Unconstitutionality in the Voice Amendment

The constitutionality of the Voice referendum turns in the first place on a fundamental distinction as to the way we relate to entities in the world: we can find and recognise them, or we can design and build them. We can find and recognise the Voice, or we can design and build it.

The draft referendum confounds this duality in a large ambivalence. The draft says, “There shall be a body to be called the Aboriginal and Torres Strait Islander Voice”; but is this an assertion in the present tense (as in “there shall right now, on the referendum passing, be a Voice”) or is it future-directed: there shall be a Voice when the Parliament makes laws “with respect to … the composition, functions, powers and procedures” of the Voice? As an assertion of a present Voice this statement can only be a recognition by the amendment process of a prior existing Voice—no design or building could possibly be done in that moment. But in an extended second moment there is design and building; and the designer and builder is not the referendum process but the Parliament. If this second reading is what is intended, the referendum is unconstitutional: the design and building of the Voice are not done by the referendum process but by unlawful delegation to the Parliament.

The simple present voice of the first moment is a Voice to Parliament (the slim Voice) and the Parliamentary design and building of the Voice in the second moment is the Parliamentary Voice to Parliament (the fat Voice). Each includes the Executive. Were the referendum limited to the slim Voice to Parliament it would be constitutionally valid. But if it carries Parliament’s design and building in the second moment—a long moment, for the Prime Minister is proposing a six-month design and building process after the constitutional amendment is passed—it will be invalid for unlawful delegation.

The referendum presents two draft formulations: first, that which is to go into and be part of the Constitution, and second, the short question that electors will find on their ballot paper. I discuss the short question at the end of this paper; it also robs the sovereign people of their right to scrutiny of the referendum question.

The constitutional formulation has three parts. The first two parts are: (1) There shall be a body to be called the Aboriginal and Torres Strait Islander Voice, and (2) The Voice may make representations to Parliament and the Executive Government on matters relating to Aboriginal and Torres Strait Islander peoples. These are the basic referendum propositions and create a simple autonomous and special entry for Aboriginal and Torres Strait Islander peoples into Australian governance.

However, the third proposition in the draft proposal, the unconstitutional one, gives Parliament power “to make laws with respect to matters relating to the Voice, including the composition, functions, powers and procedures” of the Voice. This is a voice to be designed and built by Parliament, and is an unconstitutional delegation of the amendment power. It is a fat Voice. The constitutional problem with this delegation I shall now explain.

The Final Report of the Referendum Council stated (on page 36): “No one has suggested there be an attempt to enshrine in the Constitution provisions of the kind more appropriately left to Parliament.” Well, that is undoubtedly true for matters of detail, which can—indeed, must—be delegated. But we are talking about the definition of the Voice that “shall be”—the absolutely central question which clearly must be an issue for the referendum itself and must be “enshrined” in the Constitution if passed.

The reason for the invalidity of the fat Voice to Parliament is that it is unconstitutional for Parliament to be given the charge of the definition of the Voice. The referendum would in effect be delegating its power to Parliament in a way that keeps the terms of the definition of the Voice hidden from perusal in the referendum process. More than that, it would be asserting an alternative method of constitutional amendment, and a much less exacting one than the current amendment power of section 128.

The High Court can be expected to be scrupulous in its protection of the integrity of the sovereign amending power of the Constitution. Section 128 reads (in part):

The proposed law for the alteration thereof must be passed by an absolute majority of each House of the Parliament, and … shall be submitted in each State and Territory to the electors qualified to vote for the election of members of the House of Representatives.

To be submitted to the electors the law must be a valid law, and our law cannot be valid because in breach of the constitutional amendment power it delegates the substance of the proposed amendment to the Commonwealth Parliament. The validity of this is a matter arising under the Constitution and involving its interpretation (Judiciary Act, section 30). It is highly likely that the High Court, if asked, would exercise this jurisdiction, and so hold. I describe the likely reasoning shortly. But I should now explain “invalidity”. The High Court invalidates nothing. If it finds that the delegation to Parliament is unlawful then (a) it is of no valid constitutional effect and (b) there can be no valid decision by the referendum process in place.

The prohibition of a delegate delegating is a fundamental principle of law—so fundamental that it has a Latin formulation: delegatus non potest delegare. Its point is obvious enough. The founders of the Constitution (the people) could have made the Constitution something to be amendable by Parliament. But they didn’t; they established the somewhat complex process of section 128. To allow a delegation to Parliament by that latter process would cancel the choice the founders actually made.

The principle is not against delegation: it is against a delegate delegating. The preclusion of a delegate delegating logically supposes two prior elements: a foundational power and a first delegation by that power. Upon that basis one first-delegated cannot delegate further. Is the referendum process itself a first-delegate? Yes, it is. And the prior foundational power is: “WHEREAS the people of New South Wales, Victoria, South Australia, Queensland, and Tasmania … have agreed to unite in one indissoluble Federal Commonwealth …”; and, my projection adds, have delegated our joint power of amendment to the process described in section 128.

Delegation would only be allowed to the section 128 process if it were the foundational power of the Constitution. It is not: section 128 is part of the Constitution, not its foundation.

The people who founded the Constitution were establishing a federation; but the further delegation of the referendum power removes the protection that section 128 gives to the federation when it requires that a majority of at least four states support a proposal (protecting the smaller states from the larger). No such provision applies to the passing of laws by Parliament, and so when the referendum issue is delegated to Parliament that fundamental federal protection is by-passed. This makes it unlikely in the extreme that the creation (people’s delegation) of the section 128 power contemplated a delegation to Parliament.

(Can the process of section 128 be used to amend itself? This is a difficult question—a version of the liar paradox—but is beyond the scope of this paper. The Voice referendum does not propose to amend section 128.)

Lawyers will know that the colonial parliaments could in their latter years delegate. This is because they were approaching independence, which implies their becoming foundational powers and so losing their status as delegates—it is only delegates who cannot delegate. Everything changed in 1901: the people became the foundation of the nation (called Commonwealth) and its Constitution, and they first-delegated to the section 128 process, which therefore can delegate no further.

The issue of delegation has been an active one in the politics of the referendum over the last few months. The Prime Minister, leading the Yes case, has constantly refused to discuss the details of the delegation. There is considerable evidence that the delegation is a deliberate attempt to limit the knowledge that the people bring to the ballot box, lest they learn of matters that might worry them. Now, there will be many electors who know this but are willing to accept it. Such sloppiness of thought is their privilege. But this is not a privilege that transfers itself into the duty of the High Court.

Why have the lawyers who devised the Prime Minister’s referendum draft not seen the obvious? There are lawyers who seeing the (true) fact that there are no precedents on the issue therefore conclude that there is no law governing the process. But the mistake of more serious lawyers seems to have been to think that since the referendum process is the ultimate constitutional power it can do anything at all (and so no questions can be asked about the validity of delegating to Parliament the essence of the question—the definition of the Voice). The amending power is indeed ultimate, but the one thing it cannot do is cheat on its own process. And it’s worse than this.

Suppose Parliament declined to exercise the referendum’s delegated power, or a later Parliament repealed a former Parliament’s legislation, there would be no Voice. Yet the Constitution would continue to say: “there shall be a Voice”. “There shall be” would now lack meaning. These words do not mean: “there might be a Voice” or “we hope there will be a Voice”.

So, what has been delegated to Parliament is not just the power to design and build the Voice, but also the power to cancel the constitutional amendment by depriving it of its meaning. This is not just a delegation of the Voice power; it is a delegation of the power to cancel a referendum result. There’s no way that could possibly be valid.

It follows that the constitutionality of the Voice referendum now turns on our recognising the Voice itself rather than giving power to Parliament to design and build it. We recognise the Voice by hearing it.

Designing and building a Voice is easy to describe, notwithstanding its unconstitutionality. But what is this body that right now might be? Surely such a state of being requires further explanation? No, it doesn’t, and it’s a pity anyone ever thought it did. The history of the Voice is the point, not an explanatory description. The Voice has with determined power made its way from Uluru to the Constitution (as it were, knocking at the door). “There shall be a Voice in the Constitution” means simply that the Constitution hears the Voice, and this means its powers, legislative and executive must hear the Voice, too. In this way, it makes no sense not to include the executive power.

And so does the judicial power hear the Voice. The judicial power heard the Voice in Mabo. And heard it again in Love where the High Court held that the first peoples could not be deported—not separated from their land (their religion). They are in the land, inseparable from it, and the next question is, can they be heard?

In Gulliver’s Travels Jonathan Swift described several imaginary peoples. They did not exist and therefore were designated merely by description. However, when we refer to a real country/people—Fiji/Fijians, say—we designate it by the proper name, Fiji. This is not just a point of grammar, or of lists for the naming of children: it is the most fundamental distinction in any language—the distinction between description and reality. Of any description, and “unicorn” is the usual example here, it might or might not exist; but the proper-naming of something recognises its real existence.

Power in our Constitution is usually defined by description—lighthouses, marriage, trade, courts, elections, taxation, interstate commission—wherever you like to look you see descriptive names coming into the Constitution. And once in the Constitution these items of power develop proper names—the lighthouse power, the Federal Court, the Interstate Commission. But the Voice is a different entity entirely—a remarkable constitutional insight—it comes into the Constitution as a proper name, not a description. This gives it a membership of the Constitution and an intimacy with those other members designated by historical proper name. I refer to: “WHEREAS the people of New South Wales, Victoria, South Australia, Queensland, and Tasmania … have agreed to unite in one indissoluble Federal Commonwealth…” All the states (specifically, the people of the states) are historical proper names. The rest of the Constitution isn’t. The Voice as proper name is an equivalent historical event. It is the entry of the first peoples of Australia into the Constitution, having made their way from the beginning and latterly through Uluru to join the settler peoples.

This is an entry that some would say is two centuries late; but it is an entry, voluntary on both sides, to a better relationship for all.

The wording of the proposal “there shall be a Voice” uses the indefinite article which normally indicates a description. However, what the referendum is saying is: as well as the proper-named Voice of Uluru there will be a constitutional voice immediately recognising it. And when this joinder is effected, the two Voices will be represented in one definite article (“the Voice may make representations”). For us the history is clear and understood, and no more necessary to be set out than it would be for, say, the words “New South Wales, Victoria …” in the Constitution’s preamble to be explained. And nothing could be so effective in carrying the autonomy of the first peoples. The meaning of “there shall be a Voice” is historical rather than lexical.

I have in the discussion so far left the body out of the Voice (“there shall be a body to be called the … Voice”). What I think I have shown, however, is the prior historical embodiment of the Voice. We may welcome “body” back.

And we may contrast Constitution section 101’s only other use of “there shall be”. “There shall be,” it says, “an Inter-State Commission”. Were one appointed it would be proper-named. But unlike the Voice it does not come into the Constitution as proper-named.

The Constitution brands the Aboriginal and Torres Strait Islander peoples as a race for whom there can be special laws—the races power, section 51 (xxvi), a deplorable power that should have been removed entirely in 1967 but was kept in order to pass laws that benefit the first peoples. They are a land-religion culture more than a race, but the Constitution is stuck with the word race.

The fat Voice to Parliament—anything Parliament can say in its designing the Voice—will be racially discriminatory, as many critics have said, because it picks out a certain race. But when we look beyond Parliament to the first peoples’ Voice as a proper name, there is no racial discrimination. Race is a description; the proper-named Voice is not. A proper name applied to a human being identifies them essentially as an individual beneath whatever descriptive qualities they might have. Two people might have the same description, but their proper names distinguish this one from that (as Russell said, “this” and “that” are themselves proper names). And the proper-named Voice does just that for the set of individuals making up the historical first peoples. The proper-named Voice is an historical entity, and all history is fundamentally proper-named.

There will be subsidiary legislative provisions (regulations) required concerning the Voice once established. But the power for these does not need to be explicit in the referendum. They can be special laws under the races power, section 51 (xxvi). Or, better, they can be laws under the incidental powers attaching to the Parliament and the Executive—specifically to their openness to the Voice’s representations.

The Voice is established in the Constitution and by that status becomes part of its sovereignty. This implies the cession of any other sovereignty.

The elevation of the Voice to constitutional status and the entry of its people into the Constitution must be seen as the termination of any claim to the first peoples’ original sovereignty and therefore also to the pursuit of a treaty. There is one sovereignty, the Constitution and its power of amendment; and in “there shall be a Voice” the Constitution is inviting the first peoples into that sovereignty—a beautiful idea overcoming the tired and expired idea of a treaty. That there is one sovereignty means that the first peoples’ accession to the constitutional Voice is their cession of their original sovereignty—a cession which becomes the historical stuffing of the Voice. The Voice section of the Australian Constitution will be an important part of Australian constitutional sovereignty. But this will be a merged and therefore unified sovereignty. Any treaty between A and B when A is part of B—a treaty within a constitution—is a logical nonsense. Treaties are between sovereign, that is independent, nations. This does not mean that first peoples lose by the Voice the possibility of authentic lives: it is the point of a sovereign nation to guarantee that authenticity for all. As Hobbes said, the loss of sovereignty is a descent into the war of all, where no one’s authenticity survives.

The cession of first peoples’ sovereignty into a merged and unified Australian sovereignty is a function of the Constitution and its amendment power. It is not a function of Parliament, which is answerable to the sovereignty, not part of it. This fundamental fact is completely muddled in the contradictory situation in which the Constitution says “there shall be a Voice” but the delegated Parliament denies it. The Parliament, not itself being sovereign, has neither power to give sovereignty nor deny it. This is another reason why the delegation of the Voice amendment to the Parliament is unconstitutional.

The Voice can be regulated, I have said, by laws under the races power, section 51 (xxvi), or by laws that are incidental to the existence of the Voice or incidental to the regulation of its two functions: a Voice to Parliament and to the Executive. However, this regulation by Parliament would have to be true to the constitutional status of the Voice and not detract from it—neither take from it nor load it with fanciful “invasion” politics. This does not mean that Parliament’s legislation cannot control deleterious consequences of Voice activity. In fact, Voice activity would very much be something as to which it might be in the terms of section 51 (xxvi) “deemed necessary to make special laws”. The Voice is part of the single (merged) Australian sovereignty; it is not the whole of it and has duties of moderation to the rest of the sovereignty (just as the Australian states have duties to each other). If the referendum is passed it will be because the existing sovereignty invited the Voice in: this was not an invitation to take the whole of the sovereignty.

It is normal to think that the constitutional sovereignty is hierarchically superior to legislation. The Voice is part of this constitutional sovereignty, so will it not therefore override the incidental legislation (passed, say, under the races power) that might be needed to deal with deleterious consequences in the receipt of Voice activity, such as clogging up the executive process? No. The constitutional hierarchies are more subtle than this.

If the referendum is passed, the Voice (in the newly merged sovereignty) and the races power (part of the original sovereignty) will be hierarchical equals; and this means that the Voice and legislation passed under the races power (or incidental power) will also be hierarchical equals. This is because if legislation passed under a constitutional power were to be systemically invalid the power itself could not exist. The races power does exist, therefore its legislation and the Voice are constitutional equals. Each carries their place in the Constitution, the Voice its existence, the legislation its existent power.

The High Court’s function in relation to such legislation would be one of impartiality between all parts of the Constitution. This means it must find a balance between those parts. The idea pressed from many quarters that it would find ways of augmenting the powers of the Voice is misconceived: the High Court does not make things up. On the passing of the referendum the Voice would achieve a high constitutional power; why would the Court want to make it higher? It must find a balance. And it might err, but this error can rub either way.

“There shall be a Voice”. Is this an immediate Voice? If yes, it can only be the slim proper-named sovereign Voice I have described. This is an autonomous Voice (sovereignty is always autonomous). From this it follows that Parliament’s power “to make laws with respect to matters relating to the Voice, including the composition, functions, powers and procedures” of the Voice must be invalid for inconsistency with the sovereign autonomy of the Voice established by the referendum process. On the other hand, if there is no immediate Voice—no Voice until Parliament establishes it—the referendum is invalid as an unconstitutional delegation.

It is a curious fact that many of the critics of the Voice in recent months have pictured the High Court as some sort of ogre that will cause government to be clogged by decisions expanding the constitutional status of the Voice. But the High Court holding the referendum to be invalid is the best chance they have of saving the country from that fate. The High Court as saviour!

My hope is: the High Court as saviour of the true Voice. The principle against constitutional delegation is an old, very strong, and important legal conception, which when allied with the principle of federation (as I have shown) is seemingly unanswerable. However, when the High Court holds a law to be constitutionally invalid it looks to see whether any part of it can be saved. This depends on the saved part having no logical reliance on the excised unconstitutional part (the delegation). In our case, the question would be whether the first two parts of the constitutional formulation (the slim Voice to Parliament) logically need the delegated third part (Parliament’s fat creation of the Voice). They don’t. The severance of the third part requires that the remainder constitutes a complete constitutional Voice in its own right. And it does: there shall be … an autonomous Voice doesn’t need a Parliamentary construction (indeed, it needs it to be excised).

The consequence of this is that the designed and built (fat) Parliamentary Voice to Parliament is struck out, leaving the recognised autonomous Voice to Parliament as the slim, true and sovereign Voice.

Seven voices would be six too many. The current spate of state voices created by state parliaments will trivialise the Australian sovereign Voice to Parliament; and at some point be inconsistent with it and their voices therefore will be invalid.

The state voices are a nonsense—silly playthings, threatening to be worse. They are not sovereign Voices: there is only one way for the Voice to enter Australian sovereignty; and that is by the referendum process of section 128. A state would have to secede from the Commonwealth to gain an independent sovereignty if it is to be the host of a sovereign Voice. The Australian sovereign Voice to Parliament is a single Voice of the single Australian sovereignty; voices growing as if on trees affront the dignity, quietness and focus of the single Voice, as well as its constitutionality.

Now, it’s obviously important that the single Voice of Australian sovereignty have access to state parliaments and executives as well as Commonwealth. This should be done by including them in the coming referendum. The Voice to the Commonwealth Parliament and Executive is not a merely a Commonwealth Voice, it is an Australian sovereign Voice and binds the states.

The sovereign autonomy of the Voice requires it to organise and formally constitute itself, free of Parliamentary control. Can it do this? Why not? It might begin at Uluru. Those who “gathered at the 2017 National Constitutional Convention, coming from all points of the southern sky [to make a] statement from the heart” might meet again and make a Voice constitution or plan of action.

The plan must be autonomous. The political correctness that thrives in the Parliament threatens this autonomy and is cheap: it will have no difficulty in dreaming up some corporate structure to impose on all of us including the people it seeks to help. Can the Voice resist this? The Aboriginal and Torres Strait people are not in their culture a natural corporation. But at Uluru they came together in a corporate or unified way such that there could be the Voice. The next step is their autonomous plan for the working of the Voice—it’s their problem, and if they can’t solve it, nobody else decently can.

Two things from the parliaments and executives would assist (if wanted). First, the Voice will be a constitutional entity and so must be funded by the Commonwealth: the administrative and financial cost of organising the structure and working of the Voice would be an ordinary item in the Commonwealth budget. The Voice as a working Voice would present a budget, and in that budget, it would have the autonomy to make all the decisions as to how it would work. And second, the parliaments and executives might lend their various administrative arrangements to facilitate Voice organisation.

An autonomous Voice is a free Voice. The Leader of the Opposition has objected to another layer of bureaucracy in the form of what he calls “the Canberra Voice”. “I think we can deliver better outcomes on the ground,” he said in April. The slim, autonomous Voice I have described is a Voice working at ground level. The Liberal Party, which now opposes the fat “Canberra” Voice, will fare better if it offers not just opposition but a real and palpable alternative. The slim autonomous Voice that I have described is a liberal Voice: it brings two free peoples together. The Voice itself is free (autonomous), and we settled Australians are confident enough in our own freedom to put academic definitions of democracy aside and welcome the Voice into the Constitution.

The short question in the referendum process—the one we see in the ballot box—at the moment reads as follows:

A Proposed Law: to alter the Constitution to recognise the First Peoples of Australia by establishing an Aboriginal and Torres Strait Islander Voice. Do you approve this proposed alteration?

Let us call the current three-pointed constitutional formulation of the fat Voice to Parliament the long result. Were it put to the people and carried, the long result would be inserted into the text of the Constitution with no changes. But a huge problem lurks. What is to be put to the people is not this long result; it is a short question as to whether we approve of “a proposed law: to alter the Constitution to recognise the First Peoples of Australia by establishing an Aboriginal and Torres Strait Islander Voice”.

If this latter is put to the people it will achieve (if it is passed) what I shall call the short result. What happens then? Clearly, it is not the short result that is intended to be written into the Constitution. But how can the long result be taken as the decision of the sovereign people when only the short question has been put to them? The long result is not what has been passed.

All that can be entered into the Constitution is that which has been passed by the sovereign people. Suppose, the short result having been passed, that the government directed the printing office to trash all its copies of the (old) Constitution and print a new set with the long result as section 129. Writs would fly. Neither government nor Parliament has the discretion to interpret the short result as the long result. That the short result has no immediate means of transition into the Constitution is quite obvious. The question in the High Court looking at the flying writs would be: What is it possible to say the electors voted for?

The short result question is currently headed “a proposed law”. But what has “law” got to do with the issue? In section 128 the word law refers to the proposal agreed by both Houses of Parliament to be put to the people in the referendum. But this is not a law. It binds no one and when it does bind it is part of the Constitution, not a law—at most it is an administrative direction. Electors can clearly see a certain substance of the proposal in the short version put to them. What could “a proposed law” add to that substance?

The best we can say here is that it will, by the ordinary elector, be ignored; and that they will cast their vote on the substantive short statement. But then for these electors there is no way the short statement can be connected to the long one. How can they be said to have voted for the long one?

Some electors, however, will be more knowledgeable. Those familiar with section 128’s strange usage would expect a law with a proposal in it. But this doesn’t work, either; it doesn’t match the short question which refers to a proposed law. A proposed law is by no means the same thing as a law with a proposal in it (a lawed proposal!). In fact, “A Proposed Law” begins to look like a nonsense. The “proposal” should be about the amendment, should it not?

Some might be even more knowledgeable and take the phrase “to alter the Constitution to establish an Aboriginal and Torres Strait Islander Voice” as identifying a law (statute) by its name—its long title. Well, this may be its long title, but it still gives no clue to what is contained in the statute except what the short description/long title points to.

Only if the short question were: “Do you approve of a proposed law (say, statute no. xy of 2023)” and there was no information at all as to what that statute was about, only then could the short result connect to the long. Whilst this ruse would overcome the logical problem of getting the long result’s content into the Constitution, it would be absurd to allow it. Ballot boxes would be exploding; how can we possibly vote on this statute when we know nothing about it?

The constitutional amendment process is fundamentally about the sovereign people, it is not about the non-sovereign Parliament. The rather curious use of Parliament’s “laws” is just a secure and precise way to get the referendum going, it is not a result in itself. The popular process could never be perfect; but there must come a point below which it fails to represent the sovereign will and is invalid.

The issue comes down to this: in the Voice referendum, if, as is likely, a substantial proportion of electors do not even know that there is a long result, they cannot possibly be said to have voted for it. Nothing beyond the short result (in a Yes vote) could be put into the Constitution.

But actually it couldn’t, for the obvious reason that it failed the Parliamentary part of the process—it was not what Parliament intended to put to the electorate. And so, the whole process must be invalid. But now the issue turns interesting. With a little re-drafting, the short result would, in the Voice referendum, actually be a good result, albeit unintended by the Parliament. And the removal of the long/fat excess would correspond to the substantive invalidity of the unconstitutional delegation to Parliament.

“To alter the Constitution to recognise the First Peoples of Australia by establishing an Aboriginal and Torres Strait Islander Voice” would turn into the formulation: “129: There shall be a body to be called the Aboriginal and Torres Strait Islander Voice.”

To this we should add its function to make representations, and this should also include state and territory parliaments and executives. So: “129: There shall be an autonomous body to be called the Aboriginal and Torres Strait Islander Voice which may make representations to Australian Parliaments and Executive Governments on matters relating to Aboriginal and Torres Strait Islander peoples.” (“Autonomous” slim gets in as fat Parliament is pushed out.)

Shorter even than the republic referendum’s short question, just that should be put to the people on their ballot papers. The distinction between the long and short formulations would no longer hold (this should always be the case with simple amendments). On a Yes-vote this short statement would go immediately into the Constitution with nothing else.

What was intended for the Constitution was the long form of the fat Voice to Parliament. The proponents of the short statement are often simply frightened they might lose on the real and more complex long issues. And they feel justified in this position by a notorious and unconstitutional notion that the difficulty of amendment of the Australian Constitution is a kind of enemy to overcome. It isn’t: that there have been few successful referendums in Australia is a sign of a good constitution, not a bad.

I have described what should happen with the content of the ballot papers. If that does not happen and the referendum goes ahead as currently designed, there would be no way of achieving either the long or the short result. The referendum would be null and void, and I believe the High Court, were it asked, would so rule. There is irony in the fact that between the constitutional substance and the form of the referendum process there emerges a kind of conspiracy against the sovereign people. The people are excluded from perusing the terms of the definition of the Voice by their being delegated to Parliament, and then these terms are further hidden from perusal in the referendum process by the way Parliament has set it up.

Michael Detmold is Professor of Law (emeritus), University of Adelaide.