with whom Mr. Justice Brennan joins, dissenting.
Appellee was discharged from her job with the Boston State Hospital solely because she refused to swear or affirm the following oath:1
“I will uphold and defend the Constitution of the United States of America and the Constitution of the Commonwealth of Massachusetts and ... I will oppose the overthrow of the government of the United States of America or of this Commonwealth by force, violence or by any illegal or unconstitutional method.” Mass. Gen. Laws, c. 264, § 14.
She brought this action in the United States District Court for the District of Massachusetts seeking declaratory and injunctive relief against enforcement of the oath as a condition of her employment.2 The District Court found that the oath was unconstitutionally vague and granted the relief requested by appellee. The Court now reverses the District Court and sustains the validity *692of the oath in its entirety. In my opinion, the second half of the oath is not only vague, but also overbroad. Accordingly, I dissent.
The first half of the oath, requiring an employee to indicate a willingness to “uphold and defend” the state and federal Constitutions, is clearly constitutional. It is nothing more than the traditional oath of support that we have unanimously upheld as a condition of public employment.
It is the second half of the oath to which I object. I find the language “I will oppose the overthrow of the government of the United States of America or of this Commonwealth by force, violence or by any illegal or unconstitutional method” to be impermissibly vague and overbroad.
It is vague because “men of common intelligence [must] speculate at their peril on its meaning.” Whitehill v. Elkins, 389 U. S. 54, 59 (1967). See also Connally v. General Construction Co., 269 U. S. 385, 391 (1926); Cline v. Frink Dairy Co., 274 U. S. 445, 465 (1927); Lanzetta v. New Jersey, 306 U. S. 451, 453 (1939). The most striking problem with the oath is that it is not clear whether the last prepositional phrase modifies the verb “oppose” or the noun “overthrow.” Thus, an af-fiant cannot be certain whether he is swearing that he will “oppose” governmental overthrow by utilizing every means at his disposal, including those specifically prohibited by the laws or constitutions he has sworn to support, or whether he has merely accepted the responsibility of opposing illegal or unconstitutional overthrows. The first reading would almost surely be unconstitutional since it is well established that a State cannot compel a citizen to waive the rights guaranteed him by the Constitution in order to obtain employment. See, e. g., Pickering v. Board of Education, *693391 U. S. 563 (1968); Garrity v. New Jersey, 385 U. S. 493 (1967). This reading would also make the second half of the oath inconsistent with the first half. It is far from clear to me which reading the Massachusetts Legislature intended. A reasonable man could certainly read the oath either way, and Massachusetts has not offered to make a binding clarification of its purport.
Even assuming that the second reading were unconditionally adopted by the appellants and communicated to prospective employees, the vice of vagueness is still not cured, for the affiant is left with little guidance as to the responsibilities he has assumed in taking the oath. In what form, for example, must he manifest his opposition to an overthrow? At oral argument in the District Court, the Commonwealth’s attorney asserted that citizens have three standards of obligation to their government to oppose overthrows:
“The ordinary citizen who has taken no oath has an obligation to act in extremis; a person who has taken the first part of the present oath would have a somewhat larger obligation, and one who has taken the second part has one still larger.” 300 F. Supp. 1321, 1322.3
*694I agree with the conclusion of the District Court that “[t]he very fact that such varied standards . . . can be suggested is enough to condemn the language as hopelessly vague.” Id., at 1323.
Vagueness is also inherent in the use of the word “overthrow.” When does an affiant’s undefined responsibility under the oath require action: When an overthrow is threatened? When an overthrow is likely to be threatened? When a threatened overthrow has some chance of success? Cf. Brandenburg v. Ohio, 395 U. S. 444 (1969); Yates v. United States, 354 U. S. 298 (1957); Dennis v. United States, 341 U. S. 494 (1951). The oath answers none of these questions, and for that reason, if no other, cannot stand.
The importance of clarity and precision in an oath of this kind should not be underestimated. Chapter 264, § 14, of the Massachusetts General' Laws provides that the oath is taken subject to the pains and penalties of perjury, and § 15 of that chapter specifies that the pains and penalties may amount to one year in prison and/or a $10,000 fine.
*695In concluding that this oath is vague, I rely on Baggett v. Bullitt, 377 U. S. 360 (1964). One part of the oath considered in Baggett, like the Massachusetts oath, required that the affiant assert a willingness to conform future conduct to the criteria set forth in an oath taken under penalty of perjury. The Court struck down the oath in Baggett, and Mr. Justice White’s opinion for the Court explained in great detail the inordinate difficulties employees would have in attempting to conform their actions to the oath’s criteria. Id., at 371. While the oath involved herein differs somewhat from that involved in Baggett, the considerations in both cases are the same, and the results should also be the same.
I would also strike down the second half of this oath as an overbroad infringement of protected expression and conduct.
The Court’s prior decisions represent a judgment that simple affirmative oaths of support are less suspect and less evil than negative oaths requiring a disaffirmance of political ties, group affiliations, or beliefs. Compare Connell v. Higginbotham, 403 U. S. 207 (1971); Knight v. Board of Regents, 269 F. Supp. 339 (SDNY 1967), aff’d, 390 U. S. 36 (1968); Hosack v. Smiley, 276 F. Supp. 876 (Colo. 1967), aff’d, 390 U. S. 744 (1968); Ohlson v. Phillips, 304 F. Supp. 1152 (Colo. 1969), aff’d, 397 U. S. 317 (1970), with Whitehill v. Elkins, 389 U. S. 54 (1967); Baggett v. Bullitt, supra; Cramp v. Board of Public Instruction, 368 U. S. 278 (1961); Speiser v. Randall, 357 U. S. 513 (1958); Wieman v. Updegraff, 344 U. S. 183 (1952); Garner v. Board of Public Works, 341 U. S. 716 (1951).
Yet, I think that it is plain that affirmative oaths of loyalty, no less than negative ones, have odious connotations and that they present dangers. See Asper, The Long and Unhappy History of Loyalty Testing in Mary*696land, 13 Am. J. Legal Hist. 97,104 (1969); Askin, Loyalty-Oaths in Retrospect: Freedom and Reality, 1968 Wis. L. Rev. 498, 502; Note, Loyalty Oaths, 77 Yale L. J. 739, 763 (1968). We have tolerated support oaths as applied to all government employees only because we view these affirmations as an expression of “minimal loyalty to the Government.” American Communications Assn. v. Douds, 339 U. S. 382, 415 (1950). Such oaths are merely indications by the employee “in entirely familiar and traditional language, that he will endeavor to perform his public duties lawfully.” Law Students Research Council v. Wadmond, 401 U. S. 154, 192 (Marshall, J., dissenting).
It is precisely because these oaths are minimal, requiring only that nominal expression of allegiance “which, by the common law, every citizen was understood to owe his sovereign,” Knight v. Board of Regents, 269 F. Supp., at 341, that they have been sustained. That they are minimal intrusions into the freedom of government officials and employees to think, speak, and act makes them constitutional; it does not mean that greater intrusions will be tolerated. On the contrary, each time this Court has been faced with an attempt by government to make the traditional support oath more comprehensive or demanding, it has struck the oath down. See, e. g., Connell v. Higginbotham, supra; Baggett v. Bullitt, supra; cf. Bond v. Floyd, 385 U. S. 116 (1966).
When faced with an “imminent clear and present danger,” governments may be able to compel citizens to do things that would ordinarily be beyond their authority to mandate. But, such emergency governmental power is a far cry from compelling every state employee in advance of any such danger to promise in any and all circumstances to conform speech and conduct to opposing an “overthrow” of the government. The *697Constitution severely circumscribes the power of government to force its citizens to perform symbolic gestures of loyalty. Cf. Board of Education v. Barnette, 319 U. S. 624 (1943). Since the overbreadth of the oath tends to infringe areas of speech and conduct that may be protected by the Constitution, I believe that it cannot stand. See Whitehill v. Elkins, supra; Baggett v. Bullitt, supra; Wieman v. Updegraff, supra; Shelton v. Tucker, 364 U. S. 479 (1960).
Because only the second half of the oath is invalid, I would normally favor severing the statute and striking only the second part. See Connell v. Higginbotham, supra. However, when confronted with an oath strikingly similar to that before us, the Supreme Judicial Court of Massachusetts held that the two portions of the oath were not severable. Pedlosky v. Massachusetts Institute of Technology, 352 Mass. 127, 224 N. E. 2d 414 (1967). This Court must bow to state courts in their construction of state legislation. Therefore, we must bow to the decision of the state court and strike the oath in its entirety.
Before concluding, I add one additional word about loyalty oaths in general. They have become so prevalent in our country that few Americans have not at one time or another taken an oath to support federal and state governments. Such oaths are not only required as a condition of government employment, but often as a prerequisite to entering military service, to obtaining citizenship, to securing a passport or an educational loan or countless other government offerings. Perhaps we have become so inundated with a variety of these oaths that we tend to ignore the difficult constitutional issues that they present. It is the duty of judges, however, to endeavor to remain sensitive to these issues and not to “encourage the casual taking of oaths by upholding the discharge or exclusion from public employment of *698those with a conscientious and scrupulous regard for such undertakings.” Baggett v. Bullitt, supra, at 373-374.
Loyalty oaths do not have a very pleasant history in this country. Whereas they may be developed initially as a means of fostering power and confidence in government, there is a danger that they will swell “into an instrument of thought control and a means of enforcing complete political conformity.” Asper, The Long and Unhappy History of Loyalty Testing in Maryland, 13 Am. J. Legal Hist. 97, 108 (1969). Within the limits of the Constitution it is, of course, for the legislators to weigh the utility of the oaths and their potential dangers and to strike a balance. But, as a people, we should always keep in mind the words of Mr. Justice Black, concurring in Speiser v. Randall, 357 U. S., at 532:
“Loyalty oaths, as well as other contemporary ‘security measures/ tend to stifle all forms of unorthodox or unpopular thinking or expression — the kind of thought and expression which has played such a vital and beneficial role in the history of this Nation. The result is a stultifying conformity which in the end may well turn out to be more destructive to our free society than foreign agents could ever hope to be. . . . I am certain that loyalty to the United States can never be secured by the endless proliferation of ‘loyalty’ oaths; loyalty must arise spontaneously from the hearts of people who love their country and respect their government.”
Accordingly, I would affirm the decision of the District Court.
Appellee was not requested to take the oath before she began her employment. The reasons for the failure of the hospital officials to require the oath as a prerequisite to employment are not readily apparent from the record. In any event, the oath was required of all state employees at all relevant times.
Appellee also sought damages for back wages allegedly owed. It is apparent that all back wages have now been paid. Thus, this claim is no longer in controversy. The District Court rejected ap-pellee’s belated attempt to make a claim for loss of wages due to termination, and this decision was well within its discretion under Rule 15 of the Federal Rules of Civil Procedure.
It is clear that both speech and conduct are affected by this portion of the oath. Appellants conceded as much in their brief in the court below:
“[I]n the event that a clear and present danger arose of the actual overthrow of the government, . . . the public employee [would] be required to use reasonable means at his disposal to attempt to thwart that effort. What he might do in such circumstances could range from the use of physical force to speaking out against the downfall of the government. The kind of response required would be commensurate with the circumstances and with the employee’s ability, his training, and the means available to him at the time.” (Emphasis added.) Quoted at 300 F. Supp., at 1322.
The final sentence of this quotation evidences the confusion that *694the State confesses about the responsibilities assumed by employees in taking the oath.
In light of the arguments that the appellants make, I find it impossible to agree with the Court that the second half of the oath adds nothing to the first. The appellants contend, contrary to the assertions of the Court, that a citizen who takes the first part of the oath has more of a duty to his government than one who takes no oath, and that one who takes the second part of the oath has a still greater duty. While the appellants are unsure as to where and how far that duty extends, they never have suggested that it simply does not exist. The argument is even made that the duty extends to the use of physical force.
Were we faced with merely a traditional oath of support, I would join the Court. I share the Court’s dismay at having to hold state legislation unconstitutional, but I cannot ignore the thrust that a State would give its statutes. Cf. Pedlosky v. Massachusetts Institute of Technology, 352 Mass. 127, 224 N. E. 2d 414 (1967).