dissenting: I respectfully dissent from the court’s conclusion that RSA 507-B:2,1, violates articles 2,12 and 14 of part I of the Constitution of New Hampshire insofar as it would provide municipal immunity for liability arising out of ownership, occupation, and maintenance of “public sidewalks, streets [and] highways. . . .” In explaining why, I will confine my thoughts to what I see as a mis*121application of the accepted middle-tier equal protection standard and say nothing more about article 14, given this court’s view that the article mirrors the equal protection standard for scrutinizing limitations on a generally applicable civil right of action, see Estate of Cargill v. City of Rochester, 119 N.H. 661, 667, 406 A.2d 704, 707 (1979).
At least since the date of Carson v. Maurer, 120 N.H. 925, 424 A.2d 825 (1980), the guarantee of equal protection under articles 2 and 12 has been applied to limitations on rights of civil recovery by recognizing a natural class of all people personally injured by conduct within a recognized category of ostensibly tortious behavior. Leaving aside the special problems that are said to be created by the workers’ compensation statute, see Estabrook v. American Hoist & Derrick, Inc., 127 N.H. 162, 183, 498 A.2d 741, 754 (1985) (Souter, J., dissenting), we are supposed to pass on statutory classifications restricting a right to recover for an injury so inflicted by treating the affected cause of action as “an important substantive right,” Carson v. Maurer, supra at 931-32, 424 A.2d at 830, and by requiring that any classification resulting from a restriction upon such a right “be reasonable, not arbitrary, and . . . rest upon some ground of difference having a fair and substantial relation to the object of the legislation. .. . (Emphasis added.) State v. Scoville, 113 N.H. 161, 163, 304 A.2d 366, 369 (1973), quoting F. S. Royster Guano Co. v. Virginia, 253 U.S. 412, 415 (1920).” Carson v. Maurer, supra at 932, 424 A.2d at 831 (further citations omitted).
Middle-tier equal protection scrutiny thus entered the jurisprudence of the State Constitution, id., and accompanying questions as yet unanswered will provide fodder for a good many opinions. I am concerned here, however, with what I see as the misapplication of this intermediate standard, assuming it to be truly intermediate in character and as limited in application as the court in Carson indicated by the language used to describe it.
An understanding of that intermediate character and the limits of such review can prove elusive, however, and it is well to acknowledge that Carson’s test suffers from a proven susceptibility to confusion with other standards of equal protection review, a failing perhaps portended by the derivation of Carson’s language from F.S. Royster Guano Co. v. Virginia, 253 U.S. 412, 415. Royster involved a challenge to disparate State corporate income tax treatments, as to which the latitude of State legislative discretion under the Fourteenth Amendment’s equal protection clause was admitted to be “no*122tably wide.” Id. The majority nonetheless struck down the statute in question because they could “conceive” of “no ground” for justifying the differential treatment, id. at 416; Justices Brandéis and Holmes, on the other hand, dissented because they could conceive of just such a ground, id at 418. What is clear, however, is that the entire Court treated the test of what was “reasonable, not arbitrary ... having a fair and substantial relation to the object of the legislation,” id. at 415, as what we today would call the first-tier, rational basis test. Although the federal judiciary, like this court, has subsequently tried to use Royster's formulation to provide “somewhat heightened” middle-tier scrutiny, City of Cleburne v. Cleburne Living Center, 473 U.S. 432, 441 (1985), the very opinions cited in Carson as so applying it have reverted to type, as it were, by lapsing into rational basis terminology. See, e.g., on legitimacy classifications, Lalli v. Lalli, 439 U.S. 259, 273 (1978), cited in Carson v. Maurer, supra at 932, 424 A.2d at 831 (“our inquiry [is]... on whether the statute’s relation to the state interests it is intended to promote is so tenuous that it lacks ... rationality”); and, on gender, Reed v. Reed, 404 U.S. 71, 76 (1971) (question “is whether a difference in the sex of competing applicants ... bears a rational relationship to a state objective....”). This court, indeed, has gone one step further in recognizing candidly that the rational basis test and the test derived from Royster have in some instances been treated as interchangeable, see State v. Deflorio, 128 N.H. 309, 315, 512 A.2d 1133, 1136 (1986).
At least in our State cases, this proven judicial tendency to blur any difference between the two tests doubtless reflects the further fact that the first segment of the compound standard derived from Royster is simply a concise version of the rational basis test, Carson’ s threshold requirement that a challenged classification be “reasonable, not arbitrary,” see Carson v. Maurer, supra at 932, 424 A.2d at 831, being equivalent to the usual first-tier standard that the classification must “rationally relate[ ] to a legitimate State interest,” Boehner v. State, 122 N.H. 79, 84, 441 A.2d 1146, 1149 (1982) (citation omitted). And while the court in Carson may or may not have said the last word in seeming to identify rational basis with a favorable cost-benefit balance, Carson v. Maurer, 120 N.H. at 933, 424 A.2d at 831, the court did make it clear that this prong of the test was deferential to the legislative judgment in question, “the wisdom of or necessity for” which was beyond the scope of judicial review. Id.
Indeed, as the court expressed it at one point, the second-tier test was deferential in its entirety (i.e., in the absence of suspect classi*123fication or fundamental right, courts will not second-guess legislative judgment of need, id.). This pledge of deference is a shaky one, however, thanks to uncertainty over the meaning of the second segment of the standard derived from Royster, requiring a “fair and substantial” relationship between the chosen classification and the legitimate legislative objective. This uncertainty must be seen as a further condition not only facilitating the identification (or misidentification) of the Carson standard with the rational basis test, as we have seen, but also placing temptation in the way of those inclined to impose a far stricter standard in the name of intermediate scrutiny, as we will see below.
While the definitive explanation of “fair and substantial relationship” must apparently await another day, it is fair to say here that if the phrase really is to function as a genuinely middle-tier test, and at the same time defer to ultimate legislative policy judgments, it must be understood as a substantively neutral requirement that the classification in question fit the permitted legislative objective with some minimally acceptable level of precision, or promote it with some like degree of efficiency. It is thus presumably meant to be a test that would strike down a classification barring too many people from invoking a right even when they could do so without compromising the State’s objective, and it might also take into consideration a claimant’s demonstration that a challenged classification would allow too many to invoke the right even when that would be antithetical to the State’s objective. The test, in any event, must be intended to demand that there be some appropriate level of inclusiveness in any classification selected to serve a lawful interest by disparate treatment. On any other view, indeed, the Carson test would simply be an obscurely articulated judicial commission for reviewing the merits of legislation, a role that the Carson court took pains to disclaim any authority to play in the name of intermediate scrutiny. (Whether Carson’s promise of substantively neutral middle-tier review can be a truly practical objective is, of course, another question, which would take me beyond the justifiable scope of a dissent in this case. Here, I am only trying to work with Carson’s reasoning taken at face value.)
Assuming, then, that we are to have a distinctly articulated intermediate test of equal protection review, and assuming that I am right about the function that the “fair and substantial” criterion may perform in such review consistently with Carson’s reasoning, the task confronting the court is to identify the requisite degree of efficiency, *124or fit, that intermediate scrutiny demands. While I would not suggest that this will be a simple job, both advocates before the court and the members of the court itself should confront the difficulty in the earliest possible case, for until the job is attempted, the intermediate nature of the scrutiny will remain elusive. Fortunately, however, the facts of the case before us allow for application of what the Carson opinion claims to be intended, even without a more exactly articulated standard of legislative tailoring than the court has as yet announced.
I reach this conclusion by viewing the instant case from a point on which my colleagues and I agree. The majority opinion follows this court’s earlier recognition that “ ‘there are real and vital differences between the situations of governmental units and of private parties as potential tort defendants,’ Cooperrider, The Court, the Legislature, and Governmental Tort Liability in Michigan, 72 Mich. L. Rev. 187, 272 (1973) (emphasis in original).” Estate of Cargill v. City of Rochester, 119 N.H. 661, 666, 406 A.2d 704, 706 (1979), and I join in the court’s general conclusion that it was open to the legislature to find it inappropriate to hold municipalities to an ordinary negligence standard insofar as they own and maintain highways and sidewalks.
Although this position implies that a municipal immunity defense against some tort claims arising out of sidewalk and highway defects would not necessarily deny equal protection under the standard of intermediate scrutiny, the court nonetheless hypothesizes two circumstances under which application of the municipal immunity conferred by RSA 507-B:2, I, would violate that standard: when the municipality in the course of maintenance or construction created the hazard that eventuated in the plaintiff’s injury, and when the municipality failed to repair a hazardous condition of which an official or employee had actual knowledge with adequate opportunity to act. On the basis of these possibilities the court rules that the classification in issue “is unreasonably broad, is arbitrary, and does not bear, a fair and substantial relation to the legislative objective” of relieving municipalities of negligence liability associated with road and sidewalk construction and maintenance.
I part company from the court here for two reasons. The first is principally analytical, and I will not dwell on it. The court seems to ignore Carson’s emphasis on the composite character of the test, the first segment of which is merely the first-tier rational basis test. Leaving aside the question whether anything is gained by incorporating the first-tier standard into the second-tier, it suffices to say *125that my view of applying Carson’s second segment implies that the statute in question here would pass the rational basis test, and I assume that the court would likewise give it a passing grade on that test alone.
It is over the application of the second segment of Carson’s standard that the substantial disagreement occurs, and to understand that disagreement there is need to look more carefully at the governmental objective that the court assumes to be a legitimate object of legislation. As I noted above, the court speaks of that objective as relief from an unreasonable burden of satisfying an ordinary negligence standard with respect to streets and highways that are “subject to constant and unsupervised public use.” The point of the remainder of the court’s opinion, however, is that relief from liability arising from street and highway maintenance and construction is a legitimate objective only to the extent that the liability would not arise from hazardous conditions created by a municipal employee, officer or contractor, or actually known to a municipal agent. If we accept the court’s view of the permissible objective, then, it is relief from liability stemming from road hazards neither municipally created nor known that is the end to be kept in view when answering the questions whether the means chosen are fairly and substantially related to a legitimate objective.
In assessing those means we must not, of course, view the particular immunity statute in isolation. As the court recognizes to a degree, the statutory immunity from negligence liability under RSA 507-B:2, I, is part of a total scheme, in which other statutes narrow the immunity that § 2, I, would otherwise provide. See RSA 507-B:5. Thus, RSA 231:90 and :91 provide for municipal liability for damage resulting from disrepair or travel hazards on Class IV or V highways and their bridges, if the municipality fails to begin repairs within twenty-four hours after receiving notice of the safety hazard from three or more citizens or taxpayers. RSA 231:92, moreover, subjects municipalities to liability for damage resulting from the unsuitability for travel of any bridge, culvert, sluiceway or dangerous embankment subject to town maintenance. (And, though not directly relevant here, RSA 231:75 and :77 provide for assessment of damages against the town for harm to land caused by grading and ditching a highway.)
In the main, then, there is no municipal immunity from liability when harm results from especially hazardous highway features or risky maintenance chores, or when the town fails to respond promptly to notice of any condition dangerous enough to excite com*126plaints by three people. Conversely, the remaining sources of liability are those most likely to elude municipal notice in the absence of constant systematic inspections of all municipal sidewalks and highways, which the majority agree the legislature could find it unreasonable to require.
The statutory scheme, in other words, classifies victims by barring recovery to those whose damages result from highway and sidewalk conditions and operations that are not notably hazardous per se, and that fail to excite three or more people to provide notice to the municipality. By leaving a town potentially liable in most other instances, the scheme avoids much of the overinclusiveness that might arguably result from absolute immunity from liability for any street or highway hazard, even for highway features with a generally high potential for harm, or about which actual notice might have been given repeatedly. Municipal immunity under this statutory scheme is thus tailored more narrowly than the State immunity that would have been conferred by the bill considered by Opinion of the Justices, 126 N.H. 554, 566-67, 493 A.2d 1182, 1184 (1985) (which is therefore without persuasive authority on the present facts), and it exhibits no apparent underinclusiveness, i.e., in recognizing liability in cases where a municipality cannot fairly be held to the standard appropriate for other potential defendants.
Now, the fit is admittedly not mathematically exact between this total scheme and the objective of immunity only as to hazards not municipally created or known at least to one municipal agent. The majority are correct that the legislature could have narrowed the immunity provisions even further by making a town liable whenever a hazard is known to any of its officials, or whenever the hazard was created in the course of construction and maintenance activities.
There are two reasons, however, why this less-than-mathematically-precise fit should still qualify as fair and substantial. One of those reasons is suggested in the court’s own opinion: if allegations of negligence in construction, or notice to some municipal agent, are enough to counter a municipal immunity claim in the first instance, there will be precious few street and sidewalk accidents without a subsequent lawsuit against a city or town, and the court’s-view of immunity’s proper scope will entail a serious burden of defending against such claims. Certainly the legislature could reasonably choose the scope of immunity embodied in the present statutory scheme in order to obviate the unmeritorious litigation that would attend the narrower concept favored by the court.
*127The second reason for finding a sufficiently fair and substantial fit between the court’s view of legitimate immunity and this statutory scheme is that we have no reason to believe that in practice there would be any large gap between the immunity the court would allow and the immunity the statutory scheme would produce. Because there is no indication of the number or proportion of highway and sidewalk hazards that eventuate from municipal indifference or negligence in construction, there is no indication that the statutory immunity is overinclusive to any significant degree. It is crucial to recognize this, lest we lose sight of the point that we are supposed to be applying an intermediate standard, and should not demand the precise congruence of problem and solution that would be appropriate, say, if a fundamental right recognized by the Constitution were at stake. When, on the contrary, the majority of the court rely on two possibilities of wholly uncertain significance to strike down the statute in issue, their ruling cannot be understood as an instance of intermediate equal protection review, but as one of strict scrutiny wholly inappropriate when the trigger of judicial oversight is the limitation of a right with less than fundamental constitutional significance.
And so the “fair and substantial” relation test is metamorphosed yet again. A formulation that began its juridical life as a rational basis test, and was ostensibly adopted by this court as a standard of intermediate review, is now being applied by a majority of the court to impose the strictest scrutiny known to equal protection analysis. There could be no more striking argument for the need to reexamine the Carson test and the conceptual basis underlying what passes for intermediate review.
Thayer, J., joins in the dissent.