Miron v. Trudel, [1995] 2 S.C.R. 418

John O. Miron and Jocelyne Valliere Appellants

v.

Richard Trudel, William James McIsaac

and the Economical Mutual Insurance Company Respondents

and

The Attorney General of Canada,

the Attorney General for Ontario,

the Attorney General of Quebec and

the Attorney General of Manitoba Interveners

Indexed as: Miron v. Trudel

File No.: 22744.

1994: June 2; 1995: May 25.

Present: Lamer C.J. and La Forest, L'Heureux‑Dubé, Sopinka, Gonthier, Cory, McLachlin, Iacobucci and Major JJ.

on appeal from the court of appeal for ontario

Constitutional law ‑‑ Charter of Rights ‑‑ Equality rights ‑‑ Automobile insurance ‑‑ Standard automobile policy prescribed by provincial legislation extending accident benefits to "spouse" of policy holder ‑‑ Term "spouse" not including unmarried common law spouse ‑‑ Whether limitation of benefits to married persons violates s. 15(1) of Canadian Charter of Rights and Freedoms ‑‑ If so, whether violation justifiable under s. 1 of Charter ‑‑ Insurance Act, R.S.O. 1980, c. 218, ss. 231, 233, Schedule C.

Constitutional law ‑‑ Charter of Rights ‑‑ Enforcement ‑‑ Appropriate remedy ‑‑ Standard automobile insurance policy prescribed by provincial legislation extending accident benefits to "spouse" of policy holder ‑‑ Term "spouse" not including unmarried common law spouse ‑‑ Limitation of benefits to married persons violating s. 15(1) of Canadian Charter of Rights and Freedoms ‑‑ Violation not justifiable under s. 1 of Charter ‑‑ Whether Court should retroactively "read in" more inclusive definition of "spouse" under s. 24 of Charter.

The appellants lived together with their children. While they were not married, their family functioned as an economic unit. In 1987, M was injured while a passenger in an uninsured motor vehicle driven by an uninsured driver. After the accident, the appellant M could no longer work and contribute to his family's support. He made a claim for accident benefits for loss of income and damages against V's insurance policy, which extended accident benefits to the "spouse" of the policy holder. The respondent insurer denied his claim on the ground that M was not legally married to V and hence not her "spouse". The appellants sued the insurer. The insurer brought a preliminary motion to determine whether the word "spouse", as used in the applicable portions of the policy, includes unmarried common law spouses. The motions court judge found that "spouse" meant a person who is legally married. The appellants appealed the decision to the Court of Appeal, arguing first that M is a spouse under the terms of the policy, and alternatively, that the policy terms, which are those of the standard automobile policy prescribed by the Insurance Act, R.S.O. 1980, c. 218, discriminate against him in violation of s. 15(1) of the Canadian Charter of Rights and Freedoms. The Court of Appeal dismissed their appeal.

Held (Lamer C.J. and La Forest, Gonthier and Major JJ. dissenting): The appeal should be allowed.

Per Sopinka, Cory, McLachlin and Iacobucci JJ.: The analysis under s. 15(1) of the Charter involves two steps. First, the claimant must show a denial of "equal protection" or "equal benefit" of the law, as compared with some other person. Second, he or she must show that the denial constitutes discrimination. To establish discrimination, the claimant must show that the denial rests on one of the grounds enumerated in s. 15(1) or an analogous ground and that the unequal treatment is based on the stereotypical application of presumed group or personal characteristics, although in rare cases distinctions made on enumerated or analogous grounds may prove to be, upon examination, non‑discriminatory. Once a violation of s. 15(1) is established, the onus then shifts to the party seeking to uphold the law, usually the state, to justify the discrimination under s. 1 of the Charter. This division of the analysis between s. 15(1) and s. 1 accords with the injunction that courts should interpret the enumerated rights in a broad and generous fashion, leaving the task of narrowing the prima facie protection thus granted to conform to conflicting social and legislative interests to s. 1. At the same time, it does not trivialize s. 15(1) by calling all distinctions discrimination. Proof that the enumerated or analogous ground founding a denial of equality is relevant to a legislative goal is only one factor in determining whether a distinction is discriminatory in the social and political context of each case. Relevance as the ultimate indicator of non‑discrimination suffers from the disadvantage that it may validate distinctions which violate the purpose of s. 15(1) and may lead to inquiries better pursued under s. 1.

Exclusion of unmarried partners from accident benefits available to married partners under the policy violates s. 15(1) of the Charter. Denial of equal benefit on the basis of marital status is established in this case, and marital status is an analogous ground of discrimination for purposes of s. 15(1). First, discrimination on that basis touches the essential dignity and worth of the individual in the same way as other recognized grounds of discrimination violative of fundamental human rights norms. Second, marital status possesses characteristics often associated with recognized grounds of discrimination under s. 15(1). Persons involved in an unmarried relationship constitute an historically disadvantaged group, even though the disadvantage has greatly diminished in recent years. A third characteristic sometimes associated with analogous grounds, namely distinctions founded on personal, immutable characteristics, is also present, albeit in attenuated form. While in theory, the individual is free to choose whether to marry or not to marry, in practice the reality may be otherwise. Since the essential elements necessary to engage the overarching purpose of s. 15(1) ‑‑ violation of dignity and freedom, an historical group disadvantage, and the danger of stereotypical group‑based decision‑making ‑‑ are present, discrimination is made out.

The state has failed to demonstrate that the exclusion of unmarried members of family units from motor vehicle accident benefits is demonstrably justified in a free and democratic society. The goal or functional value of the legislation here at issue, which is to sustain families when one of their members is injured in an automobile accident, is of pressing and substantial importance. The legislative goal is not, however, rationally connected to the discriminatory distinction and the law impairs the right more than reasonably necessary to achieve that goal. Marital status is not a reasonably relevant marker of individuals who should receive benefits in the event of injury to a family member in an automobile accident, having regard to available alternative criteria and the need to minimize prejudice to anomalous cases within the group. If the issue had been viewed as a matter of defining who should receive benefits on a basis that is relevant to the goal or functional values underlying the legislation, rather than marriage equivalence, alternatives substantially less invasive of Charter rights might have been found. As an appropriate remedy, the new definition of "spouse" adopted in 1990, which includes heterosexual couples who have cohabited for three years or who have lived in a permanent relationship with a child, should be retroactively "read in" to the impugned legislation.

Per L'Heureux‑Dubé J.: The following factors must be established by a rights claimant before the impugned distinction will be found to be discriminatory: (1) there must be a legislative distinction; (2) this distinction must result in a denial of one of the four equality rights on the basis of the rights claimant's membership in an identifiable group; and (3) this distinction must be "discriminatory" within the meaning of s. 15. Comparisons between different groups are necessary to discern the differential effect of the legislation and to assist the court in properly characterizing and identifying the groups that are relevant to the particular s. 15 inquiry at hand. The only appropriate comparison here is between married persons and unmarried persons who are in a relationship analogous to marriage, that is, a relationship of some degree of publicly acknowledged permanence and interdependence.

Here, assuming that the statutory interpretation of "spouse" as used in the relevant parts of the policy excludes unmarried couples who are cohabiting, this distinction is reasonably capable of either promoting or perpetuating a view amongst persons in relationships analogous to marriage that they are less worthy of recognition or value as human beings or as members of Canadian society, equally deserving of concern, respect, and consideration, and is therefore discriminatory within the meaning of s. 15 of the Charter. Discriminatory impact can be assessed by looking to the nature of both the interest and the group adversely affected by the impugned distinction. Persons in opposite‑sex relationships analogous to marriage have suffered, and continue to suffer, some disadvantage, disapproval and marginalization in society, and are therefore somewhat sensitive to legislative distinctions having prejudicial effects. Nor is marriage simply a matter of individual choice. The decision of whether or not to marry can be one of the most personal decisions an individual will ever make over the course of his or her lifetime. Although certain rights and obligations follow from this choice, it does not do it justice to reduce it to a question of contract. Moreover, there are a significant number of couples in which one person wishes to be in a relationship of publicly acknowledged permanence and interdependence and the other does not. Both the courts and the legislatures have, in recent years, acknowledged and responded to the injustices that often flow from power imbalances of this type and have thereby given increasing recognition to non‑traditional forms of relationships. The affected interest at issue here is the protection of family units from potentially disastrous financial consequences due to the injury of one of their members. Protection of "family" is, in turn, one of the most important interests imaginable in our society. While all injured persons are entitled to that part of their health care costs covered by their provincial medicare systems, actual health care costs may often represent only a small part of the total losses suffered as a result of injury in a motor vehicle accident when loss of income as well as pain and suffering are taken into account. Equally significant, although persons ineligible to claim from a private insurance company under the standard automobile policy may still claim for some compensation under the Motor Vehicle Accident Claims Act, the cost, time and difficulty of recovery by this means are significantly greater than if the person were insured by a private company. The financial consequences of these differences can be profound on a family unit, particularly if the injured party is an income‑earner who has been disabled as a result of the accident. In addition, the impugned distinction categorically excludes from joint insurance coverage all couples in a relationship analogous to marriage.

The impugned distinction cannot be saved under s. 1 of the Charter. The objective of the standard automobile policy, which is to protect stable family units by insuring against the economic consequences that may follow from the injury of one of the members of the family, is pressing and substantial. The government has not demonstrated, however, that the impugned distinction is rationally connected to the objective of the legislation. At the time of the accident, common law spouses in Ontario were bound by an obligation of mutual support yet were excluded from a standard automobile policy whose basic purpose was almost inextricably related to that mutual obligation and to the relationship of interdependency upon which that obligation is premised. The impugned distinction also fails the minimal impairment test, since the unit the legislator has decided to protect (i.e. married persons) is underinclusive of the purpose of the legislation. Although the unit deserving of protection can be defined by marriage, it can also be defined in a workable and acceptably certain way by reference to the length of the relationship or to the existence of children, as was done here when the definition of "spouse" was amended in 1990 to include common law spouses. This new definition should be retroactively "read in" to the legislation.

Per Lamer C.J. and La Forest, Gonthier and Major JJ. (dissenting): The Charter applies to the policy since the policy's terms are prescribed by the Insurance Act. A breach of s. 15(1) occurs when one of the four equality rights set out therein has been infringed in a discriminatory manner. The s. 15(1) analysis involves three steps. The first looks to whether the law has drawn a distinction between the claimant and others. The second questions whether the distinction results in disadvantage, and examines whether the impugned law imposes a burden, obligation or disadvantage on a group of persons to which the claimant belongs which is not imposed on others, or does not provide them with a benefit which it grants others. It is at this second step that the direct or indirect effect of the legislation is examined. The third step assesses whether the distinction is based on an irrelevant personal characteristic which is either enumerated in s. 15(1) or one analogous thereto. This third step thus comprises two aspects: determining the personal characteristic shared by a group and then assessing its relevancy having regard to the functional values underlying the legislation. By its very nature the s. 15(1) review rests on a comparative analysis. Context has a vital part to play in identifying the appropriate groups to be compared, determining whether prejudice flows from the distinction, and assessing the nature and relevancy of the personal characteristic upon which the distinction is drawn. More specifically, an indispensable element of the contextual approach to s. 15(1) involves an inquiry into whether a distinction rests upon or is the expression of some objective physical or biological reality, or fundamental value. This inquiry crucially informs the assessment of whether the prejudicial distinction has been drawn on a relevant basis, and therefore, whether or not that distinction is discriminatory.

Under the approach adopted by this Court in Andrews, the analysis under s. 15 encompasses a determination as to whether the prejudicial distinction is attributable to or on the basis of an enumerated or analogous ground. Such a ground is identified as one that is commonly used to make distinctions which have little or no rational connection with the subject matter, generally reflecting a stereotype. With respect to those grounds listed or enumerated in s. 15, distinctions drawn on such a basis are often but not necessarily always discriminatory, since they may be relevant as merely reflecting a fundamental reality or value. Relevancy is also at the heart of the identification of an analogous ground. The proper identification of such a ground requires a sensitive, contextual examination of its nature in order to determine whether it qualifies as a basis for irrelevant distinctions, and hence is an analogous ground. Once the analogous ground is identified and defined in terms of its nature and scope, any further issues as to relevance are to be examined not under s. 15 but under s. 1 together with any other issues as to justification.

Marriage is both a basic social institution and a fundamental right which states can legitimately legislate to foster. Married status, at least in our society, can only be acquired by the expression of the individual's personal, free choice, regardless of the reason for which that status is assumed. Marriage rests upon a contractual basis, to which the law attaches certain rights and obligations. The decision to marry includes the acceptance of various legal consequences incident to the institution of marriage, including the obligation of mutual support between spouses and the support and raising of children of the marriage. Where individuals choose not to marry, it would undermine the choice they have made if the state were to impose upon them the very same burdens and benefits which it imposes upon married persons. An additional element distinguishing marriage from other relationships is the commitment towards permanence accepted by the parties to the marriage contract. While the decision to marry or not is a joint choice, it is a choice nonetheless.

The insurance policy's limitation of accident benefits to married couples does not infringe s. 15 of the Charter. The impugned legislation draws a distinction, in that it treats married and unmarried couples in a different manner. This distinction is not prejudicial, however, when considered in the larger context of the rights and obligations uniquely and appropriately attached to marriage. Further, since the functional values underlying the legislation are relevant to marital status, marital status is not a personal characteristic which qualifies as an analogous ground. Marital status has several unique characteristics which distinguish it from the grounds enumerated in s. 15(1). In addition to resting upon a consensual, contractual basis, marriage is a status to which the legislature, as a reflection of its social policy, attaches a bundle of rights and obligations. These characteristics are not found in any of the enumerated grounds. Moreover, in contemporary society unmarried couples do not constitute a distinct group suffering from stereotypes or prejudices, although they have been the subject of such prejudices in the past. In this respect, the fostering of marriage as a social institution does not stigmatize unmarried couples nor subject them to stereotypes.

Unmarried couples are not in a situation identical to married spouses with respect to mutual support obligations. While the insurance policy clearly is concerned with economic interdependence, such interdependence is only relevant in so far as it relates to the institution of marriage. The functional value of the benefits at issue here is not to provide support for all family units living in a state of financial interdependence but rather to assist those couples who are married or, as in subsequent legislation, to assist certain prescribed couples who are in a "marriage‑like" relationship. The functional value identified in this legislation, namely the support of marriage, is not itself discriminatory. Distinctions as to the scope of the institution and the benefits which attach thereto are properly the objects of legislative definition.

Just as it is within the scope of legitimate social policy for the legislature to define the scope of "marriage‑like" relationships, there is no obligation on the legislature to extend all the attributes of marriage to unmarried couples. A legislature may as a matter of social policy choose whether and under what circumstances to extend some or all of the attributes of marriage to unmarried couples without running afoul of s. 15(1) of the Charter. The courts must be wary of second‑guessing legislative social policy choices relating to the status, rights and obligations of marriage, a basic institution of our society intimately related to its fundamental values. Barring evidence of a change in these values by a clear consensus that there should be a constitutional constraint on the powers of the state to legislate in relation to marriage, the matter must remain within the scope of legitimate legislative action.

Cases Cited

By McLachlin J.

Referred to: Andrews v. Law Society of British Columbia, 1989 2 (SCC), [1989] 1 S.C.R. 143, aff'g (1986), 1986 1287 (BC CA), 27 D.L.R. (4th) 600; Hunter v. Southam Inc., 1984 33 (SCC), [1984] 2 S.C.R. 145; R. v. Big M Drug Mart Ltd., 1985 69 (SCC), [1985] 1 S.C.R. 295; Re B.C. Motor Vehicle Act, 1985 81 (SCC), [1985] 2 S.C.R. 486; United States v. Cotroni, 1989 106 (SCC), [1989] 1 S.C.R. 1469; R. v. Turpin, 1989 98 (SCC), [1989] 1 S.C.R. 1296; McKinney v. University of Guelph, 1990 60 (SCC), [1990] 3 S.C.R. 229; R. v. Swain, 1991 104 (SCC), [1991] 1 S.C.R. 933; Symes v. Canada, 1993 55 (SCC), [1993] 4 S.C.R. 695; Weatherall v. Canada (Attorney General), 1993 112 (SCC), [1993] 2 S.C.R. 872; Egan v. Canada, 1995 98 (SCC), [1995] 2 S.C.R. 513; Bliss v. Attorney General of Canada, 1978 25 (SCC), [1979] 1 S.C.R. 183; Brooks v. Canada Safeway Ltd., 1989 96 (SCC), [1989] 1 S.C.R. 1219; Boronovsky v. Chief Rabbis of Israel, P.D. CH [25] (1), 7; Pettkus v. Becker, 1980 22 (SCC), [1980] 2 S.C.R. 834; Peter v. Beblow, 1993 126 (SCC), [1993] 1 S.C.R. 980; Canada (Attorney General) v. Mossop, 1993 164 (SCC), [1993] 1 S.C.R. 554; R. v. Oakes, 1986 46 (SCC), [1986] 1 S.C.R. 103; R. v. Edwards Books and Art Ltd., 1986 12 (SCC), [1986] 2 S.C.R. 713; Reference re ss. 193 and 195.1(1)(c) of the Criminal Code (Man.), 1990 105 (SCC), [1990] 1 S.C.R. 1123; Rocket v. Royal College of Dental Surgeons of Ontario, 1990 121 (SCC), [1990] 2 S.C.R. 232; Schachter v. Canada, 1992 74 (SCC), [1992] 2 S.C.R. 679; Rodriguez v. British Columbia (Attorney General), 1993 75 (SCC), [1993] 3 S.C.R. 519.

By L'Heureux‑Dubé J.

Referred to: Egan v. Canada, 1995 98 (SCC), [1995] 2 S.C.R. 513; Andrews v. Law Society of British Columbia, 1989 2 (SCC), [1989] 1 S.C.R. 143; Pettkus v. Becker, 1980 22 (SCC), [1980] 2 S.C.R. 834; Sorochan v. Sorochan, 1986 23 (SCC), [1986] 2 S.C.R. 38; Moge v. Moge, 1992 25 (SCC), [1992] 3 S.C.R. 813; Canada (Attorney General) v. Mossop, 1993 164 (SCC), [1993] 1 S.C.R. 554.

By Gonthier J. (dissenting)

Leroux v. Co‑operators General Insurance Co. (1991), 1991 7054 (ON CA), 4 O.R. (3d) 609; Andrews v. Law Society of British Columbia, 1989 2 (SCC), [1989] 1 S.C.R. 143; R. v. Turpin, 1989 98 (SCC), [1989] 1 S.C.R. 1296; Symes v. Canada, 1993 55 (SCC), [1993] 4 S.C.R. 695; R. v. Big M Drug Mart Ltd., 1985 69 (SCC), [1985] 1 S.C.R. 295; R. v. Hess, 1990 89 (SCC), [1990] 2 S.C.R. 906; Weatherall v. Canada (Attorney General), 1993 112 (SCC), [1993] 2 S.C.R. 872; McKinney v. University of Guelph, 1990 60 (SCC), [1990] 3 S.C.R. 229; Reference re Public Service Employee Relations Act (Alta.), 1987 88 (SCC), [1987] 1 S.C.R. 313; R. v. Généreux, 1992 117 (SCC), [1992] 1 S.C.R. 259; Bliss v. Attorney General of Canada, 1978 25 (SCC), [1979] 1 S.C.R. 183; Brooks v. Canada Safeway Ltd., 1989 96 (SCC), [1989] 1 S.C.R. 1219; Maynard v. Hill, 125 U.S. 190 (1888); Meyer v. Nebraska, 262 U.S. 390 (1923); Skinner v. Oklahoma, 316 U.S. 535 (1942); Griswold v. Connecticut, 381 U.S. 479 (1965); Loving v. Virginia, 388 U.S. 1 (1967); Boddie v. Connecticut, 401 U.S. 371 (1971); Cleveland Board of Education v. LaFleur, 414 U.S. 632 (1974); Moore v. East Cleveland, 431 U.S. 494 (1977); Zablocki v. Redhail, 434 U.S. 374 (1978); Marvin v. Marvin, 557 P.2d 106 (1976); Elden v. Sheldon, 758 P.2d 582 (1988); Beaty v. Truck Insurance Exchange, 8 Cal.Rptr.2d 593 (1992); Zurich Insurance Co. v. Ontario (Human Rights Commission), 1992 67 (SCC), [1992] 2 S.C.R. 321; Dickason v. University of Alberta, 1992 30 (SCC), [1992] 2 S.C.R. 1103; Geiger v. London Monenco Consultants Ltd. (1992), 1992 7425 (ON CA), 43 C.C.E.L. 291; Pettkus v. Becker, 1980 22 (SCC), [1980] 2 S.C.R. 834; Sorochan v. Sorochan, 1986 23 (SCC), [1986] 2 S.C.R. 38; Norman v. Unemployment Insurance Appeals Board, 663 P.2d 904 (1983); Hendrix v. General Motors Corp., 193 Cal.Rptr. 922 (1983).

Statutes and Regulations Cited

Canadian Charter of Rights and Freedoms, ss. 1, 3, 8, 11(b), 15, 24(1).

Constitution Act, 1982, s. 52.

European Convention for the Protection of Human Rights and Fundamental Freedoms, 213 U.N.T.S. 221, art. 12.

Family Law Act, R.S.O. 1990, c. F.3 [formerly S.O. 1986, c. 4], ss. 30, Part III, 53.

Family Law Reform Act, S.O. 1978, c. 2, s. 14.

Human Rights Code, 1981, S.O. 1981, c. 53, s. 9(j).

Insurance Act, R.S.O. 1980, c. 218, ss. 231, 233, Schedule C.

Motor Vehicle Accident Claims Act, R.S.O. 1990, c. M.41.

R.R.O. 1980, Reg. 535.

Universal Declaration of Human Rights, G.A. Res. 217 A (III), U.N. Doc. A/810, at 71 (1948), art. 16.

Authors Cited

Freeman, Michael D. A., and Christina M. Lyon. Cohabitation without Marriage. Aldershot, Hants., England: Gower, 1983.

Gibson, Dale. "Analogous Grounds of Discrimination Under the Canadian Charter: Too Much Ado About Next to Nothing" (1991), 29 Alta. L. Rev. 772.

Hafen, Bruce C. "The Constitutional Status of Marriage, Kinship, and Sexual Privacy ‑‑ Balancing the Individual and Social Interests" (1983), 81 Mich. L. Rev. 463.

Hogg, Peter W. Constitutional Law of Canada, 2nd ed. Toronto: Carswell, 1985.

Holland, Winifred H. "Marriage and Cohabitation ‑‑ Has the Time Come to Bridge the Gap?". In Family Law: Roles, Fairness and Equality, Special Lectures of the Law Society of Upper Canada 1993. Scarborough: Carswell, 1994, 369.

APPEAL from a judgment of the Ontario Court of Appeal (1991), 1991 7312 (ON CA), 4 O.R. (3d) 623, 83 D.L.R. (4th) 766, [1991] I.L.R. ¶ 1‑2770, 7 C.C.L.I. (2d) 317, affirming a judgment of Chilcott J. (1990), 1990 6699 (ON SC), 71 O.R. (2d) 662, 65 D.L.R. (4th) 670, [1990] I.L.R. ¶ 1‑2551, 45 C.C.L.I. 296, dismissing the appellants' action. Appeal allowed, Lamer C.J. and La Forest, Gonthier and Major JJ. dissenting.

Giovanna Roccamo and Mark Edwards, for the appellants.

Catherine L. Jones and R. Cooligan, for the respondents.

Graham R. Garton, Q.C., and James Hendry, for the intervener the Attorney General of Canada.

Rebecca Regenstreif, for the intervener the Attorney General for Ontario.

Madeleine Aubé and Kathleen McNicoll, for the intervener the Attorney General of Quebec.

Shawn Greenberg, for the intervener the Attorney General of Manitoba.

W. Ian Binnie, Q.C., and Lisa A. Clarkson, counsel appearing as amicus curiae.

The reasons of Lamer C.J. and La Forest, Gonthier and Major JJ. were delivered by

1 Gonthier J. (dissenting) -- I have had the benefit of the reasons of Justice L'Heureux-Dubé and Justice McLachlin. I agree that the statutory interpretation of the word "spouse" in the insurance policy limits accident benefits to married couples and does not extend to unmarried couples living together. With regard to the question of whether this limitation infringes s. 15 of the Canadian Charter of Rights and Freedoms, I respectfully disagree with their conclusion that these provisions are discriminatory. I would dismiss the appeal.

2 It is my position that marital status may constitute an analogous ground of discrimination under s. 15 of the Charter. However, in examining whether a law conforms to s. 15 in any given case, regard must be had to the nature of the analogous ground and its relevancy to the distinction that is being drawn by the legislation. Marriage is an institution entered into by choice which carries with it certain benefits and burdens. Among these is the obligation of mutual support. As I will explain below, the benefits at issue in this case are most appropriately characterized as coming within the scope of the support obligations which the law attaches to marriage. Where the legislature draws a distinction premised on a characteristic relevant to the institution of marriage, such as these support obligations, then the distinction is not discriminatory and is therefore permissible. In the case at bar, I find that the distinction drawn by the impugned provisions of the Insurance Act, R.S.O. 1980, c. 218, is relevant to the institution of marriage. Accordingly, the legislation does not infringe s. 15(1) of the Charter, and s. 1 need not be addressed.

I - Facts and Proceedings

3 At the outset, it should be noted that the merits of this case have not yet been heard by the lower court. The issue in this appeal as to the meaning of "spouse" under the Insurance Act, was raised by preliminary motion on an agreed statement of facts.

4 The appellants John Miron and Jocelyne Valliere have resided together in a common law relationship since May 1983. Miron is the father of two of the three children who have been born to Valliere. These children were born in 1981 and 1984. The respondent, the Economical Mutual Insurance Company, issued a policy of motor vehicle insurance to Valliere for the period December 12, 1986 to December 12, 1987. The terms of this policy were set by the Ontario Standard Automobile Policy as provided by the Insurance Act, ss. 231, 233, Schedule C, and R.R.O. 1980, Reg. 535.

5 In August 1987, Miron sustained injuries while a passenger in a motor vehicle owned by the respondent William James McIsaac, and which was driven by the respondent Richard Trudel. Neither McIsaac nor Trudel was insured. As a result, Miron claimed accident benefits for loss of income pursuant to Section B, Subsection 2, Part II of the Ontario Standard Automobile Policy incorporated in the policy issued by the respondent to Valliere. He also claimed damages pursuant to the Uninsured Motorist coverage under Section B, Subsection 3 of the same policy.

6 The respondent, the Economical Mutual Insurance Company, brought a motion to determine a question of law prior to trial, namely whether Miron was the "spouse" of Valliere for the purposes of the Section B, Subsection 2, Part II or Section B, Subsection 3 of the Ontario Standard Automobile Policy. Chilcott J., the motion judge, found that, for the purposes of the applicable portions of the policy, "spouse" meant a person who is legally married. Accordingly, Miron was found not to be a "spouse" within the meaning of those provisions of the policy, and therefore not insured under the policy.

7 The appellants appealed the decision to the Ontario Court of Appeal and based their argument solely upon s. 15 of the Charter. That argument had not been raised before the motion judge. The Court of Appeal dismissed the appeal for reasons given by the court on the same day in Leroux v. Co-operators General Insurance Co. (1991), 1991 7054 (ON CA), 4 O.R. (3d) 609. In Leroux, the Court of Appeal held that there was no violation of s. 15 of the Charter, as marital status was not a ground of discrimination analogous to those specifically enumerated in s. 15. The Court of Appeal concluded that unmarried couples were not members of a "disadvantaged group", nor did they constitute a "discrete and insular minority" which had suffered "social, political and legal disadvantage in our society". The Court of Appeal also stated that the characteristic of being an "unmarried partner" did not constitute an "immutable" characteristic (pp. 620-21).

II - Relevant Statutory Provisions

Canadian Charter of Rights and Freedoms

15. (1) Every individual is equal before and under the law and has the right to the equal protection and equal benefit of the law without discrimination and, in particular, without discrimination based on race, national or ethnic origin, colour, religion, sex, age or mental or physical disability.

Ontario Standard Automobile Policy (S.P.F. No. 1)

Section B, Subsection 2, Part II ^__ Loss of Income

Subject to the provisions of this Part, a weekly payment for the loss of income from employment for the period during which the insured person suffers substantial inability to perform the essential duties of his occupation or employment . . . .

Section B, Special Provisions, Definitions, and Exclusions of Section B

(1) "INSURED PERSON" DEFINED

In this section, the words "insured person" mean,

(b) the insured and, if residing in the same dwelling premises as the insured, his or her spouse and any dependent relative of either while an occupant of any other automobile; provided that,

(i) the insured is an individual or are husband and wife;

Section B, Subsection 3 -- Uninsured Motorist Cover

All sums that

(a) a person insured under the contract is legally entitled to recover from the owner or driver of an uninsured automobile or unidentified automobile as damages for bodily injuries resulting from an accident involving an automobile;

(b) any person is legally entitled to recover from the owner or driver of an uninsured automobile or unidentified automobile as damages for bodily injury to or the death of a person insured under the contract resulting from an accident involving an automobile; and

(c) a person insured under the contract is legally entitled to recover from the identified owner or driver of an uninsured automobile as damages for accidental damage to the insured automobile or its contents, or to both the insured automobile and its contents, resulting from an accident involving an automobile.

1.Definitions:

For the purposes of this subsection,

(b) "person insured under the contract" means,

(iii) in respect of a claim for bodily injuries or death,

b.the insured and, if residing in the same dwelling premises as the insured, his or her spouse and any dependent relative of either,

(1)while an occupant of an uninsured automobile . . . [Emphasis added.]

III - Issues

8 Two issues are presented in this appeal.

1. Is Miron, as a person living in a conjugal relationship outside of marriage, a "spouse" of Valliere and therefore an insured person within the meaning of Section B, Subsection 2, Part II or Section B, Subsection 3 of the Ontario Standard Automobile Policy?

2. In the alternative, do the provisions of Section B, Subsection 2, Part II or Section B, Subsection 3 of the Ontario Standard Automobile Policy infringe s. 15 of the Canadian Charter of Rights and Freedoms, and if so, are the provisions justified under s. 1 of the Charter?

9 As I indicated above, I agree with my colleague McLachlin J.'s reasons with regard to the statutory interpretation of the word "spouse" as being limited to married couples. Therefore, my analysis will focus on the second question.

10 The application of the Charter to this private contract is not contested by the parties. It is common ground that the Charter applies to the insurance policy since the terms of the policy are prescribed by the Insurance Act, ss. 231 and 233 and Schedule C.

11 Needless to say, the interpretation that will be given to s. 15 of the Charter by this Court will not simply add judicial gloss to what is the equivalent of ordinary legislation subject to amendment by the Legislature, but rather, will establish what the Legislature can or must do under the Constitution thereby constraining its powers. As McIntyre J. stated in Andrews v. Law Society of British Columbia, 1989 2 (SCC), [1989] 1 S.C.R. 143, at p. 175:

Both the enumerated grounds themselves and other possible grounds of discrimination recognized under s. 15(1) must be interpreted in a broad and generous manner, reflecting the fact that they are constitutional provisions not easily repealed or amended but intended to provide a "continuing framework for the legitimate exercise of governmental power" and, at the same time, for "the unremitting protection" of equality rights . . . . [Emphasis added.]

I stress this obvious fact, however, because as I see the matter, this case is ultimately about the ambit of legitimate legislative choice in defining the attributes of a fundamental social institution, namely the rights and obligations attached to marriage. This Court should be cautious in labelling as discriminatory laws which simply seek to define the rights and obligations of married persons differently from those who choose to cohabit outside of marriage.

IV - Analysis

A. General Principles Applicable to Section 15 of the Charter

12 It has been firmly established by this Court that not all distinctions infringe s. 15 of the Charter. A breach of s. 15 occurs only when one of the four equality rights set out therein has been infringed in a discriminatory manner. McIntyre J. outlined in Andrews, supra, at pp. 174-75, the concept of discrimination in the following words:

I would say then that discrimination may be described as a distinction, whether intentional or not but based on grounds relating to personal characteristics of the individual or group, which has the effect of imposing burdens, obligations, or disadvantages on such individual or group not imposed upon others, or which withholds or limits access to opportunities, benefits, and advantages available to other members of society. Distinctions based on personal characteristics attributed to an individual solely on the basis of association with a group will rarely escape the charge of discrimination, while those based on an individual's merits and capacities will rarely be so classed.

13 The analysis to be undertaken under s. 15(1) of the Charter involves three steps. The first step looks to whether the law has drawn a distinction between the claimant and others. The second step then questions whether the distinction results in disadvantage, and examines whether the impugned law imposes a burden, obligation or disadvantage on a group of persons to which the claimant belongs which is not imposed on others, or does not provide them with a benefit which it grants others (Andrews, supra). It is at this second step that the direct or indirect effect of the legislation is examined.

14 The third step assesses whether the distinction is based on an irrelevant personal characteristic which is either enumerated in s. 15(1) or one analogous thereto. As McIntyre J. emphasized in Andrews, supra, at p. 165, s. 15(1) seeks to eliminate differences based on irrelevant personal characteristics:

In other words, the admittedly unattainable ideal [of equality] should be that a law expressed to bind all should not because of irrelevant personal differences have a more burdensome or less beneficial impact on one than another.

In his separate concurring reasons, La Forest J. agreed that discrimination exists where distinctions are drawn "on the basis of `irrelevant personal differences' such as those listed in s. 15 and, traditionally, in human rights legislation" (Andrews, supra, at p. 193). The ingredient of the relevancy of the personal characteristic was also emphasized as going to the essential core of this Court's s. 15 jurisprudence by Professor Dale Gibson in his article "Analogous Grounds of Discrimination Under the Canadian Charter: Too Much Ado About Next to Nothing" (1991), 29 Alta. L. Rev. 772, at p. 780.

15 This third step thus comprises two aspects: determining the personal characteristic shared by a group and then assessing its relevancy having regard to the functional values underlying the legislation. On the first aspect of the third step of the s. 15(1) analysis, the individual's membership in a group is an essential condition, while idiosyncrasies unrelated to membership in a group do not give rise to discrimination. However, the notion of a "group" should not be confused with the concept of a "disadvantaged group" which refers to a "discrete and insular minority" which has suffered from political, historical or legal disadvantage: R. v. Turpin, 1989 98 (SCC), [1989] 1 S.C.R. 1296, at p. 1333. The identification of such a disadvantaged group in our society can serve as a meaningful indicium of discrimination, that is, it may help to identify the existence of a disadvantage attributable to an irrelevant personal characteristic. However, I agree with McLachlin J. that membership in such a disadvantaged group is not an essential precondition for bringing a claim under s. 15 of the Charter. The second aspect of the third step, that of assessing relevancy, looks to the nature of the personal characteristic and its relevancy to the functional values underlying the law. Of course, the functional values underlying the law may themselves be discriminatory. Such will be the case where the underlying values are irrelevant to any legitimate legislative purpose. Relevancy is assessed by reference to a ground enumerated in s. 15 or one analogous thereto. I will return to this aspect of the analysis below.

(1) The Importance of Context to the Comparative Analysis Under Section 15 of the Charter

16 Throughout the s. 15(1) analysis, to paraphrase Iacobucci J. in Symes v. Canada, 1993 55 (SCC), [1993] 4 S.C.R. 695, at p. 754, one must necessarily undertake a form of comparative analysis in order to determine whether particular facts give rise to inequality. This proposition was enunciated by McIntyre J. in Andrews, supra, at p. 164, in the following terms:

[Equality] is a comparative concept, the condition of which may only be attained or discerned by comparison with the condition of others in the social and political setting in which the question arises. It must be recognized at once, however, that every difference in treatment between individuals under the law will not necessarily result in inequality and, as well, that identical treatment may frequently produce serious inequality.

17 In R. v. Turpin, supra, at pp. 1331-32, Wilson J. explained how this comparative analysis is linked to the examination of the larger context. In her words:

In determining whether there is discrimination on grounds relating to the personal characteristics of the individual or group, it is important to look not only at the impugned legislation which has created a distinction that violates the right to equality but also to the larger social, political and legal context. . . . Accordingly, it is only by examining the larger context that a court can determine whether differential treatment results in equality or whether, contrariwise, it would be identical treatment which would in the particular context result in inequality or foster disadvantage.

18 Thus, Wilson J.'s words stress the importance of a contextual approach in order to prevent the s. 15 analysis from becoming a mechanical and sterile categorization process. This admonition was most memorably issued by Dickson J., as he then was, in the landmark case of R. v. Big M Drug Mart Ltd., 1985 69 (SCC), [1985] 1 S.C.R. 295, at p. 344, where he stated that "the Charter was not enacted in a vacuum" and it must "be placed in its proper linguistic, philosophic and historical contexts". Context is indispensable to identifying the appropriate groups to be compared, to determining whether prejudice flows from the distinction, and to assessing the nature and relevancy of the personal characteristic upon which the distinction is drawn. In sum, the larger context importantly informs all stages of the analysis and ensures that it is not narrowly restricted to the "four corners of the impugned legislation" (to use the words of Wilson J. in Turpin, supra, at p. 1332).

19 More specifically, an indispensable element of the contextual approach to s. 15(1) involves an inquiry into whether a distinction rests upon or is the expression of some objective physical or biological reality, or fundamental value. This inquiry crucially informs the assessment of whether the prejudicial distinction has been drawn on a relevant basis, and therefore, whether or not that distinction is discriminatory.

20 For example, in R. v. Hess, 1990 89 (SCC), [1990] 2 S.C.R. 906, at pp. 928-29, this Court held that the legitimacy of distinctions drawn on the basis of sex in the context of the criminal law will depend on the nature of the offence in issue. In that case, s. 146(1) of the Criminal Code made it an offence for a man to have sexual intercourse with a female under the age of 14 who was not his wife. In finding that the legislation did not infringe s. 15(1), the Court was careful to consider the offence in the appropriate context by taking into account certain fundamental biological realities, namely the fact that only men can commit the proscribed act. This Court took a similar approach in Weatherall v. Canada (Attorney General), 1993 112 (SCC), [1993] 2 S.C.R. 872, where a male prison inmate challenged the constitutionality of frisk searching and patrolling conducted in male prisons by female guards. The appellant argued that the practice was discriminatory since female inmates were not also subject to cross-gender frisk searches. Speaking for the Court, La Forest J. stated that the different treatment of male and female inmates does not necessarily amount to discrimination. As he explained (at p. 877), "[g]iven the historical, biological and sociological differences between men and women, equality does not demand that practices which are forbidden where male officers guard female inmates must also be banned where female officers guard male inmates". In both Hess, supra, and Weatherall, supra, this Court found that distinctions drawn on the basis of relevant biological differences between the sexes do not necessarily constitute discrimination.

21 Like biological realities, fundamental values may also be critical to an appreciation of the appropriate context in which to conduct an analysis of s. 15(1). For example, in McKinney v. University of Guelph, 1990 60 (SCC), [1990] 3 S.C.R. 229, at p. 278, to support a finding of discrimination under s. 15(1), La Forest J. approvingly quoted from Reference re Public Service Employee Relations Act (Alta.), 1987 88 (SCC), [1987] 1 S.C.R. 313, at p. 368, that employment is "one of the most fundamental aspects in a person's life". He held that a restriction on employment under the mandatory retirement policies of the respondent universities imposed a burden on the basis of the irrelevant personal difference of age and, hence, was discriminatory (at p. 278).

22 Finally, it is worth stressing that a contextual analysis may lead to fundamentally different assessments as to whether distinctions drawn on the basis of the same ground will amount to discrimination. In other words, depending on the context, the same ground may be discriminatory with respect to certain classes of distinction but not with respect to others. For example, this Court recognized in R. v. Turpin, supra, that while province of residence was not a ground of discrimination under the applicable legislative scheme in that case, it was nevertheless possible that in different circumstances a distinction based on province of residence could be discriminatory. In the words of Wilson J., at p. 1333:

I would not wish to suggest that a person's province of residence or place of trial could not in some circumstances be a personal characteristic of the individual or group capable of constituting a ground of discrimination. I simply say that it is not so here.

In the same vein, in R. v. Généreux, 1992 117 (SCC), [1992] 1 S.C.R. 259, Lamer C.J. dismissed the suggestion that the existence of a parallel system of military justice discriminated against members of the Armed Forces, but nevertheless added the following important caveat (at p. 311):

I emphasize, however, that my conclusion here is confined to the context of this appeal. I do not wish to suggest that military personnel can never be the objects of disadvantage or discrimination in a manner that could bring them within the meaning of s. 15 of the Charter. Certainly it is the case, for instance, that after a period of massive demobilization at the end of hostilities, returning military personnel may well suffer from disadvantages and discrimination peculiar to their status, and I do not preclude that members of the Armed Forces might constitute a class of persons analogous to those enumerated in s. 15(1) under those circumstances. However, no circumstances of this sort arise in the context of this appeal, and the appellant gains nothing by pleading s. 15 of the Charter. [Emphasis in original.]

In short, a sensitive, contextual approach is essential in determining whether distinctions drawn on the basis of a particular ground are discriminatory in any given case. Differently put, context is indispensable in helping to determine whether a given basis of distinction is a discriminatory ground for certain classes of cases but not for others.

(2) Relevance and this Court's "Enumerated and Analogous Grounds" Approach to Section 15 in Andrews

23 As I noted above, a concern for whether a given distinction is drawn on a relevant basis is at the core of the analysis under s. 15. An otherwise prejudicial distinction drawn on a relevant basis is not discriminatory. In determining what constitutes such a relevant basis for distinction, this Court in Andrews, supra, adopted the so-called "enumerated and analogous grounds" approach to discrimination. Under this approach, the analysis under s. 15 of the Charter encompasses a determination as to whether the prejudicial distinction is attributable to or on the basis of an enumerated or analogous ground. Such a ground is identified as one that is commonly used to make distinctions which have little or no rational connection with the subject matter, generally reflecting a stereotype.

24 With respect to those grounds listed or enumerated in s. 15, distinctions drawn on such a basis are often but not necessarily always discriminatory, since they may be relevant as merely reflecting a fundamental reality or value. As I noted above, in Hess, supra, and Weatherall, supra, this Court found that distinctions drawn on the basis of the enumerated ground of sex but relevant as reflecting certain biological realities are not discriminatory.

25 Relevancy is also at the heart of the identification of an analogous ground. The proper identification of such a ground requires a sensitive, contextual examination of its nature in order to determine whether it qualifies as a basis for irrelevant distinctions, and hence is an analogous ground. But as I pointed out above, a contextual analysis also serves to understand the nature of the ground and determine in what respect it so qualifies. It may be that a given ground is analogous in certain respects but not others. As in the case of enumerated grounds, the distinction being drawn may simply reflect certain biological or physical realities or fundamental values which are in themselves relevant. Certain distinctions may be inherently connected to the ground, indeed come within the scope of its definition and, by that very fact, may be relevant regardless of circumstances and incapable of grounding discrimination.

26 As I develop below, marital status is an example of a ground which, while analogous in certain respects, cannot be so with respect to those attributes and effects which serve to define marriage itself, which include the rights and obligations necessarily incident to the institution, and distinguish it from a state of absence of marriage. This is so, as marriage in itself is not discriminatory as it is a matter of choice and a basic institution of society.

27 To the extent, then, that a law in any given case mirrors or reflects a distinction drawn on such a basis that is relevant to its functional values which are not themselves discriminatory, the distinction drawn by the law will not be discriminatory. In the absence of such a relevant basis, the ground in issue is properly qualified as one analogous to those enumerated in s. 15 and a distinction based on it will be discriminatory and infringe s. 15.

28 I should also emphasize that the approach to s. 15 in these reasons in no way departs from this Court's approach in Andrews, supra, and in subsequent jurisprudence. My concern has only been to clarify a qualification which must be made in the application of the analogous grounds approach, a qualification which merely calls for a heightened sensitivity to the nature of the ground in issue in any given case, and a recognition that a ground which may be the basis of discrimination in one context may be innocuous in another.

29 Parenthetically, I note that a recognition of the need for sensitivity to the context in a s. 15(1) analysis is also a complete response to my colleague McLachlin J.'s concern that a focus on relevance alone is not sufficient in all cases to address whether a charge of discrimination has been proven. Surely, if both the larger context and the varieties of context are kept firmly in mind in assessing the nature of an analogous or enumerated ground, then there can be no danger that the purpose of the equality guarantees will somehow be eclipsed or overlooked in a relevance approach to s. 15. Indeed, a criterion defined in terms of stereotype based on presumed group characteristics, rather than on the basis of merit, capacity or circumstances, is but an elaboration of the concept of relevance.

30 Nor, in my opinion, does recognition of the essential role of relevance under s. 15 raise the spectre of superficial biological differences reemerging as a justification for sexual discrimination, as had been countenanced by this Court's decision in Bliss v. Attorney General of Canada, 1978 25 (SCC), [1979] 1 S.C.R. 183. The lesson to be learned from Bliss is most emphatically not that a court should never sustain legislative distinctions on the basis of relevant biological differences between the sexes. If that were the case, then the state would, for example, be barred under s. 15 of the Charter from providing pregnant women with financial assistance for the purpose of assisting them to maintain their own and their child's physical well-being during the pregnancy, which would be absurd. I would venture to say, without any hesitation, that a distinction drawn on the basis of such a fundamental biological reality would be laudable as well as relevant. No, the lesson to be drawn from Bliss is rather that discrimination analysis must be conducted with a view to the larger context. Only then can a court sensibly separate biological differences which are normatively relevant and hence benign, from those which are irrelevant and thus discriminatory. This Court was thus properly mindful of broadening the appropriate context in Brooks v. Canada Safeway Ltd., 1989 96 (SCC), [1989] 1 S.C.R. 1219, at p. 1237, when it overturned Bliss and recognized that exclusion of pregnancy from coverage under an employer benefit plan for compensation of employees unable to work for health-related reasons was "untenable", being irrelevant to the purpose of the plan, as pregnancy is such a valid health-related reason. It was therefore discrimination and also discrimination on the basis of sex. The Court rightly pointed out that while nature endowed only one of the sexes with the ability to bear children, a facile reliance on this fact would result in placing the major cost of procreation entirely on women.

(3)Relevance under Section 15(1) and its Relationship to Reasonableness under Section 1

31 Before addressing the application of s. 15(1) of the Charter to the facts of this case, I think it is important to clarify the relationship between the requirement of relevance under s. 15(1) of the Charter and that of reasonableness under s. 1. It should be emphasized that determining the relevancy of a distinction does not amount to importing under s. 15(1) the principles of justification found within s. 1 of the Charter. This approach is appropriately mindful, then, of the caution which Wilson J. issued in Turpin, supra, at p. 1328, that "[t]he equality rights must be given their full content divorced from justificatory factors properly considered under s. 1".

32 Fundamentally, s. 15(1) is concerned with the relevancy of distinctions. Relevancy goes to the determination of the existence of discrimination, whereas the s. 1 justification only arises after discrimination has been established. Thus, even where a distinction is based on an irrelevant personal characteristic, and is therefore discriminatory, it is still possible for the discrimination to be rationally connected to a pressing and substantial governmental objective. In this regard, it is useful to refer back to Andrews, supra. That decision significantly modified in at least two ways the concept of equality as it was then applied by the courts. First, the Court rejected as deficient the rule of formal equality which stated that persons in similar situations must be given similar treatment. Second, and most relevant in the present context, this Court also rejected what might be termed the "reasonableness test" (Andrews, supra, at pp. 181-82). The object pursued by the "reasonableness test" was to ascertain whether the impugned distinction was reasonable or fair, taking into account the purposes, the aims and the effect of the legislation on the person. Under this test, both the finding of whether a distinction resulted in discrimination and, to a large extent, the justification of that discrimination were done under s. 15(1).

33 Instead, under the enumerated and analogous grounds approach adopted by the Court in Andrews, the analysis under s. 15 of the Charter encompasses a determination as to whether the prejudicial distinction is attributable to an enumerated or analogous ground. Once the analogous ground is identified and defined in terms of its nature and scope as explained above, any further issues as to relevance are to be examined not under s. 15 but under s. 1 together with any other issues as to justification.

34 Differently put, the Court in Andrews, supra, held that the relevance of a basis for differential treatment under s. 15 must be clearly distinguished from its reasonableness and thus whether it can be justified under s. 1. This distinction was made forcefully by La Forest J., who stated (at p. 197):

While it cannot be said that citizenship is a characteristic which "bears no relation to the individual's ability to perform or contribute to society" (Fontiero v. Richardson, 411 U.S. 677 (1973), at p. 686), it certainly typically bears an attenuated sense of relevance to these. That is not to say that no legislative conditioning of benefits (for example) on the basis of citizenship is acceptable in the free and democratic society that is Canada, merely that legislation purporting to do so ought to be measured against the touchstone of our Constitution. It requires justification. [Emphasis added.]

35 An additional example of the proper relationship between s. 15(1) and s. 1 is found in McKinney, supra. La Forest J. found that the mandatory retirement policies of the respondent universities, though amounting to discrimination under s. 15(1) as impinging upon employment, were rationally connected to the legitimate objectives of the impugned legislation, which included the promotion of excellence in higher education and the preservation of academic freedom (p. 281). These values were not in issue. The debate bore upon the means chosen to promote them. La Forest J. examined the relationship between the needs of the universities and the tenure of faculty members. He concluded that "mandatory retirement [was] intimately tied to the tenure system" (p. 283). He further resolved that mandatory retirement "ensures continuing faculty renewal, a necessary process to enable universities to be centres of excellence. . . . In a closed system with limited resources, this can only be achieved by departures of other people" (p. 284) (emphasis in original). The evidence submitted was found to demonstrate some correlation, depending on the resource allocation of universities, between the mandatory retirement and the renewal of faculty members. These were considerations properly weighed under s. 1 of the Charter.

36 Nor do I believe that an approach to s. 15 which focuses on enumerated or analogous grounds as explained above places any additional onus on a Charter claimant. First, as I have already noted, a concern for relevance has run through this Court's s. 15 jurisprudence. In this sense, my clarification of the analogous grounds approach has obviously not imposed any additional burden on a Charter claimant. Secondly, the onus is always on a claimant to prove on the balance of probabilities that a Charter-guaranteed right or freedom has been infringed. Discharging this onus and demonstrating that legislative distinctions have been drawn on an irrelevant ground having regard to its nature and scope will, admittedly, require a knowledge of the wording of the legislation and of the larger context within which it is situated. However, these are obviously not matters to which only the enacting government is privy. Where it is not plain from the subject matter and wording of the law that a distinction based on an enumerated or analogous ground reflects some biological or physical reality or fundamental value and where no other evidence is tendered, a court will generally be obliged to find that such a distinction is discriminatory. As a consequence, in any case where legislation is challenged under s. 15 of the Charter, the enacting government will have an incentive to assist the court to determine what relevance, if any, there is in the impugned legislative distinction.

37 I would also stress that there may indeed be significant overlap between the assessment of the functional values of the legislation under s. 15, and the purpose of the legislation under s. 1. This will not necessarily always be the case, however, as the legislation in McKinney, supra, clearly illustrates. However, to the extent that there is overlap between the functional values and the purpose of the legislation, this is not cause for concern, since fundamentally different questions are asked under s. 15 and under s. 1 proceeding under fundamentally different premises: under s. 15, one asks whether the legislation is discriminatory on the basis of certain grounds, and then under s. 1, having found it to be discriminatory, one asks whether it is otherwise a reasonable limit which can be demonstrably justified in a free and democratic society. In short, as McIntyre J. stressed in Andrews, supra, while "the relationship between these two sections may well be difficult to determine on a wholly satisfactory basis" it is "important to keep them analytically distinct" (p. 178).

38 I conclude from the foregoing that a distinction deemed irrelevant because based on an enumerated or analogous ground and thus discriminatory under s. 15(1) may nevertheless be rationally connected to furthering a larger social purpose, and hence may be judged reasonable under s. 1.

B. Section 15(1) of the Charter and the Ground of Marital Status

39 The distinction drawn by the legislation at issue is made on the basis of marriage and the appellants have rested their whole argument upon the premise that their situation is identical to that of married couples, and carries with it the same consequences; it is therefore important in addressing the context of this distinction to examine briefly the concept of marriage and its place in our society, together with some implications of the unique contractual basis of marriage, even at the risk of stating the obvious.

(1)The Importance of Marriage as a Social Institution

40 The question raised by the case at bar is intimately linked with the institution of marriage, the importance of which has long been recognized in our society. Other countries, as well as international law, acknowledge both the importance of marriage and the legitimacy of state action which fosters this basic social institution.

41 For example, the United States Supreme Court has long recognized that marriage is a fundamental social institution. In Maynard v. Hill, 125 U.S. 190 (1888), at pp. 205 and 211, the court stated:

Marriage, as creating the most important relation in life, as having more to do with the morals and civilization of a people than any other institution, has always been subject to the control of the legislature. That body prescribes the age at which parties may contract to marry, the procedure or form essential to constitute marriage, the duties and obligations it creates, its effects upon the property rights of both, present and prospective, and the acts which may constitute grounds for its dissolution.

[Marriage] is an institution, in the maintenance of which in its purity the public is deeply interested, for it is the foundation of the family and of society, without which there would be neither civilization nor progress.

42 The court has also affirmed the constitutional status of the right to marry. In Meyer v. Nebraska, 262 U.S. 390 (1923), at p. 399, the court found that the liberty protected by the Fourteenth Amendment includes "the right of the individual . . . to marry, establish a home and bring up children", and in Skinner v. Oklahoma, 316 U.S. 535 (1942), at p. 541, that "[m]arriage and procreation are fundamental to the very existence and survival of the race". Nor can one forget Douglas J.'s eloquent words in Griswold v. Connecticut, 381 U.S. 479 (1965), at p. 486, in finding that freedom of choice in marriage and family relationships lies at the heart of the right to privacy:

We deal with a right of privacy older than the Bill of Rights ^__ older than our political parties, older than our school system. Marriage is a coming together for better or for worse, hopefully enduring, and intimate to the degree of being sacred. It is an association that promotes a way of life, not causes; a harmony in living, not political faiths; a bilateral loyalty, not commercial or social projects. Yet it is an association for as noble a purpose as any involved in our prior decisions.

These sentiments were repeated by Warren C.J. in Loving v. Virginia, 388 U.S. 1 (1967), at p. 12: "The freedom to marry has long been recognized as one of the vital personal rights essential to the orderly pursuit of happiness by free men." (See also Boddie v. Connecticut, 401 U.S. 371 (1971), at p. 374; Cleveland Board of Education v. LaFleur, 414 U.S. 632 (1974), at pp. 639-40; Moore v. East Cleveland, 431 U.S. 494 (1977), at p. 499; and Zablocki v. Redhail, 434 U.S. 374 (1978), at pp. 384-85.)

43 The California courts have also affirmed on several occasions the State's legitimate interest in fostering marriage as a social institution (see Marvin v. Marvin, 557 P.2d 106 (Cal. 1976), at p. 122; Elden v. Sheldon, 758 P.2d 582 (Cal. 1988), at pp. 586-87; Beaty v. Truck Insurance Exchange, 8 Cal.Rptr.2d 593 (Ct.App. 3 Dist. 1992), at p. 600).

44 Moving from domestic to international law, art. 16 of the Universal Declaration of Human Rights, G.A. Res. 217 A (III), U.N. Doc. A/810, at 71 (1948), which is binding on Canada, and art. 12 of the European Convention on Human Rights, 213 U.N.T.S. 221, provide individuals with "the right to marry". For example, art. 16 of the Universal Declaration states that "Men and women of full age, without any limitation due to race, nationality or religion, have the right to marry and to found a family. They are entitled to equal rights as to marriage, during marriage and at its dissolution".

45 This brief review of domestic and international law confirms what in my view requires little if any confirmation, namely that marriage is both a basic social institution and a fundamental right which states can legitimately legislate to foster.

(2) The Contractual Basis of Marriage

46 In considering the particular attributes of the ground at issue, it is also important to remember that married status, at least in our society, can only be acquired by the expression of the individual's personal, free choice, regardless of the reason for which that status is assumed. Marriage rests upon a contractual basis, to which the law attaches certain rights and obligations. The decision to marry includes the acceptance of various legal consequences incident to the institution of marriage, including the obligation of mutual support between spouses and the support and raising of children of the marriage. In my view, freedom of choice and the contractual nature of marriage are crucial to understanding why distinctions premised on marital status are not necessarily discriminatory: where individuals choose not to marry, it would undermine the choice they have made if the state were to impose upon them the very same burdens and benefits which it imposes upon married persons. The authors Michael D. A. Freeman and Christina M. Lyon, in Cohabitation without Marriage (1983), at p. 191, make just these points:

. . . marriage is a voluntary institution in which the parties express their willingness to commit themselves to each other for life. Whether they are completely cognisant of all the legal effects of such a commitment is immaterial; the commitment is made, nevertheless, and marital rights and obligations inevitably follow. Cohabiting couples do not make that same commitment, and rights and duties akin to marriage should not as a result follow. The danger with imposing the incidents of marriage on a cohabiting couple is that it constitutes a denial of a fundamental freedom.

47 An additional element distinguishing marriage from other relationships is the commitment towards permanence accepted by the parties to the marriage contract. This contractual term "places it in a different category of relational interests than if it were temporary" (Bruce C. Hafen, "The Constitutional Status of Marriage, Kinship, and Sexual Privacy ^__ Balancing the Individual and Social Interests" (1983), 81 Mich. L. Rev. 463, at p. 486). While I accept as a fact that some non-marital relationships do indeed endure as long as some marriages, my point is rather that the commitment towards permanence is a defining characteristic of the contract of marriage.

48 Similarly, while one may speculate that in many instances only one partner in a cohabiting couple does not want to marry, this clearly does not imply that the decision to marry is any less a matter of choice. The decision to marry or not is, admittedly, a joint choice, but a choice nonetheless. Simply because one party prefers not to marry does not entitle a couple to all the benefits which the legislature uniquely attaches to marriage.

49 In my opinion, distinctions drawn on the basis of the status, burdens and benefits acquired through marriage cannot, without more, be discriminatory under s. 15(1) of the Charter, since these attributes are acquired through contract. One cannot claim discrimination solely on the basis that the status, rights and obligations validly contracted for differ from those prevailing in the absence of contract. Distinctions premised on marital status are therefore relevant to laws which are aimed at defining marriage, its effects and the rights and obligations to which it gives rise. Stated otherwise, the attributes of married status cannot give rise to discrimination as against those who are not married. In this aspect, marital status is not an analogous ground. In the case at bar, I will conclude that the benefits claimed by the appellants, namely the rehabilitation and medical compensation and the income benefits, are most appropriately characterized as falling within the scope of the support obligations imposed by law on married spouses. In my view, marital status is therefore relevant to the purposes underlying the distinction contained in the insurance policy, and accordingly, that distinction does not infringe s. 15(1).

C. Applying the Section 15(1) Analysis to This Case

(1) Step One: Has a Distinction Been Drawn by the Legislation?

50 Under the first step of the s. 15(1) analysis, one must determine whether the impugned legislation creates a distinction between the appellants and one or more groups. In this case, the relevant group to be compared is married spouses, and indeed, the appellants, an unmarried couple, urge the Court to compare their situation with that of married spouses. They submit that, as an unmarried couple, they are in a relationship akin to marriage, that is, a relatively permanent state of interdependence and cohabitation carrying with it, in Ontario at least, an obligation of mutual support.

51 It is apparent that the insurance policy distinguishes between spouses and non-spouses. Under the definition set forth in Section B of the insurance policy, only "spouses" can claim for the loss of income and bodily injuries. As I mentioned above, I agree with McLachlin J. that the meaning of "spouse" for the purpose of this section of the insurance policy is limited to married couples. Thus, the legislation treats married and unmarried couples in a different manner.

(2)Step Two: Is There Prejudice Resulting from the Legislative Distinction?

52 I turn now to the second step of the s. 15(1) analysis, namely the determination of whether this distinction results in disadvantage. As Iacobucci J. observed in Symes, supra, at p. 761, the question at this stage is whether the differential treatment has "the effect of imposing a burden, obligation or disadvantage not imposed upon others or of withholding or limiting access to opportunities, benefits and advantages available to others".

53 It is true that the insurance policy, as regards injury and income benefits, distinguishes between married and unmarried spouses. It cannot be said, however, that this distinction is prejudicial when considered in the larger context of the rights and obligations uniquely and appropriately attached to marriage. In the words of the Ontario Court of Appeal in Leroux, supra, at pp. 620-21:

We appreciate that unmarried persons who live together do not possess some of the important rights that married persons have but, by the same token, they are not subject to many of the legal burdens and obligations of married persons. We do not think that it can be said that their overall position nets out as one of disadvantage.

While I agree with this statement of the Court of Appeal, I do not propose to dispose of the case at bar on this basis. I will, therefore, go on to consider the functional values underlying the impugned insurance legislation and their relevance to the distinction drawn between married and unmarried couples.

(2) Step Three: Are the Functional Values of the Legislation Relevant to Marital Status?

54 The third step of the s. 15(1) analysis asks the broad question of whether the alleged distinction is based upon an irrelevant personal characteristic which is enumerated in s. 15(1) or analogous thereto. In this case, this stage involves an inquiry into whether the functional values underlying the legislation are relevant to marital status. Thus, we must first inquire whether married status is a personal characteristic which qualifies as an analogous ground. In so proceeding, it is useful to turn to the enumerated grounds of s. 15(1) of the Charter. These enumerated grounds have been recognized as being the most socially and historically destructive forms of discrimination, and in many cases they are irrelevant to distinctions drawn by legislation (Andrews, supra, at p. 175).

55 While marital status is a personal characteristic, and distinctions between married and unmarried couples may be discriminatory under s. 15(1), it must be stressed that marital status has several unique characteristics which distinguish it from the grounds enumerated in s. 15(1) of the Charter. In addition to resting upon a consensual, contractual basis, marriage is a status to which the Legislature, as a reflection of its social policy, attaches a bundle of rights and obligations. In none of the enumerated grounds do we find these characteristics. For instance, while both citizenship and religion may in some instances be said to be "chosen", they do not reflect the acceptance of a legal status and rights and obligations founded in contract. Furthermore, as with other contracts, marriage must be entered into freely and voluntarily; failing this it may be annulled. This aspect also distinguishes married status from other grounds of discrimination.

56 As these comments suggest, it is imperative to be acutely sensitive to the nature of the personal characteristic and its specific attributes. This is so whether the ground of alleged discrimination is one enumerated in s. 15(1) or analogous thereto. For example, in McKinney, supra, at p. 297, La Forest J. recognized that not all enumerated grounds should be treated on the same basis. Different considerations may be applicable depending on the nature of the ground in question. For instance, in the case of age discrimination, he stated at p. 297:

. . . there are important differences between age discrimination and some of the other grounds mentioned in s. 15(1). To begin with there is nothing inherent in most of the specified grounds of discrimination, e.g., race, colour, religion, national or ethnic origin, or sex that supports any general correlation between those characteristics and ability. But that is not the case with age.

57 Furthermore, in applying this nuanced approach to the different protected grounds under s. 15(1), it may be useful in some circumstances to examine whether a party is claiming discrimination based on its membership in a group that is disadvantaged. While membership in a disadvantaged group is not an essential element, it may well be an indicium that a distinction is drawn on the basis of an irrelevant personal characteristic.

58 The appellants claim that marital status has historically been a basis for "pervasive discrimination" and, therefore, should be recognized as an analogous ground under the Charter. However, in contemporary society, unmarried couples do not constitute a distinct group suffering from stereotypes or prejudices, although they have been the subject of such prejudices in the past. In this respect, the fostering of marriage as a social institution does not stigmatize unmarried couples nor subject them to stereotypes.

59 Today, unmarried couples are not subject to legal restrictions. They may enter into binding and enforceable contracts and may agree upon their respective rights and obligations during cohabitation. Agreements may also relate to the division of assets, property, and expenses. Such agreements may be entered into by unmarried couples either during cohabitation or upon separation. Section 53 et seq. of the Ontario Family Law Act (S.O. 1986, c. 4 and R.S.O. 1990, c. F.3) expressly recognize the validity of such agreements. The validity of such contracts between unmarried couples has also been recognized in the United States (see Marvin v. Marvin, supra).

60 I also find some observations of the Attorney General of Quebec to be instructive. He explained that at the time of the broad reform of Quebec's family law in 1980, it was decided not to extend the rights and obligations attached to marriage to unmarried couples; this decision was taken with a view to respecting the choice made by unmarried couples, not to stigmatize them.

61 The appellants further submitted that marital status should be recognized as an analogous ground of discrimination in s. 15(1) of the Charter because provincial human rights legislation recognizes marital status as a prohibited ground of discrimination. In Andrews, supra, at pp. 175-76, McIntyre J. noted the usefulness of referring to provincial human rights Acts but he also emphasized their differences from s. 15(1) of the Charter. He observed, for example, that in provincial human rights statutes, the grounds of prohibited discrimination are restricted, and the prohibition itself is limited in scope though absolute in the area to which it applies. Moreover, certain limited exceptions are available by way of exemptions or defences, such as a bona fide occupational requirement. Under s. 15(1) of the Charter, by contrast, the enumerated grounds are not exhaustive nor is the scope of the anti-discrimination protection limited. Additionally, it is significant that discrimination can be justified under s. 1. It is especially important to emphasize that although human rights laws constitute a pre-eminent category of legislation and are, in some cases, described as "quasi-constitutional" in nature (Zurich Insurance Co. v. Ontario (Human Rights Commission), 1992 67 (SCC), [1992] 2 S.C.R. 321, at p. 339), they do not have the same status as s. 15 of the Charter. Most importantly, s. 15 of the Charter is a constitutional provision which constrains or compels state but not individual action. This contrast was stated clearly by Cory J. in Dickason v. University of Alberta, 1992 30 (SCC), [1992] 2 S.C.R. 1103, at pp. 1121-22, in the following terms:

Yet it must be remembered there is a crucial difference between human rights legislation and constitutional rights. Human rights legislation is aimed at regulating the actions of private individuals. The Charter's goal is to regulate and, on occasion, to constrain actions of the state.

Therefore, in my opinion, while I recognize that marital status may be an analogous ground, I would emphasize that any reference to the grounds enumerated in provincial human rights legislation for the purpose of interpreting s. 15(1) of the Charter should be made with caution.

62 In illustrating my position that marital status may be an analogous ground, I would first reiterate my earlier conclusion that the benefits and burdens which the law attaches to marriage itself cannot be considered as giving rise to discrimination as they are distinctly related to the existence of marriage. It may well be, however, that discrimination might occur from benefits granted or burdens imposed on the basis of marital status but which are not relevant to the institution of marriage. Geiger v. London Monenco Consultants Ltd. (1992), 1992 7425 (ON CA), 43 C.C.E.L. 291, illustrates the point. There, the Ontario Court of Appeal decided that an employer's policy to offer married, but not unmarried, employees at a remote job site return flights to their home was discrimination based on marital status contrary to s. 4(1) (now 5(1)) of the Ontario Human Rights Code, 1981. Robins J.A. held that marital status was unrelated to the employees' performance of their duties and responsibilities (at pp. 300-301). There was thus no connection between marital status and the nature or the functions of the employment. In such a context, marital status is an irrelevant basis of distinction between employees.

63 I would add that in Andrews, supra, at p. 196, La Forest J. also expressed this concern for the specific attributes of the personal characteristic in issue to be closely examined to determine whether and under what circumstances it might be an analogous ground. While La Forest J. found that citizenship was an analogous ground in relation to the right to practise law, he recognized that it may not be analogous in other circumstances:

There is no question that citizenship may, in some circumstances, be properly used as a defining characteristic for certain types of legitimate governmental objectives.

For example, citizenship is a relevant criterion to determine the entitlement to the right to vote in a federal election having regard to s. 3 of the Charter, and in this respect is not an analogous ground.

64 Accepting, then, that the analysis under s. 15(1) should be conducted with the specific attributes of the personal characteristic firmly in mind, I turn now to examine the nature of the benefits in the case at bar. Since it is my position that the benefits in this case are most appropriately characterized as falling within the scope of the support obligations attached to marriage, and the appellants base their argument on the similarity of those support obligations with those imposed on common law couples, I will first compare the nature and scope of the support obligations of married couples with those of common law couples. With this comparison in mind, I will then consider the relevancy of the distinction drawn by the insurance legislation in relation to its functional values.

65 One of the essential characteristics of marriage is the obligation of mutual support which rests upon the mutual lifetime commitment spouses make to each other and which endures until the dissolution of the marriage. While the support obligations of marriage are defined by legislation, it is only by freely choosing to assume married status that the parties also choose to accept those support obligations. By contrast, outside marriage, at common law, this consensual, mutual support obligation is non-existent. While it is true that the common law, through the constructive trust, may provide some relief at the end of a relationship, this is not specific to conjugal relationships and is only available with respect to property claims (Pettkus v. Becker, 1980 22 (SCC), [1980] 2 S.C.R. 834; Sorochan v. Sorochan, 1986 23 (SCC), [1986] 2 S.C.R. 38). In short, only in cases where the legislator imposes a support obligation on unmarried couples does it exist. The Ontario Family Law Act is an example of the Legislature's intervention. Under that legislation, the support obligation arises in a factual situation of cohabitation, and is therefore imposed without any regard to the will of the parties. For instance, s. 30, Part III, of the Ontario Family Law Act, which might apply to the appellants, states: "Every spouse has an obligation to provide support for himself or herself and for the other spouse, in accordance with need, to the extent that he or she is capable of doing so." For the purpose of Part III, "spouse" has been defined as including either a man and a woman who are not married to each other and have cohabited, continuously for a period of not less than three years, or who are in a relationship of some permanence if they are the natural or adoptive parents of a child. The obligation terminates when cohabitation ceases. These obligations were only imposed by the Family Law Reform Act (S.O. 1978, c. 2, s. 14) in 1978.

66 Thus, in my view, the nature of the support obligations of unmarried couples and those of married couples is in essence very different. The sources of the obligations, the conditions governing their existence and their termination differ significantly, although both obligations may be given the same title.

67 In the case at bar, however, the appellant Miron claims that he should be entitled to accident benefits and loss of wage benefits under his unmarried partner's insurance policy because they are in the same situation as married spouses. The appellants submit that since the Ontario Family Law Act treats married and unmarried individuals in the same manner with respect to support obligations, they should have access to the means allowing them to fulfil their duty.

68 I cannot accept the appellants' position that because the Family Law Act imposes support obligations in particular circumstances on some unmarried couples, then it should follow that s. 15 of the Charter requires that common law spouses must be covered by all the provisions of the Insurance Act. Such an interpretation would in effect give an advantage to unmarried couples over married spouses since they are not burdened with the same obligations.

69 Nor do I think that economic interdependence is, without further qualification, a relevant consideration in concluding that the appellants should be covered under the insurance policy. The benefits contained in the insurance policy fall within the scope of a mutual support obligation arising from marriage since they compensate for physical injury (medical and rehabilitation benefits) and material loss (income replacement benefits); without the insurance policy, a married spouse would have to provide this kind of support to his or her injured spouse. While the insurance policy clearly is concerned with economic interdependence, such interdependence is only relevant in so far as it relates to the institution of marriage. If we were to follow the appellants' reasoning, and focus on economic interdependence simpliciter in determining the scope of policy coverage required by the Charter, then even those not cohabiting for the required time or those who ceased cohabiting could claim benefits under the insurance policy. For instance, according to the appellants' arguments, if an unmarried partner suffered from a partial permanent physical injury, he or she could claim the benefits which in fact are a permanent indemnity, although cohabitation between unmarried couples unlike that of married spouses is not de jure of a permanent nature. Therefore, unmarried couples would receive an advantage from the insurance policy without having committed to mutual support obligations of the same duration.

70 I therefore conclude that, contrary to the appellants' assertion, unmarried couples are not in a situation identical to married spouses with respect to mutual support obligations.

71 Keeping the above considerations in mind, I move now to consider the relevancy of the functional values of the legislation to marital status. Both McLachlin J. and the amicus curiae identified the functional values of the benefits provided by the Insurance Act as relating to financial interdependence, and more specifically, the provision of support to families when one member is incapable of contributing to the family unit. The amicus curiae first raised this characterization of the benefit in this Court without providing any evidence.

72 I respectfully disagree with McLachlin J.'s description of the functional value of the benefits in the legislation. The very terms of the statute indicate that it was not the Legislature's intention to separate the benefit aimed at the financial well-being of families from the institution of marriage. On the contrary, in my view the Legislature was primarily concerned with defining certain benefits attached to marriage. In some cases, these benefits are extended to unmarried couples, but that does not change the essential character of the benefits, which is to provide support for marital relationships. Indeed, as the amicus curiae emphasized in relation to the amendments of this legislation since 1971, the Legislator's search "was directed towards defining a `marriage-like' conjugal relationship". Thus, the functional value of the benefits is not to provide support for all family units living in a state of financial interdependence, but rather, the Legislature's intention was to assist those couples who are married, or, as in subsequent legislation, to assist certain prescribed couples who are in a "marriage-like relationship".

73 Furthermore, in my opinion it is clearly within the range of legitimate social policy for the Legislature to define the scope of a "marriage-like relationship". In other words, the functional value identified in this legislation, namely the support of marriage, is not itself discriminatory. Distinctions as to the scope of the institution and the benefits which attach thereto are properly the objects of legislative definition; assessing the legitimacy of those definitions must necessarily take into account the fundamental position of the institution of marriage in our society.

74 In this case, the appellants and others like them do not fall within the scope of the Legislature's definition of "spouse", since as noted above, "spouse" refers exclusively to married spouses. The Insurance Act demonstrates that where the Ontario Legislature has decided to extend some of the benefits conferred on married couples to unmarried couples, it expressly so provides and defines the conditions (such as the number of years of cohabitation, or the presence of a child). Thus, while the Legislature introduced the no-fault insurance benefits in 1971, as the amicus curiae pointed out it was not necessary at that time to define "spouse" as that term would have undoubtedly been interpreted as "married spouse". Only in 1978 did the Legislature modify the Insurance Act and decide to extend death benefits to unmarried couples. This was done as a remedial measure because without that amendment, the "lawful" surviving spouse would receive such benefits, even if the married spouses had been separated for years previously and the deceased spouse had been cohabiting at the time of his or her death with an unmarried partner. The amendment thus provided the death benefit to the unmarried partner over the lawful spouse, thereby extending to unmarried cohabitees a benefit previously restricted to married spouses.

75 It should be noted that after this case was initiated, the Legislature amended the Insurance Act in 1990 to extend the definition of "spouse" to include heterosexual couples who have cohabited for three years or live in a permanent relationship with a child. In enlarging the definition of "spouse", the Legislature was again simply carrying out its legitimate function of defining a "marriage-like relationship" for the purposes of benefits legislation. As in the case of the extension of a mutual support obligation to some unmarried couples, it is a matter of social policy for legislators to decide when and to what extent the attributes of marriage and their consequences should be extended and imposed upon unmarried couples, as indeed the Ontario Legislature decided to do in 1990 in amending the Insurance Act.

76 As a corollary of my position that it is within the scope of legitimate social policy for the Legislature to define the scope of "marriage-like relationships", I would add that there is no obligation on the Legislature to extend all the attributes of marriage to unmarried couples. As La Forest J. observed in Andrews, supra, at p. 194:

. . . it was never intended in enacting s. 15 that it become a tool for the wholesale subjection to judicial scrutiny of variegated legislative choices in no way infringing on values fundamental to a free and democratic society. . . . I am not prepared to accept that all legislative classifications must be rationally supportable before the courts. Much economic and social policy-making is simply beyond the institutional competence of the courts: their role is to protect against incursions on fundamental values, not to second guess policy decisions. [Emphasis in original.]

In short, it is my view that a Legislature may as a matter of social policy choose whether and under what circumstances to extend some or all of the attributes of marriage to unmarried couples without running afoul of s. 15(1) of the Charter. Indeed, as I suggested above, to extend all the attributes of marriage to unmarried couples would interfere directly with the individual's freedom to voluntarily choose whether to enter the institution of marriage by imposing consequences on cohabitation without any regard to the will of the parties.

77 The courts must therefore be wary of second-guessing legislative social policy choices relating to the status, rights and obligations of marriage, a basic institution of our society intimately related to its fundamental values. Barring evidence of a change in these values by a clear consensus that there should be a constitutional constraint on the powers of the state to legislate in relation to marriage, the matter must remain within the scope of legitimate legislative action.

78 In closing, I note that similar views on the ambit of legitimate legislative discretion in defining the attributes of marriage have been asserted elsewhere. For instance, although the American approach to discrimination differs from that under our Charter, the Supreme Court of California in Norman v. Unemployment Insurance Appeals Board, 663 P.2d 904 (Cal. 1983), at pp. 907‑8, reaffirmed the conclusion it had reached in Marvin, supra, that "[i]t is for the Legislature to determine whether such relationships [unmarried couples], because of their commonness in today's society or for other policy reasons, deserve the statutory protection afforded the sanctity of the marriage union". The California Court of Appeal also expressed similar views in the case of Hendrix v. General Motors Corp., 193 Cal.Rptr. 922 (Ct.App. 1 Dist. 1983), at p. 925:

This strong public policy [of fostering the institution of marriage] would be thwarted if persons could gain marital legal rights without accepting the correlative marital legal responsibilities. . . . Were this court to extend to unmarried persons legal rights heretofore confined to married persons, we would overstep our authority and usurp the authority of the Legislature to set public policy. Only the Legislature responsible to the electorate should have the power to make such a radical change in the fabric of society.

79 This American approach accurately points out that it falls within the scope of the Legislature's legitimate authority to determine the conditions under which unmarried couples should benefit from rights attached to marriage without having to assume the correlative obligations. In the context of legislation defining the rights and obligations attached to marriage, marital status is a relevant ground of distinction. Moreover, s. 15 of the Charter does not compel the Legislature to extend the status, benefits or burdens of marriage to unmarried couples.

V - Conclusion

80 To conclude, in my view, the benefits the appellants claim under the insurance policy are most appropriately characterized as relating to the support obligations existing between married spouses. In that context, marital status cannot be a ground of discrimination since the distinction pertains to an inherent aspect of marriage, namely support obligations, and the function of the impugned provisions of this legislation is relevant to that status. Therefore, I would dismiss the appeal with costs.

The following are the reasons delivered by

81 L'Heureux-Dubé J. -- Although I agree with the result reached in the instant case by Justice McLachlin, I arrive at this conclusion somewhat differently. For the reasons I set out in Egan v. Canada, 1995 98 (SCC), [1995] 2 S.C.R. 513, released concurrently I prefer to focus on the group adversely affected by the distinction as well as on the nature of the interest affected, rather than on the grounds of the impugned distinction. What follows, therefore, is my application of the framework developed in Egan to the facts of this case.

82 This case raises the question of the definition of "spouse", as that term is used in Section B, Subsection 2, and Section B, Subsection 3, Part II, of the Ontario Standard Automobile Policy. Although the term "spouse" is not defined anywhere in the legislation, I will assume without deciding, for the purposes of the following analysis, that its statutory interpretation contemplates only married couples and therefore excludes unmarried couples that are cohabiting. I shall therefore pass directly to an examination of whether this distinction violates s. 15 of the Canadian Charter of Rights and Freedoms and, if so, whether it can be saved by s. 1 of the Charter.

A. Section 15

83 In Egan, I set out the following factors that must be established by a rights claimant before the impugned distinction will be found to be discriminatory within the meaning of s. 15 of the Charter: (1) there must be a legislative distinction; (2) this distinction must result in a denial of one of the four equality rights on the basis of the rights claimant's membership in an identifiable group; and (3) this distinction must be "discriminatory" within the meaning of s. 15. I shall address each of these factors below.

84 To begin with, assuming that the meaning of "spouse" in the Standard Automobile Policy is limited to married couples, then it follows that a distinction is, indeed, made in the legislation.

85 The second question requires inquiry into whether the distinction has the effect of imposing a burden, obligation, or disadvantage not imposed on others, or of withholding or limiting access to opportunities, benefits and advantages available to others. The parameters of this inquiry cannot, however, generally be established without first ascertaining the appropriate basis for comparison. In other words, only once the appropriate comparator group has been identified can it be decided that there has been a denial of one of the four equality rights.

86 The respondents argue that the fact that the cohabiting appellants seek to compare themselves to persons who are married is, essentially, returning to the similarly situated test rejected by this Court in Andrews v. Law Society of British Columbia, 1989 2 (SCC), [1989] 1 S.C.R. 143. In my view, this argument is incorrect. The similarly situated test, in essence, requires that persons who are "similarly situated be similarly treated", and that persons who are "differently situated be differently treated". It was rejected by this Court on the basis that it contemplated only formal, Aristotelian equality, and because it excluded any consideration of the nature of the impugned law itself: Andrews, supra, at pp. 165-68. In rejecting the similarly situated test as a means to identify discrimination, however, this Court did not reject in principle the process of drawing comparisons between groups. In fact, McIntyre J. noted at p. 164 that

[Equality] is a comparative concept, the condition of which may only be attained or discerned by comparison with the condition of others in the social and political setting in which the question arises.

In short, comparisons between different groups are necessary to discern the differential effect of the legislation and to assist the court in properly characterizing and identifying the groups that are relevant to the particular s. 15 inquiry at hand.

87 The question then becomes, who or what do we compare for the purpose of determining whether an equality right has been denied? In the instant case, the interpretation of "spouse" in the Standard Automobile Policy distinguishes between persons who are married and persons who are not married, and results in the extension of a benefit only to the former. The group of persons who are not married, however, comprises many subsets: for instance, persons cohabiting in a conjugal relationship (e.g. common law spouses), persons cohabiting who are not in a conjugal relationship (e.g. roommates), persons who are related (e.g. brother-sister), and single persons.

88 The guarantee of equality in s. 15 does not require that the entire, collective, heterogenous group of non-married persons be compared against the essentially homogenous group of married persons. In fact, uncritical comparison of dissimilar groups can undermine the purposes of s. 15 of the Charter rather than further them. Comparison is only a fruitful exercise when carried out between groups that possess sufficient analogous qualities to make the exercise of comparison meaningful in respect of the distinction being examined. Thus, in the present case, the only appropriate comparison is between married persons and unmarried persons who are in a relationship analogous to marriage (i.e. of some degree of publicly acknowledged permanence and interdependence). In other words, with all other things being roughly equal, the latter group is denied the equal benefit of the law for essentially one reason: the fact that they are not married.

89 What constitutes a relationship analogous to marriage is, of course, potentially a subject of considerable debate. I prefer not to engage in this debate. I am, however, satisfied that the appellants Miron and Valliere fall within this group. In August 1987, at the time of the accident which resulted in the present litigation, they had been cohabiting as common law spouses for over four years, and Miron was the father of two of Valliere's three children, aged two and five. Miron and Valliere would fall within any number of legislatively accepted definitions of "relationship analogous to marriage". For this reason, I am satisfied that the impugned legislation denies Miron and Valliere the equal benefit of the law on the basis that they are in a relationship analogous to marriage.

90 The last ingredient in a s. 15 analysis is an inquiry into whether the distinction is "discriminatory" within the meaning of s. 15 of the Charter. In Egan, supra, I stated that a distinction would be discriminatory within the meaning of s. 15 where it is capable of either promoting or perpetuating the view that the individual adversely affected by this distinction is less capable, or less worthy of recognition or value as a human being or as a member of Canadian society, equally deserving of concern, respect, and consideration. I noted, as well, that this examination should be undertaken from a subjective-objective perspective. To give form and substance to this examination, I elaborated upon how discriminatory impact can be assessed by looking to the nature of both the interest and the group adversely affected by the impugned distinction. I observed, moreover, that neither category of factors will generally yield meaningful results without consideration of the other. It is to these factors that I now turn.

1. The Nature of the Group Affected

91 The question of whether or not persons in relationships analogous to marriage have typically suffered historical disadvantage is not clear-cut, partly because the modern phenomenon of common law cohabitation as an alternative to marriage is a comparatively recent one. The subgroups within the ground of marital status that have typically suffered the most historical disadvantage and marginalization are individuals who are single parents, or are divorced or separated. The mere fact that common law spouses are not the first group that comes to mind when considering historical disadvantage does not mean, however, that such relationships have escaped completely from societal opprobrium. In fact, non-traditional relationships outside of marriage have in the past generally been frowned upon and considered undesirable by large portions of society. Only recently have they come to be increasingly accepted. That they have become more accepted does not mean, however, that they are now accepted without reservation into the mainstream of society. The assumptions and attitudes underlying the continuing societal disapproval of such relationships are, in many ways, both exemplified and perpetuated by the impugned law, which essentially deems that John Miron and Jocelyne Valliere, a couple who at the time of the accident had been living together as common law spouses for over four years and had two children together, are not an interdependent family unit worthy of protection against the potentially catastrophic financial losses flowing from injury to one of their members.

92 The group of persons in relationships analogous to marriage is, of course, not homogenous. Clearly, the effect of a distinction will be felt more severely by certain persons within that group than by others. In the present case, sensitivity to what is reasonable and representative of the group of persons in relationships analogous to marriage therefore requires that we examine the issue from the perspective of both men and women. Although I will elaborate on this point very shortly, I think it important to note at this juncture that courts and legislatures have already acknowledged that the effect of a distinction based on marital status may be felt more severely by women than by men. I therefore have no difficulty concluding that persons in opposite-sex relationships analogous to marriage have suffered, and continue to suffer, some disadvantage, disapproval and marginalization in society, and are therefore somewhat sensitive to legislative distinctions having prejudicial effects.

93 Membership in a discrete and insular minority that is politically weak is another relevant consideration identified in Andrews and subsequent jurisprudence. I am not convinced, however, that it can necessarily be said that persons in relationships analogous to marriage represent a discrete and insular minority, or a politically powerless or vulnerable group. I emphasize, though, that the presence or absence of this factor alone, like any others, cannot be dispositive of the s. 15 analysis.

94 The last important factor to be considered is whether the impugned distinction is based upon a fundamental attribute of personness' or humanness'. Many argue that marriage is a choice, and that cohabiting persons who are not married have often made a conscious choice to avoid the rights and obligations incumbent upon marriage. Consequently, since marriage is a legal status which people can assume by choice in a manner analogous to contract, it is argued that a distinction based upon this status does not touch upon anything that is sufficiently personal that it could be a basis for discrimination. A corollary of this argument is that treating common law couples in a manner similar to married couples would undermine people's "free choice" to contract in or out of these rights and obligations. Although the absence or presence of an element of "choice" should not be determinative of s. 15 analysis, my colleague Justice Gonthier addresses it in his reasons and I feel compelled to make some observations in this respect.

95 In my view, the decision of whether or not to marry can, indeed, be one of the most personal decisions an individual will ever make over the course of his or her lifetime. It can be as fundamental, as momentous, and as personal as a choice regarding, for instance, one's citizenship or even one's religion. Although certain rights and obligations follow from each one of these three diverse choices, it does not render any of these choices justice to reduce them to a question of contract. I highly doubt, for instance, that people enter the institution of marriage because it strikes them as offering an attractive package of contractual rights and obligations. By that same token, people who make a conscious decision not to subscribe to the institution of marriage may very well be motivated by very personal beliefs which have nothing whatsoever to do with the contractual rights and obligations that incidentally attach to that status.

96 Beyond this preliminary objection, however, my disagreement with the assertion that marriage is simply a matter of individual choice goes far deeper. In particular, I believe that this argument is premised upon an important and, in my mind, unchallenged assumption: that the majority of unmarried persons living in a relationship of some interdependence and duration are, indeed, exercising a "free choice". In my respectful view, this assumption may mischaracterize the reality of a significant number of persons in non-traditional relationships. This silent and oft-forgotten group constitutes couples in which one person wishes to be in a relationship of publicly acknowledged permanence and interdependence and the other does not:

The two partners in a relationship may not have similar views: while one partner may value personal autonomy, that view may not be shared by the other. One may in fact be anxious to marry, while the other resists it. Whose view of the relationship is to prevail? . . . The flip side of one person's autonomy is often another's exploitation. [Emphasis added.]

(W. Holland, "Marriage and Cohabitation -- Has the Time Come to Bridge the Gap?" in Family Law: Roles, Fairness and Equality, Special Lectures of the Law Society of Upper Canada 1993, 369, at p. 380.) It is small consolation, indeed, to be told that one has been denied equal protection under the Charter by virtue of the fact that one's partner had a choice.

97 Both the courts and the legislatures have, in recent years, acknowledged and responded to the injustices that often flow from power imbalances of this type and have thereby given increasing recognition to non-traditional forms of relationships. Why else did the Ontario legislature in 1986 extend benefits from married persons to cohabiting partners in over 30 Ontario statutes, several of which raised issues of financial interdependence that are analogous to the impugned provisions of the Insurance Act? Why else has the Ontario Family Law Act, R.S.O. 1990, c. F.3, imposed an obligation of mutual support on common law spouses since 1978? Why else has the common law doctrine of constructive trust intervened to provide relief to unmarried persons in instances where one partner is unjustly enriched by the relationship? Pettkus v. Becker, 1980 22 (SCC), [1980] 2 S.C.R. 834; Sorochan v. Sorochan, 1986 23 (SCC), [1986] 2 S.C.R. 38. Finally, why else has this Court rejected the hegemony of the "clean break" model of spousal support in both married and common law relationships of demonstrated duration and interdependence? Moge v. Moge, 1992 25 (SCC), [1992] 3 S.C.R. 813. In all of these cases, although the language has generally employed gender-neutral references to "spouses", it is indisputable that much of the impetus for these changes stemmed from courts' and legislatures' increasing recognition of the disadvantage endured by dependent spouses, most often women, within the context of those relationships.

98 If the continuing individual autonomy of the parties were the only assumption governing the formation of family units, then none of these protections would be deemed necessary. Rather, however, it is now widely accepted that one cannot speak of "autonomy" or "free choice" without first asking whose autonomy one seeks to preserve, and at what cost to others. It is extremely important that, when contemplating the impact of a particular legislative distinction on an affected group, courts make reasonable accommodation for diverse points of view and life situations.

99 I do not dispute, of course, that some cohabiting couples do, indeed, continue to agree mutually that they wish to avoid the rights and obligations of marriage. Indeed, as I noted earlier, their reasons for doing so may be intensely personal, and their alternative forms of relationships are equally deserving of respect, consideration, and protection under the Charter. My point is merely that it is dangerous, and possibly untenable, to assume for the purposes of s. 15 Charter analysis that the reasons one starts to live together are necessarily the reasons that one continues to live together. Although many couples initially cohabit as part of a "trial relationship", and may therefore initially agree not to marry, such mutual agreement may diminish over time, as the length of cohabitation increases or most particularly after the birth of a child.

100 As I noted above, in recognition of this reality, legislatures have intervened in a wide variety of contexts to protect individuals' vested interests in relationships of some permanence and interdependence. These interventions are not anti-marriage. They simply acknowledge that the family unit is evolving in response to changing times. In my respectful view, it would therefore be a significant step backwards for this Court nonetheless to conclude that "unfettered choice" is the only framework by which to measure and evaluate extramarital cohabitation. Such logic would, in effect, entail adopting a narrower approach to the realities of cohabitation under s. 15 of the Charter -- which is supposed to be interpreted broadly and purposively -- than has already been widely accepted both in the common law and in statutes throughout Canada.

101 Finally, it must be recalled that the imposition of marriage-like mutual rights and obligations upon couples in a relationship analogous to marriage need not deprive them of the autonomy required to make personal choices if these persons also have the possibility of resorting to domestic contract to exclude the effects of the legislation. Rather than placing the onus on unmarried couples to contract into any such mutual rights and obligations, inclusion within the legislation merely shifts the onus to those who wish to preserve individual autonomy to contract out.

102 To recapitulate, the decision of whether or not to marry is most definitely capable of being a very fundamental and personal choice. The importance actually ascribed to the decision to marry or, alternatively, not to marry, depends entirely on the individuals concerned. For a significant number of persons in so-called "non-traditional" relationships, however, I dare say that notions of "choice" may be illusory. It is inappropriate, in my respectful view, to condense the forces underlying the adoption of one type of family unit over another into a simple dichotomy between "choice" or "no choice". Family means different things to different people, and the failure to adopt the traditional family form of marriage may stem from a multiplicity of reasons -- all of them equally valid and all of them equally worthy of concern, respect, consideration, and protection under the law.

103 I noted earlier that persons in non-traditional relationships have suffered, and continue to suffer, some degree of disadvantage and marginalization from the mainstream. To this, I would add that I believe that it is beyond dispute that the adverse effects of distinctions which exclude persons in relationships analogous to marriage are likely to be experienced more sharply by the more dependent member of these partnerships, most often still women (in heterosexual couples). It is important not to close our eyes to that social reality. On the other hand, I believe that it is nonetheless fair to say that, on the whole, the affected group is not so sensitive as to compel the conclusion that virtually any adverse distinction would have a discriminatory impact. Therefore, an examination of the nature of the interest affected by the impugned distinction is particularly important in the present instance.

2. The Nature of the Impugned Interest

104 At issue in the present appeal is the eligibility of John Miron, the common law spouse of Jocelyne Valliere, for insurance against injury, under her automobile insurance policy. The affected interest is the protection of family units from potentially disastrous financial consequences due to the injury of one of their members. Protection of "family" is, in turn, one of the most important interests imaginable in our society. Moreover, as I observed in Canada (Attorney General) v. Mossop, 1993 164 (SCC), [1993] 1 S.C.R. 554, at p. 634:

It is possible to be pro-family without rejecting less traditional family forms. It is not anti-family to support protection for non-traditional families. The traditional family is not the only family form, and non-traditional family forms may equally advance true family values.

105 Although it goes without saying that all injured persons are entitled to that part of their health care costs covered by their provincial medicare systems, actual health care costs may often represent only a small part of the total losses suffered as a result of injury in a motor vehicle accident when loss of income as well as pain and suffering are taken into account. Equally significant, although persons ineligible to claim from a private insurance company under the Standard Automobile Policy may still claim for some compensation under the Motor Vehicle Accident Claims Act, R.S.O. 1990, c. M.41, the cost, time and difficulty of recovery by this means are significantly greater than if the person were insured by a private company. The amicus curiae points out that one of the most notable differences is that no payment whatsoever is available from the Act until the injured person has successfully obtained a formal judgment in his or her favour. Under most private insurance plans, by contrast, there is frequently no need for the injured person to bring a legal action, and the injured party will often receive advance or periodic payments. The financial consequences of these differences can be profound on a family unit, particularly if the injured party is an income-earner who has been disabled as a result of the accident. Thus, I conclude that the impugned distinction adversely affects, to an extent that is potentially very serious in economic terms, an interest which has extremely high societal value. On the other side of the equation, however, I note that the value of this interest in constitutional terms is limited, and that this distinction does not restrict access in any meaningful way to any fundamental social institution.

106 The final factor contemplated in Egan is whether the distinction constitutes a complete non-recognition of the affected group. In the instant case, we are not faced with a legislative definition of common law spouse that some would simply consider overly restrictive or underinclusive for various reasons. Rather, the impugned distinction categorically excludes from joint insurance coverage all couples in a relationship analogous to marriage. This factor must be viewed as significant, since it can be reasonably perceived as sending a clear message that society does not consider this genre of relationship to be worthy of equal protection in such contexts.

107 I noted in Egan that evaluation of either the interest or the group affected was not entirely meaningful until each was assessed in light of the other. In the instant case, I believe the impugned interest to be sufficiently pressing, the possible economic consequences to be sufficiently severe, and the manner of exclusion to be sufficiently complete to constitute a significant, though not overwhelming, discriminatory potential. I also concluded earlier that the group affected by the distinction (i.e. unmarried persons in a relationship analogous to marriage) is somewhat vulnerable and that, in a significant number of cases, persons within this group do not have meaningful control over their circumstances. I noted that the consequences of excluding unmarried persons from the benefits or protections of the law will generally be experienced more severely by the dependent spouse, who is still all too often female. Viewing all of these factors together, I conclude that the impugned distinction does, on the whole, have an impact that is discriminatory within the sense of s. 15 of the Charter. The impugned distinction is reasonably capable of either promoting or perpetuating a view amongst persons in relationships analogous to marriage that they are less worthy of recognition or value as human beings or as members of Canadian society, equally deserving of concern, respect, and consideration. On this basis, I find the impugned distinction to be in violation of s. 15(1) of the Charter.

108 Having determined that the impugned distinction violates s. 15 of the Charter, I now turn to the question of whether this distinction is relevant to a proportionate extent to a pressing and substantial objective, and may therefore be saved under s. 1.

B. Section 1

109 When characterizing the objective of the Standard Automobile Policy for the purposes of s. 1 analysis, it is important to adopt a functional and pragmatic approach which frames that purpose neither too broadly nor too narrowly. The objective of the Standard Automobile Policy, which I accept as pressing and substantial, is to protect stable family units by insuring against the economic consequences that may follow from the injury of one of the members of the family. Although I might agree that an incidental effect of this legislation, as a result of the impugned distinction, may be to encourage marriage, I cannot agree that, by reason of this distinction alone, we should conclude that the promotion of marriage is, itself, the fundamental purpose of the legislation. I cannot see people seriously considering the availability of joint automobile insurance as a factor in their decision of whether or not to marry. As such, taken as a whole, although the legislation is most certainly "pro-family", I do not believe that it would be proper to characterize it as "pro-marriage".

110 It is now trite law to observe that the onus of demonstrating a rational connection between the impugned distinction and the pressing and substantial objective lies with the government. Unfortunately, only the intervener the Attorney General of Manitoba has submitted any argument on this point. The amicus curiae appointed by this Court to make submissions on the s. 1 issue argues that the violation cannot be saved under s. 1. Thus, this Court is left in the uncomfortable position of envisioning s. 1 arguments on the government's behalf.

111 The Attorney General of Manitoba argues that restricting the meaning of "spouse" in the legislation to married couples is rationally connected to the objective because it promotes certainty and determinacy, and because it would be impossible for the legislature to settle with certainty upon a defined standard that would not be underinclusive of someone. This argument, which essentially urges judicial deference to a legislative choice, more properly goes to the question of whether or not the legislation is minimally impairing, and I shall address it in that context below.

112 It could, however, be argued that the rational connection requirement is fulfilled because it is rational for the legislature to believe that, in general, marriages are longer-lasting and more likely to engender relationships of interdependency than common law relationships. If such an argument were to be advanced, though, then this Court would require some empirical evidence to this effect from the government, since it would undermine the very purpose of s. 15 to permit the government to justify a violation of s. 15 by relying on assumptions that may, themselves, be stereotypical and discriminatory in nature.

113 In Egan, I also noted that unless the distinction relates to a right or obligation flowing from a particular legal status, it would be difficult to envision that the distinction is rationally connected to the legislative objective. As my colleague Gonthier J. points out, the common law imposes an obligation of mutual support between married partners yet does not impose that obligation upon unmarried partners. In the instant case, it could therefore be argued that the exclusion of common law couples from the Standard Automobile Policy is rationally connected to the objective of the legislation because this exclusion relates to the absence of an obligation of mutual support between unmarried individuals.

114 As I have already noted elsewhere in these reasons, however, the impugned insurance legislation is in effect in Ontario -- a jurisdiction in which the Family Law Act has, since 1978, prescribed a mutual obligation of support for common law spouses (see s. 29 of the Act). Thus, as at August 1987, common law spouses in Ontario were, indeed, bound by an obligation of mutual support yet were excluded from a Standard Automobile Policy whose basic purpose was almost inextricably related to that mutual obligation and to the relationship of interdependency upon which that obligation is premised. Other applicable laws form part of the social context of a distinction and I therefore cannot see how, in light of the fact that an obligation of support applies to common law couples, it can still be said that the impugned distinction is rationally connected to the purpose of the legislation.

115 I would therefore conclude that the distinction cannot be justified on the basis that the government has not demonstrated that the impugned distinction is rationally connected to the objective of the legislation.

116 Even if the rational connection test were found to be satisfied, I would also conclude that the impugned distinction fails the minimal impairment test. The unit that the legislator has decided to protect (i.e. married persons) is underinclusive of the purpose of the legislation. The argument could be made that the legislation is minimally intrusive because there is no other reasonably ascertainable standard which could further the purposes of the Act.

117 Although this submission may have merit in other contexts, it does not in the present case. Indeed, although the unit deserving of protection can be defined by marriage, it can also be defined in a workable and acceptably certain way by reference to the length of the relationship or to the existence of children. These two criteria have been recognized by the legislature as feasible indicia of interdependence in other statutes which confer rights or obligations upon relationships outside of marriage. In fact, the legislature adopted just such criteria in the instant case when it amended the impugned legislation to include common law spouses in 1990. Although deference should be had with respect to policy choices made by the legislature as to what duration of cohabitation is necessary to define such a relationship, courts should not feel obliged to be as deferential when the legislature has simply excluded other possibilities altogether, unless the government can demonstrate that this exclusion is, itself, the product of a reasonable attempt to balance competing social science or policy interests. In the present case, the government has once again not overcome that burden.

C. Remedy and Disposition

118 Underinclusive legislation raises special problems from a remedial perspective. The fact that legislation is underinclusive, however, does not make it any less discriminatory. Underinclusion is, in many ways, a backhanded way of permitting discrimination. I agree with McLachlin J. that the appropriate remedy in this case is to take the unusual step of retroactively "reading in" the definition of "spouse" adopted by the legislature in 1990 for the purposes of this very legislation. I find further comfort in this conclusion from the fact that the proposed remedy would not impose any additional burden upon the public purse. The insurance companies would have to absorb this additional cost (which, the Court has been informed, would have caused only a nominal increase in premiums to the public, of roughly 0.7 percent over the period from 1978 to 1989). Consequently, I would dispose of this matter in the manner proposed by McLachlin J.

The judgment of Sopinka, Cory, McLachlin and Iacobucci JJ. was delivered by

119 McLachlin J. -- This appeal requires us to decide whether exclusion of unmarried partners from accident benefits available to married partners violates the equality guarantees of the Canadian Charter of Rights and Freedoms. I conclude that it does.

120 The record before us posits the following facts. John Miron and Jocelyne Valliere lived together with their children. They were not married, yet their family functioned as an economic unit. In 1987, John Miron was injured while a passenger in a motor vehicle owned by the respondent William James McIsaac and driven by the respondent Richard Trudel. Neither McIsaac nor Trudel was insured. After the accident, Mr. Miron could no longer work and contribute to his family's support. He made a claim for accident benefits for loss of income and damages against Ms. Valliere's insurance policy which extended accident benefits to the "spouse" of the policy holder. The insurance company, Economical Mutual, denied his claim on the ground that Mr. Miron was not legally married to Ms. Valliere and hence not her "spouse".

121 Mr. Miron and Ms. Valliere sued the insurer. The insurer brought a preliminary motion to determine whether the word "spouse", as used in the applicable portions of the policy, includes unmarried common law spouses. The motions court judge found that "spouse" meant a person who is legally married. Mr. Miron and Ms. Valliere appealed the decision to the Ontario Court of Appeal, arguing first that Mr. Miron is a spouse under the terms of the policy, and alternatively, that the policy terms which are prescribed by the Insurance Act, R.S.O. 1980, c. 218, discriminate against him in violation of s. 15(1) of the Charter. The Court of Appeal dismissed their claim: (1991), 1991 7312 (ON CA), 4 O.R. (3d) 623, 83 D.L.R. (4th) 766, [1991] I.L.R. ¶ 1-2770, 7 C.C.L.I. (2d) 317. They now appeal to this Court.

Issues

A. Are the claimants "spouses" under the policy?

B. If not, does the limitation of benefits to married persons violate the equality provisions of the Charter?

1. The test for violation of s. 15(1) and the relationship between s. 15(1) and s. 1 of the Charter

2. Section 15(1) -- Discrimination and its grounds

3.Justification under s. 1 of the Charter

C.The Remedy

Analysis

A. Are the claimants "spouses" under the policy?

122 The insurance company contends that Mr. Miron was not a spouse under its policy because he was not legally married to Ms. Valliere. Mr. Miron objects. He submits that "spouse" in the policy extends to couples who live together in a common law relationship.

123 The wording and history of the provisions under which Mr. Miron claims pose difficulties for his argument. The benefits in question were governed by 1980 legislation. That legislation extended the accident and loss of income benefits in question to the "spouse" of the insured, left undefined. By contrast, the same legislation included in "spouse" for the purposes of the death benefit provisions a man and woman who were not married to each other but who had cohabited continuously for five years or lived in a relationship of some permanence and had a child. In 1990, the Legislature amended the Act and expanded the definition of "spouse" in relation to the benefits Miron and Valliere claim. The new definition of "spouse" includes a heterosexual couple who have cohabited for three years or who have lived in a permanent relationship with a child.

124 The extended definition of "spouse" in the 1980 legislation for death benefits and in the 1990 legislation for the benefits here at issue belies the suggestion that the Legislature in 1980 intended the term "spouse" to apply to unmarried partners. In fact, where the Legislature wanted to extend benefits to such persons, it expressly did so. In the face of this, the submissions that "spouse" is an ambiguous term, which ambiguity should be resolved in favour of the insured, cannot prevail.

125 I conclude that "spouse" in the 1980 provisions relating to loss of income benefits and uninsured motorist claims did not include unmarried couples living in a common law relationship.

B. Does the limitation of benefits to married persons violate the equality provisions of the Charter?

126 Mr. Miron and Ms. Valliere advance an alternative argument: that denial of benefits to them on the ground that they are not legally married and hence "spouses" violates the equality provisions of the Charter.

1. The Canadian Charter of Rights and Freedoms guarantees the rights and freedoms set out in it subject only to such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society.

2. (1) Every individual is equal before and under the law and has the right to the equal protection and equal benefit of the law without discrimination and, in particular, without discrimination based on race, national or ethnic origin, colour, religion, sex, age or mental or physical disability.

3. The test for violation of s. 15(1) and the relationship between s. 15(1) and s. 1 of the Charter

127 The early days of the Charter saw debate on the division of the equality analysis between s. 15(1) and s. 1. Some thought that denial of equality on any ground would establish discrimination under s. 15(1), propelling the analysis immediately to s. 1: Hogg, Constitutional Law of Canada (2nd ed. 1985), at p. 800. The other extreme held that s. 15(1) could be satisfied only by showing that there had been a denial of equality which was irrational or unreasonable, leaving little for s. 1: Andrews v. Law Society of British Columbia (1986), 1986 1287 (BC CA), 27 D.L.R. (4th) 600 (B.C.C.A.). This Court in Andrews v. Law Society of British Columbia, 1989 2 (SCC), [1989] 1 S.C.R. 143, rejected both approaches, charting instead a middle course.

128 The analysis under s. 15(1) involves two steps. First, the claimant must show a denial of "equal protection" or "equal benefit" of the law, as compared with some other person. Second, the claimant must show that the denial constitutes discrimination. At this second stage, in order for discrimination to be made out, the claimant must show that the denial rests on one of the grounds enumerated in s. 15(1) or an analogous ground and that the unequal treatment is based on the stereotypical application of presumed group or personal characteristics. If the claimant meets the onus under this analysis, violation of s. 15(1) is established. The onus then shifts to the party seeking to uphold the law, usually the state, to justify the discrimination as "demonstrably justified in a free and democratic society" under s. 1 of the Charter.

129 This shift of the burden through the use of s. 15 and s. 1 is appropriate. It places the duty of adducing proof upon the parties who are in the best position to adduce it. It is for the claimant to show that he or she has been denied a benefit or suffers a disadvantage compared with another person. It is also for the claimant to show the basis for imposing the burden or withholding the benefit. These matters are within the knowledge of the claimant. Once these have been made out the burden shifts to the state. It is the state's law that has violated the individual's equality on suspect grounds, and it is the state that most appropriately defends the violation. To require the claimant to prove that the unequal treatment suffered is irrational or unreasonable or founded on irrelevant considerations would be to require the claimant to lead evidence on state goals, and often to put proof of discrimination beyond the reach of the ordinary person. Nor is the resultant burden unjust to the state: while it is open to the state to attempt to differentiate on suspect stereotypical grounds, it must be prepared to justify such suspect differentiation if it wishes its law to stand. In cases such as the present, where the party upholding the law is a non-state actor, it is always open to the state to defend its law as an intervener in the proceedings. (If still in doubt as to the law's purpose and rationale, the court may also appoint an amicus curiae to assist the court by providing an impartial assessment, as was done in this appeal.)

130 This division of the analysis between s. 15(1) and s. 1 accords with the injunction to which this Court has adhered from the earliest Charter cases: courts should interpret the enumerated rights in a broad and generous fashion, leaving the task of narrowing the prima facie protection thus granted to conform to conflicting social and legislative interests to s. 1. See Hunter v. Southam Inc., 1984 33 (SCC), [1984] 2 S.C.R. 145, at p. 156; R. v. Big M Drug Mart Ltd., 1985 69 (SCC), [1985] 1 S.C.R. 295, at p. 344; Re B.C. Motor Vehicle Act, 1985 81 (SCC), [1985] 2 S.C.R. 486, at p. 509; United States v. Cotroni, 1989 106 (SCC), [1989] 1 S.C.R. 1469, at p. 1480. It is significant that where the Charter seeks to narrow rights by concepts like reasonableness, it does so expressly, as in s. 8 and s. 11(b). Section 15(1) does not contain this sort of limitation.

131 At the same time, this approach does not trivialize s. 15(1) by calling all distinctions discrimination. Unequal treatment alone -- the mere fact of making a distinction -- does not establish a breach of s. 15(1) of the Charter. The s. 15(1) guarantee relied on is ". . . equal benefit of the law without discrimination". To prove discrimination, the claimant must show that the unequal treatment is based on one of the grounds expressly mentioned in s. 15(1) -- race, national or ethnic origin, colour, religion, sex, age or mental or physical disability -- or some analogous ground. These grounds serve as a filter to separate trivial inequities from those worthy of constitutional protection. They reflect the overarching purpose of the equality guarantee in the Charter -- to prevent the violation of human dignity and freedom by imposing limitations, disadvantages or burdens through the stereotypical application of presumed group characteristics rather than on the basis of individual merit, capacity, or circumstance.

132 The enumerated and analogous grounds serve as ready indicators of discrimination because distinctions made on these grounds are typically stereotypical, being based on presumed rather than actual characteristics. Nevertheless, in some situations distinctions made on enumerated or analogous grounds may prove to be, upon examination, non-discriminatory. For example, the distinction may be found not to engage the purpose of the Charter guarantee. Thus in R. v. Turpin, 1989 98 (SCC), [1989] 1 S.C.R. 1296, at p. 1333, Wilson J., while leaving open the possibility that province of residence could be an analogous ground, held that it was not used in a way that engaged the purpose of s. 15 in that case. See also McKinney v. University of Guelph, 1990 60 (SCC), [1990] 3 S.C.R. 229, at pp. 392-93 per Wilson J.; R. v. Swain, 1991 104 (SCC), [1991] 1 S.C.R. 933, at p. 992 per Lamer C.J.; and Symes v. Canada, 1993 55 (SCC), [1993] 4 S.C.R. 695, at p. 761 per Iacobucci J. Furthermore, if the law distinguishes on an enumerated or analogous ground but does not have the effect of imposing a real disadvantage in the social and political context of the claim, it may similarly be found not to violate s. 15: Weatherall v. Canada (Attorney General), 1993 112 (SCC), [1993] 2 S.C.R. 872. Cases where a distinction made on an enumerated or analogous ground does not amount to discrimination, however, are rare. Faced with a denial of equal benefit based on an enumerated or analogous ground, one would be hard pressed to show that the distinction is not discriminatory: Andrews, supra, at pp. 174-75, and McKinney v. University of Guelph, supra, at pp. 392-93 per Wilson J.

133 My colleague Gonthier J. asserts that discrimination under s. 15(1) is conclusively rebutted by a finding that the ground on which the equal treatment is denied is relevant to the legislative goal or the functional values underlying the impugned law. With respect, I cannot agree. Proof that the enumerated or analogous ground founding a denial of equality is relevant to a legislative goal may assist in showing that the case falls into the class of rare cases where such distinctions do not violate the equality guarantees of s. 15(1), serving as an indicator that the legislator has not made the distinction on stereotypical assumptions about group characteristics. However, relevance is only one factor in determining whether a distinction on an enumerated or analogous ground is discriminatory in the social and political context of each case. A finding that the distinction is relevant to the legislative purpose will not in and of itself support the conclusion that there is no discrimination. The inquiry cannot stop there; it is always necessary to bear in mind that the purpose of s. 15(1) is to prevent the violation of human dignity and freedom through the stereotypical application of presumed group characteristics. If the basis of the distinction on an enumerated or analogous ground is clearly irrelevant to the functional values of the legislation, then the distinction will be discriminatory. However, it does not follow from a finding that a group characteristic is relevant to the legislative aim that the legislator has employed that characteristic in a manner which does not perpetuate limitations, burdens and disadvantages in violation of s. 15(1). This can be ascertained only by examining the effect or impact of the distinction in the social and economic context of the legislation and the lives of the individuals it touches.

134 In approaching the concept of relevance within s. 15(1), great care must be taken in characterizing the functional values of the legislation. My colleague Gonthier J. concedes that the distinction here at issue -- denial on the basis of marital status -- might, for some purposes, be viewed as an analogous ground. He asserts, however, that it is not used in a discriminatory manner in this case because "the functional value of the benefits is not to provide support for all family units living in a state of financial interdependence, but rather, the Legislature's intention was to assist those couples who are married" (para. 72). He concludes that distinguishing on the basis of marital status is relevant to this purpose and hence that the law is not discriminatory. On examination, the reasoning may be seen as circular. Having defined the functional values underlying the legislation in terms of the alleged discriminatory ground, it follows of necessity that the basis of the distinction is relevant to the legislative aim. This illustrates the aridity of relying on the formal test of logical relevance as proof of non-discrimination under s. 15(1). The only way to break out of the logical circle is to examine the actual impact of the distinction on members of the targeted group. This, as I understand it, is the lesson of the early decisions of this Court under s. 15(1). The focus of the s. 15(1) analysis must remain fixed on the purpose of the equality guarantees which is to prevent the imposition of limitations, disadvantages or burdens through the stereotypical application of presumed group characteristics in violation of human dignity and freedom.

135 The same criticism can be made of La Forest J.'s reasoning in Egan v. Canada, 1995 98 (SCC), [1995] 2 S.C.R. 513, released concurrently. La Forest J. characterizes the functional value of the legislation as meeting the need to support married couples who are elderly. Because, in his view, marriage is "firmly anchored in the biological and social realities that heterosexual couples have the unique ability to procreate" (para. 21), Parliament may use the relevant ground of sexual orientation as a basis for distinguishing who should receive benefits under the Act. By defining the legislative aim in terms of the alleged discriminatory ground, namely married couples, the relevance of the ground is assured. On the assumption -- misplaced in my view -- that this relevance suffices to negate discrimination, s. 15(1) is said to be met without examining the actual impact of the legislation on members of groups who may be disadvantaged by the distinction.

136 The danger of using relevance as a complete answer to the question of whether discrimination is made out, and thus of losing sight of the values underlying s. 15(1), is acute when one is dealing with so-called "biological" differences. This is the lesson of Bliss v. Attorney General of Canada, 1978 25 (SCC), [1979] 1 S.C.R. 183, and Brooks v. Canada Safeway Ltd., 1989 96 (SCC), [1989] 1 S.C.R. 1219. In Bliss, a Bill of Rights case, this Court denied benefits to pregnant women under the Unemployment Insurance Act, 1971, on the reasoning that the distinction drawn under the Act was based on relevant biological differences. Ten years later, in Brooks, this Court acknowledged that the superficial relevance of the biological difference between women and men had led it astray in Bliss. The ultimate issue was whether the impugned distinction denied benefits to a class of people -- pregnant women -- in a way which was discriminatory on the basis of sex. In the result, the Court concluded that the denial of benefits had the effect of denying equality to women, the only class of persons who could become pregnant, and unfairly placed an economic burden due to pregnancy solely on the shoulders of women. Much as this Court did in Bliss, La Forest J. relies on the biological differences between heterosexual and homosexual couples to find that the Old Age Security Act does not discriminate on the basis of sexual orientation. Following the lesson of Brooks, I would respectfully suggest that more is required; if we are not to undermine the promise of equality in s. 15(1) of the Charter, we must go beyond biological differences and examine the impact of the impugned distinction in its social and economic context to determine whether it, in fact, perpetuates the undesirable stereotyping which s. 15(1) aims to eradicate.

137 Relevance as the ultimate indicator of non-discrimination suffers from the disadvantage that it may validate distinctions which violate the purpose of s. 15(1). A second problem is that it may lead to enquiries better pursued under s. 1. As pointed out by this Court in Andrews v. Law Society of British Columbia, supra, an analysis within s. 15 of whether the distinction was reasonable leaves little to s. 1, because in determining reasonableness, one must look at the conflicting state interest and determine if its importance outweighs the denial of equality. The same difficulties arise with asking whether the unequal treatment is justified because the distinction is relevant to the legislative goal. If any professed relevance suffices, unevaluated and unweighed, then few claims would pass s. 15(1). On the other hand, an evaluation of the degree of relevance of the ground of distinction to the legislative goal necessarily involves weighing the legislative purpose against the seriousness of the unequal treatment. Under the scheme of the Charter, such questions are better posed under s. 1.

138 Dividing the analysis between s. 15(1) and s. 1 as I have suggested also corresponds to the Canadian practice under human rights codes. Typically, these codes prohibit distinctions made on specified grounds, similar to those enumerated in s. 15(1) of the Charter. To found a complaint, it suffices to show that there has been a denial of equality on one of the listed grounds. But this is not the end of the story. The person alleged to have contravened the Code may avoid liability by establishing a justificatory defence, for example, that the distinction was justified in view of the requirements of the workplace (a bona fide occupational requirement).

139 Finally, the analysis I propose does not preclude the state from making distinctions between people on grounds like race, sex, age and citizenship. The state may do so, provided it can justify its use of the suspect criterion. Citizenship, recognized as an analogous ground in Andrews, provides a ready example. The state may be justified in confining certain privileges, like carrying a passport or serving in high government office, to citizens. If it can establish that justification, it may deny the privileges to non-citizens. McKinney v. University of Guelph, supra, provides another example. While the Court found that the university's policy of mandatory retirement at age 65 constituted discrimination on the basis of age, the policy was held to be reasonable and demonstrably justified in a free and democratic society. In short, the Charter does not forbid all distinctions on the basis of the enumerated or analogous grounds; it forbids stereotypical distinctions which the state cannot justify.

140 To recapitulate, the analysis under s. 15(1) involves two steps: examination of whether there has been a denial of "equal protection" or "equal benefit" of the law, and a finding that the denial constitutes discrimination. To establish discrimination, the claimant must bring the distinction within an enumerated or analogous ground. In most cases, this suffices to establish discrimination. However, exceptionally it may be concluded that the denial of equality on the enumerated or analogous ground does not violate the purpose of s. 15(1) -- to prevent the violation of human dignity and freedom through the imposition of limitations, disadvantages or burdens through the stereotypical application of presumed group characteristics, rather than on the basis of merit, capacity or circumstance. While irrelevance of the ground of distinction may indicate discrimination, the converse is not true. Proof of relevance does not negate the possibility of discrimination. We must look beyond relevance to ascertain whether the impact of the impugned legislation is to disadvantage the group or individual in a manner which perpetuates the injustice which s. 15(1) is aimed at preventing.

141 If a violation of s. 15(1) is established, the burden shifts to the party upholding the denial of equality to justify it under s. 1 of the Charter. Section 15(1) and s. 1 of the Charter must be read together. Neither, in itself, is complete. Together, they provide a comprehensive equality analysis that provides effective remedies against discrimination while preserving the power of the state to deny protections and benefits to individuals where differences between them justify it.

2.Section 15(1) -- Discrimination and its grounds

142 As indicated, the equality analysis under s. 15(1) involves a two-step process. First, the claimant must show that the law treats the claimant unequally in relation to another person. Second, the claimant must show that the denial results in discrimination and was made on the basis of one of the grounds enumerated in s. 15(1) or an analogous ground.

Step One -- Denial of equal benefit or equal protection of law and its basis

143 In this case, the insurer concedes that the legislation-based Ontario Standard Automobile Policy held by Ms. Valliere at the time of Mr. Miron's injury grants benefits to married couples which it does not accord to couples who are unmarried. The policy denies a person in an unmarried relationship benefits granted a similar person in a married relationship. Thus denial of equal benefit on the basis of marital status is established. The alleged discrimination is direct; there is no question of indirect discrimination because of the effect of the legislation, as opposed to its facial wording. The next inquiry is whether marital status is an analogous ground under s. 15(1).

Step Two -- Is marital status an analogous ground and if so, is the distinction on the basis of marital status discriminatory?

144 Section 15(1) of the Charter forbids discrimination, and "in particular, . . . discrimination based on race, national or ethnic origin, colour, religion, sex, age or mental or physical disability". The ground upon which the distinction in this case is based -- marital status -- is not included in the list of particularized grounds. We must therefore determine whether marital status is an analogous ground.

145 Our approach must be generous, reflecting the "continuing framework" of the constitution and the need for "`the unremitting protection' of equality rights": Andrews, per McIntyre J., at p. 175. Andrews instructs us that our approach must also reflect the human rights background against which the Charter was adopted. In evoking human rights law as the defining characteristic of discrimination under s. 15(1) of the Charter, this Court in Andrews engaged the principle of equality which underlies the constitutions of free and democratic countries throughout the world. This principle recognizes the dignity of each human being and each person's freedom to develop his body and spirit as he or she desires, subject to such limitations as may be justified by the interests of the community as a whole. It recognizes that society is based on individuals who are different from each other, and that a free and democratic society must accommodate and respect these differences.

146 The corollary of the recognition of the dignity of each individual is the recognition of the wrong that lies in withholding or limiting access to opportunities, benefits, and advantages available to other members of society, solely on the ground that the individual is a member of a particular group deemed to be less able or meritorious than others. This is the evil we call discrimination. It denies to the individual the right to realize his or her potential and to live in the freedom accorded to others, solely because of the group to which the individual belongs. In the course of the past century, free and democratic societies throughout the world have recognized that the elimination of such discrimination is essential, not only to achieving the kind of society to which we aspire, but to democracy itself. "The principle of equality, which is but the other side of the coin of discrimination and to which the law of every democratic country strives to realize in pursuit of justice and decency, means that one must apply, for the purpose of the [legislative] goal in question, equal treatment for all people, where there are no real differences amongst them that are relevant to that goal": Boronovsky v. Chief Rabbis of Israel, P.D. CH [25] (1), 7, 35.

147 The grounds of discrimination enumerated in s. 15(1) of the Charter identify group characteristics which often serve as irrelevant grounds of distinction between people. The history of the human rights movement is a history of reaction against persecution and denial of opportunity on the basis of irrelevant stereotypical group classifications like race, sex, and religion. It is not surprising therefore to see these as well as other common markers of irrelevant exclusion enumerated in s. 15(1). But the categories are not closed, as s. 15(1) recognizes. Analogous grounds of discrimination may be recognized. Logic suggests that in determining whether a particular group characteristic is an analogous ground, the fundamental consideration is whether the characteristic may serve as an irrelevant basis of exclusion and a denial of essential human dignity in the human rights tradition. In other words, may it serve as a basis for unequal treatment based on stereotypical attributes ascribed to the group, rather than on the true worth and ability or circumstances of the individual? An affirmative answer to this question indicates that the characteristic may be used in a manner which is violative of human dignity and freedom.

148 The theme of violation of human dignity and freedom by imposing limitations and disadvantages on the basis of a stereotypical attribution of group characteristics rather than on the basis of individual capacity, worth or circumstance is reflected in qualities which judges have found to be associated with analogous grounds. One indicator of an analogous ground may be that the targeted group has suffered historical disadvantage, independent of the challenged distinction: Andrews, supra, at p. 152 per Wilson J.; Turpin, supra, at pp. 1331-32. Another may be the fact that the group constitutes a "discrete and insular minority": Andrews, supra, at p. 152 per Wilson J. and at p. 183 per McIntyre J.; Turpin, supra, at p. 1333. Another indicator is a distinction made on the basis of a personal characteristic; as McIntyre J. stated in Andrews, "[d]istinctions based on personal characteristics attributed to an individual solely on the basis of association with a group will rarely escape the charge of discrimination, while those based on an individual's merits and capacities will rarely be so classed" (pp. 174-75). By extension, it has been suggested that distinctions based on personal and immutable characteristics must be discriminatory within s. 15(1): Andrews, supra, at p. 195 per La Forest J. Additional assistance may be obtained by comparing the ground at issue with the grounds enumerated, or from recognition by legislators and jurists that the ground is discriminatory: see Egan v. Canada, supra, per Cory J.

149 All of these may be valid indicators in the inclusionary sense that their presence may signal an analogous ground. But the converse proposition -- that any or all of them must be present to find an analogous ground -- is invalid. As Wilson J. recognized in Turpin (at p. 1333), they are but "analytical tools" which may be "of assistance". For example, analogous grounds cannot be confined to historically disadvantaged groups; if the Charter is to remain relevant to future generations, it must retain a capacity to recognize new grounds of discrimination. Nor is it essential that the analogous ground target a discrete and insular minority; this is belied by the inclusion of sex as a ground enumerated in s. 15(1). And while discriminatory group markers often involve immutable characteristics, they do not necessarily do so. Religion, an enumerated ground, is not immutable. Nor is citizenship, recognized in Andrews; nor province of residence, considered in Turpin. All these and more may be indicators of analogous grounds, but the unifying principle is larger: the avoidance of stereotypical reasoning and the creation of legal distinctions which violate the dignity and freedom of the individual, on the basis of some preconceived perception about the attributed characteristics of a group rather than the true capacity, worth or circumstances of the individual.

150 What then of the analogous ground proposed in this case -- marital status? The question is whether the characteristic of being unmarried -- of not having contracted a marriage in a manner recognized by the state -- constitutes a ground of discrimination within the ambit of s. 15(1). In my view, it does.

151 First, discrimination on the basis of marital status touches the essential dignity and worth of the individual in the same way as other recognized grounds of discrimination violative of fundamental human rights norms. Specifically, it touches the individual's freedom to live life with the mate of one's choice in the fashion of one's choice. This is a matter of defining importance to individuals. It is not a matter which should be excluded from Charter consideration on the ground that its recognition would trivialize the equality guarantee.

152 Second, marital status possesses characteristics often associated with recognized grounds of discrimination under s. 15(1) of the Charter. Persons involved in an unmarried relationship constitute an historically disadvantaged group. There is ample evidence that unmarried partners have often suffered social disadvantage and prejudice. Historically in our society, the unmarried partner has been regarded as less worthy than the married partner. The disadvantages inflicted on the unmarried partner have ranged from social ostracism through denial of status and benefits. In recent years, the disadvantage experienced by persons living in illegitimate relationships has greatly diminished. Those living together out of wedlock no longer are made to carry the scarlet letter. Nevertheless, the historical disadvantage associated with this group cannot be denied.

153 A third characteristic sometimes associated with analogous grounds --distinctions founded on personal, immutable characteristics -- is present, albeit in attenuated form. In theory, the individual is free to choose whether to marry or not to marry. In practice, however, the reality may be otherwise. The sanction of the union by the state through civil marriage cannot always be obtained. The law; the reluctance of one's partner to marry; financial, religious or social constraints -- these factors and others commonly function to prevent partners who otherwise operate as a family unit from formally marrying. In short, marital status often lies beyond the individual's effective control. In this respect, marital status is not unlike citizenship, recognized as an analogous ground in Andrews: the individual exercises limited but not exclusive control over the designation.

154 Comparing discrimination on the basis of marital status with the grounds enumerated in s. 15(1), discrimination on the ground of marital status may be seen as akin to discrimination on the ground of religion, to the extent that it finds its roots and expression in moral disapproval of all sexual unions except those sanctioned by the church and state.

155 Of late, legislators and jurists throughout our country have recognized that distinguishing between cohabiting couples on the basis of whether they are legally married or not fails to accord with current social values or realities. As the amicus curiae has pointed out, 63 Ontario statutes currently make no distinction between married partners and unmarried partners who have cohabited in a conjugal relationship. For example, the right to spousal maintenance is not conditioned on marriage: see Part III, Family Law Act, R.S.O. 1990, c. F.3, which establishes a right to spousal support for those who have cohabited continuously for a period of not less than three years or who have cohabited in a relationship of some permanence and who have a child. Other provinces have adopted similar benefit thresholds. In the judicial domain, judges have recognized the right of unmarried spouses to share in family property through the doctrine of unjust enrichment: Pettkus v. Becker, 1980 22 (SCC), [1980] 2 S.C.R. 834; Peter v. Beblow, 1993 126 (SCC), [1993] 1 S.C.R. 980. All this suggests recognition of the fact that it is often wrong to deny equal benefit of the law because a person is not married.

156 These considerations, taken together, suggest that denial of equality on the basis of marital status constitutes discrimination within the ambit of s. 15(1) of the Charter. If the evil to which s. 15(1) is addressed is the violation of human dignity and freedom by imposing limitations or disadvantages on the basis of the stereotypical application of presumed group characteristics, rather than on the basis of individual capacity, worth or circumstance, then marital status should be considered an analogous ground. The essential elements necessary to engage the overarching purpose of s. 15(1) -- violation of dignity and freedom, an historical group disadvantage, and the danger of stereotypical group-based decision-making -- are present and discrimination is made out.

157 These observations are sufficient to dispose of the insurer's arguments based on alleged absence of historical disadvantage and the "mutable" nature of the unmarried state. It remains to consider, however, the theme underlying the whole of the insurer's submissions -- that marriage is a good and honourable state and hence cannot serve as a ground for discrimination. To most in our society, marriage is a good thing; to many a sacred thing. There is nobility in the public commitment of two people to each other to the exclusion of all others. How can it be wrong to use this commitment as the condition of receiving legal protection and benefit?

158 These sentiments, valid as they are, do not advance the insurer's case. The argument, simply put, is that marriage is good; the grounds of discrimination evil; therefore marriage cannot be a ground of discrimination. The fallacy in the argument is the assumption that the grounds of discrimination are evil. Discrimination is evil. But the grounds upon which it rests are not. Consider the enumerated grounds -- race, national or ethnic origin, colour, religion, sex, age and mental or physical disability. None of these are evil in themselves. Indeed, people rightfully take pride in their race and ethnic origin; they find identity in their colour and their sex. Even mental and physical disabilities should be regarded not as deficiencies, but differences -- differences which, while they will make some aspects of life more difficult, do not affect others, and may, moreover, contribute to society's richness and texture. What is evil is not the ground of discrimination, but its inappropriate use to deny equal protection and benefit to people who are members of the marked groups -- not on the basis of their true abilities or circumstance, but on the basis of the group to which they belong. The argument that marital status cannot be an analogous ground because it is good cannot succeed. The issue is not whether marriage is good, but rather whether it may be used to deny equal treatment to people on grounds which have nothing to do with their true worth or entitlement due to circumstance. L'Heureux-Dubé J. stated in Canada (Attorney General) v. Mossop, 1993 164 (SCC), [1993] 1 S.C.R. 554, at p. 634: "It is not anti-family to support protection for non-traditional families." One might equally say it is not anti-marriage to accord equal benefit of the law to non-traditional couples.

159 Consider this Court's decision in Andrews. The quality of citizenship there at issue is universally regarded as a good and valuable thing. Like marriage, it is the formal marker of a relationship -- the relationship between the individual and his or her country. At trial, the argument prevailed that citizenship, as the sacred marker of a relationship of commitment and permanence, could not found a claim for discrimination under s. 15(1) of the Charter. But this Court took a different view. It held that citizenship could constitute an analogous ground of discrimination. In itself citizenship was good. But it could not be used inappropriately to permit the exclusion of people from activities for which, in terms of personal merit, they were qualified. It was not citizenship which constituted discrimination, but its inappropriate use to bar Mr. Andrews from an activity to which citizenship was only tangentially relevant. Similarly marriage, however sacred, may be inappropriately used to bar individuals not belonging to the married group from the protection or benefit of laws to which the status of legal marriage has little real relevance. This potential for denial of benefit based on stereotypical characteristics attributed (or not attributed) to a group rather than on the basis of the characteristics of the individual makes marriage, like citizenship, an analogous ground. This does not mean that citizenship or marriage cannot be used as the basis of appropriate legislative distinctions. Marriage and citizenship may be used as the basis to exclude people from protections and benefits conferred by law, provided the state can demonstrate under s. 1 that they are truly relevant to the goal and values underlying the legislative provision in question.

160 I conclude that marital status may serve as an analogous ground of discrimination under s. 15(1) of the Charter.

Conclusion on s. 15

161 The legislation reflected in the insurance policy at issue denies equal benefits to partners in an unmarried relationship solely on the ground of their marital status. This ground is an analogous ground under s. 15(1). This is not one of the exceptional cases where a distinction drawn on the basis of an enumerated or analogous ground does not fall within the anti-discrimination guarantees of the Charter. It follows that discrimination under s. 15(1) is established.

3.Justification under s. 1 of the Charter

162 A finding of denial of equal benefit or protection of the law on a discriminatory ground under s. 15(1) does not mean that the law is unconstitutional. The Court must go on to examine whether, notwithstanding its discriminatory character, the law or government action in question is "demonstrably justified in a free and democratic society". The complainant bears the burden of showing discrimination under s. 15(1). This established, the burden shifts to the state or the party seeking to uphold the law to justify the discrimination.

163 Determining whether it has been demonstrated that the impugned distinction is "demonstrably justified in a free and democratic society" involves two inquiries. First, the goal of the legislation is ascertained and examined to see if it is of pressing and substantial importance. Then the court must carry out a proportionality analysis to balance the interests of society with those of individuals and groups. The proportionality analysis comprises three branches. First, the connection between the goal and the discriminatory distinction is examined to ascertain if it is rational. Second, the law must impair the right no more than is reasonably necessary to accomplish the objective. Finally, if these two conditions are met, the court must weigh whether the effect of the discrimination is proportionate to the benefit thereby achieved. See R. v. Oakes, 1986 46 (SCC), [1986] 1 S.C.R. 103.

164 Examination of the goal of the legislation is vital in discrimination cases as elsewhere. Sometimes the legislative goal is apparent on the face of the legislation. Other times it may not be. Legislation aimed at effecting a less than worthy goal may be cloaked in the rhetoric of justice and reason. The task of the court in every case is to identify the functional values underlying the law.

165 The goal or functional value of the legislation here at issue is to sustain families when one of their members is injured in an automobile accident. When an adult partner in a family unit is injured, economic dislocation may not be far behind. If the injured partner is a wage-earner, the family income may be reduced or eliminated. If the injured party works in the home, it may be necessary to hire replacement services. In either case, the result is economic dislocation. This, in turn, can work great hardship on the family and its members. The goal of the legislation is to reduce this economic dislocation and hardship. This is a laudable goal. And given the frequency of injuries from motor vehicle accidents, it assumes an importance which can without exaggeration be described as pressing and substantial.

166 The next inquiry is whether a rational connection has been shown to exist between the legislative goal and the discrimination. As we have seen, analogous grounds may be used in ways that are relevant, or rationally connected, to a valid legislative goal. For example, exclusion from a regulated activity on the basis of age or citizenship might be justified if the state can show that age or citizenship is relevant to the ability to perform the activity safely and properly. If the proponent of the law can demonstrate that the ground of denial of a protection or benefit is relevant to the goal of the legislation, the discrimination loses its sting. "[I]f the . . . differences amongst different people are relevant to the goal in question, then this will be a permissible distinction . . .": Boronovsky v. Chief Rabbis of Israel, supra.

167 At this point we meet the problem of how relevant a criterion must be in discrimination cases. This inquiry echoes the "minimal impairment" analysis undertaken in Oakes, supra, and discussed in R. v. Edwards Books and Art Ltd., 1986 12 (SCC), [1986] 2 S.C.R. 713, and Reference re ss. 193 and 195.1(1)(c) of the Criminal Code (Man.), 1990 105 (SCC), [1990] 1 S.C.R. 1123. The marker chosen by the legislator may be only tangentially relevant to the legislative goal, as citizenship was found to be to the determination of ability to practise law in Andrews. In such cases, we say it is a poor marker; one which excessively impairs the equality rights. Although it may eliminate some people who may legitimately be excluded, it also excludes many who, given the goal of the legislation, should not be excluded. In contrast, a good marker excludes most people who should be excluded given the goal of the legislation, and only a few who should not. The standard which the legislator must meet is not perfection, but reasonableness. Of necessity laws use group criteria; and of necessity there are sometimes individual members of the group chosen who do not conform to the usual profile of the group and with respect to whom, viewed individually, even a relevant legislative marker may be irrelevant. For example, a law may deny drivers' permits to persons under the age of sixteen. There may be some people under the age of sixteen who are good drivers. But if the state can show that most people under sixteen would not be competent and responsible drivers, the age of sixteen may be defended as a relevant marker of those who should be permitted to drive. Provided the group marker chosen by the state is relevant to the legislative goal, the existence of minor anomalies due to the variation of individuals within the group will not render the marker violative. On the other hand, if the number of anomalies is so high that it significantly undermines the relevance of the group marker, or if more reasonable markers are available, the law may be invalid because it impairs the right more than reasonably necessary to achieve the legislative goal.

168 Returning to the case at bar, the question is whether marital status is a reasonably relevant marker of individuals who should receive benefits in the event of injury of a family member in an automobile accident, given the goal of the legislation. The insurer defends the marker of marital status as an indicator of stability which goes to the economic interdependence of the family unit. To maintain this claim, the state (or the insurer that here stands in its stead) must show that stable, and thus economically interdependent, family units typically involve married partners, and conversely, that unmarried partners in stable relationships are but a minor anomaly. Further, given the injustice of any anomalies, one would expect a demonstration that better criteria, producing fewer anomalous cases, are not readily available. In short, it must be demonstrated that the chosen group marker is reasonably relevant to the legislative goal in all the circumstances of the case, having regard to available alternative criteria and the need to minimize prejudice to anomalous cases within the group.

169 This the insurer and the state have not done. The record suggests that the legislators recognized that marital status was at best a problematic indicator of who should receive accident benefits upon injury in a motor vehicle accident. The debate centred on marital equivalence. To quote the amicus curiae, "[the legislators'] search was directed towards defining a `marriage-like' conjugal relationship, usually in terms of mutual commitment and permanence -- a "near" marriage -- instead of trying to define the underlying functional values, e.g. financial interdependence, relevant to the legislative subject matter of the Insurance Act". Having misconstrued the issue as one of marriage equivalence, the Legislature found itself unable to agree. But this provides no justification for failing, from 1980 to 1987, to deal directly with the problem of which family units were so financially interdependent and stable as to warrant provision of the benefits in question.

170 If the issue had been viewed as a matter of defining who should receive benefits on a basis that is relevant to the goal or functional values underlying the legislation, rather than marriage equivalence, alternatives substantially less invasive of Charter rights might have been found. For example, the Legislature was able to agree in 1980 on a formula to extend death benefits to a certain class of unmarried persons. And in 1981, in the Ontario Human Rights Code, 1981, S.O. 1981, c. 53, s. 9(j), the Legislature agreed on a definition of "spouse" as the person to whom a person of the opposite sex is married or with whom the person is living in a conjugal relationship outside marriage. A modified version of the Human Rights Code definition requires marriage or cohabitation for at least one year or having a child together, or entering into a cohabitation agreement under s. 53 of the Family Law Act. This modified definition is used in 21 Ontario statutes.

171 It thus emerges that in fixing on marital status as the criterion of eligibility for family accident benefits, the Legislature chose a criterion that was at best only collaterally related to its legislative goal; a criterion, moreover, that had the effect of depriving a substantial number of deserving candidates of receipt of benefits. Better tests were available. In short, the Legislature did not choose a reasonably relevant marker.

172 It is suggested that the Legislature's choice of an inappropriate marker for family accident benefits can be defended on the ground that the legislation was passed in a period of rapidly changing family norms. Legislatures, it is argued, should not be held to standards of social perfection. As La Forest J. wrote in McKinney v. University of Guelph, supra, at p. 317:

. . . a Legislature should not be obliged to deal with all aspects of a problem at once. It must surely be permitted to take incremental measures. It must be given reasonable leeway to deal with problems one step at a time . . . .

173 I agree with these sentiments; however the need for legislative leeway is of little assistance in the case at bar. Marital status was not a reasonable criterion even in 1980, and the alternatives adopted in the years that followed belie the suggestion that the failure to adopt a more relevant criterion in the years between 1980 and 1987 can be attributed to the time required for legislative response.

174 It remains to consider whether the Legislature's choice of the inappropriate criterion of marital status to distinguish between those who receive benefits and those who do not can be justified on the ground that the resultant discrimination is proportionate to the legislative goal. Having determined that marital status is not a reasonable indicator of those who should obtain accident benefits -- that, to use the language of Oakes, the rational connection between the discrimination and the goal of the legislation is deficient and the law impairs the right more than reasonably necessary to achieve the legislative goal -- it is unnecessary to move to the final step to consider whether the effect of the infringement is "proportionate" to the benefit to be derived from using the discriminatory marker.

175 I conclude that the state has failed to demonstrate that the exclusion of unmarried members of family units from motor vehicle accident benefits is demonstrably justified in a free and democratic society. It follows that the Charter violation is established.

C. Remedy

176 Having found that the impugned statutory provisions of the Insurance Act violate the Charter, the Court is left with the choice between "reading in" appropriate amendments into the provisions, or leaving them as they are with the result that they fall as invalid under s. 52 of the Constitution Act, 1982. In the latter case, the Court may consider a declaration of suspension of the invalidity for a period of time sufficient to allow the Legislature to remedy the violation.

177 The remedy of "reading in" is available if the question of how far the benefit should be extended can be answered with "sufficient precision" to justify the Court in doing so, so as to bring the case within the guidelines laid out in Rocket v. Royal College of Dental Surgeons of Ontario, 1990 121 (SCC), [1990] 2 S.C.R. 232, and Schachter v. Canada, 1992 74 (SCC), [1992] 2 S.C.R. 679. An affirmative answer in this case is suggested by the fact that in 1990 the Ontario Legislature amended the eligibility criteria in a way which would include the appellants, thus giving an indication of what it would do if the matter were remitted to it anew. While this does not meet concerns that the social and legislative picture may have changed further in the years since 1990, or resolve the problem for the other Ontario statutes containing similar provisions, it does offer reasonably conclusive evidence of how the Legislature would have remedied the 1980 legislation had it been required to do so when the appellants' claim arose.

178 The alternative remedy entails a declaration of invalidity of the 1980 legislation. It also entails consideration of a temporary suspension of that declaration for a period of time during which the Legislature might be expected to amend the 1980 Insurance Act, in order to avoid the revocation of benefits payable under that Act. If this were done, it would still leave the appellants and others in their situation without a remedy.

179 It is suggested that the Court could fashion a remedy for the appellants under s. 24(1) of the Charter, which provides that "[a]nyone whose rights or freedoms ... have been infringed or denied may apply to a court . . . to obtain such remedy as the court considers appropriate and just in the circumstances". In Rodriguez v. British Columbia (Attorney General), 1993 75 (SCC), [1993] 3 S.C.R. 519, at p. 577, this Court (per Lamer C.J., dissenting on other grounds) suggested that an order of suspension of invalidity might be coupled with individual relief in the form of a "constitutional exemption" to the applicant who has suffered the Charter violation and has initiated court proceedings to obtain Charter relief. Three other dissenting judges agreed. Assuming the Court were inclined to grant the appellants an exemption from the 1980 legislation and insurance policy provisions, the question remains of how it could do so without creating further inequities between the appellants and others in their situation who have been denied benefits. To avoid this, any constitutional exemption would have to be extended to all similar families. This in turn would require formulation of general criteria of eligibility, thus involving the court in the very activity which would have led it to eschew "reading up" the 1980 statute in conformity with the terms legislated in 1990. Yet to deny such persons a remedy would be to perpetuate the effects of a discrimination which the Court has found to violate the Charter when the obvious remedy -- the payment of the benefits that should have been paid -- remains available.

180 Having considered the available remedies, I am persuaded that this is one of those exceptional cases where retroactively "reading up" a statute may be justified. The 1990 amendments provide the best possible evidence of what the Legislature would have done had it been forced to face the problem the appellants raise. The only claims are monetary and readily calculable and satisfied. Most importantly, the result will be to cure an injustice which might otherwise go unremedied.

181 In this case, the benefit payments are actually made available pursuant to the insurance contract between Ms. Valliere and Economical Mutual. Because the provisions in the insurance policy were mandated by the Insurance Act, the effect of reading up the Act is to import the 1990 definition of spouse into the Standard Automobile Insurance Policy.

Disposition

182 The appeal is allowed with costs on a party and party basis. The insurer's application to strike out the appellants' action is dismissed. The action is remitted for trial to determine whether Mr. Miron and Ms. Valliere meet the requirements of the 1990 legislation.

Appeal allowed with costs, Lamer C.J. and La Forest, Gonthier and Major JJ. dissenting.

Solicitors for the appellants: Nelligan ♦ Power, Ottawa.

Solicitors for the respondents: Cooligan Ryan, Ottawa.

Solicitor for the intervener the Attorney General of Canada: The Deputy Attorney General of Canada, Ottawa.

Solicitor for the intervener the Attorney General for Ontario: The Ministry of the Attorney General, Toronto.

Solicitors for the intervener the Attorney General of Quebec: Kathleen McNicoll and Madeleine Aubé, Ste‑Foy.

Solicitor for the intervener the Attorney General of Manitoba: The Department of Justice, Winnipeg.

Solicitors appointed by the Court as amicus curiae: McCarthy Tétrault, Toronto.