Melbourne University Law Review
[In Lange v Australian Broadcasting Corporation, the High Court announced a more limited view of the freedom of political communication, under which the freedom is to be interpreted solely by reference to constitutional text and structure. This article shows that this interpretive method is unsustainable. It does so by considering an undecided doctrinal issue: the formulation of a test for determining the validity of laws affecting political communication. This article shows that two approaches, a single proportionality test and a ‘two-tiered’ test, which applies stricter scrutiny to some cases, have been vying for ascendancy in the High Court. A solution to this problem is suggested. It is argued that the Court should prefer the proportionality test as the freedom of political communication develops but ultimately replace this with a more defined body of rules. However, the central point of this article is that the choice between these standards of review cannot be made simply by reference to the text and structure of the Constitution. Moreover, whichever approach prevails, the development of the standard of review will require reference to values or principles external to the Constitution. Thus, no matter how the Court proceeds, it will have to depart from the Lange interpretive method.]
The High Court’s early decisions on the freedom of political communication were marked by an adventurousness that attracted much controversy. Some argued that the Court had departed from proper methods of interpretation in favour of a view, drawn from extra-constitutional sources, that the protection of free political communication was a desirable outcome. Perhaps then it was to the relief of some that the High Court appeared to retreat from that view in Lange v Australian Broadcasting Corporation. Although the decision confirmed that the freedom of political communication is an established feature of Australian constitutional law, it also confirmed the arrival of a renewed conservatism in the High Court’s approach to constitutional interpretation and to the freedom of political communication in particular. The implication from representative and responsible government that had given rise to the freedom of political communication was considerably confined. The Court announced that the freedom of political communication would be interpreted only by reference to those aspects of representative and responsible government that could be identified in the ‘text and structure’ of the Constitution, eschewing any reference to ‘political principles or theories’ extraneous to the Constitution.
Although I have elsewhere joined in the criticism of the early freedom of political communication cases, especially the decision in Theophanous v Herald & Weekly Times Ltd, I do not share wholeheartedly in the relief felt by others in response to Lange. In this article, I argue that the High Court in Lange committed itself to an unsustainable interpretive theory. I will demonstrate this through consideration of an undecided doctrinal issue: the standard of review to be applied to laws implicating the freedom of political communication.
I should begin, then, by explaining the nature of that issue. Despite its unanimity in Lange, the High Court has not yet clearly indicated the level of protection that political communication is to be given. Although the Court has given sustained attention to the question of what kinds of communications are covered by the freedom, the question of how much protection such communication receives from the freedom has not been clearly answered. If a particular communication falls within the coverage of the freedom of political communication, does this mean that no regulation of that communication can be justified? Or can reasonable regulation pursuing some legitimate interest be justified? Or, perhaps, does political communication, or some types of it, require special protection?
The High Court’s answers to these kinds of questions has been rather unclear. Although at times the High Court appears to have settled on a proportionality test, at other times some members of the High Court have departed from this test and have held that a stricter standard of review, closely resembling the ‘strict scrutiny’ test seen in American constitutional law, should apply to certain categories of cases. These doctrinal swings, moreover, have been neither acknowledged by the Court nor the subject of sustained scholarly attention.
The resolution of this issue is a matter of considerable practical importance, but more importantly for my purposes, it demonstrates the limits of the interpretive method the Court adopted in Lange. Deciding on an answer to the standard of review question will require that the High Court depart from its commitment to text and a limited kind of structural implication and develop the freedom of political communication by reference to some values or ideas that are not, at least according the High Court’s avowed interpretive method, readily identifiable in the Constitution.
I will begin, in Part II, with a review of how the High Court has interpreted the freedom of political communication, showing the shift in Lange to a reading more closely confined by constitutional text. In Part III, I will consider the current state of the High Court’s doctrine on the standard of review question. I will show that the Court has fluctuated between adopting the relatively deferential proportionality test (or some synonymous formulation) as a single test to be applied to all regulation of political communication, and a ‘two-tiered’ test that applies strict scrutiny in certain cases.
In Part IV, I analyse these tests and the nature of the choice they present to the Court. With the help of some important debates that have occurred in the context of the First Amendment of the United States Constitution, I show that although each of these tests involves the balancing, or weighing up, of the interest protected by the freedom of political communication against some other interest, an important difference between the two tests lies in the degree of discretion each accords to the judge applying the rule. I conclude that in determining the future development of the standard of review, the Court essentially has to decide how much discretion is appropriate.
In Part V, I will show that the method of constitutional interpretation to which the Court appears to have confined itself in Lange cannot provide it with the answers to this problem. The choice actually requires it to revisit more fundamental issues about the role of judicial decision-making, specifically the relative merits of, on the one hand, an approach that leaves judges with much flexibility and, on the other, one that constrains their discretion but provides greater certainty.
I conclude by considering how the High Court ought to develop the doctrine in the future. I argue that whichever course it pursues, departure from the Lange method is inevitable. Although the proportionality test appears to be compatible with the Lange method, that test will only allow the High Court to postpone, rather than avoid, identifying the extra-constitutional values or principles underlying the freedom of political communication. The weighing of interests under the proportionality test requires the High Court to draw upon values that underlie the freedom of political communication and over time the proportionality test will produce a doctrine that reflects decisions about those values.
Thus, the only choice for the High Court is how it chooses to express those value judgments. Although the flexibility provided by the proportionality test is especially to be valued as the freedom of political communication develops, I will argue the ultimate expression of these value judgments in relatively defined rules (like the strict scrutiny part of the two-tiered test) is a desirable outcome and, moreover, is actually supported by the Court’s general interpretive approach exhibited in Lange. For although the Court’s commitment to ‘text and structure’ cannot solve the problem of choosing a standard of review, the same values that gave rise to it would be well served by the development of more defined tests.
As is well known, the freedom of political communication emerged from Australian Capital Television v Commonwealth and Nationwide News Pty Ltd v Wills, where the High Court held that the system of representative and responsible government established by the Australian Constitution impliedly prohibits legislative action that interferes with the political expression necessary for the proper operation of this system of government.
Importantly for my purposes, rather divergent interpretive approaches to the freedom of political communication are evident in these cases. On the one hand, there are the closely aligned approaches of Mason CJ and Deane, Toohey and Gaudron JJ. These Justices held that, by implication, the provisions of the Constitution establishing the Parliament and the Executive and allowing for constitutional amendment by referendum, entrenched the institutions of representative and responsible government. From this they draw the implication that the proper operation of these institutions required a certain level of free political communication.
However, from the beginning, two of the Justices who recognised the implied freedom of political communication were considerably more conservative in their formulation of the doctrine. Most significant for the ultimate development of the freedom is the judgment of McHugh J. Although in Australian Capital Television he also recognised a limit on the Parliament’s power to regulate political communication and found the law in question invalid, his analysis of the freedom of political communication is rather more closely tied to the text of the Constitution. Whereas Mason CJ and Deane, Toohey and Gaudron JJ recognised that the text gave rise to an implication of representative government which, in turn, required protection of the political communication necessary to protect representative government, McHugh J resisted the middle step: the recognition of the implication of representative government. Instead, his approach was to consider what specific provisions of the Constitution require. So, rather than considering what was required by the general concept of ‘representative government’, he returned to the specific provisions establishing elections for the House of Representatives and the Senate and considered what kind of free political communication is necessary for the proper working of those sections. In his view, this required that ‘the people have a constitutional right to convey and receive opinions, arguments and information concerning matter intended or likely to affect voting in an election for the Senate or the House of Representatives.’ Under this approach, the scope, or coverage, of the freedom of political communication is determined by reference to specific constitutional provisions rather than the more general concept of representative government.
This more conservative view of the freedom of political communication led McHugh J to dissent in the next major free speech case, Theophanous, and its companion case, Stephens v West Australian Newspapers Ltd. Moreover, after Theophanous, with the departure of Mason CJ and then Deane and Toohey JJ, a more conservative mood began to take hold in the Court and the emphasis on text seen in McHugh J’s judgments began to exert considerable influence. In McGinty v Western Australia, a majority of the Court rejected the notion that the Constitution contained what they described as a ‘free standing’ implication of representative government. Rather, as McHugh J had earlier held, they found that the constitutional concept of representative government is limited to specific aspects of representative government that can be identified in the ‘text’ or ‘structure’ of the Constitution.
The ascendancy of this kind of reasoning with respect to the freedom of political communication was ensured by Lange where the Court unanimously held that constitutional ‘text and structure’ alone govern the constitutional concept of representative government:
[T]he Constitution gives effect to the institution of ‘representative government’ only to the extent that the text and structure of the Constitution establish it ... the relevant question is not, ‘What is required by representative and responsible government?’ It is, ‘What do the terms and structure of the Constitution prohibit, authorise or require?’
The Court identified three elements of representative government discernible from the text and structure of the Constitution: the requirement discerned from ss 7 and 24 that ‘the members of the Senate and the House of Representatives ... be directly chosen at periodic elections by the people’; the requirement of responsible ministerial government found in the provisions that set out the relationship between the Executive and the Parliament; and the provision for constitutional amendment by popular referendum in s 128.
The Court’s view is that the freedom of political communication only exists to the extent necessary to enable these aspects of government to function. Therefore, it only protects communication ‘which enables the people to exercise a free and informed choice as electors’; communication ‘concerning the conduct of the executive branch of government throughout the life of a federal Parliament’; and communication of ‘information that might be relevant to the vote [electors] cast in a referendum to amend the Constitution.’ According to the High Court, then, the correct approach to a question concerning the freedom of political communication is to address these aspects of the Constitution and consider whether the impugned law interferes with communication that relates to one of these protected features.
Thus, after a period of controversy, the High Court appears to have settled upon an interpretive strategy to govern the freedom of political communication. It is a conservative approach to constitutional interpretation that draws principally from the text. Although this is supplemented by the possibility of ‘structural implication’, the Court has a narrow conception of such implications, limited to those ‘logically or practically necessary’ for the integrity of the Constitution. The method is notably adverse to drawing on broad statements of principles or values, because these cannot be located in the constitutional ‘text and structure’. That is not to say that the Court denies the influences of extra-constitutional ideas altogether. Indeed, McHugh J, perhaps the most persistent advocate of this method of constitutional interpretation, has acknowledged that the commitment to this method is itself a commitment to a theoretical position, drawn from values external to the Constitution. However, although this method of interpretation is itself based on an extra-constitutional theory, its point is to preclude other analysis of a theoretical or philosophical kind. So, except to the extent that the Court is committed to the general principle that the freedom of political communication is to be interpreted by reference to its ‘text and structure’, the Court has been quite emphatic that the freedom is not to be interpreted with reference to broad principles, overarching or underlying theories.
My point in this article is to show the limits of this interpretive method. In order to do so, I will consider an issue that has so far received rather less attention from the Court: the formulation of a standard of review.
Although the High Court has not clearly settled upon a standard of review, it has not been able to avoid the issue. From its earliest decisions, the High Court held that the freedom of political communication is not ‘absolute’, and consequently, the task of formulating a standard under which regulation of political communication is assessed immediately presented itself. That is, once it decided that not all laws affecting political communication are invalid, the High Court had to determine which regulations of political communication were valid, and which were not.
From the beginning, there were two different approaches. One applied a single test to all regulation of the freedom of political communication. One form in which this is expressed is found in the judgment of Gaudron J in Australian Capital Television. As her Honour formulated it, a law that burdens the freedom of political communication must be reasonably appropriate and adapted to a legitimate end. Brennan J’s approach in that case was similar.
This formulation of the test is, at first glance, misleading. In other circumstances, the ‘reasonably appropriate and adapted to’ formula has been used as a very minimal standard of review. When using the formula to interpret grants of Commonwealth legislative power, specifically the incidental power, the Court has shown a high level of deference to the law-makers’ selection of an appropriate means. Indeed, this deference was so great that the test has sometimes been described as a test of merely whether a law was a means to an end, as opposed to a reasonable or appropriate means to that end.
By contrast, the proportionality formula, which has also been used to interpret grants of Commonwealth power, is a more rigorous tool of judicial review. In contrast to its previous deference, when employing the language of proportionality the High Court would ask whether the end could be pursued by less drastic means, and it has been particularly sensitive to laws that impose adverse consequences unrelated to their object, such as the infringement of basic common law rights.
This kind of test resembles those employed in European Union law and in Canada. There, courts have identified three separate questions that make up the proportionality test: (1) whether a law is actually serving the end it purports to serve; (2) the availability of alternative, less drastic means by which that same end could be achieved; and (3) whether the end pursued by that law is worth the restriction or costs imposed.
To take Canada as an example, there the proportionality test is used as a way of determining whether a restriction on a freedom protected by the Canadian Charter of Rights and Freedoms is permissible under s 1 of the Charter. That section provides that each right or freedom is subject to ‘such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society.’ In determining whether a law limiting a right is justified, Canadian courts consider first whether the law serves an objective ‘of sufficient importance to warrant overriding a constitutionally protected right or freedom’ and then whether that limit is ‘reasonable and demonstrably justified.’
At this point, a proportionality test is employed. If a law limits a right or freedom, then it must satisfy each of the three requirements set out above. As the Canadian courts have expressed the test, it must be shown: (1) that the measures adopted to achieve a law’s end must be ‘carefully designed to achieve the objective of the law in question’; (2) that the measure should impair the right or freedom in question as little as possible; and (3) that there be a proportionality between the measure chosen and the ends pursued so that the importance of the law’s objective is not ‘outweighed’ by the encroachment on the right or freedom concerned.
Despite the High Court’s use of the formulation ‘reasonably appropriate and adapted to’, it is clear that in the context of the freedom of political communication, the Court does not use it to mean the minimal kind of review seen in other contexts. First, the High Court has been quite explicit that, in this context, the formulation is synonymous with proportionality. Moreover, as I will show below, the current use of this formula approximates the more rigorous proportionality test. Thus, in this context, included within the concept of ‘reasonably appropriate and adapted to’ are the various inquiries that make up the proportionality test.
The other common approach to the standard of review has been to vary the level of scrutiny applied according the nature of the regulation at issue. In Australian Capital Television, in a manner reminiscent of American First Amendment law, Mason CJ and McHugh J drew a distinction between ‘restrictions on communication which target ideas or information and those which restrict an activity or mode of communication by which ideas or information are transmitted.’ Laws that target information or ideas require a ‘compelling justification’ and ‘must be no more than is reasonably necessary to achieve the protection of the competing public interest’, whereas other laws are valid unless the ‘burden on free communication ... is disproportionate to the attainment of the competing public interest’.
Deane and Toohey JJ set out a slightly different, but also ‘two-tiered’, test. According to their Honours, a law ‘with respect to the prohibition or control of some or all communications relating to government or governmental instrumentalities’ would be particularly difficult to justify. These laws were contrasted with a law ‘with respect to some other subject and whose effect on such communications is unrelated to their nature as communications’. Laws in the former category could only be justified if either they ‘are conducive to the overall availability of the effective means of such communication in a democratic society’ or ‘do not go beyond what is reasonably necessary for the preservation of an ordered society or for the protection or vindication of the legitimate claims of individuals to live peacefully and with dignity’. Although Deane and Toohey JJ are not explicit about the meaning of the phrase ‘with respect to’, it suggests that they place rather less emphasis on the apparent purpose of a law and more on its actual operation.
It briefly appeared that this division was resolved by the Court’s unanimous opinion in Lange, which apparently endorsed the single test approach for review of all laws affecting the freedom of political communication. The Court held that where a law ‘effectively burden[s] freedom of communication about government or political matters’, the law must be ‘reasonably appropriate and adapted to serve a legitimate end’. The Court also acknowledged that this test was synonymous, in this context at least, with proportionality. However, shortly after Lange, the High Court delivered its judgment in Levy v Victoria, where it upheld regulations that limited the activities of demonstrators protesting against the recreational shooting of ducks. There, despite the apparent agreement in Lange on a single standard of review, the two-tiered tests re-emerged in some judgments.
The survival of the two-tiered tests is clearest in the judgments of Gaudron and Kirby JJ. Although it did not form part of her Honour’s analysis in earlier cases, in Levy Gaudron J drew on the distinction made by Mason CJ and McHugh J in Australian Capital Television. If the ‘direct purpose of the law is to restrict the freedom of political communication’, she would require that it be ‘necessary’ to achieve some ‘overriding public purpose’, whereas if regulation ‘only incidentally restricts political communication’, she would employ the more lenient and familiar requirement that the regulation must be ‘reasonably appropriate and adapted to’ that purpose. Kirby J took a similar view of laws targeted at a particular idea or message and also suggested that communications taking place in a traditionally public forum should receive particular protection.
Toohey and Gummow JJ also employed a two-tiered test, though it was expressed slightly differently. As with Toohey J’s position in the early cases, there is more focus on the operation of the law than on its purpose. Their Honours applied the less stringent ‘reasonably appropriate’ test to the Levy regulations because they ‘do not have, as their direct operation, the denial of the exercise of the constitutional freedom in a significant respect’, because they imposed ‘no general prohibition or regulation of communication or discussion’, and because they did not involve ‘a significant curtailment of the constitutional freedom of political communication and discussion.’ The implication seems to be that if any of these criteria had not been met, a higher standard of review would have been applied.
In summary, then, only two members of the Court in Levy applied the Lange test without qualification: Dawson and McHugh JJ. Particularly significant for these purposes, a majority of the Justices, Toohey, Gaudron, Gummow and Kirby JJ, identified circumstances in which they would apply a stricter test to regulations infringing the freedom of political communication.
Thus, despite the apparent clarity of the approach announced in Lange, the waters were immediately muddied by Levy and the precise nature of the High Court’s approach to the standard of review question is yet to be determined. The purpose of this article is to show that the solution to this doctrinal difficulty requires departure from the interpretive framework to which the High Court has committed itself. Before I make that part of the argument, it is necessary to establish some basic features of the two approaches.
I have referred above to the Canadian and European concepts of proportionality that require a law to satisfy three requirements: (1) that it be directed to a legitimate end; (2) that it employ the least restrictive means practicable to achieve that end; and (3) that the end to which the law is directed be worth the cost or restriction imposed in order to achieve it.
Although the Australian High Court has not so clearly differentiated each of these inquiries, this is essentially the process in which it engages when it applies its proportionality test. Certainly the first two steps outlined above form part of the High Court’s inquiry. By way of illustration, consider Levy, where the Court reviewed regulations that limited access to hunting areas during a particular weekend. The Court identified the end served by the law — protester safety — and then considered the proportionality of the measure chosen to pursue that end. Part of that inquiry involved a consideration of the availability of less drastic means to protect the protesters. One argument put to the Court was that the protection of protesters could have been secured without the complete exclusion of protesters from the hunting areas. Among those Justices who considered the argument, a majority rejected it, apparently on the basis that on the facts of this particular case the regulation was not unnecessarily restrictive.
Although the Court has not always been explicit about it, it is clear that the balancing of the interest pursued by the law against that pursued by the freedom does form part of its analysis. The ‘balancing’ process is certainly performed as part of the consideration of less drastic means. The ‘balancing’ arises because, although the availability of less restrictive means to achieve its end is relevant to a law’s validity, it is not determinative. To take Levy again, there obviously were less restrictive means of regulating the protesters. Perhaps the protesters could have been allowed to get closer to the hunting. Perhaps the restrictions might have operated over a more limited time. Perhaps (less plausibly) the regulations could have required an intermittent halt to the hunting which would have allowed the protesters to enter the hunting area to achieve their avowed aim of speaking publicly about recreational duck shooting and being seen on television aiding or collecting killed or injured birds. However, the availability of such means is clearly insufficient to establish that the Levy regulations were invalid. The employment of these means would impose an unacceptable cost on either the stated purpose of the regulations — protester safety — or on some other matter, such as the undisturbed continuance of hunting activities. Consequently, in considering the availability of less restrictive means the Court is really considering whether the means actually used to achieve a particular end were justified, given the alternatives. Therefore, what the Court is really doing is considering a specific aspect of the larger question, whether the end pursued is worth the restriction imposed.
Importantly, it should be noted that the Court engaged in this analysis in Levy despite the prevalence of the ‘reasonably appropriate and adapted to’ formula as an expression of the less strict standard of review. Thus it is clear that, despite the use of this expression, the Court is not engaging in the minimal kind of review that this formula entails in other contexts. By considering the availability of less drastic means, which in itself involves some balancing of means against ends, the essentials of the traditional proportionality test are subsumed within the concept of a law ‘reasonably appropriate and adapted to’ its end.
As will become apparent, the important point for the purpose of this article is that the proportionality test employed by the High Court involves this balancing of means against ends. Therefore, it is not really necessary for me to establish that the Australian proportionality test mirrors the Canadian and European test precisely. Nonetheless, it seems that the third part of the Canadian/European proportionality analysis — the consideration of whether the end pursued is worth the restriction — is a separate part of the High Court’s decisions on the freedom of political communication. If this question does play a role, then it would be possible for a law to fail the proportionality test even if it served a legitimate end and did so in the least restrictive manner practicable. It would do so if the restriction on free communication was so significant as to outweigh the competing public interest served by the impugned law. By way of illustration, if this analysis was applied in Levy, it would not be sufficient simply to show that the regulations actually served the aim of protecting protesters and that there were no reasonable or practical alternatives. As part of its determination the Court would consider whether, irrespective of the alternatives, the interest in protester safety is worth or outweighs the restriction on communication it imposed.
The High Court’s reasoning in the freedom of political communication cases indicates that this kind of balancing does form part of the proportionality test. In Australian Capital Television, several members of the Court clearly identified the end pursued by the impugned law — to reduce the reliance of political parties on money and thereby improve the fairness and integrity of the political process — as a value that competed with the freedom of political communication. Although, in part, the legislation seems to have been impugned on the ‘least restrictive means’ part of the test, the Court’s attention to the ‘free time’ provisions to invalidate the legislation suggests a balancing analysis. The objection to the free time provisions was that they favoured the established political parties. This suggests that the legislation would have failed the proportionality test even if it was accepted that the legislation served its aim of improving the political process (by reducing the dependence of political parties on money to finance electronic broadcasting) and that it did so in the least restrictive manner possible. It would be invalid because the benefits achieved by the legislation simply were not worth the entrenchment of established political parties.
Finally, some balancing also seems to be integral to Lange. The Court recognised the protection of reputation as an interest competing with free political communication and regarded the common law of defamation as balancing these two competing interests. By according more protection to free political communication, Lange resets the balance traditionally struck by the common law, out of a concern for free political communication.
With this understanding of the nature of the less stringent tests, we can see that the stricter test is actually somewhat similar. It also involves an assessment of the relationship between the means used by a law and the end pursued, but where the stricter test is employed, satisfying the Court of the legitimacy of the end pursued and, probably, the appropriateness of the means employed, will be more difficult.
Take the concept of ends. Under the less stringent form of the test, the Justices typically speak of a ‘legitimate end’. There is some uncertainty as to whether this simply requires that the impugned law be directed to an end that is within Commonwealth power or whether it imposes an additional requirement. Although some formulations of the test suggest that a legitimate interest might simply be one that is within the Commonwealth’s power to pursue, following Lange it is reasonably clear that an additional requirement that the end must itself be compatible with the freedom of political communication is imposed. But whatever the precise nature of the legitimate interest requirement under the less strict form of the test, it is clear that under the strict scrutiny test, it is replaced by something more demanding. Under the most common formulation of the test, the end pursued by the impugned law must be compelling or overriding rather than merely legitimate.
The doctrinal uncertainty I identified in Part II seems to call for a choice between two methods of weighing up the freedom of political communication against other interests. However, before I consider that choice, I should address one possible response to the debate I have just outlined: that applying a two-tiered system of tests (applying a stricter, more rule-like test to some categories of cases) rather than the simple proportionality test is not likely to make much difference in practice. Take, for example, Australian Capital Television and Levy, the two cases in which the different approaches have been most apparent. Almost without exception, the Justices reached the same conclusion, invalidating the Australian Capital Television law and validating the Levy law, no matter which test they applied.
There is something to this point. The tests have a similar structure. As I have just shown, under both forms of the test, the Court seeks an accommodation of competing interests in a manner that can be described as ‘balancing’. So one possible way to view the stricter forms of the test is that the Court has simply identified certain circumstances in which it will be harder to satisfy it that the law is proportionate to a particular end.
However, although there is some sense in this analysis, I think that it also overlooks some important differences. Although both forms of the test employ balancing, under the stricter form of the test, the balancing task is approached somewhat differently. It identifies some circumstances in which special weight is to be given to the freedom of political communication and, in those circumstances, the Court performs the balancing test with the scales already weighted in favour of the freedom of political communication. The point of the arguments against ad hoc balancing that I have just discussed is that this prior weighting of the scales is important because it reduces the discretion available to judges in each case.
Once this is understood, it can be seen that the real significance of the stricter forms of the test is that they reduce the flexibility accorded by the proportionality test. A striking feature of the proportionality test is that it gives the judge little guidance as to how the balance is to be struck. It is for the judge to determine, on a case by case basis, what weight the judge accords to each, when determining whether a law is proportionate to its purpose. In Levy, for example, those Justices who applied the proportionality test without modification simply made a judgment, with little explanation, as to whether this particular legislative goal — the safety of protesters — justified this particular imposition on free political communication. They had this freedom because the proportionality test itself does not give any guidance as to, and consequently does not place any restriction on, how judges assign weight to the competing interests. On the other hand, where the stricter forms of the test apply, the judge’s discretion is reduced. The rule requires that in certain circumstances free political communication is to be accorded a special weight.
So, the important distinction between the two tests lies not in any immediately obvious difference they will produce in result, but in the degree of flexibility or discretion that they leave in the hands of the judge applying them. Indeed, the nature of the distinction between the two tests has led one commentator to suggest that a two-tiered test which modifies the proportionality test in some circumstances is undesirable precisely because it restricts flexibility. However, the simple invocation of the flexibility of the test does not justify its use. The debate over standards of review under the Australian freedom of political communication here runs headlong into the American debate over the use of balancing tests in the law of the First Amendment, a debate essentially about the degree of discretion to respond to the facts of a particular case that a judge should have. Let me first outline that debate, before I return to consider its significance for the freedom of political communication.
At one time, certain First Amendment cases were decided by an approach described, at least by its critics, as ‘ad hoc’ balancing. That is, where the ad hoc balancing test was applicable, an American court proceeded by determining whether the interest in freedom of speech was, in the particular circumstances of the case, outweighed by the end to which the impugned law was directed.
When the debate first arose, the rival of ad hoc balancing was an approach described as ‘absolutism’. This position is associated with Black J, who rejected the balancing analysis in favour of an absolute protection in cases that would now be termed as cases concerning incidental burdens on speech.
However, balancing is now generally contrasted with a ‘conceptual’ or ‘categorical’ approach, that is, with the proliferation of doctrines that apply particular standards of review to particular categories of cases. First Amendment law is now characterised by a myriad of complex and often overlapping categories to which particular standards of review apply. In part, speech is categorised according to its subject matter. Most familiar in the Australian context is the distinct law relating to defamation of public figures that influenced the High Court in Theophanous. There are also categories relating to advocacy of illegal action, obscenity, child pornography, invasion of privacy, intentional infliction of emotional distress, ‘fighting words’, expressive conduct, and commercial speech. These are overlaid by categories drawn according to the place where speech takes place, most notably those relating to public fora, and according to the kind of regulation at issue such as the rules relating to prior restraints, and content-based and content-neutral restrictions.
The important feature of a categorical or conceptual approach is that it brings with it more defined standards of review. To each of the categories just described, a particular test of validity applies. Moreover, in many cases, the regulation of speech in that category is governed not by a flexible standard, like proportionality or an ad hoc balancing test, but by a rather determinate rule, that is, a test that is governed by the presence or absence of some specified facts. So, in applying such a rule, the judge considers whether those specified facts are present rather than whether, on balance, the regulation is justified. The presence or absence of these facts determines the result, even if a judge in an individual case might have weighed the interests differently.
One such rule is found in New York Times v Sullivan, which laid the foundation for the law relating to defamation of public figures to which I have just referred. Under that rule, a public official is prohibited from recovering damages for a defamatory falsehood relating to his or her official conduct unless the official proves that the statement was made with ‘actual malice’ — that is, with knowledge that it was false or with reckless disregard as to whether it was false or not. In applying this rule, rather than weighing up the interests at stake, the judge has to determine a series of questions. Is the plaintiff a public official? Was the statement made with actual malice or reckless disregard for the truth? The result is determined by the answers to these questions.
The principal argument in favour of ad hoc balancing is that it gives the judge applying the test much flexibility. When the judge balances the competing interests served by the impugned law and by protection of speech or communication, the judge can easily respond to the particular circumstances of the case. On the other hand, if, as under a more defined rule, protection depends on whether the speech in question has certain qualities, it is more difficult for a judge to be so flexible.
The constraint that a rule is said to impose can create problems when a court is presented with unforeseen facts, such as new technological and social developments. Professor Schauer has shown, for example, that the existing law of the First Amendment has difficulty dealing with the threatened publication of an article containing comprehensive, accurate and detailed analysis of the overall construction and operation of a thermonuclear weapon. The existing rule governing the regulation of advocacy of illegal conduct, the Brandenburg v Ohio rule, appears not to anticipate circumstances in which a publication might contain no incitement nor pose an imminent, or perhaps even probable, risk of lawless action, but might nonetheless pose such a risk that denial of First Amendment protection and the prevention of publication seems desirable.
A judge using an ad hoc balancing analysis would have little trouble responding to such facts. She would be able to conclude that the state interest in suppression in this case was so great that it outweighed the interest in freedom of speech. Earlier decisions protecting freedom of speech in the face of advocacy of illegal action would be distinguished on the basis that in those cases the interest in suppression was not nearly so weighty. So, balancing allows courts to respond to the special aspects of each case as they arise. This gives them the capacity to accommodate unforeseen circumstances, to respond to technological or social change and to attempt to exploit ‘loopholes’ to escape the operation of a rule. By contrast where a rule is applied, it may turn out that, as in the example just given, a rule is under-inclusive (does not apply to a circumstance to which it should apply) or over-inclusive (applies to facts to which it should not apply). The attempt to fit the unforeseen case into the existing rule is likely to produce anomalous results.
Just as the particularised, discretionary nature of ad hoc balancing contains the essence of its virtue, it also points to its principal flaw. An array of arguments can be put against such a particularised, discretionary approach. First, because the outcome is very dependent on the facts of a given case, it gives rise to uncertainty and, consequently, problems for those who wish to rely on the law. In its strongest form, the argument is that an ad hoc balancing approach is insufficiently certain to count as law. The argument against ‘balancing’ and similar discretionary approaches is therefore linked to the argument for ‘neutral principles’; that is, for judicial decision-making based on reasons generalisable beyond the particular case. It is based on a conception of the appropriate judicial role within a society committed to the rule of law. So even when a law is judge-made, the role of the judge is appropriately constrained by rules that reduce judicial discretion at the point of application.
Secondly, there is the argument that flexibility produces errors. The uncertainty that follows leaves judges in subsequent cases without much guidance and thus increases the probability of anomalous results. Thirdly, and related to the first two arguments, it is said that the use of such a particularised form of decision-making is especially inappropriate for high appellate courts. These, more than other courts, have a greater responsibility for providing guidance to lower courts, legislators, litigants and others who may seek to rely on, or apply, the law. Finally, there are arguments, with which I will deal in some detail below, that place a special emphasis on defined rules and certainty of outcome in the freedom of speech context.
The debate over categories and balancing illuminates the current state of Australian law on the standard of review issue, because the two approaches currently vying for ascendancy in the High Court approximate the two sides of the debate over ad hoc balancing.
First, a ‘proportionality’ test is much like the ad hoc balancing test. Like ad hoc balancing, it allows for a rather unconstrained weighing of interests on the facts of particular cases. This is not to say that proportionality tests are entirely unstructured. Indeed a feature of them is that they structure the judge’s inquiry by directing the judicial mind to identifying (at the first step) the purpose or end pursued by a law, (at the second step) the means used to achieve that end and (at the second and third steps) the overall effects of the law. Thus although, as I have explained above, the comparison of means against ends requires judges to balance competing values, the balancing task is given some structure by a proportionality test.
However, although a proportionality test encourages the systematic identification of the competing interests at stake, it does not alter the essentially ad hoc nature of the balance. That is because these tests do not in themselves give any guidance as to how the competing interests are to be weighed up. This can be seen in much of the Canadian law applying a proportionality test. A strong theme in cases that apply a proportionality test to laws limiting freedom of expression is that the application of a proportionality test must be contextual. That is, it must be done by reference to the particular circumstances, to the particular interests at stake in any given case and without reference to preconceived notions as to the resolution of those conflicts. As a consequence, one Canadian scholar has criticised the Canadian test in terms strikingly close to the American critique of ad hoc balancing:
[T]he methodology of section 2(b) [freedom of expression] adjudication is troubling and problematic. The foundation of that methodology is the contextual approach, which has ... transformed section 1 review into an ad hoc exercise that exalts flexibility at the expense of principle.
In its essential respects, then, the proportionality test as it has been used in Canada and, importantly, as it is currently employed in relation to the freedom of political communication, bears much similarity to ad hoc balancing. Given the similarity between proportionality and ad hoc balancing, it is not surprising that, as we have just seen, the advantage of proportionality is said to lie in its flexibility.
By the same token, the strict scrutiny test that we have seen some judges employ under the freedom of political communication has much in common with the more rule-like tests of the American First Amendment, and thus responds to many of the criticisms levelled at ad hoc balancing.
When the Court requires the strict scrutiny test to apply to certain categories of speech — say, where a law targets ideas or information, to use Mason CJ’s formulation — the Court announces that free political communication carries special weight in some circumstances. Although there remains an element of balancing in the application of such a test, it does constrain the exercise of judicial discretion in the particular case by nominating certain circumstances in which freedom of political communication is to be accorded special weight. So, the two-tiered approach which applies strict scrutiny to some kinds of regulation provides a more defined approach to the freedom of political communication which brings with it some benefits at the expense of flexibility.
The debate over ad hoc balancing thus suggests that, irrespective of the circumstances in which it would actually apply, the form of the test matters. It suggests, then, that in resolving this doctrinal uncertainty I identified in Part II, the High Court must engage with fundamental issues relating to the comparative merits of flexibility and certainty.
Before continuing, I should acknowledge that there is a prominent objection to the analysis of First Amendment law that I have advanced. This objection is similar to the Australian argument, which I have already addressed, that there is little difference between the two approaches. However, the American form of the argument is rather more radical. Whereas the Australian commentary has pointed to the similarity between the two tests in arguing that the tests amount to the same thing, the American argument is that the arguments for rules are misplaced because the benefits they offer are illusory. Despite appearances, the argument runs, rules cannot really constrain judges at all because, even where a test is rule-like, judges retain a significant amount of discretion.
In support of this argument, these critics point out that a judge’s capacity to determine whether factual conditions specified by a rule are actually present, or to create exceptions if a rule seems not to be applicable in all the circumstances, gives a judge a significant amount of discretion as to whether to apply these seemingly ‘determinate’ rules, or to take some other approach in response to the facts of a particular case. This is particularly likely to be the case if a rule employs some vague kind of term. The use of the concept of ‘recklessness’ in the supposedly determinate New York Times v Sullivan rule is a case in point. It is obvious that a judge has considerable discretion in determining whether particular behaviour is, or is not, reckless.
Again, there is some sense in this argument. It is undeniable that a rule cannot produce determinate results in all cases and, indeed, such absolute rigidity may not be desirable. But it is quite another thing to suggest that the form of the test does not make any difference at all. Professor Nimmer has addressed this argument in the context of his argument for ‘definitional balancing’. Definitional balancing is his term for the application of rules like the New York Times v Sullivan rule that, although they balance competing interests, remove discretion at the point of application of the rule. Although he concedes that ‘neither definitional balancing nor any other technique can offer absolute assurance that a given court under sufficient internal or external pressure in some ‘hard’ case will not depart from a definitional rule’, he argues that:
Nevertheless, definitional balancing can insulate a judge from legally irrelevant pressures to a considerable degree if the judge wishes such insulation. How much easier it would be for a conscientious judge ... to explain ... that he found as he did because that was ‘the rule,’ rather than because upon a weighing of the interests involved he found weightier the side that public opinion opposed.
So, rules can provide a ‘refuge’ for judges who want to resist popular sentiment. Indeed, rules might provide encouragement to do so, so that although judges might be able to avoid rules, it might nonetheless be true that judges are actually inclined to follow them, at least in most cases.
The American preference for rules gains added significance when viewed in the historical context of the First Amendment. American scholars, mindful that the First Amendment provided little protection from the intolerance of the McCarthy years, have argued that First Amendment doctrine ought to be constructed with such intolerant periods in mind. Rules ought to be formed now so that they might better withstand the pressure of intolerant times. Of course, there is nothing to stop a determined judge, or one who is simply weakened in the face of intolerant times, from modifying or simply abandoning a preformulated rule, but such rules, it is argued, represent the best chance of doing so.
These arguments place significant emphasis on the power of rules to protect speech. My argument does not, however, rely on the protective nature of rules. I am not suggesting that rules are preferable because they increase the protection of speech. My point is only to address the argument that expressing doctrine in the form of rules has no effect on judicial decision-making. Although it is true that rules do not necessarily create constraints, the American experience suggests that, nonetheless, the form of the test matters. Thus, although the proposition that rules do not make a difference would ultimately require empirical testing, there is good reason to suppose that rules can constrain judges even if they do not produce that result in all cases.
With the American debate in mind we can see that the standard of review issue requires the High Court to decide between the competing virtues of doctrinal flexibility and doctrinal definition. At this point, the limits on the Lange interpretive method are evident because that method precludes the kind of reasoning that might assist in making the choice. By tying the freedom of political communication so closely to the text, the High Court has precluded, and indeed intended to preclude, the development of a theoretical or philosophical basis for free political communication.
The difficulty this creates for the Court in choosing between a balancing or a categorical approach to the freedom of political communication can be seen by contrasting it with the American and Canadian positions. The experiences in these countries suggest that it might be possible to find some guidance by considering the nature of the freedom and what kind of test might best ensure the appropriate operation of the freedom.
In the First Amendment context, the tussle between flexibility and definition has been resolved in favour of definition, largely because of features of the American free speech tradition. That tradition places much emphasis on competition in an open ‘marketplace of ideas’. Regulation interfering with that marketplace is regarded with considerable suspicion, resulting in a highly protective free speech system.
For the reasons I have just explained, rules are regarded as providing more protection for the freedom of speech, particularly in times of intolerance. They are to be contrasted, in this regard, with ad hoc balancing. One concern with ad hoc balancing is that the uncertainty that accompanies it will produce self-censorship and ‘chill’ the unfettered exchange of ideas. Further, although balancing will not, in any given case, necessarily produce a less protective outcome than a more defined rule, the balancing approach is commonly criticised for a tendency toward what American commentators consider to be inadequate speech protection in the courts. This flows from two arguments as to the behaviour of law-makers and the judiciary. First, it is said that there is a tendency, a natural human instinct, to censor unpopular speech or the speech of unpopular speakers. It is almost inevitable, then, that law-makers will use their power to restrain speakers or speech of which they disapprove. This is a serious threat in a speech tradition that places a high value on the unfettered exchange of even the most unpopular ideas. Secondly, there is an argument that, under an ad hoc balancing test, judges tend to defer to legislative or executive judgment in a manner that tips the scales in favour of regulation. When considering a law regulating speech, a court is faced with a legislative or executive determination that some other interest outweighs speech protection. That is, the law-maker has already done the balancing and some interest other than freedom of speech has come out on top. In these circumstances, it is said, judges are likely to defer to the judgment of the majoritarian arms of government, particularly where it is unpopular speech that is regulated. A balancing test, which gives judges much discretion as to the weighing of speech against other values at the point of application, places an inadequate check on ‘the ingrained judicial deference’ to the elected arms of government.
However, just as the American free speech tradition seems incompatible with balancing, a different result might follow from different ideas about freedom of speech. The Canadian Supreme Court’s interpretation of the protection of freedom of expression in the Canadian Charter of Rights and Freedoms has in some respects been more sympathetic to government action. In part, this reflects textual differences. Section 1 of the Charter provides that rights are subject to ‘such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society’, thus explicitly acknowledging that freedom of expression might sometimes give way to competing interests. However, this tolerance of regulation also reflects philosophical differences. The Canadian Supreme Court seems to put less store in the unfettered exchange of ideas in ‘the marketplace of ideas’. It has allowed, for example, greater scope for the law of libel to apply to public discussion. It rejected the very protective New York Times v Sullivan rule partly out of a concern for the social cost of the deprecation of truth in public discourse to which, it is said, the rule gives rise. This scepticism of the marketplace of ideas finds support in some critiques of American free speech law that suggest that the marketplace of ideas philosophy neglects the distorting effect of existing inequalities in access to information and the capacity to communicate.
Although such a conception of freedom of speech does not necessarily entail balancing, it is more consistent with the deference to legislative and executive action that seems to follow from balancing. It is not surprising, therefore, that Canadian courts have chosen to resolve conflicts between freedom of expression and other values by a balancing analysis.
This analysis suggests that the choice between flexibility and definition, proportionality and the two-tiered system, might be assisted by reference to the values that underlie the freedom of political communication itself. However, under the Australian Constitution, this enterprise immediately runs into problems. Such an analysis is not easily performed within the bounds of the constitutional method that the High Court has set for itself. The High Court’s assertion that the freedom of political communication is governed solely by textually based interpretation would appear to preclude reference to the kinds of ideas just discussed and therefore the Court is left without much guidance as to the selection of a standard of review.
The mere fact that the ‘text and structure’ of the Australian Constitution reveal a commitment to (1) free elections for the House of Representatives and the Senate; (2) responsible government; and (3) an open referendum procedure, and the freedom of political communication necessary to sustain them, does not take the matter very far. Although it can be said that political communication must receive the kind of protection that these aspects of representative government require, this really only raises the question in another form: what kind of protection of political communication do those aspects of representative government require? Indeed, it is consistent with either of the two views that I have outlined above.
Take, for example, the requirement that the House of Representatives and Senate be ‘directly chosen by the people’ and the implication following from it that there be free elections for these bodies. What do such elections require in the way of free political communication? It might be that a level of protection similar to that provided by the First Amendment is required. If we adopt the American enthusiasm for the ‘marketplace of ideas’ and the related fear that governmental interference with it may give rise to authoritarianism, we might conclude that a truly free choice in parliamentary elections requires a commitment to unfettered exchange of political information and consequently a high level of protection from regulation. On the other hand, if it is accepted that government involvement in speech regulation can secure rather than interfere with an appropriately functioning democracy, an approach closer to the Canadian approach might be preferred. That is, it could be argued that in order to have free elections, to have a Parliament that is truly ‘chosen by the people’, a certain level of government intervention is not only permissible, but, indeed, desirable in order to correct the deficiencies of an unregulated marketplace.
To decide on this kind of basis, then, the High Court would need to make a choice between these two ideas, or develop its own ideas as to how freedom of expression relates to representative government. The High Court’s determinedly anti-theoretical approach to constitutional interpretation does not assist in that task. It says only that some freedom of political communication is necessary to protect certain institutions: free voting in elections and referenda, and responsible government. But still the questions remain: how much and what kind of protection of political communication does this entail?
What, then, should the High Court do? One rather radical solution would be to abandon the quest for a standard of review altogether. The High Court could do this by adopting instead an absolutist approach or, perhaps even more radically, by abandoning the freedom of political communication, precisely because it requires the Court to make decisions of this kind. However, I will not consider these arguments in much detail. Lange represents a firm determination that the freedom of political communication exists and that it is not absolute. Instead, I will suggest that the Court might in fact be able to avail itself of the benefits of each approach by identifying the circumstances in which each is most appropriate.
The proportionality test is an attractive option for the High Court as the freedom of political communication develops. One virtue of the proportionality test is that it allows the Court to avoid, at least for some time, the process of determining the ultimate values that the freedom of political communication ought to serve. This is because such a flexible and particularised test is ‘incompletely theorized’ in the sense explained by Professor Sunstein. That is, the test does not itself make important statements about the weight to be given to particular interests that might be pursued by an impugned law, or to the countervailing interest in free political discussion. This lack of theorisation is important for at least two reasons.
First, for a time at least, it makes the proportionality test workable within the constraints of the Lange method. In the face of such an insistently anti-theoretical stand, proportionality is very attractive precisely because, rather than requiring the High Court to formulate judgments as to how free speech is to be reconciled with other interests and to create rules expressing that judgment, proportionality allows a weighing of the competing interests only in particular circumstances.
By contrast, although the development of a categorical approach does not require the High Court to commit itself to any particular vision of the freedom of political communication, it does require it to have some idea about the values that the freedom serves and how it serves them. If the Court is going to create a rule that gives freedom of political communication special weight in particular circumstances, it needs some conception of the freedom of political communication against which to do this. Take, for example, the American distinction between content-based and content-neutral regulation that, as we have seen, has been influential in the High Court. The rule expresses a decision that, where content-based regulation is concerned, the interest in freedom of speech is to be given very great weight. Why is this? Several explanations have been advanced. Perhaps most commonly, it is explained as reflecting a traditional First Amendment suspicion of government involvement in determining the shape of public discourse. Does the Australian freedom of political communication share this concern? As we have just seen, the answer to this question is not clear and cannot readily be determined from the ‘text and structure’ of the Constitution to which the Court has apparently confined itself. Thus, to make a rule of this kind that expresses a judgment about the relative importance of free political communication and competing values inevitably involves the kind of reasoning against an overarching or underlying principle or set of values. This is just the kind of reasoning that the Court distanced itself from in Lange.
In addition to its compatibility with the Lange method, the proportionality test also brings with it the other benefits of ‘incomplete theorization’: the avoidance of error and of judicial disagreement. To develop a system of rules, the current High Court would have to decide for itself the values that the freedom of political communication ought to serve. However, as Professor Sunstein has shown, courts are ill-equipped to make decisions of this nature. Judges commit themselves to a general principle of what is right or good at the risk of error and in the face of the possibility of confusing division amongst members of a multi-member Court. And as I have argued elsewhere, these problems are particularly acute where, as with the freedom of political communication, a court is operating in the context of a new, complex and controversial area of law.
Indeed, these problems can be seen in the High Court’s limited attempts to incorporate rules into the freedom of political communication doctrine so far. Through these rules, the High Court has rather unthinkingly imported other conceptions of freedom of speech. To take an example I have discussed elsewhere, when the High Court adopted a form of the New York Times v Sullivan rule in Theophanous, it brought with it some of the American reverence for the marketplace of ideas. The Levy case, in which the strict scrutiny test approach re-emerged, provides another example. The adoption of a ‘compelling justification’ requirement in circumstances where a law targets an idea or where speech occurs in a traditionally public place imports, without much analysis, aspects of American free speech jurisprudence from which the High Court might want to differ. The importation of these ideas is not necessarily a bad result, but the unthinking importation of First Amendment values is a poor substitute for independent consideration of the values that the Australian freedom of political communication ought to serve.
Further, the phenomenon of disagreement is evident already in the various ways in which the Justices have identified the kinds of laws that require special attention under the freedom of political communication: laws that target information and ideas; laws ‘with respect to the prohibition or control of some or all communications relating to government or governmental instrumentalities’; laws whose ‘direct operation’ entail ‘the denial of the exercise of the constitutional freedom in a significant respect’; and laws that prohibit speech in a public forum. A test that leaves greater flexibility at the point of application allows the Court to avoid deciding whether any or all of these possibilities are correct. It is, therefore, less likely to promote disagreement among the members of the Court.
The immediate determination of a theory, or of a set of values, that underpins the freedom of political communication seems then to be unwise. For the moment, this seems to solve the problem of choosing between a flexible standard and a rules-based approach. Without such determinations, a rules-based approach is precluded. However, these arguments only create a case for a period of incomplete theorisation through a test like proportionality. There are several reasons why a proportionality test will prove unsatisfactory as a long term solution.
First, although under the proportionality test the Court need not immediately identify the values on which its decisions about the freedom of political communication depend, considerations of this nature will play a role in its decisions. The proportionality test will be applied in a case-by-case manner with each decision particular to the facts of each case. However, the reasoning in each case will, of course, be fleshed out by analogical reasoning, by the identification of relevant similarities and distinctions between cases. Once this is acknowledged, it is apparent that there must be some kind of value judgments that underlie the process. Sunstein himself acknowledges that:
[A]nalogizers cannot reason from one particular to another particular without saying something at least a little abstract. They must say that case A was decided rightly for a reason, and they must say that that reason applies, or does not apply, in case B.
Articulating that reason sows the seeds of a more theorised, or value-laden doctrine.
Take for example the decision in Levy that the regulations governing protests at the duck hunt were not disproportionate to their effect on political communication. In future cases concerning restrictions on political communication, the question will be ‘how like or unlike the regulations in Levy is the present law?’ or ‘how does the aim served compare with the aim of the Levy regulations?’
A simple example illustrates how this reasoning must involve recourse to values underlying the freedom. Imagine that the Commonwealth Parliament passed a law aimed at preventing protest within a certain distance of Parliament House in Canberra. Although the proportionality test gives the inquiry as to the law’s validity some structure by requiring a judge to identify the purpose of the law and then to consider whether the means used to achieve this end justified the cost imposed on free political communication, it says nothing, however, about how much weight to give to each of these. The Court would naturally refer, then, to its decision in Levy and the process of analogical reasoning would begin.
The issue could be posed as the competition between the freedom of political communication and another interest, with the competing interest being the need to protect the property of Parliament House. In this case, the Court would need to consider whether the interest in freedom of political communication is greater, or lesser, in this case than in Levy, and also whether the interest that competes with free political communication in this case is more, or less, deserving of protection than that served by the Levy law. This would raise the following questions: given the particular potency of protest in the vicinity of Parliament House, should this be given more weight than protest in more remote areas? Is the protection of the property of Parliament House a mere matter of convenience compared with the comparatively more serious matter of protecting protestors’ lives?
Alternatively, if the laws were aimed at preventing protest that disrupted the conduct of the business of the Parliament, the issue could be posed as a competition between ways of achieving appropriate political discussion. In this case, the Court would consider what kind of political communication is most valued. Is our Constitution one that values the opportunity for citizens to influence their representatives, if needs be through protest? Or does it give more value to calm deliberation by those in whose hands the power actually lies?
Of course, much would depend on the details of the law at issue. As I have discussed, the proportionality test considers the means used to achieve an end, as well as the relative importance of the law’s end, as against the cost imposed on free political communication. So, it would be relevant to consider just how much restriction was imposed by the law in search of its aim. There is no need, however, to determine whether such a law would be valid, or to provide answers here to the questions posed above. The point on which I wish to rely is that the proportionality test requires a balancing of interests and to do this the Court must make some judgments as to the values that the freedom of political communication serves.
Once again we have reached a point where the Court is faced with a question that cannot be answered by referring to the sections of the Constitution that institute elections and the like. Rather, it needs to decide how the interest in protecting those institutions is to be balanced against other interests and how, among competing possibilities, it is best protected. Thus, reasoning about the freedom of political communication will involve reference to values that are external to the Constitution and the High Court can only choose whether to express those values in rules or whether to allow them to remain unexpressed through the use of an incompletely theorised test.
This brings me to my final point. My argument is that, eventually at least, those value judgments should be expressed in propositions that are more defined and rule-like than the proportionality test. I make that argument for three reasons.
First, the development of some rule-like propositions is in fact the likely result of the long term application of the proportionality test. To return to the example I have just discussed, imagine that the High Court strikes down such a law because it places a special premium on free political expression in the public areas close to Parliament House. Such a decision could prove to be the first step in the development of a rule like the American public forum doctrine and the Court will have begun the process of determining the values that underlie the freedom and expressing those values in more defined propositions. Thus, the use of the proportionality test relies on values external to the Constitution and, over time, is likely to produce rule-like propositions that express these value judgments.
Indeed, this is an entirely familiar process. The eventual emergence of relatively defined propositions of law from the gradual process of analogical, case-by-case reasoning typifies common law reasoning. It would be a mistake, of course, to think that this is an entirely smooth, linear process. The process of formulating a rule out of existing precedent can be complicated by inconsistent, or at least apparently inconsistent, precedent. Indeed, the capacity to do so is considered the mark of a great common law judge. Nor are established rules entirely stable. Rules might prove inadequate in some way and the Court may develop exceptions that eventually lead to wholesale departure from an announced rule. But in the end, a system of relatively defined, relatively stable propositions is likely to emerge from a case-by-case analogical reasoning process. In the case of the freedom of political communication these propositions or rules will, whether the High Court likes it or not, reflect value judgments about what it is that free political communications serves.
The case for the eventual development of rules by the High Court is also supported by the traditional arguments made for rules in the First Amendment context. Although I have argued that the High Court ought to be cautious about too readily committing itself to rules that express such value judgments, the arguments for incomplete theorising, that make proportionality attractive as the freedom of political communication develops, are not enduring. As the doctrine matures and courts become more experienced with it and its effects, the arguments for an incompletely theorised approach like proportionality carry less weight and at some point are overtaken by the case for at least some degree of certainty, predictability and generalisable principle.
This is so whether or not the freedom of political communication develops, in a manner similar to the First Amendment, to be highly protective of freedom of speech. It is true that some of the arguments advanced above for rules are specific to the First Amendment context, that is, they support rules because rules provide the kind of protection of freedom of speech that the First Amendment requires. Other arguments for rules, however, apply outside the American free speech context. Indeed, the case for rules can be expressed more generally as a conception of the rule of law. One commonly advanced argument is that the rule of law is best served by an identifiable set of rules on which citizens can rely to guide their behaviour. Where judicial law-making takes the form of particularised decisions made on the facts of each case, with the law gradually changing over time, it undermines both the ideal that laws should be certain, stable and thus able to provide effective guidance to both citizens and courts, and the appearance of impartiality and fairness. Thus, there is an argument for rules that are of general application and it is on that form of the argument that I seek to rely here.
Finally, I suggest that the argument for rules is in fact supported by the general approach in Lange. For although I suggest that the Lange method is ultimately not sustainable, the ideals evident in it would be well served by the development of rules.
This argument has two steps. First, it relies on the point I have just made that the argument for rules relies on a conception of the rule of law that requires judicial review to be constrained by some ascertainable law of general application.
The second step in the argument is that the High Court’s commitment to textualism, supplemented by limited kinds of inferences from ‘structure’, also relies upon a conception of the rule of law: a conception that requires judicial review to be constrained by some legitimate law-making process.
The Court’s commitment to textualism is founded on the idea that judicial review is only legitimate when mandated by the constitutional text. McHugh J’s dissent in Theophanous makes the point by invoking the famous Engineers’ Case, which asserted the predominance of textualism in constitutional interpretation, and his Honour stated:
If this Court is to retain the confidence of the nation as the final arbiter of what the Constitution means, no interpretation of the Constitution by the Court can depart from the text of the Constitution and what is implied by the text and the structure of the Constitution. ... [T]his Court has consistently held that it is not legitimate to construe the Constitution by reference to political principles or theories that find no support in the text of the Constitution. ... [O]ne starts with the text and not with some theory of federalism, politics or political economy. The Engineers’ Case made it plain that the Constitution is not to be interpreted by using such theories ... unless those theories can be deduced from the terms or structure of the Constitution itself. It is the text and the implications to be drawn from the text and structure that contain the meaning of the Constitution.
In a similar manner, the Engineers’ Case itself grounds textualism in the political legitimacy of the Constitution:
It is ... the manifest duty of this Court to turn its earnest attention to the provisions of the Constitution itself. That instrument is the political compact of the whole of the people of Australia, enacted into binding law by the Imperial Parliament, and it is the chief and special duty of this Court faithfully to expound and give effect to it according to its own terms, finding the intention from the words of the compact, and upholding it throughout precisely as framed.
The special appeal of textualism is that it responds to both conceptions of the rule of law to which I have referred. If constitutional text really does govern a decision, then that decision is constrained by reference to a text that is ascertainable, general in its application, and the result of a legitimate law-making process. However, as I have argued, constraint by reference to text is not available in these circumstances. The question then becomes: what is the High Court to do if the text offers no constraint in the choice of a standard of review?
My argument is that, although the High Court cannot achieve constraint by reference to text, it is, nonetheless, able to achieve a measure of judicial constraint, especially on lower courts, through a system of rules. A system of rules like that which characterises the First Amendment does not respond to concerns that judicial decision-making should be constrained by legitimate sources of law. However, it does respond to these other concerns about judicial decision-making, by providing identifiable rules of general application that guide the behaviour of citizens and constrain the decision-making of courts. In short, my point is that if the High Court cannot advance all these rule of law ideals, it should seek to advance those it can.
In making this argument, I should concede that I have assumed that the High Court’s commitment to the ideal that judges should be constrained by the text brings with it a preference for defined, and therefore constraining, judge-made rules. This is not an inevitable conclusion. These ideas, which I have referred to as conceptions of the rule of law, are logically separate and a commitment to one does not necessarily require a commitment to the other. It is at least conceivable, then, that someone might value the idea that judicial decision-making is constrained by a legitimate source of law (such as constitutional text) and not hold a commitment to judicially created rules over more flexible standards and particularised decision-making. However, these two commitments have something in common: the desire that judges be constrained by some ascertainable law. It seems, then, to be at least a plausible, if not a likely, conclusion that the High Court’s commitment to textualism brings with it a preference for doctrinal definition. Given that the Court has to give up on the notion that the development of the freedom of political communication is constrained by the constitutional text, it should at least choose constraint, even when conferred by judicially created, rather than constitutionally dictated, rules.
I have launched here an attack on the interpretive conservatism seen in the High Court’s approach in Lange. I have argued that the freedom of political communication cannot be developed solely by reference to the constitutional ‘text and structure’. The unsustainability of the Lange method can be seen by considering how the High Court should choose between the proportionality test and other more defined tests, of which strict scrutiny is one example. This choice itself cannot be made without reference to ideas beyond the text and structure of the Constitution and, moreover, the future development of such tests requires departure from that method. More value-laden reasoning in the freedom of political communication doctrine might be unpalatable, even daunting, for the High Court, but it is the inevitable result of the course on which the Court set itself when it first recognised the freedom of political communication.
[*] BA, LLB (UNSW), LLM (Colum); Faculty of Law, Australian National University. Thanks are due to Vince Blasi, Tom Campbell, Michael Dorf, Kent Greenawalt, Graeme Hill, Michael Mathieson, William Ryan, George Williams and Leslie Zines for their comments on earlier drafts and other assistance with the development of this article. This article forms part of the author’s JSD dissertation at Columbia University School of Law.
 See, eg, ‘Symposium: Constitutional Rights for Australia?’ (1994) 16 Sydney Law Review 141.
 See, eg, Nicholas Aroney, Freedom of Speech in the Constitution (1998).
  HCA 25; (1997) 189 CLR 520, 559 (‘Lange’) ‘Freedom of communication on matters of government and politics is an indispensable incident of that system of representative government which the Constitution creates’. I have written about this decision and Levy v Victoria  HCA 31; (1997) 189 CLR 579 (‘Levy’) in Adrienne Stone, ‘Lange, Levy and the Direction of the Freedom of Political Communication’  UNSWLawJl 38; (1998) 21 University of New South Wales Law Journal 117.
 It is arguably a return to the textualism of Amalgamated Society of Engineers v Adelaide Steamship Co Ltd  HCA 54; (1920) 28 CLR 129 (‘Engineers’ Case’). See below nn 207–209 and accompanying text. See also George Williams, ‘Sounding the Core of Representative Democracy: Implied Freedoms and Electoral Reform’  MelbULawRw 6; (1996) 20 Melbourne University Law Review 848, 870, arguing that an earlier case, McGinty v Western Australia  HCA 48; (1996) 186 CLR 140 (‘McGinty’), signalled the ‘Lazarus-like comeback’ of the Engineers’ Case.
 Compare with the implications derived from the separation of judicial power provided by Chapter III of the Constitution. See, eg, Kable v Director of Public Prosecutions (NSW)  HCA 24; (1996) 189 CLR 51; Re Wakim; Ex parte McNally; Re Wakim; Ex parte Darvall; Re Brown; Ex parte Amann; Spinks v Prentice  VSC 227; (1999) 163 ALR 270.
 Lange  HCA 25; (1997) 189 CLR 520, 566–7.
 McGinty  HCA 48; (1996) 186 CLR 140, 232 (McHugh J): ‘[I]t is not legitimate to construe the Constitution by reference to political principles or theories that are not anchored in the text of the Constitution or are not necessary implications from its structure’. For the elaboration of this idea by the Court in Lange see below Part II(B).
 Adrienne Stone, ‘Freedom of Political Communication, the Constitution and the Common Law’ (1998) 26 Federal Law Review 219.
  HCA 46; (1994) 182 CLR 104 (‘Theophanous’).
 Lange  HCA 25; (1997) 189 CLR 520, 559–62, and see below Part II(B).
 The question of protection has been usefully distinguished, by American scholars, from the question of coverage. The coverage question concerns identifying what it is that is to be protected. In the United States Constitution, the question of coverage begins with the words of the First Amendment: the meaning of ‘speech’ or, perhaps the more general concept, ‘the freedom of speech.’ See Frederick Schauer, Free Speech: A Philosophical Enquiry (1982) 89–91; William Van Alstyne, ‘A Graphic Review of the Free Speech Clause’ (1982) 70 California Law Review 107. In Australia, the coverage of the freedom of political communication is determined by the institutions of representative and responsible government from which the freedom derives: Lange  HCA 25; (1997) 189 CLR 520. See below Part II(B).
 Police Department of Chicago v Mosely,  USSC 165; 408 US 92 (1972); Carey v Brown,  USSC 126; 447 US 455 (1980); Perry Education Association v Perry Local Educators Association,  USSC 34; 460 US 37 (1983) (‘Perry Education Association’); Simon & Schuster Inc v New York State Crimes Victims Board,  USSC 152; 502 US 105 (1991) (‘Simon & Schuster Inc’).
 There has been some consideration of the use of the ‘proportionality’ test and the High Court’s recent reluctance with respect to that formulation in the determination of the extent of grants of Commonwealth power: H P Lee, ‘Proportionality in Australian Constitutional Adjudication’ in Geoffrey Lindell (ed), Future Directions in Australian Constitutional Law (1994) 126; Jeremy Kirk, ‘Constitutional Guarantees, Characterisation and the Concept of Proportionality’  MelbULawRw 1; (1997) 21 Melbourne University Law Review 1; Brian F Fitzgerald, ‘Proportionality and Australian Constitutionalism’  UTasLawRw 16; (1993) 12 University of Tasmania Law Review 263. However, there has been little consideration of standards of review specifically directed to the freedom of political communication.
 For some of the writing on these cases, see Williams, ‘Sounding the Core of Representative Democracy’, above n 4; George Williams, Human Rights under the Australian Constitution (1999).
  HCA 45; (1992) 177 CLR 106 (‘Australian Capital Television’).
  HCA 46; (1992) 177 CLR 1 (‘Nationwide News’).
 Some members of the High Court have drawn a distinction between ‘representative government’ and ‘representative democracy’: Theophanous  HCA 46; (1994) 182 CLR 104, 130 (Mason CJ, Toohey and Gaudron JJ), 199 (McHugh J). The distinction is not important for my argument. For simplicity’s sake I will prefer the term ‘representative government’.
 These Justices defend the interpretive technique of drawing implications, that is, of recognising doctrines not expressed in the text but implied from constitutional text and structure:
[Grants of legislative power] must be read and construed in the context of, and as ‘subject to’, the Constitution as a whole including the fundamental implications of the doctrines of government upon which the Constitution as a whole is structured and which form part of its fabric. ... In that regard it has long been recognized that the ‘notion’ or ‘dogma’ which gained currency for a period after the decision in the Engineers’ Case to the effect that ... no implications can be made in construing the Constitution was mistaken.
Nationwide News  HCA 46; (1992) 177 CLR 1, 69 (Deane and Toohey JJ) (citations omitted). See also Australian Capital Television  HCA 45; (1992) 177 CLR 106, 135 (Mason CJ) (implications may be drawn from the structure of the Constitution that are ‘logically or practically necessary for the preservation of the integrity of that structure’), 208–9 (Gaudron J):
The Constitution cannot be construed in a vacuum. As Sir Owen Dixon pointed out in Jesting Pilate, it is the general law which is ‘the source of legal conceptions that govern us in determining the effect of the written instrument’ ... And, of course, the common law embraces those constitutional principles which have guided the development of democracy and responsible government in the United Kingdom.
 Nationwide News  HCA 46; (1992) 177 CLR 1, 70–3 (Deane and Toohey JJ); Australian Capital Television  HCA 45; (1992) 177 CLR 106, 137 (Mason CJ), 209–10 (Gaudron J).
 Nationwide News  HCA 46; (1992) 177 CLR 1, 72 (Deane and Toohey JJ); Australian Capital Television  HCA 45; (1992) 177 CLR 106, 138–40 (Mason CJ), 211–12 (Gaudron J).
 Australian Capital Television  HCA 45; (1992) 177 CLR 106, 227–35. The other Justice with a conservative view of the freedom of political communication was Brennan J, whose scepticism of the capacity of the Court to understand the ‘political milieu’ in which an impugned law operates led him to show rather more deference to the legislature, adopting the European notion of a ‘margin of appreciation’, as discussed in The Observer v United Kingdom  ECHR 49; (1991) 14 EHRR 153, 178. Thus, even though he recognised an implication limiting the Parliament’s power to interfere with political communication, he upheld the restrictions on political advertising challenged in Australian Capital Television  HCA 45; (1992) 177 CLR 106, 158–9.
 Australian Capital Television  HCA 45; (1992) 177 CLR 106, 232 (citations omitted).
 See above n 11.
  HCA 45; (1994) 182 CLR 211. Brennan J also dissented, though for slightly different reasons. In his view, the implied freedom operated only to limit the legislative and executive powers of the Commonwealth: Theophanous  HCA 46; (1994) 182 CLR 104, 149.
  HCA 48; (1996) 186 CLR 140, rejecting the argument that the implication from representative government required equal or equal-sized electorates for the Western Australian Parliament.
 Ibid 171 (Brennan CJ) (‘The [constitutional] principle of representative democracy ... can be no wider than — for it is synonymous with — what inheres in the text of the Constitution or in its structure’), 180–3 (Dawson J), 233 (McHugh J) (the principle of representative government is ‘a shorthand but inexact expression for an implication derived from the need to maintain the representative system of federal government that the Constitution sets up’), 281–3 (Gummow J).
  HCA 25; (1997) 189 CLR 520, 566–7 (citations omitted).
 Ibid 557. The High Court relied on s 1 (vesting the legislative power of the Commonwealth in the Parliament); ss 8 and 30 (electors for the Senate and the House of Representatives to vote only once); s 25 (persons of any race disqualified from voting at elections not to be counted in determining electorates under s 24); s 28 (duration of the House of Representatives) and s 13 (six years to be the longest term served by a Senator).
 Ibid 558–9. The High Court relied on s 6 (requiring a session of Parliament at least once a year); s 62 (executive power of the Queen exercised on the advice of ministers); s 64 (ministers required to sit in Parliament); s 49 (providing authority for each chamber to summon witnesses or to require the production of documents); and s 83 (requiring that money be appropriated from the treasury by law).
 Ibid 559.
 Ibid 560.
 Ibid 561.
 McGinty  HCA 48; (1996) 186 CLR 140, 231 (McHugh J): ‘Implications derived from the structure of the Constitution are also part of the Constitution’s meaning but such implications may be drawn only when they are “logically or practically necessary for the preservation of the integrity of that structure”.’ See also Australian Capital Television  HCA 45; (1992) 177 CLR 106, 135 (Mason CJ); Lange  HCA 25; (1997) 189 CLR 520, 566–7.
 McGinty  HCA 48; (1996) 186 CLR 140, 230: ‘The Constitution contains no injunction as to how it is to be interpreted. Any theory of constitutional interpretation must be a matter of conviction based on some theory external to the Constitution’.
 McGinty  HCA 48; (1996) 186 CLR 140, 168 (Brennan CJ), 182–3 (Dawson J), 231–2 (McHugh J), 270 (Gummow J); Lange  HCA 25; (1997) 189 CLR 520, 566–7.
 Nationwide News  HCA 46; (1992) 177 CLR 1, 51; Australian Capital Television  HCA 45; (1992) 177 CLR 106, 142–4 (Mason CJ), 159 (Brennan J), 169 (Deane and Toohey JJ), 217–18 (Gaudron J); Cunliffe v Commonwealth  HCA 44; (1994) 182 CLR 272, 299 (Mason CJ) (‘Cunliffe’).
 This discussion is drawn from Stone, ‘Lange, Levy and the Direction of the Freedom of Political Communication’, above n 3, 123–4.
 In later judgments, however, Gaudron J’s approach changed: see below nn 72–73 and accompanying text.
  HCA 45; (1992) 177 CLR 106, 218 (although note that her exact words were ‘reasonably and appropriately adapted to’). In this respect her Honour distinguished the freedom of political communication test slightly from the test used in the characterisation context described below in nn 42–44 and accompanying text. Where relevant to characterisation, the test is ‘whether the law is reasonably capable of being viewed as appropriate and adapted to achieving the purpose in question. Where the implied freedom is concerned, the test is more direct: it is whether the law is reasonably appropriate and adapted to the relevant purpose’: Cunliffe  HCA 44; (1994) 182 CLR 272, 388 (citations omitted) (emphasis in original).
 Australian Capital Television  HCA 45; (1992) 177 CLR 106, 157–8. See also Cunliffe  HCA 44; (1994) 182 CLR 272, 324–5. However, his Honour was particularly deferential to legislative action, adopting the European notion that courts should allow the Parliament a ‘margin of appreciation’ in assessing the need for regulation: Australian Capital Television  HCA 45; (1992) 177 CLR 106, 159, citing The Observer v United Kingdom  ECHR 49; (1991) 14 EHRR 153, 178.
 The Commonwealth Parliament has the powers necessary to effectuate the main purpose of the powers expressly granted to it. This power, the Australian cousin of the ‘necessary and proper’ clause (United States Constitution art I, s 8, cl 18), is described as the ‘express incidental power’. It is found in s 51(xxxix) of the Australian Constitution, which gives the Commonwealth power over ‘[m]atters incidental to the execution of any power vested by this Constitution in the Parliament’. It is also regarded as implied, by virtue of the ordinary rules of construction, in the grant of any of the enumerated powers: D’Emden v Pedder  HCA 1; (1904) 1 CLR 91, 109; Grannall v Marrickville Margarine Pty Ltd  HCA 6; (1955) 93 CLR 55, 77.
 An important statement of this approach is found in the judgment of Kitto J in Herald & Weekly Times Ltd v Commonwealth  HCA 78; (1966) 115 CLR 418, 437. In upholding restrictions on ownership in companies involved in broadcasting, even where those restrictions applied to persons with little, if any, influence over broadcasting activities, Kitto J famously remarked:
How far they should go was a question of degree for the Parliament to decide, and the fact that the Parliament has chosen to go to great lengths — even the fact, if it be so, that for many persons difficulties are created which are out of all proportion to the advantage gained — affords no ground of constitutional attack.
 This is Professor Zines’ view of Kitto J’s classic judgment in Herald & Weekly Times Ltd v Commonwealth  HCA 78; (1966) 115 CLR 418: Leslie Zines, The High Court and the Constitution (4th ed, 1997) 47.
 Its different use is partly explained by the different history of the proportionality concept. In contrast to the long acceptance of the ‘reasonably appropriate and adapted to’ formula, the concept of proportionality appears to have been imported more recently and to have come from European law, via the law of the European Community and the jurisprudence of the European Court of Justice: see Leask v Commonwealth  HCA 29; (1996) 187 CLR 579, 600–1 (Dawson J); Minister for Resources v Dover Fisheries Pty Ltd  FCA 366; (1993) 43 FCR 565, 575 (Gummow J); Lee, above n 13, 141; Kirk, above n 13, 3–4.
 Nationwide News  HCA 46; (1992) 177 CLR 1, 30 (Mason CJ). See also Zines, above n 44, 47; Kirk, above n 13, 29–30.
 Mason CJ summarised the point in Nationwide News  HCA 46; (1992) 177 CLR 1, 30–1:
[I]n determining whether [the] requirement of reasonable proportionality is satisfied, it is material to ascertain whether, and to what extent, the law goes beyond what is reasonably necessary or conceivably desirable for the achievement of the legitimate object sought to be attained and, in so doing, causes adverse consequences unrelated to the achievement of that object. In particular, it is material to ascertain whether those adverse consequences result in any infringement of fundamental values traditionally protected by the common law, such as freedom of expression.
 See Kirk, above n 13, 4–5.
 There are, of course, some questions antecedent even to this. One possible question is ‘what is the end to which this law is directed?’ For an analysis of how this question is answered, see Kirk, above n 13, 5–7. A second antecedent question is ‘is the end to which the law is directed a legitimate end?’ See text accompanying below n 95.
 In the European context, these three criteria have been dubbed ‘suitability’, ‘necessity’ and ‘balancing’. See Kirk, above n 13, 4–5.
 Part I of the Constitution Act 1982, being Schedule B to the Canada Act 1982 (UK) c 11.
 R v Big M Drug Mart Ltd 1985 CANLII 69;  1 SCR 295, 352; 18 DLR (4th) 321, 366; R v Oakes  1 SCR 103, 138; 26 DLR (4th) 200, 227.
 R v Oakes  1 SCR 103, 139; 26 DLR (4th) 200, 227.
 Ibid 139–40; 227. See also Irwin Toy Ltd v Quebec  1 SCR 927, 991–1000; 58 DLR (4th) 577, 624–30.
 Lange  HCA 25; (1997) 189 CLR 520, 567 fn 272.
 See below Part IV(A).
 See below n 86 and accompanying text.
 American free speech law distinguishes between ‘content-based’ and ‘content-neutral’ regulation. Laws in the first category — that is, laws directed at speech or communication because of its content — are generally subject to ‘strict scrutiny’. To justify such laws, it is necessary to show either that the kind of speech regulated falls into a category of ‘low value’ speech, or that they are narrowly tailored to serve a compelling government interest: Perry Education Association USSC 34; , 460 US 37 (1983); Simon & Schuster Inc USSC 152; , 502 US 105 (1991).
  HCA 45; (1992) 177 CLR 106, 143 (Mason CJ). See also 234–5 (McHugh J).
 Ibid 143 (Mason CJ). McHugh J’s view was similar: ‘[A] law ... which seeks to prohibit or regulate the content of electoral communications can only be upheld on grounds of compelling justification’: at 235.
 Ibid 143 (Mason CJ). See also 235 (McHugh J); Cunliffe  HCA 44; (1994) 182 CLR 272, 299–300 (Mason CJ).
 Nationwide News  HCA 46; (1992) 177 CLR 1, 76–7. As Deane J later expressed it, it will be difficult to justify ‘a general prohibition or regulation of communication or discussion as such or ... a significant curtailment of the freedom of political communication and discussion’: Cunliffe  HCA 44; (1994) 182 CLR 272, 339 (Deane J). See also 381, 383 (Toohey J).
 Nationwide News  HCA 46; (1992) 177 CLR 1, 76–7. See also Australian Capital Television  HCA 45; (1992) 177 CLR 106, 169.
 Nationwide News  HCA 46; (1992) 177 CLR 1, 77 (citations omitted). See also Australian Capital Television  HCA 45; (1992) 177 CLR 106, 169.
 Nationwide News  HCA 46; (1992) 177 CLR 1, 77. See also Australian Capital Television  HCA 45; (1992) 177 CLR 106, 169. Although this standard of review might appear less stringent than the ‘compelling justification’ test advanced by Mason CJ and McHugh J, the elaboration of the concept by Deane J in Cunliffe suggests otherwise. The requirement that a law must ‘not go beyond what is necessary’ means that it must be justified by a ‘pressing social need’: Cunliffe  HCA 44; (1994) 182 CLR 272, 339–40 (quoting Lord Goff in Attorney-General v Guardian Newspapers Ltd [No 2]  1 AC 109, 283–4).
 These words have been judicially considered in the interpretation of s 51 of the Constitution which grants the Commonwealth legislative power ‘with respect to’ the matters enumerated in that section. The accepted understanding of these words is that they require a determination of ‘what is the actual operation of the law in question in creating, changing, regulating or abolishing rights, duties, powers or privileges’: Bank of NSW v Commonwealth  HCA 7; (1948) 76 CLR 1, 187. See also Fairfax v Federal Commissioner of Taxation  HCA 64; (1965) 114 CLR 1, 7 (Kitto J), 16 (Taylor J). Considerable emphasis is put on the practical effect of the law: Herald & Weekly Times Ltd v Commonwealth  HCA 78; (1966) 115 CLR 418, 440; Commonwealth v Tasmania  HCA 21; (1983) 158 CLR 1, 152.
 Lange  HCA 25; (1997) 189 CLR 520, 567.
 Ibid 567 fn 272.
  HCA 31; (1997) 189 CLR 579.
 Although less relevant to this article, the peculiarly deferential approach of Brennan J, discussed above n 21, and who since Nationwide News and Australian Capital Television had become the Chief Justice, also re-emerged. Unlike the other Justices who considered, and rejected it on its merits, Brennan CJ would not entertain an argument that the regulations were invalid because they were ‘overbroad’. In his Honour’s view, ‘[u]nder our Constitution, the courts do not assume the power to determine that some more limited restriction than that imposed by an impugned law could suffice to achieve a legitimate purpose’: Levy  HCA 31; (1997) 189 CLR 579, 598.
 See above n 39 and accompanying text.
 Levy  HCA 31; (1997) 189 CLR 579, 619.
 Ibid. In Kruger v Commonwealth  HCA 27; (1997) 190 CLR 1, 128, Gaudron J expressed the distinction as one between a law the purpose of which was to restrict political communication (which requires a compelling justification) and a law that restricted it as an incident to achieving some other purpose (to which the proportionality test applies).
 Levy  HCA 31; (1997) 189 CLR 579, 645.
 In doing so, his Honour was influenced by United States authority which is especially hostile to the regulation of speech in public fora: ibid 638–42.
 Ibid 614.
 Ibid (citations omitted).
 Ibid 608–9 (Dawson J), 627 (McHugh J). For the approach of Brennan CJ, see above n 70.
 See above nn 48–54 and accompanying text.
 See Kirk, above n 13, 17–19; Williams, Human Rights under the Australian Constitution, above n 14, 88–91.
 Levy  HCA 31; (1997) 189 CLR 579, 614–15 (Toohey and Gummow JJ), 627 (McHugh J), 647–8 (Kirby J). Brennan CJ rejected the argument because, on his view, the test should accord more deference to legislative judgment: at 598. See also above n 21 for a discussion of Brennan CJ’s approach in past cases.
 This is with the exception of Mason CJ in Australian Capital Television  HCA 45; (1992) 177 CLR 106, 143, who said of the impugned law restricting political advertising on radio and television:
Whether those restrictions are justified calls for a balancing of the public interest in free communication against the competing public interest which the restriction is designed to serve, and for a determination whether the restriction is reasonably necessary to achieve the competing public interest (citations omitted) (emphasis added).
See also Cunliffe  HCA 44; (1994) 182 CLR 272, 300 (Mason CJ).
 See John Hart Ely, ‘Flag Desecration: A Case Study in the Roles of Categorization and Balancing in First Amendment Analysis’ (1975) 88 Harvard Law Review 1482, 1484–7.
 The protesters might have relied on an argument that they would not be in danger when they entered on the site because the shooting of ducks occurred while they were ‘on the wing’ and consequently guns were aimed well above the level of any human intruder. This was put in support of an argument that the regulations were entirely unnecessary, but it might also support an argument that they were overly restrictive. See Levy  HCA 31; (1997) 189 CLR 579, 599.
 Ibid 625.
 Ibid 608–9 (Dawson J), 614–15 (Toohey and Gummow JJ), 619 (Gaudron J), 627 (McHugh J), 645 (Kirby J).
 See below Part IV(E).
 See Kirk, above n 13, 9. The law might be invalid either because of the means or the ends. In the first case the impugned law would be directed to some legitimate end, but would employ means that are impermissible even though they are the least restrictive means of achieving that end. In the First Amendment context, Professor Volokh has suggested that a law preventing a presidential candidate from indicating that he or she would settle an ongoing war on terms more favourable than a sitting President might fall into this category. It is directed to a legitimate (indeed, a compelling) end and could be narrowly tailored (that is, use the least restrictive means possible) to that end. He nonetheless suggests that it would be invalid under the First Amendment because it uses impermissible means. See Eugene Volokh, ‘Freedom of Speech, Permissible Tailoring and Transcending Strict Scrutiny’ (1996) 144 University of Pennsylvania Law Review 2417, 2425–31. In the second kind of case, the law is invalid because the nature of the end to which the law is directed cannot be justified whatever the means used to achieve it. It is invalid even though the least restrictive means is employed. One way of expressing the analysis in the second kind of case is that the end pursued by the law is not ‘legitimate’. In these cases, the third step in the proportionality analysis appears to overlap with the ostensibly separate question of whether the end pursued is a legitimate end. See below n 95.
 Australian Capital Television  HCA 45; (1992) 177 CLR 106, 143, 146 (Mason CJ), 174–5 (Deane and Toohey JJ), 238–9 (McHugh J); Theophanous  HCA 46; (1994) 182 CLR 104, 131–3 (Mason CJ, Toohey and Gaudron JJ), 178–84 (Deane J).
 Australian Capital Television  HCA 45; (1992) 177 CLR 106, 175 (Deane and Toohey JJ):
The argument that to achieve the postulated objectives it is necessary effectively to exclude the people of the Commonwealth, including legitimate special interest groups, from political communication on the electronic media during an election period unless they be political parties or candidates seems to us, however, quite unconvincing.
 Under ss 95F–95R of the Political Broadcasts and Political Disclosures Act 1991 (Cth), free broadcast time was distributed to established political parties according to votes won at the previous election: Australian Capital Television  HCA 45; (1992) 177 CLR 106, 126–8.
  HCA 45; (1992) 177 CLR 106, 131–2 (Mason CJ), 175 (Deane and Toohey JJ), 237–8 (McHugh J).
 Lange  HCA 25; (1997) 189 CLR 520, 568.
 See Australian Capital Television  HCA 45; (1992) 177 CLR 106, 217–18 (Gaudron J); Nationwide News  HCA 46; (1992) 177 CLR 1, 95 (Gaudron J); Levy  HCA 31; (1997) 189 CLR 579, 645 (Kirby J).
 The precision with which the High Court described the nature of a legitimate end in Lange  HCA 25; (1997) 189 CLR 520, 567, as an end ‘the fulfilment of which is compatible with the maintenance of the constitutionally prescribed system of representative and responsible government and the procedure prescribed by s 128’, suggests that some laws, which would otherwise be within power, will fail under the freedom of political communication because of the nature of the end they pursue, no matter the means employed to achieve it. Indeed, this must be the case with respect to a law of which the very purpose is incompatible with the freedom, for example a law that is directed to, as distinct from merely having the effect of, increasing the likelihood that sitting members of parliament are returned. Moreover, this view would be consistent with the High Court’s approach to s 92 of the Constitution (‘trade, commerce ... among the States ... shall be absolutely free’), where a law that places a discriminatory burden on interstate trade and commerce with a protectionist effect must, to be valid, both pursue a non-protectionist end and use measures proportionate to that end: see Castlemaine Tooheys Ltd v South Australia (1990) 169 CLR 436, 472–3.
 Australian Capital Television  HCA 45; (1992) 177 CLR 106, 143 (Mason CJ), 235 (McHugh J); Levy  HCA 31; (1997) 189 CLR 579, 614–15 (Toohey and Gummow JJ), 619 (Gaudron J), 647 (Kirby J); Kruger  HCA 27; (1997) 190 CLR 1, 128 (Gaudron J). It may also be that the scrutiny of means is more demanding. Under the American test, a measure subject to strict scrutiny must be ‘narrowly tailored’ towards its end, leaving very little latitude to the law-maker to impose unnecessary restrictions: Perry Education Association USSC 34; , 460 US 37 (1983); Simon & Schuster Inc USSC 152; , 502 US 105 (1991). The High Court has not been as explicit about adopting this feature of the American test. As it is usually expressed in Australian law, a law subject to strict scrutiny must be ‘necessary’ or must ‘do no more than is reasonably necessary’ to achieve its nominated end: Australian Capital Television  HCA 45; (1992) 177 CLR 106, 143 (Mason CJ), 169 (Deane and Toohey JJ); Cunliffe  HCA 44; (1994) 182 CLR 272, 339 (Deane J); Levy  HCA 31; (1997) 189 CLR 579, 619 (Gaudron J); Kruger v Commonwealth  HCA 27; (1997) 190 CLR 1, 128 (Gaudron J). Although it remains to be seen how exactly this requirement will be applied, it may well be that the High Court will impose a more demanding test on the means employed by the law than that imposed by the proportionality test. The language could certainly bear this meaning and this interpretation would be compatible, moreover, with the American test that has, so far, had great influence on the High Court.
 A R Blackshield, ‘The Implied Freedom of Communication’ in Geoffrey Lindell (ed), Future Directions in Australian Constitutional Law (1994) 232, 253–4. See also Union v Laing (1998) 159 ALR 73, 91–2 (citations omitted):
The freedom of communication in relation to public affairs and political discussion ... will not invalidate a law enacted to satisfy some legitimate end if that law is compatible with the maintenance of representative and responsible government under the Constitution and is reasonably appropriate and adapted to achieving the legitimate end ... This single standard will subsume and allow for application of the distinction made in Australian Capital Television between laws directed to the content of communication and laws directed to modes of communication or activities affecting communication.
 In Australian Capital Television  HCA 45; (1992) 177 CLR 106, five Justices found the legislation invalid, one of the dissentients being Dawson J who refused to recognise a freedom of political communication at all. For the approach of Brennan J, the other dissentient, see above n 21 and accompanying text. In Levy  HCA 31; (1997) 189 CLR 579, all the Justices found the regulations valid.
 Professor Fallon makes the same point about the ‘suspect-content’ and ‘non-suspect content’ tests employed in the American law of the First Amendment. See Richard H Fallon Jr, ‘Foreword: Implementing the Constitution’ (1997) 111 Harvard Law Review 54, 79: ‘There is, undoubtedly, a sense in which these tests could count as balancing tests. Both require courts to assess whether a statute ought to be upheld, in light of the governmental interests that it serves, despite its impact on constitutionally protected values.’
 As I will argue, they resemble, therefore, tests which, in the United States, have been pejoratively described as ‘ad hoc’ balancing. See below Part IV(D).
 This was explicitly recognised by Brennan J in Australian Capital Television  HCA 45; (1992) 177 CLR 106, 150–1 (emphasis added):
[A] law which trespasses upon absolute freedom to communicate ... is a valid law provided the restrictions imposed by the law are proportionate to the interest which the law is calculated to serve. The proportionality of the restriction to the interest served is incapable of a priori definition: in the case of each law, it is necessary to ascertain the extent of the restriction, the nature of the interest served and the proportionality of the restriction to the interest served.
 Kirk, above n 13, 17.
 These cases concerned regulation that had an indirect effect on freedom of speech incurred as a by-product of regulation directed at some other object. See Communist Party v Subversive Activities Control Board,  USSC 137; 367 US 1 (1961), which upheld an order requiring the Communist Party to register as a ‘Communist-action organisation’ under the Subversive Activities Control Act of 1950 s 7, 50 USC s 786 (1994); Barenblatt v United States,  USSC 132; 360 US 109 (1959) (‘Barenblatt’), which upheld the petitioner’s conviction for failure to answer questions put to him by a Congressional Subcommittee; Konigsberg v State Bar of California,  USSC 73; 366 US 36 (1961) (‘Konigsberg’), which rejected the claim by the petitioner, who had been refused admission to the California bar, that the First Amendment provided a privilege not to respond to questions dealing with Communist Party Membership. See also Louis Henkin, ‘Infallibility under Law: Constitutional Balancing’ (1978) 78 Columbia Law Review 1022, 1045; Harry Kalven Jr, ‘Upon Rereading Mr Justice Black on the First Amendment’ (1967) 14 UCLA Law Review 428, 443.
 Laurent B Frantz, ‘Is the First Amendment Law? — A Reply to Professor Mendelson’ (1963) 51 California Law Review 729, 730; Melville B Nimmer, ‘The Right to Speak from Times to Time: First Amendment Theory Applied to Libel and Misapplied to Privacy’ (1968) 56 California Law Review 935, 938.
 Barenblatt,  USSC 132; 360 US 109, 126 (1959): ‘Where First Amendment rights are asserted to bar governmental interrogation resolution of the issue always involves a balancing by the courts of the competing private and public interests at stake in the particular circumstances shown’; Konigsberg,  USSC 73; 366 US 36, 51 (1961):
Whenever [in the context of incidental burdens on speech] these constitutional protections [of the First and Fourteenths Amendments] are asserted against the exercise of valid governmental powers a reconciliation must be effected and that perforce requires an appropriate weighing of the respective interests involved.
 Konigsberg USSC 73; , 366 US 36, 61 (1961): ‘I believe that the First Amendment’s unequivocal command that there shall be no abridgment of the rights of free speech and assembly shows that the men who drafted our Bill of Rights did all the “balancing” that there was to be done’; Barenblatt USSC 132; , 360 US 109, 139–45 (1959); Edmond Cahn, ‘Justice Black and First Amendment “Absolutes”: A Public Interview’ (1962) 37 New York University Law Review 549, 559. Black J did not, however, take the view that the First Amendment did not allow for any regulation of speech. Indeed, in some kinds of cases, those involving ‘time, manner or place’ restrictions, he took the view that balancing was an appropriate way to resolve the tension between the protection of speech and the end of an impugned law. See Kalven, above n 103, 442–3; Nimmer, above n 104, 935–6. The more radical proposition that all ‘speech’ should receive absolute protection under the First Amendment has never been seriously advanced by a member of the Supreme Court: Henkin, above n 103, 1044; Ely, above n 83, 1500 fn 75. But see the commentators discussed by Professor Smolla in Rodney A Smolla, Free Speech in an Open Society (1993) 371–2 fn 8.
 See Kent Greenawalt, Fighting Words: Individuals, Communities and Liberties of Speech (1995) 12.
 Frederick Schauer, ‘Categories and the First Amendment: A Play in Three Acts’ (1981) 34 Vanderbilt Law Review 265.
 Frederick Schauer, ‘Codifying the First Amendment: New York v Ferber’  Supreme Court Review 285.
 New York Times v Sullivan,  USSC 40; 376 US 254 (1964).
  HCA 46; (1994) 182 CLR 104.
 Brandenburg v Ohio,  USSC 141; 395 US 444 (1969).
 Roth v United States,  USSC 100; 354 US 476 (1957).
 New York v Ferber USSC 169; , 458 US 747 (1982).
 Time, Inc v Hill,  USSC 11; 385 US 374 (1967); Florida Star v BJF,  USSC 123; 491 US 524 (1989).
 Hustler Magazine v Falwell,  USSC 24; 485 US 46 (1988).
 Chaplinsky v New Hampshire,  USSC 50; 315 US 568 (1942).
 O’Brien v United States,  USSC 116; 391 US 367 (1968).
 Board of Trustees of the State University of New York v Fox,  USSC 144; 492 US 469 (1989).
 Perry Education Association USSC 34; , 460 US 37 (1983).
 Near v Minnesota,  USSC 154; 283 US 697 (1931).
 Simon & Schuster Inc USSC 152; , 502 US 105 (1991).
 That is not to say that all First Amendment law is dominated by such tests. ‘Incidental’ or ‘time, place and manner’ regulation of speech has traditionally been subjected to a different, more particularised and deferential approach. However, the deference shown to these regulations is subject to a significant number of exceptions where a heightened form of scrutiny applies. See Michael C Dorf, ‘Incidental Burdens on Fundamental Rights’ (1996) 109 Harvard Law Review 1175, 1200–10; Geoffrey R Stone, ‘Content-Neutral Restrictions’ (1987) 54 University of Chicago Law Review 46, 50–2. For the purposes of this article, it is sufficient to note the prevalence of rule-like tests in First Amendment doctrine.
 My terminology here, and throughout the rest of this article, draws on the much discussed distinction between rules and standards. As it is usually understood, a ‘rule’ binds a decision-maker to respond in a determinate way to the presence of specified triggering facts. Such a rule is based on some kind of background principle or policy but ordinarily a rule will apply even where its application might be contrary to that background justification. By contrast, a ‘standard’ seeks to apply the background justification or policy directly to a fact situation. Standards allow the decision-maker more discretion because they encourage taking into account all the relevant factors in a particular case. See Kathleen M Sullivan, ‘Foreword: The Justices of Rules and Standards’ (1992) 106 Harvard Law Review 22, 58–9. See also Pierre Schlag, ‘Rules and Standards’ (1985) 33 UCLA Law Review 379; Frederick Schauer, Playing by the Rules: A Philosophical Examination of Rule-Based Decision-Making in Law and in Life (1991) 104 fn 35.
 Sullivan, above n 124, 58.
  USSC 40; 376 US 254 (1964).
 Schauer, ‘Categories and the First Amendment’, above n 108, 301; see also Nimmer, above n 104; Greenawalt, above n 107, 12.
 New York Times v Sullivan,  USSC 40; 376 US 254, 279–80 (1964).
 It should be acknowledged, however, that the notion of ‘recklessness’ involves some kind of comparison of harm imposed by the publication as against its benefit. See Greenawalt, above n 107, 12.
 Ibid. For some arguments that criticise this understanding of rules, see below Part IV(F).
 Above Part IV(C).
 United States v The Progressive Inc, 467 F Supp 990 (WD Wis, 1979). Attempts to suppress the article were ultimately abandoned when the information emerged from other sources: Schauer, ‘Categories and the First Amendment’, above n 108, 296–8.
  USSC 141; 395 US 444, 447 (1969): ‘[T]he constitutional guarantees of free speech and free press do not permit a State to forbid or proscribe advocacy of the use of force or of law violation except where such advocacy is directed to inciting or producing imminent lawless action and is likely to incite or produce such action.’
 Schauer, ‘Categories and the First Amendment’, above n 108, 297–8. Schauer also makes this point about child pornography, itself the subject of a special First Amendment rule formulated in New York v Ferber,  USSC 169; 458 US 747 (1982), in Schauer, ‘Codifying the First Amendment’, above n 109, 311:
[W]e cannot incorporate into our standards intended to guide the future every contingency because we do not know what they will be ... Ferber itself is a perfect example because the phenomenon of child pornography is so new that it would have been impossible to predict even ten years ago. And there is no reason to believe that ten years from now we will not be presented with First Amendment issues that we have no way of foreseeing today.
The First Amendment problems posed by the internet are proof of Professor Schauer’s foresight in this regard: see Cass R Sunstein, ‘The First Amendment in Cyberspace’ (1995) 104 Yale Law Journal 1757; Charles Nesson and David Marglin, ‘The Day the Internet Met the First Amendment: Time and the Communications Decency Act’ (1996) 10 Harvard Journal of Law and Technology 113; Eugene Volokh, ‘Freedom of Speech, Shielding Children, and Transcending Balancing’  Supreme Court Review 141.
 See Henkin, above n 103, 1047:
Balancing is highly appealing. It provides bridges between the abstractions of principle and the life of facts ... It softens the rigors of absolutes, makes room for judgment and for sensitivity to differences of degree ... The flexibility it provides may have been an important ingredient in making judicial review work and rendering it acceptable.
Justice Scalia makes this point about any mode of analysis that, like ad hoc balancing, confers considerable judicial discretion to determine the result on a case-by-case basis: Justice Antonin Scalia, ‘The Rule of Law as a Law of Rules’ (1989) 56 University of Chicago Law Review 1175, 1177.
 Sullivan, above n 124, 66.
 As Schauer explains more precisely, a rule may be over- or under-inclusive with respect to its justification, that is, ‘the evil sought to be eradicated or the goal sought to be served’: Schauer, Playing by the Rules, above n 124, 26, 31–4.
 Schauer, ‘Categories and the First Amendment’, above n 108, 296–8.
 Ibid 299; Nimmer, above n 104, 939; Scalia, above n 135, 1179.
 Nimmer, above n 104, 939: ‘[A]d hoc balancing by hypothesis means that there is no rule to be applied, but only interests to be weighed’; Scalia, above n 135, 1182–3.
 Herbert Wechsler, ‘Toward Neutral Principles of Constitutional Law’ (1959) 73 Harvard Law Review 1.
 See below Part V(C)(3).
 This is Justice Scalia’s view: see Scalia, above n 135; Sullivan, above n 124, 65–6. Schauer refers to the capacity of rules to allocate power among competing decision-makers by constraining the power of those applying the rules: see Schauer, Playing by the Rules, above n 124, 231–2.
 Schauer, ‘Categories and the First Amendment’, above n 108, 299.
 Gerald Gunther, ‘In Search of Judicial Quality on a Changing Court: The Case of Justice Powell’ (1972) 24 Stanford Law Review 1001; Scalia, above n 135, 1178–9.
 See below nn 163–164 and accompanying text.
 Indeed, in setting out the test in Australian Capital Television  HCA 45; (1992) 177 CLR 106, 143 fn 25, Mason CJ cites one of the classic ad hoc balancing cases, Konigsberg USSC 73; , 366 US 36 (1961).
 See Part IV(A) above.
 For a similar analysis see Lorraine Eisenstat Weinrib, ‘The Supreme Court of Canada and Section One of the Charter’ (1988) 10 Supreme Court Law Review 469, 500. Although Professor Weinrib rejects the balancing ‘metaphor’ to explain the drawing of limits on Charter rights, in her view the proportionality test ‘transforms [the Supreme Court’s] idea of the unity of values underlying both rights and limits into a workable, intelligible judicial test’.
 That is not to say that a proportionality test could not be developed, or modified, in a manner that gives this kind of guidance. Indeed, I argue below that such a test could and should be used this way. See below Part V(C)(2).
 Edmonton Journal v Alberta  2 SCR 1326, 1355–6; 64 DLR (4th) 577, 583–4 (Wilson J):
One virtue of the contextual approach ... is that it recognizes that a particular right or freedom may have a different value depending on the context ... It seems to be more sensitive to the reality of the dilemma posed by the particular facts and therefore more conducive to finding a fair and just compromise between the two competing values ...
See also Rocket v Royal College of Dental Surgeons of Ontario  2 SCR 232, 246; 71 DLR (4th) 68, 78 (McLachlin J for the Court):
While the Canadian approach does not apply special tests to restrictions on commercial expression, our method of analysis does permit a sensitive, case-oriented approach to the determination of their constitutionality. Placing the conflicting values in their factual and social context when performing the s 1 analysis permits the courts to have regard to special features of the expression in question.
Despite some cases that seem to contradict this, the contextual approach ‘quickly became the mainstay of a methodology which is now firmly entrenched in precedent’: Jamie Cameron, ‘The Past, Present, and Future of Expressive Freedom under the Charter’ (1997) 35 Osgoode Hall Law Journal 1, 5. See also Sidney R Peck, ‘An Analytical Framework for the Application of the Canadian Charter of Rights and Freedoms’ (1987) 25 Osgoode Hall Law Journal 1, 77:
[J]ustices have considerable leeway when they balance the importance of the purposes of a limit against the severity of the effects of the limit on a right or freedom. The direction to balance these factors contains no indication of the matters judges should consider in the balancing process; it contains no standard to control the way in which judges strike the balance.
 Cameron, above n 151, 5 (emphasis added).
 Kirk, above n 13, 17.
 Australian Capital Television  HCA 45; (1992) 177 CLR 106, 143.
 Even balancing in this manner has attracted some criticism in the United States. In Simon & Schuster Inc USSC 152; , 502 US 105 (1991), Kennedy J, in a separate concurrence, attacked the use of the strict scrutiny test in respect of at least some content-based regulation (although he appears to exempt certain categories of content-based regulation, like defamation laws, from his criticism). Importantly for my purposes, his criticism of this test draws strongly on the criticisms traditionally made of the ad hoc balancing approach seen in Konigsberg USSC 73; , 366 US 36 (1961) and Barenblatt,  USSC 132; 360 US 109 (1959). Indeed, undoubtedly drawing on the similarities between the traditional ad hoc balancing test and tests requiring a compelling interest and narrow tailoring, he refers to the latter as a form of ad hoc balancing:
Here, a law is directed to speech alone where the speech in question is not obscene, not defamatory, not words tantamount to an act otherwise criminal, not an impairment of some other constitutional right, not an incitement to lawless action, and not calculated or likely to bring about imminent harm the State has the substantive power to prevent. No further inquiry is necessary to reject the State’s argument ... the compelling interest and narrow tailoring analysis is ill advised when all that is at issue is a content-based restriction, for resort to the test might be read as a concession that States may censor speech whenever they believe there is a compelling justification for doing so ... use of these traditional categories is preferable to the sort of ad hoc balancing that the Court henceforth must perform in every case if the analysis used here becomes our standard test.
Simon & Schuster Inc USSC 152; , 502 US 105, 124–7 (1991).
 Professor Sullivan makes the same point about the ‘two-tiered’ system of judicial review in American constitutional law: ‘True, these tiers employ nominal balancing rhetoric, but the Court ties itself to the twin masts of strict scrutiny and rationality review in order to resist (or appear to resist) the siren song of the sliding scale’: Sullivan, above n 124, 60.
 See above Part IV(C).
 Pierre J Schlag, ‘An Attack on Categorical Approaches to Freedom of Speech’ (1983) 30 UCLA Law Review 671; Steven H Shiffrin, The First Amendment, Democracy, and Romance (1990) 33–5.
 Nimmer, above n 104, 945. By the same token, the point has been often made that in the right hands, a balancing analysis can be very protective of speech. For example, Harlan J, who preferred a balancing analysis, sometimes accorded speech protection in cases where Black J, perhaps the leading opponent of balancing, denied it: see Street v New York,  USSC 86; 394 US 576 (1969), which reversed the appellant’s conviction for burning an American flag (Warren CJ, Black, White and Fortas JJ dissenting); Cohen v California,  USSC 114; 403 US 15 (1971), which reversed a conviction for disturbing the peace by offensive conduct (Burger CJ, Black and Blackmun JJ dissenting). See generally Gunther, above n 145, 1006: ‘[F]or Justice Harlan, balancing never served as an escape from responsibility. His balancing was always marked by an alert perception of free speech elements in a controversy — a perception often more sensitive than that of devotees of more rigid formulations.’
 Nimmer, above n 104, 945.
 See also John Hart Ely, Democracy and Distrust: A Theory of Judicial Review (1980) 109–16.
 Schauer, ‘Categories and the First Amendment’, above n 108, 303–4. See also Schauer, ‘Codifying the First Amendment’, above n 109, 315–16.
 Most notable was the successful prosecution of leaders of the Communist Party of the United States under the Smith Act s 2, 54 Stat 671 (1940), for conspiracy to advocate and to organise a group to advocate, the overthrow of the government of the United States: see Harry Kalven Jr, A Worthy Tradition: Freedom of Speech in America (1988) 191–210.
 This reasoning is behind Professor Ely’s defence of ‘clearly and narrowly defined categories’: Ely, above n 161, 110. Ely also states at 116 (emphasis added) that:
Allowing people to assault our eardrums with outrageous and overdrawn denunciations of institutions we treasure will inconvenience, annoy and infuriate us on occasion ... [that’s] exactly the price we shouldn’t think twice about paying ... In 1980 most people who have thought about the issue appreciate this. The hard part will be to sustain that appreciation through our future periods of actual or perceived crisis. Maybe we won’t be able to, but we increase the chance by using today to build protective barriers around free expression as secure as words can make them.
For the same kinds of reasons, the construction of doctrines that leave comparatively little discretion in the hands of future decision makers is part of Professor Blasi’s recommended strategy for ensuring that First Amendment protection survives periods of intolerance: Vincent Blasi, ‘The Pathological Perspective and the First Amendment’ (1985) 85 Columbia Law Review 449, 474.
 In this regard it is worth noting that the adoption of rules does appear to have made a difference to the results in First Amendment cases. The adoption of rules seems to have brought with it a greater level of First Amendment protection than was the case under the ad hoc balancing test. Thus, in the United States, a rules-based approach does appear to have constrained the discretion that judges earlier exercised in favour of law-makers: Schauer, ‘Categories and the First Amendment’, above n 108, 304.
 However, the insistence on text and structure over other values to interpret the freedom is itself a theoretical or philosophical commitment: see above n 35 and accompanying text.
 See Abrams v United States,  USSC 206; 250 US 616, 630 (1919) (Holmes J): ‘[T]he best test of truth is the power of the thought to get itself accepted in the competition of the market.’ See generally Kent Greenawalt, ‘Free Speech Justifications’ (1989) 89 Columbia Law Review 119, 130–41, 145–6; Schauer, Free Speech: A Philosophical Enquiry, above n 11, 15–34.
 See above nn 159–163 and accompanying text.
 Frederick Schauer, ‘Fear, Risk and the First Amendment: Unravelling the “Chilling Effect”’ (1978) 58 Boston University Law Review 685, 685.
 It is striking that in some cases Harlan J, employing a balancing approach, reached a result protective of speech, over the dissents of colleagues like Black J, who are commonly associated with the supposedly more protective ‘absolutist’ approach: Cohen v California,  USSC 114; 403 US 15 (1971); Street v New York,  USSC 86; 394 US 576 (1969); Smolla, above n 106, 40; Gunther, above n 145, 1106–11; Schauer, ‘Categories and the First Amendment’, above n 108, 303–4.
 Smolla, above n 106, 41.
 In his famous dissent in Abrams v United States,  USSC 206; 250 US 616, 630 (1919), Holmes J wrote that ‘[p]ersecution for the expression of opinions seems to me perfectly logical. If you have no doubt of your premises or your power and want a certain result with all your heart you naturally express your wishes in law and sweep away all opposition’.
 Smolla, above n 106, 41.
 This is likely to be often because First Amendment issues are likely to arise when a kind of speech is sufficiently unpopular as to have become the subject of a repressive law. Nimmer, above n 104, 939–40; Henkin, above n 103, 1048. See also Scalia, above n 135, 1180.
 Nimmer, above n 104, 940–1 referring to Frankfurter J’s statement in Dennis v United States, 341 US 494, 525 (1951), that: ‘Primary responsibility for adjusting the interests which compete in the situation before us of necessity belongs to the Congress ... We are to set aside the judgment of those whose duty it is to legislate only if there is no reasonable basis for it.’ See also Smolla, above n 106, 41; Scalia, above n 135, 1179–80.
 Freedom of expression is protected under s 2 of the Canadian Charter of Rights and Freedoms.
 Perhaps most notable has been the Supreme Court’s stance on the regulation of hate speech and pornography: R v Keegstra  INSC 224;  3 SCR 697; 2 WWR 1; R v Butler  1 SCR 452; 89 DLR (4th) 449. See Greenawalt, above n 107, 66.
 Hill v The Church of Scientology  2 SCR 1130, 1182–5; 126 DLR (4th) 129, 165–8. See also Cameron, above n 151, 39.
 See Owen Fiss, ‘Free Speech and Social Structure’ (1986) 71 Iowa Law Review 1405; Cass R Sunstein, The Partial Constitution (1993) 203–13.
 R v Oakes  1 SCR 103, 139–40; 26 DLR (4th) 200, 227; Irwin Toy Ltd v Quebec  1 SCR 927, 991–1000; 58 DLR (4th) 577, 624–30. See above nn 48–54 and accompanying text.
 See Cass R Sunstein, Legal Reasoning and Political Conflict (1996) 35–61.
 See above Parts II(A)(2) and II(B).
 Geoffrey R Stone, ‘Restrictions of Speech Because of Its Content: The Peculiar Case of Subject-Matter Restrictions’ (1978) 46 University of Chicago Law Review 81, 104; but see Kenneth L Karst, ‘Equality as a Central Principle in the First Amendment’ (1975) 43 University of Chicago Law Review 20, which argues that the First Amendment is governed by an equality principle that prevents distinctions based on content.
 See Sunstein, Legal Reasoning and Political Conflict, above n 181, 35–62. Sunstein argues for the preference for common law case-by-case reasoning over more ambitious styles of reasoning that more overtly advance more ambitious theoretical justification. These arguments support preference for an incompletely theorised test like proportionality over a rules-based approach where decisions as to underlying values must be more quickly announced. That is not to deny that the common law approach itself has some limitations. See Michael C Dorf, ‘Foreword: The Limits of Socratic Deliberation’ (1998) 112 Harvard Law Review 4, 26–50, arguing that the common law method is not able to adapt statutory and constitutional text to modern conditions adequately because innovations under the common law method are inspired by speculation and intuition of judges whose abilities to understand and predict the actual effects of their decisions are limited. It is therefore unlikely to produce effective solutions in a complex modern world.
 Adrienne Stone, ‘Incomplete Theorizing in the High Court — Review Essay: Cass R Sunstein, Legal Reasoning and Political Conflict’  FedLawRw 8; (1998) 26 Federal Law Review 195, 204.
 Ibid 202–3.
 See above n 75 and accompanying text.
 Australian Capital Television  HCA 45; (1992) 177 CLR 106, 143 (Mason CJ), 234–5 (McHugh J); Levy  HCA 31; (1997) 189 CLR 579, 619 (Gaudron J), 645 (Kirby J).
 Nationwide News  HCA 46; (1992) 177 CLR 1, 76–7.
 Levy  HCA 31; (1997) 189 CLR 579, 614.
 Ibid 638–42 (Kirby J).
 Sunstein, Legal Reasoning and Political Conflict, above n 181, 63 (emphasis in original). Some of Sunstein’s critics claim he fails to acknowledge this sufficiently: see Dorf, above n 184, 32; Larry Alexander, ‘Incomplete Theorizing: A Review Essay of Cass R Sunstein’s Legal Reasoning and Political Conflict’ (1997) 72 Notre Dame Law Review 531, 535–8.
 I am assuming for the purposes of this example that this law regulates protest against a Commonwealth legislative policy, to eliminate the possibility suggested by Brennan CJ and McHugh J in Levy, that the freedom of political communication might not protect discussion of state political matters unless it is possible to make a specific connection between the discussion of state matters and federal political matters. Although Brennan CJ and McHugh J rested their decisions on other grounds, both required that the discussion protected by the freedom of political communication must have some significance for federal matters and both were sceptical that discussion of the Victorian duck hunting regulations could have federal significance: Levy  HCA 31; (1997) 189 CLR 579, 596 (Brennan CJ), 626 (McHugh J). See also Stone, ‘Lange, Levy and the Direction of the Freedom of Political Communication’, above n 3, 129.
 See above nn 147–149 and accompanying text.
 See above Part III(A)(1).
 For example, the type and time of protest, and the distance from the Parliament within which protest was prohibited. Clearly a blanket prohibition on protest in any part of the parliamentary precinct is less likely to be valid than a law just aimed at some kinds of protest activity, in some places, at some times of the day.
 See above n 75. I have criticised the rather unthinking adoption of such a rule, but this criticism would not apply to a more careful, thoughtful development of a rule with such modifications, if any, as are appropriate to the Australian context.
 Sunstein, Legal Reasoning and Political Conflict, above n 181, 78: ‘Much of the common law has rulelike features, even if the governing law has emerged through analogies and encounters with particulars.’ See also Dorf, above n 184, 32, for although Professor Dorf criticises Sunstein’s account for its failure to acknowledge adequately the role of theory in common law reasoning, he shares the view that common law reasoning can produce relatively defined rules: ‘The pattern of decided cases crystallizes in doctrines that both exemplify and justify the underlying pattern.’
 The power to synthesise existing precedent is perhaps the great achievement of Lord Atkin’s famous dictum in Donoghue v Stevenson  AC 562, 580, laying the basis for the modern Anglo-Australian law of negligence. His formulation of the basis for actions in negligence rationalised and explained the principle underlying the particular categories of cases in which the law recognised a general duty of care. See R P Balkin and J L R Davis, Law of Torts (2nd ed, 1996) 200. See also Dorf, above n 184, 43–4.
 See Melvin Aron Eisenberg, The Nature of the Common Law (1988) 72–4.
 See Scalia, above n 135, 1179. See also Tom D Campbell, The Legal Theory of Ethical Positivism (1996).
 See Richard H Fallon Jr, ‘“The Rule of Law” as a Concept in Constitutional Discourse’ (1997) 97 Columbia Law Review 1, 8, 14–17; Scalia, above n 135, 1178–80.
 See Fallon, above n 202, 8–9; Scalia, above n 135, 1178.
 See above Part V(C)(3).
 See Fallon, above n 202, 11–14.
 The High Court’s commitment to interpretation according to constitutional text and structure expounded in Lange is thus based in some value external to the Constitution itself. McHugh J, at least, acknowledges this: see above n 35.
  HCA 54; (1920) 28 CLR 129.
  HCA 46; (1994) 182 CLR 104, 197–8.
  HCA 54; (1920) 28 CLR 129, 142.
 The capacity of rules to constrain others varies according to the decision-making context, depending upon some external standard that accords weight to rules. See generally Schauer, Playing by the Rules, above n 124, 118–34. Thus the capacity of the High Court to constrain others will differ in different contexts. Considering just the power of the Court’s rules to constrain other judges, it must be seen that the force of the rules formulated by the High Court will be greatest in lower courts, especially courts of first instance. In such cases, judges are compelled by the possibility of appeal to the High Court to show great deference to doctrines it formulates. The capacity of rules to constrain would be less in the High Court itself, because the Court does not regard itself as bound by its previous decisions. However, the Court’s reluctance to depart from its previous decisions gives such doctrines at least some force.
 Fallon, above n 202.