Wade Fitzsimmons Richardson was killed when a two-ton dumptruck driven by Billibob Lewis collided with the car that Richardson was operating. Lewis had become intoxicated at a bar owned by Carnegie Library, Inc.; he subsequently stole the dumptruck from the lot behind Bennett*690Cathey, Inc.; and he negligently drove and crashed the truck into Richardson’s vehicle.
The decedent’s personal representative, Gayle D. Richardson, brought a wrongful death action against Carnegie and BennettCathey. Her complaint alleged that while Lewis was intoxicated, Carnegie served alcohol to him in violation of the Dramshop Act, NMSA 1978, Section 41-11-1 (Repl. Pamp.1986); that Bennett-Cathey negligently left the keys in the ignition of the unattended dumptruck; and that the negligent acts of both defendants proximately caused Richardson’s death. The district court granted summary judgment in favor of Bennett-Cathey, entered a default judgment against Carnegie (for failure to answer), and found that Richardson suffered damages for which he would be entitled to recover $250,000 from Carnegie. The court awarded only $50,000, however, finding itself limited by the maximum recovery allowable under the Dramshop Act.
Richardson appealed the district court’s ruling to the court of appeals, claiming error in the grant of summary judgment to Bennett-Cathey, and attacking the cap on liability under the dramshop act as unconstitutional. Richardson enumerated several “special circumstances” that would justify the imposition of liability against Bennett-Cathey: Bennett-Cathey knew that the brakes on its truck were inoperative; the lot from which Lewis stole the truck was not fenced and was easily accessible; the area where the truck was parked was frequented by transients; the truck required special skills for safe operation; and the truck was large and bulky and more capable of causing serious injuries than an automobile. She asserted that the theft of the unattended vehicle, with keys left in the ignition, was not an independent, intervening, or superseding act that would exempt Bennett-Cathey from liability. The docketing statement also presented the issue that the damage limitation on dram-shop liability violated the United States and New Mexico Constitutions. In her memorandum opposing summary affirmance, Richardson argued that the statute denied equal protection because the damage cap allowed victims of a tavernkeeper’s negligence to be undercompensated although victims of other tortfeasors were entitled to obtain full recovery; and further, that the damage cap violated her right to a trial by jury as guaranteed by Article II, Section 12 of the New Mexico Constitution because the cap usurped the fundamental function of a jury to determine damages.
The court of appeals, by memorandum opinion, upheld the trial court on all issues. In its first calendar notice, the court proposed affirmance of the summary judgment on grounds that the theft was not foreseeable, but instead was an intervening, superseding, criminal act by a- third person. Acknowledging that several jurisdictions look to special circumstances to determine foreseeability of the harm to be caused by the negligence of an owner leaving the keys in an unattended vehicle, and the liability which attends that foreseeability, the appellate court noted that it could not overrule our opinion in Bouldin v. Sategna, 71 N.M. 329, 378 P.2d 370 (1963), by which it felt itself bound. See Alexander v. Delgado, 84 N.M. 717, 718, 507 P.2d 778, 779 (1973) (lower courts should not overrule precedents set by superior court).
Regarding the equal protection issue, the calendar notice considered that the damage cap concerned no fundamental rights and implicated no suspect classes. The court employed, therefore, the rational basis standard of review and looked to the purposes of the challenged statute. It then reasoned that the legislature had “created a cause of action” subsequent to this court’s “creation” of a common-law cause of action for tavernkeeper negligence, and that its purpose was to limit dramshop liability in exchange for creating that new cause of action. Impressed that the damage cap applied equally to all persons seeking to recover under the dramshop act, the court’s notice proposed affirmance on a determination that the statute did not unconstitutionally violate the equal protection clause.
In its second calendar notice, the court of appeals addressed the jury trial issue, and reiterated its conclusion that the legislature had transformed a common-law cause *691of action into a statutory one. It then concluded that Richardson had a right to a trial by jury on the question of liability, but that he had no right to have a jury determine the amount of his damages because the statutory action limited the amount of liability. The court of appeals, therefore, summarily dismissed Richardson’s appeal and affirmed the trial court by memorandum opinion.
We granted Richardson’s petition for .writ of certiorari and gave leave to file amicus curiae briefs to the New Mexico Trial Lawyers Association (NMTLA), the Defense Lawyers Association (DLA), Mothers Against Drunk Drivers (MADD), and Students Against Drunk Drivers (SADD). The only issue addressed by all of the amici briefs in support of the petitioner’s application for review is the constitutionality of the dramshop act.
MADD and SADD point out that New Mexico has one of the most severe drunk-driving problems in the United States and that every conceivable approach to resolve the drunk-driving menace is needed. They agree that dramshop liability is an effective measure in curbing drunken driving, but that the salutary impact of the dramshop act is diffused by the damage cap. Urging that the limit on recovery is inconsistent with the purpose for imposing liability, they emphasize that reinforcing dramshop liability and invalidating the damage cap would best serve the public interest.
NMTLA challenges the damage cap as unconstitutional violations of the due process and equal protection clauses, the right to trial by jury, and the doctrine of separation of powers. Regarding the separation of powers argument, NMTLA contends that the legislatively mandated damage cap prevents judges from exercising their historic procedural power to exercise discretion in reviewing the excessiveness of a jury’s award upon a motion for a new trial under SCRA 1986, 1-059; and that it compels judges to order a remittitur, another discretionary act historically inherent in a trial judge’s powers. Because procedural rules are within the sole domain of the supreme court, and because the statutory limitation on liability impinges upon the provisions of Rule 59, NMTLA insists with some logic that the damage cap constitutes a legislative usurpation of judicial power and thus violates the doctrine of separation of powers.
NMTLA proposes, too, that the damage cap violates the right to trial by jury, arguing that the legislature did not transform dramshop liability from an action at common law to a statutory cause of action but, rather, only narrowed and modified the judicially-created common-law liability that was established in Lopez v. Maez, 98 N.M. 625, 651 P.2d 1269 (1982). Thus, because a plaintiff has a fundamental right to have a jury determine liability and damages in a common-law action, NMTLA argues that the damage cap unconstitutionally infringes on a party’s right to trial by jury.
Amicus NMTLA contends that the damage cap also infringes the due process and equal protection clauses of the New Mexico Constitution because the rights to a jury trial and of access to the courts, being fundamental constitutional rights, are rendered meaningless if full and adequate recoveries are not available to all plaintiffs. NMTLA has traced these rights from their historic geneses in Spanish and Mexican laws, the Siete Partidas, the Fuero Juzgo, and the Kearney Code, and it concludes that all formed an integral part of the civil law in the days before statehood and were incorporated into the New Mexico Constitution. Accordingly, NMTLA argues that the damage cap is subject to strict scrutiny under which it surely must be invalidated, but that even under the minimal rational basis analysis, no justification exists to uphold a limitation on the award of damages.
DLA, from an opposing position, argues that Richardson has no standing to raise the jury trial issue because she never requested a jury trial. On the issue of recoverable damages, it takes the stance that the legislature changed dramshop liability to a statutory action from one at common law and, thus, because it created the liability, it can limit the amount of recovery. We are urged to disregard the separation of powers, due process, and right of access to the *692courts issues because they were not briefed or argued at trial and, therefore, are not properly before us for consideration. But DLA does respond to some of the amici arguments, urging that the separation of powers doctrine is inapplicable here because the judiciary promulgates procedural rules out of convenience and efficiency only; and the right of access to the courts merely refers to the availability of the judicial machinery to resolve disputes and is not a right guaranteed explicitly in the New Mexico Constitution.
Regarding the equal protection issue, DLA asserts that we are not here dealing with fundamental rights or suspect classes. Any rights found in the Kearney Code or any other civil law predating statehood, it says, were not adopted by or incorporated into the New Mexico Constitution. Moreover, characterizing the dramshop act as social and economic legislation reviewable by the rational basis test, it denies that any right to full compensation can be implied from the guarantee of certain inalienable rights in Article II, Section 4 of our constitution. DLA views the damage cap as rationally related to the dual legislative goals of compensating victims injured as the result of the negligent service or sale of alcohol but not overburdening tavernkeepers, conjecturing that recovery under the dramshop act probably will not be the only source of recovery available to such a plaintiff.
Responding to whether Richardson properly preserved certain constitutional issues for appeal, NMTLA points to Richardson’s broad claim that the damage cap was unconstitutional, which opened the door for amici to explain in more detailed and specific analyses under the right to jury trial, ■due process, and equal protection clauses of the New Mexico Constitution, exactly why the statute is invalid. But DLA is correct in asserting that two of the issues, separation of powers and due process, cannot be raised for the first time on appeal. See Romero v. Sanchez, 86 N.M. 55, 56, 519 P.2d 291, 292 (1974) (court will not consider claim offered for first time on appeal); State ex rel. Brown v. Hatley, 80 N.M. 24, 25, 450 P.2d 624, 625 (1969) (same). It is not enough for a party to make a broad, general assertion that a statute is unconstitutional and then leave it to amici to develop and refine her arguments. The complainant must specify in what manner his or her constitutional rights are affected adversely. State v. Hines, 78 N.M. 471, 474, 432 P.2d 827, 830 (1967). Richardson did not request resolution, in either the trial court or the court of appeals, of the separation of powers and due process claims raised by NMTLA, and we will not consider new issues presented for the first time on appeal through amicus briefs. St. Vincent Hosp. v. Salazar, 95 N.M. 147, 149, 619 P.2d 823, 825 (1980).
DLA is likewise correct in observing that Richardson did not request a jury trial. Failure to demand a jury trial in a timely manner constitutes a waiver of a trial by jury. SCRA 1986, 1-038(D). Even though this issue was addressed by the court of appeals in its notices of proposed affirmance, we will not consider in this review the jury trial issue. We do not consider, therefore, whether the limit on dramshop liability violates Richardson’s right to a jury’s determination of damages in a common-law cause of action.
NMTLA, however, correctly analyzes right of access to the courts as an implicit fundamental right entitled to the equal protection guarantees of the Constitution. Richardson claimed an equal protection violation in the court below; consequently, whether a right of access to the courts is violated by the damage cap is a relevant question in determining whether a fundamental right is contravened. That, in turn, is likewise relevant in determining which standard of review to apply in analyzing the equal protection right guaranteed by Article II, Section 18 of the New Mexico Constitution.
Recently we discussed, in Meyer v. Jones, 106 N.M. 708, 749 P.2d 93 (1988), that in equal protection attacks upon statutes, at least three tests for reviewing such challenges have been recognized and applied. Traditionally, the United States Supreme Court long had employed a two-ti*693ered analysis: minimum scrutiny, or the rational basis test, when reviewing social and economic legislation, McGowan v. Maryland, 366 U.S. 420, 81 S.Ct. 1101, 6 L.Ed. 2d 393 (1961), and strict scrutiny when analyzing legislation that infringed fundamental constitutional rights or made distinctions directed toward suspect classes.1 The tests for reviewing equal protection challenges generally are the same under New Mexico and federal law.2
We have observed that a statute infringing fundamental rights or involving suspect classes must support a compelling state interest to escape judicial invalidation. State v. Edgington, 99 N.M. 715, 718, 663 P.2d 374, 377 (Ct.App.), cert. denied, 99 N.M. 644, 662 P.2d 645, cert. denied, 464 U.S. 940, 104 S.Ct. 354, 78 L.Ed.2d 318 (1983). We have also said that legislative acts are presumptively valid and normally are subjected to the rational basis test; it is well-settled that they will not be declared invalid unless the court is clearly satisfied that the legislature went outside the constitution in enacting them. Espanola Hous. Auth. v. Atencio, 90 N.M. 787, 788, 568 P.2d 1233, 1234 (1977); Board of Trustees v. Montano, 82 N.M. 340, 343, 481 P.2d 702, 705 (1971). The burden of proof is on the plaintiff to demonstrate that the challenged legislation is clearly arbitrary and unreasonable, not just that it is possibly so. Sanchez v. M.M. Sundt Constr. Co., 103 N.M. 294, 296, 706 P.2d 158, 160 (Ct.App. 1985); Gallegos v. Homestake Mining Co., 97 N.M. 717, 722, 643 P.2d 281, 286 (Ct. App.1982). The fact that a statute appears unreasonable to the courts is not decisive; that is not enough to invalidate an act. Hutcheson v. Atherton, 44 N.M. 144, 149, 99 P.2d 462, 465 (1940). Only when a statutory classification is so devoid of rational support or serves no valid governmental interest, so that it amounts to mere caprice, will it be struck down under the rational basis test. Montano, 82 N.M. at 343, 481 P.2d at 705; Hutcheson, 44 N.M. at 149, 99 P.2d at 465; Edgington, 99 N.M. at 719, 663 P.2d at 378. When employing the minimal scrutiny test, the courts neither will inquire into the wisdom, policy, or justness of legislation, nor will they substitute their views for that of the legislature, but rather will uphold the statute if any state of facts reasonably can be conceived that will sustain the challenged classification. Garcia v. Albuquerque Pub. Schools Bd. of Educ., 95 N.M. 391, 393, 622 P.2d 699, 701 (Ct. App.1980), cert. quashed, 95 N.M. 426, 622 P.2d 1046 (1981). The rational basis test, therefore, employs no independent review or analysis of the factual basis of the state’s goal, or of the means designated by the statute to attain that goal. Nowak, Realigning the Standards of Review Under the Equal Protection Guarantee— Prohibited, Neutral, and Permissive Classifications, 62 Geo.LJ. 1071, 1094 (1974).
An intermediate equal protection standard of review, somewhere between the rational basis and strict scrutiny standards, arose more recently out of the Supreme Court's dissatisfaction with the traditional, two-tiered analysis. See Craig v. Boren, 429 U.S. 190, 210 n. *, 97 S.Ct. 451, 463 n. *, 50 L.Ed.2d 397 (1976) (Powell, J., concurring); Gunther, In Search of Evolving Doctrine on a Changing Court: A Model for a Newer Equal Protection, 86 Harv.L. Rev. 1, 17-19 (1972). Accordingly, the third test has been aimed at legislative classifications infringing important but not fundamental rights, and involving sensitive but not suspect classes. L. Tribe, American Constitutional Law § 16-33, at 1610, 1613 (2d ed. 1988). The Court first enunciated the intermediate (or “heightened scrutiny”) test in F.S. Royster Guano Co. v. Virginia, 253 U.S. 412, 415, 40 S.Ct. 560, 561, 64 L.Ed. 989 (1920), when it declared that a classification must be reasonable, *694not arbitrary, and must rest upon some ground of difference having a fair and substantial relation to the object of the legislation. See Reed v. Reed, 404 U.S. 71, 76, 92 S.Ct. 251, 254, 30 L.Ed.2d 225 (1971) (employing Royster’s intermediate scrutiny test to invalidate statute based on gender classification); see also City of Cleburne v. Cleburne Living Center, Inc., 473 U.S. 432, 441, 105 S.Ct. 3249, 3255, 87 L.Ed.2d 313 (1985) (under heightened standard of review, classification fails unless it is substantially related to sufficiently important or legitimate governmental interest). The Court has applied the intermediate analysis principally to statutes that classify according to gender and illegitimacy. See City of Cleburne, 473 U.S. at 440-41, 105 S.Ct. at 3254-55.
Although we have referred to the Supreme Court’s use of the third, intermediate standard of review, see McGeehan v. Bunch, 88 N.M. 308, 310, 540 P.2d 238, 240 (1975), on occasion we have muddied the constitutional waters in New Mexico by interchangeably using the rational basis and intermediate tests as if they were identical. For example, in McGeehan, the court considered the validity of an automobile guest statute, construed the act as social and economic legislation, and cited the applicable standard of review as the intermediate test that was enunciated in Reed. See McGeehan, 88 N.M. at 310, 540 P.2d at 240. The court described the facets of the rational basis test, declared the legislative classification unreasonable and arbitrary, id. at 311, 540 P.2d at 241, but said also that the statute had no “fair and substantial relation” to its goal. Id. at 313-14, 540 P.2d at 244. In the end, the court invalidated the guest statute as violative of the equal protection clause, but it is not clear on which standard of review it relied to do so; and if the court employed both the rational basis and intermediate tests to strike the statute, the opinion is not clear why the court used both instead of either.
The imprecision was perpetuated in Pruey v. Department of Alcoholic Beverage Control, 104 N.M. 10, 715 P.2d 458 (1986). In considering an equal protection challenge to regulations prohibiting the sale of alcohol on Sundays, the court quoted the rational basis test as outlined in McGowan, and then cited the intermediate test and Reed and McGeehan in support. Id. at 12, 715 P.2d at 460. The court seemed to consider the two tests as different manifestations of the same principle; but the court upheld the statute as having a rational basis. Id. at 13, 715 P.2d at 461.
The confusion probably is a result of a misinterpretation of the longstanding precedent that legislative classifications must be based upon substantial distinctions. See State v. Atchison, T. & S.F. Ry., 20 N.M. 562, 570, 151 P. 305, 307 (1915). That rule is found in Gruschus v. Bureau of Revenue, 74 N.M. 775, 399 P.2d 105 (1965), an opinion often cited for its explication of the rational basis standard of review, wherein it was said that the equal protection clause “does not prohibit classification for legislative purposes, provided that there is a rational and natural basis therefor, that it is based on a substantial difference between those to whom it does and those to whom it does not apply * * Id. at 778, 399 P.2d at 107. In Gruschus, the challenged statute was found reasonable and not arbitrary, affording substantially equal treatment to all persons similarly situated. Id. at 779, 399 P.2d at 108. The test might better be stated as one assuring that classifications are based on real differences bearing a rational and proper relationship to the classification. See Community Pub. Serv. Co. v. New Mexico Pub. Serv. Comm’n, 76 N.M. 314, 317-18, 414 P.2d 675, 677, cert denied, 385 U.S. 933, 87 S.Ct. 292, 17 L.Ed. 2d 213 (1966); Burch v. Foy, 62 N.M. 219, 224, 308 P.2d 199, 202 (1957). The Española Housing court said that the question is whether the reasons advanced for validity of a statute were “real and pertinent differences or merely artificial differences * * * not relevant to the classification involved.” 90 N.M. at 789, 568 P.2d at 1235.
Thus, the rational basis test, which requires classifications to be based on substantial or real distinctions and be rationally related to the legislative goal, is *695different from the intermediate test, which requires a classification to be more than just rationally related to the statutory purpose; it requires also that the classification be substantially related to an important state interest. Additionally, a key difference in the tests is that under rational basis the party objecting to the legislative classification has the burden of demonstrating that the classification bears no rational relationship to a conceivable legislative purpose whereas under heightened scrutiny the party maintaining the validity of the classification must prove that the classification is substantially related to an important governmental interest. See Craig v. Boren, 429 U.S. 190, 97 S.Ct. 451, 50 L.Ed.2d 397 (1976).
Our research discloses that only the Supreme Court of Minnesota has considered the constitutionality of a limitation on damages for dramshop liability. Employing the rational basis test, McGuire v. C & L Restaurant, Inc., 346 N.W.2d 605 (Minn. 1984), invalidated the liability limitations as violative of the state equal protection clause. Id. at 613. In similar challenges to medical malpractice damage caps, however, several jurisdictions have considered the equal protection argument. We have found those cases most instructive in that for all practical purposes the constitutional analysis of medical malpractice limited liability legislation is identical to an equal protection analysis of limited dramshop liability. Thus, we discuss some of those decisions.
Some courts have construed the damage caps as social and economic legislation and have upheld them after reviewing the legislation under the rational basis test.3 Three separate intermediate courts in Texas have invalidated legislation that limited liability for medical malpractice actions, purportedly using the rational basis standard of review.4 The Supreme Court of Texas, also applying the minimum standard in a yet-unreleased opinion, recently affirmed the invalidity of the statutory liability limitation in Lucas v. United States, 757 S.W.2d 687 (1988), holding the cap to be “unreasonable and arbitrary” when balanced against the purpose and basis of the legislation. Id. at 690. In Lucas, the Texas Supreme Court had no difficulty in determining under the rational basis standard that an unreasonable and arbitrary cap on medical malpractice damages was an unconstitutional denial of a “ ‘remedy by due course of law.’ ” 757 S.W.2d at 690.
Three other jurisdictions have invalidated damage limitation provisions as violative of a plaintiff’s explicit, state constitutional right to full recovery in a tort action, employing a strict scrutiny analysis to do so. See Kenyon v. Hammer, 142 Ariz. 69, 83, 688 P.2d 961, 975 (1984); Smith v. Department of Ins., 507 So.2d 1080, 1088 (Fla. 1987); Pfost v. State, 713 P.2d 495, 503 (Mont.1985). Several other courts have held that the right to recover damages for personal injuries is not a fundamental right and that the class of victims denied full recovery is not a suspect class; but those courts have declared further that the rights infringed by medical malpractice legislation are sufficiently important and substantive, and the class of persons affected sufficiently sensitive, to justify invoking an intermediate standard of review to invalidate the statutes.5
*696In determining which standard of review to apply in the equal protection analysis of the damage cap in the present case, we note first that no suspect class is involved. A suspect class has been defined as a discrete group “saddled with such disabilities, or subjected to such a history of purposeful unequal treatment, or relegated to such a position of political powerlessness as to command extraordinary protection from the majoritarian political process.” San Antonio Independent School Dist. v. Rodriguez, 411 U.S. 1, 28, 93 S.Ct. 1278, 1294, 36 L.Ed.2d 16 (1973); see Lyng v. Castillo, 477 U.S. 635, 106 S.Ct. 2727, 2729, 91 L.Ed. 2d 527 (1986); Plyler v. Doe, 457 U.S. 202, 216 n. 14, 102 S.Ct. 2382, 2394 n. 14, 72 L.Ed.2d 786 (1982). Only statutory classifications based on race, national origin, or alienage so far have been treated as suspect. Wilkinson, The Supreme Court, The Equal Protection Clause, and the Three Faces of Constitutional Equality, 61 Va. L.Rev. 945, 951 (1975). The class of tort victims denied full recovery for dramshop liability does not rise to the level of “suspectness” under existing precedent and, therefore, does not trigger strict scrutiny.
Secondly, we address the question of whether the damage cap infringes upon any fundamental constitutional rights. A fundamental right is that which the Constitution explicitly or implicitly guarantees. Rodriguez, 411 U.S. at 33-34, 93 S.Ct. at 1296-1297. The petitioner and amicus NMTLA argue that the damage cap violates her right of access to the courts and her right to full recovery in tort. Neither of these “rights” is guaranteed explicitly in our constitution. We have declared, however, that the right of access to the courts is one aspect of the right to petition for redress of grievances, and we have acknowledged that right as one guaranteed by the first amendment to the federal constitution and also protected by both the United States and New Mexico Constitutions by the prohibitions against “depriving a person of life, liberty or property without due process of law.” Jirón v. Mahlab, 99 N.M. 425, 426, 659 P.2d 311, 312 (1983). We once again recognized a “plaintiffs constitutional right to petition for redress” in Otero v. Zouhar, 102 N.M. 482, 486, 697 P.2d 482, 486 (1985).
With regard to whether the right to full recovery reaches fundamental status, the argument is that Article II, Section 4 of the New Mexico Constitution guarantees a fundamental right to be compensated fully and adequately for injuries that result from negligent behavior. That provision reads: “All persons are born equally free, and have certain natural, inherent and inalienable rights, among which are the rights of enjoying and defending life and liberty, * * * and of seeking and obtaining safety and happiness.” N.M. Const, art II, § 4. Some commentators assert that this is “textual evidence of an intent on the part of the constitutional ratifiers to afford substantive protection against the power of the state to impair economic interests.” Developments in The Law — The Interpretation of State Constitutional Rights, 95 Harv.L.Rev. 1324, 1480 (1982). But because we do not think it necessary, we decline at this time to interpret this provision as implicitly guaranteeing a fundamental right to full recovery in tort actions, so as to trigger a strict scrutiny analysis.
Acknowledging, also, our recognition of a constitutional right of access to our courts, we do not apply strict scrutiny to the issue of full recovery, principally because we conclude that the damage cap is constitutionally invalid under the lesser, intermediate scrutiny test. It is thus unnecessary to impose the highest level of review.
We are aware that in the history of the interpretation of the federal equal protection clause, the rational basis test generally has been minimal scrutiny in theory and has amounted to virtual judicial abdication in fact, whereas maximum scrutiny has been strict in theory and almost always fatal in fact. Gunther, 86 Harv.L.Rev. at 8. Strict scrutiny has operated as an anti*697majoritarian safeguard. Tribe, § 16-31, at 1588; Learner, Restrictive Medical Malpractice Compensation Schemes: A Constitutional “Quid Pro Quo” Analysis to Safeguard Individual Liberties, 18 Harv. J. on Legis. 143, 152 (1981). Accordingly, the application of the strict scrutiny test has resulted in the virtual immunization of certain liberties from legislative affliction. “The very purpose of a Bill of Rights was to withdraw certain subjects from the vicissitudes of political controversy, to place them beyond the reach of majorities and officials and to establish them as legal principles to be applied by the courts. One’s * * * fundamental rights may not be submitted to vote; they depend on the outcome of no elections.” West Virginia State Bd. of Educ. v. Barnette, 319 U.S. 624, 638, 63 S.Ct. 1178, 1185, 87 L.Ed. 1628 (1943).
By contrast, the rational basis test affords minimal scrutiny because of the concept that “it is constitutionally appropriate to ‘fight out the wise use of legislative authority in the forum of public opinion and before legislative assemblies rather than to transfer such a contest to the judicial arena,’ since all the ‘effective means of inducing political changes are left free.’ ” Id.; see Cleburne, 473 U.S. at 441-42, 105 S.Ct. at 3255-56 (because of doctrine of separation of powers, courts should be reluctant to closely scrutinize economic and social legislation, but rather should employ rational basis test). The primary theoretical basis for deferring to the legislature when applying the rational basis test, then, is that political entities can respond best to the electorate and can experiment with and allocate the state’s often limited resources in a manner that best reflects the concerns of their constituencies over social and economic issues. See Redish, Legislative Response to the Medical Malpractice Insurance Crisis: Constitutional Implications, 55 Tex.L.Rev. 759, 761 (1977); Tussman & tenBroek, The Equal Protection of the Laws, 37 Calif.L.Rev. 341, 366 (1949). In our own jurisprudence, we also have observed that courts should be hesitant to overturn a statute other than on fundamental rights grounds because the separation of powers doctrine mandates deference to a legislative determination of reasonableness. Edgington, 99 N.M. at 718, 663 P.2d at 377.
In advancing the intermediate test as a third level of review, the Supreme Court has “recognized that certain forms of legislative classification, while not facially invidious, nonetheless give rise to recurring constitutional difficulties” and in those “limited circumstances” the Court seeks “assurance that the classification reflects a reasoned judgment consistent with the ideal of equal protection by inquiring whether it may fairly be viewed as furthering a substantial interest of the State.” Plyler, 457 U.S. 202, 217-18, 102 S.Ct. 2382, 2395. And although even the Supreme Court has presented the heightened scrutiny test in a myriad of fashions, it has been characterized, in whatever form, at least by a “sharper focus” on legislative classifications “poised between the largely toothless invocation of minimum rationality and the nearly fatal invocation of strict scrutiny.” Tribe, § 16-32, at 1601; see Cleburne, 473 U.S. at 451, 105 S.Ct. at 3260 (Stevens, J., concurring) (standards of review for equal protection challenges reflect “a continuum of judgmental responses to differing classifications which have been explained in opinions by terms ranging from ‘strict scrutiny’ at one extreme to ‘rational basis’ at the other”).
Some critics have said that when courts elect to apply the intermediate test, they abandon judicial objectivity and make subjective judgments that lack constitutional support, thereby succumbing to the temptation to usurp the legislature’s function by making highly political decisions about certain social and economic issues. See, e.g., Redish, 55 Tex.L.Rev. at 782; Note, Fein v. Permanente Medical Group: Future Trends in Damage Limitation Adjudication, 80 Nw.U.L.Rev. 1643,1663-64 & 1673 (1986). But judicial scrutiny always requires judgments about legislative decisions, and that is particularly so when heightened scrutiny is called for. See Cleburne, 473 U.S. at 443, 105 S.Ct. at 3256; Coburn, 627 F.Supp. at 991. See generally Haines, General Observations on the *698 Effects of Personal, Political, and Economic Influences in the Decisions of Judges, 17 Ill.L.Rev. 96, 112-14 (1922). We do not consider that the intermediate constitutional review process necessarily constitutes “abandonment” of any judicial responsibilities but, instead, hones the indispensable requirement of detached analytical examination of competing interests between legislative power and constitutional restraints.
To support our application of the intermediate test we are impressed with Professor Tribe’s observation that the heightened, intermediate standard of review is a judicial response to an awareness that the
all-or-nothing choice between minimum rationality and strict scrutiny ill-suits the broad range of situations arising under the equal protection clause, many of which are best dealt with neither through the virtual rubber-stamp of truly minimal review nor through the virtual death-blow of truly strict scrutiny, but through methods more sensitive to risks of injustice than the former and yet less blind to the needs of governmental flexibility than the latter. [Emphasis added.]
Tribe, § 16-32, at 1609-10; see Cleburne, 473 U.S. at 460,105 S.Ct. at 3265 (Marshall, J., dissenting) (“level of scrutiny employed in an equal protection case should vary with ‘the constitutional and societal importance of the interest adversely affected and the recognized invidíóúsness of the basis upon which the particularXclassification is drawn’ ”). We agree that implementing, the intermediate test in appropriate circumstances narrows the wide 'gap between strict and minimal scrutiny, “not by abandoning the strict but by raising the level of the minimal from virtual abdication to genuine judicial inquiry.” Gunther, 86 Harv.L. Rev. at 24.
It is clear from the foregoing discussion that the limitation of a full tort recovery at issue here under Section 41-11-1(1) implicates a substantial and important individual interest. For substantial and important individual interests, we invoke an intermediate standard of review because we think it best strikes the balance between the legislature’s constitutional prerogative to deliberate over and counterbalance the variety of interests involved in social and economic issues, and the judiciary’s constitutional responsibility to strictly scrutinize legislation that either infringes upon fundamental rights or impacts upon suspect classes. Viewing this constitutional balance within the separation of powers context, which is the gist of opposition to it, we are satisfied that we neither trample arbitrarily upon the legislature’s preferred position of direct, political accountability to the electorate, nor do we forsake our duty to protect individuals from the deleterious effects of controversial social and economic legislation that, in this case at least, could result in economic devastation of innocent victims simply by the fortuitous happenstance .of the tortfeasor’s status. We see no usurpation of power in a heightened scrutiny of legislation in those limited circumstances when the class implicated is so sensitive to injustice and the rights affected are so substantial and important that they warrant special judicial attention.
Section 41-11-1 tacitly makes three separate classifications: a class of victims suffering from injuries resulting from the negligence of a tavernkeeper as distinguished from victims of another tortfeasor’s negligent conduct; a class of victims suffering from the negligence of tavernkeepers whose injuries amount to less than $50,000 lumped together with those victims whose injuries resulting from the same cause are in excess of that damage limitation; and a class of tortfeasors accorded the benefit of the $50,000 cap as distinguished from all other tortfeasors, most of whom are liable for the full amount of damages they cause. We believe that these classifications effect a substantial injustice in this case. The classifications infringe an individual’s important interest to be compensated fully for his injuries, especially when, as is alleged in the instant case, they are a result of no fault of his own. This interest, in our view, certainly is amply important and substantial to justify the invocation of at least the heightened, intermediate test instead of the minimum rationality test. We are per*699suaded also that the class of tort victims affected by the damage cap is “sensitive” enough to the injustice wrought to warrant applying the heightened test. Consequently, we take the intermediate approach and analyze the constitutional challenge in this case under heightened scrutiny.
We commence our examination by repeating that the court of appeals erred in its equal protection analysis of the damage limitation. A legislative classification not only must affect equally all persons within the class to which the legislation applies but, to begin with, the legislature must have a legitimate purpose for creating the class, and a constitutionally permissible reason for treating persons within that class differently from those without. See McLaughlin v. Florida, 379 U.S. 184, 190, 85 S.Ct. 283, 287, 13 L.Ed.2d 222 (1964). In light of those considerations, the court of appeals erred in concluding that the damage cap did not violate the equal protection clause because it applied equally to all persons affected by the dramshop act. “Judicial inquiry under the Equal Protection Clause * * * does not end with a showing of equal application among the members of the class defined by the legislation.” Id. at 191, 85 S.Ct. at 288. No argument has been presented to us to persuade us that the classifications created by the legislation are constitutionally legitimate and, under the McLaughlin dictate, we have been unable to discern or discover any by our own reasoning processes.
Plaintiff has presented a prima facie showing of an arbitrary and unreasonable denial of equal protection and of a restriction on a plaintiffs right of access to the courts. On the other hand, respondent completely failed to carry its burden of demonstrating that any substantial interest of the state is furthered by the legislation. In the absence of any contrary showing by respondent, we cannot think of legitimate public good or supportive policy reasons that are promoted by the special protection of tavernkeepers in the dispensation of intoxicating liquor. We are distinctly unable to rationalize a legitimate or substantial reason for limiting the liability of a tavernkeeper who has a duty not to place drunks behind the wheel of a vehicle on the highway when, by contrast, a rancher or farmer is fully liable for negligently allowing his livestock to meander dumbly into the path of oncoming vehicles. See NMSA 1978, §§ 30-8-13 (Repl.Pamp.1984) & 66-7-363 (Repl.Pamp.1987).
Even though we agree that the legislature most often is better suited to make such policy determinations, a heightened scrutiny of legislation that infringes substantial and important individual interests, such as we have here, compels us to the conviction that the liability cap works a manifest injustice on innocent tort victims and lacks any of the redeeming features entitling it to constitutional validity. Absolutely nothing was shown sufficient to overcome plaintiffs arguments, or to demonstrate that the damage limitation in Section 41-11-1(1) has a substantial relationship to a legitimate or important governmental purpose and we have been unable to fathom one. The cap on damages mandated by Section 41-11-1(1) simply does not withstand heightened scrutiny, and we hold it to be constitutionally invalid as violative of the equal protection clause.
Turning to the other issue, we acknowledge that, as in Bouldin v. Sategna, 71 N.M. 329, 378 P.2d 370 (1963), a substantial number of courts has not held owners liable for leaving the keys in their unattended vehicles and for the injuries to third persons as a result of the thefts and subsequent negligent operation of those vehicles. Those courts have concluded either that an owner owes no duty to the general public to guard against the risk of a thief’s negligent operation of a vehicle in which the owner left his keys; that the theft and subsequent negligence of the thief could not reasonably be foreseen by the owner as a natural or probable consequence of leaving the keys in the ignition of the car; or have concluded that even if the owner was negligent, his actions were not the proximate cause of the injuries because the *700thief’s actions constituted an independent, intervening cause.6
An emerging group of jurisdictions, on the other hand, has rejected the contention that an intervening criminal act automatically breaks the chain of causation as a matter of law, concluding instead that a reasonable person could foresee a theft of an automobile left unattended with the keys in the ignition and reasonably could foresee the increased risk to the public should the theft occur.7 In addition, a few courts, including some of those that earlier denied liability, have indicated a willingness to impose liability upon the owner under “special circumstances.”8 Courts looking at special circumstances seek to determine whether an owner’s conduct enhanced the probability that his car would be stolen and thus increased the hazard to third persons. Considering special circumstances, then, is just another way of examining the degree of foreseeability of injury and whether the owner is subject to a duty to exercise reasonable care.9 Vadala v. Henkels & McCoy, Inc., 397 A.2d 1381 (Del.Super. 1979), listed some of the circumstances aiding the court in its resolution of a similar case:
(a) the vehicle in question is of a type which may attract potential intermeddlers who are unlikely to have the necessary knowledge and skill to operate it safely;
(b) that vehicle is capable of inflicting more serious injury and damage than an ordinary vehicle when not properly controlled;
(c) no security measures were taken after it became evident that the lock which secures the gate to the truck yard had been partially cut and an *701intoxicated individual was loitering nearby. * * *
NMSA 1978, Section 66-7-353, which prohibits leaving a motor vehicle to stand unattended without “first stopping the engine, locking the ignition, [and] removing the key,” was enacted for the purpose of promoting public welfare and safety. Bouldin, 77 N.M. at 332, 378 P.2d at 372. Prevention of the kinds of unfortunate circumstances that occurred in this case from failure to comply with the statute would be conducive to promoting public safety. When “a person by his own negligence produces a dangerous condition of things, which does not become active for mischief until another person has operated upon it by the commission of another negligent act, which might not unreasonably be foreseen to occur, the original act of negligence is then regarded as [a] proximate cause of the injury which finally results.” Thompson v. Anderman, 59 N.M. 400, 412, 285 P.2d 507, 515 (1955).
Some of the members of this Court believe that our adoption of comparative negligence as the rule of law in this jurisdiction, Scott v. Rizzo, 96 N.M. 682, 634 P.2d 1234 (1981), commits us to the principles there expressed that if a jury finds more than one party to have been negligent, a verdict “requiring wrongdoers to share the losses caused, at the ratio of their respective wrongdoing, * * * fairly distributes the burden of fault” and “holds all parties fully responsible for their own respective acts to the degree that those acts have caused harm.” Id. at 689-90, 634 P.2d at 1241-42. See, e.g., St. Sauver v. New Mexico Peterbilt, Inc., 101 N.M. 84, 87, 678 P.2d 712, 715 (1984); Ramirez v. Armstrong, 100 N.M. 538, 542-43, 673 P.2d 822, 827-28 (1983); Bartlett v. New Mexico Welding Supply, Inc., 98 N.M. 152, 158-59, 646 P.2d 579, 585-86 (Ct.App.), cert. denied, 98 N.M. 336, 648 P.2d 794 (1982). There is a divergence in the opinions of members of this Court, however, whether questions of fact are presented in any inquiry into whether an owner reasonably could foresee that his vehicle might be stolen if he left it unattended, unlocked, and with the keys in its ignition, and whether he reasonably could anticipate that the thief might drive negligently and injure someone, see Ney v. Yellow Cab Co., 2 Ill.2d 74, 83, 117 N.E.2d 74, 80 (1954), or whether Bouldin was correct in holding, as a matter of law, that such ensuing theft and subsequent negligence resulting in injury were not natural, foreseeable events attendant upon leaving one’s keys in the vehicle.
Consequently, a majority of the Court being unable to reach agreement on the Bouldin issue at this time, we do not disturb the summary judgment entered by the trial court in favor of Bennett-Cathey.
We remand the case to the district court for entry of judgment in the amount of $250,000 against defendant Carnegie. IT IS SO ORDERED.
SOSA, Senior Justice, concurs.
RANSOM, J., specially concurring.
STOWERS, J., dissenting.
SCARBOROUGH, C.J., not participating.