444 U.S. 507 (1980)
Supreme Court of United States.
Decided February 19, 1980.[*]
ON PETITION FOR WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT.PER CURIAM.
In No. 78-1871, Frank W. Snepp III seeks review of a judgment enforcing an agreement that he signed when he accepted employment with the Central Intelligence Agency (CIA). He also contends that punitive damages are an inappropriate remedy for the breach of his promise to submit all writings about the Agency for prepublication review. In No. 79-265, the United States conditionally cross petitions from a judgment refusing to find that profits attributable to Snepp's breach are impressed with a constructive trust. We grant the petitions for certiorari in order to correct the judgment from which both parties seek relief.
Based on his experiences as a CIA agent, Snepp published a book about certain CIA activities in South Vietnam. Snepp published the account without submitting it to the Agency for prepublication review. As an express condition of his employment with the CIA in 1968, however, Snepp had executed an agreement promising that he would "not . . . publish . . . any information or material relating to the Agency, its activities or intelligence activities generally, either during or after the term of [his] employment . . . without specific prior approval by the Agency." App. to Pet. for Cert. in No. 78-1871, p. 59a. The promise was an integral part of Snepp's concurrent undertaking "not to disclose any classified information relating to the Agency without proper authorization." Id., at 58a. Thus, Snepp had pledged not to divulge classified information and not to publish any information without prepublication clearance. The Government brought this suit to enforce Snepp's agreement. It sought a declaration that Snepp had breached the contract, an injunction requiring Snepp to submit future writings for prepublication review, and an order imposing a constructive trust for the Government's benefit on all profits that Snepp might earn from publishing the book in violation of his fiduciary obligations to the Agency.
The District Court found that Snepp had "willfully, deliberately and surreptitiously breached his position of trust with the CIA and the  secrecy agreement" by publishing his book without submitting it for prepublication review. 456 F. Supp. 176, 179 (ED Va. 1978). The court also found that Snepp deliberately misled CIA officials into believing that he would submit the book for prepublication clearance. Finally, the court determined as a fact that publication of the book had "caused the United States irreparable harm and loss." Id., at 180. The District Court therefore enjoined future breaches of Snepp's agreement and imposed a constructive trust on Snepp's profits.
The Court of Appeals accepted the findings of the District Court and agreed that Snepp had breached a valid contract. It specifically affirmed the finding that Snepp's failure to submit his manuscript for prepublication review had inflicted "irreparable harm" on intelligence activities vital to our national security. 595 F. 2d 926, 935 (CA4 1979). Thus, the court upheld the injunction against future violations of Snepp's prepublication obligation. The court, however, concluded that the record did not support imposition of a constructive trust. The conclusion rested on the court's perception that Snepp had a First Amendment right to publish unclassified information and the Government's concession for the purposes of this litigationthat Snepp's book divulged no classified intelligence. Id., at 935-936. In other words, the court thought that Snepp's fiduciary obligation extended only to preserving the confidentiality of classified material. It therefore limited recovery to nominal damages and to the possibility of punitive damages if the Governmentin a jury trialcould prove tortious conduct.
Judge Hoffman, sitting by designation, dissented from the refusal to find a constructive trust. The 1968 agreement, he wrote, "was no ordinary contract; it gave life to a fiduciary relationship and invested in Snepp the trust of the CIA." Id., at 938. Prepublication clearance was part of Snepp's undertaking to protect confidences associated with his trust. Punitive damages, Judge Hoffman argued, were both a speculative and inappropriate remedy for Snepp's breach. We agree with Judge Hoffman that Snepp breached a fiduciary obligation and that the proceeds of his breach are impressed with a constructive trust.
Snepp's employment with the CIA involved an extremely high degree of trust. In the opening sentence of the agreement that he signed, Snepp explicitly recognized that he was entering a trust relationship. The trust agreement specifically imposed the obligation not to publish any information relating to the Agency without submitting the information for clearance. Snepp stipulated at trial thatafter undertaking this obligationhe had been "assigned to various positions of trust" and that he had been granted "frequent access to classified information, including information regarding intelligence sources and methods." 456 F. Supp., at 178. Snepp published his book about CIA activities on the basis of this background and exposure. He deliberately and surreptitiously violated his obligation to submit all material for prepublication review. Thus, he exposed the classified information with which he had been entrusted to the risk of disclosure.
Whether Snepp violated his trust does not depend upon whether his book actually contained classified information. The Government does not denyas a general principle Snepp's right to publish unclassified information. Nor does it contendat this stage of the litigationthat Snepp's book contains classified material. The Government simply claims that, in light of the special trust reposed in him and the agreement that he signed, Snepp should have given the CIA an opportunity to determine whether the material he proposed to publish would compromise classified information or sources. Neither of the Government's concessions undercuts its claim that Snepp's failure to submit to prepublication review was a breach of his trust.
Both the District Court and the Court of Appeals found that a former intelligence agent's publication of unreviewed material relating to intelligence activities can be detrimental to vital national interests even if the published information is unclassified. When a former agent relies on his own judgment about what information is detrimental, he may reveal information that the CIAwith its broader understanding of what may expose classified information and confidential sourcescould have identified as harmful. In addition to receiving intelligence from domestically based or controlled sources, the CIA obtains information from the intelligence services of friendly nations and from agents operating in foreign countries. The continued availability of these foreign sources depends upon the CIA's ability to guarantee the security of information that might compromise them and even endanger the personal safety of foreign agents.
Undisputed evidence in this case shows that a CIA agent's violation of his obligation to submit writings about the Agency for prepublication review impairs the CIA's ability to perform its statutory duties. Admiral Turner, Director of the CIA, testified without contradiction that Snepp's book and others like it have seriously impaired the effectiveness of American intelligence operations. He said:
"Over the last six to nine months, we have had a number of sources discontinue work with us. We have had more sources tell us that they are very nervous about continuing work with us. We have had very strong complaints from a number of foreign intelligence services with whom we conduct liaison, who have questioned whether they should continue exchanging information with us, for fear it will not remain secret. I cannot estimate to you how many potential sources or liaison arrangements have never germinated because people were unwilling to enter into business with us." 456 F. Supp., at 179-180.
In view of this and other evidence in the record, both the District Court and the Court of Appeals recognized that Snepp's breach of his explicit obligation to submit his material classified or notfor prepublication clearance has irreparably harmed the United States Government. 595 F. 2d, at 935; 456 F. Supp., at 180.
The decision of the Court of Appeals denies the Government the most appropriate remedy for Snepp's acknowledge wrong. Indeed, as a practical matter, the decision may well leave the Government with no reliable deterrent against similar breaches of security. No one disputes that the actual damages attributable to a publication such as Snepp's generally are unquantifiable. Nominal damages are a hollow alternative, certain to deter no one. The punitive damages recoverable after a jury trial are speculative and unusual. Even if recovered, they may bear no relation to either the Government's irreparable loss or Snepp's unjust gain.
The Government could not pursue the only remedy that the Court of Appeals left it without losing the benefit of the bargain it seeks to enforce. Proof of the tortious conduct necessary to sustain an award of punitive damages might force the Government to disclose some of the very confidences that Snepp promised to protect. The trial of such a suit, before a jury if the defendant so elects, would subject the CIA and its officials to probing discovery into the Agency's highly confidential affairs. Rarely would the Government run this risk. In a letter introduced at Snepp's trial, former CIA Director Colby noted the analogous problem in criminal cases. Existing law, he stated, "requires the revelation in open court of confirming or additional information of such a nature that the potential damage to the national security precludes prosecution." App. to Pet. for Cert. in No. 78-1871, p. 68a. When the Government cannot secure its remedy without unacceptable risks, it has no remedy at all.
A constructive trust, on the other hand, protects both the Government and the former agent from unwarranted risks. This remedy is the natural and customary consequence of a breach of trust. It deals fairly with both parties by conforming relief to the dimensions of the wrong. If the agent secures prepublication clearance, he can publish with no fear of liability. If the agent publishes unreviewed material in violation of his fiduciary and contractual obligation, the trust remedy simply requires him to disgorge the benefits of his faithlessness. Since the remedy is swift and sure, it is tailored to deter those who would place sensitive information at risk. And since the remedy reaches only funds attributable to the breach, it cannot saddle the former agent with exemplary damages out of all proportion to his gain. The decision of the Court of Appeals would deprive the Government of this equitable and effective means of protecting intelligence that may contribute to national security. We therefore reverse the judgment of the Court of Appeals insofar as it refused to impose a constructive trust on Snepp's profits, and we remand the cases to the Court of Appeals for reinstatement of the full judgment of the District Court.
MR. JUSTICE STEVENS, with whom MR. JUSTICE BRENNAN and MR. JUSTICE MARSHALL join, dissenting.
In 1968, Frank W. Snepp signed an employment agreement with the CIA in which he agreed to submit to the Agency any information he intended to publish about it for prepublication review. The purpose of such an agreement, as the Fourth Circuit held, is not to give the CIA the power to censor its employees' critical speech, but rather to ensure that classified, nonpublic information is not disclosed without the Agency's permission. 595 F. 2d 926, 932 (1979); see also United States v. Marchetti, 466 F. 2d 1309, 1317 (CA4 1972), cert. denied, 409 U. S. 1063.
In this case Snepp admittedly breached his duty to submit the manuscript of his book, Decent Interval, to the CIA for prepublication review. However, the Government has conceded that the book contains no classified, nonpublic material. Thus, by definition, the interest in confidentiality that Snepp's contract was designed to protect has not been compromised. Nevertheless, the Court today grants the Government unprecedented and drastic relief in the form of a constructive trust over the profits derived by Snepp from the sale of the book. Because that remedy is not authorized by any applicable law and because it is most inappropriate for the Court to dispose of this novel issue summarily on the Government's conditional cross-petition for certiorari, I respectfully dissent.
The rule of law the Court announces today is not supported by statute, by the contract, or by the common law. Although Congress has enacted a number of criminal statutes punishing the unauthorized dissemination of certain types of classified information, it has not seen fit to authorize the constructive trust remedy the Court creates today. Nor does either of the contracts Snepp signed with the Agency provide for any such remedy in the event of a breach. The Court's per curiam opinion seems to suggest that its result is supported by a blend of the law of trusts and the law of contracts. But neither of these branches of the common law supports the imposition of a constructive trust under the circumstances of this case.
Plainly this is not a typical trust situation in which a settlor has conveyed legal title to certain assets to a trustee for the use and benefit of designated beneficiaries. Rather, it is an employment relationship in which the employee possesses fiduciary obligations arising out of his duty of loyalty to his employer. One of those obligations, long recognized by the common law even in the absence of a written employment agreement, is the duty to protect confidential or "classified" information. If Snepp had breached that obligation, the common law would support the implication of a constructive trust upon the benefits derived from his misuse of confidential information.
But Snepp did not breach his duty to protect confidential information. Rather, he breached a contractual duty, imposed in aid of the basic duty to maintain confidentiality, to obtain prepublication clearance. In order to justify the imposition of a constructive trust, the majority attempts to equate this contractual duty with Snepp's duty not to disclose, labeling them both as "fiduciary." I find nothing in the common law to support such an approach.
Employment agreements often contain covenants designed to ensure in various ways that an employee fully complies with his duty not to disclose or misuse confidential information. One of the most common is a covenant not to compete. Contrary to the majority's approach in this case, the courts have not construed such covenants broadly simply because they support a basic fiduciary duty; nor have they granted sweeping remedies to enforce them. On the contrary, because such covenants are agreements in restraint of an individual's freedom of trade, they are enforceable only if they can survive scrutiny under the "rule of reason." That rule, originally laid down in the seminal case of Mitchel v. Reynolds, 1 P. Wms. 181, 24 Eng. Rep. 347 (1711), requires that the covenant be reasonably necessary to protect a legitimate interest of the employer (such as an interest in confidentiality), that the employer's interest not be outweighed by the public interest, and that the covenant not be of any longer duration or wider geographical scope than necessary to protect the employer's interest.
The Court has not persuaded me that a rule of reason analysis should not be applied to Snepp's covenant to submit to prepublication review. Like an ordinary employer, the CIA has a vital interest in protecting certain types of information; at the same time, the CIA employee has a countervailing interest in preserving a wide range of work opportunities (including work as an author) and in protecting his First Amendment rights. The public interest lies in a proper accommodation that will preserve the intelligence mission of the Agency while not abridging the free flow of unclassified information. When the Government seeks to enforce a harsh restriction on the employee's freedom, despite its admission that the interest the agreement was designed to protectthe confidentiality of classified informationhas not been compromised, an equity court might well be persuaded that the case is not one in which the covenant should be enforced.
But even assuming that Snepp's covenant to submit to prepublication review should be enforced, the constructive trust imposed by the Court is not an appropriate remedy. If an employee has used his employer's confidential information for his own personal profit, a constructive trust over those profits is obviously an appropriate remedy because the profits are the direct result of the breach. But Snepp admittedly did not use confidential information in his book; nor were the profits from his book in any sense a product of his failure to submit the book for prepublication review. For, even if Snepp had submitted the book to the Agency for prepublication review, the Government's censorship authority would surely have been limited to the excision of classified material. In this case, then, it would have been obliged to clear the book for publication in precisely the same form as it now stands. Thus, Snepp has not gained any profits as a result of his breach; the Government, rather than Snepp, will be unjustly enriched if he is required to disgorge profits attributable entirely to his own legitimate activity.
Despite the fact that Snepp has not caused the Government the type of harm that would ordinarily be remedied by the imposition of a constructive trust, the Court attempts to justify a constructive trust remedy on the ground that the Government has suffered some harm. The Court states that publication of "unreviewed material" by a former CIA agent "can be detrimental to vital national interests even if the published information is unclassified." Ante, at 511-512. It then seems to suggest that the injury in such cases stems from the Agency's inability to catch "harmful" but unclassified information before it is published. I do not believe, however, that the Agency has any authority to censor its employees' publication of unclassified information on the basis of its opinion that publication may be "detrimental to vital national interests" or otherwise "identified as harmful." Ibid. The CIA never attempted to assert such power over Snepp in either of the contracts he signed; rather, the Agency itself limited its censorship power to preventing the disclosure of "classified" information. Moreover, even if such a wide-ranging prior restraint would be good national security policy, I would have great difficulty reconciling it with the demands of the First Amendment.
The Court also relies to some extent on the Government's theory at trial that Snepp caused it harm by flouting his prepublication review obligation and thus making it appear that the CIA was powerless to prevent its agents from publishing any information they chose to publish, whether classified or not. The Government theorized that this appearance of weakness would discourage foreign governments from cooperating with the CIA because of a fear that their secrets might also be compromised. In support of its position that Snepp's book had in fact had such an impact, the Government introduced testimony by the Director of the CIA, Admiral Stansfield Turner, stating that Snepp's book and others like it had jeopardized the CIA's relationship with foreign intelligence services by making them unsure of the Agency's ability to maintain confidentiality. Admiral Turner's truncated testimony does not explain, however, whether these unidentified "other" books actually contained classified information. If so, it is difficult to believe that the publication of a book like Snepp's which does not reveal classified information, has significantly weakened the Agency's position. Nor does it explain whether the unidentified foreign agencies who have stopped cooperating with the CIA have done so because of a legitimate fear that secrets will be revealed or because they merely disagree with our Government's classification policies.
In any event, to the extent that the Government seeks to punish Snepp for the generalized harm he has caused by failing to submit to prepublication review and to deter others from following in his footsteps, punitive damages is, as the Court of Appeals held, clearly the preferable remedy "since a constructive trust depends on the concept of unjust enrichment rather than deterrence and punishment. See D. Dobbs, Law of Remedies § 3.9 at 205 and § 4.3 at 246 (1973)." 595 F. 2d, at 937.
The Court's decision to dispose of this case summarily on the Government's conditional cross-petition for certiorari is just as unprecedented as its disposition of the merits.
Snepp filed a petition for certiorari challenging the Fourth Circuit's decision insofar as it affirmed the entry of an injunction requiring him to submit all future manuscripts for prepublication review and remanded for a determination of whether punitive damages would be appropriate for his failure to submit Decent Interval to the Agency prior to its publication. The Government filed a brief in opposition as well as a cross-petition for certiorari; the Government specifically stated, however, that it was cross petitioning only to bring the entire case before the Court in the event that the Court should decide to grant Snepp's petition. The Government explained that "[b]ecause the contract remedy provided by the court of appeals appears to be sufficient in this case to protect the Agency's interest, the government has not independently sought review in this Court." In its concluding paragraph the Government stated: "If this Court grants [Snepp's] . . . petition for a writ of certiorari in No. 78-1871, it should also grant this cross-petition. If the petition in No. 78-1871 is denied, this petition should also be denied." Pet. for Cert. in No. 79-265, p. 5.
Given the Government's position, it would be highly inappropriate, and perhaps even beyond this Court's jurisdiction, to grant the Government's petition while denying Snepp's. Yet that is in essence what has been done. The majority obviously does not believe that Snepp's claims merit this Court's consideration, for they are summarily dismissed in a footnote. Ante, at 509, n. 3. It is clear that Snepp's petition would not have been granted on its own merits.
The Court's opinion is a good demonstration of why this Court should not reach out to decide a question not necessarily presented to it, as it has done in this case. Despite the fact that the Government has specifically stated that the punitive damages remedy is "sufficient" to protect its interests, the Court forges ahead and summarily rejects that remedy on the grounds that (a) it is too speculative and thus would not provide the Government with a "reliable deterrent against similar breaches of security." ante, at 514, and (b) it might require the Government to reveal confidential information in court, the Government might forgo damages rather than make such disclosures, and the Government might thus be left with "no remedy at all," ante, at 515. It seems to me that the Court is foreclosed from relying upon either ground by the Government's acquiescence in the punitive damages remedy. Moreover, the second rationale is entirely speculative and, in this case at least, almost certainly wrong. The Court states that
"[p]roof of the tortious conduct necessary to sustain an award of punitive damages might force the Government to disclose some of the very confidences that Snepp promised to protect." Ante, at 514.
Yet under the Court of Appeals' opinion the Government would be entitled to punitive damages simply by proving that Snepp deceived it into believing that he was going to comply with his duty to submit the manuscript for prepublication review and that the Government relied on these misrepresentations to its detriment. I fail to see how such a showing would require the Government to reveal any confidential information or to expose itself to "probing discovery into the Agency's highly confidential affairs." Ante, at 515.
The uninhibited character of today's exercise in lawmaking is highlighted by the Court's disregard of two venerable principles that favor a more conservative approach to this case.
First, for centuries the English-speaking judiciary refused to grant equitable relief unless the plaintiff could show that his remedy at law was inadequate. Without waiting for an opportunity to appraise the adequacy of the punitive damages remedy in this case, the Court has jumped to the conclusion that equitable relief is necessary.
Second, and of greater importance, the Court seems unaware of the fact that its drastic new remedy has been fashioned to enforce a species of prior restraint on a citizen's right to criticize his government. Inherent in this prior restraint is the risk that the reviewing agency will misuse its authority to delay the publication of a critical work or to persuade an author to modify the contents of his work beyond the demands of secrecy. The character of the covenant as a prior restraint on free speech surely imposes an especially heavy burden on the censor to justify the remedy it seeks. It would take more than the Court has written to persuade me that that burden has been met.
I respectfully dissent.
[*] Together with No. 79-265, United States v. Snepp, also on petition for certiorari to the same court.
 Upon the eve of his departure from the Agency in 1976, Snepp also executed a "termination secrecy agreement." That document reaffirmed his obligation "never" to reveal "any classified information, or any information concerning intelligence or CIA that has not been made public by CIA . . . without the express written consent of the Director of Central Intelligence or his representative." App. to Pet. for Cert. in No. 78-1871, p. 61a.
 At the time of suit, Snepp already had received about $60,000 in advance payments. His contract with his publisher provides for royalties and other potential profits. 456 F. Supp. 176, 179 (ED Va. 1978).
 The Court of Appeals and the District Court rejected each of Snepp's defenses to the enforcement of his contract. 595 F. 2d 926, 931-934 (CA4 1979); 456 F. Supp., at 180-181. In his petition for certiorari, Snepp relies primarily on the claim that his agreement is unenforceable as a prior restraint on protected speech.
When Snepp accepted employment with the CIA, he voluntarily signed the agreement that expressly obligated him to submit any proposed publication for prior review. He does not claim that he executed this agreement under duress. Indeed, he voluntarily reaffirmed his obligation when he left the Agency. We agree with the Court of Appeals that Snepp's agreement is an "entirely appropriate" exercise of the CIA Director's statutory mandate to "protec[t] intelligence sources and methods from unauthorized disclosure," 50 U. S. C. § 403 (d) (3). 595 F. 2d, at 932. Moreover, this Court's cases make clear thateven in the absence of an express agreementthe CIA could have acted to protect substantial government interests by imposing reasonable restrictions on employee activities that in other contexts might be protected by the First Amendment. CSC v. Letter Carriers, 413 U. S. 548, 565 (1973); see Brown v. Glines, ante, p. 348; Buckley v. Valeo, 424 U. S. 1, 25-28 (1976); Greer v. Spock, 424 U. S. 828 (1976); id., at 844-848 (POWELL, J., concurring); Cole v. Richardson, 405 U. S. 676 (1972). The Government has a compelling interest in protecting both the secrecy of information important to our national security and the appearance of confidentiality so essential to the effective operation of our foreign intelligence service. See infra, at 511-512. The agreement that Snepp signed is a reasonable means for protecting this vital interest.
 The Government's concession distinguished this litigation from United States v. Marchetti, 466 F. 2d 1309 (CA4), cert. denied, 409 U. S. 1063 (1972). There, the Government claimed that a former CIA employee intended to violate his agreement not to publish any classified information. 466 F. 2d, at 1313. Marchetti therefore did not consider the appropriate remedy for the breach of an agreement to submit all material for prepublication review. By relying on Marchetti in this litigation, the Court of Appeals overlooked the difference between Snepp's breach and the violation at issue in Marchetti.
 The first sentence of the 1968 agreement read: "I, Frank W. Snepp, III, understand that upon entering duty with the Central Intelligence Agency I am undertaking a position of trust in that Agency of the Government. . . ." App. to Pet. for Cert. in No. 78-1871, p. 58a.
 Quite apart from the plain language of the agreement, the nature of Snepp's duties and his conceded access to confidential sources and materials could establish a trust relationship. See 595 F. 2d, at 939 (Hoffman, J., concurring in part and dissenting in part). Few types of governmental employment involve a higher degree of trust than that reposed in a CIA employee with Snepp's duties.
 Every major nation in the world has an intelligence service. Whatever fairly may be said about some of its past activities, the CIA (or its predecessor the Office of Strategic Services) is an agency thought by every President since Franklin D. Roosevelt to be essential to the security of the United States andin a sensethe free world. It is impossible for a government wisely to make critical decisions about foreign policy and national defense without the benefit of dependable foreign intelligence. See generally T. Powers, The Man Who Kept the Secrets (1979).
 In questioning the force of Admiral Turner's testimony, MR. JUSTICE STEVENS' dissenting opinion suggests that the concern of foreign intelligence services may not be occasioned by the hazards of allowing an agent like Snepp to publish whatever he pleases, but by the release of classified information or simply the disagreement of foreign agencies with our Government's classification policy. Post, at 522-523. MR. JUSTICE STEVENS' views in this respect not only find no support in the record, but they also reflect a misapprehension of the concern reflected by Admiral Turner's testimony. If in fact information is unclassified or in the public domain, neither the CIA nor foreign agencies would be concerned. The problem is to ensure in advance, and by proper procedures, that information detrimental to national interest is not published. Without a dependable prepublication review procedure, no intelligence agency or responsible Government official could be assured that an employee privy to sensitive information might not conclude on his owninnocently or otherwisethat it should be disclosed to the world.
The dissent argues that the Court is allowing the CIA to "censor" its employees' publications. Post, at 522. Snepp's contract, however, requires no more than a clearance procedure subject to judicial review. If Snepp, in compliance with his contract, had submitted his manuscript for review and the Agency had found it to contain sensitive material, presumably if one accepts Snepp's present assertion of good intentionsan effort would have been made to eliminate harmful disclosures. Absent agreement in this respect, the Agency would have borne the burden of seeking an injunction against publication. See Alfred A. Knopf, Inc. v. Colby, 509 F. 2d 1362 (CA4), cert. denied, 421 U. S. 992 (1975); United States v. Marchetti, 466 F. 2d 1309 (CA4), cert. denied, 409 U. S. 1063 (1972).
 Although both the District Court and the Court of Appeals expressly found otherwise, MR. JUSTICE STEVENS says that "the interest in confidentiality that Snepp's contract was designed to protect has not been compromised." Post, at 516-517. Thus, on the basis of a premise wholly at odds with the record, the dissent bifurcates Snepp's 1968 agreement and treats its interdependent provisions as if they imposed unrelated obligations. MR. JUSTICE STEVENS then analogizes Snepp's prepublication review agreement with the Government to a private employee's covenant not to compete with his employer. Post, at 518-520. A body of private law intended to preserve competition, however, simply has no bearing on a contract made by the Director of the CIA in conformity with his statutory obligation to "protec[t] intelligence sources and methods from unauthorized disclosure." 50 U. S. C. § 403 (d) (3).
 Judge Hoffman's dissent suggests that even this remedy may be unavailable if the Government must bring suit in a State that allows punitive damages only upon proof of compensatory damages. 595 F. 2d., at 940. The Court of Appeals majority, however, held as a matter of federal law that the nominal damages recoverable for any breach of a trust agreement will support an exemplary award. See id., at 936, and n. 10, 937-938.
 See id., at 939 (Hoffman, J., concurring in part and dissenting in part).
MR. JUSTICE STEVENS concedes that, even in the absence of a written contract, an employee has a fiduciary obligation to protect confidential information obtained during the course of his employment. Post, at 518. He also concedes that all personal profits gained from the exploitation of such information are impressed with a constructive trust in favor of the employer. Post, at 521. In this case, he seems to think that the common law would not treat information as "confidential" unless it were "classified." See, e. g., post, at 518. We have thought that the common-law obligation was considerably more expansive. See, e. g., Restatement (Second) of Agency §§ 396 (c), 400 and Comment c, 404 and Comments b, d (1958); 5 A. Scott, Trusts § 505 (3d ed. 1967). But since this case involves the breach of a trust agreement that specifically required the prepublication review of all information about the employer, we need not look to the common law to determine the scope of Snepp's fiduciary obligation.
 Snepp also signed a termination agreement in 1976 in which he made substantially the same commitment.
 In response to an interrogatory asking whether it contended that "Decent Interval contains classified information or any information concerning intelligence or CIA that has not been made public by CIA," the Government stated that "[f]or the purpose of this action, plaintiff does not so contend." Record Item No. 24, p. 14. Because of this concession, the District Judge sustained the Government's objections to defense efforts to determine whether Decent Interval in fact contains information that the Government considers classified. See, e. g., the testimony of Admiral Stansfield Turner, Director of the CIA, Tr. 135; and of Herbert Hetu, the CIA's Director of Public Affairs, Tr. 153.
 See, e. g., 18 U. S. C. § 798, which imposes a prison term of 10 years and a $10,000 fine for knowingly and willfully publishing certain types of classified information; 18 U. S. C. § 794, which makes it a criminal offense punishable by life in prison to communicate national defense information to a foreign government; and 5 U. S. C. § 8312, which withdraws the right to Government retirement benefits from a person convicted of violating these statutes. See also Exec. Order No. 12065, 3 CFR 190 (1979), note following 50 U. S. C. § 401 (1976 ed., Supp. II), which provides administrative sanctions, including discharge, against employees who publish classified information. Thus, even in the absence of a constructive trust remedy, an agent like Snepp would hardly be free, as the majority suggests, "to publish whatever he pleases." Ante, at 513, n. 8.
 In both his original employment agreement and the termination agreement Snepp acknowledged the criminal penalties that might attach to any publication of classified information. In his employment agreement he also agreed that a breach of the agreement would be cause for termination of his employment. No other remedies were mentioned in either agreement.
 In a footnote, see ante, at 515, n. 11, the Court suggests that it need not look to the common law to support its holding because the case involves a written contract. But, inasmuch as the contract itself does not state what remedy is to be applied in the event of a breach, the common law is the only source of law to which we can look to determine what constitutes an appropriate remedy.
 See, e. g., Sperry Rand Corp. v. A-T-O, Inc., 447 F. 2d 1387, 1392 (CA4 1971) (Virginia law), cert. denied, 405 U. S. 1017; Tlapek v. Chevron Oil Co., 407 F. 2d 1129 (CA8 1969) (Arkansas law); Structural Dynamics Research Corp. v. Engineering Mechanics Research Corp., 401 F. Supp. 1102, 1120 (ED Mich. 1975) (Michigan law); Restatement (Second) of Agency § 396 (c) (1958) ("Unless otherwise agreed, after the termination of the agency, the agent: . . . (c) has a duty to account for profits made by the sale or use of trade secrets and other confidential information, whether or not in competition with the principal . . .").
 As the court held in Herbert Morris, Ltd. v. Saxelby,  A. C. 688, 704, the employer's interest in protecting trade secrets does not outweigh the public interest in keeping the employee in the work force:
"[A]n employer can[not] prevent his employee from using the skill and knowledge in his trade or profession which he has learnt in the course of his employment by means of directions or instructions from the employer. That information and that additional skill he is entitled to use for the benefit of himself and the benefit of the public who gain the advantage of his having had such admirable instruction. The case in which the Court interferes for the purpose of protection is where use is made, not of the skill which the man may have acquired, but of the secrets of the trade or profession which he had no right to reveal to any one else. . . ."
 See, e. g., Briggs v. R. R. Donnelley & Sons Co., 589 F. 2d 39, 41 (CA1 1978) (Illinois law); American Hot Rod Assn., Inc. v. Carrier, 500 F. 2d 1269, 1277 (CA4 1974) (North Carolina law); Alston Studios, Inc. v. Lloyd V. Gress & Associates, 492 F. 2d 279, 282 (CA4 1974) (Virginia law); Mixing Equipment Co. v. Philadelphia Gear, Inc., 436 F. 2d 1308, 1312 (CA3 1971) (New York law); Water Services, Inc. v. Tesco Chemicals, Inc., 410 F. 2d 163, 167 (CA5 1969) (Georgia law); Restatement (Second) of Contracts § 330 (Tent. Draft No. 12, Mar. 1, 1977).
 The covenant imposes a serious prior restraint on Snepp's ability to speak freely, see n. 17, infra, and is of indefinite duration and scope factors that would make most similar covenants unenforceable. See, e. g., Alston Studios, Inc. v. Lloyd V. Gress & Associates, supra, at 283 (holding void under Virginia law a covenant with no geographical limitation); American Hot Rod Assn., Inc. v. Carrier, supra, at 1279 (holding void under North Carolina law a covenant with no durational or geographical limitation); E. L. Conwell & Co. v. Gutberlet, 429 F. 2d 527, 528 (CA4 1970) (holding void under Maryland law a covenant with no durational or geographical limitation).
 The Court correctly points out that the Government may regulate certain activities of its employees that would be protected by the First Amendment in other contexts. Ante, at 509, n. 3. But none of the cases it cites involved a requirement that an employee submit all proposed public statements for prerelease censorship or approval. The Court has not previously considered the enforceability of this kind of prior restraint or the remedy that should be imposed in the event of a breach.
 If he had submitted the book to the Agency and the Agency had refused to consent to the publication of certain material in it, Snepp could have obtained judicial review to determine whether the Agency was correct in considering the material classified. See United States v. Marchetti, 466 F. 2d 1309, 1317 (CA4 1972), cert. denied, 409 U. S. 1063. It is noteworthy that the Court does not disagree with the Fourth Circuit's view in Marchetti, reiterated in Snepp, that a CIA employee has a First Amendment right to publish unclassified information. Thus, despite its reference in footnote 3 of its opinion to the Government's so-called compelling interest in protecting "the appearance of confidentiality," ante, at 509, n. 3, and despite some ambiguity in the Court's reference to "detrimental" and "harmful" as opposed to "classified" information, ante, at 511-512, I do not understand the Court to imply that the Government could obtain an injunction against the publication of unclassified information.
 The District Judge sustained the Government's objections to questions concerning the identity of other agents who had published the unauthorized works to which Admiral Turner referred. Tr. 136. However, Admiral Turner did testify that the harmful materials involved "[p]rimarily the appearance in the United States media of identification of sources and methods of collecting intelligence. . . ." Id., at 143. This type of information is certainly classified and is specifically the type of information that Snepp has maintained he did not reveal in Decent Interval. See, e. g., Snepp's December 7, 1977, interview on the Tomorrow show, in which he stated; "I have made a very determined effort not to expose sources or methods. . . ." Government's Requests for Admissions, Record Item 19, Exhibit I, p. 5.
 Snepp's attorneys were foreclosed from asking Admiral Turner whether particular foreign sources had stopped cooperating with United States' authorities as a direct result of the publication of Decent Interval. Tr. 138. Thus, it is unclear whether or why foreign sources may have reacted unfavorably to its publication. However, William E. Colby, the CIA's former Director, did indicate in his testimony that foreign nations generally have a stricter secrecy code than does the United States. Id., at 175-176.
 One of the Court's justifications for its constructive trust remedy is that "it cannot saddle the former agent with exemplary damages out of all proportion to his gain." Ante, at 516. This solicitude for Snepp's welfare is rather ironic in view of the Draconian nature of the remedy imposed by the Court today.
 I have been unable to discover any previous case in which the Court has acted as it does today, reaching the merits of a conditional cross-petition despite its belief that the petition does not merit granting certiorari.
 Which, it should be noted, does not appear anywhere in the Government's 5-page cross-petition.
 The mere fact that the Agency has the authority to review the text of a critical book in search of classified information before it is published is bound to have an inhibiting effect on the author's writing. Moreover, the right to delay publication until the review is completed is itself a form of prior restraint that would not be tolerated in other contexts. See, e. g., New York Times Co. v. United States, 403 U. S. 713; Nebraska Press Assn. v. Stuart, 427 U. S. 539. In view of the national interest in maintaining an effective intelligence service, I am not prepared to say that the restraint is necessarily intolerable in this context. I am, however, prepared to say that, certiorari having been granted, the issue surely should not be resolved in the absence of full briefing and argument.