Averette were himself the plaintiff here, but we do find significance in Congress' special deference toward the Court of Military Appeals where that court rules against the Government. This being so, we think that civilian courts should normally follow, without the usual independent examination, a Court of Military Appeals ruling on an issue of jurisdiction under military law which favors the court-martial defendant." Id. at 771.
This discussion seems well reasoned and wise. It should be noted that the court did not intimate that it owed an absolute duty of obedience to the Court of Military Appeals, however. The Court of Military Appeals has no direct decisional responsibility for the civil implications of its rulings. To the court system which must bear the responsibility for determining and enforcing those civil implications must go the theoretical right to make independent decisions on all relevant issues of law as they apply to those civil effects. However, it is clear that the Court of Military Appeals should usually be followed. Only under the most unusual circumstances should the civil court substitute its judgment for that of the Court of Military Appeals. The substantive questions raised by the Ortiz and Greenwell decisions do not present such circumstances. The retrospectivity determination inherent in the Ferry decision does.
This case is a bit of an anomoly. In the context of civilian criminal law, a mis-convicted person might sue the state only for the correction of records, and the return of fines. The money impact of fines in a civilian criminal context is very much less. Fines are not often levied, and less often collected in the average civilian case. In the military, *345 since the sovereign is the employer and the paymaster, and since absent a sentence of forfeiture a soldier must be paid even in jail, forfeitures are levied as the rule. It should be noted that the impact of a sentence to jail by a civilian court is the same loss of income but no right of action against the government can accrue from it. In the military context, because the criminal jurisdiction arises from and has implications for a quasi-employment contract, a misconvicted person can sue the state for such loss of income.
In the civilian context, a retroactive ruling that does not throw open the doors of the jails has little probability of precipitating a dislocating impact anywhere. In the military context, it has little such impact within the confines of the military justice system itself. However, the impact may be massive in the civil courts bailiwick. The almost off-the-cuff determination of retroactivity in Ferry cannot have been the product of a thorough examination of the impact that it might have on the civilian courts. It must have been merely the product of an incantation of the traditional Blackstonian effects of jurisdictional failure which raised no real problems within the Court of Military Appeals field of vision at the time. Indeed, it could be argued that the Court of Military Appeals could not properly have denied retrospective application to Greenwell within its own bailiwick because the proper magnitude of impact simply was not present within the military system itself to overcome the general rule of retrospectivity of decision. It is not so obvious that the traditional Blackstonian effects of jurisdictional failure are appropriate in the determination of the civil implications of Greenwell which are the sole responsibility of this Court to decide.
All concepts of jurisdiction spring from the assumption that the lawful power of all men or bodies of men in official capacities should be limited. The concept of limited power and therefore the concept of jurisdiction is necessary to prevent any single man or body of men from boldly arrogating to themselves tyrannical power, from consolidating such power by degrees, or from exercising arbitrary or arrogant power outside the framework of a rule limited system, and then being able to call their actions legal. It is centrally necessary to any system of checked and balanced power.
Our ideas of jurisdiction and its attributes have been formed over a thousand years of legal evolution. They serve an immensely important role in describing to us the ways in which power is limited. Their role is so important that they should be tampered with only for the most cogent reasons, for they are like a living organism: It is difficult to predict all the consequences that even a small change in the traditional incidents of jurisdiction may have, or its effect on the vitality of those doctrines in protecting each of us from the arrogance of tyrannical power.
Still, it must be admitted that sometimes the rigidity of traditional concepts of jurisdiction can have extreme effects which do not seem warranted. Jurisdiction traditionally either exists or does not exist. The common law made no distinction between a kangaroo court, and a court whose judge had by mistake been issued the improper commission. Both tribunals lacked jurisdiction equally. The proceedings of both tribunals were totally void. The common law looked not to the reasonableness or unreasonableness of the decisions which had led to the exercise of power, but to its metaphysical presence or absence.
This Draconian conceptual approach gives maximum protection against the improper assumption and exercise of power. No circumstance can ratify its exercise, no appearance of reasonable mistake can vindicate its assumption, and consequently there are no rewards to be gained and held by its improper exercise. And this approach was well suited to the conditions of most of our legal history. Although the decision that a particular action was done without *346 jurisdiction controlled in the case of the parties litigant, and conceptually controlled other cases as well, the vast bulk of the effected population was bound to be too poor and ignorant to take advantage of the ruling. The last half-century has seen this Country take great strides in providing legal mechanisms to insure that mere ignorance or poverty are not the sole determining factors of how a person is treated by the process of the law. The class action device is often such a mechanism, for although it helps with judicial administration, it also helps in many cases make reality the promise that persons similarly situated will not be differently treated merely because some of them are ignorant of their rights.
Moving to make our best myths reality has caused some problems, however. This case is one of those problems. The real reason for the idea of jurisdiction is to protect us all, not merely in the individual case, but over the long run, from unchecked power. Any application of prospectivity to a jurisdictional failure erodes that concept to some extent. Yet perhaps some very limited exception based on a sound analysis of the actual policies behind jurisdictional limitations, is more in keeping with both the approaches and the realities of 20th Century jurisprudence.
The plaintiffs claim that the courts-martial in question did not provide proper guilt-determining agencies. Yet, had the Secretary of the Navy signed the order authorizing the convening authority, as he could have, and as it appears from his actions since Greenwell, he would have, had he known he had to, everything would have been the same. If there is outcome-affecting command influence in such small units, it is still there. Either it is enough of a factor in itself to constitute a denial of due process or else the trials of the plaintiffs met due process standards on all outcome-affecting grounds, because the actual size of the units exercising special court-martial authority has not been changed by post Greenwell practice. Sailors and Marines are still subject to the same system on the operational level, which is where it counts in terms of outcome-determining due process.
Of course, some failures of "jurisdiction" logically require complete retroactivity in all its aspects. For example, if we deal with conduct which could not be made criminal, or has not been made criminal, or which could not be constitutionally punished, retroactivity must apply. There has been no crime, and no court could properly try the defendants in question. The failure of jurisdiction runs to the jurisdiction not merely of the courts but society as a whole. On the other end of the spectrum is the present case. Not only was the behavior complained of properly criminal, so that some tribunal could hear the case, the very tribunal that heard the case could have heard it if the Secretary had made personal delegation. Further, the Secretary in every case almost beyond doubt would have signed the convening authority delegation personally if that had been thought necessary at the time. The Secretary's en bloc method of designation was a continuous Navy practice, predating the adoption of the Uniform Code of Military Justice, and was not obviously forbidden as a means of "empowering" a commanding office. The *347 mistake was a good faith error not springing from a stiff-backed resistance to the limitations which a reasonable man would or should have seen imposed by the law. The procedure was followed under the Uniform Code of Military Justice for nearly 20 years before it was challenged.
This does not mean the Court disagrees with Greenwell. In fact, the Court has made its own study of the scheme of Article 23 in the context of the Uniform Code of Military Justice and the materials available indicating the legislative intent behind the words, and believes that personal designation is more in line with the statutory scheme, and was probably intended by the word "empowered." However, the Navy was not unreasonable in assuming that its past practices were still acceptable under the meaning of the word "empowered." The drafters probably intended to halt such practice, but did not make that intention clear enough to make the Navy aware that it should adopt a positive change of practice. The issue did not arise until Greenwell. The circumstance that made the issue surface was the Court of Military Appeals' awareness, after the decision of Ortiz, of the strong congressional intent that small commands not be given special court-martial power as a matter of routine. The *348 vices of bloc designations under 23(a) (7) in rendering such grants routine only then became apparent. Only then did the Court of Military Appeals analyze closely 23(a) (7) and decide that the Secretary's practice was forbidden by it. In doing so, it adopted the construction of the language more clearly comporting with congressional intent over that adopted of long-standing by the Secretary. That is a valid exercise of judicial power, and the special expertise of the Court of Military Appeals on the actual tension between the Secretary's practice and the purpose of Section 823 render the decision in Greenwell unarguable.
The Secretary, according to his attorneys, has continued to empower small units as routinely and widely as before by his power under § 823(a) (7). Since such routine exercise, while arguably bad practice, does not produce courts-martial so shot with command influence as to be unconstitutionally unfair, that is not a problem for this Court's consideration. At least the Secretary must sign each order now, and the policy of routine and widespread authorization will not be continued merely by inertia and abdication to the military, but by the continuing conscious decision of the civilian entrusted by Congress with checking the military.
Because Greenwell was right does not mean it should be retroactive. It was at least as much a prophylactic as a remedial decision.
But there is a major unresolved issue lurking behind any consideration of prospectivity. The main consideration militating for prospectivity in this case is the impact of retroactivity on the public treasury, both in direct refunds to the serviceman involved, and also in the payment of sufficient clerical personnel to make the corrections which would be necessitated by retroactivity, and which would constitute a massive undertaking. But can a refusal to grant retroactive effect to a decision be based primarily on "mere" considerations of public expense. That is exactly what is at stake in this case. The Court of Military Appeals has already ruled that the Navy can make no use of Greenwell-type records in the military justice system. This Court's decision will in no way diminish that ruling. Indeed, it could not. There is little benefit to the Navy or the cause of general law enforcement that these specific records of petty crimes be maintained, and petty the crimes must be, since a special court-martial may not impose a sentence of more than six months imprisonment.
Yet the potential liability to the public treasury is massive, in terms of the cost of correction, refund of forfeitures and payment of damages in back pay suits. It might be argued that the forfeitures are merely being returned to their rightful owners, but that begs the question. If the total burden on the treasury for correction, refund and damages is merely $1000 per man, this is a current drain of $30,000,000 from the public monies, and a massive disturbance of the financial status quo.
These factors were thought proper for consideration by the pro-prospectivity plurality in Gosa v. Mayden, 413 U.S. 665, 93 S. Ct. 2926, 37 L. Ed. 2d 873 (1973). But the usual factor of the public interest in being protected from mass release of felons was present there also, to an indeterminate degree. May the Court properly play money changer with the integrity of the concept of jurisdiction, and with the usual rule of retrospectivity of decisions? It is a dangerous course yet the answer is necessarily yes. If a technical defect of jurisdiction were found which existed in every court-martial in the last century, the courts would not be required to bankrupt the government to pay homage to the traditional implications of a jurisdictional failure. That massive public expense may alone justify prospectivity is in the extreme case self-evident.
Yet is $5,000,000 or $30,000,000 or $100,000,000 enough. Absent the precipitation of a gross public financial crisis, *349 is any amount of mere money enough?
Gosa v. Maiden is of little help, and in fact probably renders the decision more difficult. Gosa and its companion case of Warner v. Flemmings dealt with the issue of whether or not the Supreme Court's decision in O'Callahan v. Parker, 395 U.S. 258, 89 S. Ct. 1683, 23 L. Ed. 2d 291 (1969), was to be applied retroactively.
O'Callahan had held that Congress did not have the power under the Constitution to empower the military court system to try a member of the military for a non-service related crime if the defendant was not thereby accorded the usual Constitutional rights of trial by jury, indictment by a grand jury, et al. The decision in O'Callahan was couched in "jurisdictional" terms, and obviously does deal with a limitation upon the power of Congress. Yet, decisions holding that certain procedural safeguards must be accorded in certain contexts, do not fit comfortably within the traditional concept of jurisdiction, and are closer in many ways to the concept of jurisdiction-in-quotes which was elaborated on to some extent by Mr. Justice Douglas in his separate opinion in Gosa, and which is traceable more or less directly to the decision of Johnson v. Zerbst, 304 U.S. 458, 58 S. Ct. 1019, 82 L. Ed. 1461 (1938).
In Gosa, the Fifth Circuit held O'Callahan not to be retroactive, using the usual test of Stovall v. Denno, 388 U.S. 293, 87 S. Ct. 1967, 18 L. Ed. 2d 1199 (1967). In Warner, the Second Circuit held that O'Callahan was retroactive because it involved a jurisdictional issue.
Although the four Justice plurality in favor of prospectivity spoke in its opinion of a "holding", their views were not in fact a holding. Four Justices favored retrospectivity on the basis that the issue involved was jurisdictional. Four Justices favored prospectivity after a Stovall v. Denno analysis. One Justice voted for further argument. It is true that the Fifth Circuit was affirmed in Gosa, and that the Second Circuit's result was reversed in Warner. But the vote on the crucial issue was a tie, and the Second Circuit was reversed because three Justices who did not vote for prospectivity found the Second Circuit case to involve a service related crime. If both cases had involved a clearly non-service related crime, both would have been affirmed on a 4-4 tie vote, with Justice Douglas voting for more argument. Exactly what will happen in the future is anybody's guess, literally.
If the relevant variable in the Supreme Court proves to be how clearly a given case concerns an issue of traditional jurisdiction, Greenwell will be held retroactive, for it involves more clearly what Blackstone would have called a jurisdictional issue than did O'Callahan. However Stovall v. Denno may be applied en grosse to an examination of jurisdictional issues, or, more likely, it may be adopted in a narrowed form. If the relevant variable proves to be an analysis based on the importance of the issue in the scheme of properly allocated and limited power, and its impact on outcome determination viewed in light of the dislocating impact of a holding of retrospectivity, then Greenwell should seriously be considered for prospective application, for it is more clearly a candidate for prospectivity on such an analysis than was O'Callahan.
This area of the law is currently in the early stages of evolution, and the proper course to follow is not clearly marked. It is a difficult area, and not without its dangers, both for the image of the judicial process in general, as has been noted quite persuasively by Professor Mishkin in his forward to the Harvard Review of the 1964 Supreme Court Term, 79 Harvard Law Review 56, and also to the vitality of jurisdictional limitations to protect us from arbitrary power. The Court does not believe that the test of Stovall v. Denno will be applied broadly to jurisdictional questions, but that ultimately some narrow and separate test will be so applied, and that *350 jurisdictional questions will be allowed prospectivity in limited circumstances.
This Court will attempt to frame no general rule. It merely holds that, when a jurisdictional defect springs from a technical failure to empower a tribunal, and that tribunal could have been properly empowered, and that when the mistake which led to the jurisdictional failure was a reasonable mistake, made in good faith, and with no hint of an intentional improper assumption of power, or a stiff-backed resistance to the clear mandate of existing law, that then a court may consider applying the decision which strikes down such jurisdiction only prospectively. Further, this Court finds that under all the circumstances of this case, and in consideration of the magnitude of the financial impact upon the public monies that a holding of retroactivity in this case would impose, the decision of the Court of Military Appeals in U. S. v. Greenwell should be accorded only prospective application in the civil courts of the United States.
Only one housekeeping issue remains to be dealt with. The plaintiffs in this case have sought to represent a class comprised, not only of those persons tried by courts-martial convened pursuant to the authority of JAG Manual Section .0103(b) (5), which was struck down in Greenwell, and pursuant to which both of them were tried, but also persons tried pursuant to the authority of other en bloc designations under § 823(a) (7) struck down by the Court of Military Appeals after Greenwell. The Court entered an Order drawing the class over-broadly. The Court does not now feel that the named plaintiffs have standing to challenge any court-martial brought under any Section other than .0103(b) (5) of the JAG Manual, and further feels that the issues raised in applying the later decisions retroactively for civil purposes may be different than the issues raised in this case, at least as regards retroactivity to the date of the Greenwell decision. Accordingly, the Court's previous order defining the class will be amended to cover and bind only those people tried by special courts-martial convened under Section .0103(b) (5) of the JAG Manual of the Navy between the decision in Ortiz and the decision in Greenwell, and summary judgment will be entered for the defendants and against the plaintiffs on the issue of liability in this case.NOTES
 The scope of mandamus relative to the return of illegally levied fines and forfeitures is discussed infra.
 Although the opinions in Burns v. Wilson reflect a very deferential attitude toward military sovereignty, some writers have traced the abandonment of the traditional "lack of jurisdiction" standards for collateral review by habeas corpus to this case. See, e. g. Hosmer, Justice & the Military, § 6-320.
 This Court does not reach the same conclusion as the Court in Parrish, but has found the discussion there very helpful nonetheless.
 See also Parrish v. Seamans, 343 F. Supp. 1087 (D.S.C., 1972), a case best explained as following Carter.
 Since the joinder of all the class members, absent the class action device, is not essential to the determination of the claims of the named plaintiffs in this action, the claims of the class are not joint or common. This is the test of a true class suit given by Professor Moore at ¶ 23.08(1) of Moore's Federal Practice (2d ed.) Under Snyder v. Harris, 394 U.S. 332, 89 S. Ct. 1053, 22 L. Ed. 2d 319 (1969), the claims of individual class members in such a situation may not be aggregated for purposes of determining the amount in controversy. This applies to actions under the Tucker Act as well. United States v. Louisville and Nashville Railroad Co., 221 F.2d 698 (6th Cir. 1955); Northern Natural Gas Co. v. Grounds, 292 F. Supp. 619 (D.Kan., 1968), reversed on other grounds 441 F.2d 704 (10th Cir., 1971).
 Since this case concerns jurisdiction in the technical sense, the defendants have made no claim that that branch of the doctrine of exhaustion which works a waiver of an issue for failure to pursue a remedy available in the past is at all applicable in this case. But see Mr. Justice Douglas' separate opinion in Gosa v. Mayden, 413 U.S. 665, at 686, 93 S. Ct. 2926 at 2939, 37 L. Ed. 2d 873 at 891 (1973).
 This is the parties' ballpark estimate of the size of the class on the present case. The size of the class may have been constricted somewhat by the Court's action in redefining the class infra.
 This is well substantiated by the observation of Lord Camden in Entick v. Carrington, 95 Eng.Rep.R. 807 (1765), that the exercise by the Crown of the assumed power to issue general warrants for 70 years can only be explained by the fact that prior to the Entick case, such warrants had only been used against those without the means or the sophistication to resort to the courts: "It must have been the guilt or poverty of those upon whom such warrants have been executed, that deferred or hindered them from contending against the power of the Secretary of State and the Solicitor of the Treasury, or such warrants could never have passed for lawful till this time." id at 818.
 It is somewhat ironic that the economic impact which is the major factor militating in favor of prospectivity in this case might have gone unnoticed if the issue had not been raised in the context of a class action.
 The relevance of the Ortiz case to this case should be clarified at this point. The defendants have contended that even if Greenwell is correct and applied retroactively, recovery must be denied plaintiffs because Ortiz was incorrectly decided and 10 U.S. § 823(a) (6) provides an independent direct congressional grant of authority to the courts-martial in question. Because of the way this case has been decided, we need not discuss this issue at length except to indicate that, as in the case of Greenwell, this Court has made an independent evaluation of the Ortiz decision and is in substantial agreement with it. Section 823(a) (6) may not originally have been intended to grant special court-martial convening authority to any command completely within a single service. The structure of § 823(a) (6) suggests that it is a unification provision, a saving clause to grant special court-martial convening authority to joint units analogous to that granted units of the individual services in § 823(a) (2), (a) (3), (a) (4) and (a) (5). Thus read, the word "command" should not be read to indicate a unit smaller than the "battalion or corresponding unit" of the other sections. Consequently, the pre-Ortiz use of § 823(a) (6) was mistaken. There is no direct legislative history on the role of § 823(a) (6) as a unification provision, but the following quotation from the prepared statement of Prof. E. M. Morgan of Harvard, drafter of the Uniform Code of Military Justice, to the Senate Armed Services Subcommittee is helpful in elucidating the reasons for and uses of unification provisions, even though it relates to 10 U.S.C. § 817, not § 823:
"Most of the articles consist of a rewording and revision of provisions found at present in both the Articles of War and the Articles for the Government of the Navy. Article 17, however, is new in that it provides reciprocal jurisdiction of courts martial. By its terms, each armed force shall have court martial jurisdiction over all persons subject to the uniform code. There is thus provided authority for any Army court martial to try either its own personnel or the personnel of the Navy, the Air Force, or the Coast Guard. It is felt that this provision is necessary in the light of unification and by virtue of the tendency to have military operations undertaken by joint forces. Inasmuch as it is not possible at this time to forecast the different forms of joint operation which will take place in the future, the exercise of the reciprocal jurisdiction of one armed force over the personnel of other services has been left to the regulations of the President. In this way a desirable flexibility is attained which will enable the President to prescribe the types of operations in which reciprocal jurisdiction will be exercised." Hearings on S. 857 and H.R. 4080 (UCMS) before a Subcommittee of the Senate Committee on the Armed Services, 81st Congress, 1st Session, p. 35 (1949).
It appears to this Court that the main purpose of § 823(a) (6) was to foreclose the argument by persons tried by a special court-martial of a joint operations unit analogous to a detached battalion that, since the unit was not a unit of any one of the armed forces individually, no authority for special courts-martial existed under § 823(a) (2)-(a) (5). It was not intended, as the Ortiz Court properly concluded, to extend special court-martial convening authority below the detached battalion level within a single service.