*1017Opinion
WERDEGAR, J.For the better part of a century, California law has guaranteed to employees wage and hour protection, including meal and rest periods intended to ameliorate the consequences of long hours. For most of that time, only injunctive remedies were available for violations of meal and rest period guarantees. In 2000, however, both the Legislature and the Industrial Welfare Commission (IWC) adopted for the first time monetary remedies for the denial of meal and rest breaks. (Murphy v. Kenneth Cole Productions, Inc. (2007) 40 Cal.4th 1094, 1105-1106 [56 Cal.Rptr.3d 880, 155 P.3d 284].) These remedies engendered a wave of wage and hour class action litigation, including the instant suit in which the trial court granted class certification and the Court of Appeal then issued writ relief and ordered three subclasses decertified.
We granted review to consider issues of significance to class actions generally and to meal and rest break class actions in particular. We conclude, contrary to the Court of Appeal, that trial courts are not obligated as a matter of law to resolve threshold disputes over the elements of a plaintiffs claims, unless a particular determination is necessarily dispositive of the certification question. Because the parties have so requested, however, we nevertheless address several such threshold disputes here. On the most contentious of these, the nature of an employer’s duty to provide meal periods, we conclude an employer’s obligation is to relieve its employee of all duty, with the employee thereafter at liberty to use the meal period for whatever purpose he or she desires, but the employer need not ensure that no work is done.
On the ultimate question of class certification, we review the trial court’s ruling for abuse of discretion. In light of the substantial evidence submitted by plaintiffs of defendants’ uniform policy, we conclude the trial court properly certified a rest break subclass. On the question of meal break subclass certification, we remand to the trial court for reconsideration. With respect to the third contested subclass, covering allegations that employees were required to work “off-the-clock,” no evidence of common policies or means of proof was supplied, and the trial court therefore erred in certifying a subclass. Accordingly, because the Court of Appeal rejected certification of all three subclasses, we will affirm in part, reverse in part, and remand for further proceedings.
Factual and Procedural Background
Defendants Brinker Restaurant Corporation, Brinker International, Inc., and Brinker International Payroll Company, L.P. (collectively Brinker), own and operate restaurants throughout California, including Chili’s Grill & Bar *1018and Maggiano’s Little Italy. Brinker previously has owned and operated additional chains in California, including Romano’s Macaroni Grill, Comer Bakery Cafe, Cozymel’s Mexican Grill, and On the Border Mexican Grill & Cantina. Name plaintiffs Adam Hohnbaum, Iliya Haase, Romeo Osorio, Amanda June Rader, and Santana Alvarado (collectively Hohnbaum) are or were hourly nonexempt employees at one or more of Brinker’s restaurants.
State law obligates employers to afford their nonexempt employees meal periods and rest periods during the workday. (See Lab. Code, §§ 226.7, 512; IWC wage order No. 5-2001 (Cal. Code Regs., tit. 8, § 11050), hereafter Wage Order No. 5.)1 Labor Code section 226.7, subdivision (a)2 prohibits an employer from requiring an employee “to work during any meal or rest period mandated by an applicable order of the Industrial Welfare Commission.” In turn, Wage Order No. 5, subdivision 12 prescribes rest periods, while subdivision 11, as well as section 512 of the Labor Code, prescribes meal periods. Employers who violate these requirements must pay premium wages. (§226.7, subd. (b); Wage Order No. 5, subds. 11(B), 12(B); see Murphy v. Kenneth Cole Productions, Inc., supra, 40 Cal.4th at p. 1114.)
In 2002, the Division of Labor Standards Enforcement (DLSE) launched an investigation into whether Brinker was complying with its obligations to provide rest and meal breaks, maintain proper records, and pay premium wages in the event required breaks were not provided. The DLSE filed suit and eventually settled in exchange for Blinker’s payment of $10 million to redress injuries suffered by employees between 1999 and 2001 and the stipulation to a court-ordered injunction to ensure compliance with meal and rest break laws. In connection with the settlement, Brinker disclaimed all liability.
In the aftermath of the DLSE’s suit, Hohnbaum filed this putative class action, seeking to represent the cooks, stewards, buspersons, wait staff, host staff, and other hourly employees who staff Brinker’s restaurants. The operative complaint, the first amended complaint, alleges in its first cause of action that Brinker failed to provide employees the rest breaks, or premium wages in lieu of rest breaks, due them under law. (See § 226.7; Wage Order No. 5, subd. 12.) The second cause of action alleges Brinker failed to provide employees the meal breaks, or premium wages in lieu of meal breaks, required by law. (See §§ 226.7, 512; Wage Order No. 5, subd. 11.) In the *1019course of litigation, two distinct theories underlying the meal break claim have emerged: (1) Brinker provided employees fewer meal periods than required by section 512 and Wage Order No. 5; and (2) Brinker sometimes required “early lunching,” a single meal period soon after the beginning of a work shift followed by six, seven, eight, or more hours without an additional meal period. Finally, Hohnbaum contends Brinker required employees to work off-the-clock during meal periods and engaged in time shaving, unlawfully altering employee time records to misreport the amount of time worked and break time taken.3
In aid of a court-ordered mediation, the parties stipulated to the trial court’s resolving the legal issue central to the early lunching theory: whether state law imposes timing requirements on when a meal period must be provided and, if so, what it requires. Hohnbaum contended governing law obligates an employer to provide a 30-minute meal period at least once every five hours. Brinker countered that no such timing obligation is imposed, and an employer satisfies its meal period obligations by providing one meal period for shifts over five hours and two meal periods for shifts over 10 hours.
The trial court generally agreed with Hohnbaum, holding that an employer’s obligations are not satisfied simply by affording a meal period for each work shift longer than five hours, and that affording a meal period during the first hour of a 10-hour shift, with nothing during the remaining nine hours, would violate the obligation to provide a meal period for each five-hour work period. This advisory opinion subsequently was confirmed as a court order. Brinker filed a writ petition in the Court of Appeal, which was denied.
Hohnbaum then moved for class certification, defining the class as “[a]ll present and former employees of [Brinker] who worked at a Brinker owned restaurant in California, holding a non-exempt position, from and after August 16, 2000.”4 The class definition included several subclasses, three of which are pertinent here: (1) a “ ‘Rest Period Subclass’ ” comprising all “Class Members who worked one or more work periods in excess of three and a half (3.5) hours without receiving a paid 10 minute break during which the Class Member was relieved of all duties, from and after October 1, 2000”; (2) a “ ‘Meal Period Subclass’ ” covering all “Class Members who worked one or more work periods in excess of five (5) consecutive hours, without receiving a thirty (30) minute meal period during which the Class Member was relieved of all duties, from and after October 1, 2000”; and (3) an “ ‘Off-The-Clock’ Subclass” for all “Class Members who worked ‘off-the-clock’ or without pay from and after August 16, 2000.”
*1020Hohnbaum argued class certification was warranted because, inter alia, common legal and factual issues predominated. He contended Brinker applied common meal and rest break policies to all nonexempt employees, the legality of these common policies was most appropriately decided on a classwide basis, and computer shift records maintained by Brinker could be used to identify violations and establish classwide liability. Hohnbaum supported the motion with numerous declarations from proposed class members asserting that Brinker had failed to provide individuals with meal and rest breaks or provided breaks at allegedly improper times during the course of an employee’s work shift. He also submitted survey evidence of ongoing meal and rest break violations even after settlement with the DLSE.
Brinker opposed class certification, arguing that individual issues predominated. Specifically, Brinker argued that a rest break subclass should not be certified because an employer’s obligation is simply to permit such breaks to be taken, as Brinker did, and whether employees in fact chose to take such breaks is an individualized inquiry not amenable to class treatment. Brinker contended a meal period subclass should not be certified because an employer is obliged only to make meal breaks available and need not ensure that employees take such breaks. Brinker asserted it had complied with its legal obligation to make meal breaks available, many employees took those breaks, and inquiry into why particular employees did not take meal breaks raised individual questions precluding class treatment. Brinker also contended plaintiffs’ early lunching claims were legally unfounded and, in any event, individual issues again predominated, rendering the meal period claims unsuitable for litigation on a class basis. Finally, Brinker argued the off-the-clock subclass should not be certified because no Brinker policy permitted such alteration of time records, Brinker did not suffer or permit off-the-clock work, and any such off-the-clock work would require individualized employee-by-employee proof. Brinker submitted hundreds of declarations in support of its opposition to class certification.
Following a full hearing, the trial court granted class certification, finding that common issues predominated over individual issues: “[Cjommon questions regarding the meal and rest period breaks are sufficiently pervasive to permit adjudication in this one class action. [][] [Brinker’s] arguments regarding the necessity of making employees take meal and rest periods actually point[] toward a common legal issue of what [Brinker] must do to comply with the Labor Code. Although a determination that [Brinker] need not force employees to take breaks may require some individualized discovery, the common alleged issues of meal and rest violations predominate.” A class proceeding was also superior: “Adjudicating plaintiffs’ allegations in one litigation” would be “much more efficient” than resolving it in 60,000 separate administrative or judicial proceedings, as Brinker had suggested.
*1021The Court of Appeal granted writ relief and reversed class certification as to the three disputed subclasses. We granted review to resolve uncertainties in the handling of wage and hour class certification motions.
Discussion
I. Class Certification Principles
Originally creatures of equity, class actions have been statutorily embraced by the Legislature whenever “the question [in a case] is one of a common or general interest, of many persons, or when the parties are numerous, and it is impracticable to bring them all before the court . . . .” (Code Civ. Proc., § 382; see Fireside Bank v. Superior Court (2007) 40 Cal.4th 1069, 1078 [56 Cal.Rptr.3d 861, 155 P.3d 268]; City of San Jose v. Superior Court (1974) 12 Cal.3d 447, 458 [115 Cal.Rptr. 797, 525 P.2d 701].) Drawing on the language of Code of Civil Procedure section 382 and federal precedent, we have articulated clear requirements for the certification of a class. The party advocating class treatment must demonstrate the existence of an ascertainable and sufficiently numerous class, a well-defined community of interest, and substantial benefits from certification that render proceeding as a class superior to the alternatives. (Code Civ. Proc., § 382; Fireside Bank, at p. 1089; Linder v. Thrifty Oil Co. (2000) 23 Cal.4th 429, 435 [97 Cal.Rptr.2d 179, 2 P.3d 27]; City of San Jose, at p. 459.) “In turn, the ‘community of interest requirement embodies three factors: (1) predominant common questions of law or fact; (2) class representatives with claims or defenses typical of the class; and (3) class representatives who can adequately represent the class.’ ” (Fireside Bank, at p. 1089, quoting Richmond v. Dart Industries, Inc. (1981) 29 Cal.3d 462, 470 [174 Cal.Rptr. 515, 629 P.2d 23].)
Here, only a single element of class suitability, and a single aspect of the trial court’s certification decision, is in dispute: whether individual questions or questions of common or general interest predominate. The “ultimate question” the element of predominance presents is whether “the issues which may be jointly tried, when compared with those requiring separate adjudication, are so numerous or substantial that the maintenance of a class action would be advantageous to the judicial process and to the litigants.” (Collins v. Rocha (1972) 7 Cal.3d 232, 238 [102 Cal.Rptr. 1, 497 P.2d 225]; accord, Sav-On Drug Stores, Inc. v. Superior Court (2004) 34 Cal.4th 319, 326 [17 Cal.Rptr.3d 906, 96 P.3d 194].) The answer hinges on “whether the theory of recovery advanced by the proponents of certification is, as an analytical matter, likely to prove amenable to class treatment.” (Sav-On, at p. 327.) A court must examine the allegations of the complaint and supporting declarations (ibid.) and consider whether the legal and factual issues they present are such that their resolution in a single class proceeding *1022would be both desirable and feasible.5 “As a general rule if the defendant’s liability can be determined by facts common to all members of the class, a class will be certified even if the members must individually prove their damages.” (Hicks v. Kaufman & Broad Home Corp. (2001) 89 Cal.App.4th 908, 916 [107 Cal.Rptr.2d 761]; accord, Knapp v. AT&T Wireless Services, Inc. (2011) 195 Cal.App.4th 932, 941 [124 Cal.Rptr.3d 565].)
On review of a class certification order, an appellate court’s inquiry is narrowly circumscribed. “The decision to certify a class rests squarely within the discretion of the trial court, and we afford that decision great deference on appeal, reversing only for a manifest abuse of discretion: ‘Because trial courts are ideally situated to evaluate the efficiencies and practicalities of permitting group action, they are afforded great discretion in granting or denying certification.’ [Citation.] A certification order generally will not be disturbed unless (1) it is unsupported by substantial evidence, (2) it rests on improper criteria, or (3) it rests on erroneous legal assumptions. [Citations.]” (Fireside Bank v. Superior Court, supra, 40 Cal.4th at p. 1089; see also Hamwi v. Citinational-Buckeye Inv. Co. (1977) 72 Cal.App.3d 462, 472 [140 Cal.Rptr. 215] [“So long as [the trial] court applies proper criteria and its action is founded on a rational basis, its ruling must be upheld.”].) Predominance is a factual question; accordingly, the trial court’s finding that common issues predominate generally is reviewed for substantial evidence. (Sav-On Drug Stores, Inc. v. Superior Court, supra, 34 Cal.4th at pp. 328-329.) We must “[p]resum[e] in favor of the certification order ... the existence of every fact the trial court could reasonably deduce from the record . . . .” (Id. at p. 329.)
The appellate judgment reversing certification rests on two separate grounds. First, the Court of Appeal held the trial court committed error per se by ruling on certification without first resolving legal disputes over the scope of Brinker’s duties to provide meal and rest periods. Second, it held that any court, upon resolving those disputes, could only have concluded certification was inappropriate. We consider the first of these grounds in part II., post, and the second of them in parts IV. through VI., post. As we shall explain, the first ground does not support the judgment, while the second supports it only partially.
*1023II. Class Certification and Disputes over a Claim’s Elements
The trial court concluded it could certify a class without resolving disputes over the scope of Brinker’s duty to provide breaks because common questions would predominate even if Brinker’s legal positions were correct. According to the Court of Appeal, this was error: the trial court “was required to determine the elements of plaintiffs’ claims” because the court “could not determine whether individual or common issues predominate in this case, and thus whether a class action was proper, without first determining this threshold issue.” While we agree trial courts must resolve any legal or factual issues that are necessary to a determination whether class certification is proper, the Court of Appeal went too far by intimating that a trial court must as a threshold matter always resolve any party disputes over the elements of a claim. In many instances, whether class certification is appropriate or inappropriate may be determined irrespective of which party is correct. In such circumstances, it is not an abuse of discretion to postpone resolution of the disputed issue.
“The certification question is ‘essentially a procedural one that does not ask whether an action is legally or factually meritorious.’ ” (Sav-On Drug Stores, Inc. v. Superior Court, supra, 34 Cal.4th at p. 326, quoting Linder v. Thrifty Oil Co., supra, 23 Cal.4th at pp. 439-440; see also Eisen v. Carlisle & Jacquelin (1974) 417 U.S. 156, 178 [40 L.Ed.2d 732, 94 S.Ct. 2140] [“ ‘In determining the propriety of a class action, the question is not whether the plaintiff or plaintiffs have stated a cause of action or will prevail on the merits, but rather whether the requirements of [class certification] are met.’ ”].) A class certification motion is not a license for a free-floating inquiry into the validity of the complaint’s allegations; rather, resolution of disputes over the merits of a case generally must be postponed until after class certification has been decided (Fireside Bank v. Superior Court, supra, 40 Cal.4th at pp. 1083-1086), with the court assuming for purposes of the certification motion that any claims have merit (Linder, at p. 443).
We have recognized, however, that “issues affecting the merits of a case may be enmeshed with class action requirements ...” (Linder v. Thrifty Oil Co., supra, 23 Cal.4th at p. 443; see also Wal-Mart Stores, Inc. v. Dukes (2011) 564 U.S._,_[180 L.Ed.2d 374, 131 S.Ct. 2541, 2551] [Analysis of a class certification’s propriety “[frequently . . . will entail some overlap with the merits of the plaintiff’s underlying claim. That cannot be helped.”]; Coopers & Lybrand v. Livesay (1978) 437 U.S. 463, 469, fn. 12 [57 L.Ed.2d 351, 98 S.Ct. 2454] [“ ‘Evaluation of many of the questions entering into determination of class action questions is intimately involved with the merits of the claims.’ ”].) When evidence or legal issues germane to the certification question bear as well on aspects of the merits, a court may properly evaluate *1024them. (Wal-Mart Stores, 564 U.S. at pp. _-_ & fn. 6 [131 S.Ct. at pp. 2551-2552 & fn. 6]; Ghazaryan v. Diva Limousine, Ltd. (2008) 169 Cal.App.4th 1524, 1531 [87 Cal.Rptr.3d 518]; Caro v. Proctor & Gamble Co. (1993) 18 Cal.App.4th 644, 656 [22 Cal.Rptr.2d 419].) The rule is that a court may “consider[] how various claims and defenses relate and may affect the course of the litigation” even though such “considerations . . . may overlap the case’s merits.” (Fireside Bank v. Superior Court, supra, 40 Cal.4th at p. 1092; see Szabo v. Bridgeport Machines, Inc. (7th Cir. 2001) 249 F.3d 672, 676 [if the considerations necessary to certification “overlap the merits .. . then the judge must make a preliminary inquiry into the merits”].)
In particular, whether common or individual questions predominate will often depend upon resolution of issues closely tied to the merits. (Coopers & Lybrand v. Livesay, supra, 437 U.S. at p. 469, fn. 12; Linder v. Thrifty Oil Co., supra, 23 Cal.4th at p. 443.) To assess predominance, a court “must examine the issues framed by the pleadings and the law applicable to the causes of action alleged.” (Hicks v. Kaufman & Broad Home Corp., supra, 89 Cal.App.4th at p. 916.) It must determine whether the elements necessary to establish liability are susceptible of common proof or, if not, whether there are ways to manage effectively proof of any elements that may require individualized evidence. (See Sav-On Drug Stores, Inc. v. Superior Court, supra, 34 Cal.4th at p. 334.) In turn, whether an element may be established collectively or only individually, plaintiff by plaintiff, can turn on the precise nature of the element and require resolution of disputed legal or factual issues affecting the merits. For example, whether reliance or a breach of duty can be demonstrated collectively or poses insuperable problems of individualized proof may be determinable only after closer inspection of the nature of the reliance required or duty owed and, in some instances, resolution of legal or factual disputes going directly to the merits. (See, e.g., Erica P. John Fund, Inc. v. Halliburton Co. (2011) 563 U.S. _, _-_ [180 L.Ed.2d 24, 131 S.Ct. 2179, 2184-2186]; Bartold v. Glendale Federal Bank (2000) 81 Cal.App.4th 816, 829-831 [97 Cal.Rptr.2d 226].)
Such inquiries are closely circumscribed. As the Seventh Circuit has correctly explained, any “peek” a court takes into the merits at the certification stage must “be limited to those aspects of the merits that affect the decisions essential” to class certification. (Schleicher v. Wendt (7th Cir. 2010) 618 F.3d 679, 685.) While the Schleicher defendants urged that the trial court had erred by failing to resolve disputes over the falsity and materiality of their statements, the Seventh Circuit affirmed class certification without inquiry into such matters, concluding no element of the certification determination hinged on their resolution. (Ibid.) Likewise, in Jaimez v. Daiohs USA, Inc. (2010) 181 Cal.App.4th 1286, 1303-1305 [105 Cal.Rptr.3d 443], the Court of Appeal reversed the trial court’s refusal to certify a wage and hour class without deciding contested legal issues concerning the defendant’s meal *1025break policy because common questions predominated in any event. (See also Medrazo v. Honda of North Hollywood (2008) 166 Cal.App.4th 89, 97-98 [82 Cal.Rptr.3d 1] [trial court erred in resolving the merits of an affirmative defense divorced from consideration of the specific criteria for class certification].)
We summarize the governing principles. Presented with a class certification motion, a trial court must examine the plaintiff’s theory of recovery, assess the nature of the legal and factual disputes likely to be presented, and decide whether individual or common issues predominate. To the extent the propriety of certification depends upon disputed threshold legal or factual questions, a court may, and indeed must, resolve them. Out of respect for the problems arising from one-way intervention, however, a court generally should eschew resolution of such issues unless necessary. (See Fireside Bank v. Superior Court, supra, 40 Cal.4th at p. 1074; Schleicher v. Wendt, supra, 618 F.3d at p. 685.) Consequently, a trial court does not abuse its discretion if it certifies (or denies certification of) a class without deciding one or more issues affecting the nature of a given element if resolution of such issues would not affect the ultimate certification decision.6
In support of its conclusion that a trial court must always first decide upon the applicable law and resolve legal issues surrounding each element of a proposed class claim, the Court of Appeal relied principally on our decision in Washington Mutual Bank v. Superior Court (2001) 24 Cal.4th 906 [103 Cal.Rptr.2d 320, 15 P.3d 1071]. We disagree with the Court of Appeal’s reading of our decision. In Washington Mutual, the plaintiffs sought certification of a nationwide class. Although members of the plaintiff class were subject to choice-of-law agreements, the trial court granted certification without first determining whether the agreements were enforceable and would result in the application of different state laws, and whether any applicable state laws varied in ways that would render the class proceeding unmanageable. We reversed, explaining that it was not possible to intelligently assess predominance and the manageability of claims asserted on behalf of nonresidents without those determinations. (Washington Mutual, at pp. 915, 922, 927-928.) Washington Mutual involves an unexceptional application of the principles we have articulated: if the presence of an element necessary to certification, such as predominance, cannot be determined without resolving a particular legal issue, the trial court must resolve that issue at the certification *1026stage. That the failure to resolve disputed legal issues affecting the elements of a claim is always reversible error does not follow.
III. Wage Orders and the Labor Code
We turn to the Court of Appeal’s alternate basis for reversing class certification—that if one considers the substance of the parties’ various legal disputes and the elements of Hohnbaum’s claims, one must conclude as a matter of law that common questions do not predominate. In assessing that conclusion, at the parties’ request we examine the merits of their substantive legal disputes. (See Linder v. Thrifty Oil Co., supra, 23 Cal.4th at p. 443 [“[W]e see nothing to prevent a court from considering the legal sufficiency of claims when ruling on certification where both sides jointly request such action.”].) Because those disputes derive in part from conflicting visions of the respective roles statutes and wage orders play in establishing the state’s wage and hour law, we begin by examining those roles.
Nearly a century ago, the Legislature responded to the problem of inadequate wages and poor working conditions by establishing the IWC and delegating to it the authority to investigate various industries and promulgate wage orders fixing for each industry minimum wages, maximum hours of work, and conditions of labor. (Martinez v. Combs, supra, 49 Cal.4th at pp. 52-55; see Cal. Const., art. XIV, § 1 [confirming the Legislature’s authority to establish a commission and grant it legislative and other powers over such matters].) Pursuant to its “broad statutory authority” (Industrial Welfare Com. v. Superior Court (1980) 27 Cal.3d 690, 701 [166 Cal.Rptr. 331, 613 P.2d 579]), the IWC in 1916 began issuing industry- and occupation-wide wage orders specifying minimum requirements with respect to wages, hours, and working conditions (id. at p. 700). In addition, the Legislature has from time to time enacted statutes to regulate wages, hours, and working conditions directly. Consequently, wage and hour claims are today governed by two complementary and occasionally overlapping sources of authority: the provisions of the Labor Code, enacted by the Legislature, and a series of 18 wage orders, adopted by the IWC. (Reynolds v. Bement (2005) 36 Cal.4th 1075, 1084 [32 Cal.Rptr.3d 483, 116 P.3d 1162]; see IWC wage orders Nos. 1-2001 to 17-2001 and MW-2007 (Cal. Code Regs., tit. 8, §§ 11000-11170).)
We apply the usual rules of statutory interpretation to the Labor Code, beginning with and focusing on the text as the best indicator of legislative purpose. (Murphy v. Kenneth Cole Productions, Inc., supra, 40 Cal.4th at p. 1103.) “[I]n light of the remedial nature of the legislative enactments authorizing the regulation of wages, hours and working conditions for the protection and benefit of employees, the statutory provisions are *1027to be liberally construed with an eye to promoting such protection.” (Industrial Welfare Com. v. Superior Court, supra, 27 Cal.3d at p. 702; see also Murphy, at p. 1103 [given the Legislature’s remedial purpose, “statutes governing conditions of employment are to be construed broadly in favor of protecting employees”].)
In turn, the IWC’s wage orders are entitled to “extraordinary deference, both in upholding their validity and in enforcing their specific terms.” (Martinez v. Combs, supra, 49 Cal.4th at p. 61.) When a wage order’s validity and application are conceded and the question is only one of interpretation, the usual rules of statutory interpretation apply. (Collins v. Overnite Transportation Co. (2003) 105 Cal.App.4th 171, 178-179 [129 Cal.Rptr.2d 254]; see Cal. Drive-in Restaurant Assn. v. Clark (1943) 22 Cal.2d 287, 292 [140 P.2d 657].) As with the Labor Code provisions at issue, the meal and rest period requirements we must construe “have long been viewed as part of the remedial worker protection framework.” (Murphy v. Kenneth Cole Productions, Inc., supra, 40 Cal.4th at p. 1105.) Accordingly, the relevant wage order provisions must be interpreted in the manner that best effectuates that protective intent. (Martinez, at pp. 61-62; see Industrial Welfare Com. v. Superior Court, supra, 27 Cal.3d at p. 724; Bono Enterprises, Inc. v. Bradshaw (1995) 32 Cal.App.4th 968, 974 [38 Cal.Rptr.2d 549].)
The IWC’s wage orders are to be accorded the same dignity as statutes. They are “presumptively valid” legislative regulations of the employment relationship (Martinez v. Combs, supra, 49 Cal.4th at p. 65), regulations that must be given “independent effect” separate and apart from any statutory enactments (id. at p. 68). To the extent a wage order and a statute overlap, we will seek to harmonize them, as we would with any two statutes. (Cal. Drive-in Restaurant Assn. v. Clark, supra, 22 Cal.2d at pp. 292-293.)
Here, Wage Order No. 5, governing the public housekeeping industry, applies.7 We consider in turn both the scope of the duties it and several related statutes (see §§ 226.7, 512, 516) impose on restaurant employers to afford rest and meal periods, and whether in light of those duties the Court of Appeal erred in reversing as an abuse of discretion the trial court’s certification of three subclasses.8
*1028IV. Rest Period Class Certification
A. The Scope of an Employer’s Duty to Provide Rest Periods
Preliminary to its assessment of the trial court’s certification of a rest period subclass, the Court of Appeal addressed two threshold legal questions: the amount of rest time that must be authorized, and the timing of any rest periods. We consider these same two questions.
1. The rate at which rest time must be authorized and permitted
Brinker’s rest period duties are defined solely by Wage Order No. 5, subdivision 12. To determine the rate at which rest time must be authorized, we begin, as always, with the text. (See Reynolds v. Bement, supra, 36 Cal.4th at p. 1086 [“The best indicator of [the IWC’s] intent is the language of the [wage order] provision itself.”].) Subdivision 12(A) provides in relevant part: “Every employer shall authorize and permit all employees to take rest periods, which insofar as practicable shall be in the middle of each work period. The authorized rest period time shall be based on the total hours worked daily at the rate of ten (10) minutes net rest time per four (4) hours or major fraction thereof. However, a rest period need not be authorized for employees whose total daily work time is less than three and one-half (SVi) hours.”
The text of the wage order is dispositive; it defines clearly how much rest time must be authorized. Under Wage Order No. 5, subdivision 12(A)’s second sentence, employees receive 10 minutes for each four hours of work “or major fraction thereof.” Though not defined in the wage order, a “major fraction” long has been understood—legally, mathematically, and linguistically—to mean a fraction greater than one-half.9 The term “majority fraction” was first introduced in 1947 and then amended to “major fraction” in 1952;10 the contemporaneous historical evidence suggests the IWC in the 1940’s understood the term in just such a sense. (See, e.g., IWC meeting mins. (June 14, 1943) p. 22 [interpreting “ ‘any fraction of fifteen minutes’ ” to mean “the majority fraction thereof, or eight minutes or *1029more”].) The Division of Labor Standards Enforcement (DLSE) has so interpreted the phrase as well, construing “major fraction thereof’ as applied to a four-hour period to mean any amount of time in excess of two hours—i.e., any fraction greater than half. (Dept. Industrial Relations, DLSE Opn. Letter No. 1999.02.16 (Feb. 16, 1999) p. l.)11
It follows that Wage Order No. 5, subdivision 12(A)’s second sentence defines the rest time that must be permitted as the number of hours worked divided by four, rounded down if the fractional part is half or less than half and up if it is more (a “major fraction”), times 10 minutes. Thus, under the initial calculation called for by this part of the wage order, an employee would receive no rest break time for shifts of two hours or less, 10 minutes for shifts lasting more than two hours up to six hours, 20 minutes for shifts lasting more than six hours up to 10 hours, and so on.
Though under the basic calculation the right to 10 minutes’ rest would accrue for any shift lasting more than two hours, the third sentence of Wage Order No. 5’s rest period subdivision modifies this entitlement slightly. Under the third sentence, “a rest period need not be authorized for employees whose total daily work time is less than three and one-half (3½) hours.” (Wage Order No. 5, subd. 12(A).) Thus, employees working shifts lasting over two hours but under three and one-half hours, who otherwise would have been entitled to 10 minutes’ rest, need not be permitted a rest period. The combined effect of the two pertinent sentences, giving full effect to each, is this; Employees are entitled to 10 minutes’ rest for shifts from three and one-half to six hours in length, 20 minutes for shifts of more than six hours up to 10 hours, 30 minutes for shifts of more than 10 hours up to 14 hours, and so on.
The Court of Appeal, however, construed the third sentence of the subdivision as supplying the definition of “major fraction thereof,” reasoning that otherwise the three and one-half hour proviso and the preceding language would be irreconcilable. In its view, employees are entitled to 10 minutes’ rest for shifts of three and one-half hours or more, to 20 minutes’ rest for shifts of seven and one-half hours or more, and so on. An employee working a seven-hour shift thus would be entitled to only 10 minutes’ rest.
*1030This reading cannot be reconciled with either the wage order’s text or its adoption history. First, the express language of the three and one-half hour proviso speaks only to the circumstance where an employee’s “total daily work time is less than three and one-half (3½) hours” (Wage Order No. 5, subd. 12(A), italics added); it does not speak to the circumstance where an employee’s total daily work time is less than seven and one-half hours, or less than 11½ hours. “[M]ajor fraction” can be applied to, and must be defined for, each four-hour period, not just the first four hours of an employee’s shift: How much time must be worked to earn a second 10 minutes, or a third? What does it mean to work four hours plus a “major fraction” of another four hours? The three and one-half hour proviso cannot answer those questions.
Second, the Court of Appeal’s interpretation disregards the use of the word “However” at the beginning of the three and one-half hour proviso, which signals that what follows is a deviation from or exception to the previous rule, not an amplification of it. Though the Court of Appeal perceived an inconsistency, there is nothing inconsistent in reading the three and one-half hour proviso as a specific exception to the general rule that working for a “major fraction” of four hours is sufficient to entitle one to rest time: to earn the first 10 minutes, one must be scheduled for a work shift of at least three and one-half hours, while to earn the next 10 minutes, one must be scheduled to work four hours plus a major fraction, to earn the next, eight hours plus a major fraction, and so on.
The IWC’s explanatory remarks at the time the three and one-half hour proviso was adopted reveal the proviso was intended as just such a limited exception: “ ‘The rest period provision was clarified to indicate that an employee working less than 3½ hours for the entire day would not need to have a rest period.’ ” (IWC meeting mins. (May 16, 1952) p. 34.) The three and one-half hour proviso thus was not inserted as a definition of the phrase “major fraction,” but simply as a limit on the shift length that would warrant any break at all.
Finally, the Court of Appeal attached great significance to a different 1952 change, the substitution in the wage order of “major” for “majority,” but the two terms are essentially synonymous when used as modifiers, and the change appears to have been the product of an idiomatic choice, rather than an intended semantic distinction. (See also IWC wage order No. 5-57, subd. 15(a) (Nov. 15, 1957) [amending toilet requirements to mandate “one toilet for every twenty-five (25) female employees or major fraction thereof’ in lieu of “. . . majority fraction thereof’ with no evident change in meaning].)
*1031Having resolved the amount of rest time an employer must authorize and permit, we turn to the question of when it must be afforded.
2. Rest period timing
Hohnbaum asserts employers have a legal duty to permit their employees a rest period before any meal period. Construing the plain language of the operative wage order, we find no such requirement and agree with the Court of Appeal, which likewise rejected this contention.
Wage Order No. 5, subdivision 12(A) provides in relevant part: “Every employer shall authorize and permit all employees to take rest periods, which insofar as practicable shall be in the middle of each work period.” Neither this part of the wage order nor subdivision 11, governing meal periods, speaks to the sequence of meal and rest breaks. The only constraint on timing is that rest breaks must fall in the middle of work periods “insofar as practicable.” Employers are thus subject to a duty to make a good faith effort to authorize and permit rest breaks in the middle of each work period, but may deviate from that preferred course where practical considerations render it infeasible. At the certification stage, we have no occasion to decide, and express no opinion on, what considerations might be legally sufficient to justify such a departure.
The difficulty with Hohnbaum’s argument that we should read into the wage order an absolute obligation to permit a rest period before a meal period can be illustrated by considering the case of an employee working a six-hour shift. Such an employee is entitled (in the absence of mutual waiver) to a meal period (Wage Order No. 5, subd. 11(A)) and, as discussed above, to a single rest period. Either the rest period must fall before the meal period or it must fall after. Neither text nor logic dictates an order for these, nor does anything in the policies underlying the wage and hour laws12 compel the conclusion that a rest break at the two-hour mark and a meal break at the four-hour mark of such a shift is lawful, while the reverse, a meal break at the two-hour mark and a rest break at the four-hour mark, is per se illegal.
Hohnbaum seeks to overcome the lack of textual support for his position by offering a DLSE opinion letter interpreting the identical language in a different wage order. (Dept. Industrial Relations, DLSE Opn. Letter No. 2001.09.17 (Sept. 17, 2001) [interpreting IWC wage order No. 16-2001 (Cal. Code Regs., tit. 8, § 11160)].) Responding to a hypothetical about an employer who affords employees a meal break at the five-hour mark of an *1032eight-hour shift, the DLSE opined that “absent truly unusual circumstances,” placing both rest breaks before the meal break, and none after, would not comport with the wage order requirement that rest breaks “ ‘insofar as practicable, shall be in the middle of each work period.’ ” (DLSE Opn. Letter No, 2001.09.17, supra, at p. 4.) We have no reason to disagree with the DLSE’s view regarding the scenario it considered, but that view does not establish universally the proposition that an employee’s first rest break must always come sometime before his or her first meal break. Rather, in the context of an eight-hour shift, “[a]s a general matter,” one rest break should fall on either side of the meal break. (Ibid.) Shorter or longer shifts and other factors that render such scheduling impracticable may alter this general rule.
B. Certification of a Rest Period Subclass
In granting class certification, the trial court accepted without modification the proposed class and subclass definitions. The rest period subclass covers “Class Members who worked one or more work periods in excess of three and a half (3.5) hours without receiving a paid 10 minute break during which the Class Member was relieved of all duties, from and after October 1, 2000 (‘Rest Period Subclass’).”
That the trial court did not apply improper criteria, i.e., decide certification on a basis other than whether superiority of the class action mechanism, commonality of issues, and other relevant factors had been shown, is undisputed. (See Sav-On Drug Stores, Inc. v. Superior Court, supra, 34 Cal.4th at p. 332; Walsh v. IKON Office Solutions, Inc. (2007) 148 Cal.App.4th 1440, 1451 [56 Cal.Rptr.3d 534].) Nor, as we have discussed, was the trial court obligated as a matter of law to resolve all legal disputes concerning the elements of Hohnbaum’s rest break claims before certifying a class. (Ante, pt. II.) Hence, the only remaining question is whether the court abused its discretion in concluding that common questions predominate. We conclude it did not.
The issue for the trial court was whether any of the rest break theories of recovery advanced by Hohnbaum were “likely to prove amenable to class treatment.” (Sav-On Drug Stores, Inc. v. Superior Court, supra, 34 Cal.4th at p. 327.) The complaint alleges Blinker failed “to provide rest periods for every four hours or major fraction thereof worked per day to non-exempt employees.” Though Hohnbaum briefs multiple theories of liability, to conclude class certification was not an abuse of discretion we need consider only one; the theory that Blinker adopted a uniform corporate rest break policy that violates Wage Order No. 5 because it fails to give full effect to the “major fraction” language of subdivision 12(A).
*1033Hohnbaum presented evidence of, and indeed Brinker conceded at the class certification hearing the existence of, a common, uniform rest break policy. The rest break policy was established at Brinker’s corporate headquarters; it is equally applicable to all Brinker employees. Under the written policy, employees receive one 10-minute rest break per four hours worked: “If I work over 3.5 hours during my shift, I understand that I am eligible for one ten minute rest break for each four hours that I work.”13 Classwide liability could be established through common proof if Hohnbaum were able to demonstrate that, for example, Brinker under this uniform policy refused to authorize and permit a second rest break for employees working shifts longer than six, but shorter than eight, hours. Claims alleging that a uniform policy consistently applied to a group of employees is in violation of the wage and hour laws are of the sort routinely, and properly, found suitable for class treatment. (See, e.g., Jaimez v. Daiohs USA, Inc., supra, 181 Cal.App.4th at pp. 1299-1305; Ghazaryan v. Diva Limousine, Ltd.., supra, 169 Cal.App.4th at pp. 1533-1538; Bufil v. Dollar Financial Group, Inc. (2008) 162 Cal.App.4th 1193, 1205-1208 [76 Cal.Rptr.3d 804].)
In reversing class certification, the Court of Appeal concluded that because rest breaks can be waived—as all parties agree—“any showing on a class basis that plaintiffs or other members of the proposed class missed rest breaks or took shortened rest breaks would not necessarily establish, without further individualized proof, that Brinker violated” the Labor Code and Wage Order No. 5. This was error. An employer is required to authorize and permit the amount of rest break time called for under the wage order for its industry. If it does not—if, for example, it adopts a uniform policy authorizing and permitting only one rest break for employees working a seven-hour shift when two are required—it has violated the wage order and is liable. No issue of waiver ever arises for a rest break that was required by law but never authorized; if a break is not authorized, an employee has no opportunity to decline to take it. As Hohnbaum pleaded and presented substantial evidence of a uniform rest break policy authorizing breaks only for each full four hours worked, the trial court’s certification of a rest break subclass should not have been disturbed.
We observe in closing that, contrary to the Court of Appeal’s conclusion, the certifiability of a rest break subclass in this case is not dependent upon resolution of threshold legal disputes over the scope of the employer’s rest break duties. The theory of liability—that Brinker has a uniform policy, and that that policy, measured against wage order requirements, allegedly violates the law—is by its nature a common question eminently suited for class treatment. As noted, we have at the parties’ request addressed the merits of *1034their threshold substantive disputes. However, in the general case to prematurely resolve such disputes, conclude a uniform policy complies with the law, and thereafter reject class certification—as the Court of Appeal did— places defendants in jeopardy of multiple class actions, with one after another dismissed until one trial court concludes there is some basis for liability and in that case approves class certification. (See Fireside Bank v. Superior Court, supra, 40 Cal.4th at p. 1078.) It is far better from a fairness perspective to determine class certification independent of threshold questions disposing of the merits, and thus permit defendants who prevail on those merits, equally with those who lose on the merits, to obtain the preclusive benefits of such victories against an entire class and not just a named plaintiff. (Id. at p. 1083.)
V. Meal Period Class Certification
As with the rest break subclass, the Court of Appeal addressed two threshold legal issues before assessing whether certification of a subclass was proper: (1) the nature of an employer’s duty to provide employees with meal periods; and (2) the timing requirements applicable to the provision of meal periods. We likewise begin with these issues.
A. The Scope of the Employer’s Duty to Provide Meal Periods
1. The nature of the duty
We consider what it means for an employer to provide a nonexempt employee a meal period. Hohnbaum contends an employer is obligated to “ensure that work stops for the required thirty minutes.” Brinker, in a position adopted by the Court of Appeal, contends an employer is obligated only to “make available” meal periods, with no responsibility for whether they are taken. We conclude that under Wage Order No. 5 and Labor Code section 512, subdivision (a), an employer must relieve the employee of all duty for the designated period, but need not ensure that the employee does no work.
Historically, an employer’s meal period obligations were governed solely by the language of the IWC’s wage orders, and so we begin there. Under Wage Order No. 5, subdivision 11(A), “[n]o employer shall employ any person for a work period of more than five (5) hours without a meal period of not less than 30 minutes . . .” absent a mutual waiver in certain limited circumstances. The wage order employs no verb between “without” and “a meal period” (e.g., providing, requiring, offering, allowing, granting) to specify the nature of the employer’s duty. Rather, the order identifies only the condition triggering the employer’s duty (employment of any person for at least five hours) and the employee’s concomitant entitlement (a meal period of at least 30 minutes).
*1035In the absence of a verb, the key language giving content to the employer’s duty comes from the wage order’s further definition of what an employee is to receive. Under Wage Order No. 5, subdivision 11(A), “[u]nless the employee is relieved of all duty during a 30 minute meal period, the meal period shall be considered an ‘on duty’ meal period and counted as time worked. An ‘on duty’ meal period shall be permitted only when the nature of the work prevents an employee from being relieved of all duty and when by written agreement between the parties an on-the-job paid meal period is agreed to. The written agreement shall state that the employee may, in writing, revoke the agreement at any time.”
Parsed, the order’s text spells out the nature of “ ‘on duty’ ” meal periods and the precise circumstances in which they are permitted. It follows that absent such circumstances, an employer is obligated to provide an “off-duty” meal period. The attributes of such off-duty meal periods are evident from the nature of their reciprocal, on-duty meal periods. An on duty meal period is one in which an employee is not “relieved of all duty” for the entire 30-minute period. (Wage Order No. 5, subd. 11(A).) An off-duty meal period, therefore, is one in which the employee “is relieved of all duty during [the] 30 minute meal period.” (Ibid., italics added.) Absent circumstances permitting an on-duty meal period, an employer’s obligation is to provide an off-duty meal period: an uninterrupted 30-minute period during which the employee is relieved of all duty.
The IWC’s wage orders have long made a meal period’s duty-free nature its defining characteristic. The 1943 version of the wage order governing restaurant employees first introduced the principle: “No employer shall employ any woman or minor for a work period of more than five (5) hours without an allowance of not less than thirty (30) minutes for a meal. If during such meal period the employee can not be relieved of all duties and permitted to leave the premises, such meal period shall not be deducted from hours worked.” (IWC wage order No. 5 NS, subd. 3(d) (June 28, 1943).) The 1947 wage order retained the duty-free concept, but more clearly specified the circumstances under which an employer would be excused from relieving an employee: “An ‘on duty’ meal period will be permitted only when the nature of the work prevents an employee from being relieved of all duty, and such ‘on duty’ meal period shall be counted as hours worked without deduction from wages.” (IWC wage order No. 5 R, subd. 10 (June 1, 1947).) In 1963, the operative language was amended almost to its current form (IWC wage order No. 5-63, subd. 11(a) (Aug. 30, 1963)), save for the requirement that on-duty *1036meals be agreed to in writing, which was added in 1976 (IWC wage order No. 5-76, subd. 11(A) (Oct. 18, 1976)).14
As the IWC explained plainly in 1979: “A ‘duty free’ meal period is necessary for the welfare of employees. The section is sufficiently flexible to allow for situations where that is not possible,” i.e., by establishing conditions for an on-duty meal period. “The Commission received no compelling evidence and concluded that there was no rationale to warrant any change in this section, the basic provisions of which date back more than 30 years.” (IWC, Statement as to the Basis for Wage Order No. 5-80 (Sept. 7, 1979); accord, IWC, Statement as to the Basis for Wage Order No. 5-89 (Sept. 7, 1989); IWC statement of findings in support of 1976 wage order revisions (Aug. 13, 1976) p. 14.)
The DLSE’s contemporaneous opinion letters reflect the same understanding. In 1988, the DLSE noted it “has historically taken the position that unless employees are relieved of all duties and are free to leave the premises, the meal period is considered as ‘hours worked.’ ” (Dept. Industrial Relations, DLSE Opn. Letter No. 1988.01.05 (Jan. 5, 1988) p. 1.) Three years later, in response to a question concerning employees working in the field free of direct supervision and control, it advised that if “the employee has a reasonable opportunity to take the full thirty-minute period free of any duty, the employer has satisfied his or her obligation. The worker must be free to attend to any personal business he or she may choose during the unpaid meal period.” (Dept. Industrial Relations, DLSE Opn. Letter No. 1991.06.03 (June 3, 1991) p. 1.)15 As these opinion letters make clear, and as the DLSE argues in its amicus curiae brief, the wage order’s meal period requirement is satisfied if the employee (1) has at least 30 minutes uninterrupted, (2) is free to leave the premises, and (3) is relieved of all duty for the entire period. (DLSE Opn. Letter No. 1988.01.05, supra, at p. 1; Dept. Industrial Relations, DLSE Opn. Letter No. 1996.07.12 (July 12, 1996) p. 1.) We agree with this DLSE interpretation of the wage order.
It was against this background that in 1999 the Legislature first regulated meal periods, previously the exclusive province of the IWC. New *1037section 512 made meal periods a statutory as well as a wage order obligation: an employer must “provid[e] the employee with a meal period of not less than 30 minutes” for workdays lasting more than five hours, and provide two meal periods for workdays in excess of 10 hours, subject to waiver in certain circumstances. (Former § 512, enacted by Stats. 1999, ch. 134, § 6, p. 1823.)16
The duty to provide meal periods is not further defined by section 512, but the nature of the duty is evident from surrounding indicia of legislative intent. As discussed, when the Legislature entered the field of meal break regulation in 1999, it entered an area where the IWC and DLSE had, over more than half a century, developed a settled sense of employers’ meal break obligations. In such circumstances, we begin with the assumption the Legislature did not intend to upset existing rules, absent a clear expression of contrary intent. (Industrial Welfare Com. v. Superior Court, supra, 27 Cal.3d at p. 734; see also Cal. Drive-in Restaurant Assn. v. Clark, supra, 22 Cal.2d at p. 292 [statutes should be construed insofar as possible to avoid implied repeal of wage orders].) Section 512’s mandate that employers “provid[e]” 30-minute meal breaks can be read as shorthand for the requirement contemplated in subdivision 11 of most of the IWC’s wage orders: Employers must afford employees uninterrupted half-hour periods in which they are relieved of any duty or employer control and are free to come and go as they please.
Examination of the relevant legislative history confirms this reading. The origins of section 512 trace to the late 1990’s, when the IWC amended five wage orders to abolish daily overtime, limiting overtime compensation to hours worked in excess of 40 per week, rather than hours worked in excess of eight per day, as had previously been the case. (See Johnson v. Arvin-Edison Water Storage Dist. (2009) 174 Cal.App.4th 729, 735 [95 Cal.Rptr.3d 53].) Troubled by this weakening of employee protections, the Legislature enacted the Eight-Hour-Day Restoration and Workplace Flexibility Act of 1999 (Stats. 1999, ch. 134, p. 1820, enacting Assem. Bill No. 60 (1999-2000 Reg. Sess.)), which restored daily overtime, nullified IWC-approved alternative workweek schedules, and directed the IWC to readopt conforming wage orders. (Sen. Rules Com., Off. of Sen. Floor Analyses, 3d reading analysis of Assem. Bill No. 60 (1999-2000 Reg. Sess.) as amended July 1, 1999, pp. 2-5; see also Stats. 1999, ch. 134, § 2, p. 1820; Assem. Conc. Sen. Amends, to Assem. Bill No. 60 (1999-2000 Reg. Sess.) as amended July 1, 1999, p. 6.)
As part of its response to the IWC’s rollback of employee protections, the Legislature wrote into statute various guarantees that previously had *1038been left to the IWC, including meal break guarantees. (§ 512, subd. (a).) The declared intent in enacting section 512 was not to revise existing meal period rules but to codify them in part. (See, e.g., Assem. Com. on Appropriations, Analysis of Assem. Bill No. 60 (1999-2000 Reg. Sess.) as amended Mar. 22, 1999, p. 3; Legis. Counsel’s Dig., Assem. Bill No. 60 (1999-2000 Reg. Sess.) 5 Stats. 1999, Summary Dig., p. 62.) It follows that the duty the Legislature intended to impose was the duty as it had existed under the IWC’s wage orders. Thus, under what is now section 512, subdivision (a), as under Wage Order No. 5, an employer’s obligation when providing a meal period is to relieve its employee of all duty for an uninterrupted 30-minute period.
Hohnbaum contends that an employer has one additional obligation: to ensure that employees do no work during meal periods. He places principal reliance on a series of DLSE opinion letters. In 2001, in the course of discussing rest breaks, the DLSE distinguished an employer’s meal break duties and observed that for meal breaks “[an] employer has an affirmative obligation to ensure that workers are actually relieved of all duty, not performing any work, and free to leave the worksite . . . .” (Dept, of Industrial Relations, DLSE Opn. Letter No. 2001.09.17, supra, p. 4, italics added.) In 2002, the DLSE reiterated the point: with regard to meal periods, “an employer has an affirmative obligation to ensure that workers are actually relieved of all duty, not performing any work, and . . . free to leave the employer’s premises.” (Dept, of Industrial Relations, DLSE Opn. Letter No. 2002.01.28 (Jan. 28, 2002) p. 1, italics added; see also Dept, of Industrial Relations, DLSE Opn. Letter No. 2002.09.04 (Sept. 4, 2002) p. 2 [“[A]s a general rule the required meal period must be an off-duty meal period, during which time the employee ... is not suffered or permitted to work . . . .”].)
We are not persuaded. The difficulty with the view that an employer must ensure no work is done—i.e., prohibit work—is that it lacks any textual basis in the wage order or statute. While at one time the IWC’s wage orders contained language clearly imposing on employers a duty to prevent their employees from working during meal periods,17 we have found no order in the last half-century continuing that obligation. Indeed, the obligation to ensure employees do no work may in some instances be inconsistent with the fundamental employer obligations associated with a meal break: to relieve the employee of all duty and relinquish any employer control over the employee *1039and how he or she spends the time. (See Morillion v. Royal Packing Co., supra, 22 Cal.4th at pp. 584-585 [explaining that voluntary work may occur while not subject to an employer’s control, and its cessation may require the reassertion of employer control].)
For support, Hohnbaum focuses on the phrase “No employer shall employ any person [without the specified meal period] . . .” (Wage Order No. 5, subd. (11)(A)), contending that “employ” includes permitting or suffering one to work, and so the employer is forbidden from permitting an employee to work during a meal break. Although Hohnbaum is entirely correct about the broad meaning the wage order gives the term “employ” (see Wage Order No. 5, subd. 2(E) [“ ‘Employ’ means to engage, suffer, or permit to work.’ ”]),18 his argument misconstrues the role that broad definition plays in the structure of subdivision 11(A). The provision identifies both an employer obligation (relieving employees of all duty for 30 minutes) and a condition precedent or trigger for that obligation. “No employer shall employ” is part of the definition of the trigger, not of the obligation. If an employer engages, suffers, or permits anyone to work for a full five hours, its meal break obligation is triggered. “Employ” relates to what must transpire during the five-hour work period; it does not relate to what must transpire next.
What must transpire after the meal break obligation is triggered is covered by later parts of the subdivision relating to waiver, on-duty meal periods (and by negative implication off-duty meal periods), and premium pay. When someone is suffered or permitted to work—i.e., employed—for five hours, an employer is put to a choice: it must (1) afford an off-duty meal period; (2) consent to a mutually agreed-upon waiver if one hour or less will end the shift; or (3) obtain written agreement to an on-duty meal period if circumstances permit. Failure to do one of these will render the employer liable for premium pay. (§ 226.7, subd. (b); Wage Order No. 5, subd. 11(A), (B).) As earlier discussed, because the defining characteristic of on-duty meal periods is failing to relieve an employee of duty, not simply “suffering or permitting” work to continue, it follows that off-duty meal periods are *1040similarly defined by actually relieving an employee of all duty: doing so transforms what follows into an off-duty meal period, whether or not work continues.19
Proof an employer had knowledge of employees working through meal periods will not alone subject the employer to liability for premium pay; employees cannot manipulate the flexibility granted them by employers to use their breaks as they see fit to generate such liability. On the other hand, an employer may not undermine a formal policy of providing meal breaks by pressuring employees to perform their duties in ways that omit breaks. (Cicairos v. Summit Logistics, Inc. (2005) 133 Cal.App.4th 949, 962-963 [35 Cal.Rptr.3d 243]; see also Jaimez v. Daiohs USA, Inc., supra, 181 Cal.App.4th at pp. 1304—1305 [proof of common scheduling policy that made taking breaks extremely difficult would show violation]; Dilts v. Penske Logistics, LLC (S.D.Cal. 2010) 267 F.R.D. 625, 638 [indicating informal anti-meal-break policy “enforced through ‘ridicule’ or ‘reprimand’ ” would be illegal].) The wage orders and governing statute do not countenance an employer’s exerting coercion against the taking of, creating incentives to forgo, or otherwise encouraging the skipping of legally protected breaks.
To summarize: An employer’s duty with respect to meal breaks under both section 512, subdivision (a) and Wage Order No. 5 is an obligation to provide a meal period to its employees. The employer satisfies this obligation if it relieves its employees of all duty, relinquishes, control over their activities and permits them a reasonable opportunity to take an uninterrupted 30-minute break, and does not impede or discourage them from doing so. What will suffice may vary from industry to industry, and we cannot in the context of this class certification proceeding delineate the full range of approaches that in each instance might be sufficient to satisfy the law.
On the other hand, the employer is not obligated to police meal breaks and ensure no work thereafter is performed. Bona fide relief from duty and the *1041relinquishing of control satisfies the employer’s obligations, and work by a relieved employee during a meal break does not thereby place the employer in violation of its obligations and create liability for premium pay under Wage Order No. 5, subdivision 11(B) and Labor Code section 226.7, subdivision (b).
2. Meal period timing
We turn to the question of timing. To determine whether the IWC or the Legislature intended to regulate meal period timing, we consider the language and history of both Labor Code section 512 and Wage Order No. 5. We conclude that, absent waiver, section 512 requires a first meal period no later than the end of an employee’s fifth hour of work, and a second meal period no later than the end of an employee’s 10th hour of work. We conclude further that, contrary to Hohnbaum’s argument, Wage Order No. 5 does not impose additional timing requirements.
We begin with the text of section 512, subdivision (a). On the subject of first meal periods, it provides: “An employer may not employ an employee for a work period of more than five hours per day without providing the employee with a meal period of not less than 30 minutes, except that if the total work period per day of the employee is no more than six hours, the meal period may be waived by mutual consent of both the employer and employee.” This provision could be interpreted as requiring employers either to provide a meal break after no more than five hours of work in a day, absent waiver, or simply to provide a meal break at any point in scheduled shifts that exceed five hours.
The first interpretation is the correct one: the statute requires a first meal period no later than the start of an employee’s sixth hour of work. Section 512, subdivision (b) resolves the ambiguity. It provides: “Notwithstanding subdivision (a), the Industrial Welfare Commission may adopt a working condition order permitting a meal period to commence after six hours of work if the commission determines that the order is consistent with the health and welfare of the affected employees.” The provision employs the language of timing: the IWC may adopt a rule “permitting a meal period to commence after six hours,” i.e., as late as six hours into a shift. {Ibid., italics added.) By beginning with “Notwithstanding subdivision (a),” the provision further indicates that any such timing rule would otherwise contravene subdivision (a). Only if subdivision (a) was intended to ensure that a first meal period would commence sooner than six hours, after no more than five hours of work, would this be true. (See Assem. Republican Caucus, analysis of Sen. Bill No. 88 (1999-2000 Reg. Sess.) as amended Aug. 10, 2000, p. 1 [prior to the addition of § 512, subd. (b), noting that “[e]xisting law requires *1042that the meal period begin no later than 5 hours after work begins”].) Accordingly, first meal periods must start after no more than five hours.20
We turn to the matter of second meal periods. Section 512, subdivision (a) provides in its second sentence: “An employer may not employ an employee for a work period of more than 10 hours per day without providing the employee with a second meal period of not less than 30 minutes, except that if the total hours worked is no more than 12 hours, the second meal period may be waived by mutual consent of the employer and the employee only if the first meal period was not waived.” As with the first sentence of subdivision (a), this language is susceptible of two readings: it could be interpreted as requiring employers to provide a second meal break after no more than 10 hours of work in a day, absent waiver, or as simply requiring employers to provide at least two separate breaks at any point in scheduled shifts that exceed 10 hours. Significantly, however, the language is parallel to subdivision (a)’s first sentence. Hence, if the first sentence was intended to ensure a first meal period no more than five hours into a shift, as subdivision (b) reveals it was, it follows that the second, parallel, sentence should be read to require a second meal period after no more than 10 hours of work in a day, i.e., no later than what would be the start of the 11th hour of work, absent waiver.
Hohnbaum contends section 512 should be read as requiring as well a second meal period no later than five hours after the end of a first meal period if a shift is to continue. The text does not permit such a reading. It requires a second meal after no more than 10 hours of work; it does not add the caveat “or less, if the first meal period occurs earlier than the end of five hours of work.” Because the statutory text is conclusive, we need not consider extrinsic sources on this point. (Beal Bank, SSB v. Arter & Hadden, LLP (2007) 42 Cal.4th 503, 507-508 [66 Cal.Rptr.3d 52, 167 P.3d 666].)
The further issue is whether Wage Order No. 5 imposes any additional requirement. We agree with Brinker that it does not.
The IWC has long been understood to have the power to adopt requirements beyond those codified in statute. (Industrial Welfare Com. v. Superior Court, supra, 27 Cal.3d at p. 733; Cal. Drive-in Restaurant Assn. v. Clark, supra, 22 Cal.2d at pp. 292-294; see also ante, at p. 1027.) Section 516 creates an exception; it bars the use of this power to diminish section 512’s protections: “Except as provided in Section 512, the Industrial Welfare *1043Commission may adopt or amend working condition orders with respect to break periods, meal periods, and days of rest for any workers in California consistent with the health and welfare of those workers.” (Italics added). While the Legislature in section 516 generally preserved the IWC’s authority to regulate break periods, it intended to prohibit the IWC from amending its wage orders in ways that “conflict[] with [the] 30-minute meal period requirements” in section 512. (Legis. Counsel’s Dig., Sen. Bill No. 88 (1999-2000 Reg. Sess.) 6 Stats. 2000, Summary Dig., p. 212; see Bearden v. U.S. Borax, Inc. (2006) 138 Cal.App.4th 429, 438 [41 Cal.Rptr.3d 482].) In the absence of a conflict, however, the IWC may still augment the statutory framework with additional protections on matters not covered by section 512; that is, the Legislature did not intend to occupy the field of meal period regulation. (See, e.g., Sen. 3d reading analysis of Sen. Bill No. 88 (1999-2000 Reg. Sess.) as amended Aug. 10, 2000, p. 5 [authorizing the IWC to regulate so long as the orders it adopts are “consistent” with § 512]; § 226.7 [imposing premium wages for violations of the IWC’s meal period provisions, rather than § 512].)
The text of Wage Order No. 5 is ambiguous. Subdivision 11(A) provides in relevant part: “No employer shall employ any person for a work period of more than five (5) hours without a meal period of not less than 30 minutes, except that when a work period of not more than six (6) hours will complete the day’s work the meal period may be waived by mutual consent of the employer and employee.” This language may be read to mirror our interpretation of section 512: employees are due a first meal period after no more than five hours of work, a second meal period after no more than 10 hours, and so on. Alternatively, it may be read more restrictively, as allowing an employer to schedule no more than five hours of work between a first meal period and either another meal period or the end of the shift. Thus, for example, an employee given a meal period after three hours of work would become entitled after eight hours of work either to end the shift or to take a second meal period, even though 10 hours of work were not yet complete.21 In the face of this textual ambiguity, we consider the relevant adoption history. (See Manriquez v. Gourley (2003) 105 Cal.App.4th 1227, 1235 [130 Cal.Rptr.2d 209].)
Evidence in the historical record suggests the IWC’s meal period language originally was intended to limit employees to five-hour work intervals without a meal. In 1943, the first version of the current language appeared: *1044“No employer shall employ any woman or minor for a work period of more than five (5) hours without an allowance of not less than thirty (30) minutes for a meal.” (IWC wage order No. 5 NS, subd. 3(d) (June 28, 1943).) The provision’s intended function was to ensure workers were not required to go too long without a meal break; the IWC found “that it is necessary to insure a meal period after not more than 5 hours of work in order to protect the health of women and minors.” (IWC meeting mins. (Feb. 5, 1943) p. 19; accord, IWC meeting mins. (Apr. 14, 1943) p. 6.)22
At the time, this provision was understood to apply to the work intervals that conclude shifts, as well as those that begin shifts. In response to a request from a regulated store that, for shifts running from 9:00 a.m. to 6:00 p.m., employees be permitted to lunch between 11:00 a.m. and noon, i.e., with six hours between the end of the meal period and the shift end, the IWC adopted an exception to the five-hour limit, allowing work periods of up to six hours at the end of shifts. (IWC wage order No. 5 NS, subd. 3(d) (June 28, 1943) [“However, if the employee’s work for the day will be completed within six (6) hours, such meal period need not be given.”]; IWC meeting mins. (Jan. 29, 1943) p. 15; IWC meeting mins. (Feb. 5, 1943) p. 19.)
In 1947, the IWC briefly departed from its original formulation, rewriting the timing requirement to apply only to the beginning of each work shift. (IWC wage order No. 5 R, subd. 10 (June 1, 1947) [“No employee shall be required to work more than five (5) consecutive hours after reporting for work, without a meal period of not less than (30) minutes.”].) In 1952, however, it returned to its previous approach, adopting language that has been carried forward to today without significant change.23 (IWC wage order No. 5-52, subd. 11 (Aug. 1, 1952) [“No employer shall employ any woman or minor for a work period of more than five (5) hours without a meal period of not less than thirty (30) minutes; except that when a work period of not more than six (6) hours will complete the day’s work, the meal period may be waived.”].) The IWC explained this revision was intended to expand the right to meal periods from a single break in the first five hours to one at least every five hours through the day: “ ‘The meal period provision was amended to permit a 6-hour work period without a meal when such a work shift would complete the day’s work, [with] the additional provision that a meal period shall be every 5 hours rather than providing only one meal period within the *1045first 5 hours.’ ” (IWC meeting mins. (May 16, 1952) p. 33 [adopting summary of findings], italics added.)
The IWC’s descriptions of its meal period requirement in the ensuing years similarly reflected an understanding that work periods before and after meals were to be limited to five hours absent waiver. For example, the commission, discussing IWC wage order No. 12-63 (Aug. 30, 1963) (identically worded to Wage Order No. 5 save for a longer permissible period between meals), explained that the meal provision “requires the employer to provide meal periods at intervals of no more than five and one-half hours within the work period.” (Wage Board for IWC Wage Order No. 12—Motion Picture Industry, Rep. & Recommendations (Oct. 21, 1966) p. 6; see also Margaret T. Miller, IWC executive officer, letter to Klaus Wehrenberg (July 13, 1982) p. 2 [under the IWC’s wage orders, “meal periods must be provided ‘at such intervals as will result in no employee working longer than five consecutive hours without an eating period’ ”].)
In 1999, however, the Legislature passed Assembly Bill No. 60 (1999-2000 Reg. Sess.), which among other things repudiated the IWC’s actions in adopting a series of wage orders that had eliminated daily overtime. (Harris v. Superior Court (2011) 53 Cal.4th 170, 177 [135 Cal.Rptr.3d 247, 266 P.3d 953]; ante, at p. 1037.) Assembly Bill No. 60 repealed five wage orders, including IWC wage order No. 5-98 (Jan. 1, 1998), and required the IWC to review its wage orders and readopt orders conforming to the Legislature’s expressed intentions. (§517; Stats. 1999, ch. 134, § 21, p. 1829.) It also enacted section 512, which for the first time set out statutory meal period requirements.
The IWC complied with the directive to adopt new wage orders. Pending completion of plenary review, it issued an interim wage order applicable to all industries, including those previously and subsequently covered by Wage Order No. 5. Notably, the interim order mirrored section 512’s language, spelling out that a second meal period was required after 10 hours of work, rather than leaving the timing of second meal periods to implication, as previous wage orders generally had. (IWC interim wage order—2000 (Mar. 1, 2000) subd. 9.)24 The IWC also explained its intention that, absent waiver, employees were entitled to “a thirty-minute meal period for every 5 hours of work" (IWC official notice, Summary of Interim Wage Order—2000, italics *1046added), a lesser requirement than the IWC’s prior view that “ ‘a meal period shall be every 5 hours’ ” (IWC meeting mins. (May 16, 1952) p. 33). From the text of the interim order and the official explanation, it is apparent the IWC intended a requirement parallel to that of the Legislature’s section 512, with a second meal period due after 10 hours, rather than after an interval of no more than five hours following a first meal period.
Thereafter, the IWC held public hearings and adopted revised wage orders for each industry, including the current version of Wage Order No. 5, wage order No. 5-2001. From our review of the text of the various wage orders, the IWC’s official explanations of its intent behind these orders, and the transcripts of the IWC’s numerous hearings, we conclude the IWC abandoned any requirement that work intervals be limited to five hours following the first meal break.
With only limited exceptions, the IWC intended its 2001 wage orders to embrace section 512’s meal period requirements, not impose different ones. Having borrowed the provisions of Assembly Bill No. 60 (1999-2000 Reg. Sess.), including section 512, for its interim wage order, the IWC simply copied the interim wage order’s meal provision into most of its industry-specific wage orders. (IWC public hearing transcript (June 30, 2000) pp. 7-10 [explaining intent to mirror Assem. Bill No. 60 on meals]; see, e.g., IWC wage order No. 2-2001 (Jan. 1, 2001); IWC wage order No. 3-2001 (Jan. 1, 2001); IWC wage order No. 6-2001 (Jan. 1, 2001); Cal. Code Regs., tit. 8, §§ 11020, subd. 11(A), (B), 11030, subd. 11(A), (B), 11060, subd. 11(A), (B).) The IWC explained that under these wage orders, first meals would continue to be assured for employees “working for a period of more than five (5) hours,” while second meal periods would now be provided “in accordance with Labor Code § 512(a).” (IWC, Statement as to the Basis (Jan. 1, 2001) p. 19; see also IWC summary of amends, to wage orders Nos. 1-13, 15 & 17 (Jan. 1, 2001) [except as specified in wage orders Nos. 4, 5 & 12, employees are entitled to “a 30-minute meal period for every 5 hours of work”].) Thus, as to the majority of its 2001 wage orders, the IWC did not intend to impose a different meal period requirement than that spelled out in section 512; specifically, it did not intend to require employers to provide employees a second meal period no more than five hours after a first meal period. These orders and the statute are congruent; under each, a first meal period is guaranteed after five hours of work, while a second meal period is required only after 10 hours of work.
The IWC varied slightly the language of wage orders Nos. 4-2001 and 5-2001. These two orders retained the same subdivision 11(A) language *1047requiring a meal period for every five hours, as in other wage orders, but they omitted the subdivision 11(B) language used elsewhere to define the conditions for receiving, and for waiving, a second meal period. As we shall explain, this omission was for reasons related to meal period waivers, not meal timing. The IWC did not intend in Wage Order No. 5 to depart from the timing requirements contained in other wage orders or section 512.
The IWC had originally modified the meal waiver requirements in wage orders Nos. 4 and 5 in 1993, in response to a health care industry petition to permit its employees to waive a second meal period on longer shifts in order to leave earlier. (See IWC petn. 93-1 (Jan. 25, 1993) pp. 31-32; IWC wage order No. 4-89, subd. 11(C) (as amended Aug. 21, 1993); IWC wage order No. 5-89, subd. 11(C) (as amended Aug. 21, 1993); IWC, Statement as to the Basis of amends, to §§ 2, 3 & 11 of IWC wage order No. 5-89 (June 29, 1993).) The IWC later extended similar waiver rights to all employees covered by these wage orders and three others, but that extension was among many wage order changes repealed by the Legislature in 1999. (IWC, Statement as to the Basis, overtime and related issues (Apr. 11, 1997) pp. 7-8; Stats. 1999, ch. 134, § 21, p. 1829.)
Thereafter, health care representatives persuaded the IWC to at least preserve expanded waiver rights for their industry, along the lines of those originally afforded in 1993. (See IWC, Statement as to the Basis (Jan. 1, 2001) pp. 19-20.) Accordingly, wage orders Nos. 4-2001 and 5-2001 each contain a provision absent from other wage orders, permitting health care employees to waive one of two meal periods on longer shifts. (IWC wage order No. 4-2001 (Jan. 1, 2001) (Cal. Code Regs., tit. 8, § 11040, subd. 11(D)); Wage Order No. 5, subd. 11(D).)25 Notably, the waiver provisions permit meal waivers even on shifts in excess of 12 hours and thus conflict with language in the standard subdivision regulating second meal periods in other wage orders that limits second meal waivers to shifts of 12 hours or less (see, e.g., IWC wage order No. 2-2001 (Jan. 1, 2001) (Cal. Code Regs., tit. 8, § 11020, subd. 11(B))). For this reason, the IWC elected to omit that standard subdivision from these two wage orders. (See IWC, Statement as to the Basis (Jan. 1, 2001) pp. 19-20.) Because the omission related to *1048waiver and was not the product of any intent to include different meal timing requirements in Wage Order No. 5, we interpret that order as imposing the same timing requirements as those in most of the IWC’s other wage orders and in section 512.26
Hohnbaum contends he does not seek to require earlier second meal periods than provided for by section 512 (and, as we have determined, by Wage Order No. 5); rather, he seeks only to interpret Wage Order No. 5, subdivision 11(A) as requiring that first meal periods be timed to prevent work periods, before or after, exceeding five hours. While we agree that the period before a first meal is limited to five hours (see § 512, subd. (a)), we cannot agree that the current version of Wage Order No. 5 limits to five hours the amount of work after a meal.
First, such a reading of subdivision 11(A) in the IWC’s current wage orders would render the subdivision 11(B) guarantee of a second meal period after 10 hours of work, included in most of those same orders, superfluous. (See, e.g., IWC wage order No. 2-2001 (Jan. 1, 2001) (Cal. Code Regs., tit. 8, § 11020, subd. 11(A), (B)).) We avoid such constructions whenever possible. (Department of Alcoholic Beverage Control v. Alcoholic Beverage Control Appeals Bd. (2006) 40 Cal.4th 1, 14 [50 Cal.Rptr.3d 585, 145 P.3d 462].)27
Second, Hohnbaum’s argument rests on the contention that as used by the IWC, “ ‘work period’ is a term of art meaning ‘a continuing period of hours worked,’ ” and thus the five-hour “work period” limit in subdivision 11(A) of the wage orders must preclude more than five hours of continuous work after a meal period. “Work period” is not defined in any wage order. If the IWC’s wage orders once informally adhered to Hohnbaum’s usage, its 2001 orders no longer do. Subdivision 11(B) in most of the current orders refers to a “work period of more than ten (10) hours per day” before a second meal period. (E.g., IWC wage order No. 1-2001 (Jan. 1, 2001) (Cal. *1049Code Regs., tit. 8, § 11010, subd. 11(B)).) Any such work period must have been broken by a first meal period and thus is not “ ‘a continuing period of hours worked.’ ”
Third, there is no evidence the IWC intended to supplement the requirements of section 512 in the fashion Hohnbaum suggests. The implication is to the contrary. Having received a legislative rebuke, the IWC sought to make its orders track Assembly Bill No. 60 (1999-2000 Reg. Sess.) as closely as possible and expressed hesitance about departing from statutory requirements. (See, e.g., IWC public hearing transcript (May 5, 2000) pp. 52-56.) What departures it made appear to have been conscious choices, expressly identified in the IWC’s Statement as to the Basis, and frequently justified by explicit reliance on its authority to augment the Labor Code. (See IWC, Statement as to the Basis (Jan. 1, 2001) pp. 19-20.) In contrast, the prospect of preserving any meal timing requirement previously implicit in Wage Order No. 5, beyond the requirements of section 512, was never discussed in the agency’s 2000 hearings nor in its publications describing and explaining its 2001 wage orders. In the absence of any such discussion, we conclude the IWC did not intend to preserve Hohnbaum’s posited requirement.28
Accordingly, we conclude that Wage Order No. 5 imposes no meal timing requirements beyond those in section 512. Under the wage order, as under the statute, an employer’s obligation is to provide a first meal period after no more than five hours of work and a second meal period after no more than 10 hours of work.
B. Certification of a Meal Period Subclass
We return to the question of certification. The proposed meal period subclass includes all “Class Members who worked one or more work periods in excess of five (5) consecutive hours, without receiving a thirty (30) minute *1050meal period during which the Class Member was relieved of all duties, from and after October 1, 2000.” The trial court accepted the subclass without modification and concluded: “Although a determination that defendant need not force employees to take breaks may require some individualized discovery, the common alleged issues of meal and rest violations predominate.” Thus, it reasoned that even if Brinker were correct about the nature of its duties, to treat the case as a class action would still be the better course.
One aspect of the class definition is notable: It sweeps in not only every Brinker employee who might have a claim under Hohnbaum’s failure to provide meal periods theory, but also every employee who might have had a claim under the theory that a meal period must be provided every five hours. Consequently, because we have concluded neither Wage Order No. 5 nor section 512 imposes such a timing requirement, the class definition as presently drawn includes individuals with no possible claim.
That aspect of the class definition is notable for a second reason. In an unusual action requested by the parties, the trial court before deciding certification issued an explicit ruling on Hohnbaum’s meal timing theory, agreeing with Hohnbaum that section 512 required a meal period every five hours. That the meal subclass definition thereafter incorporated Hohnbaum’s timing theory thus raises the specter that the certification may have been influenced, in part, by the trial court’s legal assumption about the theory’s merits. Any such assumption would have been incorrect, given our ruling on the actual requirements of Wage Order No. 5 and section 512. (See ante, at pp. 1041-1049.) A grant or denial of class certification that rests in part on an erroneous legal assumption is error; without regard to whether such a certification might on other grounds be proper, it cannot stand. (Linder v. Thrifty Oil Co., supra, 23 Cal.4th at p. 436 [“[A]n order based upon improper criteria or incorrect assumptions calls for reversal ‘ “even though there may be substantial evidence to support the court’s order.” ’ ”].)
Under the unique circumstances of this case, however, we need not decide whether or not the trial court erred. Our subsequent ruling on Hohnbaum’s meal timing theory, solicited by the parties, has changed the legal landscape; whether the trial court may have soundly exercised its discretion before that ruling is no longer relevant. At a minimum, our ruling has rendered the class definition adopted by the trial court overinclusive: The definition on its face embraces individuals who now have no claim against Brinker. In light of our substantive rulings, we consider it the prudent course to remand the question *1051of meal subclass certification to the trial court for reconsideration in light of the clarification of the law we have provided.
VI. Off-the-clock Claims Class Certification
The third disputed subclass covers “Class Members who worked ‘off-the-clock’ or without pay from and after August 16, 2000.” As with the rest period subclass, we consider only whether substantial evidence supports the trial court’s conclusion that common questions predominate. None does.
Hohnbaum’s off-the-clock claims are an offshoot of his meal period claims. He contends Brinker required employees to perform work while clocked out during their meal periods; they were neither relieved of all duty nor afforded an uninterrupted 30 minutes, and were not compensated. Hohnbaum further contends Brinker altered meal break records to conceal time worked during these periods.
Unlike for the rest period claim and subclass, for this claim neither a common policy nor a common method of proof is apparent. The rest period claim involved a uniform Brinker policy allegedly in conflict with the legal requirements of the Labor Code and the governing wage order. The only formal Brinker off-the-clock policy submitted disavows such work, consistent with state law.29 Nor has Hohnbaum presented substantial evidence of a systematic company policy to pressure or require employees to work off-the-clock, a distinction that differentiates this case from those he relies upon in which off-the-clock classes have been certified. (See, e.g., Salvas v. Wal-Mart Stores, Inc. (2008) 452 Mass. 337 [893 N.E.2d 1187,- 1210-1211]; Hale v. Wal-Mart Stores, Inc. (Mo.Ct.App. 2007) 231 S.W.3d 215, 220, 225-228; Iliadis v. Wal-Mart Stores, Inc. (2007) 191 NJ. 88 [922 A.2d 710, 715-716, 723-724].)
Moreover, that employees are clocked out creates a presumption they are doing no work, a presumption Hohnbaum and the putative class members have the burden to rebut. As all parties agree, liability is contingent on proof Brinker knew or should have known off-the-clock work was occurring. (Morillion v. Royal Packing Co., supra, 22 Cal.4th at p. 585; see, e.g., White v. *1052Starbucks Corp. (N.D.Cal. 2007) 497 F.Supp.2d 1080, 1083-1085 [granting the defense summary judgment on an off-the-clock claim in the absence of proof the employer knew or should have known of the employee’s work].) Nothing before the trial court demonstrated how this could be shown through common proof, in the absence of evidence of a uniform policy or practice. Instead, the trial court was presented with anecdotal evidence of a handful of individual instances in which employees worked off-the-clock, with or without knowledge or awareness by Drinker supervisors. On a record such as this, where no substantial evidence points to a uniform, companywide policy, proof of off-the-clock liability would have had to continue in an employee-by-employee fashion, demonstrating who worked off-the-clock, how long they worked, and whether Drinker knew or should have known of their work. Accordingly, the Court of Appeal properly vacated certification of this subclass.
Disposition
For the foregoing reasons, we affirm the Court of Appeal’s judgment as to the off-the-clock subclass. We reverse its judgment as to the rest period subclass. Finally, as to the meal period subclass, we reverse the Court of Appeal’s judgment insofar as it directed the trial court to enter denial of certification with prejudice. We remand to the Court of Appeal with directions to, in turn, remand to the trial court for it to reconsider meal period subclass certification in light of the clarification of the law we have provided.
Cantil-Sakauye, C. J., Kennard, J., Daxter, J., Chin, J., Corrigan, J., and Liu, J., concurred.
The IWC issues wage orders on an industry-by-industry basis. (Martinez v. Combs (2010) 49 Cal.4th 35, 57 [109 Cal.Rptr.3d 514, 231 P.3d 259].) Wage Order No. 5 governs restaurant employees, inter alia, while other wage orders impose similar meal and rest period requirements for all other nonexempt employees in California. (See generally Cal. Code Regs., tit. 8, §§ 11010-11170.)
All further statutory references are to the Labor Code unless otherwise specified.
This claim is not expressly set forth in the complaint, but the trial court approved a stipulated amendment deeming the complaint to include allegations that employees worked off-the-clock during meal periods and Brinker engaged in time shaving.
The putative class is estimated to include just under 60,000 Brinker employees.
As one commentator has put it, “what really matters to class certification” is “not similarity at some unspecified level of generality but, rather, dissimilarity that has the capacity to undercut the prospects for joint resolution of class members’ claims through a unified proceeding.” (Nagareda, Class Certification in the Age of Aggregate Proof (2009) 84 N.Y.U. L.Rev. 97, 131.)
See also, e.g., In re Initial Public Offering Securities Litigation (2d Cir. 2006) 471 F.3d 24, 41 (endorsing similar principles under Fed. Rules Civ.Proc., rule 23, 28 U.S.C.); Gariety v. Grant Thornton, LLP (4th Cir. 2004) 368 F.3d 356, 365-367 (same); Nagareda, Class Certification in the Age of Aggregate Proof, supra, 84 N.Y.U. L.Rev. at page 114 (under federal law, a “court must inquire into the merits ... if that inquiry pertains to the satisfaction of a [Federal Rules of Civil Procedure,] Rule 23 requirement,” but “oversteps its proper bounds if it conducts” a “free-floating merits inquiry .. . untethered to a Rule 23 requirement”).
By its terms, Wage Order No. 5 expressly covers restaurant employees such as Hohnbaum and the proposed class members. (See Wage Order No. 5, subd. 2(P)(1).)
We observe that because the IWC’s funding was restricted in 2004 (see Murphy v. Kenneth Cole Productions, Inc., supra, 40 Cal.4th at p. 1102, fn. 4), the agency has been rendered essentially unavailable in recent years to cast light on the intended significance of its wage orders.
See, e.g., Department of Commerce v. Montana (1992) 503 U.S. 442,450-451 [118 L.Ed.2d 87, 112 S.Ct. 1415]; Hovet v. Myers (1971) 260 Or. 152 [489 P.2d 684, 685-686]; 7 Code of Federal Regulations part 1205.322(b)(1) (2012); Balinski and Young, The Quota Method of Apportionment, 82 American Mathematical Monthly 701, 704 (Aug.-Sept. 1975); Oxford English Dictionary Online (3d ed. 2000; online version Mar. 2012) <http://oed.com/view/ Entry/112621> (as of Apr. 12, 2012) (“major” is that which “constitutes the majority or larger part; usually with part, portion, or other similar nouns”).
IWC wage order No. 5 R, subdivision 11 (June 1, 1947); IWC wage order No. 5-52, subdivision 12 (Aug. 1, 1952).
“The DLSE ‘is the state agency empowered to enforce California’s labor laws, including IWC wage orders.’ ” (Morillion v. Royal Packing Co. (2000) 22 Cal.4th 575, 581 [94 Cal.Rptr.2d 3, 995 P.2d 139].) The DLSE’s opinion letters, “ ‘ “ ‘while not controlling upon the courts by reason of their authority, do constitute a body of experience and informed judgment to which courts and litigants may properly resort for guidance.’ ” ’ ” (Seymore v. Metson Marine, Inc. (2011) 194 Cal.App.4th 361, 369, fn. 5 [128 Cal.Rptr.3d 13]; see Morillion, at p. 584 [relying on DLSE opinion letters to inform its interpretation of the IWC’s wage orders].)
See generally Murphy v. Kenneth Cole Productions, Inc., supra, 40 Cal.4th at page 1113 (discussing the health and safety considerations behind meal and rest periods).
Other evidence the trial court was entitled to credit suggested employees may not have been permitted even that much rest break time.
We declared IWC wage order No. 5-76 invalid for failure to include an adequate Statement as to the Basis (see California Hotel & Motel Assn. v. Industrial Welfare Com. (1979) 25 Cal.3d 200, 216 [157 Cal.Rptr. 840, 599 P.2d 31]; § 1177, subd. (b)), but the IWC thereafter cured the omission and reissued identical meal period language (see IWC, Statement as to the Basis for Wage Order No. 5-80 (Sept. 7, 1979); IWC wage order No. 5-80, subd. (11)(A) (Jan. 1, 1980)).
See also Bono Enterprises, Inc. v. Bradshaw, supra, 32 Cal.App.4th at page 975 (emphasizing absence of duty and freedom from employer control as central to unpaid meal periods); Wage Order No. 5, subdivision 2(K) (defining “ ‘Hours worked’ ” as including all time when subject to employer control).
As enacted in 1999, the text of section 512 consisted of what is now section 512, subdivision (a). (Stats. 1999, ch. 134, § 6, p. 1823.) The original text was recodified without change in 2000 when a new subdivision was added. (Stats. 2000, ch. 492, § 1, p. 3500.)
See, e.g., IWC wage order No. 2, subdivision 1(20) (Apr. 14, 1916) (“no woman or minor shall be permitted to return to work in less than one-half hour’’); IWC wage order No. 3a, subdivision 12 (lune 4, 1928) (same); IWC wage order No. 18, subdivision 10 (Feb. 26, 1932) (same); id., subdivision 11, footnote * (“It is recommended that. . . without exception where [lunch room] space is provided, all women and minors shall be required during the meal period to leave and remain out of the room in which they are regularly employed.”); IWC wage order No. 9 Amended, subdivision 9(a) (Aug. 28, 1933) (“The employer is responsible for seeing that [the meal period] time is taken.”).
Wage Order No. 5’s definition of “employ” was first inserted in 1943. (TWC wage order No. 5 NS, subd. 2(c) (June 28, 1943).) At the time, the IWC explained that “in order to overcome the possibility of subterfuge being resorted to by unscrupulous employers under which subterfuge such employers claim that they did not hire the employee, it becomes necessary to find that ‘employment’ means ‘suffering or permitting’ a woman or minor to perform services . . . .” (IWC meeting mins. (Apr. 14, 1943) p. 4.) The adopted definition was consistent with how the IWC had used the term from its earliest wage orders forward. (See Martinez v. Combs, supra, 49 Cal.4th at pp. 57-58.)
If work does continue, the employer will not be liable for premium pay. At most, it will be liable for straight pay, and then only when it “knew or reasonably should have known that the worker was working through the authorized meal period.” (Dept. Industrial Relations, DLSE Opn. Letter No. 1991.06.03, supra, p. 1; see Morillion v. Royal Packing Co., supra, 22 Cal.4th at p. 585.) The DLSE correctly explains the distinction in its amicus curiae brief: “The employer that refuses to relinquish control over employees during an owed meal period violates the duty to provide the meal period and owes compensation [and premium pay] for hours worked. The employer that relinquishes control but nonetheless knows or has reason to know that the employee is performing work during the meal period, has not violated its meal period obligations [and owes no premium pay], but nonetheless owes regular compensation to its employees for time worked.” (See also Wage Order No. 5, subd. 2(K) [defining “ ‘[h]ours worked,’ ” for which compensation is owed, to include “all the time the employee is suffered or permitted to work, whether or not required to do so”].)
The IWC has permitted first meal periods after six hours of work for all employees under wage order No. 12 and those under wage order No. 1 who collectively bargain for such a variance. (IWC wage order No. 12-2001 (Jan. 1, 2001); IWC wage order No. 1-2001, as amended July 1, 2002; Cal. Code Regs., tit. 8, §§ 11120, subd. 11(A), 11010, subd. 11(A)). It has made no similar allowance in the wage order applicable here, Wage Order No. 5.
A third interpretation, that the provision guarantees only the rate at which meal periods are owed and not the timing of them—one meal period for each five full hours worked, to be provided as early or as late as the employer chooses—is advocated by neither party. As our examination of the IWC’s records, post, makes clear, it is also not an interpretation the IWC ever contemplated.
This was consistent with historical practice, as the IWC had long regulated both the number and timing of meals. (See, e.g., IWC wage order No. 2, subd. 1(20) (Apr. 14, 1916) [guaranteeing a “noon day meal”]; IWC wage order No. 4, subd. 1(21) (June 15, 1917) [adding a provision for an “evening meal” on shifts extending into the night]; IWC wage order No. 3a, subd. 12 (June 4, 1928) [providing for an evening meal if work continued beyond 7:30 p.m.].)
In 1963, the IWC clarified that waiver of a meal period required mutual consent, and in 1976 it expanded coverage to men. (See IWC wage order No. 5-63, subd. 11(a) (Aug. 30, 1963); IWC wage order No. 5-76, Subd. 11(A) (Oct. 18, 1976).)
Subdivision 9 provided: “9. MEAL PERIODS, [ft] (A) No employer shall employ any person for a work period of more than five (5) hours without a meal period of not less than thirty (30) minutes, except that when a work period of not more than six (6) hours will complete the day’s work the meal period may be waived by mutual consent of employer and employee, [ft] (B) An employer may not employ an employee for a work period of more than ten (10) hours per day without providing the employee with a second meal period of not less than thirty (30) minutes, except that if the total hours worked is no more than twelve (12) *1046hours, the second meal period may be waived by mutual consent of the employer and the employee only if the first meal period was not waived.” (IWC interim wage order—2000 (Mar. 1, 2000) subd. 9.)
Hohnbaum argues that because the right to waive a second meal period was extended to employees working in excess of eight hours, the IWC must have contemplated employees accruing a right to a second meal before 10 hours of work. The historical record does not support this inference. The language was originally prepared by industry officials, not the IWC. (IWC, Statement as to the Basis (Jan. 1, 2001) p. 20.) In public discussion of the second meal waiver provision, neither in 1993 nor in 2000 did the issue arise of a second meal period being owed five hours after a first meal period. Instead, discussion focused on whether health care employees working 12-hour shifts, as opposed to eight-hour shifts, should be allowed to waive a second meal period. (See, e.g., IWC public hearing transcript (Apr. 7, 1993); IWC public hearing transcript (Jan. 28, 2000); IWC public hearing transcript (Apr. 14, 2000).)
IWC wage order No. 12-2001 stands in contrast. There, the IWC, in response to film industry concerns that section 512 conflicted with industry collective bargaining agreements setting meal breaks at intervals of six hours, preserved language setting first and subsequent meals at six-hour intervals, rather than after five hours of work and 10 hours of work. (See IWC wage order No. 12-2001 (Jan. 1, 2001) (Cal. Code Regs., tit. 8, § 11120, subd. 11(A)); IWC public hearing transcript (May 5, 2000) pp. 44-50.) Had the IWC intended in Wage Order No. 5 to preserve timing requirements additional to those provided by section 512, it could similarly have varied that wage order’s language. It did not.
While Wage Order No. 5 does not include a second meal period guarantee in its subdivision 11(B), that omission is not, as just discussed, the result of any different intent with respect to second meal periods under the wage order. Most other wage orders include both a subdivision 11(A) mirroring Wage Order No. 5’s subdivision 11(A) and a subdivision 11(B) guaranteeing a second meal period after 10 hours, and these provisions must be read to avoid surplusage.
In the ordinary case, proof that statutory or regulatory language was adopted with a particular intent, and that the adopting body did not thereafter express a contrary intent, may suffice to show that the language should be interpreted today in line with that original intent. Hohnbaum develops such an argument, trying to establish the original intent behind Wage Order No. 5, subdivision 11(A)’s language and afford a basis to presume the same intent must hold today. But the inference Hohnbaum relies on—that this language should be read today to mean what it meant when adopted—collapses in the presence of evidence that the IWC in 2000 selected the current language to replicate the rules enacted by the Legislature in 1999, while preserving an expanded right to waive second meal periods for certain workers, and in the absence of evidence that it selected the language to also preserve a limitation on the duration of work after first meal periods.
Brinker’s “Hourly Employee Handbook” states in part: “It is your responsibility to clock in and clock out for every shift you work. . . . [Y]ou may not begin working until you have clocked in. Working ‘off the clock’ for any reason is considered a violation of Company policy. If you forget to clock in or out, or if you believe your time records are not recorded accurately, you must notify a Manager immediately, so the time can be accurately recorded for payroll purposes.”