THE PROBLEM OF INTERNAL LEGAL PLURALISM

Dworkin’s local difficulty in substantiating his law as integrity thesis highlights some generic problems for the constitutional theory project, and which support the case for internal legal pluralism.

O’Connor J’s majority judgment sets out to clarify the holding in Bakke, it is questionable that it succeeds in this aim. Scalia J argues in his partial dissent (joined by Thomas J), that the compelling state interest identified by the majority is less the ‘“educational benefit” that emanates from the fabled “critical mass” of minority students, but rather Michigan’s interest in maintaining a “pres­tige” law school whose normal admissions standards disproportionately exclude blacks and other minorities’ (ibid at 347). For Scalia J, ‘[i]f that is a compelling state interest, everything is.’ As such, he suggests that the majority judgment ‘seems perversely designed to prolong the controversy’ (ibid at 348) over the constitutionality of racial preferences in public educational institutions.

⁴⁵ Above n 4, at 397 (emphasis added).

⁴⁶ The locus classicus of this position is of course JH Ely, Democracy and Distrust: A Theory of Judicial Review (Cambridge, MA, Harvard University Press, 1980). For a discussion of the poten­tial similarities between Ely and Dworkin, despite their ostensible differences, see N Duxbury, Patterns of American Jurisprudence (Oxford, Oxford University Press, 1995) 293—94.

⁴⁷ See, eg, T Campbell, The Legal Theory of Ethical Positivism (Aldershot, Dartmouth, 1996).

⁴⁸ See, eg, M Cappelletti, ‘The “Mighty Problem” of Judicial Review and the Contribution of Comparative Analysis’ (1980) 53 Southern California Law Review 409 and D Beatty, Constitutional Law in Theory and Practice (Toronto, University of Toronto Press, 1995).

Our discussion of Dworkin’s theoretical difficulties suggests two major prac­tical problems for this project, which I will explore in more detail. First, consti­tutional theorists are on the horns of a dilemma: if their theories are formulated in too general terms, this will invite a choice among competing values, and open up the scope for reasonable judges to disagree about the requirements of the constitution; however, if the theory is so specific as to remove any judicial dis­cretion, then it is nothing more than the bare a priori assertion of the theorist’s ideology. In practice, constitutional theorists retreat from the charge of ad hoc consequentialism by showing how their accounts give us a ‘single more general requirement for which [constitutional rights] could plausibly be said to be standing in.’^[362] My argument is that this generality is unable to satisfy liberal legalism’s criterion of coherence as it does not, and indeed cannot, prevent judges from exercising their law-making capacity.

Secondly, it is by no means clear that even more specific theories would inexorably guide judges to the same result^[363]: this brings us to the problem that constitutional theories must be tested in the context of actual adjudication.

Internal Legal Pluralism and Constitutional Theory

The first generic problem raised above is whether constitutional theory, by operating on the general plane, possesses the wherewithal to settle definitively every constitutional controversy. The challenge here is legal pluralism’s claim that not only are judges law-creating subjects, but that given their individual make-ups, they will create different laws, and so give a range of doctrinal answers to the same constitutional issue: in other words, can constitutional theory overcome the possibility of internal legal pluralism? Here I agree with Dworkin that ‘detail is more illuminating than range,’^[366] and consider the debate between Dworkin and David Dyzenhaus as to whether the constitutional principle of equal concern and respect is offended when legislatures regulate pornography. This debate is particularly interesting, not only because each works within liberal legalist methodology to show what is the correct constitu­tional response, but also because Dyzenhaus’s argument is that his is the better Dworkinian interpretation of constitutional equality.

Dworkin’s position,^[367] applying the tests of fit and value, is that the constitu­tional protection of freedom of expression prevents the criminalisation of private consumption of materials deemed obscene.^[368] He finds that the positive state of US constitutional law is intolerant of restrictions on the content of speech, but also that this is the best justification, since (following Isaiah Berlin) there is no definitive way of choosing the ideal form of society. Rather, we must ‘choose among possible combinations of ideals, a coherent set [to define our individual way of life],’^[369] and so censorship of pornography cannot be justified since freedom of expression is ‘at the core of the choice’^[370] democracies have made with regard to a modus vivendi among incommensurable values. Dyzenhaus objects that Dworkin is not being true to the primary constitutional value of equality.^[371] For Dyzenhaus, Dworkin’s significance lies in his attempt to move away from the priority of negative liberty. On this account, autonomy does not consist solely in a right to non-interference with the private sphere, but in an individual right to equal concern and respect. Making equality ‘the organ­izing principle of liberal political theory’^[372] opens up the possibility of restrain­ing expression when this undermines autonomy by perpetuating inequality. Dyzenhaus argues that this is the case with pornography because it eroticises inequality.

Dyzenhaus takes the example of the Supreme Court of Canada’s decision in R v Butler,^[373] which involved a free speech challenge under the Canadian Charter against a conviction for distributing obscene material.^[374] In upholding the con­viction, Sopinka J measured the protection to be afforded to Butler’s expression in terms of whether it furthered the ‘core values’ of free speech, which included individual self-fulfilment.^[375] In his examination of the ‘core values,’ he empha­sised the centrality of individual self-fulfilment; as pornography was far from this core to the extent that it victimises women and dehumanises their sexual- ity,^[376] it was not a form of expression deserving of constitutional protection.

Dyzenhaus’s criticism of Hercules’s presumed dissent in Butler is that if Dworkin genuinely values equality over liberty, he must drop the commitment to global neutrality which leads to his strong defence of free speech.^[379] However, we can imagine a judge faithfully implementing constitutional equality reason­ing as follows: ‘equal concern and respect means I should not discriminate as to conceptions of the good life; pornography invokes these conceptions, therefore I must not censor it if I am to promote the equal autonomy of all.’ In other cases, the same judge may hold that equality does not compromise ideas of the good life so clearly, eg affirmative action. In other words, one could reason from equality to dissent in Butler without embracing the absolute position on global neutrality associated with negative liberty.

Dyzenhaus’s second criticism is that if Dworkin relies on this argument, then ‘the most this would show is that he has a moral position that differs from that of pro-censorship feminists.’^[380] I agree with Dyzenhaus that this reduces the force of Dworkin’s claims for law as integrity; however, I do not agree that we can instead demonstrate that there is a principled case in favour of censorship which a true Dworkinian would adopt.^[381] It is not that Dworkin’s is the constitutionally wrong application of equality, merely that it is different. That is also the most that can be said for Dyzenhaus’s position as well (in other words, he cannot with­stand his own critique of Dworkin).

To return this discussion to its broader context, we see here two important protagonists in liberal legal theory grappling with the key contemporary issue of the requirements of constitutional equality for adjudication. If Dworkin and Dyzenhaus are unable to reach any coherent consensus, then we might reason­ably ask whether we have any reason to hope for anything better from judges?^[383] However, legal pluralism makes a stronger point here than simply highlighting differential levels of judicial ability: rather, the discussion confirms the generic problems of achieving doctrinal coherence. Dyzenhaus’s critique, just as much as Dworkin’s account, is firmly within the liberal legalist tradition—his ambi­tion is to show how a more general requirement, as an instantiation of equality, if properly applied, can lead judges to correct or better decisions. However, in setting out his argument, he demonstrates that arguments such as his are unable to constrain judges while operating at the level of generality necessary to qual­ify as constitutional theory. Thus, to the extent there is reasonable disagreement over what equality requires, we should expect to find this in the jurisprudence. Moreover, his only alternative, a more particularist argument, is both ineligible as a constitutional theory given its inability to distinguish judicial and political reasoning, and ipso facto, no help in producing coherence. Accordingly, legal pluralism raises serious theoretical doubts over the assumption that constitutional doctrine can be made into a coherent and systematic whole.

Internal Legal Pluralism and Constitutional Practice

Our discussion of Dworkin and Dyzenhaus highlights the difficulty of norma­tive approaches to constitutional law in overcoming the theoretical objection of internal legal pluralism. It is important, though, also to consider to what extent the practice of adjudication confirms (or otherwise) the inadequacy of liberal legalist epistemological assumptions. Accordingly, I now focus on whether the principled search for coherence is the best explanation of what judges are doing in practice. Dworkin, and other constitutional theorists, might here object that what they are concerned with is what judges ought to do, and that the mapping of what is the case is the concern of positivist, not normative, legal theory.^[384] This, though, is too easy an escape: Dworkin’s normative exhortations must surely imply that there is a sufficient basis in practice for his ‘ought’ to be readily turned into an ‘is’. As Frederick Schauer puts it, ‘there is an important empirical question whether there are many actual judges who are in fact like Hercules,’^[385] with important empirical and normative implications if answered in the negative.^[386] Accordingly, it is appropriate for our inquiry to test liberal legalist assumptions against the real world, with real judges subject to real pressures.

The major problem for liberal legalism’s attempt to portray adjudication in systematic terms is that judges often (and sharply)^[387] disagree with each other.^[388] How do we explain this? Is it plausible to maintain, as liberal legalism would have it, that all these judges are engaged in the search for constitutional prin­ciple, with only a handful ever succeeding? My argument is that, following Sampford, this is better explained by emphasising the plural forms of judicial behaviour:

[J]udges do not tend to build systems of any kind... Rather they see themselves using (conveniently ill-defined) judicial techniques to deal with the material presented to them during the arguing of cases. Some of these techniques involve authority- and source-based arguments. Some involve content- and principle-based arguments. Some involve consequentialist and pragmatic arguments. Each technique is available, and judicial discretion exists as much in the choice of technique as the different uses to which it is put.^[389]

In this regard, we can draw on a number of studies which, taking the empirical record as their focus, show the variety of ways in which judges reach their deci­sions. Some retain a doctrinal focus, and highlight the variety of interpretive moves open to judges in constitutional cases.^[390] Philip Bobbitt, for example, distinguishes between (at least) five ‘archetypes’ of constitutional argument from his reading of US constitutional jurisprudence: historical arguments, textual arguments, structural arguments, prudential arguments and doctrinal arguments.⁷⁸ Bobbitt’s argument is not just that these are different theories of interpretation, but that they are representative of the actual argumentative tech­niques employed by different judges at different times when deciding constitu­tional cases.⁷⁹ Thus, some judges can employ different techniques in the same case, or between cases.⁸⁰ Other explanations of judicial discord step outside the internal doctrinal perspective. This is first associated with the legal realists,⁸¹ who posited a number of ‘extra-legal’⁸² influences on the disposal of cases. This position has been updated by studies that have sought to show how judges are either horse-traders,⁸³ or (slightly more loftily) strategists,⁸⁴ seeking to advance their own position in the deeply political process of hammering out doctrinal compromise. Others have argued that adjudication is marked by judicial discretion exercised for ideological reasons,⁸⁵ which, to the extent judges are informed by different beliefs and philosophies, produces a further source of doctrinal conflict.⁸⁶

theories,’ whose claims reside in their consonance with, for example, previous judicial practice. He also identifies a further sub-division between ‘substantive’ theories of interpretation, which speak to the values that adjudication should promote (here he includes Rawlsian liberalism) and ‘formal’ or ‘methodological’ theories of interpretation, such as originalism, which seek an understanding inde­pendent of the values of the interpreter.

⁷⁸ See P Bobbitt, Constitutional Fate: Theory of the Constitution (New York, Oxford University Press, 1982) 7, chs 2—6.

⁷⁹ It should be stressed that Bobbitt himself does not draw the sceptical conclusion advanced in this book—rather, he works within the interpretive question, and regards ‘the next frontier in American constitutional scholarship’ as determining what judges should do when these modalities of interpretation conflict: P Bobbitt, ‘Methods of Constitutional Argument’ (1989) 23 University of British Columbia Law Review 449, 457.

⁸⁰ For an account of the different argumentative techniques employed by the Supreme Court of Canada, see S Peck, ‘An Analytical Framework for the Application of the Canadian Charter of Rights and Freedoms’ (1987) 25 Osgoode Hall Law Journal 1. Peck’s conclusion (ibid at 3) from his survey of Canadian Charter jurisprudence is that ‘judges are led to the conclusions they reach as much by their choice of values and their choice of role, that is their choice whether to adopt judicial activism or restraint, as by the doctrine they invoke.’

⁸¹ See Duxbury, above n 46, at 68.

⁸² Schauer, above n 72, at 620. Schauer lists (ibid at 619) the various ‘[extra-doctrinal] causal influences on judicial decision-making’ identified by the realists as including ‘the judge’s views about the immediate equities of the case at hand, the judge’s less particularistic views about wise public policy, or the judge’s array of philosophical, and policy views, an array that is nowadays called “ideology.”’

⁸³ B Woodward and S Armstrong, The Brethren: Inside the Supreme Court (New York, Avon Books, 1981).

⁸⁴ L Epstein, The Choices Judges Make (Washington, CQ Press, 1998).

⁸⁵ For a representative sample of this scholarship, see JAG Griffith, ‘The Political Constitution’ (1979) 42 Modern Law Review 1 (United Kingdom); A Stone, The Birth of Judicial Politics in France (Oxford, Oxford University Press, 1992) (France); M Mandel, The Charter of Rights and the Legalization of Politics in Canada, rev edn (Toronto, Thompson Educational Press, 1994) (Canada); and MV Tushnet, Red, White and Blue: A Critical Analysis of Constitutional Law (Cambridge, MA, Harvard University Press, 1988) (United States).

⁸⁶ See, eg, J Bakan, B Ryder, D Schneiderman and M Young, ‘Developments in Constitutional Law: the 1993—94 Term’ (1995) 6 Supreme Court Law Review (2d) 67.

It is important not to use the above studies to reduce judging to any one determinant—in particular, it is not necessary to reject liberal legalist notions of coherence by accusing judges of being purely tools in the service of a higher power,^[391] or ideologues acting in bad faith.^[392] Rather, the point is that even if we accept judges are acting in good faith ‘to interpret and deploy legal rules as the argumentative resources... for their decisions,’^[393] the disordering influences enumerated above show that this does not lead to a coherent rule-based model of adjudication.^[394] The strength of these studies is accordingly cumulative, and lies in providing us with a range of factors which sit more plausibly with the complexity of adjudication in practice. Thus, sometimes judges will deploy dif­ferent argumentative styles in the same case (giving contrary results), sometimes their decisions will be affected by tactical concessions in exchanges with their colleagues, and at other times informed by implicit value-choices—or all three at once. The key point here is that identifying this mixture of factors affecting judicial decision-making refutes the idea (necessary for theories such as Dworkin’s to succeed) that judges, as a collectivity, are motivated by the search for overarching constitutional principle. The most we can say is that adjudica­tion represents a series of individual searches for principle, the problem here being, as Sampford puts it, that such systems ‘are created in the mind of an indi­vidual,’ and remain there.^[395]

In the remainder of this section, I test the argument that a more pluralistic approach is a better explanation of constitutional adjudication by focusing on two pairs of Canadian Supreme Court decisions on freedom of expression. These cases are particularly interesting examples of the tendency noted above towards doctrinal incoherence. In the first pair, Keegstra^[396] and Zundel,^[397] the Supreme Court heard appeals against convictions for propagating anti-Semitic views: in the former, by a 4-3 vote, it rejected the appeal, in the latter, again by a 4-3 vote, it struck down the conviction. The second two cases, Irwin Toy^[398] and RJR-Macdonald^[399] concerned challenges to statutes restricting the plaintiff companies’ rights to advertise their products: both cases were again decided by a single vote,^[400] but only one statute passed the constitutional test. In each pair, although there was a certain similarity of subject matter, the impugned provisions were slightly different: Keegstra concerned section 319^[401] of the Criminal Code criminalising ‘hate speech’, while Zundel involved the section 181^[402] offence of spreading false news; the Quebec statute^[403] in Irwin Toy banned advertising aimed at under 13-year-olds, while in RJR- Macdonald, the federal statute^[404] dealt with tobacco products. Accordingly, these cases provide highly relevant tests for our rival epistemologies: do they confirm a view of adjudication based on the search for principle, or one characterised by relative disorder?

For the liberal legalist view to prevail, we need to be convinced not only that there is a coherence between these judgments—that the Charter properly entails such fine distinctions as to uphold section 319(2) of the Criminal Code, but to strike down section 181,^[405] or to ban advertising when it is aimed at children, but not smokers—but also that the judges are attempting to weave these differ­ences into an integrated narrative. In her majority opinion in Zundel, McLachlin J (as she then was) sets out to show how this case can be distin­guished from Keegstra, and so attempts to reconcile both decisions. The main doctrinal issue in these cases was presented as the proximity of the impugned speech to the ‘core values’ of expression (which had been elaborated in Irwin Toy as seeking the truth, participation in the democratic process, and individ­ual self-fulfilment).^[406] McLachlin J’s argument in Zundel was that the activities targeted by section 181 could be seen to lie at the heart of freedom of expression values.^[407] Given this proximity to the core values, there has to be a stringent proportionality examination, which section 181 failed as it was overbroad, catching ‘virtually all controversial statements of apparent fact.’^[408] She distin­guishes this judgment from Keegstra by contrasting the ‘low or negative value’ of the hate speech caught by section 319, with the ‘broad spectrum’ of speech, ‘much of which may be argued to have value’ affected by section 181.^[409] In RJR-Macdonald, the majority (again per McLachlin J) pointed to a number of contrasting factors with Irwin Toy: for example, only in the former had the legislature resorted to use of the criminal law (as opposed to a regulatory regime), which, as it involved ‘a contest between the state and the accused,’ was not entitled to a high degree of deference,^[410] and also that in the former case the ban was total, but only partial in the latter.^[411]

Accordingly, the majorities in both Zundel and RJR-Macdonald seek to jus­tify the difference between them and Keegstra and Irwin Toy respectively in terms of coherence (although, if they are correct, we might wonder if an average of barely over.500 is good enough). However, if we look below the surface of these cases, we can dispute that they represent the elaboration of coherent prin­ciples. In this regard, I will make two related arguments: first, that if we look more closely at what (basically) the same set of judges say in each pair, these cases are not so readily reconcilable, and secondly, that these contradictory judgments embody a range of disordering influences which militate against coherence in adjudication.

The principal argument against coherence here is that the same arguments that win in Zundel and RJR-Macdonald, are essentially the same that lose in Keegstra and Irwin Toy. In other words, there is greater consonance between the respective majorities and dissents of both pairs, than between the actual dis­posals. Thus, in her dissent in Keegstra, McLachlin J regarded section 319(2) with the same opprobrium as she later did section 181 in Zundel, and argued that the former ‘invokes all of the values upon which section 2(b) of the Charter rests,’^[412] leading it to be subjected to a heightened proportionality analysis, which, in her view, it failed. Whereas both Dickson CJC’s opinion in Keegstra, and Cory and lacobucci JJ’s dissent in Zundel placed the speech in question far from the core values of expression. Similarly, in Irwin Toy, McIntyre J’s dissent characterised the Consumer Protection Act as placing the state in the same antagonistic role^[413] later assigned to it with regard to the Tobacco Products Control Act by McLachlin J in RJR-Macdonald. In contrast, the Irwin Toy majority and RJR-Macdonald dissent depicted both statutes as involving the competing rights of different (including vulnerable) groups in society. The argument that what we see in these cases is less an elaboration of principle, but more a clash between two different views on how strictly state restrictions on expression should be viewed (with each winning out once in both pairs of cases), finds further support if we examine the judicial voting record. In Keegstra and Zundel, we can identify two groups of three judges, one in favour of, one against, legislative deference, who maintain their broad positions in each case.^[414] There is also a certain overlap between the judicial positions taken in the commercial expression and hate speech cases.^[415] However, it should also be noted that although these broad positions recur throughout the cases, not all judges can be easily assigned to one camp or the other^[416]: the most obvious example being L’Heureux-Dube J, the only judge to endorse both the Keegstra and Zundel opinions.

My argument is that a judge’s stance between these positions in a particular case reflects his or her resolution of a number of issues which are then expressed as the requirements of the constitutional rule, not the other way around. One such issue is how they conceive of expression. On one view, speech is a linear and transparent medium, whose prime characteristic is its common form, and where the benefits of free expression are seen in terms of quantity; on another, speech is social and relational, which is best understood in terms of its content, and where its benefits are seen in more qualitative terms.^[417] If judges rely on the former view, they are likely to be suspicious of all regulation of speech, whereas if they tend towards the latter, the regulation of some types of speech, judged to be socially harmful,^[418] will not pose the same constitutional danger. In other words, it is judges’ underlying conception of communication which determines their location of an expressive action to the core values of free speech. Related to this is how judges conceive of individuals: as Richard Moon has catalogued, in free speech cases, the court oscillates between seeing people as autonomous agents, to be left alone, and as susceptible to manipulation, thereby needing con­stitutional protection.^[419] How this is resolved leads not only to plural disposals of similar issues, but to the differences in judicial methodology, which ranges between philosophical exploration of the bounds of freedom, and causal inquiries into the harms caused by restrictions on speech.^[420] How judges resolve these issues is of course question-begging, and points us to the role ideological choices,^[421] reflecting the judges’ (however implicit) views of the proper role of the state and its relation to the individual,^[422] play in the exercise of judicial discretion in adjudication.^[423]