The Panglossian Peril

The dangers of naïve optimism in thinking about constitutional constraint

In a provocative paper recently posted on SSRN (and based on the HLA Hart Memorial Lecture delivered last year at Oxford), Frederick Schauer challenges a fairly common tendency to argue that apparent conflicts between rights and important interests, or among rights, are illusory, and that, properly understood, these rights and interests can be reconciled so as to avoid the conflict. Prof. Schauer calls this tendency “Panglossianism”, after the obstinately and obliviously optimistic character of Voltaire’s Candide, and argues that it makes for muddled thinking that will end up compromising the rights that Panglossians purport to value. Prof. Schauer makes important points, although I am not persuaded by his takeaway.

* * *

Professor Schauer wants us to recognize that we cannot have it all ― socially desirable policies fully implemented and rights fully protected at the same time. He laments

the common but nonetheless troubling tendency of many people to perceive (or distort) the empirical aspects of various interests in a manner that eliminates the conflict between them and other interests, or between those interests and the rights with which they may conflict. And although people sometimes thus perceive interests in ways that make rights appear cost-free, they also indulge in the equally common tendency to define rights in a way that similarly eliminates the constraints that rights sometimes impose on legitimate interests. (1-2)

To relate just one of the examples he uses, when it comes to prohibitions on hate speech, those who oppose them will often insist that hate speech is not especially harmful, or is not harmful in ways that anyone should really care about, so that upholding the right to freedom of expression has no real cost. Conversely, many of those who support the criminalization of hate speech invoke the mantra of “hate speech is not free speech”, similarly insisting that their preferred resolution of this issue is costless. “Panglossianism” can accordingly involve either a reading of the data (or speculation) about the effects of policies that minimizes their impact on rights, or a redefinition of rights or other constitutional rules that narrows them so as to ensure that a favoured policy is not precluded.

Professor Schauer argues that Panglossianism is a mechanism people deploy to deal with the threat of cognitive dissonance that people might experience if they acknowledge that their preferred policies and constitutional commitments are in tension, and even in conflict. It is easier to believe, and to say, that such conflicts are not real, or can easily avoided, than to deal with them, which would mean taking sides, recognizing that one is wrong about rights or that one’s preferred policy cannot be implemented.

Yet Panglossiansim is a problem, for two reasons. First, intellectual honesty requires us “to avoid attempting to see the empirical world through the rose-colored glasses of one’s own normative desires”. (18) And second, rights are safer if they are not justified on the basis of empirical claims that are either  weak to begin with, or at best “temporally or culturally contingent [so that] the right may turn out to be weakened or inapplicable under different empirical conditions”. (19) Rights are stronger if they are grounded in pure moral principle than if their continued existence depends on whether it is, or is not, relatively harmless. Indeed, if only harmless rights deserve protection, then not much is going to be protected at all. Professor Schauer warns that policy-makers (whether in the legislative or in the executive branch) can Panglossians too; they will tempted to insist that there is no conflict between their preferred policy and constitutional restrictions on their action. But

[u]nlike the rights-concerned Panglossian commentator or advocate who interprets – or distorts – the empirical data so as to eliminate a conflict between policies and rights, here the official is more likely to attempt to eliminate the conflict between policy preferences and constitutional constraints by understanding the constraints in a way that makes them inapplicable to the issue at hand. (22)

Panglossianism, Professor Schauer notes, can undermine not only rights protections, but all manner of intended constitutional rules. Resorting to it may be psychologically comforting, but it will weaken the very idea of constitutional constraints on governments’ pursuit of their preferred policies.

* * *

I think that Professor Schauer describes a real problem. It is indeed tempting to say that the enforcement of one’s favoured right does not compromise the attainment of valuable policy goals or the respect of other rights; it is similarly tempting to insist the implementation of one’s preferred policy conflicts with no real rights, properly understood. Debates about free speech are one area where this dynamic is especially visible, as Professor Schauer notes, but there are any number of others. It is arguable (which is not necessarily to say true) that the controversy over the federal government’s demand that religious groups “attest” to the compatibility between their “core mandate” and (some) Charter rights, about which I’ve written here, also involves Panglossian arguments on both sides.

And Professor Schauer is quite right to point out that Panglossianism can affect thinking about structural constitutional rules, and not just rights. Indeed, I would suggest that in Canadian constitutional law, Panglossianism is an especially strong danger in federalism jurisprudence. In Charter cases, section 1, which authorizes the imposition of “limits” to rights, channels the analysis into a more explicit consideration of the conflict between rights (which tend to be defined in broad and abstract terms) and policy reasons for restricting them. By contrast, the movement towards the erosion of the exclusivity of federal and provincial heads of power under the banner of “co-operative federalism” proceeds from the Supreme Court’s unwillingness to acknowledge the existence of conflict between what it sees as desirable policy and the federal division of powers. Federalism analysis simply makes no room for the acknowledgement of this conflict. This is not to say that we should change the way we approach it ― but we should beware the Panglossian dangers inherent in what we do.

Yet while I think that there is a great deal of truth to Professor Schauer’s diagnosis of the pathologies of Panglossianism, his prescriptions against it may not be especially salutary. Professor Schauer does not tell us much about how to assess what he sees as potentially-Panglossian claims about the effects of policies or the scope of rights. He warns against thinking, for instance, that not punishing hate speech is costless because such speech does not really cause any social evils. Fair enough ― those of us opposed to bans on hate speech on normative grounds will be tempted to downplay its effects. But what if it really doesn’t have any? Conversely, if hate speech really is socially harmful, that happens to align with the preferences of those who want to ban it. Both sides in this particular debate cannot, I think, be wrong at the same time. The mere fact that an empirical claim aligns with someone’s prior normative preferences cannot mean that the claim is wrong. The same applies to claims about the scope of rights (to the extent that these can be said to be correct or incorrect at all).

So while we should be wary of the dangers described by Professor Schauer, he has not convinced me to give up on empirical or otherwise contextualized thinking about rights in favour of a priori philosophizing. This is all the more so in the numerous cases that concern what might be described as marginal (possible) infringements of rights. Perhaps the hate speech question, which is about whether people can be prevented from saying certain things at all can be sufficiently resolved by an a priori insistence that such bans are never permissible. Note, though, that the argument wouldn’t work the other way: a case for banning hate speech can only be made if one is allowed to rely on empirical considerations (unless of course one takes the position that there is no right to free speech at all and anything can be banned). But what about, say, restrictions on financing political parties? Most people accept that at least some restrictions are acceptable (most people in North America, anyway; New Zealand has no limit on how much one can give, and seems to be doing just fine!); many ― most, I hope ― would also agree that some restrictions are too extreme and cannot be justified. The issue is where to draw the line, and where to err in doubt. I don’t think that we can give remotely interesting answers to these questions without knowing something about the current practices of political fundraising and the likely effects of raising or lowering the existing restrictions. Again, Professor Schauer’s warnings about Panglossianism are relevant, but his suggestion that we resolve our questions by reference to first principles alone is not helpful.

Now, Professor Schauer is right, of course, that any empirically contingent answers might be inapplicable under different circumstances. He might be overstating the extent to which this is a problem: I’m not sure, for instance, that cultural contingency of rights protections is objectionable; it’s not obvious that rights must be the same everywhere and at all times. However, to the extent that, within a legal order, rights are implemented through judicially articulated constitutional doctrine, this doctrine risks being destabilized if the empirical or normative premises on which it is based are challenged by the evolution of society and of what we know about it. How to deal with this risk of instability (and its converse, the risk of a static doctrine divorced from reality) is a difficult question, to which I have no very good answers. But I doubt that we can avoid trying to get at some answers, at least, if only mediocre (and contingent!) ones.

Thinking about constitutional rules and their relationship with policy is a difficult business. Professor Schauer is right to remind us that we are too often tempted to oversimplify it by pretending that contradictions between our normative commitments and policy preferences are less significant than they really are. Unfortunately, he doesn’t offer us much by way of useful advice for identifying the exact situations where our thinking is so sidetracked, and his suggestion that we think more about abstract principle than about the real-world effect of policy does not strike me as especially helpful. Nevertheless, Professor Schauer’s warning is an important one, and we should heed it even if we conclude that we must continue exposing ourselves to the dangers he highlights.

Section 0

I wrote yesterday about the possibility that the relationship between sections 7 and 1 of the Charter might change in the wake of the Supreme Court’s assisted-suicide decision, Carter v. Canada (Attorney General), 2015 SCC 5. If I am right, however, the changes will only be relevant in a limited number of cases (albeit potentially significant ones). So I think it is also worth saying something about the relationship as it developed in the pre-Carter days, when justifying a section 7 violation under section 1 was generally regarded as well-nigh impossible. (Maxime St-Hilaire has posted a detailed review of the relevant jurisprudence as a comment to yesterday’s post. As he makes clear, while the Supreme Court has often enough analyzed section 7 infringements under section 1, this seems usually to have been pro forma.)

Section 7 is an unusual provision in the Charter because, unlike most others, it has a built-in internal limitation. It does not merely protect rights to “life, liberty and the security of the person,” but also, in effect, provides that it is possible to deprive one of these  ― so long as this is done “in accordance with the principles of fundamental justice.” What those principles are has been controversial (and remains so, as the criticism of the Supreme Court’s recent decision to recognize the lawyers’ duty of commitment to their clients as a new principle of fundamental justice in Canada (Attorney General) v. Federation of Law Societies of Canada, 2015 SCC 7 demonstrates.) However, as the Supreme Court pointed out in Carter,

three [principles] have emerged as central in the recent s. 7 jurisprudence: laws that impinge on life, liberty or security of the person must not be arbitrary, overbroad, or have consequences that are grossly disproportionate to their object. [72]

As the Court further explained, “[e]ach of these potential vices involves comparison with the object of the law that is challenged.” [73] Arbitrariness occurs “where there is no rational connection between the object of the law and the limit it imposes on life, liberty or security of the person.” [83] Overbreadth is a situation where, a provision that is not arbitrary nevertheless “goes too far by denying the rights of some individuals in a way that bears no relation to the object.” [85] Finally, gross disproportionality happens when “the impact of the restriction on the individual’s life, liberty or security of the person is grossly disproportionate to the object of the measure.” [89]

Now here’s the thing. These “central” principles look an awful lot like the three prongs of the “proportionality” part of the Oakes test used to decide whether the infringement of a Charter right is justified under section 1. Arbitrariness is like “rational connection”; overbreadth seems similar to “minimal impairment”; and gross disproportionality is reminiscent of the balancing over prejudicial and beneficent effects, or “proportionality stricto sensu.” It might seem, then, that section 7, although infringements to it cannot really be justified under section 1, actually comes with its own section 1 built in.

Except that the Supreme Court has been insisting that it really doesn’t; that, as the Chief Justice put it in Canada (Attorney General) v. Bedford, 2013 SCC 72, [2013] 3 S.C.R. 1101,  the “parallels” between these provisions “should not be allowed to obscure the crucial differences between the two sections.” [124] I have to admit that when I first read Bedford, this disclaimer seemed difficult to countenance. The parallels really seemed too flagrant. But with an additional explanation given in Carter, I think that there really is a meaningful difference between the “central” principles of fundamental justice and the “proportionality” analysis under section 1.

What makes the distinction between the two so elusive that the Supreme Court is forced to return to and try to clarify it in case after case is that, despite its insistence that “[t]he question of justification on the basis of an overarching public goal is at the heart of s. 1, but it plays no part in the s. 7 analysis,” [Bedford, 125] the “overarching public goal” is, in fact, the reference point for the section 7 analysis as well. And it doesn’t really help to say that, for the purpose of section 7, the objective must be “taken at face value,” [Bedford, 125] because that seems to be exactly what is going on in section 1 cases as well. Nor does it help to say that under section 7

[t]he inquiry into the impact on life, liberty or security of the person is not quantitative — for example, how many people are negatively impacted — but qualitative. An arbitrary, overbroad, or grossly disproportionate impact on one person suffices to establish a breach of s. 7. [Bedford, 127]

In theory, at least, it should be the same under section 1. The number of people affected might be a consideration, if at all, at the last, balancing stage, but if the law is not rationally connected to its objective, or if less impairing alternatives exist, it will not pass the Oakes test even if that number is small.

However, a passage in Carter clarifies things at last. The Court explains that

[i]n some cases the government, for practical reasons, may only be able to meet an important objective by means of a law that has some fundamental flaw. But this does not concern us when considering whether s. 7 of the Charter has been breached. [82]

In theory, at least, that is a (partly) plausible distinction. I say partly, because I still fail to see how a finding of arbitrariness could ever be rescued by the assertion of an “important objective” ― to which the impugned law is, ex hypothesi, not connected. But it seems possible that, in some cases, in would be impossible to draft a law sufficiently precise that it would not be overbroad ― yet, precisely because no possible alternatives exist, it would still survive the “minimal impairment” stage of the Oakes test. This was precisely the government’s argument in Carter ― and, although it did not succeed, it fared better than just about any other attempt to justify a section 7 infringement ever had. It might even be possible to show that a “grossly disproportionate” law has such enormous social benefits that it would survive the balancing stage of the section 1 analysis, though I’m pretty skeptical about that. In this way, the Court’s assertions that the principles of fundamental justice in section 7 are not about public or collective interests make sense and, as the Court has been telling us all along, the “principles of fundamental justice” in section 7 are not exactly an internal section 1.

But, keeping in mind the way in which sections 7 and 1 have interacted until Carter, and are likely to keep interacting in the future, I want to suggest that they are, in effect, a sort of section 0 ― by analogy to Isaac Asimov’s “Zeroth law of robotics.” Asimov’s original “three laws,” go like this (the lower-numbered laws over-riding the higher-numbered ones):

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey the orders given it by human beings, except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Eventually, however, Asimov added an additional law, which he call the “Zeroth,” because it had priority over the First:

0. A robot may not harm humanity, or, by inaction, allow humanity to come to harm.

The First law is accordingly re-written, allowing for its violations in order to respect the Zeroth.

The principles of fundamental justice are similar in that they are considered as logically prior to and, in their actual application, arguably capable of over-riding the “First Law” of the Charter, because they protect an interest even more important than it does.

In a way, however, the relationship between “section 0” and section 1 is the exact opposite of that between the Zeroth Law and the First. The Zeroth Law reflects Asimov’s conclusion that a collective interest was superior to the individual one ― indeed, to the individual interest in life and the security of the person itself. “Section 0” reflects the opposite view, which is logical considering that the Charter, despite its strong acknowledgement of collective interests in section 1, is still the product of an individualist philosophy. Section 1 is important (as is Asimov’s Fist Law) but, at least when it comes to the most fundamental rights of all, it must give yield to an interest that is more important still. Indeed, as I have said yesterday, the Supreme Court’s greater openness to considering a section 1 justification of  a section 7 infringement may well be consistent with, rather than a departure from, this principle, insofar as it still serves to uphold the very same rights protected by section 7.

And as for the question of who, as between the Supreme Court and Asimov, has got the priority of the individual and the collective right, I’ll let you decide. The contest might be a bit unfair though, considering that Asimov’s novels tend to be better written that Supreme Court opinions.

Seven and One

I want to come back to Carter v. Canada (Attorney General), 2015 SCC 5, the Supreme Court’s decision striking down an absolute prohibition on assisting a person to commit suicide, to comment on an aspect of the Court’s reasoning that seems, as best I can tell, to have attracted little attention. The Court found that, as applied to “physician‑assisted dying for competent adults who seek such assistance as a result of a grievous and irremediable medical condition that causes enduring and intolerable suffering,” the prohibition is contrary to section 7 of the Charter. Over at ABlawg, Jennifer Koshan has pointed out that “the Court’s reasons do not add much to the existing jurisprudence defining the scope of section 7.” That is true, so far as it goes. But Carter is a very unusual section 7 case all the same ― not because of what the Court says about section 7 itself, but because of what it says, and does, about the interaction between section 7 and section 1.

In Re B.C. Motor Vehicle Act, [1985] 2 S.C.R. 486, Justice Lamer (as he then was) alluded to “war, natural disasters or epidemics” as circumstances in which section 1 could justify infringements of section 1 (albeit that Justice Lamer was specifically discussing infringements motivated by “administrative convenience,” and not some weightier objective). We have (mostly) escaped such ordeals, and all the way up to the (relatively) recent landmark decisions in the Insite case, Canada (Attorney General) v. PHS Community Services Society, 2011 SCC 44, [2011] 3 S.C.R. 134, and Canada (Attorney General) v. Bedford, 2013 SCC 72, [2013] 3 S.C.R. 1101, section 1 has played virtually no role in section 7 cases. Unlike with most other provisions of the Charter, a finding that section 7 is infringed is as good as the end of the matter. Often, the government will not even make more than a perfunctory attempt to invoke section 1.

Carter changes that. The Court points out that

It is difficult to justify a s. 7 violation … .  The rights protected by s. 7 are fundamental, and not easily overridden by competing social interests … .  And it is hard to justify a law that runs afoul of the principles of fundamental justice and is thus inherently flawed. [95; internal quotation marks omitted]

Nevertheless the Court says that

in some situations the state may be able to show that the public good — a matter not considered under s. 7, which looks only at the impact on the rights claimants — justifies depriving an individual of life, liberty or security of the person under s. 1 of the Charter. [95]

Indeed, the Court embarks on a section 1 analysis which, although in some ways conclusory (as I have argued here), is unusually lengthy by the standards of previous section 7 jurisprudence. More importantly, it states that

The question in this case comes down to whether the absolute prohibition on physician-assisted dying, with its heavy impact on the claimants’ s. 7 rights to life, liberty and security of the person, is the least drastic means of achieving the legislative objective. … This question lies at the heart of this case. [103-104]

And so we must ask the question: is the Court is really opening the door to violations of section 7 being justified under section 1? This would be a big deal, and would force us to revise some widely held assumptions ― for example, about the unconstitutionality of the former Bill C-36, the anti-sex-work legislation Parliament enacted in response to the Court’s judgment in Bedford. (Though Michael Plaxton, for one, had cautioned against making such assumptions well before Carter.)

In Carter, the Court suggested that section 1 might save a law that infringes section 7 “particularly[] in cases … where the competing societal interests are themselves protected under the Charter” [95] ― Carter being one such case of course. This is, arguably, a logical line to draw. To violate “principles of fundamental justice” in the name of administrative convenience is one thing; to do it in order to protect life, liberty, or the security of individuals is something different.

Perhaps remarkably, few if any of the big section 7 cases actually dealt with laws that aimed squarely at protecting Charter rights rather than at some other, more pedestrian objective. At best, they aim at furthering diffuse interests such as “public health” and “public safety” ― not the individual health and safety of the people whose rights they restrict. (Criminal law would usually fall in that category. It does protect our right to the security of the person, and even our rights to life and liberty in some cases, but only indirectly, mostly through the rather uncertain mechanism of deterrence.) At worst, they are justified by administrative convenience. The restrictions on sex work struck down in Bedford, which the Court found sought to prevent nuisances, were arguably somewhere in the middle of that spectrum.

Assuming, then, that section 1 will (mostly?) be used to justify infringements of section 7 where the purpose is to protect defined rights of specific persons (and perhaps especially, the rights of more or less the same persons whose rights are being infringed), how common will such cases be? The experience of the last 30 years suggests that they will be few and far between. But at least one such future case readily comes to mind ― that of C-36, which ostensibly aims at protecting sex-workers from exploitation and violations of their dignity.

Carter may, thus, prove a rather more significant part of the Supreme Court’s section 7 jurisprudence than we originally supposed. It calls into question what used to be a safe assumption about that jurisprudence ― that section 1 was not really relevant to it. Of course, in Carter, the Court found that even a full-scale application of section 1 was not enough to save the infringement of section 7. Still, this innovation may yet have consequences ― not very positive consequences, either ― in important future cases.