Article body


Canadian constitutional law regarding freedom of association for workers is a mess. As the title of this article suggests, it is a mess that we have gotten ourselves into and one that we can get ourselves out of. As usual, the getting out part depends on understanding the mess we are in and how we got there.

This paper focuses on the recent decision of the Supreme Court of Canada in Health Services and Support—Facilities Subsector Bargaining Association v. British Columbia.[1] This decision explicitly overruled a quartet of the Court’s recent holdings on freedom of association and has now replaced those judgments as the key to our constitutional law on this topic. In BC Health, the Court struck down, as violating the Canadian Charter of Rights and Freedom’s section 2(d) guarantee of “freedom of association”,[2] certain provisions in a statute passed by the government of British Columbia to curtail health-care costs. The government adopted a number of strategies in these provisions, including rewriting existing collective agreements and forbidding renegotiation of the resulting changes, and did so without consulting or negotiating with the unions involved.

BC Health has been widely hailed as a step forward for workers’ freedom of association. This essay espouses a different and less celebratory view of BC Health. The result in that case, to the extent that it holds that our Charter guarantee of freedom of association entails constitutional protection of some notion of collective bargaining, is surely correct. However, this essay criticizes both the particular conception of collective bargaining the Court promoted and the reasons it deployed in support of its holding. In particular, this essay argues that the Court misread Canadian labour law history, misunderstood Canada’s international labour law obligations, and misstated “Charter values”; that it very unhelpfully perpetuated discussions of labour rights as inherently “collective”; and that it ignored the distinction between “freedoms” and “rights” that is critical to understanding Charter freedoms in general and freedom of association in particular.

Following this critique of BC Health, this essay considers a very basic issue that this sort of internal critique cannot address: how to deal with abstract freedoms that are incorporated in detailed statutes passed by legislatures. The Charter declares that “[e]veryone” has a number of “fundamental freedoms”, including “freedom of association”.[3] Full stop. But what is the relationship between this sort of abstract statement of a basic freedom or human “right” (as found in the Charter and in international treaties) on the one hand, and the detailed statutes and regulations that articulate and instantiate that freedom at an enforceable level (such as Canadian labour relations statutes) on the other? And how should we respond to arbitrary exclusions from such detailed instantiations of the fundamental freedom? These are fundamental questions that the Court in BC Health neither asks nor answers directly in these terms.

This omission is unfortunate, since the Court’s (albeit mostly implicit) approach to these issues is now the key to constitutional labour law in this country. The Court was not able to ask these larger questions because of its prior and unsatisfactory commitments on a series of other constitutional points concerning equality, state action, and positive obligations to legislate. Briefly, these commitments boxed the Court out of the best available approach to both the legal issues in BC Health and the more general questions at stake. Furthermore, these prior commitments boxed the Court into a very unattractive, alternative approach that involved, among other things, the Court constructing its own parallel labour code for Canadian workers. The idea of a judicially created labour code has never been a good one and it is not a good one now. This code is also objectionable because it is partial, in the sense that it is available to workers in an irrational set of subsectors of the economy.

Many more unfortunate stops on this jurisprudential line are already in view. BC Health has been taken by constitutional litigators as an invitation to “start your engines”.[4] This is the freedom of association mess we are in. This essay suggests that we stop heading along this track and shift to another, existing line of thought that leads us to much more appropriate constitutional destinations via a better and more direct route.

At the time that BC Health was argued, and as a result of prior cases such as Dunmore v. Ontario (A.G.)[5] and Delisle v. Canada (A.G.),[6] the set of rules in place regarding whose freedom of association was protected and whose was not was at odds with basic Canadian and international values. I like to think that the Court agreed and that it was attempting to address this problem in BC Health. But the path that the Court set for itself is not the way forward: BC Health is an expansion of the problem, not the beginning of a solution. We will remain in the current mess as long as the Court’s prior constitutional commitments, which forced its hand in BC Health, remain unaltered. The only way out of the freedom of association mess is to revisit these prior commitments. I believe that our legal reality post-BC Health is so unsatisfactory as to be unsustainable, and that this truth offers a realistic hope for a thoughtful reconsideration of those prior commitments which have brought us to our current state of affairs. This essay is written in that spirit.

Part I offers a four-part, internal critique of the decision in BC Health. Part II provides an overview of the Court’s other constitutional positions and commitments, which at once underwrite and undermine the approach to freedom of association that the Court advanced in BC Health. From this broader perspective, external to the logic of BC Health, we can identify the real mess we are in, as well as the way out of it.

The key to understanding the mess we are in and the way out of it lies in three ideas. First, there are many ways in which domestic labour law systems can and do respect abstract international law and domestic constitutional guarantees of freedom of association while making these guarantees concrete and legally enforceable. A moment of reflection upon the very different labour relations systems of the Organisation for Economic Co-operation and Development’s thirty member nations, let alone those of the other 160 or so members of the International Labour Organization (ILO), should make this variety very evident. Second, Canadian labour relations statutes are best conceived as our domestic version of a detailed and legally enforceable instantiation of the fundamental freedom of association. Third, arbitrary exclusion from these statutes is, as a result, simply arbitrary exclusion from the guarantee of freedom of association. We cannot, as decent Canadians, make this freedom meaningful and legally enforceable for most workers while leaving others out, unless we have a very good reason for doing so. Although our particular domestic system is not the only way of applying the abstract expression of freedom, if this is how we have chosen to give effect to this freedom, then in the absence of compelling reasons for exclusion, everyone is entitled to the same protection. This is the demand of the constitutional idea of equality and it is this constitutional idea that must be deployed to get us out of the mess we are in.

I. The Four Propositions

The holding in BC Health is set out in paragraph 19 of the judgment:

At issue in the present appeal is whether the guarantee of freedom of association in s. 2(d) of the Charter protects collective bargaining rights. We conclude that s. 2(d) of the Charter protects the capacity of members of labour unions to engage, in association, in collective bargaining on fundamental workplace issues. This protection does not cover all aspects of “collective bargaining”, as that term is understood in the statutory labour relations regimes that are in place across the country. Nor does it ensure a particular outcome in a labour dispute, or guarantee access to any particular statutory regime. What is protected is simply the right of employees to associate in a process of collective action to achieve workplace goals. If the government substantially interferes with that right, it violates s. 2(d) of the Charter. We note that the present case does not concern the right to strike, which was considered in earlier litigation on the scope of the guarantee of freedom of association.[7]

The Supreme Court of Canada articulated four propositions upon which this holding was based:

Our conclusion that s. 2(d) of the Charter protects a process of collective bargaining rests on four propositions. First, a review of the s. 2(d) jurisprudence of this Court reveals that the reasons evoked in the past for holding that the guarantee of freedom of association does not extend to collective bargaining can no longer stand. Second, an interpretation of s. 2(d) that precludes collective bargaining from its ambit is inconsistent with Canada’s historic recognition of the importance of collective bargaining to freedom of association. Third, collective bargaining is an integral component of freedom of association in international law, which may inform the interpretation of Charter guarantees. Finally, interpreting s. 2(d) as including a right to collective bargaining is consistent with, and indeed, promotes, other Charter rights, freedoms and values.[8]

This paper takes the following view: the first proposition is correct on its face, but the reasoning behind it is not, while the validity of the remaining three propositions depends upon the substance and content of the “right”[9] to collective bargaining that the Court finally articulated.[10] Yet this articulation was only offered after all of these propositions were deployed and defended in support of the Court’s holding. When the Court later revealed its understanding of the content of the “right” (especially that the right includes an employer’s duty to bargain), it became clear that the final three propositions were false. The structure of the reasoning in BC Health is thus unsound and subject to the criticism that it involves sleight of hand or, more abstractly, elides critical steps in reasoning. One cannot move from the claim that international law, Canadian labour law history, or Charter values support some notion of a right to collective bargaining, to the idea that international law, Canadian labour history, or Charter values support this particular conception of the right to collective bargaining. Yet such was the very move made in, and the essential difficulty with, BC Health.

I turn now to the four propositions upon which the holding in BC Health is said to rest.

A. The First Proposition: The Historic Refusal to Apply Section 2(d) to Collective Bargaining Was Incorrect

a review of the s. 2(d) jurisprudence of this Court reveals that the reasons evoked in the past for holding that the guarantee of freedom of association does not extend to collective bargaining can no longer stand.[11]

This proposition has the virtue of being true (subject to the small point that those “reasons” never stood in the first place). Unfortunately, the good news stops there.

Earlier decisions of the Court, namely the trilogy of cases dealing with the application of section 2(d)[12] and Professional Institute of the Public Service of Canada v. Northwest Territories (Commissioner),[13] held that section 2(d) did not extend to a right to collective bargaining or a right to strike. In BC Health, the Court overruled those decisions concerning collective bargaining and left the issue of a right to strike for another day. The Court was correct to overrule those decisions—but not for the reasons given. I do not propose to review the many criticisms of those earlier decisions which had been put on offer before BC Health. Rather, I articulate very briefly what a good set of reasons for overruling those decisions would look like. I start from the premise that the best account of the idea of freedom of association is a very simple one: it is the freedom to do in combination with others what one is free to do alone. Laws that unjustifiably hobble this freedom violate the Charter. Further, freedom of association so conceived is remarkably important in open societies.

This position is precisely where at least some members of the Court started in the trilogy and PIPS. The problem with the trilogy and PIPS lies not in the Court’s ability to grasp the essence of the constitutional guarantee of freedom of association. Rather, the problem lies in the Court’s inability to apply that idea. The Court disregarded the plain fact that an individual is legally free, subject to statutory minimum terms, to bargain for the terms of an employment contract and to refuse to go to work until these terms have been agreed upon. These notions are very basic; when we add them to the idea that freedom of association means the freedom to do in combination what one is free to do alone, we arrive at the conclusion that there is a right to collective bargaining and a right collectively to refuse to work until the terms of the bargain are settled (that is, to strike). This approach is simple, direct, coherent, and understandable—and was available to the Court in BC Health. However, this path was not taken.[14]

Rather than going directly to the heart of the matter, the Court continued the project begun in Delisle and Dunmore, and spent its time tilting at some odd windmills, which it sees as critical to the errors in the trilogy and PIPS. This misplaced zeal leads, in my view, to the reaffirmation of some very mistaken ideas and modes of thinking.

In support of the first proposition, the Court identified and rejected five rationales underlying the earlier decisions:

  1. thinking that collective bargaining and the right to strike are “modern” rights, creations of statute, and not “fundamental”;

  2. applying a notion of judicial restraint regarding laws dealing with labour relations;

  3. thinking that freedom of association protects doing in combination what one is free to do as an individual (that is, building section 2(d) on the foundation of individual rights);

  4. thinking in terms of whether section 2(d) protects the “objects” of the association; and

  5. taking a decontextualized approach to section 2(d).[15]

I discuss the first rationale below, in connection with the Court’s second proposition. The second rationale is not troubling on its face: labour law cannot be, as the court says, a “‘no go’ zone”[16] for the Charter. But in applying the Charter to labour law, the legality and legitimacy of the Court’s holdings must flow from general constitutional principles that apply beyond this field. This progression establishes the proper limits of judicial review when the constitution is used as a vehicle to enter and to occupy territory long and correctly assigned to others at the level of (contextual) system design and statutory detail. This idea is directly connected to the identification and “correction” of the fifth rationale—a decontextualized approach. Here, I do not think that the Court demonstrated sufficient awareness of the very heavy chains that it was rattling. When the Court takes on the task of weighing the harm of banning book clubs as compared to banning collective bargaining and relegates the former to a lower level of concern,[17] I find its approach chilling and believe that it is time to speak up. The Court is missing the point of the “fundamentalness” of fundamental rights. I return to this issue below, in my discussion of Charter values.

The central problem with BC Health lies in the Court’s rejection of the third rationale from the trilogy and PIPS: the view that an account of freedom of association can be constructed from the starting point of what each of us is free to do alone. The rejection of this simple line of thought gained considerable momentum in Dunmore and is asked to carry a lot of freight in BC Health. This rejection finds its origins in Chief Justice Dickson’s dissent in the trilogy’s Alberta Reference.[18] Such a view will haunt us for some time and fuel more of the sort of judicial effort evidenced in Dunmore and BC Health. A discussion of the fourth rationale (that is, whether talk of “objects” of association is helpful) is best undertaken as part of an analysis of this particular confusion.

How did the idea of invoking “inherently collective” activities (which do exist: think of singing in a choir or formulating a collective position) as a way of critiquing our best account of freedom of association (that is, the freedom to do together with others what you are free to do as an individual) ever make it off the ground? If there is an answer to this question, it lies neither in law nor in logic. Yet it is precisely this idea that underlies the basic understanding of freedom of association articulated in Dunmore and BC Health.

The Court’s reasoning seems to be as follows. There are some things that an individual cannot do (for example, be a one-person choir). This is true. It is in the nature of certain activities that they cannot be done alone. This is how, to use our example, the concept of a “choir” works. As the Court stated repeatedly, there are some activities that are “inconceivable on the individual level”[19] and “by their very nature, ... incapable of being performed by an individual.”[20] The Court then observes that some of the important functions of unions fall into this category of individual impossibility. So, to quote BC Health quoting Dunmore, “the law must recognize that certain union activities—making collective representations to an employer, adopting a majority political platform, federating with other unions—may be central to freedom of association even though they are inconceivable on the individual level.”[21] This is again true. It is simply nonsense to say that a single employee can make collective representations to an employer.

However, true as these observations may be, they are unhelpful to anyone with a reasonable command of the English language. Of course one person cannot be a collective (or a choir for that matter). But this obvious truth does not have any bearing on the idea and nature of rights. The fact that collective bargaining can only be done as a group has no connection to the question “is the ‘right’ to collective bargaining a right that I (as an individual citizen) have, or can have?” The answer to this question is, simply, “yes”. There are certain (human) rights that are rights of individuals “though they may be rights of individuals which can only be exercised collectively.”[22] Or, as Clyde Summers perfectly put it:

Although commonly asserted by the organization, freedom of association is not simply a collective right vested in the organization for its benefit. Freedom of association is an individual right vested in the individual to enable him to enlarge his personal freedom. Its function is not merely to grant power to groups, but to enrich the individual’s participation in the democratic process by his acting through those groups.[23]

If I have a right to bargain collectively, the fact that collective bargaining is “an inherently collective activity” does not and cannot enter into the matter. Once I have exercised my right to freedom of association, I have indeed ended up in the very circumstance that the right is designed to secure: participation in a group or collective, which by definition I cannot do alone. This individual conception of the right can, and does, perform all of the work required. Indeed, without a mooring in concern for real people, it is hard to see why groups are important. As Justice Iacobucci put it in Delisle:

The human animal is inherently sociable. People bind together in a myriad of ways, whether it be in a family, a nation, a religious organization, a hockey team, a service club, a political party, a ratepayers association, a tenants organization, a partnership, a corporation, or a trade union. By combining together, people seek to improve every aspect of their lives. Through membership in a religious group, for example, they seek to fulfill their spiritual aspirations; through a community organization they seek to provide better facilities for their neighbourhood; through membership in a union they seek to improve their working conditions. The ability to choose their organizations is of critical importance to all people. It is the organizations which an individual chooses to join that to some extent define that individual.[24]

It is true that the choir, “the press”, the “language community”, and “the union” (to use other examples that Justice Bastarache deploys in Dunmore[25]) may be distinguished from the individuals in them. But this distinction cuts no conceptual or constitutional ice at all.

On the other hand, grounding the right in individuals leads to simple and straightforward conclusions that resonate deeply with basic constitutional values. For example, if we base the right in what individuals are free to do, we arrive at the following line of thought: I have a legal freedom to bargain for a contract of employment.[26] Therefore, because I also have the right to freedom of association, I have a right collectively to bargain such terms of employment. Now, it is perfectly true (albeit unhelpful to point out) that such collective bargaining is “inconceivable on the individual level.” However, we understand this premise perfectly well already: there is a difference between individual and collective bargaining. But the idea that the right to collective bargaining cannot be based upon the right to individual bargaining is false.[27] On the contrary, it is the very point of the right to collective bargaining to guarantee that we can make the transition to what would be inconceivable on the individual level. Discussions of “collective rights” obscure this important point.

Such discussions obscure other fundamental issues as well. I think that it is safe to predict, to take a very important example, that this discussion about some activities being “inherently collective” will come back to haunt the question of the right to strike, which the Court was able to put aside in BC Health. The right to strike, on the view adopted here, is just as simple as the right to collective bargaining. I have the freedom not to work until I have agreed to a contract; because I also enjoy a freedom of association, I can do the same in combination with others. That is what a strike is. The danger, however, is that the Court has detached collective bargaining from its obvious individual-bargaining analogue and will do the same for the right to strike.

This detachment can only cause problems. On my view, these freedoms are basic, and we do not interfere with them lightly. It is vital that we see the bargaining analogue in clear terms in order to understand Canadian collective labour law. Collective bargaining is, simply, bargaining. There is nothing mysterious about it. So, too, striking has a perfectly obvious analogue in individual bargaining: the freedom to hold out for better terms before agreeing to go to work. This is an idea that cannot be avoided. Of course we know that for centuries, judges have been hostile to the idea of freedom of contract when it comes to workers relying upon it. My prediction is that slipping (or, rather, intentionally throwing away) the moorings of the individual analogue will come at a cost in terms of the perceived fundamental nature of the right to strike and will make this right more amenable to inappropriate judicial—and administrative, and statutory—(mis)management.

Many more profound difficulties will follow from a flawed understanding of the basis for the protection of collective activity. The most basic risk is that we will lose sight of the fact that the primal aspiration of our labour law is to protect and to advance the interests of workers.[28] Workers create and choose unions. That is why unions are important. It is this freedom that Canadian workers fought for and that domestic and international labour laws set out to protect and to promote. These laws are not directly concerned with promoting or protecting unions per se, unconnected to the workers who constitute them. If workers’ interests are not the basis of the freedom, then it is not worth fighting for or spending one’s life worrying about. In BC Health, the Court neglected this fundamental idea. Once on this slippery slope, it is a quick ride to a point at which the true nature of our rights and freedoms is obscured and becomes merely another factor in an undisciplined “balancing” of policies or societal interests.

So the earlier cases deserved to be overruled—but not for the reasons given.

B. The Second Proposition: The Exclusion of Collective Bargaining from the Ambit of Section 2(d) Is Inconsistent with Canada’s Labour History

an interpretation of s. 2(d) that precludes collective bargaining from its ambit is inconsistent with Canada’s historic recognition of the importance of collective bargaining to freedom of association.[29]

In the “Labour Law Casebook”[30] that for many years has been used in law schools across the country, there has long been included an excerpt from Stan D. Hanson’s essay “Estevan 1931.”[31] The essay is an historian’s account of a real episode in Canadian labour history: a story of violent resistance by coal mine owners to the efforts of miners to engage in collective bargaining near Estevan, Saskatchewan, in the early 1930s. The miners refused to negotiate individually; the owners refused to negotiate with the union. This impasse led to a strike, public demonstrations, violence, and three deaths. The reason that the “case” is so important is that it reminds students (and teachers) of labour law of what life was like in the days before the importation into Canada, in the 1940s, of a modified Wagner Act model and the creation of comprehensive statutory schemes of collective labour relations across the country.[32] The essay displays one of the great changes, or trade-offs, that those statutes brought about. In this case, the trade-off was the complete removal of the workers’ “self-help remedy” (such as it was), or what we would now call “striking for recognition”, and its substitution by the peaceful, democratic, and legally administered (by the labour relations board) process of “certification”.

There are other “pre-statutory” cases in the Casebook that establish the historical context of Canadian labour law. For example, Young v. C.N.R.[33] illustrates another of the statutes’ great trade-offs. At common law (and still in many places around the world) the way to enforce a collective agreement is, as the Privy Council patiently explained to Mr. Young, for the workers to strike (what we would call a “wildcat” strike). Under our statutory scheme, this “right”—such as it was—is removed completely; in its place is the peaceful, legally administered (by labour arbitrators) process of arbitration. The removal of the “rights” to both recognition strikes and wildcat strikes, and their substitution by non-disruptive legal processes, are formidable changes. These changes are at the core of the Canadian model, and they put great pressure on labour boards and arbitrators to “get it right” because there is no other way for workers to secure their rights now that “self-help” is illegal. The statutory changes that these cases are meant to highlight are fundamental. They are part of the grammar of our collective labour law system. They are also almost completely idiosyncratic to North America.[34]

In the statutory scheme, there are two other basic elements that go perhaps even deeper; in Ontario, these elements are contained in sections 5 and 17 of the OLRA. I turn first to section 17, which obliges both employers and unions to “bargain in good faith.” Again, the casebook deploys pre-statutory cases to try to drive home just what the statute really accomplished, or at least tried to accomplish, in terms of legal change. The central case that haunts the casebook is the 1940 decision of the Supreme Court of Canada in Christie v. York Corp.,[35] which now finds its expression in the Court’s 1981 decision in Seneca College v. Bhaduria.[36] Here, the Court gives unbridled expression to the common law freedom to decide with whom one will contract.[37] It was precisely this common law freedom that drove the conflict in Estevan: the employers were perfectly free to decide with whom they would and would not bargain. After certification, the statute removes the common law freedom of the employer to refuse to bargain. This development is the basic and innovative point of the statute. However, no one can deny that this innovation is statutory. The Supreme Court of Canada was deciding Christie at the same time as the Wagner Act model was being deployed in Canada. In fact, overruling Christie is the central point of the statute.

The other fundamental aspect of the OLRA lies in section 5, which simply states that “[e]very person is free to join a trade union of the person’s own choice and to participate in its lawful activities.”[38] Labour law casebooks no longer spend much time contrasting this simple articulation of a freedom with the complex matrix of legal rules that existed just prior to the statute. This change is evident on a scan through Bora Laskin’s 1947 casebook, which allocated about 200 of its approximately 300 pages to parsing these legal rules.[39] It would, in my view, take a very brave labour lawyer or legal historian to say that this legislative provision does no work (that is, that it reflected legal reality the moment before the statute was enacted); I know of no Canadian labour lawyer who would agree with such an assertion. However, I also believe that few have a command of the legal details required to prove their belief. Once statutes like the OLRA were passed, our residual concerns about unions or their activities as illegal in and of themselves completely evaporated—but our worries about the legality of participating in union activity remained. Above, I was careful to describe the employees’ “right”, at common law, to strike for recognition or enforcement of the collective agreement “such as it was”. When I went to law school many years ago, labour law professors like Innis Christie knew the common law cases creating the economic torts that restricted this right inside and out.[40] For the most part, current labour law professors have not had to read these cases. The torts have, to a large extent, and as was demonstrated by the Court in International Longshoremen’s Association, Locals 273, 1039, 1764 v. Maritime Employers’ Association,[41] evolved away, leaving it to the statute to regulate the legality of strikes. However, most of us do have at least some foggy memories of the “labour injunction”. Furthermore, modern labour lawyers actually know quite a bit about the picketing cases, which long continued to cabin any effective legal recognition of a right to strike; Hersees of Woodstock Ltd. v. Goldstein[42] was, after all, decided in 1963, and Harrison v. Carswell in 1976.[43]

As clear and as murky as the legal reality was at the point just before the enactment of the statute, both the clarity and the murkiness point in one direction: there was no clearly protected legal right to belong to a union and to participate in its lawful activities, the most critical of which in a system of collective bargaining is the strike. Beyond any doubt, there was no duty imposed on an employer to bargain with a union—even if, contrary to all legal indications, there was an effectively protected right to belong to a union and to participate in a strike. It is also evident that, even if unions entered into collective agreements, they were not legally binding. Once again, the right to strike, which is the common law’s equivalent to arbitration to enforce such agreements, was subject to the judicial hostility expressed in the economic torts, aided and abetted by the attending processes of the labour injunction.

The Supreme Court of Canada’s reading of Canadian legal history seems different. I do not review this reading in detail; I leave it to the real students of Canadian labour law history, including Judy Fudge and Eric Tucker,[44] to parse this aspect of the case. It is true that direct criminal liabilities were gradually removed for the act of unionization. However, the idea that the history of Canadian labour law just prior to the statutory innovations of the 1940s reveals that “[a]ssociation for the purposes of collective bargaining has long been recognized as a fundamental Canadian right”[45] cannot, in my view, be sustained. Nor does the more recent law sustain this idea given the continued, judicially imposed limitations on the right to try to mount an effective strike.

Leaving aside these more general problems, about which others may have more to say, it is important to identify a critical flaw in the Court’s reasoning regarding the second proposition. This flaw comes in the form of a very damaging elision. The Court wishes to employ the second proposition in support of its conclusion that section 2(d) comprehends a right to collective bargaining. In its efforts to achieve this result, the Court associates itself with three subpropositions:

  1. The right to collective bargaining in Canada is “pre-statutory”. The Court begins its discussion of the second proposition by writing that “[a]ssociation for purposes of collective bargaining has long been recognized as a fundamental Canadian right which predated the Charter.”[46] The idea that this “right” was not in any way dependent upon the statutes found very strong expression in Dunmore. The Court in BC Health put it this way:

    The respondent argues that the right to collective bargaining is of recent origin and is merely a creature of statute. This assertion may be true if collective bargaining is equated solely to the framework of rights of representation and collective bargaining now recognized under federal and provincial labour codes. However, the origin of a right to collective bargaining in the sense given to it in the present case (i.e., a procedural right to bargain collectively on conditions of employment), precedes the adoption of the present system of labour relations in the 1940s. The history of collective bargaining in Canada reveals that long before the present statutory labour regimes were put in place, collective bargaining was recognized as a fundamental aspect of Canadian society. This is the context against which the scope of the s. 2(d) must be considered.[47]

  2. The pre-statutory law was clear. There was no duty upon an employer to bargain. The holding in Christie still ruled. It was the statutes that brought “compulsory” bargaining to Canada.

  3. The content of the section 2(d) guarantee not only “means that employees have the right to unite, to present demands to health sector employers collectively and to engage in discussions in an attempt to achieve workplace-related goals” but also that “[s]ection 2(d) imposes corresponding duties on government employers to agree to meet and discuss with them.”[48] Or “the employees’ right to collective bargaining imposes corresponding duties on the employer. It requires both employer and employees to meet and to bargain in good faith, in the pursuit of a common goal of peaceful and productive accommodation.”[49] Furthermore, there is also “the fundamental precept of collective bargaining—the duty to consult and negotiate in good faith.”[50] And so on.

A, B, and C cannot all be true.

B is true.

If B is true, we cannot get from A to C.

The Court clearly accepts the validity of B. Yet the impact of the truth of B is never commented upon. This omission is only possible because C is not put on the table in the decision until the discussion of the second proposition is long over. The whole discussion of the second proposition seems to establish that Canadian pre-statutory law supported a right to collective bargaining without the duty to bargain.[51] The idea that a duty to bargain is part of the section 2(d) right negates all of the force that the Court wished to derive from its account of the legal history of that right.[52] This loss occurs because the legal history was invoked to suggest that the Court’s conception of the right to collective bargaining is historically (and not merely contingently, recently, and statutorily) embedded in domestic Canadian law. The Court’s third proposition is meant to show that its conception of collective bargaining is not only deeply true in Canadian law but also widely true in virtue of international labour law. However, this is not true: we have just seen that it is not true of Canadian domestic law without the statutes, and it is not true of international law at all. I now turn to this latter point.

C. The Third Proposition: International Law Treats Collective Bargaining as a Component of Freedom of Association

collective bargaining is an integral component of freedom of association in international law.[53]

Monitoring and commenting upon the use of international law by domestic courts is a global growth industry. In Canada, the use of international norms in Charter cases has given birth to an important local branch of this global legal enterprise. As the Supreme Court of Canada put it in BC Health, “Canada’s international obligations can assist courts charged with interpreting the Charter’s guarantees ... Applying this interpretative tool here supports recognizing a process of collective bargaining as part of the Charter’s guarantee of freedom of association.”[54]

Before turning to the substance of this issue, one should note the same potential for elision exists here as it did in the second proposition. Canada’s obligations under international law may or may not support the idea that freedom of association includes a right to collective bargaining. However, even locating such support will prove pointless if we later specify the content of that right in a way that international law does not support.

The argument offered here is an abridged version of that presented in “Can We Rely on the ILO?”[55] In that paper, I set out at some length the way in which the ILO and ILO law work, and those interested in the full argument should refer to that essay. The key point, which so many overlook, is that under the ILO’s constitution, ratification of ILO treaties (called “conventions”) is an entirely voluntary matter. The ILO’s constitution expressly states that a member state’s sole obligation is to place a convention, newly minted in Geneva, before the appropriate domestic authority for possible, voluntary ratification. If a country fails or decides not to ratify then, according to the constitution, “no further obligation shall rest upon the Member.”[56] This point is very important; understanding it helps avoid all sorts of difficulties and, for example, misstatements to the effect that the 1998 ILO Declaration on Fundamental Principles and Rights at Work[57] binds non-ratifying members to the “fundamental” conventions, or that the Committee on Freedom of Association (CFA) process performs the same constitutionally impossible magic.[58] We miss the whole point of these innovations, and their potential and limitations, if we do not attend to this fundamental legal truth.

In connection with the issues at stake in BC Health (that is, the relationship of the Charter guarantee of freedom of association to collective bargaining), the ILO has generated two key conventions, namely Convention 87: Freedom of Association and Protection of the Right to Organise Convention, 1948,[59] and Convention 98: Right to Organise and Collective Bargaining Convention, 1949.[60] Canada has ratified C87 but not C98.

There are three significant problems with the way the Court actually relies upon ILO law in BC Health. In its account of Canada’s international law obligations, the Court begins with a discussion of two other, more general human rights treaties to which Canada is a party and that address freedom of association. The Court then adds Canada’s ratification of C87 in 1972 to the list. Nowhere in the judgment does the Court mention that Canada has not ratified C98, which is, after all, the convention dealing with collective bargaining. This is the first problem.

The Court states in BC Health that “Convention No. 87 has also been understood to protect collective bargaining as part of freedom of association”[61] but does not cite a single word of C87.[62] The decision then proceeds: “Convention No. 87 has been the subject of numerous interpretations by the ILO’s Committee on Freedom of Association, Committee of Experts and Commissions of Inquiry”[63] and quotes an academic commentator who describes these interpretations as the “cornerstone of the international law on trade union freedom and collective bargaining.”[64] However, the Court cites neither a single decision of the CFA nor a single comment by the Experts or Commissions. What the decision does instead is to take an important turn to a “recent review by ILO staff [that] summarized a number of principles concerning collective bargaining.”[65] The article in question is by Bernard Gernigon, Alberto Odero, and Horacio Guido—all very experienced current and former ILO lawyers who, inter alia, have been involved with servicing committees on the issues of freedom of association.[66] The Court then selectively quotes some of the general principles that the authors place at the end of their article as a way of summing up.[67]

What no one seems to have noticed is that C87, on freedom of association, is not the focus of the article. The article is, as the title “ILO Principles Concerning Collective Bargaining” indicates, largely all about C98, on collective bargaining. All of the critical passages relied upon by the Court concern C98, which is the convention, it will be recalled, that Canada has not ratified. Thus, the article by Gernigon, Odero, and Guido is not a summary of Canada’s international law obligations. This is the second problem.

Then comes a third problem: the selective reading of Gernigon, Odero, and Guido’s article. The background to this issue is that the jurisprudence the authors elaborate is not terribly coherent and it is very hard to fit with the wording of C98 (let alone C87, which is not actually considered). This conceptual incoherence is not surprising given the CFA’s pragmatic and functional approach to its mandate. Even leaving the CFA aside, however, the critical issue is the failure of the Court to consider the whole article and to confront what its authors clearly say. Gernigon, Odero, and Guido write that freedom of association not only comprehends the freedom of workers to bargain collectively (which in my view it does) but also imposes a duty upon employers to bargain (which in my view it does not). If the Court had taken a look at C98, on collective bargaining, it would have found that the convention’s central demand, made after setting out requirements regarding what we would identify in Canada as unfair labour practices, is that ratifying states shall take measures to encourage and to promote the full development and utilization of machinery for “voluntary negotiation between employers ... and workers’ organisations.”[68]

The key word here is “voluntary”. The interesting question is, as becomes clear in BC Health, how could the commitment to voluntary negotiation in C98 provide interpretative support for a legal, indeed constitutional, ‘duty’ to bargain? The answer is that it simply cannot. It would have been helpful if the Court had, in its reading of the article by the three distinguished ILO officials, looked beyond the authors’ efforts to summarize at a general level and considered what Gernigon, Odero, and Guido had to say specifically about the vital question of a duty to bargain. On this issue, the authors state—as we would expect in light of C98’s central commitment to voluntary negotiation—the following:

The voluntary nature of collective bargaining is explicitly laid down in Article 4 of Convention No. 98 and, according to the Committee on Freedom of Association, is “a fundamental aspect of the principles of freedom of association” ... Thus, the obligation to promote collective bargaining excludes recourse to measures of compulsion. ... It cannot therefore be deduced from ILO’s Conventions on collective bargaining that there is a formal obligation to negotiate or to achieve a result (an agreement).[69]

One would think that these words have some bearing on the issue of whether international law imposes a duty to bargain upon the employer. In fact, a footnote following the text just quoted reads: “The obligation to negotiate is imposed in certain countries.” Those countries include Canada (and the USA), where we have, as discussed above at length, imposed upon employers a statutory duty to bargain. However, the footnote does not apply to the vast majority of ILO members.[70] A duty to bargain not only does not appear in C98, but is also inconsistent with the core idea of the convention: the voluntariness of collective bargaining.[71] This is the third problem.

So, to summarize: the relevant ILO law is not what the Court thinks it is—and, in any event, Canada is not a party to it.[72]

D. The Fourth Proposition: The Court’s Interpretation of Section 2(d) in BC Health Is Consistent with Other Charter Rights

interpreting s. 2(d) as including a right to collective bargaining is consistent with, and indeed, promotes, other Charter rights, freedoms and values.[73]

The Court’s reasoning in favour of the fourth proposition has the same structural problem as that identified in the second and third propositions. The unelaborated concept of the “right” to collective bargaining is presented as fitting with other Charter values. This presentation is unexceptional precisely because, once again, at this stage in the judgment the substance of that right as understood by the Court is not yet apparent. For the reader not anticipating what the Court will actually say in this regard, the judgment’s elaboration of the fourth proposition reads in a familiar way. It contains a recitation of what every labour law student knows is the “received wisdom” concerning the purposes and virtues of labour law in general and collective bargaining in particular. However, once the Court reveals the content of the right as involving an employer’s duty to bargain, it is no longer possible, in my view, to invoke Charter values to underwrite that outcome. The theoretical arms that the fourth proposition is meant to provide for the Court’s campaign towards its conclusion are, in the end, incapable of reaching the desired target.

At the core of the Court’s reasoning on this issue lies the constitutional mystery of why and how section 2(d)’s guarantee of freedom of association for employees imposes a duty upon an employer. At the risk of sounding overly Hohfeldian, my freedoms are not obviously connected to direct reciprocal duties imposed upon you;[74] they would seem to embody an odd idea of freedom if they did. The basic idea, which I defend below, is that my freedom to bargain individually or collectively does not lead to a duty upon you to bargain with me. We can say that I have a “right” to bargain collectively, meaning that I have a right that you bargain with me, and that you have a corresponding duty to bargain with me. This duty exists by statute in Canada. However, this idea of a right cannot be conjured directly from my constitutional freedom. This point needs to be argued, and I return to it below.

The idea that the Charter requires legislatures to impose a duty to bargain is constitutionally unsound for several other reasons as well. Among other things, to impose such a duty is in fact to do exactly what the Court said section 2(d) does not do: to impose a “particular model of labour relations”[75] upon Canada; in other words, to constitutionalize an idiosyncratic element of North American labour relations. Such constitutionalization is probably a bad thing to do. It is certainly a bad thing for a court to do.

The basic idea is this: as with other freedoms, like freedom of speech, of religion, or of thought, the freedom does not and should not mean that anyone else has a duty to listen or even to hear. Let alone to converse. Let alone to agree. This is equally true for association—in all its forms, including collective bargaining. The Court’s focus on the “inherently collective” nature of section 2(d) rights obscures this important point.

To understand these points, it is necessary to go back to some very basic legal ideas. Above, I mentioned Hohfeld who, for most writers who have concerned themselves with freedom of association or rights in general, lies at the centre of clear thinking about these issues.[76] It is critical, however, to understand Hohfeld’s role in legal thinking.[77] Hohfeld cannot solve our political and moral controversies; he can merely make sure that we are talking clearly and about the same things when we speak of “rights”.

Hohfeld wanted to clear up what he saw as evident confusion in legal discussions of “rights”. He correctly saw that the word has very different uses in legal discourse. He sought to bring clarity to this matter and to show the various and very different uses to which “rights” are put. His key distinction was between “claim rights” (I shall just use the word “right”) on the one hand, and “privileges” (what we would call “liberties” or “freedoms”; here I use the word “freedom”, as does our Charter) on the other. Sometimes, when we use the word “right”, what we mean is what Hohfeld meant by “right”, and sometimes we mean what Hohfeld meant by “freedom”. They are very different things—but we slide back and forth between the two usages of the word “right” with seeming abandon. In fact, I have done so throughout this essay. We should not be so ambiguous—or perhaps more realistically given common usage, we should be aware when we use the word “right” that it can be taken as meaning two very different legal animals, and we should be able to specify which one we mean.[78]

The whole question in BC Health is whether collective bargaining, in Hohfeld’s terms, is a “right” or a “freedom”. Merely invoking the word “right” cannot entail the answer, for we are perfectly comfortable using “right” to mean both in everyday parlance. However, when the very constitutional issue is “which one is it?”, we need to be more careful. Hohfeld cannot tell us the answer to this question, but he can clearly identify what is at stake when we ask it. Better, he allows us to ask the question by explicating the two possible answers that remain, seemingly to many, obscure. What then is the distinction between a Hohfeldian right and a Hohfeldian freedom?

The answer is as follows. My rights implicate the action or inaction of someone else. I have the right that you not assault me; you have the duty to not assault me. The subject matter of my freedoms, on the other hand, is my own action or inaction. I have the freedom to think, to speak my mind, to follow a certain creed if I am so inclined—and you have, as Hohfeld puts it, no right that I not do these things. It makes no sense to say “I have the right to do or not to do something” in Hohfeld’s terms. I have freedoms to do or not to do things.

Once we bear these distinctions firmly in mind, we are in a better position to approach the issues at stake in BC Health. We can begin by noting that the Charter actually deploys Hohfeld’s basic distinction in its very name—the Charter of Rights and Freedoms. And there is more. Section 2 of the Charter is about “fundamental freedoms”. It is followed by provisions dealing with “democratic rights”, “mobility rights”, “legal rights”, “equality rights”, and so on. This structure seems, easily and generally, to fit with Hohfeld’s insights. Look at what the fundamental freedoms include: conscience, thought, expression, and association. It makes no sense to think of freedom of conscience or expression, for example, as a right that I have, corresponding to your duty to think (or not to think) or to say (or not to say) anything.[79] That would be crazy. Rights, however, are of a different nature. My Charter right not to be arbitrarily imprisoned is not about my doing or not doing anything; it is about someone else (the state) not doing something (and having a duty to that effect). That right is a Hohfeldian right. Of course, my freedom of expression is backed up by a perimeter of rights that requires you to refrain from certain actions (for example, putting your hand over my mouth as I speak). My freedom is also limited by the freedoms and rights of others (for example, to exclude me from their property when I say things that they do not wish to hear). There is a very basic legal grammar at work here, and Hohfeld helps us identify it.[80]

The question in BC Health is whether the Charter is correct, in Hohfeldian terms, when it describes the constitutional guarantee of “freedom of association” as a “freedom” rather than a “right” (strictly speaking). I think that it clearly is—and if this is so, then the substance of the freedom is not about another person having a duty to do or not to do anything. It must also be true that any “right” (loosely speaking) to collective bargaining that follows from freedom of association is also a “freedom” (strictly speaking) and does not impose reciprocal duties upon others. It is not a “right” that I have that someone else do or not do something; it is my freedom to do or not to do something.

Earlier, I was so bold as to say that to characterize the freedoms in section 2 of the Charter as rights in a true Hohfeldian sense would be “crazy”. That word is not a technical legal term, nor is it an argument. But the assertion is correct. Why? I think that the answer lies in our very basic political and moral commitments—commitments so basic that they seldom break the surface of legal debate.

Here is a short version of how this thinking goes. To conceive of freedom of conscience, belief, religion, expression, or association as Hohfeldian rights (that is, as true rights that I have, which impose corresponding duties upon others to think or to act in a certain way) is to miss the point of such freedoms. To think in this manner would not simply delimit the freedom wrongly; it would constitutionalize the very opposite of what we wish to preserve.

However, we know intuitively that the task of constitutional interpretation lies in just the opposite direction, in constructing a perimeter of protections around the freedom. Freedoms give us claims that others not interfere with our exercise of them; they do not give us claims that anyone else join in our exercise of them.[81]

It is possible to say more, which leads me to the heart of the idea of “Charter values”. The Court thinks that imposing a duty to bargain on others is a necessary part of my freedom. It does not mention Hohfeld, nor the logical, moral, political, and legal intersection that must be crossed in imposing such a duty. It is not clear that the Court sees that this crossing exists. But if we come to the edge of this conceptual divide, and see it, what should we do—cross it or not?

We should begin with a reminder that this is, after all, a constitution that is being interpreted. It is an entrenched bill of rights and freedoms. The Charter value of freedom of association is a basic one. It applies to all Canadians, and the role of the Court is to interpret it in a principled way. There is, in my view, not one freedom of association for Nova Scotia and another for Ontario, nor one for students and one for tenants, nor one for the service sector and another for the manufacturing sector. If there were, this freedom would not be a matter of fundamental justice. It would be, to use the Court’s word, “contextual”.

Here is where the Court’s idea that it should “contextualize” its approach to section 2(d) comes home to roost. There are at least two sorts of concerns arising from this approach: one about understanding fundamental rights and the other about the pragmatic dangers of resulting judicial overreach. First, some of the pragmatics. BC Health constitutionalizes what most reasonable Canadian labour lawyers would agree is a sensible part of the statutory scheme. This is deeply worrying at the level of constitutional practice. It takes the Court off the path of legitimacy in constitutional matters and into the thicket of what generations of Canadian labour lawyers have railed against (and with good reason): judicial review of the details of our labour law. The path that Canadian labour law has taken is littered with the wrecks of ill-considered judicial wanderings. The Court has now committed itself to such outings on an ongoing basis. We have no reason to think that a judicially drafted labour law will be a good one. Then there is the very difficult question of where the process will end. What will the complete, judicially mandated “agricultural labour code of Ontario” (the construction of which began in Dunmore) look like? What about the complete, judicially imagined federal law of “RCMP collective bargaining” (which began life in Delisle)? Do we want to know? [82] Enough has been said about this type of experiment throughout the history of Canadian labour law. I need only say here that the Charter provides no justification for such an undertaking.

We also have deeper reasons for concern. There is only so much that we can and should ask our constitution to do for us. A constitution’s effort to construct a system of equal liberty or sovereignty involves the idea that its rules must be basic and universal. This understanding rightly places powerful constraints on “contextualization”: one cannot contextualize the basic grammar of the idea of a freedom. This insight saves us, or should save us, from the other, more pragmatic problems that we have noted. However, the key is to see that freedom of association, freedom of collective bargaining, and freedom to form, to join, and to participate in other trade union activities cannot and do not impose direct, correlative duties on others. The fact that I cannot participate in a union unless someone else joins me cannot, does not, and should not impose a duty on anyone else to join me. Everyone else has the same freedom as I do to join or not to join. If we get this wrong, we have missed a very large point.

Freedom of association does, on the basic logic relied upon in this paper, lead to a freedom to engage in collective bargaining. But again, the fact that I (or we) cannot bargain without someone to bargain with cannot impose a duty on others to bargain with me (or us). Just as the horrible truth that I cannot have a union unless at least one other person joins with me cannot impose a duty on anyone else to join with me.

If freedoms did not work this way, they would not be freedoms. Freedoms can be respected, protected, and promoted. But it is the promotion of freedom, not its negation, that must be the object of this exercise—even if the refusal of others to exercise their freedom means that I cannot exercise mine. Important and fundamental as constitutional freedoms are, they are important precisely because they are about what we are free to do, not what we can demand of others as of right.

However, as we shall see, this problem does not arise in a tolerably well-functioning democracy. This is because there are other constitutional ideas available that do not share this grammar and that, given what we know about North American labour relations, attend to the instinct to impose a duty to bargain in order to do “the right thing”. The main constitutional ideal that has to carry the freight here wrongly loaded onto section 2(d) is, as I noted early on, the idea of equality. Rather than constitutionalize key elements of the statutory scheme in order to extend it to those who are denied its promises, rather than confuse freedoms with rights, rather than allow judges to draft labour codes, rather than overrule the very important idea that there is no constitutional right to contract, a better constitutional approach is to ask the question: is there a rational reason why certain people are excluded from the statutory regime that instantiates a fundamental freedom open to others? Asking this question is all that is required to avoid the pitfalls of Dunmore, Delisle, and BC Health. It also avoids many crises, including that of legitimacy, that the path actually taken renders inevitable. It is to this idea that I now turn.

II. Thinking About Constitutional Labour Rights

A. How the Court Boxed Itself into the Freedom of Association Mess

The internal critique in Part I advances the claim that the second, third, and fourth propositions in BC Health do not provide support for the Court’s conclusion that section 2(d)’s freedom of association guarantee protects collective bargaining as the Court conceives of it. When the Court finally reveals its ideas about the substance of the right to collective bargaining, one major component is a duty to bargain. However, as we have seen, Canadian pre-statutory law, Canada’s international law obligations, and Charter values do not, in fact, support such an idea.

In this Part, I wish to abandon this internal critique, which takes BC Health on its own terms, and to address directly the issue that I believe lies at the heart of the decision: the relationship between the Charter’s guarantee of freedom of association and Canada’s labour relations statutes. This issue is never explicitly addressed or even identified in BC Health. Instead, the Court tethers itself to a series of constitutional commitments that simultaneously block it out of a sensible approach to the issues in BC Health and box it into the undesirable approach that it adopts in that case. The key to understanding the mess we are in lies in taking a broader view of the constitutional ideas in play. I begin this analysis by returning briefly to these prior commitments and the resulting legal circumstances—circumstances all of its own making—in which the Court finds itself in BC Health. Here is a brief review of how we got into the freedom of association mess:

  1. The Charter came into force in 1982. The equality provision, section 15, entered into force in 1985.

  2. In 1986, we learned in Retail, Wholesale and Department Store Union, Local 580 v. Dolphin Delivery Ltd.[83] that the Charter’s guarantee of freedom of expression does not extend to private litigation over situations in which employers use common or civil law rules to obtain injunctions to restrict picketing. (This is the “state action” requirement.)

  3. In 1987, in the trilogy of so-called “right to strike” cases, the Court held that the Charter’s section 2(d) guarantee of freedom of association does not include protection of a “right to strike”.[84]

  4. In 1990, in PIPS, the Court informed us, to the surprise of careful readers of the trilogy, that the trilogy excludes not only a right to strike but also a right to collective bargaining.[85] (Combining the trilogy and PIPS, we get what I refer to as the “thin theory of freedom of association”: it contains the right to establish, to belong to, and to maintain a union, but not the right to bargain collectively or to strike.)

  5. In a series of important cases in the 1980s and 1990s, the Court “read down” section 15 of the Charter from a true equality provision to a much narrower nondiscrimination provision. Section 15, in the Court’s view, only applies to cases in which differential treatment is based on what we now refer to as “enumerated or analogous grounds” (and then only if this treatment is also discriminatory). In 1989, in Reference ReWorkers’ Compensation Act, 1983 (Nfld.),[86] we learned that employment status is not an analogous ground. (This is the “thin theory of equality.”)

  6. In 1999, in Delisle, we were told by the Court that the exclusion of RCMP officers from any statutory scheme of collective bargaining results did not violate section 2(d) because the state has no duty to legislate to protect freedom of association from private (employer) interference.[87] The only restriction is that the state itself not interfere. As well, the underinclusiveness of the legislation does not violate section 15 as understood by the thin theory of equality. However, RCMP officers are free to set up their own association outside of the statute, and as they are government employees they have “direct” section 2(d) protection in court against interference from their employer (the government), because this interference would be state action. In other words, the courts are now in the unfair-labour-practice-adjudication business.

Thus, after the trilogy and the PIPS and Delisle decisions, the Court found itself, as noted above, inside a box of its own making. The four sides of that box were the Court’s equality jurisprudence (the “thin theory of equality”); the “state action” requirement; a basic position against “positive obligations” on governments to protect section 2 rights; and the “thin theory of freedom of association.” After Delisle, these four results conspired to render labour law a virtual constitutional “‘no go’ zone”, as the Court put it in BC Health.[88] This rendering was justified by reminders of the need for judicial restraint in labour law matters.

The important question is whether we can have real (thick, not thin) constitutional rights for labour without judicial overreach and without misunderstanding our past, our international obligations, and the nature of freedoms. The answer is yes, and the primary vehicle for this answer is the idea of equality. In fact, this is the real explanation for what is wrong in Delisle and what is right in Dunmore and BC Health.

Post-Delisle, and as a result of the four-sided box of the Court’s own construction, we were left with a dismal set of rules. Employer interference with employees’ exercise of freedom of association rights is generally beyond the reach of the constitution, unless your employer happens to be the government. Workers exercising freedom of association rights are fair game for employers, as per the “state action” side of the box established in Dolphin Delivery and Delisle. However, Delisle also extended Charter protection to government employees “directly”, without any explanation as to how judges can change the law which governs here (i.e., Christie v. York and so on) for some workers but not others. But this protection for government employees covers only a “thin” content of section 2(d) and does not extend to a right to collective bargaining or the right to strike, as determined by the trilogy and PIPS.

The courts are now in the business of adjudicating unjust labour practice disputes, but only for workers in the public sector. Moreover, the failure of the government to pass unfair labour practice laws to extend the same protections to private-sector workers is not a constitutional wrong because section 2 does not give rise to positive obligations on legislatures. Even more radically, the legislature can arbitrarily provide such protection to some workers and not to others. Under the “thin theory of equality”, this is not a violation of the Court’s understanding of the Charter’s equality guarantee unless the arbitrariness happens to map onto a ground of discrimination. Not only is there no obligation to protect employees’ exercise of freedom of association rights in all of these ways, but the government is also free to interfere with the “thick” content of that right—including the right to collective bargaining and to strike—because the “thin theory of freedom of association” that makes up the fourth side of the box does not extend the right’s protection to these employees’ group activities.

All of these rules are justified by the idea that labour law is a subject matter requiring judicial restraint. These rules represent the situation pre-Dunmore and pre-BC Health. The key to BC Health lies not in the thin theory of freedom of association; it lies in the conjunction of the Court’s four prior constitutional commitments.[89] Even the Court recognized this sad state of affairs, and undertook the process of rewriting these rules in Dunmore.

B. Dunmore’s Judicially Imposed Labour Code

Dunmore dealt with agricultural workers who were excluded from the private-sector collective bargaining laws that apply to almost everyone else. We learn several interesting things in this case. We learn that despite the absence of any government obligation to legislate the protection of any workers’ section 2(d) freedoms from private-employer interference, if such legislation is passed for some workers and not others, this underinclusion can be a Charter violation—of section 2(d), not section 15.[90]

We also learn that the Court rejected the theory of section 2(d) set out in the earlier cases: specifically, that the right in question is the right to do in concert what we can do as individuals. According to Chief Justice Dickson’s dissent in the Alberta Reference, there are some union activities that are inherently collective and that have no individual analogues. As I have already said, this is an example of poor and dangerous thinking, yet it was followed in BC Health and will create many problems (with the right to strike, for example).

Dunmore imposed a positive obligation to legislate in a case of underinclusion to protect the exercise of section 2(d) freedoms by Ontario agricultural workers. However, because this case is said to concern section 2(d) and not section 15, the remedy provided was to not include the complainants in the OLRA despite the fact that their exclusion from this act formed the basis of their claim. Instead, the Court demanded that the legislature provide statutory protection of the exercise of the (thin) freedom of association. That is, the Court drafted an agricultural workers’ labour code, even though the very code that these workers required was already on the statute books and formed the basis of their claim (i.e., exclusion from that scheme), and even though the Court’s remedy was to cut-and-paste unfair labour practice provisions directly from that scheme. Dunmore hinged upon an equality argument, an equality holding, and an equality remedy—but the Court could not recognize these as such, given its commitment to its “thin theory of equality”.

So, after Dunmore, we must amend the rule that the legislature may arbitrarily provide protection for the section 2(d) rights of some workers while excluding others, by adding the words “unless the workers are employed in the agriculture industry.”

However, it is obvious that the Court went to great lengths so as not to revisit its commitment to the thin theory of equality, the real culprit in this case. This is why the reasoning in Dunmore is incoherent. The Court attempted to stuff what was really a section 15 case into section 2(d), even though this is impossible and even though readers of the decision can discern what was really going on. This particular rabbit cannot come out of this particular hat, and everyone can see from which hat it actually did emerge.

This abuse of sections 2(d) and 15 is highly undesirable. In Delisle, we see the Court getting directly into the unfair-labour-practice-adjudication business. In Dunmore, we see the Court getting into the business of drafting an agricultural workers’ labour code. These are obviously bad ideas. And where will it all end?[91] (For example, I take it that, after BC Health, the agricultural workers’ labour code now contains a duty to bargain.) Why didn’t the Court just say that these were equality cases—that these workers were excluded from the relevant statute’s instantiation of the section 2(d) freedom without good reason? And why not simply remedy these section 15 violations by including these workers in the statutory scheme? That would be a direct and simple solution.[92]

C. BC Health’s Impact

BC Health is, on its face, a different sort of case. It primarily engages the fourth side of the box, the thin theory of freedom of association. It is a case that seems not to be about the right to demand protection from the state against employer interference, but about the state itself, as the legislator, attacking this right.[93] However, it becomes clear almost as soon as the issue is thus reframed that all of these cases are about attacks on this right. It is only because we are inured to certain ideas regarding “state action” and “positive obligations” that a distinction between an attack and a failure to protect is possible. If employer and legislative attacks are seen as equally threatening to the right, then the legislative failure to protect against the former and the legislative action in the latter are seen to be cases of exclusion from a constitutional right as instantiated in the relevant statute—cases that raise very important equality claims. But this path was not taken. The Court in BC Health, as in Dunmore, was forced to revisit another side of the box. As I have noted, in this case it is the fourth side, the thin theory of freedom of association. And as we have seen, focusing on side four results in further, and in my view unhappy, theorizing about the nature of the section 2(d) right as a collective one.

The basic idea of equality was designed to carry important freight. The Court’s failure to make use of section 15 has required the transfer of that freight to section 2(d) and the construction of a detailed parallel system, adequate to its conception of the constitutional idea of freedom of association. In the most recent version of this model (i.e., the Court’s approach in BC Health), workers get collective bargaining, the centrepiece of which is the peculiar, North American statutory duty to bargain. This duty is taken hook, line, and sinker from the statute and the detailed jurisprudence regarding those statutory provisions. This complex, roundabout way of doing business expends too many resources, including basic institutional legitimacy. And when will the Court stop introducing fine adjustments to borrowed statutory concepts regarding, in BC Health for example, the distinction between mandatory and permissive bargaining? I suspect that the depth of trouble invited here has yet to be fathomed.

We are left with the following set of arbitrary rules regarding one of our most fundamental freedoms in an open society. BC Health only changes the theory of freedom of association from a thin to a thicker one; that is, it tinkers with another non-equality side of the box, this time overtly. Everything else remains in place. As a result, we would seem to be in a situation that forces us to restate the rules discussed in Part II.B slightly, as follows.

Workers exercising the fundamental right to freedom of association are still fair game for employers, unless the employer happens to be the government. BC Health again offers no explanation as to how the law that governs here—that is, Christie and so on—can be changed by judges for some workers but not others. However, this protection for government employees now covers not only the “thin” content of section 2(d)—that is, a right to form an association—but also extends to part of the “thick” content—that is, a right to collective bargaining, but not (yet) the right to strike.

The rule that section 2 of the Charter does not give rise to positive obligations on governments to pass laws protecting the rights of private-sector employees still stands, as does the rule that the legislature can arbitrarily provide such protection to some workers and not to others. Although there is still no obligation to pass laws protecting employee exercise of freedom of association rights, thick or thin, now the government cannot interfere directly (regarding its own employees), or by legislation (regarding private-sector employees), with the content of that right, thick or thin.

The idea that labour law is legislative subject matter requiring judicial restraint is abandoned.

If this dispensation is a happier one than that post-Dunmore, it is certainly only marginally so. Public-sector workers are guaranteed protection against attacks on their freedom of association. Private-sector workers are not, unless they are in agriculture. How would we explain that to the ILO? BC Health, while changing the content of the “thin theory of freedom of association” side of the box, leaves everything else in place. We have the freedom, it seems, but no right to protection of it—even if others do—unless we work in select sectors of the economy.

D. The Way Out

Our predicament post-BC Health, is that, in general, private-sector workers have no constitutional right to protection from employer interference with their efforts to organize nor do they have a right to collective bargaining (with or without an employer duty). As far as the constitution is concerned, private-sector workers can be fired or otherwise discriminated against for organizing, for making real demands, and certainly for striking—even if they are determined to exercise a constitutional right in the end. This situation seems backwards: the constitution should tell us when workers are entitled to statutory protection, not the other way around. The constitutional idea of equality gives us our first answer to this question: workers have an entitlement to the real, concrete instantiation of the right when others have it and they have been excluded without good reason.

The Canadian labour relations statutes have given expression to these constitutional rights for most workers. We can avoid all sorts of difficulties, including the incoherencies just noted, by taking this expression as our starting point and applying the idea of equality. We may not have a right to a particular instantiation of the freedom of association (as already noted, a study of the domestic systems of ILO members shows that there are many legitimate models), but once we have this instantiation for some, we must extend it to all unless there is a good reason not to do so.

This approach has many virtues. It does not constitutionalize our local statutory version of the freedom. It also avoids the judicial construction of a parallel labour code. There is an added bonus: the approach to restrictions upon freedom of association that is set out in BC Health could not withstand an all-out assault on collective rights if the legislature were to dismantle collective representation schemes and revert to common law rules of individual bargaining. This incapacity should tell us a great deal. When and if we really need section 2(d) and not section 15, it will not be there for us. The result would be a case of government inaction. It would be a global version of what the government of Ontario did in Dunmore—and, on the Court’s view, nothing could be done in response. Overcoming this inability would require revisiting the “no positive obligations” side of the box.

Given our current circumstances, the problem is very different. We do have a local statutory version of a comprehensive protection of the right, just not for all. The equality side of the box needs to be liberated to correct this constitutional wrong, without any of the side effects that we have noted.

If the Court does not employ the idea of equality to deal with cases of statutory exclusion, then the slack will, as in Dunmore, be taken up elsewhere. Once the equality argument is off the table, our chance of doing the right thing (that is, moving workers into the statutory scheme) is taken from us. The only other way to protect them is to construct a parallel universe outside of the statute. This is the project that the Court has taken on—or more accurately, been driven to—by virtue of boxing itself out of a more direct approach.

The Court’s own jurisprudence drives us to the curious results set out above. Its unwillingness to let section 15 do its job explicitly has put great pressure on other constitutional provisions. This unwillingness has also required the Court to reject the ideas of judicial restraint and humility in labour relations matters.

The Court’s view of the relationship between the statutes and the constitutional guarantee leads to some odd statements. For example, in BC Health, we are told that “there is nothing in the statutory entrenchment of collective bargaining that detracts from its fundamental nature.”[94] I find this sentence revealing. Who would ever think that in passing the statutory regimes into law in the 1940s, repudiating the almost-simultaneous decision in Christie and putting in place processes and safeguards that made the right “real”, Canadian legislatures were detracting from the “fundamental nature” of labour rights to freedom of association and collective bargaining? This idea seems very odd; it seems to have things inverted. The statutes are Canada’s indigenous legal expression of the right. Many experienced people think that this expression did a passable job, at least for quite a long time. It may be that we will need to change it.

The legislation is inextricably related to the international human (and now Charter) “right”. Labour relations statutes are not the expression of an unrelated state interest that somehow intrudes or impinges upon a freedom or right, which is the structure of almost all other constitutional litigation. On the contrary, the statutes are our way of actually making that freedom available to Canadian workers. The instantiation of the freedom is the state interest. However, if we start off without this idea of the relationship of the basic human freedom (expressed in three words), and the statute’s detailed efforts (expressed in thousands of words), we are likely to end up in some odd places. Among other things, we will not see exclusion from the statute as exclusion from the fundamental freedom. Recall that this argument does not say that the statute is written in stone, nor that the statute has it exactly right, nor that the right is constitutionally guaranteed (all of which are views to which the Court now seems unfortunately committed in its construction of the parallel world). The argument is simply that when a government extends a detailed statutory instantiation of a fundamental freedom to some, it has to justify not giving the freedom to others. This is the demand of the idea of equality.


To those who have been critical of the Canadian instantiation of freedom of association, Dunmore and BC Health are exactly what the doctor ordered. Roy J. Adams is right to speak of the “revolutionary potential”[95] of such cases. He sees the construction of the parallel system outside of the statute, flowing directly from international human-rights norms, as a new beginning—and he celebrates it. I take a different view. I see what Adams sees and think that this project is a dangerous distortion of the judicial role in labour law and in a just society. I see it as an utter failure to appreciate the nature of fundamental freedoms. I see it as a failure to grasp the important relationship between the abstract and concrete expressions of a basic freedom. I see it as driving us to an incoherent and dangerous account of freedom of association. I see it as opening up the idea of a constitutional “right” to contract. I see it as the continuation of an obvious and pathological desire to avoid the demands of one of our most basic constitutional ideals: equality. I see it as a recipe for continuing the set of irrational distinctions (public vs. private vs. agriculture, and so on) that we now have in place. I see it as constitutionalizing (writing in stone) a very particular statutory scheme. I see it as a recipe for endless, incoherent, and needless litigation. All without end. And the most frustrating part is that all of these problems are unnecessary. The tools to do the right thing in these sorts of cases and in a way that would avoid all of these problems were at hand—and they still are.