(dissenting). The issue presented in this case, one of first impression in this Court,1 concerns the proper interpretation of the first sentence of § 29 of the Headlee Amendment, Const 1963, art 9, §§ 25-34. Contrary to the reasoning and conclusions reached in the opinions of my Sister Boyle and my Brother Levin, I conclude that the "statewide-to-statewide” interpretation of this provision is the most persuasive, and I would accordingly affirm on that basis the judgment of the Court of Appeals, and dismiss without prejudice the plaintiffs’ complaint.
I. THE LANGUAGE OF § 29
As this Court has held, in order to properly interpret the meaning of the Headlee Amendment, "we must ascertain the intent of the voters who passed [it].” Durant v State Bd of Ed, 424 Mich *265364, 378; 381 NW2d 662 (1985). The starting point for that inquiry is "the language of the constitution itself.” Id.
Section 29 of the Headlee Amendment provides:
The state is hereby prohibited from reducing the state financed proportion of the necessary costs of any existing activity or service required of units of Local Government by state law. A new activity or service or an increase in the level of any activity or service beyond that required by existing law shall not be required by the legislature or any state agency of units of Local Government, unless a state appropriation is made and disbursed to pay the unit of Local Government for any necessary increased costs. The provision of this section shall not apply to costs incurred pursuant to Article VI, Section 18.
The dispositive issue in this case is the proper meaning of the language in the first sentence of § 29, "the state financed proportion of the necessary costs of any existing activity or service required of units of Local Government by state law.” The question is whether that language refers to the state financed proportion of (1) the necessary costs of all activities and services required by state law of all units of local government, calculated on a statewide basis, (2) the necessary costs of each state-required activity or service taken individually, but calculated on a statewide basis with regard to all relevant units of local government, or (3) the necessary costs incurred by each unit of local government, taken separately, with regard to each state-required activity or service.
The first of these alternatives is clearly precluded by § 29’s reference, in the singular, to "any existing activity or service.” Clearly, the state must, with regard to each state-required activity *266or service, maintain at least the same proportionate level of financing of that activity or service that prevailed in 1978-79. But does it follow that the state must maintain the same proportionate financing of the costs incurred by each unit of local government with regard to each activity or service, as under the third alternative? Nothing in § 29 suggests that. The relevant sentence refers to "units of Local Government” in the plural, not in the singular.2 That reference occurs as part of the adjectival phrase, "required of units of Local Government by state law,” which defines "existing activity or service.” That adjectival phrase reflects the obvious fact that any given state-required activity or service will apply to a large number of local units across the state.3 Because the language *267refers to "the state financed proportion of the necessary costs of any existing activity or service” it seems eminently logical to calculate the "necessary costs” on the basis of the aggregate costs incurred by all "units of Local Government” that are required to provide each activity or service.
Section 29 does not refer to "the state financed proportion of the necessary costs of each unit of Local Government.” If such a local-to-local interpretation had been intended, that meaning could have been conveyed in any number of clearer and more sensible ways. Section 29, for example, could have been phrased in relevant part as:
The state is hereby prohibited from reducing the state financed proportion of the necessary costs incurred by any unit of Local Government in providing any existing activity or service required of units of Local Government by state law. [Emphasized words added.]
Or, perhaps less clearly:
The state is hereby prohibited from reducing the state financed proportion of the necessary costs of any existing activity or service required of any unit of Local Government by state law. [Emphasized words added.]
In sum, because § 29 focuses on "the state financed proportion of the necessary costs of any existing [state-required] activity or service,” not on the state financed proportion of the necessary costs of each unit of local government, and because § 29 refers in the relevant sentence to "units” in the plural, the statewide-to-statewide interpretation *268appears to be the most logical and persuasive way to read the provision.4
II. THE PURPOSE OF § 29
This Court has held that in attempting to determine the meaning of a constitutional provision, " 'the circumstances surrounding the adoption of [the] provision and the purpose sought to be accomplished may be considered.’ ” Livingston Co v Dep’t of Management & Budget, 430 Mich 635, 642; 425 NW2d 65 (1988), quoting Traverse City School Dist v Attorney General, 384 Mich 390, 405; 185 NW2d 9 (1971).
The plaintiffs argue that the statewide-to-statewide interpretation undermines the purposes of § 29 and of the Headlee Amendment as a whole. There can be little dispute about those purposes. As this Court said in Livingston Co, 430 Mich 644:
Th[e] plan is quite obvious. Having placed a limit on state spending, it was necessary to keep the state from creating loopholes either by shifting more programs to units of local government without the funds to carry them out, or by reducing the state’s proportion of spending for "required” programs in effect at the time the Headlee Amendment was ratified.
I disagree, however, with the contention that the *269statewide-to-statewide interpretation in any way conflicts with the plan or purposes of the Headlee Amendment. The plaintiffs argue essentially that the statewide-to-statewide interpretation would permit the state to reduce its proportionate funding of the costs incurred by some school districts,5 while increasing its proportionate funding of others. This may be true, but it is irrelevant to the broad antishifting purpose of § 29. As the state notes in its brief, "[t]he state does not realize one nickel in savings if it allocates funds for a particular activity or service to school district a instead of school district b.”6
As long as the state is prevented from shifting its funding responsibilities onto local governments on an overall, statewide basis, the goal of § 29 is fully realized and effectuated. Because that is precisely what the statewide-to-statewide interpretation does, in fact, prevent, that interpretation does not open up any loophole in the Headlee Amendment.7
*270I do not suggest that there are no potential legal grounds upon which to challenge any particular school funding equalization scheme. There may be any number of grounds for such a challenge. But we are not presented in this case with any such issues. The sole issue before us is whether the plaintiffs have stated a claim under § 29. That provision simply has no bearing on the validity of any equalization or redistribution of state funds among school districts, as long as the state does not attempt to shift its aggregate, statewide funding responsibilities, with regard to "any existing [state-required] activity or service,” onto units of local government. Because the plaintiffs in this case have not alleged any such aggregate, statewide shifting, their § 29 claim fails as a matter of law.8
*271III. THE RELEVANCE OF § 30
The plaintiffs, in urging the local-to-local interpretation of § 29, place heavy reliance on a comparison with § 30 of the Headlee Amendment, which provides:
The proportion of total state spending paid to all units of Local Government, taken as a group, shall not be reduced below that proportion in effect in fiscal year 1978-79. [Emphasis added.]
The plaintiffs argue essentially that because the words "taken as a group” do not appear in § 29, there is no basis for interpreting § 29 in a way that treats "units of Local Government” in the aggregate, as under the statewide-to-statewide interpretation.
In my view, the § 30 comparison does not support the local-to-local interpretation of § 29, nor does it undermine the statewide-to-statewide interpretation. It was obviously necessary to include the words "taken as a group” in § 30 to avoid a potentially serious ambiguity in that provision. Without the words "taken as a group,” § 30’s reference to "state spending paid to all units of Local Government” might have been interpreted to mean state spending paid to each and every unit of local government. The effect would have been to freeze the allocation of state funds among all the units of local government.
No comparable ambiguity exists in § 29, and there was therefore no need to add words such as "taken as a group” to convey a statewide meaning already clearly evident from the text of § 29. A comparable ambiguity in § 29 might perhaps have existed if § 29 had been phrased as follows:
The state is hereby prohibited from reducing the *272state financed proportion of the necessary costs of all units of Local Government. [Emphasized words added.]
But as enacted, § 29 refers to "the state financed proportion of the necessary costs of any existing activity or service required of units of Local Government by state law.” As we have already noted, the concluding phrase, "required of units of Local Government by state law,” is simply an adjectival phrase modifying "any existing activity or service,” which is the operative heart of the sentence.
It would not have been logical or coherent for § 29 to use the words "taken as a group,” because § 29, unlike § 30, was not designed simply to preserve a proportionate level of funding for all units of local government, for all purposes. Rather, § 29 was designed to preserve a proportionate level of funding with respect to each "existing [state-required] activity or service.” As I have already demonstrated, however, nothing in the first sentence of § 29 suggests that its requirements apply not only to each state-required activity or service but also to each unit of local government. To the contrary, as I have noted, the plural reference to "units of Local Government” suggests the opposite, as does the fact that each given activity or service will obviously apply to a large number of local units.
Finally, while the plaintiffs vest paramount significance in the absence of the words "taken as a group” from the first sentence of § 29, it is difficult to see where the words could have been inserted to produce the sought-after clarification. If they had been inserted following the words "units of Local Government,” they would only have created ambiguity and confusion by suggesting that only activities or services required of all units of local gov*273ernment were implicated. If they had been inserted at the end of the first sentence, their referent would have been ambiguous, and if the referent were taken to be "any existing activity or service,” the plural implications of "taken as a group” would have conflicted inexplicably with the singular specification of "any existing activity or service.”
In sum, the absence of the words "taken as a group” from § 29 does not undermine the statewide-to-statewide interpretation, because that interpretation already flows clearly from the existing language of §29, because the words "taken as a group” serve a clarification purpose unique to § 30, and because adding the words "taken as a group” to § 29 would not have achieved any further clarity in that provision.9 The relevant issue is not whether the requirements of the first sentence of § 29 apply to units of local government "taken as a group,” but whether those requirements apply to state-required activities and services, taken individually. They clearly do, and this, in my view, leads with inexorable logic to the statewide-to-statewide interpretation.10_
*274IV. POLICY IMPLICATIONS
This Court’s interpretation of § 29 should not, of course, be swayed by the perceived policy implications. Our task is simply to arrive at the most persuasive and intellectually honest interpretation of § 29 that we can. But I find it nevertheless appropriate to point out, in view of the policy arguments raised by the parties, the amici curiae, and my colleagues, that the local-to-local interpretation of § 29, aside from being inherently unpersuasive, would, as the state argues, lead to irrational and inequitable consequences.
Under the local-to-local approach, the pattern of state funding of individual school districts would, to a large extent, be permanently frozen according to the pattern existing in 1978, when the Headlee Amendment was adopted. Not only would the local-to-local approach tend to preserve whatever funding inequities may have existed in 1978, but it would hinder the state from responding to even more pronounced inequities that may have developed over the last fourteen years, and that may continue to develop into the foreseeable future, as a result of changing economic and demographic circumstances within districts across the state.11
*275By adopting the Headlee Amendment, the voters clearly intended to place strict limits on the growth of state taxation and spending. By including § 29, the voters clearly intended to prevent the state from evading the Headlee Amendment’s restrictions by shifting its overall funding responsibilities onto local governments. But there is nothing in the language or background of the Headlee Amendment to support the conclusion that the voters intended to permanently freeze the pattern of state funding among school districts or other units of local government that happened to exist in 1978. I do not believe that Michigan voters intended to permanently immunize relatively wealthy school districts from the possible effects of funding equalization or redistribution schemes. I am aware of no evidence that such issues played any role at all in the debate or adoption of the Headlee Amendment. In any event, the amendment itself cannot fairly be read to require such a result.
V. THE STATEWIDE-TO-LOCAL INTERPRETATION
My Sister Boyle’s opinion adopts the statewide-to-local interpretation first suggested as an option in our order granting leave to appeal, and urged before us by the mea. While that interpretation might, at first glance, appear to be a superficially attractive compromise, it is clearly untenable as a *276construction of the language of § 29. It would require that the phrase, "the state financed proportion of the necessary costs of any existing activity or service required of units of Local Government by state law,” be interpreted in a fundamentally different manner depending on whether the base year (1978-79) or the year of challenged funding was at issue.12
The relevant phrase, however, appears only once in § 29 and can have only one meaning. Whichever meaning is most persuasive must clearly apply both to the base year and the year of challenged funding.13 I would therefore reject the statewide-to-local interpretation as no more persuasive than the local-to-local interpretation.14
*277My colleague attempts to evade the logical impossibility of the statewide-to-local interpretation by stressing that the first and second sentences of § 29 must be read together, and by stressing that the first sentence refers to "the state financed proportion . . . See Boyle, J., ante, pp 249-251. I agree that it is generally appropriate to read a constitutional provision as a whole. In this instance, however, the different language of the first and second sentences of § 29 does not contain the key to the puzzle sought by my colleague. While the first and second sentences are obviously related, they address distinct and separable issues. The first sentence is a self-contained provision governing the funding of programs that were required by state law at the time the Headlee Amendment was passed. The second sentence is a self-contained provision governing the funding of new programs mandated by post-Headlee Amendment state laws. I have already acknowledged and discussed the obvious fact that the second sentence, in contrast to the ñrst sentence, does "apply on an individual unit basis.” See n 2. Indeed, as I have noted, that very contrast in language supports the statewide-to-statewide interpretation with regard to the first sentence. See n 2.
My colleague asserts that "[t]he only language in § 29 that speaks directly to payment occurs in the second sentence . . . .” Boyle, J., ante, p 251. But this is not really true, and does not support the conclusion that my colleague draws. It is true that the second sentence in § 29 contains a more *278explicit reference to the state’s duty to "appropriat[e]” and "disburse[]” the required funding to local governments, but this refers only to the separate requirements imposed by the second sentence. The first sentence independently refers to the state’s payment responsibility in the year of challenged funding, by providing that the state shall not "reduc[ej” the proportionate funding level existing when the Headlee Amendment was passed, which is simply another way of saying that the state shall maintain and continue paying that proportionate amount of funding.15
The fact that the first sentence in § 29 refers to "the state financed proportion” does not support the statewide-to-statewide, the statewide-to-local, or the local-to-local interpretation, and simply does not provide the guidance discerned by my colleague. It is impossible to say, on the sole basis of the use of the word "the,” whether "the state financed proportion” refers to the proportion as measured on a statewide or local unit basis. Consider a hypothetical constitutional provision stating: "The state shall annually appropriate to each unit of local government the amount of funds required by each unit to provide adequate police services.” The use of "the” does not necessarily imply or require that the term modified by "the” be a unique, unitary, or statewide entity. "[T]he *279state financed proportion” in § 29, like "the amount of funds” in our hypothetical provision, could refer to the multiplicity of varying amounts or proportions of funding relevant to different local units. The argument for the statewide-to-statewide interpretation rests not on the use of the word "the,” but primarily, as a textual matter, on the grounds I have stated in part i.
Shifting from textual to more broadly theoretical and policy-based arguments, my Sister Boyle joins my Brother Levin in asserting that the statewide-to-statewide interpretation might theoretically permit the state to engage in some bizarre, extreme, or irrational form of funding redistribution, such as the hypothesis that the state might "fund[ ] the cost of all mandated services in one county and none in the other eighty-two,” Boyle, J., ante, p 258, or the notion that the state might inadvertently, through the operation of the complex factors governing funding redistribution, "shift funds from poorer districts to wealthier . . . .” Ante, p 258, n 25.
This "parade-of-horribles” type of argument misses the point in my view. I have already noted that "[t]here may be any number of grounds,” apart from the Headlee Amendment provision at issue, "upon which to challenge any particular school funding equalization scheme.” See part n, ante, p 270. More to the point, it would simply be unreasonable to suppose that the Headlee Amendment was designed or intended to prohibit or forestall any conceivably unwise, unfair, or even evil state fiscal policy that might be imagined. The first sentence of § 29 of the Headlee Amendment, as its text reveals, has the specific and limited objective of preventing any overall shifting of funding responsibilities for state-mandated programs from the state onto local governments. As I *280have discussed in part n, the statewide-to-statewide interpretation fully vindicates this objective. That there may be many unwise, unfair, or even evil state policies that would not be prohibited by § 29 under the statewide-to-statewide interpretation is a meaningless and irrelevant observation that would be true of any constitutional provision. As a wise judge once said:
The process of Constitutional adjudication does not thrive on conjuring up horrible possibilities that never happen in the real world and devising doctrines sufficiently comprehensive in detail to cover the remotest contingency. [New York v United States, 326 US 572, 583; 66 S Ct 310; 90 L Ed 326 (1946) (opinion of Frankfurter, J., joined by Rutledge, J.).]
My Sister Boyle takes me to task for what she describes as my "assumption that the drafters of the [Headlee] [Ajmendment were content to let the acceptability of discrete decisions within the aggregate responsibility, be evaluated in the polling place.” Ante, p 259. Translated into plain English, this amounts to the proposition that, to the extent the state is not governed by constitutional restraints, it is free to do whatever it wants to do, subject only to the democratic verdict of the voters. I would heartily embrace this proposition, which I had thought was fundamental to our form of government. While I recognize that my colleague and I have an honest difference of opinion regarding the scope of the constitutional restraint at issue, I reject the notion that my analysis is flawed simply because it would leave some funding and distribution decisions by the state, within the overall confines of the Headlee Amendment, to the democratic judgment of the political process. As Chief Justice John Marshall once said:
The wisdom and the discretion of [the members *281of the Legislature], their identity with the people, and the influence which their constituents possess at election, are, in this, as in many other instances, . . . the sole restraints on which they have relied, to secure them from [the] abuse [of power]. They are the restraints on which the people must often rely solely, in all representative governments. [Gibbons v Ogden, 22 US (9 Wheat) 1, 197; 6 L Ed 23 (1824).]
VI. THE SOCIAL SECURITY ISSUE
I turn finally to the plaintiffs’ social security claim, as described in my Sister Boyle’s opinion, pp 246-247. This claim suffers from a number of infirmities. First of all, the obligation of units of local government to pay the employer share of social security taxes is directly imposed by federal law, see ante, p 246, n 10, and thus is not "required ... by state law” within the meaning of § 29.
Second, social security coverage is not itself an "activity or service” within the meaning of § 29. MCL 21.232(1); MSA 5.3194(602)(1) provides:
"Activity” means a specific and identifiable administrative action of a local unit of government. The provision of a benefit for, or the protection of, public employees of a local unit of government is not an administrative action.
MCL 21.234(1); MSA 5.3194(604)(1) provides:
"Service” means a specific and identifiable program of a local unit of government which is available to the general public or is provided for the citizens of the local unit of government. The provision of a benefit for, or the protection of, public employees of a local unit of government is not a program.
I agree with these definitions of "activity or ser*282vice” for purposes of § 29,16 and they clearly exclude a general employee benefit such as social security coverage.
It is also clear, however, that the cost of social security coverage for an employee engaged in carrying out any "activity or service required of units of Local Government by state law” does constitute part of the "necessary costs” of such an activity or service. This follows from the definition of "necessary costs” adopted in Durant, 424 Mich 389-391. Indeed, the state concedes this point in this case.17
Thus, the state would violate § 29 if any reduction of the state financed proportion of local government social security coverage costs resulted in a reduction, as compared to 1978-79, of "the state financed proportion of the necessary costs of any existing activity or service required of units of Local Government by state law.” However, as I have noted, the plaintiffs in this case have not *283alleged any such reduction under the statewide-to-statewide interpretation of § 29 that I would adopt; nor, for that matter, have they alleged any such reduction under the statewide-to-local interpretation adopted by my Sister Boyle. Thus, the social security component of the plaintiffs’ complaint does not state a claim under § 29.
VII. CONCLUSION
I would adopt the statewide-to-statewide interpretation of the first sentence of Const 1963, art 9, § 29. Under that approach, the proper comparison for purposes of assessing the state’s compliance is between the state financed proportion in 1978-79, and in the year of challenged funding, of the total necessary costs incurred by all relevant units of local government, statewide, in providing the state-required activity or service at issue.18 In accordance with that interpretation, I would affirm the judgment of the Court of Appeals and dismiss without prejudice the plaintiffs’ complaint._
My Sister Boyle suggests that the dispositive issue in this case is actually governed by Durant v State Bd of Ed, 424 Mich 364; 381 NW2d 662 (1985). See ante, pp 256-260. My colleague admits, however, that Durant does not speak directly to the issue; rather, my colleague invokes certain alleged "implicit assumptions” in Durant. See ante, p 256. My colleague, however, departs from an express holding of Durant, by relying on the drafters’ notes of the Headlee Amendment. See ante, p 257, n 24. But see Durant, 424 Mich 382, n 12 (finding the drafters’ notes to be "of no value in this casé,” and citing the well-settled rule that "comments made after the adoption or passage of a statute or constitutional provision are given little weight”).
In a noteworthy contrast, the second sentence of § 29, dealing with state financing of new or increased state-required activities or services, refers to "unit of Local Government” in the singular, by way of prohibiting the state from adding new program obligations or increasing the level of existing program obligations on local governments, "unless a state appropriation is made and disbursed to pay the unit of Local Government for any necessary increased costs.” Thus, the framers of § 29 were perfectly capable of referring to "unit of Local Government” in the singular when they wished to do so, and it would disregard the careful phrasing of § 29 to treat the plural reference to "units of Local Government” in the first sentence as if it were in the singular. Of course, it is entirely logical for the requirements of the second sentence of § 29 to apply on a local unit basis. The essence of the second sentence is to require that all new, increased costs resulting from new or increased state program requirements be one hundred percent funded by the state. This could not occur unless every unit’s increased costs in that regard were funded one hundred percent.
Contrary to an argument asserted by the plaintiffs, there is no logical conflict between the statewide-to-statewide interpretation and the fact that many state-required activities and services may not apply to all units of local government in the state. Section 29 refers to programs "required of units of Local Government,” not programs "required of all units of Local Government.” For example, the state-required programs primarily at issue in this case, see n 2 and accompanying text, apply to all school districts but not to other types of units of local government. The focus under the statewide-to-statewide interpretation is simply on the state financed proportion of the aggregate "necessary costs” of each activity or service, regardless of *267which category or categories of units of local government are required to provide that particular activity or service.
Inasmuch as I agree with my Brother Levin that "unit of Local Government,” as used in § 29, refers to individual units, including individual school districts, I do not follow the logic of my colleague’s argument on pp 299-301, and I fail to see how that argument undercuts my analysis or advances that of my colleague. The issue is not the definition of "unit of Local Government,” which is not in dispute. Rather, as my colleague concedes, see post, p 303, the primary issue is the scope of the term, "the state financed proportion.” While I rely in part on the fact that "units” is rendered in the plural in the first sentence of § 29, that has nothing to do with the definition of "unit of Local Government” itself, and could not possibly conflict with the “geographically limited” nature of a "unit of Local Government.” Cf. post, p 302.
Specifically, of course, they argue that the state has reduced its proportionate funding of the plaintiff districts themselves.
My Sister Boyle asserts that this is a "fallacy,” ante, p 259, n 27, because it might, in fact, be more expensive for the state to finance any given equalization scheme if the state were simultaneously required to maintain certain wealthier school districts at a higher level of subsidy pursuant to § 29. The point of the state’s argument, however, is that the state realizes no savings with regard to financing any particular state-required program, and thus, no shifting of the overall financial burden with regard to that program. That the statewide-to-statewide interpretation might make it easier for the state to pursue broader goals such as general funding equalization is not a reason to reject that interpretation. On the contrary, the fact that the local-to-local and statewide-to-local interpretations might make it effectively impossible for the state to pursue such equalization, by requiring the state to subsidize some wealthier districts in perpetuity at an unduly high level, is a very strong reason to doubt that the voters intended such unjust and irrational results. See part IV.
The Headlee Amendment itself, in its introductory section, summarizes its intended effect by stating, in relevant part:
*270Property taxes and other local taxes and state taxation and spending may not be increased above the limitations specified herein without direct voter approval. The state is prohibited from requiring any new or expanded activities by local governments without full state financing, from reducing the proportion of state spending in the form of aid to local governments, or from shifting the tax burden to local government. [Const 1963, art 9, § 25.]
The references to "local governments” or "local government” are in generalized terms suggesting the overall balance between state and local responsibilities in the aggregate, not the balance between state responsibilities and the responsibilities of any individual unit of local government. Section 25 does not indicate that the Headlee Amendment’s purpose is to prohibit the state “from shifting the tax burden to any unit of local government.”
My Brother Levin asserts that the recapture provision in § 21(5) of the School Aid Act does, in fact, result in a statewide shifting of funding responsibilities in violation of § 29. Post, pp 294-295 (part iii[a]). My colleague misses the point. I neither deny nor concede the possibility that the recapture provision may operate in such a manner. Certainly, an allegation of such improper statewide shifting would state a claim under § 29. But the plaintiffs have not made any such allegation in this case. No such claim or issue is before this Court in this case, and my colleague’s discussion of it therefore constitutes dicta.
For the reasons I have stated in this section, my analysis of § 29 does not rest on "reading] into § 29” the words "taken as a group,” Levin J., post, p 304; nor, unlike my colleague, do I read the state’s argument to rest on any such theory. I have no need to "read” anything "into” § 29 in order to arrive at the statewide-to-statewide interpretation. The whole point of my discussion in this section is to demonstrate that the appearance of the words "taken as a group” in § 30 does not, ultimately, have any relevance to the proper interpretation of § 29.
The plaintiffs argue further that the statewide-to-statewide interpretation of § 29 would merely impose upon the state the identical obligations imposed by § 30, and thus render § 29 redundant and nugatory. This argument, however, is obviously erroneous. Nothing in § 30, standing alone, would prevent the state from reducing its proportionate level of financing of any given state-required activity or service, as long as the state continued to devote the same overall percentage of its budget as in 1978-79 to funding for all units of local government, taken as a group, for all purposes; yet § 29, under the *274statewide-to-statewide interpretation, obviously does prohibit any reduction in proportionate funding of any state-required activity. Conversely, nothing in § 29, under any of the interpretations urged in this case, would necessarily prevent the state from dramatically increasing funding in areas of its budget other than aid to local governments, while freezing or increasing at a lower rate the amount of local aid, thereby reducing the overall proportion of local aid within the state budget. Yet § 30 obviously does prohibit any reduction in the overall proportion of local aid within the state budget.
Under the local-to-local approach, if the state financed thirty percent and seventy percent, respectively, of the costs incurred by school districts A and b with regard to a particular state-required program in 1978-79, the state would remain obligated, from 1978 to the present, and into the foreseeable future, to continue financing both districts’ costs at the same guaranteed minimum rate. This *275would remain the case even if, over time, school district a became the poorest district in the state, and school district b the wealthiest. This growing inequity might be further exacerbated if, as might be expected, the poorer district’s costs with regard to certain state-required programs (such as bilingual education and lunch and supplemental milk programs) tended to increase as its economic and demographic profile shifted — for example, as a result of an influx of economically disadvantaged, non-English-speaking immigrants. Cf. Plyler v Doe, 457 US 202; 102 S Ct 2382; 72 L Ed 2d 786 (1982) (holding that even illegal immigrant children are constitutionally guaranteed equal access to public education).
With regard to the year of challenged funding, the statewide-to-local interpretation, like the local-to-local interpretation, looks to the state financed proportion of the necessary costs incurred by each individual district with regard to each required program. With regard to the base year, however, the statewide-to-local interpretation looks to the statewide average proportionate financing of the necessary costs of each required program.
The required comparison between the base year and the year of challenged funding is conveyed by § 29’s language providing that "[t]he state is hereby prohibited from reducing . . . .” (Emphasis added.) What it is prohibited from "reducing” is set forth only once in the phrase that follows. The language simply does not allow for a different basis for calculating "the state financed proportion” in the year of challenged funding, as compared to the base year. A provision consistent with the statewide-to-local interpretation might perhaps have been phrased as follows:
The state is hereby prohibited from reducing the state financed proportion of the necessary costs incurred by any unit of Local Government in providing any existing activity or service required of units of Local Government by state law, below the level of the overall state financed proportion, in the year of the adoption of this amendment, of the total necessary costs of any existing activity or service required of units of Local Government by state law.
Furthermore, I doubt very much whether the statewide-to-local interpretation is, in fact, a desirable compromise even in practical policy terms. It appears that it would, like the local-to-local interpretation, lead to irrational and inequitable consequences. By perma*277nently freezing every individual school district at the same guaranteed minimum proportion of state funding — the statewide average proportion prevailing in the base year — the statewide-to-local interpretation would take no account of the widely differing economic circumstances of districts throughout the state, both in the base year and in the face of potentially dramatic changes in later years in the relative wealth and poverty of different districts.
My colleague suggests that the essence of her approach is to somehow blend or combine the elements of the first and second sentences of § 29, as opposed to, in her words, "viewing these two approaches as embodying conflicting principles requiring a choice Boyle, J., ante, pp 253-254. I do not see any "conflict” between the first and second sentences under my interpretation, however, and I do not "choose” between their respective "approaches,” because, as noted in the text, the two sentences are each self-contained, address distinct issues, and impose distinct requirements. This case involves only the issue governed by, and the requirements imposed by, the first sentence. To mix and match the language of different portions of the Headlee Amendment results in a distortion of the precise and carefully wrought meaning of that language.
The quoted statutory provisions are part of the definitional sections of the legislation enacted by 1979 PA 101 to implement the Headlee Amendment. Legislation enacted subsequent to the adoption of a constitutional amendment cannot, of course, be given much, if any, weight in ascertaining the intent of the voters who adopted the amendment. Cf. Durant, 424 Mich 382, n 12 (disclaiming reliance on the drafters’ notes published after the adoption of the Headlee Amendment). This Court, not the Legislature, has the ultimate authority and responsibility to interpret the meaning of any law or constitutional provision, see Romein v General Motors Corp, 436 Mich 515, 539, n 20; 462 NW2d 555 (1990), aff’d 503 US —; 112 S Ct 1105; 117 L Ed 2d 328 (1992), and we have not hesitated in the past to reject legislative interpretations of the Headlee Amendment that we find to be inconsistent with the true meaning intended by the voters. See Durant, 424 Mich 391-392 (rejecting the legislative definition of "necessary costs” under the Headlee Amendment). In this instance, however, I find the legislative definition persuasive and I would adopt it.
By the same token, the cost of social security coverage for employees engaged in general educational services apart from any specific state-required program would not be counted as part of the "necessary costs” of any such program. As this Court held in Durant, 424 Mich 377-387, to whatever extent the mandate of Const 1963, art 8, § 2, guaranteeing free public elementary and secondary education, applies either to the state or local governments, it does not constitute a requirement of "state law” within the meaning of § 29.
My Brother Levin mischaracterizes the state’s argument and my analysis by repeatedly asserting that they are premised on the theory that the state may comply with § 29 by simply "tentatively allocat[ing]” funding to units of local government, rather than actually providing such funding. See post, pp 287-288, 290, 295-296, 298. Nothing in my opinion even remotely suggests such a theory, nor do I understand the state to suggest such a theory. Obviously, the state’s compliance with § 29 must be measured by funding that is actually provided, not that which is merely "tentatively allocated.” Obviously, if the "music” were indeed "stop[ped] . . . immediately after the allocation,” as my colleague suggests, post, p 287, and if the subsequent adjustments in funding pursuant to the recapture provision were deemed legally irrelevant, the plaintiffs’ entire argument, and my entire analysis responding to that argument, would be moot and pointless. As I have made perfectly clear, my analysis is premised, not on the state’s "tentative allocation” of funding, but on the fact that the plaintiffs have not alleged that the recapture provision has caused any actual reduction of proportionate state funding on a statewide basis.
The statewide-to-statewide versus local-to-local issue, far from being a "red herring,” post, p 295, is the central and dispositive issue in this case, to which the parties have devoted virtually all their briefing and argument.