(dissenting). The centerpiece of the majority’s analysis is two examples from Schmidt v Dep’t of Ed, 441 Mich 236, 253; 490 NW2d 584 (1992), a case that addressed the first sentence of Const 1963, art 9, § 29.1 While acknowledging that the Schmidt examples are, “by definition, dicta,” ante at 600, the majority believes that they are representative of both the language and purpose of § 29. Because the language of § 29, its underlying purpose, and our opinion in Schmidt all belie the majority’s contentions, I respectfully dissent. I would hold that the Headlee Amendment obligations contained in the second sentence of § 29 run from the state to each unit of local government individually. I would also hold that the Court of Appeals erred in defining the relevant activity or service at issue in this case as “trial court operations.”
I. THE LANGUAGE OF CONST 1963, ART 9, § 29
The starting point for any constitutional analysis is the language of the constitution itself. The primary *612rule of constitutional construction is that of the “common understanding” of the voters approving the constitution. The rule has been described by Justice Cooley as follows:
A constitution is made for the people and by the people. The interpretation that should be given it is that which reasonable minds, the great mass of the people themselves, would give it. “For as the Constitution does not derive its force from the convention which framed, but from the people who ratified it, the intent to be arrived at is that of the people, and it is not to be supposed that they have looked for any dark or abstruse meaning in the words employed, but rather that they have accepted them in the sense most obvious to the common understanding, and ratified the instrument in the belief that that was the sense designed to be conveyed.” [Traverse City School Dist v Attorney General, 384 Mich 390, 405; 185 NW2d 9 (1971), quoting Cooley, Constitutional Limitations, p 81 (emphasis added).]
Moreover, the Court must give force to all the words used in a constitutional provision. “Neither the legislature, nor this Court, has any right to amend or change a provision. in the Constitution.” Pillon v Attorney General, 345 Mich 536, 547; 77 NW2d 257 (1956). Thus, the Court must avoid any construction that would invalidate portions of the language used in the provision under review.
Unfortunately, the majority has strayed from these time honored rules of construction. The majority’s conclusion that “the relevant frame of reference [under the second sentence of § 29] is whether a mandated activity is new (or at an increased level) to local units collectively,” ante at 599-600, is contradicted by the specific language of the second sentence:2
*613A 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. [Emphasis added.]
The italicized portion of this sentence explicitly refers to local unit in the singular. This use of the singular “suggests that the section was intended to impose an obligation on the state vis-a-vis each unit of local government with respect to mandated activities or services as well as new requirements.” Schmidt, 441 Mich 251 (emphasis added). The Court addressed this very issue in Schmidt:
The first sentence [of art 9, § 29] focuses on a single proportionate obligation by the state measured during the base year, i.e., the year that the Headlee Amendment became effective. The only language in § 29 that speaks directly to payment occurs in the second sentence, which requires a “state appropriation ... to pay the unit of Local Government for any necessary increased costs.” (Emphasis added.) This language references a singular unit and a singular governmental body. It suggests that the framers intended, and the voters understood, that the state’s obligation ran to each unit of local government. [Id.][3]
*614The majority believes that “[t]he use of the plural ‘units’ in the first phrase of this sentence indicates that the state is not to require activities that are new or at an increased level from the perspective of local units collectively.” Ante at 601. However, such a reading of § 29 does not comport with the common understanding of the terms selected by the drafters and ratified by the public. To reiterate, the second sentence of § 29 states:
*615A new activity or service of 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. [Emphasis added.]
Read in the context of the entire sentence, the use of the plural “units” in the first part of the sentence is merely a recognition that there are a variety of different units of local government.4 Unlike the singular state government, there are different units of local government, including those governing cities, counties, townships, school districts, and the like. Indeed, to read the reference to “units” any other way would require one to effectively ignore other language appearing in the sentence.
Section 29 requires in part that the state pay for “an increase in the level of any activity or service beyond that required by existing law . . . .” However, the majority’s interpretation effectively eliminates the verbiage. Under its analysis, it does not seem possible for a locality to recover for an “increased” obligation foisted on it by the state. If the required activity has already been undertaken by the collective local gov*616emments in the past, “§ 29 only guarantees that each local unit will receive the state-financed proportion of funding provided on a statewide basis in 1978.” Ante at 603. Therefore, even if the city of Detroit’s expenses increased tenfold because of the court reorganization plan, the city would not be able to recover “for any necessary increased costs” apart from the percentage the state was already paying before 1978. In other words, there is never an “increase” in an activity; rather, there is only a change in the proportionate share required by the state. This approach is simply incompatible with the specific language of § 29, and it violates the rule of constitutional construction that we must give force to the actual language of the constitutional provision.5
In sum, I believe the majority’s interpretation of the second sentence of § 29 strays from the actual text of the sentence. Its interpretation gives the word “units” exactly the “dark or abstruse meaning” that Justice Cooley warned against. Moreover, it seems to ignore the language pertaining to an “increase” in a required activity or service.
*617n. THE PURPOSE OF CONST 1963, ART 9, § 29
Not only do I find the majority’s interpretation contrary to the language of § 29,1 also find it contrary to the purposes underlying the Headlee Amendment. The Court has previously recognized that the Headlee Amendment was proposed as part of a nationwide “taxpayer revolt.” Durant v State Bd of Ed, 424 Mich 364, 378; 381 NW2d 662 (1985). The amendment was intended to prohibit unfunded mandates so that the state could not require new or expanded activities of local units without fully funding them. Id. at 391. In order to effectuate this intent, § 29 was designed to prohibit the state from shifting the costs of activities or services required by state law to various units of local government:
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. [Livingston Co v Dep’t of Management & Budget, 430 Mich 635, 644; 425 NW2d 65 (1988).][6]
*618Unfortunately, however, the majority’s analysis would allow the state to shift certain activities or services to local governments with no financial consequences. As long as some unit of local government undertook a particular activity or service in 1978, then the state would be able to require any or all other units of local government to undertake the task without providing any additional funding for this new mandate. For example, the state could require that every locality establish a separate “Recorder’s Court” for criminal prosecutions. Although this separate recorder’s court would be new to every locality except the city of Detroit, the majority’s analysis would not require the state to fund any of these new courts, since a single recorder’s court existed in the city of Detroit in 1978. Allowing the state to foist such an unfunded mandate on localities is, without question, diametrically opposed to the purpose behind Headlee.
Even so, the majority argues that it is reading § 29 in a way that is consistent with the purposes of Headlee. In particular, it argues that its interpretation meets the underlying purposes of “1) balancing state discretion in allocating funds and minimum funding guarantees for local units, and 2) uniform allocation of resources . . . .” Ante at 602. Assuming for sake of argument that these are the most significant purposes underlying § 29, I believe that the majority’s approach gives wide latitude to the state to require localities to undertake services that they have never undertaken before.7 As mentioned above, as long as *619some localities were undertaking the service in 1978, the state can require other localities to undertake the same function without providing funding for the new activity. More importantly, it allows the state to shift responsibilities between different units of local government so that resources axe not distributed uniformly. For example, the state could shift a variety of activities formerly performed by several different units of local government to a single unit of local government, while only paying for a portion of the new cost to the locality. The single unit of government would have to shoulder the financial burden of all these new activities while the other units of local government would be alleviated of the financial burden accompanying the activity. In other words, some units of local government would receive a windfall at the expense of another.
For these reasons, I cannot accept the majority’s reasoning. I believe that the state’s obligations under the second sentence of § 29 flow to each individual unit of local government and not local governments in the collective. Therefore, I would hold that, if the state should mandate a program shift from one local unit to another, that program is “new” to the unit that must now provide the activity. The local unit is entitled to an appropriation for any necessary increased costs as a result of the switch. Similarly, if the state should mandate a program shift from one local unit to another that requires the unit to provide a greater level of an activity than the unit already provides, the *620unit is likewise entitled to a state appropriation for any necessary increased costs resulting from the program shift.
Does this impose significant restrictions on state government? Does this, time and again, thwart the state’s ability to enact good public policy? The obvious answer to each question is “yes.” That, however, is the result, intended or unintended, of the Headlee Amendment that Michigan voters adopted.
in. THE RELEVANT ACTIVITY OR SERVICE
Examining Act 374 from the standpoint of the individual units of local government involved in this case, I believe that the court reorganization mandated by the act does violate the Headlee Amendment. While I believe that the Court of Appeals properly viewed this case from the perspective of the individual units of local government rather than local governments in the collective, I believe it erred in defining the relevant activity or service in this case as “trial court operations.” Rather, I believe the relevant activities or services in this case are “former Common Pleas Court functions” and “former Recorder’s Court functions.”8
*621The Court of Appeals holding is aptly summarized by the majority:
The Court of Appeals concluded that “trial court operations” was the relevant activity or service under § 29. 228 Mich App [386] 401-402 [579 NW2d 378 (1998)]. It found that Act 374 requires counties to fund the operation of the circuit courts, district units to fund the operation of district courts, and the state to fully fund circuit and district court judicial salaries. Id. at 402-403. It found that in 1978, state law mandated that local units fund and operate circuit and district courts and that the state subsidized a portion of judicial salaries. Id. at 405. It then concluded that Act 374 neither mandated new activities for local units nor increased the level of any activity required of local units. Id. Finally, it concluded that, despite incomplete information to perform the relevant equations, it was able to determine that Act 374 did not decrease the state-financed proportion of necessary costs of trial court operations to either Detroit or Wayne County from that provided on a statewide basis in 1978. Id. at 407-408. [Ante at 596-597.]
However, defining the relevant activity or service in question as “trial court operations” is simply too broad a category. First, the circumstances surrounding the adoption of the Headlee Amendment clearly indicate that the people of Michigan intended the word “activity” to apply to more specific functions than the umbrella term “trial court operations.” A primary aim of those who ratified the Headlee Amendment was to prevent the state from enacting laws and *622regulations that create financial burdens on local units, unaccompanied by financial support to alleviate those burdens. Schmidt, 441 Mich 251. If all trial court functions fall under the broad category of “trial court operations,” the state is free to impose a multitude of laws and regulations on a local unit without an appropriation for the resulting increased costs. To accept the Court of Appeals definition would create a situation in which a local unit can be depicted as having provided practically any activity in 1978, thereby precluding a finding that a “new” activity had been imposed.9 This construction effectively nullifies the second sentence of the amendment. The people, in adopting the Headlee Amendment, did not intend to offer so little protection against the very problem that they aimed to prevent.10
*623Moreover, this Court’s decision in Durant reflects that the proffered definition is too broad. In Durant, one issue presented to the Court was whether education, as a whole, is a state-mandated “activity” within the meaning of the Headlee Amendment. Plaintiffs attempted to characterize the relevant activity as “education.” They argued that because the state has delegated to local school districts the duty to provide education, all the functions performed by a school district are required by state law within the meaning of the Headlee Amendment. The Court rejected this reasoning and held that § 29 applies only to specific requirements imposed on the school districts by state statutes and state agencies. 11
By analogy, just as “education” is too broad a concept for a Headlee examination, so too is the broad concept of “trial court operations.” The Court of Appeals reviewed the statutory delineations of various educational functions in order to determine if the activities were specific and identifiable. We should similarly examine the statutory delineations of trial court functions. For example, the Common Pleas Court (the predecessor of the 36th District Court) *624was a distinct judicial entity, established and operated pursuant to 1929 PA 260 and vested with specific judicial authority different from any other court in Wayne County. The same can be said of Recorder’s Court. Thus, it is logical to define the relevant activity by the statutory parameters set for its delivery. To define the activities as “former Common Pleas Court functions” and “former Recorder’s Court functions” comports with the purpose and intent of the Headlee Amendment as well as the applicable case law.
Thus, I would hold that the relevant activities or services in this case are “former Common Pleas Court functions” and “former Recorder’s Court functions.” Such a holding would provide the level of specificity required by the Headlee Amendment. Viewing the relevant activities or services as such, I would find that Act 374 violates § 29 of the Headlee Amendment by requiring the city of Detroit to undertake the new or increased responsibility for “former Common Pleas Court functions,” and Wayne County for undertaking the new or increased responsibility for “former Recorder’s Court functions.”
IV. CONCLUSION
The majority finds no Headlee violation in this case by examining units of local government in the collective for purposes of determining whether the state has required them to undertake “new” or “increased” activities. However, I believe such an approach is simply inconsistent with the language and purpose of Const 1963, art 9, § 29. Act 374 requires that the city of Detroit provide “former Common Pleas functions” and that Wayne County provide “former Recorder’s Court functions.” These activities were previously *625performed by other units of local government. Nevertheless, I believe that the state’s funding obligation under the Headlee Amendment is implicated because the activities are “new” to these particular units of local government. Consequently, the state should be required to make appropriations for any necessary increased costs resulting from the requirements of Act 374.
Brickley and Kelly, JJ., concurred with Cavanagh, J.The first sentence of the majority opinion in Schmidt explicitly stated that “[t]he question presented is the proper interpretation of the first sentence of § 29 of the Headlee Amendment . . . .” 441 Mich 241 (emphasis added).
The complete language of § 29 is as follows:
*613The 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.
Ironically, the majority examines certain examples from the Schmidt opinion, but it ignores the remainder of the opinion’s logic, which, in discussing the first sentence of § 29, repudiates the exact view that is adopted by the majority. See id. at 259-260 (“[T]he statewide approach *614permits the state to shift funding from some districts to others while still requiring those districts to carry out mandated programs. This separation of the taxing question from the programing question is contrary to one of the primary purposes of the Headlee Amendment”).
The idea that the second sentence should be approached from the perspective of an individual unit of local government was also clearly expressed in both of the Schmidt dissents. In my own dissent, I explained:
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.” . . . [I]t 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. [Id. at 266, n 2, (opinion of Cavanagh, C.J.).]
Justice Levin’s dissent also came to a similar conclusion, albeit for a different reason:
The phrase “taken as a group” set forth in § 30 does not appear in § 29; for purposes of § 29, units of local government are not to be taken as a group. [Id. at 291, n 21.]
Thus, it seems the only thing that seven justices could agree on in Schmidt was that the second sentence required consideration from the perspective of each individual unit of local government. I believe this clear consensus undermines any reliance on Schmidt by the majority.
The language of the second sentence of § 29 is in notable contrast with that of § 30 of the Headlee Amendment:
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.]
If the drafters had meant for the second sentence to apply to units of local governments collectively, they could have made § 29 abundantly clear by adding the modifying phrase “taken as a group” after units of local government, just as they did in § 30. See Schmidt, 441 Mich 272-273 (Cavanagh, C.J., dissenting) (adding the words “taken as a group” after “units of local government” would have suggested “that only activities or services required of all units of local government were implicated”).
Moreover, the second sentence of § 29 specifically requires the state to pay an individual unit of local government “for any necessary increased costs.” (Emphasis added.) This is in direct contrast with the first sentence, which prohibits the state from reducing its “proportion of the necessary costs of any existing activity or service required of units of Local Government by state law.” However, the majority requires that the state only pay for a proportion of the necessary cost of any increase in an expense resulting from a shift in responsibilities among different units of local government. While I agree with the majority’s quotation of Schmidt explaining that the two sentences of § 29 “ ‘must be read together’ ” when interpreting the meaning of those two sentences, ante at 598, n 2, that does not mean that the two sentences should be read as requiring the same thing. The two sentences should be consistent with one another, but they still provide distinct requirements on the state and cover different situations. Any other interpretation would render the second sentence merely surplusage.
The need for § 29’s antishifting protections is underscored by § 31 of the Headlee Amendment, which specifically limits a local unit of government’s ability to raise revenue for any new or expanded activities. In particular, § 31 does not allow a locality to raise property or other local taxes “without the approval of a majority of the qualified electors.” See Bolt v City of Lansing, 459 Mich 152; 587 NW2d 264 (1998). Thus, Headlee’s antishifting purpose, embodied in § 29, is critical to prevent the state government from requiring new or increased activities of a locality without giving the locality the means to pay for that new mandate. Otherwise, the local government would be forced into the Hobson’s choice of trying to convince its constituency to approve by referendum a tax increase to pay for the unfunded state mandate, or to cut other areas of the locality’s budget to pay for the mandate. Either choice is not likely to be popular with the voting public.
It is worth noting that both these underlying principles come from Schmidt, which dealt with the first sentence of § 29. Since the first sentence requires that the state continue to fund its proportionate share of preexisting state mandates, it certainly makes sense that the Court was *619concerned about “minimum funding guarantees for local units” and “uniform allocation of resources." However, the second sentence requires funding for all new or increased activities. Thus, it is questionable whether these underlying purposes identified in Schmidt should apply to the second sentence of § 29 at all.
In 1978, the three trial courts located within Wayne County were organized differently than trial courts in other counties in the state of Michigan. In most Michigan counties there existed a district court and a circuit court. In Wayne County, there existed Recorder’s Court, the Common Pleas Court, and the Third Circuit Court. Detroit Recorder’s Court jurisdiction included felony, misdemeanor, and traffic and ordinance violations arising within the city of Detroit. Additionally, the civil functions normally performed by a district court were performed by the Common Pleas Court. Its jurisdiction included civil disputes less than $10,000 and landlord-tenant disputes. The Third Circuit Court’s organization and jurisdiction was the same as other circuit courts statewide.
In 1980, the Legislature enacted 1980 PA 438 and 440, reorganizing the three DetroitAVayne County trial courts. The acts created the 36th District Court and transferred to it the functions performed by the Common Pleas *621Court along with the criminal misdemeanor and traffic and ordinance divisions of Recorder’s Court. The Common Pleas Court was abolished. However, Recorder’s Court continued to have felony criminal jurisdiction.
It is worth noting that the Legislature conditioned the 1980 court reform legislation on the city of Detroit and Wayne County agreeing to assume responsibility for the expenses required of them by the court reform legislation. As a result of negotiations between the local units and the state, both the city and the county expressly waived any potential claims under the Headlee Amendment.
For example, as plaintiffs argue, local units also have the obligation to provide education, prosecuting attorneys, police officers, fire protection, etc. Every new local mandate will, by definition, fall under one of these categories since these are only the broadly defined categories under which local units of government are vested with the power to act. To define the relevant activity in such a broad manner means that every new mandate would be determined for Headlee purposes to be merely a continuation of broader, preexisting activities.
Further support is found in the definitions contained in the Headlee implementation statute:
“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 21.232(1); MSA 5.3194(602)(1).]
“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. [MCL 21.234(1); MSA 5.3194(604)(1).]
These definitions reinforce the idea that the Court of Appeals characterization of the relevant activity as “trial court operations” is too broad. Trial court functions as a whole have historically been broken down to specific *623areas of jurisdictional authority. Thus, to characterize the activity or service that a local unit provides as “trial court services” would do little to shed light on what the unit “specific[ally] and identifiabl[y]” provides.
Durant v State Bd of Ed, 424 Mich 375. This portion of the decision affirmed the decision of the Court of Appeals on remand, which stated:
Only those specific and identifiable programs which the state requires the school districts to provide by state statute or state agency regulation fall within the state financing requirements of § 29. [129 Mich App 517, 528; 342 NW2d 591 (1983).]
Examples of such state-mandated programs as stated in Durant are: courses in the constitution, history, government, and civics, courses in health and physical education, bilingual instruction in school districts with an enrollment of twenty or more children of limited English speaking ability, special education, etc.