More than a half century ago, Justice Felix Frankfurter observed that “[t]he one stark fact that emerges from a study of the history of [legislative] apportionment is its embroilment in politics, in the sense of party contests and party interests.” (Colegrove v. Green (1946) 328 U.S. 549, 554 [90 L.Ed. 1432, 66 S.Ct. 1198] (plur. opn. of Frankfurter, J.).) Faced with entreaties by litigants seeking judicial intervention in the redistricting process, Justice Frankfurter famously warned that “[c]ourts ought not to enter this political thicket.” (Id. at p. 556.) Although the law has not adopted the uncompromising version of this principle urged by Justice Frankfurter (see, e.g., Reynolds v. Sims (1964) 377 U.S. 533 [12 L.Ed.2d 506, 84 S.Ct. 1362]; Baker v. Carr (1962) 369 U.S. 186 [7 L.Ed.2d 663, 82 S.Ct. 691]), his admonition continues to resonate each decade when courts are asked to decide what are fundamentally political disputes. Judicial restraint is especially important in the context of legislative redistricting because, as the high court recently observed, “experience has shown the difficulty of defining neutral legal principles in this area.” (Perry v. Perez (2012) 565 U.S._,_ [181 L.Ed.2d 900, 905, 132 S.Ct. 934, 941].)
In this case, I agree with the court’s bottom line: the Senate district map certified by the Citizens Redistricting Commission (Commission) is the interim map that should be used in the event that petitioner’s referendum qualifies for the ballot. As the court explains (maj. opn., ante, at pp. 464-485), we need not exclude the Commission map from consideration as a possible interim map, and the Commission map is superior to the proposed alternatives when evaluated against applicable federal and state redistricting criteria. I write separately, however, because I believe the court’s discussion of our authority to decide cases such as this leaves too much to “prudence” (see maj. opn., ante, at pp. 451, 457, 458, 459, 460) and places insufficient emphasis on language in the California Constitution that channels and checks our discretion.
*487Today’s opinion concludes that “under California Constitution, article VI, section 10, this cotut is authorized to issue an order to show cause and decide which districts should be used in the event a proposed referendum directed at a Commission-certified redistricting map qualifies for the ballot, even in the absence of a showing that the proposed referendum is likely to qualify for the ballot.” (Maj. opn., ante, at p. 463.) The court will find a mandate action ripe for decision “when we conclude that, in light of the relative probability that the proposed referendum will qualify for the ballot and the time limitations and potential detrimental consequences of refusing to consider a mandate petition at that point in time, it is prudent to issue an order to show cause and decide which districts should be used in the upcoming elections in the event the proposed referendum does qualify for placement on the ballot.” (Id. at p. 457.)
I believe these statements, by maximizing our discretion, will have the unintended consequence of inviting future litigants to bring their grievances with the redistricting process to this court. Of course, the court will exercise prudence in addressing those litigants’ claims. However, as this court’s own experience shows, redistricting controversies are fraught with political peril. Where one judge sees prudence, another may see partisanship. I respectfully disagree with the court’s unduly broad assertion of authority because it underestimates the risks of political entanglement that are inherent to redistricting disputes and because the all-things-considered discretion it contemplates gives insufficient weight to language in our Constitution that can help the court avoid such entanglement.
I.
As an initial matter, I agree with the court that we have jurisdiction under California Constitution, article VI, section 10 to entertain a petition for writ of mandate and to issue an order to show cause in this matter. Our jurisdiction, in the fundamental sense of lawful authority to hear the case, is not in question.
The question we face is under what circumstances this court should decide the merits of a petition seeking relief in the form of an interim map to be used in the event that a referendum challenging a Commission map qualifies for the ballot. Uncertainty as to whether a proposed referendum will qualify for the ballot presents a timing problem with two competing dimensions. On one hand, the court risks acting too late if it waits for the signature verification process to indicate whether the referendum will qualify. “[I]f this court were to conclude that other districts should be used in the event the referendum qualifies, it may be too late at that time to apply the court’s decision to implement those other districts.” (Maj. opn., ante, at p. 456.) On *488the other hand, the court risks acting too early if it decides the merits of the petition and selects an interim map before knowing whether the referendum will qualify. “[T]he issuance of a court decision favoring an alternative map over the Commission-certified map might—in the event the referendum does not qualify—unnecessarily cast a cloud over the legitimacy of that Commission-certified map for the ensuing decade . . . .” (Id. at p. 459.)
The risk of acting too early is not present where, as here, the court concludes that the Commission map should be the interim map if petitioner’s referendum qualifies for the ballot. “Even if, after the court issues its opinion, the referendum ultimately does not qualify for the ballot and the Commission-certified map is not stayed, issuance of the court’s decision— approving the Commission’s map—could have no adverse effect upon the Commission-certified map.” (Maj. opn., ante, at p. 459.) Under such circumstances, there is no need to assess the referendum’s likelihood of qualification, and the court “should issue its decision as early as possible so as to eliminate the uncertainty that inevitably arises from the ongoing signature verification process and the pendency of the writ proceeding in this court.” (Ibid.)
Today’s decision could have resolved the timing issue on that basis and left matters there. However, in an effort “to provide guidance on this procedural point for the future” (maj. opn., ante, at p. 448), the court goes further to broadly hold that we have discretion and maximum flexibility to determine whether and when we may decide which map should be used upon qualification of a referendum, even where the interim map we choose is not the Commission map (see id. at pp. 456-457, 459^-60). In that scenario, where the court concludes that an alternative map is superior to the Commission map, the problematic consequences of acting too early or too late come to the fore. This case does not present that scenario, but it is clear that the court’s broad holding applies to that scenario and is expressly intended to address it. (See id. at p. 456.) Accordingly, the discussion from here forward addresses the court’s understanding of our authority in circumstances where we find or contemplate finding that an alternative to the Commission map should be the interim map.
In my view, the court’s articulation of its authority raises two concerns. First, the court refers to the harm from acting too early as a “possibility” that “might” come to pass. (Maj. opn., ante, at p. 460.) I see no reason for such tentative acknowledgment of the substantial harm at issue. If the court issues a decision favoring an alternative to the Commission map but the referendum does not end up qualifying, the court’s decision would have needlessly burdened election officials with using a dual-track planning process until completion of the signature verification process. More importantly, serious *489concerns about the fairness and legitimacy of the electoral process would arise throughout the next decade if a Commission map were to go into effect in the shadow of an opinion by this court favoring another map even if only on an interim basis. These problems are hardly speculative. The consequences of acting too early are at least as serious and foreseeable as the consequences of acting too late, yet the court’s opinion seems far less concerned with the former than with the latter. (See id. at pp. 456-458.)
This imbalance contributes to the second concern, and that is the expansive discretion the court claims for itself to decide whether and when to act on a mandate petition in cases where it finds that an alternative map is superior to the Commission’s. The court says it may act “when we conclude that, in light of the relative probability that the proposed referendum will qualify for the ballot and the time limitations and potential detrimental consequences of refusing to consider a mandate petition at that point in time, it is prudent” to act. (Maj. opn., ante, at p. 457.) Elsewhere the court says it “may properly grant relief . . . when the court is of the view that there is a sufficient chance that the proposed referendum will qualify to make it prudent for the court” to do so. (Id. at pp. 457-458.) And “[i]n determining whether it is prudent to issue its decision in advance of the proposed referendum’s qualification notwithstanding [the] potential adverse consequence [of acting too early] (and if so, how far in advance), this court would undoubtedly take into account the particular reasons underlying its analysis and determination that an alternative map is more consistent with the constitutionally based criteria than the Commission-certified map, and then decide if and when to issue its opinion based in part on such considerations.” (Id. at pp. 459-460.)
These verbal formulations all boil down to the same thing: the court will act when the court feels it is prudent to do so. (See maj. opn., ante, at p. 460 [“[T]his court properly retains broad discretion to take into account all such considerations as well as any other relevant factor in deciding what relief is appropriate in such a proceeding and when it should be ordered.”].) It casts no doubt on the court’s collective wisdom to worry that such open-ended discretion offers little in the way of an objective standard for determining—in the face of uncertainty and the serious risks it entails—whether and when to issue a decision as important as declaring that a map other than the Commission’s should be used as an interim map. In effect, the court’s opinion invites future litigants to make their best case for or against judicial intervention without supplying a determinate legal principle that can help keep our decisionmaking, in appearance and reality, above the political fray. We need not place such a heavy burden on the dictates of our own prudence, for the text of our Constitution provides a better approach.
*490II.
The voters of California fundamentally reformed the redistricting process when they passed Proposition 11 in 2008 and Proposition 20 in 2010. Those initiatives created the Commission, defined its membership, procedures, and responsibilities, established a prioritized list of redistricting criteria, and provided that Commission-certified maps may be challenged by referendum. In addition, the recent reform contemplates this court’s intervention in the redistricting process in four circumstances. First, a registered voter may challenge the lawfulness of a Commission map within 45 days after it is certified. (Cal. Const., art. XXI, § 3, subd. (b)(2), (3).) Second, when the Commission does not approve a final map by the requisite votes, it is this court’s duty to supply a map with the aid of a special master. (Id., § 2, subd. (j).) Third, when voters disapprove a Commission map in a referendum, the court also must supply a map with the aid of a special master. (Ibid.)
The fourth circumstance is the one relevant here: “Any registered voter in this state may also file a petition for a writ of mandate or writ of prohibition to seek relief where a certified final map is subject to a referendum measure that is likely to qualify and stay the timely implementation of the map.” (Cal. Const., art. XXI, § 3, subd. (b)(2) (hereafter section 3(b)(2)).) This sentence of section 3(b)(2) addresses situations where a referendum challenging a Commission map has not yet qualified for the ballot, and it provides guidance on the timing problem we face.
Section 3(b)(2) was enacted as part of Proposition 20 along with another provision that moved up the date by which the Commission must release its final maps from September 15 to August 15. (Cal. Const., art. XXI, § 2, subd. (g).) The evident purpose of these provisions was to give this court flexibility to act in situations where a referendum is likely to qualify but where the signature verification and official qualification process has not reached completion. Further, article XXI, section 3, subdivision (b)(3) of the California Constitution (hereafter section 3(b)(3)) says: “The California Supreme Court shall give priority to ruling on a petition for a writ of mandate or a writ of prohibition filed pursuant to [section 3(b)(2)].”
At a minimum, section 3(b)(2) means that a petition for relief in the form of an interim map is properly before the court and ripe for adjudication on the merits where the petitioner has shown that a referendum challenging a Commission map is likely to qualify. Section 3(b)(3) makes clear that the court must act expeditiously to decide the merits of such a petition. The import of these constitutional provisions is that when a petitioner has shown that a referendum is likely to qualify, the probability of qualification is sufficiently high that this court must promptly provide guidance to election *491officials on the interim map that will be used in the event that the referendum ultimately does qualify. Although a referendum deemed likely to qualify may still end up not qualifying, sections 3(b)(2) and 3(b)(3) imply that the risk is sufficiently low that the court should decide the merits of the petition.
While acknowledging that a petition is “sufficiently ripe ... to be entertained” when a referendum is likely to qualify (maj. opn., ante, at p. 462), today’s opinion stops short of saying that the court should promptly decide the merits of such a petition. Instead, , the court says that “the ‘likely to qualify’ language of article XXI, section 3(b)(2) ... is not directed to the time when the court may or should decide the merits of the mandate petition, but rather to the time when a registered voter may file such a petition in this court.” (Id. at pp. 462-463, fn. 25.) Even if this reading of section 3(b)(2) were correct (and I am doubtful that it is (see post, at pp. 491-493)), section 3(b)(3) leaves no ambiguity as to what this court is supposed to do. Where a petitioner has shown that a referendum is likely to qualify, the “ruling” contemplated by section 3(b)(3)’s mandate that this court “shall give priority to ruling on a petition . . . filed pursuant to [section 3(b)(2)]” can only be understood as a ruling on the merits.
Sections 3(b)(2) and 3(b)(3) thus answer part of the timing question we face: when a referendum is likely to qualify, the court must act. But what should the court do when available information does not support a finding that a referendum is likely to qualify? If a petitioner cannot show that a referendum is likely to qualify, what significance does that circumstance have for whether and when the court should act? I agree with today’s opinion that because section 3(b)(2) “does not purport to limit this court’s jurisdiction” (maj. opn., ante, at p. 461), a petitioner’s inability to show that a referendum is likely to qualify does not deprive the court of its prerogative to decide whether and when to act. But I do not agree that failure to show that a referendum is likely to qualify has no particular significance to the exercise of our authority, beyond being a factor that it may be “prudent” to consider. (Id. at p. 460.)
Once again, section 3(b)(2) says: “Any registered voter in this state may also file a petition for a writ of mandate or writ of prohibition to seek relief where a certified final map is subject to a referendum measure that is likely to qualify and stay the timely implementation of the map.” Textually, the phrase “where a certified final map is subject to a referendum measure that is likely to qualify” is sensibly read as a condition precedent to the availability of “relief’ under section 3(b)(2). Just as the “likely to qualify” standard serves to mitigate the risk of acting too late, it also serves to mitigate the risk of acting too early. Where the probability of qualification is sufficiently low that a referendum cannot be deemed likely to qualify, the court generally should *492stay its hand because of the sufficiently high risk that issuing a decision on the merits will prove unnecessary and injurious to the electoral process. Section 3(b)(2) addresses the uncertainty where a referendum has not yet qualified by balancing the public interest in giving the referendum’s supporters a timely opportunity to seek relief in the event that the referendum does qualify against the public interest in having a Commission-certified map implemented without disruption in the event that the referendum does not qualify. In other words, the “likely to qualify” standard strikes the very balance that the court would relegate to its sense of prudence. Because section 3(b)(2) does not limit this court’s jurisdiction over original writ proceedings, it is always possible that exceptional circumstances may compel the court to act even where a petitioner has not shown that a referendum is likely to qualify. However, I would adopt as a general rule—indeed, a presumption— that where a petitioner has not shown that a referendum is likely to qualify, the court should not decide the merits of the mandate petition.
The court contends that the “likely to qualify” language in section 3(b)(2) has no bearing on when judicial action is warranted and, as a textual matter, speaks “to the time when a registered voter may file” a petition for writ of mandate or writ of prohibition. (Maj. opn., ante, at pp. 462-463, fn. 25.) But it makes little sense to read section 3(b)(2) as a timing provision that pertains to filing in light of the court’s conclusion that under article VI, section 10 of the California Constitution, a petition for writ of mandate or prohibition may be filed whether or not the petitioner can show that a referendum is likely to qualify. (See maj. opn., ante, at pp. 450-451 & fn. 18, 451.) Instead of indicating when a registered voter may file a petition, the “likely to qualify” language in section 3(b)(2) is better read as specifying when relief is available and may be granted by this court before a referendum has qualified for the ballot.
More fundamentally, the court objects to this reading of section 3(b)(2) on the ground that it would limit our authority to entertain a mandate petition, even just to issue an order to show cause, unless a petitioner has shown that a referendum is likely to qualify. (Maj. opn., ante, at p. 462, fn, 25.) But my interpretation would not have “this type of limiting effect on this court’s authority.” (Ibid.) Where, as here, an original writ petition is properly filed pursuant to article VI, section 10 of the California Constitution, the court has jurisdiction and may issue an order to show cause, entertain briefing and oral argument, and deliberate. If the court concludes that the Commission map should be the interim map, it may and should say so as early as possible.1 *493Otherwise, the court generally should not decide the merits of the petition or grant relief before it finds that the referendum is likely to qualify. The inability to make that finding at the moment the petition is filed does not mean we must dismiss the petition. The court may retain jurisdiction and await further information on the referendum’s likelihood of qualification. If additional information shows that the referendum is likely to qualify, the court will be poised to act expeditiously.2
In essence, the court’s objection to my interpretation of section 3(b)(2) confuses the issue of whether the court has authority to entertain a mandate petition with the separate and distinct issue of how that authority should be exercised. The first issue is settled by article VI, section 10 of the California Constitution. (See ante, at pp. 487-488.) Section 3(b)(2) speaks only to the second issue. In exercising its proper authority to entertain a mandate petition seeking relief in the form of an interim map, under what circumstances may the court grant relief before a referendum has qualified for the ballot? Section 3(b)(2) answers that question by stating a condition precedent to the availability of relief—namely, a showing that the referendum is likely to qualify.
The advantage of this approach is precisely what the court sees as its disadvantage; it limits our discretion. By using the phrase “likely to qualify,” section 3(b)(2) provides an objective and determinate standard for balancing the competing risks of acting too early versus acting too late. In ordinary usage, the word “likely” is commonly understood to mean “more likely than *494not.” (See, e.g., Merriam-Webster’s Collegiate Dict. (11th ed. 2003) p. 721 [defining “likely” to mean “having a high probability of occurring or being true”]; Webster’s International Dict. (3d ed. 2002) p. 1310 [defining “likely” to mean “having a better chance of existing or occurring than not”]; Garner, A Dict, of Modern Legal Usage (2d ed. 1995) p. 530 [“likely has different shades of meaning” but “[m]ost often it indicates a degree of probability greater than five on a scale of one to ten”]; Black’s Law Dict. (6th ed. 1990) p. 925 [defining “likely” to mean “probable and having better chance of existing or occurring than not”].) Although the court in People v. Superior Court (Ghilotti) (2002) 27 Cal.4th 888, 916-917 [119 Cal.Rptr.2d 1, 44 P.3d 949] said that the meaning of “likely” may depend on context, Ghilotti interpreted the phrase “ ‘likely to engage in acts of sexual violence’ ” as part of an intricate statutory scheme enacted by the Legislature to provide for civil commitment of inmates previously convicted of a sexually violent offense. (See id. at pp. 915-929.) The particularized meaning of words in complex, legislatively enacted statutes has little bearing on the interpretation of words in an initiative, which we construe according to their ordinary meanings as understood by “the average voter.” (Robert L. v. Superior Court (2003) 30 Cal.4th 894, 902 [135 Cal.Rptr.2d 30, 69 P.3d 951].)
To be sure, the “likely to qualify” standard does not fully immunize the court from the risks of acting too early or too late. A referendum shown to be likely to qualify may end up not qualifying, and a referendum not shown to be likely to qualify may end up qualifying. But the risk of error in one direction or the other is present in any approach to the timing problem. That is the nature of uncertainty. Rather than address the uncertainty through the prudential exercise of discretion, I would make use of the objective constitutional standard that already balances the competing risks.
As a practical matter, the “likely to qualify” standard does not impose a heavy burden on the petitioner or on this court. In most cases, determining whether a referendum is “likely to qualify” will be a simple matter. The random sampling that takes place when a referendum has gathered a sufficient number of signatures will usually resolve the issue in a timely fashion. As the court explains: “Under the governing statutory provisions, if the random sampling projected a number of total valid signatures that was less than 95 percent of the required number of valid signatures, the petition would fail without any further count. If the projection of valid signatures was 110 percent or more of the required number, the petition would qualify without any further count. If the projection of valid signatures was between 95 and 110 percent, the Secretary of State would notify counties that a full count of all submitted signatures would be required to verify the number of valid signatures that had been submitted. (Elec. Code, §§ 9030, subds. (f) & (g), 9031, subd. (a).)” (Maj. opn., ante, at p. 454.)
*495According to documentation attached to petitioner’s declaration, of the 48 initiatives and referenda submitted to the Secretary of State between 2005 and 2010, 44 of them were projected to receive more than 110 percent of the needed signatures. Those measures qualified for the ballot without any need to count all the signatures submitted. In such cases, even before the referendum is officially certified by the Secretary of State, this court can conclude that the referendum is likely to qualify. The random count in this case was completed by January 10, and it is likely that such counts will be completed around the same time in the future. That date is well before “the end of January 2012,” which is the latest time by which the Secretary of State and county election officials need to be informed of this court’s decision in order “to implement any changes in the state Senate districts in the event the proposed referendum qualifies and automatically stays the Commission-certified state Senate redistricting map.” (Maj. opn., ante, at p. 456.)
If random sampling yields a projection between 95 percent and 110 percent of the signatures needed and a full count is required, this court will still be able to determine whether a referendum is likely to qualify in most cases. In supplemental briefing, the Secretary of State claimed that “to use the completed sampling process to determine at what point a petition becomes likely to qualify ... is beyond the capacity of the process” and that “the sampling technique is not designed to give reliable results at a greater level of precision” than determining “whether the number of valid signatures on petitions is within a broad range, 95 [percent] to 110 [percent].” But the Secretary of State also acknowledged in the same briefing that the experience of the four initiatives within the last five years that required a full count in order to qualify for the ballot “suggests that the sampling process is reasonably accurate within a margin of about 1.5%.”3 Although the Secretary of State said she was “reluctant to draw firm conclusions from a sample that consists of only four examples,” the examples do offer some indication, however limited, that the random sampling process can predict the full count within a fairly small margin of error.
Contrary to the Secretary of State’s suggestion, the Legislature’s determination that random sampling must yield a projection of at least 110 percent of the signatures needed before a referendum will be deemed qualified does not *496preclude a petitioner from citing a projection less than 110 percent as evidence that a referendum is likely to qualify. Indeed, because a referendum actually qualifies with a projection equal to or greater than 110 percent, logic dictates that a lesser projection may support a finding that a referendum is merely likely to qualify. For example, where random sampling yields a projection of 105 percent of the total signatures needed, the referendum does not qualify on that basis and must go to a full count. But the projection would still be credible evidence, given the past relationship between random counts and full counts, that the referendum is likely to qualify.
The task of timely determining whether a referendum is likely to qualify is more difficult when random sampling yields a projection that is very close to the minimum number of signatures required. Petitioner asserts that whenever random sampling projects 100 percent or more of the needed signatures, the referendum should be deemed likely to qualify. But petitioner offers no analysis or expert declarations in support of this claim. At oral argument, petitioner asserted that more refined analysis of whether a referendum is likely to qualify is within the competence of various experts and experienced consultants. Although we need not apply the “likely to qualify” standard in this case because of our ultimate disposition (see ante, at pp. 487-488), future litigants would be well-advised to bring expert analysis to bear where the issue is a close call. Statistical certainty is not required in order to render a legal judgment applying the “likely to qualify” standard (presumably, a petitioner need only show that a referendum is “likely to qualify” by a preponderance of the evidence), but the court would benefit from expert interpretation of available information.
Finally, it is worth noting that, although my view of section 3(b)(2)’s significance for the timing of judicial intervention differs from the court’s, nothing I have said is technically inconsistent with the court’s broad holding that the issue of timing should be resolved according to the dictates of prudence. My sense of prudence, which subsumes my reading of the law, impels me to assign particular significance to whether a referendum is likely to qualify in deciding whether and when we should act on the merits of a mandate petition. But whether the prudence of my colleagues would lead each of them to the same conclusion or to different conclusions in a case where the issue really matters is an open question. And that, in a sense, illustrates the problem with the court’s approach.
*497III.
Because our disposition in this case is unanimous, the concerns I have expressed may seem speculative. But the court typically speaks on redistricting only once a decade, and today’s opinion deliberately paints with a broad brush. It is of course anyone’s guess what the future will bring. But history provides a cautionary tale.
Thirty years ago, this court had a very different experience with a redistricting controversy. After the 1980 census, the Democratic-controlled Legislature enacted and the Governor, also a Democrat, signed in September 1981 three reapportionment statutes revising the boundaries of the state’s congressional, Senate, and Assembly districts. The Republican Party initiated a referendum against each of these reapportionment statutes. By December 15, 1981, these referenda had qualified for the June 1982 ballot. Various members of the Assembly, Senate, and United States House of Representatives filed mandate proceedings claiming that defects in the referendum petitions rendered the petitions invalid. They also claimed that even if the referenda did qualify for the ballot, they should not stay implementation of the new legislative maps for the June 1982 election.
In Assembly v. Deukmejian (1982) 30 Cal.3d 638 [180 Cal.Rptr. 297, 639 P.2d 939], this court rejected the challenges to the validity of these referenda and affirmed that the referenda stayed the reapportionment statutes, as they would any other statute. (Id. at pp. 656-657.) The court then considered the appropriate remedy. The referenda proponents argued that the court should order use of the old maps in the interim, as the court had done in Legislature v. Reinecke (1972) 6 Cal.3d 595 [99 Cal.Rptr. 481, 492 P.2d 385], a case in which the new legislative maps had been vetoed by the Governor and never became law. By a four-to-three majority, the court in Assembly v. Deukmejian declined to order use of the old maps, instead concluding that the new maps should be used in the interim primarily because they were drawn to comply with the one-person, one-vote requirement of the Fourteenth Amendment’s equal protection clause. (Assembly v. Deukmejian, supra, 30 Cal.3d at pp. 665-668.)
The court’s holding in favor of the new maps prompted three separate dissenting opinions. Justice Richardson, in an opinion joined by Justice Mosk and Justice Kaus, explained that it was improper to use new maps that had been stayed by qualification of the referenda and that using the old maps until *498the referenda were voted on would not violate the equal protection clause. (Assembly v. Deukmejian, supra, 30 Cal.3d at pp. 680-685 (conc. & dis. opn. of Richardson, J.).) According to Justice Richardson, “the majority completely disregards [the] stay [of the new maps] and imposes upon the people of California a state legislative reapportionment plan which has been stopped dead in its tracks by operation of law and which is heavily veiled in a cloud of political uncertainty. The majority’s adoption of this plan prejudges the result and its action can only be perceived as an official alignment of the court with one side in a partisan dispute as to which we should remain scrupulously neutral.” (Id. at p. 680.)
In a separate opinion, Justice Mosk wrote that “a bare majority of this court have become entangled in the ‘political thicket’ by ignoring their obligation of neutrality on a partisan issue, a neutrality that can be observed only by maintenance of the status quo in legislative districting until the people speak at the forthcoming election.” (Assembly v. Deukmejian, supra, 30 Cal.3d at p. 693 (conc. & dis. opn. of Mosk, J.).) And Justice Kaus wrote separately to say that “the course chosen by the majority involves greater judicial intrusion into the legislative process laid out by the California Constitution.” (Id. at p. 694 (conc. & dis. opn. of Kaus, J.).)
I express no view on which side was correct in Assembly v. Deukmejian. But I am confident that each of the four justices who voted in favor of the new maps, as well as each of the three justices who voted in favor of the old maps, cast his or her vote on the basis of a well-informed assessment of the lawful and prudent course. (Compare Assembly v. Deukmejian, supra, 30 Cal.3d at pp. 675-676 [use of new maps would be the least disruptive remedy] with id. at p. 692 (conc. & dis. opn. of Richardson, J.) [use of old maps would be least disruptive].) Even so, the court ultimately fractured in a series of strongly worded opinions laced with charges of partisanship.
Assembly v. Deukmejian did not present the timing issue we face here. But the case confirms that “[l]ogic, as well as experience, tells us . . . that there can be no total sanctuaries in the political thicket.” (Dixon, The Court, The People, and “One Man, One Vote,” in Reapportionment in the 1970s (Polsby edit., 1971) p. 32.) Today the court unanimously agrees that the Commission map is superior to the proposed alternatives. In a future case, the court may be divided with regard to which map should serve as an interim map and, closely related, whether and when to issue a decision on that important issue. *499Those questions will inevitably play out against a backdrop of partisan interests. I hope the court is correct that prudence will be sufficient to guide us out of the thicket. But I believe the language of our Constitution already provides the guidance we need.
Doing so without determining whether the referendum is likely to qualify does not “departf] • • • from the language of article XXI, section 3(b)(2)” (maj. opn., ante, at pp. 462-463, fn. 25) because in cases such as this, where we deny a petitioner’s request to order interim use of a non-Commission map, we are not granting any “relief’ sought by the *493petitioner under section 3(b)(2). In light of section 3(b)(2)’s text (“Any registered voter in this state may ... file a petition ... to seek relief’), the illogic of the court’s assertion that we are granting “relief’ within the meaning of section 3(b)(2) “whether or not the particular outcome ordered by the court is the relief petitioner is seeking” (maj. opn., ante, at pp. 462-463, fn. 25) speaks for itself.
The practice of retaining jurisdiction in a mandate proceeding and postponing action in light of possible developments is not unfamiliar to this court. (See Legislature v. Reinecke (1973) 10 Cal.3d 396, 400 [110 Cal.Rptr. 718, 516 R2d 6].) As a result of litigation after the 1970 census and reapportionment, we adopted temporary maps for the 1972 elections and retained jurisdiction to draw new maps for subsequent elections if the Legislature failed to enact valid maps during the 1972 regular session. (Ibid, [citing Legislature v. Reinecke (1972) 6 Cal.3d 595, 603-604 [99 Cal.Rptr. 481, 492 P.2d 385]].) Subsequently, “at the request of the Senate of the State of California, we postponed the time for further court action” in order to allow the Legislature an opportunity to act in special session. (Ibid [citing Legislature v. Reinecke (1972) 7 Cal.3d 92, 93 [101 Cal.Rptr. 552, 496 P.2d 464]].) When the Legislature failed to enact valid maps in 1972, we appointed three special masters to propose new maps, even as we made clear that “ ‘If at any time during the proceedings contemplated by this order valid congressional and legislative reapportionment measures are enacted the court will entertain an application to dismiss these proceedings.’ (Legislature v. Reinecke (1973) 9 Cal.3d 166, 168 [107 Cal.Rptr. 18, 507 P.2d 626].)” (Ibid) Although Legislature v. Reinecke did not present the same issues we face here, the case illustrates the flexibility the court possesses to retain jurisdiction and defer decision where the exercise of our authority depends on developments occurring after we initially assume jurisdiction.
The Secretary of State’s supplemental briefing identified four initiatives in the past five years that went to a full count. Measure No. 1226, a 2009 initiative concerning community colleges, had a random sample validity rate of 71.38 percent and a full count validity rate of 71.08 percent. Measure No. 1271, a 2009 initiative concerning a children’s hospital bond, had a random sample validity rate of 69.30 percent and a full count rate of 70.84 percent. Measure No. 1311, a 2011 term limits measure, had a random sample validity rate of 75.58 percent and a full count rate of 74.36 percent. Finally, a 2011 attempt by the Americans Elect political party to qualify for the ballot through a signature drive had a random sample validity rate of 69.81 percent and a full count rate of 68.08 percent.