What is wrong with top-down legal reasoning?


Justice Keith Mason

Sir Maurice Byers Memorial Lecture
26 February 2004

"Top-down legal reasoning" is not a term of art. In recent years it has become a term of abuse. On my researches, it entered Australian legal discourse in 1996 in the judgment of McHugh J in McGinty v Western Australia.[1]

1996 was also the year in which the High Court delivered judgments in the last two cases argued before it by Sir Maurice Byers QC, The Wik Peoples v Queensland[2] and Kable v Director of Public Prosecutions (NSW).[3] I had the uncomfortable privilege of being opposed to Maurice in Kable. I also had the pleasure of representing a plaintiff in similar interest to his client in the Political Advertising Case, his antepenultimate High Court foray.[4] These three decisions stand as remarkable tributes to his innovative and persuasive advocacy. They also illustrate legitimate judicial creativity that surfaces from time to time in every age. It is practised by all leading jurists however much some of them deny its universality or castigate those who admit it.

Both Political Advertising and Kable are connected with the topic of my address, although I hasten to add that Maurice was never accused of top-down reasoning - at least not to his face.
As you know (a beguiling preamble much beloved by Maurice), McGinty involved a challenge to Western Australian electoral laws that ensured significant disparities in the numbers of voters as between rural and metropolitan regions. The claim of constitutional invalidity was dismissed by a majority of the High Court comprising Brennan CJ, Dawson, McHugh and Gummow JJ. Toohey and Gaudron JJ dissented. Kirby J joined the Court after McGinty was argued. One gets the impression from his remarks in the 2003 Marquet decision that he would have been a McGinty dissenter.[5]

Each justice in the McGinty majority declined to find any constitutional basis for a principle of "equal value" of votes. They accepted the correctness of the Political Advertising Case, but regarded its implication of "representative democracy" as too narrow a toehold to support a constitutional proscription against grossly disproportionate electorates.

The Engineers' Case[6] established that Dicey's notion of Parliamentary sovereignty underlay the express grants of legislative power to the Commonwealth Parliament. For a time, it seemed that the Engineers' juggernaut would carry all before it, with its emphasis upon the plain meaning of the constitutional text and its rejection of federal or other unstated limitations upon the powers of the Commonwealth Parliament. But, as you know, Sir Owen Dixon found implied limitations in various areas. The most notable and enduring were to be the principles expounded in Melbourne Corporation v Commonwealth [7] a modified version of which was applied very recently in Austin v Commonwealth.[8]

Those who thought that the list of constitutional implications had closed (like the canon of scripture) upon the death of Sir Owen Dixon were surprised, even angered, by the discovery of another implied limitation upon Commonwealth legislative power, in the Political Advertising Case. The Samuel Griffith Society put the case on its blacklist, along with Mabo and other post-Dixonian heresies.

Mason CJ uttered pure orthodoxy in the Political Advertising Case, when, citing Dixon J, he said:[9]

It is essential to keep steadily in mind the critical difference between an implication and an unexpressed assumption upon which the framers proceeded in drafting the Constitution. The former is a term or concept which inheres in the instrument and as such operates as part of the instrument, whereas an assumption stands outside the instrument. Thus, the founders assumed that the Senate would protect the States but in the result it did not do so. On the other hand, the principle of responsible government ... is not merely an assumption upon which the actual provisions are based; it is an integral element in the Constitution.

In Political Advertising, this integral element of responsible government was held to have given rise to a secondary implication that a right of freedom of political speech was necessary to ensure that elected governments would continue to be responsible through Parliament to the people of Australia.

The four justices in the majority in Political Advertising (Mason CJ, Deane, Toohey and Gaudron JJ) held that the free speech right was an implication from the doctrine of representative or responsible government. The secondary implication was drawn because freedom of communication was indispensable to the efficacy of such a system of government.[10] This is the implication that attracted the hostile attention of the critics.

In Political Advertising, Mason CJ suggested a distinction between textual and structural implications when he said:[11]
It may not be right to say that no implication will be made unless it is necessary. In cases where the implication is sought to be derived from the actual terms of the Constitution it may be sufficient that the relevant intention is manifested according to the accepted principles of interpretation. However, where the implication is structural rather than textual it is no doubt correct to say that the term sought to be implied must be logically or practically necessary for the preservation of the integrity of that structure.

Brennan CJ[12] and McHugh J[13] cited this passage with approval in McGinty. Dawson J doubted the helpfulness of Mason CJ's distinction between textual and structural implications, but endorsed implications so long as they were necessary to accommodate the text of the Constitution.[14]

The McGinty majority included the three justices who dissented or partially dissented in the Political Advertising case (ie Brennan, Dawson and McHugh JJ). In McGinty, their Honours accepted the correctness of the earlier decision, but were at pains to construe its ratio narrowly. I imply no criticism by this observation. This is common law method at its purest.

McHugh J was not one of the majority in Political Advertising who had declared Part IIID of the Broadcasting Act 1942 (Cth) wholly invalid. His Honour would have struck much of it down, but for reasons considerably narrower than those adopted by Mason CJ, Deane, Toohey and Gaudron JJ. When, in McGinty, McHugh J addressed the ratio decidendi of Political Advertising he therefore had the difficult task of describing a recent decision that bound the Court of which he was a member, but that rested upon reasoning with which he disagreed. The epiphany of the joint judgment in Lange v Australian Broadcasting Corporation[15] lay yet in the future.

I have already indicated that McHugh J endorsed Mason CJ's test in Political Advertising for deriving constitutional implications. But he drew a sharp line of disagreement with two other justices who had, with Mason CJ and Gaudron J, formed the majority in that case. He said:[16]
However, I cannot accept, as Deane and Toohey JJ held in Nationwide News Pty Ltd v Wills, that a constitutional implication can arise from a particular doctrine that "underlies the Constitution". Underlying or overarching doctrines may explain or illuminate the meaning of the text or structure of the Constitution but such doctrines are not independent sources of the powers, authorities, immunities and obligations conferred by the Constitution. Top-down reasoning is not a legitimate method of interpreting the Constitution. ... [A]fter the decision of this Court in the Engineers' Case, the Court had consistently held, prior to Nationwide News and [Political Advertising], that it is not legitimate to construe the Constitution by reference to political principles or theories that are not anchored in the text of the Constitution or are not necessary implications from its structure.

Gummow J endorsed these remarks when he said that:[17]
... as McHugh J explains in his judgment, the process of constitutional interpretation by which the principle [of an implied constitutional freedom of political discussion] was derived (being an implication at a secondary level), and the nature of the implication ... departed from previously accepted methods of constitutional interpretation.

Brennan CJ[18] and Dawson J[19] were also critical of attempts to find any content in the concept of "representative democracy" from sources outside the text and structure of the Constitution itself.

I should say at the outset that, in my respectful view, McHugh J did less than justice to his two former colleagues. In the passage that he cited, Deane and Toohey JJ had referred to doctrines "which underlie the Constitution and form part of its structure" (emphasis added).[20] Furthermore, they had instanced the doctrine of representative government, which was a primary implication accepted by the entire McGinty Court. It is also unclear why McHugh J said nothing about Gaudron J's judgment in Political Advertising. Gaudron J recognised explicitly that "fundamental constitutional doctrines" could be assumed in the Constitution[21] and she included the common law as the source of revelation about the constitutional importance of free speech.[22]

There was a time when Sir Owen Dixon's views about s92 of the Constitution were contrary to the trend of existing authority. In this context he once remarked that:[23]
It is better that I should not attempt any restatement for myself of the principles upon which the decisions rest. Probably my grasp of those principles is imperfect and, as a rule, it is neither safe nor useful for a mind that denies the correctness of reasoning to proceed to expound its meaning and implications.

We may be unsure whether Sir Owen's humility was feigned, but we know for certain that these remarks were an early gambit in a quest by that great jurist to persuade his brethren to overturn existing orthodoxy on s92 of the Constitution. In this, Dixon would succeed entirely - for a time.

I have digressed, and I cannot for the life of me think why Sir Owen's observation occurred to me in the context of discussing McHugh J's critique of Political Advertising in McGinty, to which I return.

Sir Maurice's argument in Political Advertising, that the Court adopted, had invoked orthodox statements about necessary implications in aid of beguiling submissions that led the Court into the previously uncharted waters of a constitutional guarantee of freedom of speech. His final submission, as reported in the Commonwealth Law Reports, contended that "in a democracy the right to freedom of speech is part of the fabric of society. There cannot be democracy if the voters are gagged and blindfolded."[24] This appeal to principles external to the text and structure of the Constitution left the justices wide leeways of choice as to the means whereby they might bridge the gap between assumption and implication.

The concept of "top-down reasoning" proscribed by McHugh J in McGinty had been identified by Judge Richard Posner. In a frequently cited article,[25] to which McHugh J referred, Posner explained top-down and bottom-up reasoning as follows:
In top-down reasoning, the judge or other legal analyst invents or adopts a theory about an area of law - perhaps about all law - and uses it to organise, criticise, accept or reject, explain or explain away, distinguish or amplify the existing decisions to make them conform to the theory and generate an outcome in each new case as it arises that will be consistent with the theory and with the canonical cases, that is, the cases accepted as authoritative within the theory. The theory need not be, perhaps never can be, drawn "from" law; it surely need not be articulated in lawyers' jargon. In bottom-up reasoning, which encompasses such familiar lawyers' techniques as "plain meaning" and "reasoning by analogy", one starts with the words of a statute or other enactment, or with a case or a mass of cases, and moves from there - but doesn't move far, as we shall see. The top-downer and the bottom-upper do not meet.

Posner pointed out that legal reasoning from the bottom up was the "more familiar, even the more hallowed, type". But, as we shall see, he was critical of bottom-up reasoning as a genuine explanation of what happens, and he strongly endorsed the inevitability and legitimacy of top-down reasoning.[26]

Sir Maurice's advocacy was an appeal to top-down reasoning by these criteria. A political principle or theory about the importance of free speech was said to be an assumption necessarily underlying the constitutional concept of responsible government, reflected in the common law's support for free speech.

Soon after McGinty, anathemas against any form of constitutional top-down reasoning entered the currency.

McHugh J was not being complimentary when, in Gould v Brown, the Commonwealth Solicitor General's argument in favour of the validity of the cross-vesting scheme was described as involving "a lot of top-down reasoning".[27]

But just as quickly, it emerged that the charge could be hurled from different quarters. One year after McGinty, during argument in Ha v New South Wales,[28] a submission supporting the broader view of "excise" in the Constitution as including a tax on distribution was castigated by Dawson J in the following terms:[29]
The majority judgment in Capital Duplicators simply asserts that... what was achieved was a customs union and then it asserts that was an economic union and from that it asserts that it was the purpose of that union to secure control over commodities and the taxing of commodities to the Commonwealth, but it is a perfect example of top down reasoning that we have been talking of before. A customs union is not an integrated economy.

The Capital Duplicators majority railed against by Dawson J included Brennan and McHugh JJ (admittedly in the dangerous company of Mason CJ and Deane J). But the real villain in Sir Daryl Dawson's sights was Dixon. Constitutional reasoning that moved from the idea of an Australian customs union, to an integrated economy, to "excise" being a tax on distribution came directly from the observations of Dixon J in Parton v Milk Board (Vic).[30]

When the Solicitor General for the Commonwealth, Mr Griffith QC argued in support of the broader view of "excise", there was the following exchange:
DAWSON J: That is top-down reasoning and you have no basis on which to support it except the assumption that Sir Owen Dixon made.

MR GRIFFITH: Your Honour, our submission is it is not top-down reasoning. It is based on the words of the Constitution itself.

DAWSON J: That is to make an assumption as to their meaning.

The Dixonian view was to prevail in Ha's Case, albeit that what Dixon J had asserted was now underpinned by historical references, including references to the Convention Debates that the Dixon Court never openly admitted to consulting. I respectfully share Dawson J's view about this being a species of top-down reasoning, but (in light of the decision in Ha) am simply content to add Ha to the list of cases showing that top-down reasoning is not bad root and branch.

If you read the key passages from Parton, Capital Duplicators and Ha you may, I think, be forced to acknowledge the following three propositions:
(A) Top-down reasoning is part and parcel of constitutional discourse and has always been so;
(B) A top-down argument consistent with the text and structure of the Constitution may take legitimate root if adopted by an authoritative jurist or in a leading precedent; and
(C) Constitutional reasoning that invokes assumptions is facilitated by reference to common law cases and the modern practice of referring to Convention Debates and historical materials.

Dixon J once warned against confusing "the unexpressed assumptions upon which the framers of the [Constitution] supposedly proceeded with the expressed meaning of [a constitutional] power".[31] But this was a warning against sloppy thinking, not a command to disregard all assumptions.

There is a famous passage in Dixon J's judgment in the Communist Party Case.[32] Speaking of the power in s51(xxxix) to make laws with respect to "matters incidental to the execution of any power vested by this Constitution in ... the Government of the Commonwealth", Sir Owen said:
The power is ancillary or incidental to sustaining or carrying on government. Moreover, it is government under the Constitution and that is an instrument framed in accordance with many traditional conceptions, to some of which it gives effect, as, for example, in separating judicial power from other functions of government, others of which are simply assumed. Among these I think that it may fairly be said that the rule of law forms an assumption.

The passage has been frequently cited, most notably in Cheatle v The Queen,[33] where the unanimous Court pointed out that:
It is well settled that the interpretation of a constitution such as ours is necessarily influenced by the fact that its provisions are framed in the language of the English common law, and are to be read in the light of the common law's history.

I repeat, Dixon J said there are constitutional conceptions some of which are "simply assumed". The rule of law is one such assumption. Dixon mentioned two others in his paper on The Law and the Constitution,[34] namely parliamentary sovereignty and the supremacy of the Crown as a formal concept. He wrote:[35]
The fundamental conceptions, which a legal system embodies or expresses, are seldom grasped or understood in their entirety at the time when their actual influence is greatest. They are abstract ideas usually arrived at by generalization and developed by analysis. .... Sometimes indeed they are but instinctive assumptions of which at the time few or none were aware. But afterwards they may be seen as definite principles contained within the ideas which provided the ground of action. Further, when such conceptions have once taken root they seldom disappear. They persist long after the conditions in which they originated have gone. They enter into combinations with other conceptions and contribute to the construction of new systems of law and of government.

This surely is authoritative recognition of top-down constitutional reasoning.

Like many bedrock principles, the concept of the rule of law is protean. To recognise that it lies behind the Constitution leaves much room for movement (including further leeway for top-down reasoning). It is therefore perhaps unsurprising that McHugh and Gummow JJ recently observed that:[36]
In Australia, the observance by decision-makers of the limits within which they are constrained by the Constitution and by statutes and subsidiary laws validly made is an aspect of the rule of law under the Constitution. It may be said that the rule of law reflects values concerned in general terms with abuse of power by the executive and legislative branches of government. But it would be going much further to give those values an immediate normative operation in applying the Constitution.

Yet an "immediate normative operation" is surely the horse that bolted in the much-lauded Communist Party Case. Top-down theories that gain judicial acceptance cannot easily be returned to their stable or bridled.

Those who have struggled to frame an instrument hope that the hard won final text will cover all eventualities. But the best-drawn contracts, statutes and constitutions may throw up unconsidered issues. Close examination of text and context may reveal clear answers, but not always. Gaps in private contracts redound to the disadvantage of those who would enforce them. It is not so easy with statutes and constitutions that are framed as enduring instruments of governance. The language may be opaque and the area of application may become more and more removed from the original context with the passing of time. But the judicial imperative to find a workable meaning is necessarily stronger with such instruments.

Sir Owen Dixon wrote to Chief Justice Latham in 1937 suggesting that:[37]
In [s92] cases relating to transport ... I think it is almost clear that we must proceed by arbitrary methods. No doubt there will be limits but political and economic consideration will guide the instinct of the court chiefly. In time the thing will work back to some principle or doctrine.

This surely was a call to road-test theories lying outside the constitutional text and structure in order to check their consistency, with a view to adoption if accepted by the Court as an institution. It is a thousand miles away from a search for strictly necessary implications. It is driven by the unavoidable judicial function of resolving justiciable disputes in relation to a working instrument of government. In short, this was further recognition of an appropriate role for top-down reasoning.

McHugh J's anathema in McGinty stated that "underlying or overarching doctrines... are not independent sources of the powers, authorities, immunities and obligations conferred by the Constitution". Since he did so in the context of criticising Deane and Toohey JJ for finding that a constitutional implication could arise from a particular doctrine that "underlies the Constitution", it seems reasonable to add "implications" to the concepts that McHugh J said could never be sourced in underlying or overarching doctrines. I do not think that this view can stand in light of Sir Owen Dixon's compelling analysis and the case law to which I have made reference.

To that case law I would add the Political Advertising Case itself. The High Court unanimously endorsed its legitimacy in Lange, albeit underpinned by different reasoning that seems to track McHugh J's approach rather than that of the majority in the earlier case. The constitutional implication of free speech is now grounded in the interstices of the phrase "directly chosen by the people"[38] as much as in the structural concept of responsible government. But I venture to suggest that the top-down theories based upon the desirability of free speech are still quite visible. There is discussion in Lange about communications between electors and representatives being "central to the system of representative government, as it was understood at federation".[39] I submitted earlier that resort to history and common law are at times the way of pointing to the assumptions of the framers of the Constitution before moving quickly to finding a necessary implication. There is still a leap beyond logic - an entirely legitimate leap in my respectful view - from words such as "directly elected" to the free speech implication.

The relevant part of the joint judgment in Lange concludes with the following statement:[40]
To the extent that the requirement of freedom of communication is an implication drawn from ss7, 24, 64, 128 and related sections of the Constitution, the implication can validly extend only so far as is necessary to give effect to these sections. Although some statements in the earlier cases might be thought to suggest otherwise, when they are properly understood, they should be seen as purporting to give effect only to what is inherent in the text and structure of the Constitution.

This obliquely acknowledges that the ratio of Political Advertising has been completely reworked. But it does not, in my most respectful submission, reveal the processes whereby the free speech implication is found to inhere in the text and structure of the Constitution.

Something more is still at work. I dare not repeat its name.

The list of underlying or overarching principles that may come to bear upon constitutional issues will not be a large one. Any that do emerge will have to be hammered out through the dialectic, collegiate processes of decision-making in the High Court. Some ideas will surface and be rejected, others will be refined over time. Those that achieve acceptance will have been tested in the fire and beaten thin like gold. Hopefully the debate will take place without sloganeering about judicial activism or attacks ad hominem or ad feminam.

One likely contender for a constitutional assumption that will develop into a constitutional implication is a rule for resolving inconsistencies between the statute laws of different States where they clash at the margin.[41]

I come now to Kable's Case. Sir Maurice's submission about the Community Protection Act 1994 (NSW) not being a law at all was rejected. So too was a submission that a Boilermakers-style separation of powers was part of the New South Wales constitutional polity.

It was Maurice's third argument that succeeded. Its major premise was the proposition that State courts had to be kept pure vessels to receive invested federal judicial power. Its minor premise was that the 1994 Act sullied the Supreme Court of New South Wales by requiring it to exercise jurisdiction that would lower public confidence in the integrity of the judiciary.

I am only concerned with identifying the type of reasoning adopted in support of the major premise. In what follows, I imply no criticism of the premise itself.

In Kable, Sir Maurice cited the very passage from Dixon J's judgment in Australian Communist Party to which reference has already been made. He argued that the rule of law required that a citizen may only suffer loss of liberty upon conviction of an offence. From this, he moved to Chapter III's scheme for investing Commonwealth judicial power in State courts, arguing that no legislature, State or federal, might impose jurisdiction on State courts incompatible with the potential exercise of that federal judicial power.[42]

This argument prevailed at least as regards State Supreme Courts, and with some refinements.

The Constitution's express terms provided for what Gaudron J described as "an integrated Australian judicial system for the exercise of the judicial power of the Commonwealth".[43] This is found in the very text of Chapter III, so it was (with respect) an easy, though novel, step to imply the pure vessel requirement for State Supreme Courts.

But whence came the secondary implication that courts (State or Federal) in an integrated system are constitutionally required to conduct themselves in a manner consistent with "traditional judicial process"[44] lest they bring justice itself into disrepute? These are important and commendable values and I am not for a minute criticising anyone for taking them into account. Much of the burgeoning Chapter III jurisprudence proceeds from a similar proposition. My point is that this commendable notion is an assumption standing outside the constitutional text and structure. It is a principle about how judges ought to conduct themselves, how the common law sometimes required them to act and historically how they usually conducted themselves around the time that the Constitution was formed.

In Chu Kheng Lim v The Minister for Immigration, Local Government and Ethnic Affairs[45] Brennan, Deane and Dawson JJ said that the legislative power of the Commonwealth does not extend:
... to the making of a law which requires or authorizes the courts in which the judicial power of the Commonwealth is exclusively vested to exercise judicial power in a manner which is inconsistent with the essential character of a court or with the nature of judicial power.

In recent times there have been many statements by High Court justices to similar effect, some of them identifying particular matters as constituting essential characteristics of the judicial process that Parliament may not infringe.[46]

But these now constitutional desiderata are not to be found in the text or structure of the Constitution. They are not nestling inside the meaning of words like "court" or "matter". Nor are they implications that are logically or practically necessary for the preservation of the integrity of the constitutional structure.[47] One can readily point to constitutional democracies that function without the underpinning of the entrenched principles of our growing Chapter III jurisprudence.

In my opinion, this Chapter III jurisprudence should be recognised for what it is, a species of top-down reasoning that has received legitimate acceptance through the time-honoured processes of constitutional litigation.

Recently, in Roxburgh v Rothmans of Pall Mall Australia Ltd[48] Gummow J cited with approval McHugh J's hostile reference to top-down reasoning in McGinty. Gummow J applied it outside the realm of constitutional law, when cautioning against judicial acceptance of "any" all-embracing theory of restitutionary rights and remedies founded upon a notion of "unjust enrichment". Gummow J continued:
To the lawyer whose mind has been moulded by civilian influences, the theory may come first, and the source of the theory may be the writing of jurists not the decisions of judges. However, that is not the way in which a system based on case law develops; over time, general principle is derived from judicial decisions upon particular instances, not the other way around.

This is not the place for me to engage in debate about the concept of unjust enrichment. On the particular issue, I content myself with the observation that the varieties of restitutionary theory deriving from the case-based scholarship of jurists like Goff & Jones, Birks and Burrows are as entitled to compete for acceptance in the judicial market-place as property-based theories, or theories based upon unconscionability, or theories that strive to maintain at all costs the separate integrity of Equity (with a capital-E).

Some scholars distinguish between "high theory" and "middle theory", using the latter as a description of a construct that is "case-law-focussed".[49] Presumably Gummow J had the former in his sights. I suggest, however, that the difference is only one of degree. No one in the real world of judicial decision-making seeks to make everything "all tidy and four-square like the Marx brother who took shears to the bits of clothes which stuck out of his suitcase", to use a telling phrase of Professor Tony Weir.[50] Conversely, there can be nothing wrong per se in using a theoretical construct to criticise a precedent that does not fit: this happens frequently in appellate advocacy and decision-making.

We need theories for road maps or hypotheses and for deciding whether an existing authority is to be applied, distinguished or overruled. Even the professed incrementalist is confronted with deciding what is "an increment too far".[51] Like Monsieur Jourdain and prose, we judges may be ignorant of any or all of the theories that shape our reasoning, but we are truly ignorant if we deny their existence. We look to scholars like Salmond, Fleming, Treitel, Birks and Stapleton to map out structures for understanding fields of law or the essence of particular causes of action. As with the grand summaries of our greatest jurists (for example Dixon J's exposition of estoppel), these theories help explain the jumble of existing case law. They also point the way towards orthodox developments and offer guidance in knowing when to distinguish or overrule apparent departures from orthodoxy.

Such assistance is prized in the modern era where "legal coherence" is valued highly.[52] Naturally, we must guard against theories turning into a dogma that may "tend to generate new fictions in order to retain support for its thesis", as Gummow J put it in Roxburgh.[53] But this problem should not be exaggerated. I suspect that all theories used by judges as working tools have exceptions and qualifications.

I do most firmly join issue with Gummow J's suggested antipathy between "civilian" theory-based discourse on the one hand and the "case law" system on the other; implying that it is a mark of the latter that general principle is always derived from judicial decisions upon particular instances.

Posner's famous article demonstrated roles for top-down and bottom-up reasoning outside the realm of constitutional interpretation. But a categorical assertion about top-down theoretical reasoning being alien to bottom-up common law method is as startling as it is at variance with Posner's views.

I do not suggest for a moment that Gummow J represented his views as consonant with Posner. Only Posner's definition of "top-down reasoning" was cited. But I do respectfully submit that Gummow J's dichotomy is a startling misdescription of what happens day in and day out in the High Court of Australia, and on red letter days in inferior Australian courts.

I start by explaining Posner's thesis in a little more detail.

Posner gives as examples of familiar and hallowed bottom-up reasoning the principle that interpretation of a statute must start from its words, and the technique of reasoning by analogy from decided cases. But Posner wrote that "there isn't much to bottom-up reasoning" and he was highly critical of those who see the top-down and bottom-up approaches as dichotomous.

Unlike McHugh J, Judge Posner did not condemn top-down reasoning.

For Posner, decided cases might offer material for creating or testing a theory about a field of law. But without a theoretical template to view them or to know when an analogy is close and legitimate they are no more than decided cases. In his words: "But there must be a theory. You can't just go from case to case, not responsibly anyway."

Thus, constitutional and textual interpretation will involve suppositions or theories about original intent, legislators' intent, stare decisis, the role of context, the relevance of international norms, presumptions against overturning deeply-held values of the common law etc etc etc. Judges must work through these and many other issues when addressing disputes presented for resolution.

It is no different in the realm of case law. To say that Donoghue v Stevenson[54] is canonical tells you little about when and how its principles are to be applied in later cases. And when another major planet enters the solar system (Hedley Byrne & Co v Heller & Partners[55] for instance), we need theories and techniques to know how to respond. The answer may differ between England and Australia, because different forces may be at work. How these are discerned and applied by judges involves techniques at the highest levels of abstraction, ie theories.

In Posner's words, "bottom-up reasoning is not reasoning but is at best preparatory to reasoning ... legal reasoning worthy of the name inescapably involves the creation of theories to guide decision". Posner listed examples of top-down theories associated with well-known scholars, including one or two of his own. He naturally acknowledged the contestability of all theories.

Unlike scholars, judges tend not to enunciate the theoretical underpinnings of their judicial worldview. We are simply too busy deciding cases to step back and contemplate the broader patterns that underlie our words and actions. And we are reluctant to offer a broadside to scholars and other jurists unless it is really necessary. Some of us get snippets of time during sabbaticals to pursue studies in particular areas. We may open windows that show bigger pictures that help shape our understanding of the daily task. On these occasions some of us may range over wide fields of law or legal theory. For others, refreshing and occasionally useful insights may come through religious studies, history, the philosophy of the mind or probability theory.

We judges, and those whom we serve, are (I believe) the better for these glimpses into a world that is broader than the law viewed as a closed circle of self-referential ideas. That world is a reality that impacts upon the law at its every step. Why should we turn it away at the door for fear that it may enter our deliberations from the top down?

Some big picture ideas are plain wacky and others may be irrelevant or harmful to legal discourse within the confines of the judicial oath. But not all. All ideas, theories and concepts must be contestable and available for scrutiny in accordance with the processes of judicial accountability to which all are subject in differing ways.

My concern is with those who deny the universality and legitimacy of "top-down" reasoning that is part of the common law tradition. I am not advocating a role for the judicial superman (or woman) who does nothing but bring extra-legal theories or concepts into legal discourse. I have yet to meet such a character. He or she is in the class of the unicorn, as non-existent as the legal purist who is said to bring nothing to the task but a high judicial technique that finds everything within the four corners of a revealed but closed canon of legal scripture.

Top-down and bottom-up reasoning are not converse ways of approaching a single problem. As Posner puts it, "top-downer and bottom-upper do not meet". Rather, the two concepts seek to capture clusters of different types of legal reasoning each of which is widely practised by everyone (including those who sometimes profess denial). If you don't believe me, I suggest that you read Kirby J's "I told you sos" in dialogue with some of his more legalistic brethren in the footnotes to his reasons in Cattanach v Melchior.

Some types of top down reasoning are illegitimate and their very method of introduction offends orthodox judicial method. Philosophies and concepts that flaunt established principle, or that are applied by individual judges in the teeth of existing authority or the plain text of statutes or constitutions must be rejected. But that is because they are poor theories, or conflict with binding precedent, or fail to gain judicial acceptance. It is not because they may originate in academic writings, or decisions from overseas legal systems, or the insight of an individual judge.

Theories may, in Posner's words, be invented or adopted. Some can be traced to their birthplace which may be a single judge or an academic writer. None, I suggest, can truly be described as "deriv[ing] from judicial decisions upon particular instances, not the other way around", as if the two were mutually exclusive.

Lord Atkin had the parable of the Good Samaritan as much as the existing case law in mind when he enunciated the morality-based neighbour principle that turned much of the earlier law on its head.[56] Lord Wright imported an exotic plant into English jurisprudence when he introduced the principles of the American Restatement of Restitution.[57] This was a grand top-down theory (of still debatable content) that would displace the implied contract theory of quasi-contract and may yet do further damage to inherited certainties.[58] Many ideas have entered the common law when a judge picked up a theory from an academic article, road-tested it and ran with it.

My difficulty with Gummow J's description of judicial method is that it offers a false dichotomy and presents only half the picture.

The last 30 years has been an era in which the High Court has generally welcomed the insights of comparative law, international law, academic theory, social history, moral discourse etc etc etc. Read the judgment of Gleeson CJ in Cattenach for an instance where all of these factors are brought into focus in addressing a novel legal issue. But even for less exotic topics than the one considered in Cattenach, the connections and disconnections between existing precedents within our judicial system are often perceived "top-down" through such lenses. It is obviously true to say that general principles derive from existing judicial decisions. But that is only part of the picture. Other factors are at work, from the top-down as it were. They may provide the tools for "deriving" principles from existing case law. They may assist in "prioritising" conflicting decisions or introducing or spurning legal ideas from abroad. Top-down theories (of the finest kind) are the usual spur for a major shift in legal reasoning.

Let me illustrate by two examples drawn from the recent law of negligence.

Thirty years ago the leading courts in Australia and England treated causation issues as questions of fact to be answered with no more than a hearty dose of robust common sense. How things have changed. In March v E & MH Stramare Pty Ltd,[59] Mason CJ said that the "but for" test, applied as an exclusive criterion of causation, "yields unacceptable results and ... the results which it yields must be tempered by the making of value judgments and the infusion of policy considerations". Today, even this profound acknowledgement is just the signpost to further pathways for approaching causation issues in a principled (ie theory-based) manner. In the last decade you can hardly read an appellate decision on the topic that fails to acknowledge the insights of Professor Jane Stapleton. Many of her views stem from pure philosophy mediated to lawyers through Hart & Honore's Causation in the Law. Of course, Stapleton has worked with the caselaw, but in a highly critical manner. Theories have been tested against the decided cases. The decided cases have been tested against the theories. In turn, Stapleton's constructs have been tested and applied (to a degree) by the High Court and the intermediate appellate courts.

This judicial reception is top-down reasoning of the highest legitimacy.

My second example relates to concepts that have simply been introduced over the top of existing precedent because of what Holmes described as the "felt necessities of the time, the prevalent moral and political theories, intuitions of public policy, even the prejudices which judges share with their fellow men".[60]

A lawyer who gave advice today about the law of negligence by reference to principles derived from High Court decisions when Sir Gerard Brennan was Chief Justice would be an easy target in a professional negligence claim. There have been tremendous changes over the last decade. For example, the notion of general reliance has been rejected. Some justices of the High Court now pay open regard to the availability and cost of insurance as a factor relevant to imposition of a duty of care.[61] There is much emphasis upon taking responsibility for one's own actions, so much so that it is now built into the duty of care owed to footpath pedestrians.[62] Categories of strict liability are disappearing. Non-delegable duties are harder to find. The seismic shifts have not all been in the one direction, as Brodie and Tame demonstrate.

I am not concerned to debate whether these trends show that the earlier law was wrong. That is an irrelevant question for anyone who is not a member of the High Court. For everyone operating in the law below the High Court, what that Court said was right in 1984 was right in 1984 and what it says is right in 2004 is right in 2004.

My point is that these dramatic swings of the negligence pendulum didn't just emerge by deductive reasoning from the earlier case law. If this were the whole story, one would not expect to see the violent shifts that are now the norm in tort law. Policy issues crop up frequently - and I don't mean just the "policy of the law" found in the reports of decided cases.[63] Several extra-legal policy factors have entered the recent law of tort from the top-down.[64]

In my submission, these changes have been the product of entirely legitimate species of top-down reasoning adopted by the High Court. They were derived from much more than reading earlier precedents. In Tame's Case, the learning from psychiatry and the philosopher's call of coherence were too strong for old distinctions to hold. The enthusiastic judicial reception of the notion of taking care for one's own safety reflects a public mood of impatience against the culture of ambulance-chasing and blaming, as well as concern about the prohibitive cost of state-run and private insurance. It will be obvious that I imply no criticism of the judicial method that has influenced these fundamental shifts in tort law. I refrain from suggesting that they are instances of judicial "activism" that is widely-applauded, but only because "judicial activism" is an overworked cliché that lies mainly in the eye of the beholder.

Judges must listen to counsel and each other. And they must bow to superior judicial authorities and the ineluctable texts of statutes and constitutions. A judge who, in Posner's words, "invents or adopts a theory about an area of law" will always have an uphill battle to achieve its acceptance. No single jurist, not even a Dixon urged on by a Byers, can work in isolation or free of the constraints of judicial method.

To revert to Holmes, "we have too little theory in the law rather than too much".[65] Theories are essential, including those introduced from outside local case-law. There is nothing wrong with theories, even grand theories. They must of course gain acceptance through the proper exercise of judicial power, ultimately by the High Court of Australia.

In the final analysis, the justices of the High Court will decide what theories bear upon the structures of the law from time to time. They are the keepers at the gate that leads to and from the vast world of ideas. It would be sadly misleading if they saw themselves as no more than guardians within an enclosed cave.[66]

1 (1996) 186 CLR 140.
2 (1996) 187 CLR 1.
3 (1996) 189 CLR 51.
4 Australian Capital Television Pty Ltd v The Commonwealth; New South Wales v The Commonwealth (1992) 177 CLR 106.
5 See Attorney-General (WA) v Marquet [2003] HCA 67, 202 ALR 233 at [160]-[164]. He observed at [104] (ruefully?) that no attempt was made to reopen the holding on the constitutional implications decided in McGinty.
6 Amalgamated Society of Engineers v Adelaide Steamship Co Ltd (1920) 28 CLR 129.
7 (1947) 74 CLR 31.
8 (2003) 77 ALJR 491.
9 (1992) 177 CLR 106 at 135. Mason CJ cited Dixon J in Australian National Airways Pty Ltd v The Commonwealth (1945) 71 CLR 29 at 81 in support of the first sentence.
10 (1992) 177 CLR 106 at 138-40 per Mason CJ, at 168 per Deane and Toohey JJ (who incorporated their reasons in Nationwide News Pty Ltd v Wills (1992) 177 CLR 1 at 72-75), at 208-214 Gaudron J.
11 (1992) 177 CLR 106 at 135.
12 (1996) 186 CLR 140 at 168-9.
13 Ibid at 231.
14 Ibid at 184-5.
15 (1997) 189 CLR 520.
16 McGinty v Western Australia (1996)186 CLR 140 at 231-2, footnotes omitted.
17 Ibid at 291.
18 Ibid at 169-71.
19 Ibid at 184-5.
20 Nationwide News Pty Ltd v Wills (1992) 177 CLR 1 at 70.
21 (1992) 177 CLR 106 at 209). Gaudron J cited Australian Communist Party v The Commonwealth (1951) 83 CLR 1 at 193 per Dixon J, a passage discussed below.
22 (1992) 177 CLR 106 at 211-212.
23 Riverina Transport Pty Ltd v Victoria (1937) 57 CLR 327 at 362-3.
24 177 CLR 106 at 123.
25 Richard A Posner, "Legal Reasoning from the Top down and from the Bottom Up: The Question of Unenumerated Constitutional Rights" (1992) 59 U Chi L Rev 433. A revised version appears as a chapter in Posner's Overcoming Law, 1995 Harvard University Press.
26 Cf McHugh J's footnote in McGinty at 232 (247).
27 Transcript 8 April 1997.
28 (1997) 189 CLR 465.
29 Transcript 11 March 1997.
30 (1949) 80 CLR 229 at 260. See Capital Duplicators Pty Ltd v Australian Capital Territory (No 2) (1993) 178 CLR 561 at 585-7.
31 Australian National Airways Pty Ltd v The Commonwealth (1945) 71 CLR 29 at 81, cited by Dawson J in McGinty at 184. See also n9 above.
32 Australian Communist Party v Commonwealth (1951) 83 CLR 1 at 193.
33 (1993) 177 CLR 541 at 552.
34 Reproduced in Jesting Pilate, 1965 Law Book Company pp38-9, 42.
35 Ibid at p38.
36 Re Minister for Immigration and Multicultural Affairs; Ex parte Lam (2003) 77 ALJR 699, 195 ALR 502 at [72].
37 Letter 1 June 1937. Quoted by Bennett, Keystone of the Federal Arch p67.
38 Constitution, s24.
39 Lange at 560 per curiam.
40 (1997) 189 CLR 520 at 567.
41 See Port MacDonnell Professional Fisherman's Association Inc v South Australia (1989) 168 CLR 340 at 374, Pearce v The Queen (1998) 194 CLR 610 at 630, Lipohar v The Queen (1999) 200 CLR 485 at 553-4.
42 189 CLR at 54-55, 61-62.
43 Ibid at 102.
44 Ibid at 98 per Toohey J.
45 (1992) 176 CLR 1 at 27. See also per Gaudron J at 55, per McHugh J at 68.
46 See eg Nicholas v The Queen (1998) 193 CLR 173 at 185 [13]-[14] per Brennan CJ, 208 [73] per Gaudron J, 220-221 [112] per McHugh J, 232 [146] per Gummow J. See generally, Justice McHugh, "Does Chapter III of the Constitution protect substantive as well as procedural rights?" (2001) 21 Aust Bar Rev 235 esp at 237-241, The Hon JJ Spigelman AC, "The Truth can cost too much: The principle of a fair trial" (2004) 78 ALJ 29 esp at pp32-33.
47 Cf the passage cited at fn 11 above.
48 (2001) 208 CLR 516 at 544.
49 See Jane Stapleton, "Comparative Economic Loss: Lessons from Case-law-focussed 'middle theory'" 50 UCLA Law Review 531 (2002).
50 Tony Weir, Book Review of An Analysis of the Economic Torts (2002) 118 LQR 164 at 165, quoted by Stapleton, op cit p580.
51 The phrase is that of Beldam LJ in Barrett v Ministry of Defence [1995] 3 All ER 87 at 95.
52 See Sullivan v Moody (2001) 207 CLR 562, Cattanach v Melchior (2003) 199 ALR 131 at [35] (Gleeson CJ).
53 (2001) 208 CLR 516 at 545 [74].
54 [1932] AC 562.
55 [1964] AC 465.
56 See Geoffrey Lewis, Lord Atkin, Butterworths 1983 pp57-8.
57 See Fibrosa Spolka Akcyjna v Fairbairn Lawson Combe Barbour Ltd [1943] AC 32 at 61-64. The Restatement was published in 1937.
58 See Justice Keith Mason, "Where has Australian restitution law got to and where is it going?" (2003) 77 ALJ 358.
59 (1991) 171 CLR 506 at 516.
60 Holmes, the Common Law, 1881, Little Brown & Co, p1.
61 See Esanda Finance Corporation Ltd v Peat Marwick Hungerfords (1997) 188 CLR 241 at 282-3 (McHugh J), Brodie v Singleton Shire Council (2001) 206 CLR 512 at 595-6 (Kirby J), New South Wales v Lepore (2003) 195 ALR 412 at 424-5 [36] (Gleeson CJ).
62 Brodie at 581 [163].
63 Cf Cattenach at [73]-[75] per McHugh and Gummow JJ.
64 See generally D A Ipp, "Policy and the swing of the negligence pendulum" (2003) 77 ALJ 732.
65 Oliver Wendell Holmes, "The Path of the Law" 10 Harv L Rev 457 at p476 (1897).
66 I am indebted to Michael Coper, David Ipp and Leslie Katz for their suggestions about earlier drafts. I also acknowledge the research assistance of Michael Rehberg and Tim Breakspear.