*92Opinion

Petitioner seeks writ of mandate to compel the Attorney General to title and prepare summary of a proposed initiative measure.

Petitioner submitted a proposed initiative measure to the Attorney General for title and summary preparatory to obtaining signatures to qualify the measure for the ballot. (See Elec. Code, §§ 3502, 3503.) The measure would (a) make it unlawful for any teacher to strike; (b) prohibit campaign contributions by teachers’ organizations; and (c) prevent tax revenues from being used to provide transportation for purpose of racially balancing public schools.

The Attorney General refused to issue the title on the ground that the proposed measure concerns more than one subject thereby violating article II, section 8, subdivision (d), of the California Constitution. The subdivision provides: “An initiative measure embracing more than one subject may not be submitted to the electors or have any effect.”

The right of the initiative is “precious to the people and is one which the courts are zealous to preserve to the fullest tenable measure of spirit as well as letter.” (McFadden v. Jordan (1948) 32 Cal.2d 330, 332 [196 P.2d 787].) “To preserve the full spirit of the initiative the submission of issues to the voters should not become bogged down by lengthy litigation in the courts.” (Perry v. Jordan (1949) 34 Cal.2d 87, 91 [207 P.2d 47].)

In furtherance of the people’s power we have narrowly circumscribed the rights of ministerial officials to impede or delay the initiative process. Speaking of an acting registrar in Farley v. Healey, 67 Cal.2d 325, 327 [62 Cal.Rptr. 26, 431 P.2d 650], this court stated: “It is not his function to determine whether a proposed initiative will be valid if enacted or whether' a proposed declaration of policy is one to which the initiative may apply. These questions may involve difficult legal issues that only .a court can determine. The right to propose initiative measures cannot properly be impeded by a decision of a ministerial office, even if supported by the advice of the city attorney, that the subject is not appropriate for submission to the voters. ” (Italics added.)

The duty of the Attorney General to prepare title and summary for a proposed initiative measure is a ministerial one and mandate will *93lie to compel him to act when the proposal is in proper form and complies with statutory and constitutional procedural requirements. (Warner v. Kenny (1946) 27 Cal.2d 627, 630-631 [165 P.2d 889].)

The single subject requirement of article II, section 8, subdivision (d), involves difficult legal questions that only a court may resolve. (Cf. Perry v. Jordan, 34 Cal.2d 87, 92-93 [207 P.2d 47].) We are satisfied that a claim of violation of subdivision (d) is not merely a formal one, but is based on the effects of the contents of the proposed measure. Absent judicial authorization, the Attorney General may not urge violation of the single subject requirement to justify refusal to title and prepare summary of a proposed measure.

This does not mean that the Attorney General may not challenge the validity of the proposed measure by timely and appropriate legal action. We hold only that without prior judicial authorization he may not delay or impede the initiative process while claims of the measure’s invalidity are determined, Petitioner is entitled to have his proposal titled and summarized so that he may commence seeking signatures to qualify it for the ballot.

We express no view as to the merits of the claim that the proposed measure concerns more than one subject.

Let a peremptory writ of mandate issue as prayed.

Bird, C. J., Mosk, J., Richardson, J., and Newman, J., concurred.

I must respectfully dissent. The views expressed in the majority opinion are at odds with the recognized powers of the Attorney General derived from constitutional and common law sources and would compel the Attorney General to perform unnecessary acts with respect to an initiative measure which does not meet the requirement for placement on the ballot.

We have recognized the power of the Attorney General to protect the public interest. Recently we reiterated that power in D'Amico v. Board of Medical Examiners, 11 Cal.3d 1, 14-15 [112 Cal.Rptr. 786, 520 P.2d 10] when we said; “ . . . The Attorney General... is the chief law officer of *94the state (Cal. Const., art. V, § 13). As such he possesses not only extensive statutory powers but also broad powers derived from the common law relative to the protection of the public interest. (See Pierce v. Superior Court (1934) 1 Cal.2d 759, 761-762 [37 P.2d 460, 96 A.L.R. 1020], and cases there cited, especially People v. Stratton (1864) 25 Cal. 242, 246-247.) ‘[H]e represents the interest of the people in a matter of public concern.’ (Savings Bank v. Superior Court (1894) 103 Cal. 27, 32 [36 P. 1015].) Thus, ‘in the absence of any legislative restriction, [he] has the power to file any civil action or proceeding directly involving the rights and interests of the state, . . . the preservation of order, and the protection of public rights and interest.’ (Pierce v. Superior Court, supra, at pp. 761-762.) Conversely, he has the duty to defend all cases in which the state or one of its officers is a party. (Gov. Code, § 12512.) In the course of discharging this duty he is often called upon to make legal determinations both in his capacity as a representative of the public interest and as statutory counsel for the state or one of its agencies or officers....”

The majority turns away from this view by its holding today. The nature of the office of the Attorney General has been described as follows: “As guardian of royal prerogative, the Attorney General of England possessed a broad range of powers .... Unlike after the Colonial Period when state governments were organized and recognized in this country, there was no monarch in whom the governmental prerogatives were vested. Since the essential power of government resided and emanated from the people, the prerogatives had to be exercised on their behalf. Just as the Attorney General safeguarded royal prerogatives at common law, similarly, the official authority, an obligation to protect public rights and enforce public duties on behalf of the general public became vested by the states in the Attorney General, and it is this obligation inherited from the common law to represent the public interest which has shaped and colored the role which the Attorney General fulfills today.” (Rep. on the Off. of the Atty. Gen., National Association of Attorneys General (Feb. 1971) p. 33.)

If the measure sought to be titled contains more than one subject, as in the instant case, then the Attorney General not only has the right not to proceed, but indeed, he has the duty not to proceed. Any other course would not only be contrary to his constitutional duties and powers qua Attorney General, but contrary to his oath of office (Cal. Const., art. X, § 3). The position the Attorney General must take here is directly connected with the initiative process, a process constitutionally *95prescribed and which provides that any “initiative measure embracing more than one subject may not be submitted to the electors or have any effect.” (Cal. Const., art. II, § 8, subd. (d).)

If it be assumed that the measure before this court does not meet the test of article II, section 8, subdivision (d), then not only should it not be submitted to the voters, but it shall not “have any effect.” This provision does not say that the initiative shall have no effect only after judicial determination, but rather the Constitution renders such measure void, and not merely voidable. Of course, the Attorney General, in concluding that the section has no effect, proceeds at his risk that others, especially the judiciary, will disagree with him. However, such risk taking is inherent in eveiy legal prognostication and marks the very process by which law is practiced in this country, whether by lawyer, trial court or intermediate courts of appeal. That such risks exist, perhaps, may encourage the Attorney General to seek refuge in the safe course of litigation as suggested by the majority, but such risk cannot make effective that which is ineffective. Defense of the Constitution has never been the preoccupation of the timid. This court cannot properly condemn the action of the Attorney General where, as here, it refuses to determine the correctness of his decision.

The characterization of the office of the Attorney General by such terms as “ministerial officer” is not a proper substitute for analysis or understanding of that office, nor is that characterization correct. It is true that Warner v. Kenny (1946) 27 Cal.2d 627, 630-631 [165 P.2d 889], with respect to the case before it used this unfortunate expression, but in that case, the Attorney General refused to issue a summary and title on the ground that the measure was substantially the same as a measure which had been duly titled and summarized earlier but which failed to qualify for the ballot within the time allowed by law. No basis could be found in law for the Attorney General’s refusal. Accordingly, this court stated: “[n]o showing has been made to justify his refusal to prepare a title and summary for the proposed measure, and since it is in proper form and was submitted to him in accordance with the constitutional and statutory requirements as to procedure, petitioners are entitled to have furnished to them a title and summary thereof.” (Id., at p. 631.) Unlike Warner, by demonstrating that the measure was not in proper form and that it violated article II, section 8, subdivision (d) of the Constitution, the Attorney General made such a showing as to justify his refusal to prepare a title and summary. Obviously, where the Attorney General has no legal basis to refuse to title and summarize a measure his role, is *96ministerial, but where he has a legal basis to refuse to act, this court cannot ignore its responsibility to decide the issue by simply uttering “ministerial duty.”

Moreover, this case may well return to haunt the constitutional jurisprudence of this state as a result of the majority’s holding that a “ministerial officer” must perform an unconstitutional act. I know of no rule of law which excepts a “ministerial officer” from the operation of the Constitution until his decision receives this court’s imprimatur that the Constitution applies to his duty. Indeed, a classic means for a public officer to test the constitutionality of a statute is to threaten to disobey it (e.g., California Educational Facilities Authority v. Priest (1974) 12 Cal.3d 593 [116 Cal.Rptr. 361, 526 P.2d 513]). The majority turns its back on this valid test.

Here the Attorney General refused to act, petitioner seeks relief, the issue is joined and the petition has not been circulated. There is no reason why this court should not now speak on the issue of the validity of the measure. This is not a case where a sizeable body of the electorate have signed the petition. At this point this is simply a lawsuit between the Attorney General and petitioner. Thus, cases expressing our reluctance to become involved in the initiative process are not relevant. (Mulkey v. Reitman (1966) 64 Cal.2d 529, 535 [50 Cal.Rptr. 881, 413 P.2d 825]; Kevelin v. Jordan (1964) 62 Cal.2d 82 [41 Cal.Rptr. 169, 396 P.2d 585]; and Wind v. Hite (1962) 58 Cal.2d 415, 417 [24 Cal.Rptr. 683, 374 P.2d 643]; but see McFadden v. Jordan (1948) 32 Cal.2d 330 [196 P.2d 787].) This case involves not the substantive validity of the measure, but rather the question whether the measure, failing to meet the procedural requirements of the Constitution ought to be processed at all.

Moreover, this court must be aware that the present proceeding is a mandamus proceeding. As it is an equitable proceeding, there must be, among other things, a clear and present duty on the part of the respondent to act (Code Civ. Proc., § 1085; 5 Witkin, Cal. Procedure (2d ed.) Extraordinary Writs, § 61, p. 3838.) Mandamus does not lie to compel performance of an act where the act would be illegal. (5 Witkin, Cal. Procedure (2d ed.) Extraordinary Writs, § 66, p. 3844.)

Notwithstanding that the Attorney General has the power to decide the procedural sufficiency of a proposed initiative, the measure is defective as it contains more than one subject. The contested proposal includes provisions on teacher strikes, campaign contributions from *97teacher organizations, and the use of tax revenues to support the use of busing to achieve racial balance in public schools. It presents a classic example of the very evils that the one subject limitation was designed to prevent, that is, the dangers of confusing or misleading voters and of subverting the will of the majority. In order to satisfy the one subject limitation, an initiative’s provisions must be functionally related in furtherance of a common underlying purpose. The proposed initiative fails to satisfy this requirement. Accordingly, it was properly within the Attorney General’s discretion to refuse to prepare a title and summary of the initiative.

Article II, section 8, subdivision (d), provides that “an initiative measure embracing more than one subject may not be submitted to the electors or have any effect.”1 In order to interpret this provision we first consider the intent behind the amendment as reflected in the ballot arguments presented in the official voter’s pamphlet for the 1948 election and in relevant case law.2 Secondly, we must compare the limitation on initiative measures with a similar limitation on legislative enactments. As we shall see, the initiative limitation was intended to be a stricter limitation than the legislative limitation.

Designed to limit each initiative to “one subject and one subject only,” the initiative restriction was intended to eliminate the danger of voter confusion and deception. (Cal. Voters Pamp, General Election (Nov. 2, 1948) pp. 8-9.) The ballot argument in favor of the limitation explained that a proposition containing more than one subject could be misleáding and confusing to the voter. “The busy voter does not have the time to devote to the study of long, wordy propositions and must rely upon such sketchy information as may be received through the press, radio, or picked up in general conversation. If improper emphasis is placed upon one feature and the remaining features [of the initiative] ignored, or if there is a failure to study the entire proposed amendment, the voter may be misled as to the overall effect of the proposed amendment.” (Cal. Voters Pamp, supra, at p. 8.)

*98In addition to protecting voters from confusing or misleading initiatives, the initiative amendment was intended to protect signers of the sponsoring petition that is required to qualify an initiative for the ballot. “People requested to sign the sponsoring petition will readily understand just what the entire proposition is and not be confused or misled by a maze of unrelated matters some of which are inadequately explained, purposely distorted, or intentionally concealed.” (Cal. Voters Pamp., supra, at p. 8.)

Aside from these dangers, a multisubject initiative presents the additional danger of subverting the will of the voters. An initiative including several unrelated subjects, each controversial in itself, could result in the passage (or defeat) of the measure by the accumulation of minority votes even though none of the provisions could command a majority vote on its own. The same year the initiative limitation was adopted, this court explained that a multisubject initiative “does not give the people an opportunity to express approval or disapproval severally as to each major change suggested; rather [such initiatives] have the purpose of aggregating for the measure the favorable votes from electors of many suasions who, wanting strongly enough any one or more propositions offered, might grasp at that which they want, tacitly accepting the remainder. Minorities favoring each proposition severally might, thus aggregated, adopt all.” (McFadden v. Jordan (1948) 32 Cal.2d 330, 346 [196 P.2d 787], cert. den., 336 U.S. 918 [93 L.Ed. 1080, 69 S.Ct. 640].)

It is clear, then, that this limitation was intended to protect voters from misleading and confusing initiatives as well as to guarantee that each subject would receive proper and independent voter consideration. These concerns indicate that the one subject provision was intended to function as substantive limitation on the scope of initiatives. Further, as the amendment history indicates, the initiative provision was intended to ' be more narrowly applied than the similar limitation on legislative measures.

A one subject limitation has long existed for statutes proposed in the state Legislature.3 Although the legislative limitation has been construed *99broadly to uphold proposed bills,4 the special nature of the initiative process requires a narrower construction.5 First, the dangers presented by a multisubject proposal are much more limited in the legislative context than in the initiative context. A proposed bill is closely scrutinized by legislators and their staffs in both houses of the Legislature. Each bill is assigned to a standing committee that receives testimony from interested parties and makes recommendations concerning the bill. Finally, the Governor examines the bill before signing it into law. By contrast, no such scrutiny is afforded a proposed initiative by the voter. Voters have neither the time nor the resources to mount an in depth investigation of a proposed initiative. Often voters rely solely on the title and summary of the proposed initiative and never examine the actual wording of the proposal. (Cal. Voters Pamp., supra, p. 8; The California Initiative Process: A Suggestion for Reform (1975) 48 So.Cal.L.Rev. 922, 931-932.)

Second, if legislators do not agree with portions of a proposed bill, they can introduce their own bills, propose amendments to the bill, or attempt through committee to get the provisions severed from the bill. Unlike the Legislature, in which amendment and modification are customary, the initiative process permits no such flexibility. (See, e.g., Couch et al., Cal. Government and Politics (3d ed. 1964) p. 102.) The result of this inflexibility is that more often than not a proposed initiative represents the most extreme form of law which is considered politically expedient. “[T]he only expression left to all other interested parties who are not proponents is the ‘yes’ or ‘no’ vote they cast.” (Note; The California Initiative Process: A Suggestion for Reform, supra, 48 So.Cal.L.Rev. at pp. 932-933; see also Taschner v. City Council (1973) 31 Cal.App.3d 48, 64 [107 Cal.Rptr. 214].)

*100It is because of the voters’ lesser ability to scrutinize a proposal and their total inability to propose modifications, that the multisubject initiative presents greater dangers than a similar multisubject legislative bill.

This greater danger of confusion, deception and vote amalgamation compels a more restrictive interpretation of the limitation on initiatives than the limitation on legislative bills. Accordingly, to satisfy the one-subject requirement, an initiative’s provisions must be functionally related in furtherance of a common underlying purpose. This standard, although stricter than the legislative limitation, still permits complex but focused initiatives. What this standard does not permit, however, is the amalgamation of unrelated provisions that could mislead the voters or result in the frustration of the will of the majority of voters as may the present measure.

Indeed, the dangers that inspired the passage of the one-subject initiative limitation are especially manifest in the proposed initiative before us. This proposed initiative would amend the California Constitution to (1) make it unlawful for any teacher to strike, mandating the dismissal and suspension of the credential of a striking teacher; (2) prohibit campaign contributions by any teachers’ organization; and (3) prohibit the use of tax revenues to provide transportation for the purpose of racially balancing public schools.

This proposed measure involves three different and unrelated subjects, labor relations, campaign contributions, and integration,6 affecting the rights of different classes of people.7 Each of the provisions in the proposed initiative concerns a distinct, controversial, and highly emotional subject. A voter with strong feelings about any one of these matters would not be likely to assess the other provisions independently. Intentionally or innocently, the initiative would tend to cause the amalgamation of the votes of separate minorities in favor of separate provisions to achieve the measure’s enactment.

Nor are the proposed initiative’s provisions a “general scheme” or in furtherance of a common practice which might ameliorate the measure’s *101danger. Petitioner claims that the initiative deals with the single subject of “educational reform” with the purpose of eliminating the “political influence that has crept into the educational system.”8 If the broad umbrella of “educational reform” can embrace such diverse and unrelated subjects as are present here, then the one-subject limitation becomes meaningless. Another proponent could just as legitimately and persuasively argue that a welter of unrelated proposals relate to the single subject of “governmental” reform.

The dangers of confusing or misleading voters are manifest in this type of initiative,9 and it was this type of abuse that the one-subject restriction was adopted to eliminate.

Accordingly, the Attorney General correctly determined that the initiative violated the one-subject limitation of article II, section 8, subdivision (d), and because the Attorney General has the power to assess the validity of the proposed initiative and did not abuse his discretion in that determination, the petition for writ of mandate should be denied.

Tobriner, J., concurred.

Respondent’s petition for a rehearing was denied May 25, 1978. Manuel, J., was of the opinion that the petition should be granted.