‘The spectre was raised in argument of a Government seeking to rule without Parliament or, at the least, dispense with its sitting for very lengthy periods. … We do not believe that it is helpful to consider the arguments by reference to extreme hypothetical examples…’ R (Miller) v Prime Minister [2019] EWHC 2381, [65]-[66] (The Lord Burnett of Maldon CJ, Sir Terence Etherton MR and Dame Victoria Sharp P) (emphasis added)

‘The sovereignty of Parliament would, however, be undermined as the foundational principle of our constitution if the executive could, through the use of the prerogative, prevent Parliament from exercising its legislative authority for as long as it pleased. … In our view, it is no answer to these points to say, as counsel for the Prime Minister argued, that the court should decline to consider extreme hypothetical examplesR (Miller) v Prime Minister [2019] UKSC, [42]-[43] (Lady Hale and Lord Reed giving the judgment of the Court) (emphasis added)

What is the proper role of hypotheticals in constitutional law? Harvard Professor Adrian Vermeule recently commented that ‘[a] sign of intellectual maturity in law, theology and morals is not to argue by means of extremely remote, low-probability hypotheticals’. But some of the most senior judges in the United Kingdom recently divided on this question, in assessing the lawfulness of Prime Minister Boris Johnson’s attempt to prorogue the UK Parliament last August. In Australia, it is well established that the Commonwealth Constitution should not be construed by reference to ‘extreme examples and distorting possibilities’ (NSW v Commonwealth (2006) 229 CLR 1, [188] (‘Work Choices’)). Yet the High Court still analyses constitutional issues by asking, for example, whether a State could abolish its entire court system (Kable v DPP (NSW) (1996) 189 CLR 51) or create an executive body modelled on the ‘ill-fated Inter-State Commission’ (Burns v Corbett (2018) 92 ALJR 423, [100]).

So how are hypotheticals used and abused in Australian constitutional law? Two principles emerge from the case law.

The first principle

The first principle is: the possibility that a power might be abused does not entail that the power itself is unconstitutional. Consider Forge v Australian Securities and Investments Commission (2006) 228 CLR 45. The key question was whether a law giving the executive an apparently unlimited power to appoint acting judges undermined the institutional integrity of the New South Wales Supreme Court, contrary to the Kable doctrine. A majority of the High Court found that it did not. It was easy to imagine how the executive might exercise this power to undermine the Court’s integrity (e.g. by letting permanent judges retire and populating the Court with acting judges who might be more susceptible to political pressure). But this did not entail that the power itself was unconstitutional. As Gleeson CJ noted: ‘Possible abuse of power is rarely a convincing reason for denying its existence’ (at [46]).

The same issue arose more recently in Spence v Queensland [2019] HCA 15. One of the issues was whether the Commonwealth Parliament had an implied exclusive power in respect of federal elections. One strand of the Commonwealth’s argument rested on the following premises:

  • The Constitution aims to ‘secure Commonwealth control over the entire subject matter of federal elections’.
  • If the States and the Commonwealth shared legislative power in respect of federal elections, there might be ‘diverse State-by-State regulation’ contrary to that constitutional objective.
  • Practically speaking, the Commonwealth might be unable to invalidate, by legislation operating with s 109 of the Constitution, inconsistent State laws of that kind.

In support of (3), the Commonwealth suggested that ‘a State could wait until the Commonwealth Parliament had been prorogued in advance of an election, and then enact legislation that may have a profound impact on the result in the pending federal election’. That ‘possibility’ could be avoided only by accepting the Commonwealth’s claim of exclusive legislative power.

Justice Edelman expressly rejected this argument. Even assuming that the Constitution contemplates a ‘uniform federal scheme’ for federal elections, the Commonwealth could, for all practical purposes, achieve that end through s 109. It was wrong to use ‘extreme and unlikely examples’, like the risk of sneaky State legislation while the Commonwealth Parliament was prorogued, to test whether it was necessary for the Commonwealth to have exclusive legislative power over federal elections (at [304]).

We can generalise the kind of argument rejected in Forge and in Spence in the following way:

  • If the government has power to do X, then it is possible that it might exercise that power to do Y.
  • Y would be constitutionally unacceptable.
  • Therefore, it is unconstitutional for the government to have the power to do X.

One problem with this form of argument is that it proves too much. Almost every power carries with it a risk of abuse. In Forge, for example, Gleeson CJ noted (at [20]) that ‘the possibility of abuse of the power to appoint permanent judges is just as obvious as the possibility of abuse of the power of appointing acting judges’. Yet it would be absurd to suggest that the executive’s apparently unlimited power to appoint permanent judges is invalid. This form of argument could render many ordinary powers unconstitutional.

A related problem is that this form of argument is highly risk averse. It asks the court to strike down a law not because it is actually unconstitutional, but because it creates even a negligible constitutional risk. It thus places too little emphasis on allowing repositories of constitutional power to exercise that power, and too much on ensuring that they do so within constitutional limits.

Finally, this form of argument rests too heavily on the hypothetical. In several cases, Gageler J has emphasised that a court should not prefer one interpretation of a statute over another ‘merely to avoid constitutional doubt’ (NAAJA v Northern Territory (2015) 256 CLR 569, [76]). Otherwise, the court would be free to develop constitutional principles without actually having to apply them to invalidate a particular law: an anomaly in a system which rejects advisory opinions. For similar reasons, a court should not invalidate a law merely because of a ‘judicially articulated doubt as to validity’ (NAAJA at [78]) based on a vanishing possibility. Such an approach would be in tension with our constitutional tradition which requires ‘particular answers to particular questions arising in a live controversy between parties’ (Singh v Commonwealth (2004) 222 CLR 322, [152] (Gummow, Hayne and Heydon JJ)).

The second principle

The second principle is: it is permissible to use hypotheticals to draw out a constitutional rule which might render the actual exercise of power in issue unconstitutional. Consider the following two examples.

In Kable, the NSW Parliament had empowered the Supreme Court to make orders for the continuing detention of Gregory Kable, a convicted murderer who was nearing the end of his sentence. Mr Kable argued that the law was unconstitutional.

To test this proposition, McHugh J began with two broad hypotheticals. First, could a State abolish its court system altogether? No. If it could, ‘the powers conferred by s 77(ii) and (iii) would be rendered useless and the constitutional plan of a system of State courts invested with federal jurisdiction, as envisaged by Ch III, would be defeated’ (at [7]). Second, could a State abolish its Supreme Court? Again, no. Such a power would render ‘nugatory’ the right of appeal to the High Court under s 73(ii) (at [8]). These hypotheticals illustrated the underlying constitutional principle: a State cannot ‘alter or undermine the constitutional scheme set up by Ch III of the Constitution’ (at [17]), including the existence of a Supreme Court and a system of State courts capable of exercising federal jurisdiction. Justice McHugh went on to show how the impugned legislation fatally compromised the Supreme Court’s impartiality, and thus undermined the ‘constitutional scheme’ of Ch III.

In Roach v Electoral Commissioner (Cth) (2007) 233 CLR 162, the question was whether the Commonwealth Parliament could exclude all prisoners from voting in federal elections. Chief Justice Gleeson began by holding that the words ‘directly chosen by the people’ in ss 7 and 24 of the Constitution now mandate universal adult suffrage. But universal suffrage seems to admit of certain exceptions. How do we determine the limits of Parliament’s capacity to define those exceptions? Chief Justice Gleeson considered some examples. On the one hand, it would be unacceptable for Parliament to ‘disenfranchise people on the ground of adherence to a particular religion’ and it was ‘easy to multiply examples of possible forms of disenfranchisement that would be identified readily as inconsistent with choice by the people’ (at [8]). On the other hand, it seems tolerable for Parliament to link the franchise to citizenship. The constitutional principle underlying these intuitions, Gleeson CJ held, is that the exclusion must ‘have a rational connection with the identification of community membership or with the capacity to exercise free choice’ (at [8]). Because the blanket disenfranchisement of prisoners did not have a rational connection with either of these ends, it was unconstitutional.

We can generalise the argument used in Kable and in Roach in the following way:

  • If the government has power to do X, then it is possible that it might also have power to do Y.
  • Y would be constitutionally unacceptable.
  • The rule underlying (2) is A.
  • If the government did X, it would contravene A.
  • Therefore, it is unconstitutional for the government to do X.

One potent response to this form of argument is that the relevant hypothetical is not unconstitutional but simply undesirable, and as such is to be guarded against by political rather than legal means. In Roach, for example, Heydon J noted that the plaintiff’s argument ‘contained many assumptions as to whether it would be possible now to narrow the franchise on the basis of race, age, gender, religion, educational standards or political beliefs’. He concluded that ‘[i]t is enough to say that narrowing the franchise in any of these ways may be highly undesirable; it does not follow that it is unconstitutional’ (at [179]).

This response goes back at least to the Engineers Case (1920) 28 CLR 129. The doctrines of implied intergovernmental immunities and reserved state powers, according to Isaacs J, were based on fears that unrestricted constitutional powers ‘might be abused to the point of destruction’.  But ‘the extravagant use of the granted powers in the actual working of the Constitution is a matter to be guarded against by the constituencies and not by the Courts’.

The High Court mobilised the same response in the First Territorial Senators Case (1975) 134 CLR 201. The Commonwealth legislated for the Northern Territory and the Australian Capital Territory to each have two representatives in the Senate. Several States argued that this law was not supported by the territories power (s 122) because, among other reasons, a broad reading of the power would enable the Commonwealth to swamp the Senate with Territorial senators, diluting the power of the States.

The majority rejected this argument. As Mason J noted, the States were ideally placed to defend their interests through the Senate itself, as a deluge of Territorial senators would require ‘the willing participation of the senators representing the States in such an enterprise, notwithstanding that it would hasten their journey into political oblivion’ (at [25]). Justice Jacobs similarly concluded that ‘Parliament itself was the safeguard against the absurd possibility’ (at [6]).

The High Court gave the same response in Work Choices. The plaintiffs argued that a broad construction of the corporations power (s 51(xx)) might distort the ‘federal balance’, by enabling the Commonwealth to create a new system of libel laws applicable only to newspapers owned by corporations, and a new system of liquor licensing applicable only to venues owned by corporations. The majority held that these were merely ‘possible social consequences that it is said could flow if further legislation is enacted, and which it is said are to be seen as absurd or inconvenient’ (at [188]). They could not ground any constitutional principle limiting Commonwealth legislative power.

Legal and political questions

How do we determine whether a hypothetical raises a legal matter or simply a political one? This question is made more difficult by the fact that, as Dixon J pointed out in the Melbourne Corporation Case (1947) 74 CLR 31, the Constitution is a ‘political instrument’. Virtually every question arising from the Constitution is ‘political’ in some sense.

There are at least two ways to answer this question. The first is to examine the constitutional text and structure. If the hypothetical is closely grounded in the constitutional text and structure, then it is more likely to have legal, rather than merely political, significance. Drawing on the examples above, the Constitution expressly contemplates the continued existence of State Supreme Courts, but says nothing about a system of corporation-specific libel laws. This supports the conclusion that the High Court is responsible for ensuring the former, as in Kable, while the latter can be left to the ordinary political process, as in Work Choices.

The second way is to examine the effectiveness of the political checks against the hypothetical scenario. In the First Territorial Senators Case, the political process was clearly capable of protecting the States’ interests in the Senate, because the States’ representatives had a decisive voice in that process. The situation was different in Roach. Political checks on disenfranchisement are weak because those affected have no direct political power, by definition, and tend to be ‘unwanted by, or inconvenient to, those who currently form majorities’ (Murphy v Electoral Commissioner (2016) 261 CLR 28, [95]). This counted in favour of a legal check on Parliament’s power to define the franchise.


To conclude, let’s return briefly to the hypothetical raised by the plaintiff in R (Miller) v Prime Minister. The UK Government contended that Johnson’s decision to prorogue Parliament was non-justiciable. The plaintiff responded with a hypothetical: what if the Prime Minister ‘were to advise prorogation for 6 months or 1 year or longer’? What if he were to do so for the express purpose of frustrating parliamentary sovereignty?

The first principle discussed above does not apply to this argument. The plaintiff was not making an argument from constitutional risk, i.e. ‘because the power to prorogue is prone to abuse, the power itself is unconstitutional’. Rather, the plaintiff was using the hypothetical to draw out an underlying principle: it is unconstitutional for the executive to seriously frustrate Parliament’s exercise of its constitutional functions without reasonable justification. Because Johnson’s actual decision to prorogue Parliament also violated that underlying principle, so the plaintiff argued, it should be held invalid. This is an application of the second principle.

The High Court and the Supreme Court disagreed over, among other things, whether the plaintiff’s hypothetical raised a legal or merely political matter. Of course, the UK has no codified constitution to guide this assessment. The High Court suggested that there may be effective political checks on an extreme prorogation: ‘it is impossible to predict how the flexible constitutional arrangements of the United Kingdom, and Parliament itself, would react in such circumstances’ (at [66]). But this is speculative and unconvincing. First, once it occurs, prorogation by definition disables Parliament from checking the executive. Second, Parliament may have little or no time to act affirmatively before prorogation occurs. Third, the majority in Parliament may well decide not to intervene, leaving the minority unable to perform its constitutional function of holding the executive to account, and without any direct political power to change that state of affairs. It was clearly open to the Supreme Court to see the plaintiff’s hypothetical as raising a matter of constitutional principle warranting the court’s intervention.

Jack Maxwell is a public lawyer from Melbourne, now based in London.

Suggested citation:  Jack Maxwell, ‘Extreme Examples in Constitutional Law’ on AUSPUBLAW (05 February 2020) <>