Against Pure Pragmatism in Statutory Interpretation III: A Way Forward and Walsh (ONCA)

About a month ago, I wrote two posts attacking the concept of “pragmatism” in Canadian statutory interpretation. So my argument goes, the seminal Rizzo case, while commonly said to herald a “purposive” approach to interpretation, is actually methodologically pragmatic This is because the famous paragraph from Rizzo, which contains a list of things an interpret must take into account, does not assign ex ante weights to these factors. That is, it is up the interpreter to choose, in the circumstances of particular cases, which factors will be most relevant. In short, while everyone in theory agrees on what the goal of interpretation is, that agreement rapidly breaks down in the context of particular cases.

In these circumstances, methodological pragmatism is attractive because it permits interpreters to use an entire array of tools as they see fit. So the story goes, this freedom leads to “flexibility.” But it can also lead to a number of pathologies in interpretation that should be avoided. In this final post of the series, I outline these pathologies, sketch a path forward, and then highlight a recent example case (Walsh) from the Ontario Court of Appeal that demonstrates why methodological pragmatism unleashes judges to an unacceptable degree. The point here is that interpretation is designed to determine what the legislature meant when it enacted words. Purpose is important in ascertaining that meaning, but ascertaining purpose is not the point of interpretation. This leads to an approach that prefers some ordering among the relevant interpretive tools (for want of a better phrase), rather than a flexible doctrinal standard motivated by methodological pragmatism.

The Pathologies of Pragmatism

By now, and as I have outlined above and in my previous posts, Canada’s approach to statutory interpretation is oddly enigmatic. On one hand, everyone (seems) to agree on the goal of the enterprise: when courts interpret statutes, they are seeking to discover what Parliament intended when it enacted a particular provision or provisions. Putting aside thorny issues of what “legislative intent” might mean (and see here Richard Ekins’ important work), in practical terms, we are seeking to discover the legal meaning and effect of language enacted by Parliament; we are, put differently, seeking to discover what change has been effected in the law (either common law or existing statute law) by Parliament’s intervention (see Justice Miller’s opinion in Walsh, at para 134).

When a law is adopted, one can speak of ends and means, and it’s this framework that guides the discussion to follow. It would be strangely anodyne to claim that Parliament speaks for no reason when it legislates. We presume, in fact, that every word enacted by Parliament means something (represented in canons like the presumption against surplusage, see also Sullivan at 187). And so it only makes sense to take account of a particular provision’s purpose when considering interpretation. Those are the ends for which Parliament strove when adopting the legislation. Selecting the proper ends of interpretation—at the proper level of abstraction, bearing on the actual text under consideration—is an integral part of interpretation. To avoid a strictly literal approach, text must be read in this context.

But, importantly, this is not the end. What about means? In some ways, and as I will show through the example case, means are the real subject of debate in statutory interpretation. Parliament can achieve an objective in many different ways. In general, Parliament can enact broad, sweeping, mandatory language that covers off a whole host of conduct (within constitutional limits). It could leave it at that. Or it could enact permissive exceptions to general mandatory language. It can enact hard-and-fast rules or flexible standards. Administrative schemes can delegate power to “independent” actors to promulgate its own rules. The point here is that Parliament can decide to pursue a particular, limited purpose, through limited or broad means. This is Parliament’s choice, not the court’s.

While free-wheeling pragmatism can lead to all sorts of pathologies, I want to focus here on the relationship between ends and means, between purpose and text. Pragmatism can distort the proper ascertainment of ends and means. In some cases, the problem will be that the court, without any doctrinal guidance, chooses a purpose at an unacceptably high level of abstraction (see, for example, the debate in Telus v Wellman, and Hillier), perhaps even to achieve some pre-ordained result. The courts can do so because, if one simply follows Rizzo, there is no requirement that a judge seek textual evidence for the establishment of a purpose. Yet we know that, as a descriptive matter, it is most common that purpose is sourced in text (see Sullivan, at 193): an interpreter can usually glean the purpose of the legislation, not from legislative history, subsequent legislative enactments, or even the judge’s own imagination, but from the text itself.

This descriptive state of affairs is normatively desirable for two reasons. First, the point of interpretation is not to establish the purpose or mischief the legislature was intending to solve when it legislated (despite Heydon’s Case). The point is to discover the intent of the legislature as represented in the meaning of the words it used. The words are the law. Purpose assists us in determining the meaning of those words, but it cannot be permitted to dominate the actual goal of the enterprise. A pragmatist approach permits, at least in some cases, for that domination to exist: if purpose is better evidence of intention than text, in some cases, then it can be permitted to override text. But this undermines the point of interpretation.  

Secondly, for all we might say about legislative intentions, the best practical evidence of intention is what has been reduced to paper, read reasonably, fairly, and in context. Since statutory interpretation is not a theoretical exercise but a problem solving-one, the practicality of doctrine is central. For this reason, purpose can best assist us when it is related and grounded in text; when the text can bear the meaning that the purpose suggests the words should carry. To the extent pragmatism suggests something else, it is undesirable.

  Sometimes, however, the problem will lie in the means; while the relevant purpose may be common ground between the parties, there may be a dispute over the meaning of language used to achieve those ends. Such disputes tend to focus on, for example, the choice between ordinary and technical meanings, the role of particular canons of interpretation, and (importantly for our purposes) the relationship between the properly-scoped purpose and the language under interpretation. It is the job of the interpreter to work among these tools synthetically, while not replacing the means Parliament chose to accomplish whatever purpose it set out to accomplish. But with pragmatism, no matter the means chosen by Parliament, there is always the chance that the court can dream up different means (read: words) to accomplish an agreed-upon purpose. Often, these dreams begin with a seemingly benign observation: for example, a court might simply conclude that it cannot be the case that a posited interpretation is the meaning of the words, because it would ineffectually achieve some purpose.

These pathologies can work together in interesting ways. For example, an expansive purpose can cause distortions as the means selection stage of the analysis; a court entranced by a highly abstract purpose could similarly expand the means chosen by the legislature to achieve those means. But even in absence of a mistake at the sourcing stage, courts can simply think that Parliament messed up; that it failed to achieve the purpose it set out to achieve because the means it chose are insufficient, in the court’s eyes.

A Way Forward

When constructing doctrine, at least two considerations to keep in mind pertain to flexibility and formality, for want of better words. Flexibility is not an inherently good or bad thing. Being flexible can permit a court to use a host of different tools to resolve disputes before it, disputes that sometimes cannot be reduced to a formula. Too much flexibility, however, and the judicial reasoning process can be hidden by five-part factorial tests and general bromides. Ideally, one wants to strike a balance between formal limits on how courts must reason, with some built-in flexibility to permit courts some room to react to different interpretive challenges.

The point I have made throughout this series is that Rizzo—to the extent it is followed for what is says—is pragmatic, methodologically. Whatever the benefits of pragmatism, such a model fails to establish any real sequencing of interpretive tools; it does not describe the relationship between the interpretive tools; and leaves to the judge’s discretion the proper tools to choose. While subsequent Supreme Court cases might have hemmed in this pragmatic free-wheeling, they have not gone far enough to clarify the interpretive task.

The starting point for a way forward might begin with the argument that there must be some reasons, ex ante, why we should prefer certain interpretive tools to others. This starting point is informed by a great article written by Justice David Stratas, and his Law Clerk, David Williams. As I wrote here:

The piece offers an interesting and well-reasoned way of ordering tools of interpretation. For Stratas & Williams,  there are certain “green light” “yellow light” and “red light” tools in statutory interpretation. Green light tools include text and context, as well as purpose when it is sourced in text. Yellow light tools are ones that must be used with caution—for example, legislative history and social science evidence. Red light tools are ones that should never be used—for example, personal policy preferences.

In my view, this sort of approach balances formalism and flexibility in interpretation. For the reasons I stated above, the legislative text is really the anchor for interpretation (this is distinct from another argument, often made, that we “start with the text” in interpretation). That is, the text is the best evidence we have of intention, often because it contains within it the relevant purpose that should guide us in discovering the meaning of the text. For this reason, legislative text is a green light consideration. Purpose is also a green-light consideration, but this is because it is sourced in text; if it was not, purpose would be misused in a way that might only be recognizable to a methodological pragmatist. Other tools of interpretation, such as legislative history and social science evidence, can be probative in limited circumstances.

The key innovation here is the Stratas & Williams approach does not rule out so-called “external sources” of meaning, but it does structure the use of various tools for interpretation. For example, the approach does not raise a categorical bar to the consideration of legislative history. But it does make some ex ante prediction about the value of various tools, reasoning for example that purpose is most relevant when it is sourced in text.

This is an immediate improvement over the pragmatist methodology, at least when it comes to my core area of concern, the relationship between purpose and text. In the pragmatist model, purpose can be erroneously sourced and then used to expand the means chosen by the legislature; in other words, it can be used to override the language chosen by the legislature. Under the Stratas & Williams model, such a situation is impossible. Any purpose that is helpful and relevant to the interpretive task will be contained within the language Parliament chose, even if that language is limited, imperfect, or unclear.

An Example Case: Walsh

Much of this can be explained by a recent case, Walsh, at the Ontario Court of Appeal. While Walsh is a very interesting case for many reasons, I want to focus here on a key difference between the majority decision of Gillese JA and the dissent of Miller JA. Gillese JA seems to implicitly adopt a pragmatic approach, arguably making purpose rather than text the anchor of interpretation—presumably because the case called for it. Miller JA, instead, makes text the anchor of interpretation. The difference is subtle, but immensely important, because each opinion takes a different view of the “means” chosen by Parliament.

At issue in Walsh was s.162.1(1) and (2) of the Criminal Code. Section 162.1(1), in short, “makes it an offence for a person to knowingly disseminate an ‘intimate image’ of a person without their consent” [61]. An “intimate image” is defined by s.162.1(2), and relates to a “visual recording of a person made by any means including a photographic, film or video recording.”

Stripping the dispute down to brass tacks, the issue in this case was whether a FaceTime call that displays certain explicit content could constitute a recording. The problem, of course, is that FaceTime video calls cannot be conventionally saved and reproduced, like a photo (putting aside, for a moment, the possibility of recording a FaceTime video call). The Crown, at trial, argued that the language of the provisions are written broadly, and must be read “in the context of the harm that s.162.1 was enacted to address: sexual exploitation committed through technology, including cyberbullying and revenge porn” [23, 55]. For the Crown, the answer was found by reasoning from this general “mischief” that the statute was designed to address: the harm would still exist even despite “the recipient’s inability to further share or preserve the moment…” [23]. The defense, on the other hand reasoned from the ordinary meaning of the word “recording,” concluding that “recording” alludes to the “creation of an image that can be stored, viewed later, and reproduced” [57].

Gillese JA for the majority agreed with the Crown’s argument. She listed five reasons for her agreement, but one is particularly relevant on the issue of the relationship between text and purpose. Gillese JA writes, at paras 68 and 70:

[68] Fourth, restricting the meaning of “recording” to outdated technology—by requiring that it be capable of reproduction—would fail to respond to the ways in which modern technology permits sexual exploitation through the non-consensual sharing of intimate images. In so doing, it would undermine the objects of s.162.1 and the intention of Parliament in enacting it.

[…]

[70] …Giving “visual recording” a broad and inclusive interpretation best accords with the objects of s.162.1 and Parliament’s intention in enacting it.

As we will see, this is precisely backwards.

Miller JA’s dissent should be read in whole. It is a masterclass in statutory interpretation, and it is particularly representative of the approach I favour. But most importantly, Miller JA outlines why the majority’s approach demonstrates a means problem, as described above. For Miller JA, there is no purpose-sourcing problem here, since, as he says, there is common ground about the mischief that these provisions were designed to address [179]. However, for Miller JA, a proper application of the various tools of interpretation counselled an approach that did not rewrite the terms of the statute; the means chosen by the legislature. This approach is supported by a number of considerations. First, as Miller JA says, the term “recording” must be given its ordinary meaning. This is the going-in presumption, absent good reasons otherwise. But for Miller JA, the Crown offered no objective support for its assumption that the term “recording” must encompass the FaceTime video at issue. While dictionary meaning and ordinary meaning are two different things, dictionary meaning can shed light on ordinary meaning, and Miller JA noted that there was no instance of the term “recording” being used to describe a “visual display created by any means” [159].

This might have been enough, but the Crown offered another argument: that the term “recording” must be understood as encompassing new forms of technology [162]. Of course, because of the original meaning canon, it could not be said that any linguistic drift in the term “recording” is legally relevant in this case [166]. However, it is a common application of the original meaning rule that where words are written in a broad and dynamic manner, they could capture phenomena not known to drafters at the time of enactment. For Miller JA, however, this argument failed when it comes to the word “recording.” For him, FaceTime was clearly a phenomenon that existed at the time these provisions were drafted, and in fact, the context of the provisions indicated that Parliament had actually distinguished, in other places, recordings versus “visually observing a person…” [174-176]. The term “recording,” then must rely on the concept of reproducibility, as distinguished from other sorts of displays that cannot be saved and reproduced. This latter category of displays was known by Parliament when it crafted these provisions, but it is conspicuously absent from the provisions themselves.

Miller JA, having disposed of these arguments, then clearly contrasts his approach to Gillese JA’s:

[171] What the Crown is left with is the proposition that a reauthoring of the provision would better achieve s.162.1’s purpose….But where Parliament chooses specific means to achieve its ends, the court is not permitted to choose different means any more than it would be permitted to choose different ends. The interpretive question is not what best promotes the section’s purpose, such that courts can modify the text to best bring about that result, but rather how Parliament chose to promote its purpose

[172] …Although the Crown’s argument is framed in ascertaining the conventional, ordinary meaning of language, it is actually an argument about what meaning ought to be imposed on s.162.1, so as to best achieve the purposes of this section.

These paragraphs are remarkable because they clearly set up the difference between Gillese JA’s approach and Miller JA’s approach; the difference between a methodologically pragmatic approach, and an approach that roots ends in means, purpose in text. For Gillese JA, one of her five reasons for accepting the Crown argument pertained to the fact that the defense’s offered interpretation would fail to achieve the agreed-upon purpose of the provisions. This sort of reasoning is only possible under a pragmatic approach, which permits courts to prioritize different interpretive tools as they see fit. The result is a Holy Trinity abomination: where purpose is the anchor for interpretation, and the text is massaged to achieve that purpose, in the court’s view.

Miller JA’s approach is better, if one follows the argument in this post. His approach clearly sees text as an interpretive “tool” that is prior to all the others, in the sense that it is (1) what the legislature enacted to achieve some goal (2) it, practically, is the best evidence we have of what the purpose of the legislation is. Under this formulation, it is not up to the courts to decide whether better means exist to achieve the purpose of the legislation. If this were the case, the point of interpretation would be to identify the meaning of purpose, rather than the meaning of language as evidence of intention. Miller JA explicitly assigns more weight to the text in cabining the purposive analysis.

The Walsh case illustrates the problem that pragmatism has created. While all agree on the point of interpretation, that agreement tends to break down when we begin to apply the tools we have to determine the meaning of the text. Methodological pragmatism offers no hope for solving this problem, because it fails to take a stand on which tools are best. The Stratas & Williams approach, and the approach offered by Miller JA in Walsh, envisions some ranking of the interpretive tools, with text playing a notable role. This approach is better. It moves us away from the endless flexibility of pragmatism, while still leaving the judge as the interpreter of the law.

The Top Statutory Interpretation Cases of 2020

A banner year for interpretation

Introduction

To say that one believes in “purposive interpretation” has been the calling card of Canadian legal scholars for some time. Saying this, as some do, is radically incomplete. That is because competing schools of thought also look to purpose. Textualists, for example, look to the context in which words are used, as well as the purpose evident in those words (Scalia & Garner, at 20). To say that one is a purposivist might as well mean nothing, because everyone—even textualists—“routinely take[] purpose into account…” (Scalia & Garner, at 20).

Far from just being a lazy turn of phrase, though, the routine deployment of the term “purposivism” as a distinct school of thought blocks us from a clearer conversation about what should matter in statutory interpretation. For example, the real division between textualists and others is how purpose is sourced in statutory interpretation: textualists are wary of importing some abstract purpose to subvert a “close reading” of the text (see Scalia & Garner, at 20; see also the opinion of Côté Jin West Fraser), while others might source purpose differently. Saying that one is a “purposivist” also does not answer an important question: which purpose should count more in interpretation, since statutes often pursue multiple purposes at different levels of abstraction? (see, for an example of this, Rafilovich). These are real interpretive questions that are only now receiving any sort of sustained attention in the case law.

I should not hide my priors here. I too think that purpose is a relevant consideration in statutory interpretation, because it assists in the task of reading text to mean all it fairly encompasses. But purpose can be abused: indeed, “[t]he most destructive (and most alluring) feature of purposivism is its manipulability” (Scalia & Garner, 20). Because purposes can be stated in all sorts of ways, it is up to the judge, in many cases, to choose the most appropriate purpose to assist in interpreting the text. Sometimes, purpose can subvert text—which, of course, is problematic if the purpose is not sourced in text (McLachlin CJC’s opinion in West Fraser is a classic example of this).  Put simply: purpose informs text, it does not supplant it (Placer Dome, at para 23).

For that reason, we must come to sound and principled ways of sourcing purpose, rather than simply stating that we look to purpose. It is this theme that defined, in my view, the task for judicial interpreters in 2020. The following three cases are, to my mind, exemplars of dealing with some of these deeper questions in statutory interpretation. Rather than simply reciting the Rizzo & Rizzo formula and taking an “anything goes” approach to interpretation, these cases delve deeper and answer some knotty interpretive questions in a way that furthers a discussion about statutory interpretation in Canada—particularly with reference to the so-called “purposive” approach. Because these cases start a conversation on these issues (and because I happen to agree with the methodology employed by the judges writing the lead opinions in each case), these are the top statutory interpretation cases of 2020, in no particular order:

Michel v Graydon, 2020 SCC 24

In this case, the Supreme Court of Canada dealt with the question whether it is “possible to vary a child support order under the [Family Law Act] after the order has expired, and after the child support beneficiary ceases to be a “child” as defined in the [Family Law Act]” [2]. This seemingly technical question of family law, however, gave rise to all sorts of interpretive problems: the role of social science evidence in statutory interpretation, the problem of unbridled consequential analysis in statutory interpretation, and the problem raised when judges invoke both “liberal” and “purposive interpretation” in the same breath.

For Brown J, the answer to question in the case was found relatively confined to the legislative text and scheme. Starting from the text of the provision, Brown J concluded that the relevant text of the Family Law Act “creates an avenue for courts to retroactively change any child support order, irrespective of the beneficiary’s dependent status and irrespective of whether the order is extant at the time of the application” [20]. This was because of the placement of the relevant statutory scheme. Among other things, s.152 contained no textual restriction on the courts—for example, s.152(1) “contains no reference to the defined term ‘child’ that might serve to qualify the authority of a court to vary child support” [22]. The scheme of the Family Law Act supported this conclusion [23].

For Brown J, this textual conclusion was basically the end of the story (see also schematic considerations at paras 24, 26, 27). Importantly, though, Brown J’s textual conclusion was supported by a properly-scoped purpose. Brown J identified that one of the dominant features of the Family Law Act—given the statute it replaced—was a desire to “expan[d] on the circumstances under which a court may vary a child support order” [28]. Read in light with the text, the result was clear.

Martin J concurred in the result, but conducted a policy analysis to support her concurrence. In Martin J’s view, child support cases called for (that old standard) of a “fair, large, and liberal construction” [40]. For Martin J, this sort of construction required a “contextual and purposive reading of s.152” that looks to “its wider legislative purposes, societal implications, and actual impacts” [40] in a way that “takes into account the policies and values of contemporary Canadian society” [70]. Martin J concluded that “a jurisdictional bar preventing these cases from being heard not only rests in unsound legal foundations, it is inconsistent with the bedrock principles underlying child support and contributes to systemic inequalities” [40].

I agree with both judges that the text and context in this case supports this reading of s.152. But while both judges agreed on the ultimate result, the method they used to reach the result differs in important ways. While Brown J focuses largely on a contextual reading, Martin J incorporates other information, statistics, and an evaluation of the consequences of the interpretation to the result. As I will note, in this case, these approaches do not lead to dramatically different conclusions, because the tools all pointed to a certain result: text, social science, context, consequences. But where text and such other factors conflict, Martin J’s opinion raise a number of problems, in my view.

There are three comments to make about this case, and why it is important. First, Brown J’s opinion avoids the pitfalls that might be associated with external aids to interpretation.  Specifically, Martin J looked to various social science data related to poverty, family relationships, and marginalization. These are important topics, and in this case, the evidence supported the interpretation that Brown J undertook on the text. But the question arises: what to do when current social science evidence contradicts an analysis undertaken on the text? Put differently, if the text points in one direction, and that direction exacerbates problematic trends in social science evidence, which governs?

It is one thing to suggest that where the text is ambiguous, an interpretation which solves the supposed “mischief” the statute was aimed to solve should be preferred. One could make a case for that argument. But where the text and the evidence are directly contradictory, courts must follow the text because that is what the legislature enacted. This may sometimes lead to interpretations that do not make sense to contemporary society, or are unjust in face of empirical evidence, because the text was enacted at a particular time. But this is simply a function of the task of statutory interpretation, which is to determine what the legislature meant at the time of enactment (as I note below, this itself is a rule of interpretation). It must be remembered that external aids can be used to assist in interpreting the text. They cannot be used to subvert it. Martin J’s approach could lead to that result—though, as I note, the problem does not arise in this case because the text and evidence pointed to the same interpretive result.

Secondly, both opinions could be read as cabining the role of pure policy or consequential analysis in statutory interpretation, which could be an invitation for results-oriented reasoning. It is true that evaluating the competing consequences of interpretive options is a fair part of statutory interpretation (see Sullivan at 212 et seq; see also Atlas Tube, at para 10; Williams, at para 52). But there is a caveat: consequences cannot be used to dispense with the written text. This most arises in the context of the absurdity canon, where absurd interpretations of statutes are to be avoided. However, an overapplication of the absurdity canon can lead to many “false-positives” where consequences are labelled absurd in the judge’s opinion, even if those consequences are arguably a product of the text. This undermines the legislature’s role in specifying certain words. Instead, consequences can only be used to determine which of various “rival interpretations” are most consistent with the text, context, and purpose of the statute (see Williams, at para 10). In this way, consequences are not used to determine which interpretation is just or unjust in an abstract sense, but which interpretations are most consistent with the statute’s text, context, and purpose.

Brown J clearly used this sort of justified consequential analysis in his opinion. In connecting his preferred interpretation to the properly-scoped purpose of s.152(1), it was clear that his interpretation furthered that purpose. This is a proper use of consequences consistent with the text as the dominant driver of purpose.

On the other hand, Martin J’s opinion could be read in two ways: one undesirable, one not. First, it could be read as endorsing a wide-ranging assessment of consequences, at a high level of abstraction (for example, justifying her consequential analysis with reference to the need to abolish systemic inequalities: see paras 40, 70, 101).  This might be a very good thing in the abstract, but not all legislation is designed to achieve such lofty goals. If interpreted in such a way to reach a result the statute does not reach, statutes can be conceived as addressing or solving every societal problem, and therefore as resolving every unjust consequence—and this could lead to overextensions of the text beyond its ordinary meaning (see Max Radin, “Statutory Interpretation” 43 Harv L Rev 863, 876 (1930)).  This reading of Martin J’s opinion is not desirable, for that reason.

Another reading of Martin J’s opinion is that she roots her consequential analysis in the purpose of the statute as she sees it. For example, Martin J notes that her approach interprets s.152 with its “underlying purposes in mind” which includes the best interests of the child [76]. Martin J also notes that her interpretation favours access to justice, under-inclusivity, and socio-economic equality [72]. These factors may or may not be rooted to the statute under consideration.

If Martin J’s opinion is rooted in the recognized purpose of the best interests of the children, one can make the case that her opinion is justified as Brown J’s is. However, if read more broadly, Martin J could fairly be seen as addressing issues or consequences that may not fall within the consideration of the text. In the circumstances, I prefer to read Martin J’s opinion as consonant with Brown J’s. If that is done, there is no warrant to look to consequences that fall outside the purpose of the statute. But note: much will depend, as I note below, on how the purpose of a statutory provision is pitched.

Finally, Brown J’s opinion is tighter than Martin J’s in the sense that it does not raise conflicts between statutory interpretation principles. Martin J’s opinion arguably does so in two ways. First, it is well-known (despite the controversy of this practice in constitutional interpretation) that statutes must generally be interpreted as they would have been the day after the statute was passed (Perka, at 264-5). While there is some nuance on this point (see Sullivan, at 116-117), words cannot change legal meaning over time—but note that broad, open-textured terms can be flexibly applied to new conditions if the words can bear that meaning (see here). The key is that words can only cover off the situations that they can fairly encompass. But the injunction—repeated throughout Martin J’s opinion—that statutes must be interpreted in light of the “policies and values of contemporary Canadian society” [72] at least facially conflicts with the original meaning canon. To Martin J’s defense, she is not the first to say this in the context of family law and child support (see Chartier, at paras 19, 21). But nonetheless, the court cannot have it both ways, and Martin J’s opinion cannot be taken to mean that the legal meaning of texts must be interpreted to always be consistent with contemporary Canadian society.

At best, it might be said that Martin J’s opinion in this respect permits the taking into account of contemporary considerations where the text clearly allows for such considerations, or perhaps where the text is ambiguous and one interpretation would best fit modern circumstances in a practical sense. But these modern circumstances cannot be shoehorned into every interpretation.

Secondly, there is a conflict in Martin J’s opinion, in a theoretical sense, between her invocation of a “fair, large and liberal” interpretation (see paras 58, 71) and her invocation of a “purposive” interpretation  (see para 71). As Karl Llewellyn pointed out long ago, it is not unheard of for tools of interpretation to conflict. But as much as possible, judges should not invite such conflicts, and I fear Martin J did this in her opinion by conflating liberal interpretation with purposive interpretation. As I have written before, these things are not the same—in fact they are opposites. The Interpretation Act does instruct a “large and liberal” interpretation, but only as the objects of a statute permit. The Supreme Court continues to insist on an approach to statutory interpretation that uses text to ground the selection of purpose (see here). As such, text and purpose read synthetically governs—not some judge-made conception of what constitutes a “large and liberal interpretation.” This statement cannot be used to overshoot the purposes of a statute, properly scoped.

Perhaps in this case the purposes permit a large and liberal interpretation, in which case Martin J can use both of these tools interchangeably. As I said, the problem isn’t this case specifically, but what would happen if Martin J’s approach is used in the general run of cases. But it is far from clear that purposive and generous interpretation will always–or even often–lead to similar results. More likely, purpose will limit the ways in which text can be read—it will not liberate the judge to take into account any policy considerations she wishes.

Michel v Graydon raises all sorts of interesting issues. But taking Brown J’s opinion on its own terms, it is a clinic in how to clearly interpret a statute in light of its properly-scoped purpose. While Martin J’s opinion could also be read in this way, it could be read to permit a more free-flowing policy analysis that subverts legislative language. In this sense, Martin J’s opinion should be affixed with a “caution” label.

Entertainment Software Association, 2020 FCA 100

ESA will stand, I think, for some time as the definitive statement in the Federal Courts on how to conduct statutory interpretation, and the role of international law in that endeavour.

In this case, the facts of which I summarized here,  the Copyright Board offered an interpretation of the Copyright Modernization Act that arguably placed extraneous materials ahead of the governing text. Here is what I wrote about the Board’s conduct at the time:

The Board’s chosen materials for the interpretive exercise were stated, according to the Court, at a high level of generality (see paras 53-54). For example, the Board focused on the preamble to the Copyright Modernization Act to divine a rather abstract interpretation that supported its view on international law (paras 53-54). It also invoked government statements, but the Court rightly noted that these statements construed s.2.4(1.1) as a “narrow, limited-purpose provision” [56], not as an all-encompassing provision that permitted the collection of tariffs in both instances. The use of these materials was used by the Board to herald a different, broader interpretation than what the text and context of the provision indicated. 

The Court rebuffed the Board’s effort in this regard. By noting that the provision under interpretation was a “narrow, limited-purpose provision,” the Court rejected attempts by the Board to drive the interpretation higher than the text can bear.  This is a worthy affirmation of the importance of text in the interpretive process, and a warning about the malleability of purposive interpretation.

Why is this opinion so important? It makes a now oft-repeated point that purposive interpretation is not conducted “at large.” That is, it matters how judges state the purposes they hope to use in the interpretive task. As the Supreme Court noted in Telus v Wellman, courts cannot use abstract purposes to “distort the actual words of the statute” (see Telus, at para 79). This counts as an endorsement of the traditional separation of powers, under which “…the responsibility for setting policy in a parliamentary democracy rests with the legislature, not the courts” [79].

ESA is important because it implements what the Supreme Court has now repeated in Telus, Rafilovich, and other cases. It is now clear law that purposes cannot be used to subvert text; that text is the starting point in legislative interpretation, and that in sourcing purpose, text confines the scope of the exercise. In my view, ESA (expertly written by Justice Stratas) makes the clearest case yet for a sort of text-driven purposivism in the context of Canadian statutory interpretation.

Canada v Kattenburg, 2020 FCA 164

One underlying theme of much of what I have written thus far is a worry about results-oriented reasoning in statutory interpretation. To some, this might not be a risk at all. Or it might be a desirable feature: after all, if all law should simply be an adjunct of politics, then the policy preferences of judges are fair game. Of course, I readily admit that no legal system can reduce the risk of subjective policy-driven interpretation to 0; nor should it. But the Rule of Law, at its most basic, means that the law governs everyone—including judges. Part of the law judges must apply are the rules of statutory interpretation. Those rules are designed not to vindicate what the “just” result is in the abstract, what is “just” at international law (except where international law and domestic legislation meet in defined ways), or even what is “just” to the judge at equity or common law–except, of course, when statutes implicate common law rules. Statutory interpretation is a task that requires determining what the legislature thought was just to enact. As such, the rules of interpretation are guided towards that goal, and are necessarily designed to limit or exclude the preferences of judges or others, even if we reach that goal only imperfectly.

This important theoretical point was made in relation to the ascertainment of legislative purpose and international law in Kattenburg by Stratas JA. In Kattenburg, the underlying substantive issue was simple and narrow, as I wrote in my post on the case:

The Canadian Food Inspection Agency decided that certain wine imported to Canada from the West Bank are “products of Israel” (see the Federal Court’s decision in 2019 FC 1003 at para 3). The judicial review, among other issues, concerned whether the wine could be labelled as “products of Israel.” That’s it. Under ordinary administrative law principles, the court will assess whether the decision of the CFIA is reasonable. A typical legal task.

However, on the intervention motions in Kattenburg, Stratas JA noted that some intervenors ttempted to further bootstrap the record with “hyperlinks to find reports, opinions, news articles and informal articles to buttress their claims about the content of international law and the illegality of Israel’s occupation of the West Bank” (Kattenburg, at para 32). Stratas JA rejected such efforts.

Stratas JA’s rejection of these intervenors, and his strong words in denouncing them, raised the ire of some on law twitter. But anything worth doing won’t be easy, and Stratas JA said what needed to be said, particular when he noted that, with respect to the intervenors “[s]o much of their loose policy talk, untethered to proven facts and settled doctrine, can seep into reasons for judgment, leading to inaccuracies with real-life consequences” (Kattenburg, at para 44). 

 There’s no denying Stratas JA is pointing to an important methodological problem that is deserving of our attention. One way that purposes can be misstated, or used to subvert clear text, is by advancing broad understandings of international law to expand the purpose. As I’ve noted before, it is true that “international law can…be relevant to the interpretation of Canadian law where it is incorporated in domestic law explicitly, or where there is some ambiguity” (see here). But in many cases, international law will simply not be relevant to the interpretation of legislative texts, or the ascertainment of legislative purposes.

The attempt in Kattenburg to cast the legislative purpose to encompass some statement—any statement—on the legality of Israel’s conduct in the Middle East is a classic end-run around legislative text. While some of the intervenors may have wanted the Court to interpret the legislation in a particular way to encompass substantive policy goals encompassed in international law not only runs afoul of fundamental principle—international law only enters the task in defined, narrow ways—but it is contrary to precedent (see Vavilov, at para 121 and the litany of Federal Court of Appeal and Supreme Court cases on this point). Such efforts should be rejected.

Conclusion

In many ways, the three cases I have chosen as important for interpretation in 2020 are all representative of a broader theme of which lawyers should be aware. That is, there is much more happening behind the curtains in Canadian statutory interpretation than might appear at first blush. “Purposive interpretation” is not the end of the story. What matters is how we source purpose, the sources we assess to assist the interpretive task, and the role of text in grounding the interpretive process. These cases all come to defensible conclusions on these questions. The insights of these cases can be distilled into a few key propositions:

  1. Purpose must be sourced in relation to the relevant text under consideration. In this way, we are interpreting text as the legislature enacted it, and we are not using purpose to subvert that authentic reading of the text.
  2. There are reasons to be worried about consequential analysis, to the extent it could permit an expansion of legislative purposes beyond text.
  3. There are reasons to be worried about international law, to the extent it could permit the expansion of legislative purposes beyond text.

All for the better.

A Happy New Year for interpretive nerds!

“Purposive” Does Not Equal “Generous”: The Interpretation Act

It is often said in Canada that statutes must be interpreted “purposively” and “generously.” Many cite the federal Interpretation Act’s s.12, which apparently mandates this marriage between purposive and generous interpretation:

12 Every enactment is deemed remedial, and shall be given such fair, large and liberal construction and interpretation as best ensures the attainment of its objects.

The Supreme Court has also accepted this general principle in the context of the judge-made rule that benefits-conferring legislation should be interpreted liberally (see Rizzo, and more recently, Michel v Graydon).

Putting aside the judge-made rule itself, which raises similar but somewhat separate questions, I write today to make a simple point: this injunction in the Interpretation Act cannot be read so as to render purposive interpretation the same as a “generous” interpretation. Doing so could violate the Supreme Court’s statutory interpretation jurisprudence, which promotes an authentic determination of purpose according to the legislative language under consideration (see my post on Rafilovich). Indeed, as is clear in the constitutional context, purposive interpretation will often lead to the narrowing of a right, rather than a generous interpretation of that right (see, for a recent example, R v Poulin). Similarly, a purposive interpretation in statute law will lead to a narrowing of the meaning of a particular statutory provision to its purposes. Those purposes will best be reflected in text (see Sullivan, at 193; see also here). For that reason, the Interpretation Act can only mandate a simple canon of interpretation: “The words of a governing text are of paramount concern, and what they convey, in their context is what the text means” (Scalia & Garner, at 56). Words should be interpreted fairly but only insofar as purpose reflected in text dictates.

One cannot read the Interpretation Act to mandate a generous interpretation over a purposive one. The text of the provision in question says that “fair, large and liberal construction” must be rendered in a way that “best ensures the attainment of the [enactment’s] objects.” This means that purpose is the anchor for a “generous” interpretation within those purposes. Put differently, we should read words to mean all that they can fairly mean, but we cannot use some injunction of “generosity” to supplant the words or the purposes they reflect.

Prioritizing “generosity” over the natural reading of text in its context would lead to all sorts of practical problems. For one, it is difficult to determine what a “generous” interpretation of a statute would mean in practical terms (see Scalia & Garner, at 365). Does this simply mean that “[a]ny doubt arising from difficulties of language should be resolved in favour of the claimant”? (see Rizzo, at para 36). This could be defensible. But the risk is that using the language of “generosity” could invite judges to expand the scope of language and purpose to suit policy outcomes/parties they prefer.

We should be careful of this language for this reason. More importantly, if “generosity” means that the legitimately-sourced purpose of legislation can be abrogated, the language is quite inconsistent with the Supreme Court’s actual approach to interpretation in recent cases (see Telus v Wellman and Rafilovich).

Rather, the reading of the relevant section of the Interpretation Act must be taken to conform with the Supreme Court’s governing approach to statutory interpretation.  In this sense, the “fair, large, and liberal” interpretive approach mandated by the Interpretation Act might be explained by contrasting it to an old form of interpretation that virtually no one adopts now: strict constructionism. Strict constructionism, most commonly adopted in the adage that “statutes in derogation of the common law were to be strictly construed” (Scalia & Garner, at 365) was unjustified because it violated the “fair meaning rule”; the text, in its context, must be interpreted fairly. No one today—not even textualists—are strict constructionists, because everyone accepts the idea that text must be interpreted fairly. If the Interpretation Act is a response to strict constructionism, its language could perhaps be forgiven. But it should be taken no further than the fair-meaning rule, which rests on identifying relevant purposes in text and using those purposes to guide textual interpretation.

An example of a party attempting to use the Interpretation Act is a manner I consider impermissible occurred in Hillier. There, Ms. Hillier relied on the Interpretation Act and the general canon of interpretation that benefits-conferring legislation is to be liberally interpreted. Putting aside this canon (dealt with in Hillier, at para 38), the Interpretation Act was marshalled by Ms. Hillier to suggest that the court should rule in her favour. Stratas JA rejected this erroneous reliance on the Interpretation Act, concluding (at para 39):

[39]  To similar effect is the interpretive rule in section 12 of the Interpretation Act. It provides that “[e]very enactment is deemed remedial, and shall be given such fair, large and liberal construction and interpretation as best ensures the attainment of its objects.” Section 12 is not a licence for courts and administrative decision-makers to substitute a broad legislative purpose for one that is genuinely narrow or to construe legislative words strictly for strictness’ sake—in either case, to bend the legislation away from its authentic meaning. Section 12 instructs courts and administrative decision-makers to interpret provisions to fulfil the purposes they serve, broad or narrow, no more, no less.

This is an accurate description of the function of the Interpretation Act, which finds agreement with the Supreme Court’s statutory interpretation jurisprudence, such as I can discern it. Purpose—usually sourced in text—guides textual interpretation. Purpose and text should be read synthetically together to render a fair meaning of the language at hand. But broad notions of “generosity” or “fairness” should be not be used to supplant the authentic purpose(s) of legislation, derived in text. And “generosity” is not an end-round around the language the legislature actually uses.