State of NSW v Kable

FRENCH CJ, HAYNE, CRENNAN, KIEFEL, BELL, GAGELER AND KEANE JJ

State of NSW v Kable [2013] HCA 26 5 June 2013 S352/2012

2. Set aside the orders of the Court of Appeal of the Supreme Court of New South Wales made on 8 August 2012 and, in their place, order that the appeal to that Court be dismissed with costs.

3. The appellant pay the respondent's costs of the application for special leave to appeal and of the appeal to this Court.

On appeal from the Supreme Court of New South Wales

M G Sexton SC, Solicitor-General for the State of New South Wales and M J Leeming SC with J E Davidson for the appellant (instructed by Crown Solicitor (NSW))

P W Bates with P G White for the respondent (instructed by Armstrong Legal)

J T Gleeson SC, Solicitor-General of the Commonwealth with A M Mitchelmore for the Attorney-General of the Commonwealth, intervening (instructed by Australian Government Solicitor)

W Sofronoff QC, Solicitor-General of the State of Queensland with G J D del Villar for the Attorney-General of the State of Queensland, intervening (instructed by Crown Law (Qld))

S G E McLeish SC, Solicitor-General for the State of Victoria with R J Orr for the Attorney-General for the State of Victoria, intervening (instructed by Victorian Government Solicitor)

G R Donaldson SC, Solicitor-General for the State of Western Australia with K H Glancy for the Attorney-General for the State of Western Australia, intervening (instructed by State Solicitor (WA))

Notice: This copy of the Court's Reasons for Judgment is subject to formal revision prior to publication in the Commonwealth Law Reports.

Constitutional law - Judicial power - Respondent detained pursuant to order of Supreme Court of New South Wales made under Community Protection Act 1994 (NSW) ("CP Act") - CP Act subsequently held invalid - Respondent sought damages for false imprisonment - Whether order of Supreme Court valid until set aside - Whether order of Supreme Court judicial order.

Torts - False imprisonment - Defences - Lawful authority - Respondent held under order of Supreme Court subsequently set aside - Whether officers of appellant could rely on order made under invalid legislation as lawful authority.

Words and phrases - "judicial order", "lawful authority", "superior court of record", "void ab initio", "void or voidable".

Constitution, ss 76, 77. Community Protection Act 1994 (NSW), s 9.

FRENCH CJ, HAYNE, CRENNAN, KIEFEL, BELL AND KEANE JJ. The issue A State Act empowered the State's Supreme Court to order the preventive detention of Gregory Wayne Kable if satisfied that otherwise he would probably commit a serious act of violence. The Supreme Court ordered Mr Kable's detention for six months. After the six months had elapsed, the detention order was set aside on appeal to this Court and the State Act held invalid. Did the detention order provide lawful authority for Mr Kable's detention? Procedural history The Community Protection Act 1994 (NSW) ("the CP Act") provided for "the preventive detention (by order of the Supreme Court [of New South Wales] made on the application of the Director of Public Prosecutions) of Gregory Wayne Kable". On 23 February 1995, on the application of the Director of Public Prosecutions, Levine J made an order pursuant to s 9 of the CP Act that Mr Kable be detained in custody for a period of six months. Mr Kable appealed against this order to the Court of Appeal but his appeal was dismissed. By special leave, Mr Kable appealed to this Court. After the grant of special leave, but before the appeal to this Court was heard, the six month period fixed by the order of Levine J expired and Mr Kable was released from detention. In September 1996, this Court held that the CP Act was invalid. This Court allowed Mr Kable's appeal, set aside the order which the Court of Appeal had made, and, in its place, ordered that the appeal to that Court be allowed with costs, the order of Levine J be set aside and, in its place, order that the application of the Director of Public Prosecutions be dismissed with costs. It will be convenient to refer to this decision as Kable (No 1). After this Court decided Kable (No 1), Mr Kable commenced proceedings in the Supreme Court of New South Wales, claiming damages for false imprisonment. Initially the proceedings were brought against the State of New South Wales ("the State"). Later, the Director of Public Prosecutions ("the DPP") was joined as a defendant. Ultimately three causes of action were pleaded: abuse of process, malicious prosecution and false imprisonment. Before the action was tried, the DPP was dismissed from the proceedings by consent. The primary judge (Hoeben J) determined a number of issues as preliminary questions. Those issues were decided against Mr Kable and judgment entered for the State. In particular, the primary judge rejected Mr Kable's argument that the detention order made by Levine J was a nullity when made and held that the order was valid until it was set aside. Mr Kable appealed to the Court of Appeal. That Court (Allsop P, Basten, Campbell and Meagher JJA and McClellan CJ at CL) allowed the appeal in part. The Court of Appeal held that the primary judge had been right to dismiss Mr Kable's claims for collateral abuse of process and malicious prosecution but that Mr Kable should have judgment against the State for damages to be assessed on his claim for false imprisonment. All members of the Court of Appeal held that the order of Levine J was no answer to Mr Kable's claim for false imprisonment. Allsop P held that the reasons given by this Court in Kable (No 1) required the conclusion that, in making the detention order, "the Supreme Court was not exercising judicial power or authority and was not acting, institutionally, as a superior court but was acting, effectively, in an executive function (beyond that which is permissibly ancillary to the exercise of judicial power), as an instrument of the executive". Basten JA also held that the detention order made by Levine J "did not constitute a judicial order" and that "[a]ny contrary conclusion would contradict the findings" of this Court in Kable (No 1). Basten JA noted that, because constitutional questions had been raised in the proceedings before Levine J, the Supreme Court had exercised federal jurisdiction, but concluded that the detention order was none the less "an invalid non‑judicial order". By special leave, the State appealed to this Court against the orders made by the Court of Appeal. The Attorneys‑General of the Commonwealth, Queensland, Victoria and Western Australia intervened in support of the State's appeal. These reasons will show that the detention order made by Levine J provided lawful authority for Mr Kable's detention and that the State's appeal should be allowed. The competing arguments The central question in the appeal to this Court was whether the order made by Levine J, until it was set aside, had provided lawful authority for Mr Kable's detention. There was (and could be) no dispute that the Supreme Court of New South Wales is a "superior court of record". There was (and could be) no dispute that the CP Act was invalid. On its face, the detention order was made by the Supreme Court in the exercise of a jurisdiction given to it by a New South Wales Act. The order was expressed to require Mr Kable's detention in the manner and for the time specified. In the course of deciding whether to grant the order sought, Levine J had been asked to hold that the CP Act was constitutionally invalid but had rejected that submission. The State, and the interveners, submitted that the order made by Levine J was made by a superior court of record and accordingly was effective until it was set aside. Because the order was not set aside until after Mr Kable's release from detention, it followed, so the State submitted, that the order provided lawful authority for Mr Kable's detention. By contrast, Mr Kable submitted that the bases on which this Court held the CP Act invalid required not only the conclusion that the CP Act was invalid, but also the conclusion that the Supreme Court could not make (and had not made) a "judicial" order requiring his detention. He submitted that it followed that the principle requiring that effect be given to an order of a superior court until it was set aside was not engaged. Either the order made by Levine J was "void ab initio" or, when set aside by order of this Court, the order was "annulled ab initio". On either footing, the argument continued, the order provided no lawful authority for Mr Kable's detention. Kable (No 1) Consideration of the competing arguments must begin by identifying what was decided in Kable (No 1). The legislative powers of each of the State Parliaments are necessarily subject to the federal Constitution. The CP Act was held to be beyond the legislative power of the New South Wales Parliament because its enactment was contrary to the requirements of Ch III of the Constitution. The exercise of the jurisdiction which the CP Act purported to give to the Supreme Court was held to be incompatible with the institutional integrity of the Supreme Court. The incompatibility with institutional integrity which was identified in Kable (No 1) lay in the Supreme Court being required to act as a court in the performance of a function identified as not being a function for the judicial branch of government. The majority in Kable (No 1) described the function which the CP Act required the Supreme Court to undertake in several different ways. All of those descriptions emphasised that the function which the CP Act required the Court to fulfil was not judicial. So, for example, Gaudron J said that the power given by the CP Act "is not a power that is properly characterised as a judicial function, notwithstanding that it is purportedly conferred on a court and its exercise is conditioned in terms usually associated with the judicial process" and that "except to the extent that the [CP] Act attempts to dress them up as legal proceedings ... they do not in any way partake of the nature of legal proceedings". But these and other similar statements made in the reasons of the majority in Kable (No 1) proceeded from the premise that the CP Act required the Supreme Court to act as a court in performing the function prescribed by the CP Act. As Gummow J later said, in Fardon v Attorney‑General (Qld), the "legislative plan" of the CP Act was "to conscript the Supreme Court of New South Wales to procure the imprisonment of [Mr Kable] by a process which departed in serious respects from the usual judicial process". It is, therefore, to misstate the effect of the decision in Kable (No 1) to hold, as the Court of Appeal did, that in exercising power under the CP Act, "the Supreme Court was not exercising judicial power or authority and was not acting, institutionally, as a superior court". The majority in Kable (No 1) held that the CP Act was invalid because it required the Supreme Court to exercise judicial power and act institutionally as a court, but to perform a task that was inconsistent with the maintenance (which Ch III of the Constitution requires) of the Supreme Court's institutional integrity. Two further aspects of the proceedings which culminated in this Court's decision in Kable (No 1) should be noted. Those proceedings were conducted on the premises, first, that the proceedings at first instance and in the Court of Appeal engaged the judicial power of the Commonwealth, and second that the order made by the Court of Appeal (and the order made by Levine J) engaged this Court's appellate jurisdiction conferred by s 73 of the Constitution. That the proceedings before Levine J and in the Court of Appeal engaged federal jurisdiction (at least to the extent the proceedings were a matter arising under the Constitution or involving its interpretation) is not open to doubt. And it is equally beyond doubt that the orders of the Court of Appeal, dismissing Mr Kable's appeal to that Court, were within this Court's appellate jurisdiction as a particular example of a species of the genus "all judgments, decrees, orders, and sentences ... of the Supreme Court of any State". It is then to be observed that in Kable (No 1) this Court ordered that, in place of the order made by the Court of Appeal, the appeal to that Court was allowed and the order of Levine J was set aside, as distinct from quashed, or declared invalid. That is, the order of Levine J was treated in this Court's orders in a manner consistent with it having been valid until set aside. "Void" or "voidable"? Mr Kable submitted that the detention order of Levine J was void or was later avoided ab initio. In support of these submissions, he pointed to statements made in this Court (referring to earlier English decisions) distinguishing between what is "void" or "voidable" and what is an "irregularity" or a "nullity". It is necessary to exercise great care in using words like "void", "voidable", "irregularity" and "nullity" in connection with the issues that arise in this matter. Each word was used in Mr Kable's argument in this appeal to state a conclusion about the legal effect of the order of Levine J. More often than not, each word was used in a way which expressly or impliedly sought to convey a meaning identified by its opposition to another word (void versus voidable, nullity versus irregularity). Used in that way, each of the words, void, voidable, nullity and irregularity, suggests that the whole of the relevant universe can be divided between two realms whose borders are sharply defined and completely closed. None is used in a way which admits (or readily appears to admit) of the possibility that the legal effect to be given to an act affected by some want of power may require a more elaborate description which takes account not only of who may complain about the want of power, but also of what remedy may be given in response to the complaint. The difficulties associated with using words like "void" and "voidable" in connection with administrative actions have long been recognised. Writing in 1967, H W R Wade said that: "[T]here is no such thing as voidness in an absolute sense, for the whole question is, void against whom? It makes no sense to speak of an act being void unless there is some person to whom the law gives a remedy. If and when that remedy is taken away, what was void must be treated as valid, being now by law unchallengeable. It is fallacious to suppose that an act can be effective in law only if it has always had some element of validity from the beginning. However destitute of legitimacy at its birth, it is legitimated when the law refuses to assist anyone who wants to bastardise it. What cannot be disputed has to be accepted." Although directed to administrative actions, these statements may find some reflection in connection with the acts of courts and judges. If a curial decision cannot be disputed, it must be accepted. To the extent to which the orders of a superior court are valid until set aside, there seems little point in attempting to classify those orders as void or voidable. But it is not necessary to pursue those analogies to their conclusion. It is enough to notice that the legal system provides (and must provide) the rules which govern what legal effect is to be given to the decisions of, and the orders made by, courts. And, as later explained, those rules are more complex than the central proposition which underpinned Mr Kable's arguments: that want of jurisdiction for constitutional reasons necessarily entails the complete invalidity for all purposes of whatever is done in purported exercise of that jurisdiction. One further point should be made about the asserted distinction between proceedings that are "irregular" and those that are "null and void" or a "nullity". This distinction has sometimes been made in connection with applying rules of court which provide for the consequences of non‑compliance with the rules, and in that context speak of setting aside proceedings "as irregular" or "for irregularity". The distinction has been made in determining whether and when complaint could be made about a departure from procedural requirements. That is not the question which arises in this matter. There can be no direct application of that kind of distinction to the present matter. A judicial order? The point at which analysis must begin in this matter is, as the parties' submissions recognised, with the order that was made for Mr Kable's detention. Was that an order to which effect was to be given unless or until it was set aside? Mr Kable submitted that the order was of no effect because it was not a "judicial order". The premise for the submission was that the CP Act did not validly authorise the making of the detention order because the power which the CP Act sought to give to the Supreme Court was not a form of judicial power. And, at first sight, the submission that the detention order was not a "judicial order" may be thought to invoke consideration of the principles governing what is judicial power. It is important, however, to recognise the nature of the distinction being drawn in the submission that the detention order was not a "judicial order". The positive proposition which underpinned the submission was that, although made by a judge of the Supreme Court, the detention order was an exercise of administrative power by the Court. Thus, the distinction upon which the submission depended fastened upon how the power which the CP Act purported to give to the Supreme Court was exercised, not upon whether the power was given validly to the Supreme Court. In Love v Attorney‑General (NSW), this Court examined whether a warrant issued by a judge of the Supreme Court of New South Wales, pursuant to a State Act, was issued in exercise of judicial power. This Court concluded that, although the power to issue warrants was conferred upon the Supreme Court rather than the several judges of that Court, "the exercise of the power is essentially administrative in nature". The judge who issued a warrant was under a duty to act judicially, but this Court concluded that: "the issue of the warrant ... is not a 'judicial act in the same sense as is an adjudication to determine the rights of parties', to use the words of Windeyer J in [Electronic Rentals Pty Ltd v Anderson]. It is not an order inter partes from which a party whose conversations may be overheard has a right of appeal. To adapt the language used in Hilton v Wells, under [the State Act] a judge makes no order and nothing that he or she does is enforced as an order of the court." Accordingly, a warrant issued under the State Act was "an instrument made pursuant to a circumscribed statutory authority", and its effect depended entirely upon the State Act. By contrast with the warrant considered in Love, the order made by Levine J was more than an authority granted on the application of a person seeking authority and without notice to the person against whom the authority would be exercised. The order made by Levine J was the result of an adjudication determining the rights of Mr Kable and the order both authorised and required his detention for a fixed term. The order was made following proceedings which were conducted inter partes. Subject to some exceptions, the rules of evidence applied. Witnesses were examined and cross‑examined and the opposing parties made submissions. The order was enforced as a court order. Mr Kable could and did appeal against the order. All of these features of the proceedings and the order that was made disposing of the proceedings point to the order being made by a judge of the Supreme Court in his judicial capacity. None suggests that the order was, like the issue of the warrant considered in Love, "a step in the administrative process and ... thus an administrative function". The order made by Levine J was a judicial order. The order of a superior court of record As has already been noticed, there was and could be no dispute that the Supreme Court of New South Wales was and is a "superior court of record". It is necessary, however, to approach what meaning is conveyed by that expression with some fundamental principles at the forefront of consideration. First, and foremost, there can be no unthinking transplantation to Australia of what has been said in English cases about the consequences of a court being established as a "superior court of record". The constitutional context is altogether different. Due regard must be paid to those differences. Second, there is no Australian court with unlimited jurisdiction. Hence, although it is sometimes suggested that, in England, the prerogative writs of mandamus, prohibition and certiorari were not available to provide relief against a judgment or orders of a judge of a superior court, that suggestion, even if accurate, could have no direct or immediate application in Australia. Third, all courts, whether superior or inferior, have the authority to decide whether a claim that is made in the court is within its jurisdiction. That power can be described as a court having jurisdiction to decide its own jurisdiction. But because there is no Australian court with unlimited jurisdiction, a decision that a court does, or does not, have authority to decide a particular claim will be subject to review and correction. Sometimes that will be by the grant of prohibition or certiorari; sometimes, as exemplified by Kable (No 1), it will be by the process of appeal, and ultimately by appeal to this Court. And if it is said that a superior court is presumed to act within its jurisdiction, that is best understood as a statement about the effect that is to be given to its orders unless or until they are set aside. It is now firmly established by the decisions of this Court that the orders of a federal court which is established as a superior court of record are valid until set aside, even if the orders are made in excess of jurisdiction (whether on constitutional grounds or for reasons of some statutory limitation on jurisdiction). It was not submitted that any of these decisions should be reopened and there would be powerful reasons not to disturb such a long‑established stream of authority. Nor was it submitted that these principles did not apply equally to the judicial orders of a State Supreme Court. Rather, as already noted, the principles were said not to apply because the order made by Levine J was not a judicial order. And, for the reasons already given, that submission must be rejected. The roots of the doctrine, that the orders of a superior court of record are valid until set aside even if made in excess of jurisdiction, lie in the nature of judicial power. First, any court must decide whether it has authority to decide the claim that is made to it. And, as Gaudron J said in Re Macks; Ex parte Saint: "In establishing the Federal Court as a 'superior court of record', the Parliament has, at the very least, validly authorised that Court to make a binding determination on the question whether or not it has jurisdiction in a matter, subject only to the parties' right to appeal or to seek relief pursuant to s 75(v) of the Constitution." (emphasis added) Second, giving the orders of a court created by the Parliament these characteristics is within legislative power, either as incidental to the power to create the court or as an exercise of the legislative powers given by ss 76 and 77 of the Constitution. And giving these characteristics to the orders of a court by designating it to be a superior court of record reflects the distinction between the exercise of judicial power (by the final quelling of controversies according to law) and the exercise of executive power (subject to law). As Gummow J said in Re Macks: "That does not mean that the stream [of judicial power] has risen above its source. Rather, it is to recognise the relationship between Chs II and III of the Constitution and the reach of s 51(xxxix) in conjunction with ss 71 and 77(i)." Contrary to the view expressed by Basten JA in the Court of Appeal, and supported by Mr Kable in argument in this Court, these conclusions present no "logical conundrum". Nor do these conclusions require consideration of whether "where a judicial function and an incompatible non‑judicial function are purportedly exercised in one proceeding, the incompatible non‑judicial function is not thereby incorporated into a single exercise of federal jurisdiction". The conundrum identified by Basten JA was expressed as being that the law on which the effect of the judicial order depended gave it "an effect extending beyond the constitutional limits of that jurisdiction". But, as has been explained, the effect which is given to the order made beyond jurisdiction comes not from the law which purported to confer the relevant jurisdiction but from the status or nature of the court making the order (as a superior court of record). The effect which is given to the order is for only so long as it remains in force. Once set aside on appeal, the order is spent. There is then no occasion to attempt to divide the exercise of jurisdiction by Levine J in the manner considered by Basten JA. The division suggested was between the (valid) exercise of jurisdiction conferred by s 39(2) of the Judiciary Act 1903 (Cth) to hear and determine the question about the validity of the CP Act (as a question arising under the Constitution or involving its interpretation) and the (invalid) exercise of jurisdiction to decide whether to make an order under the CP Act. There being no occasion to consider this division, it is neither necessary nor desirable to examine whether the proceedings conducted by Levine J constituted the hearing and determination of one or more than one "matter", or what may have been the boundaries of the relevant matter or matters. More fundamental considerations The conclusions reached in these reasons about the effect of the order made by Levine J accord with fundamental considerations about the operation of any developed legal system. There must come a point in any developed legal system where decisions made in the exercise of judicial power are given effect despite the particular decision later being set aside or reversed. That point may be marked in a number of ways. One way in which it is marked, in Australian law, is by treating the orders of a superior court of record as valid until set aside. Were this not so, the exercise of judicial power could yield no adjudication of rights and liabilities to which immediate effect could be given. An order made by a superior court of record would have no more than provisional effect until either the time for appeal or review had elapsed or final appeal or review had occurred. Both the individuals affected by the order, and in this case the Executive, would be required to decide whether to obey the order made by a court which required steps to be taken to the detriment of another. The individuals affected by the order, and here the Executive, would have to choose whether to disobey the order (and run the risk of contempt of court or some other coercive process) or incur tortious liability to the person whose rights and liabilities are affected by the order. In this case, if the detention order made by Levine J was not effective until set aside, those apparently bound by the order were obliged to disobey it, lest they be held responsible for false imprisonment. On Mr Kable's argument, the order was without legal effect and should not have been obeyed. The decision to disobey the order would have required both the individual gaoler and the Executive Government of New South Wales to predict whether this Court would accept what were then novel constitutional arguments. More fundamentally, as the legal philosopher Hans Kelsen wrote, "[a] status where everybody is authorized to declare every norm, that is to say, everything which presents itself as a norm, as nul, is almost a status of anarchy". Finally, it will be recalled that Mr Kable submitted that the effect of this Court's orders in Kable (No 1) was to render the order of Levine J void ab initio. This Court did not declare the order made by Levine J to be void. As a judicial order of a superior court of record, the order of Levine J was valid until set aside. It was not "void ab initio". Questions not reached These conclusions make it unnecessary to consider whether, as the State submitted, the common law provides protection from tortious liability by the defence of legal justification to those who execute an order of a court that is valid on its face. Further, because Mr Kable was detained pursuant to an order that had not been set aside during the period of detention, and remained valid during that time, it is also unnecessary to examine whether, as Mr Kable contended, the Court of Appeal should have held that the State was "directly liable in false imprisonment to [Mr Kable], in addition to being vicariously liable for that tort". Conclusion and orders For these reasons the appeal should be allowed, the orders of the Court of Appeal set aside and in their place there should be orders that the appeal to that Court is dismissed with costs. In accordance with the terms on which special leave to appeal to this Court was granted, the State should pay Mr Kable's costs of the application for special leave to appeal and of the appeal to this Court.

GAGELER J. Introduction On 23 February 1995, Levine J in the Supreme Court of New South Wales made an order that: "Gregory Wayne Kable be detained in custody for a period of six months commencing 23 February 1995 and expiring 22 August 1995 pursuant to s 9 of the Community Protection Act 1994." The joint reasons for judgment conclude that the order so made provided lawful authority for the State of New South Wales to detain Mr Kable in custody between 23 February 1995 and 22 August 1995. That conclusion is reached notwithstanding the subsequent setting aside of the order by the High Court, on appeal from the Court of Appeal of the Supreme Court of New South Wales, in Kable v Director of Public Prosecutions (NSW) ("Kable (No 1)"), on the ground that the Community Protection Act 1994 (NSW) ("the CP Act") was invalid because the purported conferral by the CP Act of jurisdiction on the Supreme Court of New South Wales to make an order of that kind was incompatible with Ch III of the Constitution. I agree with that conclusion and join in the orders proposed. Precepts Chapter III of the Constitution is framed against certain precepts, both as to the nature of judicial power and as to the essential and permitted characteristics of courts as institutions administering or capable of administering judicial power. The nature of judicial power is that, as a general rule, its exercise is directed to the resolution of a question as to a legal right or a legal obligation between defined persons or classes of persons so as to produce, by a final judicial order, "a new charter by reference to which that question is in future to be decided as between those persons or classes of persons". Its "unique and essential function … is the quelling of … controversies by ascertainment of the facts, by application of the law and by exercise, where appropriate, of judicial discretion". Its "hallmark" is "the making of binding declarations of rights and obligations arising from the operation of the law upon past events or conduct". Within the framework of the Australian Constitution, there has never been any doubt that application of the law to produce a final judicial order in the exercise of judicial power can permit and require the making of a judicial determination of the constitutional validity or invalidity of a statute. That principle, and its application to the determination by a court of the constitutional invalidity of a statute purporting to confer jurisdiction or power on the court itself, was established 100 years before the establishment of the High Court, in Marbury v Madison. The principle as so established "in our system … is accepted as axiomatic, modified in varying degree in various cases (but never excluded) by the respect which the judicial organ must accord to opinions of the legislative and executive organs". Within the framework of the Australian Constitution, and consistently with the justification for that principle stated in Marbury v Madison, there has never been any doubt that a purported law that is determined in the exercise of judicial power to be invalid, as beyond legislative power or as infringing an express or implied constitutional prohibition, is no law at all and is therefore of no legal force. The notion that a law may be invalid only prospectively from the time of the making of a judicial order has been firmly rejected on the basis that "it would be a perversion of judicial power to maintain in force that which is acknowledged not to be the law". A judicial determination of validity or invalidity is a determination of what the law applicable to the rights or duties of the persons or classes of persons in controversy is or is not - not of what the law is to be. The settled position is captured in the explanation given by Latham CJ in South Australia v The Commonwealth. After pointing out that "[c]ommon expressions, such as: 'The courts have declared a statute invalid,' sometimes lead to misunderstanding", he said: "A pretended law made in excess of power is not and never has been a law at all. Anybody in the country is entitled to disregard it. Naturally he will feel safer if he has a decision of a court in his favour - but such a decision is not an element which produces invalidity in any law. The law is not valid until a court pronounces against it - and thereafter invalid. If it is beyond power it is invalid ab initio." Yet a purported but invalid law, like a thing done in the purported but invalid exercise of a power conferred by law, remains at all times a thing in fact. That is so whether or not it has been judicially determined to be invalid. The thing is, as is sometimes said, a "nullity" in the sense that it lacks the legal force it purports to have. But the thing is not a nullity in the sense that it has no existence at all or that it is incapable of having legal consequences. The factual existence of the thing might be the foundation of rights or duties that arise by force of another, valid, law. The factual existence of the thing might have led to the taking of some other action in fact. The action so taken might then have consequences for the creation or extinguishment or alteration of legal rights or legal obligations, which consequences do not depend on the legal force of the thing itself. For example, money might be paid in the purported discharge of an invalid statutory obligation in circumstances which make that money irrecoverable, or the exercise of a statutory power might in some circumstances be authorised by statute, even if the repository of the power acted in the mistaken belief that some other, purported but invalid exercise of power is valid. One of the ways in which the existence in fact of a purported but invalid law, or the existence in fact of a thing done invalidly in the purported exercise of a power conferred by law, might lead to the taking of action in fact is that the purported law or other thing might be relied on to found an action in a court which results in the making of a judicial order. Where that occurs, the judicial order will have independent legal force as a judicial order. That is illustrated by Victorian Stevedoring and General Contracting Co Pty Ltd and Meakes v Dignan, where convictions by a magistrate for offences under regulations were sustained notwithstanding the subsequent disallowance of those regulations with retrospective effect. Having held that "after a regulation has been disallowed, no one is liable to conviction for an offence committed while it was in force" in that "[h]is liability ceases when the law is revoked that imposed it", Dixon J explained: "But if he has already been convicted, then because his liability has merged in the conviction, it no longer depends upon the law under which it arose, and it does not lapse with the revocation of the law. The conviction has become the source of his liability for his offence, and the conviction continues in force because its operation does not depend upon the law creating the offence, but upon the authority belonging to a judgment or sentence of a competent Court." Within the framework of the Australian Constitution, the authority belonging to a judicial order of a court varies between categories of courts. A distinction is drawn between superior courts and inferior courts. The Supreme Courts of the States, to which reference is made in s 73 of the Constitution, were at federation and necessarily remain superior courts. The High Court, referred to in s 71 as the "Federal Supreme Court", is inherently a superior court. Other State courts as may be invested with federal jurisdiction under s 77(iii) of the Constitution can be created by State Parliaments (subject to State Constitutions) as either superior courts or inferior courts, just as other federal courts can be created by the Commonwealth Parliament under s 71 of the Constitution as either superior courts or inferior courts. A judicial order of any court, whether superior or inferior, is valid and effective if it is made within jurisdiction. Any judicial order, whether of a superior court or an inferior court and whether made within or without jurisdiction, is a judgment, decree, order or sentence from which an appeal may lie to the High Court under s 73 of the Constitution and, where such an appeal lies, a judicial order made without jurisdiction may be set aside by the High Court in determining the appeal. Any judicial order made in excess of jurisdiction by a federal court, whether the court be created as a superior court or an inferior court, may be set aside by a writ of certiorari issued under s 32 of the Judiciary Act 1903 (Cth) in the exercise of the original jurisdiction of the High Court conferred by s 75 or under s 76 of the Constitution. There is, however, a critical distinction between a superior court and an inferior court concerning the authority belonging to a judicial order that is made without jurisdiction. A judicial order of an inferior court made without jurisdiction has no legal force as an order of that court. One consequence is that failure to obey the order cannot be a contempt of court. Another is that the order may be challenged collaterally in a subsequent proceeding in which reliance is sought to be placed on it. Where there is doubt about whether a judicial order of an inferior court is made within jurisdiction, the validity of the order "must always remain an outstanding question" unless and until that question is authoritatively determined by some other court in the exercise of judicial power within its own jurisdiction. In contrast: "It is settled by the highest authority that the decision of a superior court, even if in excess of jurisdiction, is at the worst voidable, and is valid unless and until it is set aside". The constitutional robustness of the principle that a judicial order of a superior court made without jurisdiction has legal force as an order of that court unless and until it is set aside is illustrated by Re Macks; Ex parte Saint. There the principle was applied to judicial orders of the Federal Court of Australia which had been made in fact, but in the exercise of jurisdiction only purportedly conferred by Commonwealth and State statutes held to be invalid as infringing Ch III of the Constitution. The legal force of those orders, unless and until set aside, was held to derive not from the constitutionally impermissible purported conferral of jurisdiction by the Commonwealth Parliament or by State Parliaments but from the constitutionally permissible conferral by the Commonwealth Parliament on the Federal Court of the status of a superior court. The capacity of a superior court, conformably with Ch III of the Constitution, to make a judicial order that is valid unless and until it is set aside in the purported exercise of a jurisdiction purportedly conferred by a statute itself invalid because it infringes Ch III of the Constitution might at first glance seem anomalous. The justification for that capacity lies ultimately in the unique and essential function constitutionally reposed by Ch III in courts: that of rendering final resolutions of controversies about legal rights and legal obligations. The nature of that function makes some degree of self-reference in its performance inevitable. That is because a question about the validity or invalidity of a statute purporting to confer jurisdiction on a court is itself a question that can be resolved to finality only in the exercise of judicial power. The making of a final judicial order in the purported exercise of a jurisdiction conferred by statute ordinarily involves an explicit or implicit determination by the court making that order that the court has jurisdiction to determine a dispute about a legal right or a legal obligation between the persons or classes of persons who are parties to proceedings in which the jurisdiction is invoked or who are represented by those parties. The making of such a final judicial order therefore ordinarily involves, explicitly or implicitly, a resolution of the question of jurisdiction as between those persons or classes of persons. Extension of the capacity of a superior court to determine a question about its own jurisdiction to the determination of a question about the validity or invalidity of a statute purporting to confer jurisdiction on that court is therefore consistent with the finality which it is the function of the judicial power to provide. When a judicial order of a superior court explicitly or implicitly resolves a question of the validity or invalidity of a statute purporting to confer jurisdiction on that court, the finality of that order and the finality of that resolution are incidents of the nature of that court as a repository of judicial power. The same capacity, with the same justification, has been held to inhere in courts created by Congress under Art III of the United States Constitution, which are "inferior courts" only in the sense that they are subordinate to the Supreme Court of the United States. The Supreme Court has held that a judgment of such a court entered without jurisdiction, while open to direct review, cannot be challenged collaterally, "[w]hatever the contention as to jurisdiction may be, whether it is that the boundaries of a valid statute have been transgressed, or that the statute itself is invalid". The Supreme Court earlier explained: "A court does not have the power, by judicial fiat, to extend its jurisdiction over matters beyond the scope of the authority granted to it by its creators. There must be admitted, however, a power to interpret the language of the jurisdictional instrument and its application to an issue before the court … Every court in rendering a judgment, tacitly, if not expressly, determines its jurisdiction over the parties and the subject matter. An erroneous affirmative conclusion as to the jurisdiction does not in any proper sense enlarge the jurisdiction of the court until passed upon by the court of last resort, and even then the jurisdiction becomes enlarged only from the necessity of having a judicial determination of the jurisdiction over the subject matter." (footnotes omitted) The Supreme Court added: "Courts to determine the rights of parties are an integral part of our system of government. It is just as important that there should be a place to end as that there should be a place to begin litigation. After a party has his day in court, with opportunity to present his evidence and his view of the law, a collateral attack upon the decision as to jurisdiction there rendered merely retries the issue previously determined. There is no reason to expect that the second decision will be more satisfactory than the first." The same theme was later taken up by Frankfurter J. He said: "Every act of government may be challenged by an appeal to law, as finally pronounced by this Court. Even this Court has the last say only for a time. Being composed of fallible men, it may err. But revision of its errors must be by orderly process of law." He added: "No one, no matter how exalted his public office or how righteous his private motive, can be judge in his own case. That is what courts are for. And no type of controversy is more peculiarly fit for judicial determination than a controversy that calls into question the power of a court to decide. Controversies over 'jurisdiction' are … hardly fit for final determination by the self-interest of a party." The persuasive force of those observations is not undermined by the more recent adoption for a time in some judgments of the Supreme Court of the United States of an approach to prospective overruling rejected in Australia. Problem The Court of Appeal did not disregard those precepts in the present case: it rather encountered a particular problem in seeking to apply them. The CP Act purported to confer power on the Supreme Court to make a "preventive detention order", being an order "that a specified person be detained in prison for a specified period", on application by the New South Wales Director of Public Prosecutions ("the DPP") if satisfied on reasonable grounds that specified criteria were met. Proceedings under the CP Act were proceedings between the DPP and the person against whom the preventive detention order was sought, to be commenced by summons in accordance with rules of court, and to be conducted in accordance with the law (including the rules of evidence) relating to civil proceedings. The CP Act purported to provide that the jurisdiction of the Supreme Court under the CP Act "is exercisable by a single Judge", and that "[a]n appeal to the Court of Appeal lies from any determination of the [Supreme] Court to make, or to refuse to make, a preventive detention order", such an appeal able to be "on a question of law, a question of fact or a question of mixed law and fact". The appeal was "by way of rehearing", the Court of Appeal having the powers and duties of Levine J. The CP Act purported to provide that a preventive detention order was "sufficient authority for the person against whom it is made to be held in custody in accordance with the terms of the order". It purported to provide for that person to be "taken to be a prisoner" for the purposes of the Prisons Act 1952 (NSW). It purported to add: "No action lies against any person (including the State) for or in respect of any act or omission done or omitted by the person so long as it was done or omitted in good faith for the purposes of, or in connection with the administration or execution of, this Act." Those provisions of the CP Act, purporting to confer power and jurisdiction on the Supreme Court, purporting to attach statutory consequences to a preventive detention order, and purporting to confer immunity from civil action taken in good faith in the execution of the CP Act, were all held in Kable (No 1) to be invalid. The Court of Appeal in the present case recognised that the order made by Levine J on 23 February 1995, if it was a judicial order, nevertheless provided an independent source of lawful authority for the State of New South Wales to have detained Mr Kable in custody in the same way as a sentence imposed by a superior court is sufficient authority for its execution. The problem was that the Court of Appeal considered itself bound by the reasons for judgment in Kable (No 1) to hold that the order made by Levine J could not have been made in the exercise of judicial power and was therefore not a judicial order. As put by Allsop P, "[i]t would be in the teeth of the majority's views" in Kable (No 1) to describe the order as a judicial order "or the act of making [it] judicial, in any relevant sense of judicial capacity". As put by Basten JA, it was "not open" to the Court of Appeal "to conclude that the detention order was otherwise than an invalid non-judicial order". The Court of Appeal correctly recognised that a purported exercise of non-judicial power by a superior court, like any other purported exercise of a non-judicial power, lacks legal force if made without jurisdiction. Taking the view that it did of Kable (No 1), the Court of Appeal considered itself bound to hold that to be this case. Resolution There are certainly passages in the reasons for judgment of members of the majority in Kable (No 1) which strongly support the view taken by the Court of Appeal of the holding in that case. However, what was said in that case must be interpreted in the context of what was done in that case. The appellate jurisdiction of the High Court under s 73 of the Constitution is limited to the hearing and determination of an appeal from a decision made in the exercise of judicial power. Kable (No 1) came to the High Court as an appeal under s 73 of the Constitution from an order of the Court of Appeal dismissing an appeal from the order of Levine J. That "bespeaks an exercise of judicial power" by the Court of Appeal. The High Court in determining the appeal in Kable (No 1) was limited to standing in the position of the Court of Appeal at the time when the Court of Appeal dismissed the appeal to it from the order of Levine J and making such order as the Court of Appeal could and should have made in the exercise of such judicial power as was exercisable by the Court of Appeal in the appeal to it from the order of Levine J. What the High Court did in Kable (No 1), having allowed the appeal to it and having set aside the order of the Court of Appeal dismissing the appeal from the order of Levine J, was: allow the appeal to the Court of Appeal, set aside the order of Levine J and in its place order that the application by the DPP be dismissed. The High Court could have set aside the order of Levine J and ordered that the application by the DPP for a preventive detention order be dismissed only in the exercise of judicial power. What is more important for present purposes is that the High Court could have done so only on the basis that the Court of Appeal could and should have made those orders also in the exercise of judicial power. The Court of Appeal has capacity to entertain an appeal by way of rehearing from a person or body exercising administrative power. However, there is no basis for considering the nature of the power exercisable by the Court of Appeal on the appeal to it from the order of Levine J to have been different from the nature of the power exercised by Levine J in making that order. Both Levine J and the Court of Appeal were exercising power purportedly conferred on the Supreme Court by the CP Act. The power as so conferred on the Supreme Court was in each case to be exercised only in proceedings between parties in which incidents of civil process were required to be observed. What the High Court did in Kable (No 1) is therefore consistent with the jurisdiction to make a preventive detention order, purportedly conferred on the Supreme Court by the CP Act, being judicial in character, albeit having features which made the conferral of that jurisdiction incompatible with Ch III of the Constitution. Four particular aspects of the reasoning of Gummow J, one of the majority in Kable (No 1), are of particular significance in that light. The first is his description of the CP Act as performing the "double function" of imposing substantive liabilities and conferring on the Supreme Court jurisdiction with respect to those substantive liabilities. The second is his description of the submission he accepted as being that "the jurisdiction conferred by the [CP] Act upon the Supreme Court is of such an extraordinary nature as to be incompatible with the exercise by that institution of federal jurisdiction". The third is his acknowledgement that "if a State court be invested with … a non-judicial power, no exercise of that power can found an appeal to this Court" because "this Court has no power to make a non-judicial order in place of any non-judicial order which the State court ought to have made at first instance". The fourth aspect of the reasoning of Gummow J that is of present significance is his observation that the raising of a question as to the validity of the CP Act in the proceeding before Levine J had the effect of bringing the whole of the matter in issue in that proceeding and in the Court of Appeal within the federal jurisdiction of the Supreme Court conferred under s 77(iii) of the Constitution by s 39(2) of the Judiciary Act. As Basten JA pointed out in the Court of Appeal in the present case, a single proceeding in a Supreme Court might conceivably involve discrete questions as to the exercise of judicial power and as to the exercise of non-judicial power with the consequence that the former might be encompassed within a matter within federal jurisdiction while the latter could not. However, that is not the way in which the proceeding before Levine J and the Court of Appeal was analysed in Kable (No 1). It was said that "the jurisdiction exercised by the Supreme Court was wholly federal". The better view of Kable (No 1) is that there was a single matter before the Supreme Court constituted by the disputed entitlement of the DPP to a preventive detention order under the CP Act. That matter encompassed but was not confined to whether the Supreme Court had jurisdiction and whether the CP Act was invalid as incompatible with Ch III of the Constitution. The matter extended to whether the preventive detention order applied for by the DPP, if it could be made, should be made. The order made by Levine J, upheld in the Court of Appeal, resolved the whole of that matter in the exercise of judicial power within federal jurisdiction. The order was "an adjudication to determine the rights of parties". The order made by Levine J in the purported exercise of jurisdiction invalidly conferred on the Supreme Court by the CP Act was therefore a judicial order.