I concur in the judgment. I write separately because I disagree with the majority’s nonliteral interpretation of Proposition 64 (Gen. Elec. (Nov. 2, 2004)), which forecloses a variety of representative actions the measure clearly permits. Unlike the majority, I do not believe we would frustrate the voters’ intent by enforcing the measure according to its plain language.
The unfair competition law (Bus. & Prof. Code, § 17200 et seq.) (UCL), as amended by Proposition 64, requires persons who wish to pursue claims on others’ behalf to “compl\y] with Section 382 of the Code of Civil Procedure . . . .” (Bus. & Prof. Code, § 17203, italics added.) The majority construes the italicized language “to mean that such an action must meet the requirements for a class action.” (Maj. opn., ante, at p. 980.) The problem with this conclusion is that the UCL, even as amended by Proposition 64, does not refer to class actions. Instead, it refers to Code of Civil Procedure section 382 (section 382). Section 382, which also does not refer to class actions, long predates that modem procedural device. Borrowed from New York’s 1848-1849 Field Code (see Comrs. on Practice and Pleadings, Code of Civ. Proc. of the State of N.Y. (1850) § 610, p. 249), the language of section 382 entered California law in 1850 with California’s first civil practice act (Stats. 1850, ch. 142, § 14, p. 429) and was reenacted in 1872 with its current designation as part of our original Code of Civil Procedure. Since then, section 382 has been amended only once, in 1971 (Stats. 1971, ch. 244, § 12, p. 375), to delete a reference to compulsory joinder. The statute remains ancient in language and intent, without significant intervening legislative attention.
Section 382 actually codifies not class action procedure but the common law doctrine of virtual representation. (Weaver v. Pasadena Tournament of Roses (1948) 32 Cal.2d 833, 837 [198 P.2d 514].) Under the doctrine, a person who was not a party to an action was deemed to have been virtually represented, and thus bound by the judgment, if his or her interests had received adequate representation by a party.1 (See, e.g., Bernhard v. Wall *989(1921) 184 Cal. 612, 629 [194 P. 1040].) The modem law of class actions evolved out of virtual representation. In 1948, we held that the doctrine, as codified in section 382, provided courts with sufficient authority to use the class action procedural mechanism. (Weaver v. Pasadena Tournament of Roses, supra, at pp. 836-837.) Over time, encouraged by the adoption in 1966 of rule 23 of the Federal Rules of Civil Procedure, class actions multiplied and began to displace other types of multiparty representative actions. California courts, lacking any other statutory basis for class actions,2 simply continued to cite section 382 as authority and, when specific guidance was required, looked to federal decisions applying mle 23. (E.g., Green v. Obledo (1981) 29 Cal.3d 126, 146 [172 Cal.Rptr. 206, 624 P.2d 256]; Vasquez v. Superior Court (1971) 4 Cal.3d 800, 821 [94 Cal.Rptr. 796, 484 P.2d 964]; Daar v. Yellow Cab Co. (1967) 67 Cal.2d 695, 708-709 [63 Cal.Rptr. 724, 433 P.2d 732].) Today, its history largely forgotten, section 382 is commonly but inaccurately described as setting out the requirements for class certification.3 The majority adopts this shorthand description, as did the Attorney General and the Legislative Analyst in the ballot pamphlet. (See maj. opn., ante, at pp. 979-980, citing Voter Information Guide, Gen. Elec. (Nov. 2, 2004) official title and summary of Prop. 64, p. 38; id., ballot measure summary, Prop. 64, p. 6; id., analysis by the Legislative Analyst of Prop. 64, pp. 38-39.) In a non-UCL case I, too, have described the statute in the same way. (Fireside Bank v. Superior Court (2007) 40 Cal.4th 1069, 1092, fn. 9 [56 Cal.Rptr.3d 861, 155 P.3d 268].) The common shorthand, however, is not accurate. Even today, more remains of section 382 than just a makeshift citation for the proposition that California law authorizes class actions.
What remains of section 382 is best understood by reference to Taylor v. Sturgell (2008) 553 U.S. 880 [171 L.Ed.2d 155, 128 S.Ct. 2161] (Taylor), in which the United States Supreme Court comprehensively examined the federal courts’ use of virtual representation, the common law doctrine section *990382 embodies. Taylor’s basic holding is that a judgment’s binding effect is to be determined not under common law doctrines but instead under the established rules of res judicata and collateral estoppel, which typically require that a person, to be bound, must have been made a party, received service of process, and had a full and fair opportunity to litigate. (Taylor, at p. _ [128 S.Ct. at p. 2171].) Taken together, these rules form what the high court has called a general “rule against nonparty preclusion.” (Id., at p. _ [128 S.Ct. at p. 2172].) To encourage clarity in determining the preclusive effect of judgments on nonparties, the high court instructed the lower federal courts not to use the term “virtual representation.” (Id., at p. _ [128 S.Ct. at p. 2178].) In so doing, however, the court observed that to discard the term was “unlikely to occasion any great shift in actual practice” or any significant “change in outcomes” (ibid.) because the term encompasses six categories of valid, established exceptions to the rule against nonparty preclusion (ibid.; see also id., at pp. _-_ [128 S.Ct. at pp. 2172-2173]).
The propriety of any given representative action obviously depends on whether the nonparties assumed to be represented will in fact be bound by the judgment. Of the six categories of exceptions to the rule against nonparty preclusion identified in Taylor, supra, 553 U.S. 880, _-_ [128 S.Ct. 2161, 2172-2173], three might well, but for today’s decision, support non-class representative actions under the UCL: (1) preclusion because a person has agreed to be bound by the determination of issues in an action between others; (2) preclusion based on a variety of preexisting substantive legal relationships arising from the needs of property law, such as the relationships between preceding and succeeding owners of property, bailee and bailor, and assignee and assignor; and (3) preclusion because a nonparty was adequately represented by someone with the same interests who was a party, as in properly conducted class actions and in suits brought by trustees, guardians, and other fiduciaries. (Taylor, at pp. _-_ [128 S.Ct. at pp. 2172-2173].)
Taylor’s third exception to the rule of nonparty preclusion—cases in which a nonparty was adequately represented, as in “properly conducted class actions” (Taylor, supra, 553 U.S. 880, _ [171 L.Ed.2d 155, 128 S.Ct. 2161, 2172], italics added)—will undoubtedly comprise the vast majority of multiparty actions brought under the UCL. The consumers on whose behalf UCL actions are brought typically have no relationship with the representative plaintiff other than the fact that they purchased the same product or service from the defendant. Still, actions brought under the first (consent) and second (relationships based on property law) exceptions to the general rule of nonparty preclusion, and actions brought under the third exception by “trustees, guardians, and other fiduciaries” (Taylor, supra, at p. _ [128 S.Ct. at p. 2173]), fall squarely within the language and intent of section 382, remain valid under federal law (see Taylor, at pp. _-_ [128 S.Ct. at *991pp. 2172-2173]), and might well be invoked as the basis for non-class-representative actions under the UCL. One can easily imagine, for example, an action by a homeowners’ association on behalf of its members—a type of representative action California courts have consistently held to be proper under section 382 even without class certification. (E.g., Del Mar Beach Club Owners Assn. v. Imperial Contracting Co. (1981) 123 Cal.App.3d 898, 906-908 [176 Cal.Rptr. 886]; Raven’s Cove Townhomes, Inc. v. Knuppe Development Co. (1981) 114 Cal.App.3d 783, 793-794 [171 Cal.Rptr. 334].)
The majority, by simplistically construing Proposition 64’s reference to “Section 382” (Bus. & Prof. Code, § 17203, as amended by Prop. 64) as requiring class certification in every instance, forecloses these other possibilities. I acknowledge that the practical difference between the majority’s construction of Proposition 64 and my literal one is small. As I have explained, the vast majority of representative plaintiffs in UCL actions cannot hope to comply with section 382 except through class certification. Thus, my disagreement with the majority affects very few cases.
Nevertheless, strict fidelity to the language of voter initiatives is important. The specific language of an initiative measure typically represents “ ‘a delicate tightrope walk designed to induce voter approval . . .’ ” (Ross v. RagingWire Telecommunications, Inc. (2008) 42 Cal.4th 920, 930 [70 Cal.Rptr.3d 382, 174 P.3d 200], quoting People v. Galambos (2002) 104 Cal.App.4th 1147, 1152 [128 Cal.Rptr.2d 844])—a balance that judges too easily upset by reading their own policy preferences into a measure’s language. Thus, “the initiative power is strongest when courts give effect to the voters’ formally expressed intent . . . .” (Ross, supra, at p. 930.) The majority’s only justification for giving Proposition 64 a nonliteral interpretation is that the voters were told—albeit not in the text of the statute on which they were asked to vote—that the measure would compel representative plaintiffs to meet the requirements of class actions. (See maj. opn., ante, at pp. 979-980, citing Voter Information Guide, supra, official title and summary of Prop. 64, p. 38; id., ballot measure summary, Prop. 64, p. 6; id., analysis by the Legislative Analyst of Prop. 64, pp. 38-39.) The majority reasons that “[a] literal construction of an enactment . . . will not control when such a construction would frustrate the manifest purpose of the enactment as a whole.” (Maj. opn., ante, at p. 979.) I agree with the principle but not its application. In this case, to interpret Proposition 64 literally would not frustrate the voters’ intent, given the expected rarity in UCL cases of constitutionally permissible representative actions other than class actions. To prefer language in ballot pamphlets to the formal, operative text of an initiative renders the initiative process susceptible to bait-and-switch tactics. To do so even once without the plainest compulsion sets a potentially dangerous precedent.
*992Accordingly, I cannot join the majority in construing Proposition 64 according to its subjective, court-declared “spirit” rather than its “letter” (maj. opn., ante, at p. 979) without a better reason to believe the voters did not really intend to be bound by language they voted to enact. Nevertheless, I agree with the majority that the Court of Appeal correctly struck plaintiff’s representative claims under the circumstances of this case because plaintiff cannot otherwise “compl[y] with Section 382” (Bus. & Prof. Code, § 17203) and, thus, satisfy Proposition 64.
California’s Code Commissioners, in recommending section 382 to the Legislature as part of the 1872 Code of Civil Procedure, offered the examples of an action by a joint association composed of many individuals, an action by one stockholder on behalf of all against a corporation to compel an accounting, an action by one person on behalf of many claiming title to property from a single source, and an action by one partner on behalf of others to redress an *989indivisible injury to the partnership. (Code commrs., notes foll., Ann. Code Civ. Proc., § 382 (1st ed. 1872, Raymond & Burch, commrs. annotators) pp. 242-244.)
The Legislature in 1970 filled this void only incompletely with the enactment of the Consumers Legal Remedies Act. (Civ. Code, § 1750 et seq.; see id., § 1781.) Lacking further legislative guidance, the Judicial Council has adopted rules governing some aspects of class action procedure, such as notice, but not the standards for class certification. (See Cal. Rules of Court, rule 3.760 et seq.)
Three of the criteria for virtual representation set out in section 382—common interest, numerosity and the impracticability of joinder—have found their way into the modem jurisprudence of class actions. (See Fed. Rules Civ.Proc., rule 23(a)(l)-(2), (b)(1)(B), 28 U.S.C.) But, as section 382 was never intended to codify class action procedure, it says nothing about other important requirements such as the existence of common questions of law, the typicality of claims, the ability of the named plaintiff to provide fair and adequate representation, the superiority of a class action over other methods of adjudication, the likely difficulties of managing a class action, and the requirement of notice. (See Fed. Rules Civ.Proc., rule 23(a)(3), (4), (b)(3), (b)(3)(D), (c)(2)(B), 28 U.S.C.)