Opinion of the Court
COOK, Judge:Tried by special court-martial, military judge alone, the accused was convicted, despite his pleas, of conspiracy to commit larceny, in violation of Article 81, Uniform Code of Military Justice, 10 U.S.C. § 881, and several other offenses.1 The adjudged sentence was approved by the convening authority. One month later, the accused’s only alleged co-conspirator was acquitted of the same conspiracy charge. Upon being advised of that event, the United States Army Court of Military Review, on the basis of our decision in United States v. Nathan, 12 U.S.C.M.A. 398, 30 C.M.R. 398 (1961), set aside the findings of guilty of conspiracy and reassessed the sentence.2 The Judge Advocate General of the Army then certified the following issue to this Court:
*53WHETHER MILITARY LEGAL AUTHORITY, UNITED STATES V. NATHAN, 12 USCMA 398, 30 CMR 398 (1961), AS APPLIED BY THE UNITED STATES ARMY COURT OF MILITARY REVIEW AND WHICH REQUIRES THE DISMISSAL OF AN OTHERWISE VALID CONVICTION FOR CONSPIRACY ONLY BECAUSE THE REMAINING SOLE CO-CONSPIRATOR HAS BEEN ACQUITTED, IS STILL VALID AND SHOULD BE APPLIED WHERE THE CURRENT TREND IN CIVILIAN LEGAL AUTHORITY WOULD WARRANT AFFIRMANCE OF THE CONVICTION.
In United States v. Nathan, supra, we held:
The heart of a conspiracy charge is the allegation of an illegal agreement between two or more persons____ A person cannot conspire with himself. Consequently, it has long been the rule that the acquittal of one of two conspirators requires the acquittal of the other.
We agree that the concepts stated therein are not in accord with current legal authority and should no longer be followed.
In Nathan, both of the accused’s co-conspirators were acquitted of the conspiracy at earlier trials, but they appeared as government witnesses at Nathan’s trial and testified against him as to the existence of the illegal agreement. We determined that “such a collateral attack on ... the validity of what has been judicially determined in the prosecution to which they were parties” was not permissible. Id. at 399, 30 C.M.R. at 399. In reaching this conclusion, we declined to follow the decision of the Nebraska Supreme Court in Platt v. State, 143 Neb. 131, 8 N.W.2d 849 (1943).
Following this same rationale in United States v. Kidd, 13 U.S.C.M.A. 184,32 C.M.R. 184 (1962), we held that the subsequent acquittal on the merits of all the alleged co-conspirators required the acquittal of the-sole convicted conspirator. In United States v. Fisher, 16 U.S.C.M.A. 78, 36 C.M.R. 234 (1966), we held that even where a different overt act was alleged in furtherance of the conspiracy, the acquittal on the merits of the only other conspirator required appellate reversal of the accused’s conviction. Finally, in United States v. Smith, 20 U.S.C.M.A. 589, 44 C.M.R. 19 (1971), we held that even though there were approximately 15 persons engaged in the conspiracy to assault the victim, the fact that the accused was alleged to have conspired with only two named individuals, who were acquitted of the conspiracy charge, required reversal of his conviction.
In reaching these decisions, we followed the so-called “bilateral theory” of conspiracy, or the “rule of consistency,” a rule brought to this country from the British common law and generally followed by most of the jurisdictions here at that time.3 See 4 Wharton’s Criminal Law § 733 (C. Torcia, 14th ed. 1981); Annot., 91 A.L.R.2d 700 (1963); Wharton’s Criminal Law and Procedure, Conspiracy § 91 (Anderson ed. 1957; United States v. Espinosa-Cerpa, 630 F.2d 328, 330-33 (5th Cir.1980). The logic of Platt v. State, supra, however, has had remarkable vitality and has persuaded many courts to depart from the older rule.
In that case, Platt, an employee of the University of Nebraska, was charged with *54conspiring with his secretary to defraud the State by making claims for services of employees which were never performed. Platt was tried first and convicted. Thereafter, his secretary was tried, but the jury found her not guilty. Platt then argued on appeal that since the only other alleged conspirator had been acquitted, he, too, was entitled to be acquitted. His argument was based on an earlier decision of the Nebraska Supreme Court to that effect, see Sherman v. State, 113 Neb. 173, 202 N.W. 413 (Neb. 1925). The Nebraska Supreme Court determined that a review of the earlier decision was needed. It traced the history of the rule of consistency to the early English cases where persons charged with conspiracy were tried together before the same court and jury, and with the same evidence. The principle applied there was that verdicts must be consistent and devoid of repugnancy. Consequently, the conviction of one and the acquittal of the other would result in inconsistent verdicts. However, the court could find “no previous case holding that verdicts on separate trials must be consistent as to result in such cases.” 8 N.W.2d at 855. Instead, the court concluded that the earlier cases were based on a fallacy: that the acquittal of one conspirator at a separate trial necessarily meant that there was, in fact, no conspiracy. To the contrary, the court concluded that the conviction of the one conspirator required that the fact of the conspiracy and the guilt of the other conspirator must have been established to the satisfaction of the convicting jury, at least for the purpose of that trial. The subsequent acquittal by another jury could as well be occasioned by a failure of proof as to the second conspirator, which could be caused by:
[T]he death or absence of an important state witness, the incompetency of a confession of the convicted conspirator in the second trial, the incompetency of a plea of guilty entered by the convicted conspirator at his trial, or for any other reason that would amount to a failure of proof.
8 N.W.2d at 855. Absent a joint trial upon identical evidence, the court could find no logic in requiring consistency of results in separate trials of conspirators. They, therefore, overruled Sherman v. State, supra, and affirmed Platt’s conviction.
Over a decade later, the drafters of the American Law Institute’s Model Penal Code adopted the “Unilateral Approach” to conspiracy. Rejecting “the traditional view of conspiracy ... cast in terms of ‘two or more persons,”’ the Model Penal Code instead focused on “each individual’s culpability by framing the definition [of conspiracy] in terms of the conduct which suffices to establish the liability of any given actor, rather than the conduct of a group of which he is charged to be a part.” Modern Penal Code § 5.03, p. 104 (Tentative Draft # 10) (1960). The drafters of the Model Penal Code defined “conspiracy” in this manner:
A person is guilty of conspiracy with another person or persons to commit a crime if with the purpose of promoting or facilitating its commission he:
(a) agrees with such other person or persons that they or one or more of them will engage in conduct which constitutes such crime or an attempt or solicitation to commit such crime; or
(b) agrees to aid such other person or persons in the planning or commission of such crime or of an attempt or solicitation to commit such crime.
Id. at 19-20.
Section 5.04 of the Model Penal Code supplements this definition as follows:
(1) [I]t is immaterial to the liability of a person who solicits or conspires with another to commit a crime that:
(b) the person whom he solicits or with whom he conspires is irresponsible or has an immunity to prosecution or conviction for the commission of the crime.
Id. at 22.
The concept underlying the Model Penal Code definitions has been adopted, in whole or in part, in a least 26 states, and its adoption is under consideration in most of the others. Garcia v. State, 394 N.E.2d 106, 109 (Ind.1979).
*55The British also reversed their older precedents and rejected the “bilateral theory” of conspiracy in the case of Director of Public Prosecutions v. Shannon, 59 Crim. App.R. 250 (House of Lords 1974). There, Shannon pleaded guilty to conspiring only to handle stolen goods with T. Shannon was sentenced based on his plea. T was tried separately for a related offense and was acquitted. The Crown decided to offer no evidence against him on the conspiracy count, and the recorder entered a verdict of not guilty. Shannon then appealed his conviction and sentence to the Court of Appeal. That Court, following earlier precedents, quashed the conviction, and the Crown appealed to the House of Lords. Although the decision was based, in part, on statutory revisions of the right of criminal appeal, it does provide clear insight into the reasons for the adoption of the “bilateral” rule (or “rule of consistency”) and for its rejection:
It must be a self-evident proposition that one man acting alone cannot conspire and therefore cannot be guilty of conspiring. It is the very essence of a conspiracy that a person agrees with some other person or persons.
Consequently, if two conspirators are tried together on the same evidence and before the same jury, the conspiracy must be proven against them both; “there would be something wrong if a jury returned different verdicts.” On the other hand, if the two were tried separately, or if the evidence was not equally admissible against both, then there would be no reason for applying the rule. However, in the days when the only form of appeal was by writ of error (i.e., prior to the abolition of the writ of error by the Criminal Appeal Act of 1907), the only “record” which was brought forward for appeal consisted of “the arraignment, the plea, the issue and the verdict”; and there was “no account of some of the most material parts of the trial where error was most likely to occur — viz. the evidence and the direction of the judge to the jury.” Id. at 257 (Opinion of Lord Morris of Borth-y-Gest, concurred in by Lord Reid). Thus, “an inconsistency on the record pointed to the conclusion that a conviction was wrong.” Id. at 261. But, since here Shannon was tried separately from T, the reasons for the rule would not lie.
The other Law Lords went even further. Viscount Dilhome would not apply the rule where there were separate cases and one conspirator pleaded guilty “and it cannot be said that he has done so mistakenly.” Id. at 265. Lord Simon of Glaisdale would abolish the rule entirely and follow Platt v. State, supra, because:
The respondent pleaded Guilty, fully understanding the situation and professionally advised. His conviction was not relevant to (and therefore not admissible) evidence to prove Tracey’s guilt: why should Tracey’s acquittal be relevant to the respondent’s conviction? ... Such absurdities bring the law into discredit, and mean that rogues escape society’s sanctions.
Id. at 271.4 The net result was to overrule the older cases and to disapprove application of the rule in a per se manner.
In similarly rejecting the “bilateral” concept, the court in United States v. Espinosa-Cerpa, supra at 332-33 n. 5, made these comments about Director of Public Prosecutions v. Shannon, supra, and the basis for the doctrine of consistency:
*56A brief examination of the origin and evolution of the rule in ... [the English] system provides perhaps the most damning indictment of its continued use in our system.
* * * * * *
This exposition in R. v. Shannon [sic] of the historical basis for the rule makes abundantly clear its inappropriateness to a modern American criminal justice system in which all verdicts obviously are, and always have been, subject to independent review for evidentiary support and where apparently inconsistent verdicts are almost uniformly tolerated whether explainable by variances in proof or merely as jury license.
Since the enactment of the Uniform Code of Military Justice, there has been a verbatim record of trial available for review, and the mandatory review by the Boards of Review and Courts of Military Review require an examination of the evidence and a ruling that the findings and sentence were “correct in law and fact” as a condition to their affirmance. See Article 66(c), UCMJ, 10 U.S.C. § 866(c). Consequently, there was not then, nor is there now, any need to slavishly impose a per se standard to the inconsistent verdicts5 as to several conspirators. Further, the concept that the conviction of two or more conspirators is necessary to prove that a conspiracy did exist is based on the often incorrect premise that the evidence against all the conspirators was identical,6 the instructions were the same, and each jury’s reasoning was indistinguishable from the others. The contrary conclusion that different juries may believe or disbelieve the same witness, give more or less weight to the evidence, or even act with differing degrees of compassion or logic when trying different accused seems to be far better grounded in empirical logic.7 *57A verdict of “not guilty” may as well mean only “not proven” as to the particular accused.
Our present system of judicial review is quite capable of considering the evidence against each and all the named conspirators, even where they are tried separately, and deciding the criminal responsibility of each without resort to a rule8 that must often result in a miscarriage of justice simply to avoid the mere possibility of a “wrong verdict.” The law does not now demand such “foolish consistency.” 9 We, therefore, will no longer follow the decision and rationale of United States v. Nathan, supra, and its progeny.10 In the future, the conviction of an individual conspirator in a separate trial11 will be considered to ensure that the evidence supports beyond a reasonable doubt his complicity in the conspiracy; but, the acquittal of his co-conspirators will not serve to avoid his conviction in the absence of some compelling reason of record in the other cases.12
The Court unanimously answers the certified question in the negative. Since the Court of Military Review correctly applied the judicial and Manual authority in effect at the time of its decision, I would affirm the decision. However, Chief Judge EVERETT and Judge FLETCHER disagree and would reverse that decision. Accordingly, the decision of the United States Army Court of Military Review is reversed. The record of trial is returned to the Judge Advocate General of the Army for submission to that court for further review under Article 66.
. Contrary to his pleas, the accused was convicted of conspiracy to commit larceny, two absences without proper authority, larceny, and unlawful entry, in violation of Articles 81, 86, 121, and 134, Uniform Code of Military Justice, 10 U.S.C. §§ 881, 886, 921 and 934, respectively. The military judge sentenced him to a bad-conduct discharge, confinement at hard labor for 6 months, forfeiture of $334.00 pay per month for 6 months, and reduction to private (E-l).
. The Court of Military Review reduced the period of forfeitures and confinement at hard labor by 1 month.
. The rule was followed by the United States Supreme Court in Hartzel v. United States, 322 U.S. 680, 64 S.Ct. 1233, 88 L.Ed. 1534 (1944). Without elaboration or citation of authority, the Court dismissed a conspiracy conviction where the other two co-conspirators’ convictions had been set aside (apparently for evidentiary-insufficiency). Id. at 682 n. 3, 64 S.Ct. at 1234 n. 3. However, in Harris v. Rivera, 454 U.S. 339, 345 n. 13, 102 S.Ct. 460, 464 n. 13, 70 L.Ed.2d 530 (1981), the Court contrasted the Hartzel holding with the decision in Standefer v. United States, 447 U.S. 10,100 S.Ct. 1999, 64 L.Ed.2d 689 (1980), which sustained the conviction of an accused for “aiding and abetting in the commission of a federal offense ... after the named principal ha[d] been acquitted of ... (the same] offense in a previous trial.” This is the rule in military justice. See United States v. Marsh, 13 U.S.C.M.A. 252, 32 C.M.R. 252 (1962). However, there is a significant difference in the theories underlying this holding and those underlying conspiracy since aiding and abetting is not a crime in and of itself while conspiracy is. See United States v. Pearson, 667 F.2d 12 (5th Cir.1982).
. In Director of Public Prosecutions v. Shannon, 59 Crim.App.R. 250, 268 (House of Lords 1974), Lord Simon of Glaisdale stated:
The foundations of this legal superstructure contained, however, a flaw. The law in action is not concerned with absolute truth, but with proof before a fallible human tribunal to a requisite standard of probability in accordance with formal rules of evidence (in particular, rules relating to admissibility of evidence). No doubt, in the realm of the absolute, A could not conspire with B without B also conspiring with A. But it by no means follows that it cannot be proved forensically that A conspired with B (necessarily involving that it is also proved, as against A, that B conspired with him) notwithstanding a total failure of forensic proof, as against B, that B conspired with A (necessarily involving a failure of proof, as against B, that A conspired with him)____ The fallacious basis of the old rule was also forcibly demonstrated in the judgment of the Supreme Court of Nebraska in Platt v. State (1943) 8 N.W.2d 849.
. No doubt because of the availability of much more of the record of trial for consideration on appeal, the courts of this country have accepted the inevitability of inconsistent verdicts as part of the jury system.
Consistency in the verdict is not necessary.
He * * * * *
That the verdict may have been the result of compromise, or of a mistake on the part of the jury, is possible. But verdicts cannot be upset by speculation or inquiry into such matters.
Dunn v. United States, 284 U.S. 390, 393-94, 52 S.Ct. 189, 190-191, 76 L.Ed. 356 (1932). Even purportedly inconsistent verdicts as to co-defendants in a bench trial do not amount to constitutional error absent the indication that the inconsistencies “might constitute evidence of arbitrariness that would undermine confidence in the quality of the judge’s conclusion.” Harris v. Rivera, 454 U.S. 339, 346, 102 S.Ct. 460, 464-65, 70 L.Ed.2d 530 (1981).
. See Director of Public Prosecutions v. Shannon, supra at 268-69:
[A]dultery is an unlawful meeting of bodies as conspiracy is an unlawful meeting of minds. In the realm of the absolute A cannot commit adultery with B without B also committing adultery with A. But it may well be proved forensically that A committed adultery with B (necessarily involving that it is also proved, as against A, that B committed adultery with her), notwithstanding a total failure of forensic proof, as against B, that B committed adultery with A (necessarily involving a failure of proof, as against B, that A committed adultery with him).
. Steckler v. United States, 7 F.2d 59 (2d Cir. 1925), states:
The most that can be said in such cases is that the verdict shows that either in the acquittal or the conviction the jury did not speak their real conclusions, but that does not show that they were not convinced of the defendant’s guilt. We interpret the acquittal as no more than their assumption of a power which they had no right to exercise, but to which they were disposed through lenity.
Id. at 60, quoted in Dunn v. United States, supra 284 U.S. at 393, 52 S.Ct. at 190.
Even the unlikely possibility that the acquittal is the product of a lenity that judges are free to exercise at the time of sentencing but generally are forbidden to exercise when ruling on guilt or innocence, would not create a constitutional violation____ The Constitution does not prohibit state judges from being excessively lenient.
Harris v. Rivera, supra 454 U.S. at 348, 102 S.Ct. at 465. Thus, whatever may be the cause of the inconsistent verdicts, be it logical or illogical, erroneous or not, an acquittal does not constitutionally mandate reversal of an otherwise valid legal conviction, even in multiparty crimes.
That the conviction may have been the result of some compromise is, of course, possible; but to consider so is to consider too curiously, unless all verdicts are to be upset on speculation. That it represented their de*57liberate judgment seems to us beyond any reasonable doubt.
Steckler v. United States, supra at 60.
. The so-called doctrine of “nonmutual collateral estoppel” is sometimes used to rationalize the “bilateral theory.” However, its use in criminal cases has been disapproved. See United States v. Espinosa-Cerpa, 630 F.2d 328, 330 (5th Cir.1980). In Standefer v. United States, supra, the Supreme Court found no constitutional impediment to conviction of an aider and abettor after acquittal of the principal in the crime. Rejecting the doctrine of “nonmutual estoppel,” the Court concluded that:
[T]his criminal case involves “competing policy considerations” that outweigh the economy concerns that undergird the estoppel doctrine. [Citation omitted.] This case does no more than manifest the simple, if discomforting, reality that “different juries may reach different results under any criminal statute. That is one of the consequences we accept under our jury system.” [Citation omitted.] While symmetry of results may be intellectually satisfying, it is not required. [Citation omitted.]
Id. 447 U.S. at 25, 100 S.Ct. at 2008.
. “A foolish consistency is the hobgoblin of little minds.” R.W. Emerson, Essays, “Self Reliance.”
. Reason is the soul of law; the reason of law being changed, the law is also changed. 7 Coke 7 (translated).
. We do not now decide the question of whether inconsistent verdicts as to co-conspirators in a joint trial may be treated separately.
. Our decision here undermines the ninth paragraph of paragraph 160, Manual for Courts-Martial, United States, 1969 (Revised edition), which was added to conform to our decisions in United States v. Doughty, 14 U.S. C.M.A. 540, 34 C.M.R. 320 (1964); United States v. Kidd, 13 U.S.C.M.A. 184, 32 C.M.R. 184 (1962); and United States v. Nathan, 12 U.S.C.M.A. 398, 30 C.M.R. 398 (1961). Analysis of Contents, Manual for Courts-Martial, United States, 1969 (Revised edition), para. 160.