Aronson, Mark --- "Unreasonableness and Error of Law" [2001] UNSWLawJl 26; (2001) 24(2) UNSW Law Journal 315

	Home \| Databases \| WorldLII \| Search \| Feedback University of New South Wales Law Journal

[1] Judges have long denied the existence of a fixed distinction between errors of law and errors of fact, and legal literature abounds with derision and scorn for those who attempt it.[1] The topic nevertheless holds an endless fascination for practising and academic lawyers alike, and even the Commonwealth Parliament has attempted a definition in the context of the Federal Court’s former migration jurisdiction.[2] Almost everyone concedes that there is a degree of manipulability in the distinctions between legal and factual errors. The High Court said that the distinction between questions of law and fact is ‘vital’ in many legal contexts, but it also acknowledged that ‘no satisfactory test of universal application has yet been formulated’.[3] The clear implication was that the law-fact distinction produces different results in different fields of law.

[2] If the distinction depends on the context, it follows that philosophical[4] distinctions are as likely to distract as to assist. Some of the jurisprudential literature starts out confidently enough, with definitions of law and fact. A question of fact involves an inquiry into whether something happened or will happen, and is quite separate from any assertion as to its legal effect.[5] A question of law involves the identification and interpretation of a norm which is usually of general application. That distinction quickly becomes blurred, however, by the difficulties of classifying the interactions between norm and fact. The difficulties in classifying the application of law to the facts as found occupy the bulk of this article. However, there are further difficulties in classifying the processes leading up to the finding of facts.

[3] Fact-finding inevitably involves a prior knowledge of which facts might be legally relevant. We cannot know which facts to look for unless we know why we are looking, and it is the law which tells us that. The law also requires the conversion of doubts, uncertainties and hidden or explicit assumptions as to how the world works into positive findings of fact.[6] For example, a finding in a contract dispute that promises were probably exchanged becomes a finding that they were exchanged, because the legal process itself helps reduce highly contingent factual possibilities into relatively straightforward factual propositions before declaring the legal result.[7] Furthermore, in any practical or moral sense, no court could contemplate finding the facts without having some idea of the possible legal consequences of a finding one way or the other. Those consequences, therefore, set the level of persuasion for the fact finder in every case.[8]

[4] The contingent nature of the process leading to a seemingly straightforward finding of facts is sometimes matched by the equally contingent process of finding the law.[9] The issues raised by the choices offered between legal rules, or different versions or interpretations of legal rules, are fairly clearly questions of law themselves. However, the ability to choose between competing rules or formulations of rules is often a product of the degree of specificity or generality with which the facts may have been found. None of this is intended as an argument for fact scepticism,[10] or for dismissing the law-fact distinction as a philosophical illusion or lawyer’s charade.[11] It does, however, point to the distinct possibility that for legal purposes, questions of fact can contain elements of law.

[5] Causation, for example, is a question of fact in negligence, even though it may require a selection between causal conditions according to value judgments and policy factors, for the purpose of attributing legal responsibility.[12] Some of those policy choices clearly have legal elements to them.[13] Causation is also a question of fact in the law of obligations in calculating losses, although the same caveat as to policy content applies.[14] Causation may have a different meaning for the purposes of determining refugee status, but once again, this is because refugee law has its own legal, factual and policy contexts, and once again, it is still a question of fact.[15] If one can take causation’s policy elements as including legal questions, the net result is that causation can become a question of law when a court limited to such questions wishes to lay down legal parameters to guide the tribunal of fact; hence the large number of appellate decisions on an issue bearing the overall classification of a question of fact.[16]

[6] The law’s assessment of the reasonableness of conduct or decisions exhibits the same multiple features. The standard of reasonableness performs different functions within the tort of negligence. It is generally classified as an issue of fact, but the large number of judicial expositions as to how the reasonableness standard is to operate in particular areas is clearly designed to set guidelines or parameters for future cases. As with causation, the reasonableness standard can become a question of law to the extent that a court limited to such questions intervenes to lay down general guidelines governing the tribunals of fact. The last 20 years or so have seen the general tort of negligence replace more specific tort liability rules covering highway authorities,[17] occupiers of land and premises,[18] and the escape of fire or dangerous things from a defendant’s property.[19] This represents a shift from more complex liability regimes to a general negligence regime, with fewer hard and fast rules. That should produce more questions of fact and fewer questions of law. However, it will not impede the exposition of general legal guidelines for these areas. In the area of medical negligence claims, the High Court has insisted on setting its own standard of care, rather than deferring to the standards and opinions of reputable medical practitioners. At the same time, it has laid down general legal principles to structure this question of fact.[20]

[7] Causation and reasonable care, therefore, can be both fact and law, depending on context. They arise in contexts where a court has to assess what happened and why, and what ought to have happened and who should be responsible for not having brought that about. So far as reasonableness is relevant, it is relevant to the act of judging the legal outcome of acts or conduct, rather than the acceptability of an official decision. In that context, an unreasonable want of care, for example, takes the defendant’s balance of precaution and risk beyond the realm where the court merely disagrees with it, to the realm where the court concludes that there is no room for reasonable minds to differ as to what should have been done. The reasonableness standard, therefore, is premised on the notion of relative autonomy. In negligence, it is the relative autonomy of defendants to judge for themselves how to act. The autonomy is relative, because the transition from reasonable to unreasonable marks the crossing of the boundary from judicial abstention to intervention.

[8] The reasonableness standard performs a similar function in the law of judicial review, but it is more difficult to explain or justify in that context. Judicial review consistently denies review on the merits, which is a matter left for the judgment or discretion of the bureaucrat or agency concerned. All of the judicial review grounds emphasise the relative autonomy of the bureaucracy. Provided bureaucrats do the job which Parliament has given them, it is not the role of judicial review to intervene simply because the court might have viewed the facts differently or preferred a different outcome. Legality and merits are strictly separate. Review for so-called Wednesbury[21] unreasonableness, however, challenges the dividing line between merits and legality, because its trigger is the poor quality of the bureaucratic decision. It has, for that reason, been the subject of considerable debate and uncertainty. Many cases say that the unreasonableness has to be extreme.[22] They have yet to say that it should also be obvious. Unless they take that extra step, Wednesbury will require courts to follow counsel for the challenger through the minutiae of the impugned decision.[23]

[9] As originally formulated in the Wednesbury case itself,[24] a bureaucratic decision was reviewable if it was so unreasonable that no reasonable bureaucrat similarly placed would have made the same decision. As with the reasonableness standard in medical negligence cases, Wednesbury’s standard is in fact set by the judges, who do not defer to some notional body of reputable bureaucrats similarly placed to the one whose decision is challenged.[25] English case law has suggested that something might be unreasonable in a Wednesbury sense if it is morally outrageous or defies logic,[26] although there are some doubts about the latter element, at least, for Australia.[27] There is some debate in Australia as to whether Wednesbury should be confined to reviewing a bureaucrat’s choice of options in the exercise of a statutory discretion,[28] or whether it can still cover the fact finding process.[29] Other aspects of Wednesbury are still uncertain,[30] but the biggest debate concluded some considerable time ago.

[10] No doubt prompted by the threat which review for unreasonableness posed to the classic distinction between the merits of a bureaucratic decision and its legality, Professor de Smith had opposed the evolution of unreasonableness into a ground of review in its own right. Unreasonableness had long been used as a basis for inferring, where appropriate, that the bureaucrat must have misunderstood the governing law. De Smith thought that it should be no more than evidence of some other error, rather than constitute an error in its own right.[31] He lost that debate, both at common law,[32] and in Australia’s principal statutory codification of judicial review, the Administrative Decisions (Judicial Review) Act 1977 (Cth).

[11] For the most part, the sheer pressure of cases using unreasonableness as a fully fledged ground of judicial review not only overwhelmed, but also overlooked, de Smith’s opposition. The exception makes an interesting connection with doctrine largely outside the area of judicial review. Lord Diplock said in Council of Civil Service Unions v Minister for the Civil Service:

[12] Lord Diplock’s use of ‘irrationality’ equated in this passage to Wednesbury unreasonableness.[35] His Lordship was concerned to reduce the more specific grounds of judicial review to three over-arching principles,[36] and that led him to perform two elisions. The first was to move from the treatment of an unreasonable outcome as evidence that the decision-maker might well have misunderstood the relevant law or overlooked a critical and required criterion, to treating the unreasonable outcome as a legal error in its own right. The second elision was to treat unreasonable outcomes as simply one manifestation of breaching a requirement that all administrators act rationally, a requirement he regarded as one of the pillars of the administrative state under law.

[13] The slide from unreasonableness as evidence of legal error to unreasonableness as error per se has been urged in two other areas, both closely related. The issue has arisen in the context of challenges to findings of primary facts – what happened, who did it, and why? It has also arisen in the context of applying legal standards to the primary facts as found.

[14] The general principles concerned in challenging primary findings of fact are well established. The task of determining the primary facts may be shaped by legal requirements as to natural justice,[37] procedure or evidence, but on the assumption that there is no issue as to adherence to those requirements, fact-finders commit no legal error simply by getting their facts wrong, even drastically wrong. There are exceptions, but they are strictly limited.

[15] The common law of judicial review recognises the possibility that Parliament may have intended the validity of bureaucratic or inferior court action to be dependent upon the judicial review court’s opinion as to the existence of a requisite fact. Such cases used to be rare, because they require the judicial review court to redetermine the critical fact from scratch, no matter how carefully the administrator may have performed that task.[38] There is recent evidence of a revival of judicial interest in the possibility of imputing the relevant intention to Parliament where the relevant decision-maker is not an inferior court. The author has argued elsewhere that any such exercise in implying parliamentary intent has as much of the judge in it as an actual intent of Parliament. The danger lies in converting the judge’s sense of the importance of the issues at stake into an implied statutory requirement that administrators cannot get their facts wrong.[39]

[16] Statutory restatements of judicial review grounds sometimes allow review for factual error per se, where the decision-maker’s belief in the existence of a fact was critical to their reasoning process. However, this ground is tightly circumscribed,[40] and is not presented as an error of law in its own right. It amounts simply to a statutory extension of judicial scrutiny into the realm of fact-finding.

[17] Australia’s intermediate courts of appeal typically have even broader extensions of judicial power beyond mere error of law, but once again, that is a function of the statutory grounds of appeal deliberately extending into the realm of fact.[41] The net result is that the test for the ‘no case’ submission imposes considerably stricter limits on the trial judge than apply to appeal courts. Whether (taken at its highest) there is evidence which could sustain a verdict is the legal test. It is a question of law, and not much more than a mere fragment of evidence on the material issues is needed to meet the test. There must be some evidence, of course, because fact-finding without any evidence to support it has always been an error of law in its own right. However, if there is some evidence, then the test is not whether it is tenuous or comes from an untrustworthy source, or whether the evidence to the contrary is overwhelming, or whether a verdict would be unreasonable, or even perverse.[42] Indeed, whilst there have been some formulations in terms of whether evidence ‘ought reasonably satisfy’ a jury, the importation of ‘reasonably’ has been decried as a worrying and potentially misleading distraction.[43] Chief Justice Spigelman said that ‘reasonably’ is a ‘weasel word’ in this context, liable to misuse in the hands of another Lord Denning.[44] Justice Phillips similarly declared it ‘a distraction’,[45] and their Honours have cited impressive lines of authority in support. This includes Chief Justice Mason’s judgment in Australian Broadcasting Tribunal v Bond (‘Bond’),[46] in which he doubted the applicability of English decisions[47] importing the word ‘reasonable’ into the ‘no evidence’ ground of judicial review.

[18] Subject, then, to special considerations relating to jurisdictional facts, and perhaps, prescribed legal standards of persuasion, it is an error of law to make a wrong finding of primary fact only if there was no evidence to support it, or if there may have been fragments but not enough to support the finding on any view of it.

[19] Justice Kirby has long thought that this is too demanding. His Honour said in Azzopardi v Tasman UEB Industries Ltd (‘Azzopardi’)[48] that ‘perverse’ findings of primary facts should also constitute errors of law, in contexts where the decision-maker gave reasons or was obliged to give reasons. His Honour stated his proposed test for such contexts thus:

[20] His Honour failed to persuade Glass JA (with whom Samuels JA agreed). Justice Glass stuck to the old divisions between the three stages, and said as regards the first stage (namely, primary fact-finding) that the test was the same, regardless of whether there was a jury. As regards the second stage, Glass JA reaffirmed that any materially relevant misstatement of the governing law is an error of law per se. As regards the third stage, Glass JA said that the application of the law to the primary facts as found can sometimes amount to such a misfit as to justify an inference that the decision-maker must have misunderstood the law. It can also amount to an error of law in its own right ‘if the primary facts found are necessarily within or outside a statutory description and a contrary decision has been made’.[50]

[21] The reasonableness standard can therefore be both fact and law in negligence and administrative law, but it is still fact alone in primary fact-finding.[51]

[22] The English judiciary’s appellate jurisdiction over the taxation tribunal system[52] was limited by statute in Edwards v Bairstow[53] to questions of law. There was no challenge to the tribunal's detailed findings of primary fact recounting what the taxpayers had done, and why. The only issue was whether a profit on the purchase and resale of machinery was in the nature of a capital gain, and therefore not taxable as income. The answer to that question turned on whether the transaction constituted ‘an adventure in the nature of trade’, and there had been numerous precedents indicating the sort of features which could give asset sales the characteristics of trade. The tribunal had held initially that it could not be taxable, simply because it was an isolated transaction. Justice Upjohn told it to reconsider, on the basis that even an isolated transaction could be trade. Viscount Simonds agreed that the tribunal’s initial view had shown a misunderstanding of the breadth of the term ‘trade’, and that this was in itself an error of law.[54] However, the appeal to the House of Lords came after the tribunal had reconsidered.

[23] The tribunal adhered to its original conclusion, but this time made the jump from the facts to the legal conclusion without further explanation. Could legal error nevertheless be inferred? The House of Lords thought so, because the facts could not have been clearer.[55] Their Lordships acknowledged that the legal definition of trade was blurred at the edges. They said that the tax tribunals could therefore characterise facts falling within the penumbra as trade or not without legal error.[56] This, however, was not a case where the facts could reasonably permit an outcome either way. Here, the facts were within the core of the legal notion of trade, despite only an isolated transaction being involved.

[24] The amounts of tax in dispute were not huge, but the revenue authorities had thought it necessary to take their appeal as far as the House of Lords to resolve what had been thought to be a conflict between the English and Scottish cases. There were some broad statements of principle from the English courts that whether transactions amounted to trade was a pure question of fact, and some equally broad Scottish statements that they were either pure law, or at least mixed law and fact. Their Lordships reconciled the two bodies of authorities, by observing that only a question of fact was involved if the facts allowed a conclusion either way, and that a question of law was involved if the facts led irresistibly to only one legal conclusion. The tax tribunal’s legal error, therefore, was apparent in its initial decision, and inferred when it had tried to hide its legal reasoning.

[25] As an alternative to expressing this as a process of inferring legal error from the mismatch between facts and conclusion, Viscount Simonds said that one could ‘take a short cut and say that [the tribunal members] have made a wrong inference of law’.[57] His Lordship’s preference,[58] however, was to say that the conclusion that no trade was involved had been an inference of fact, drawn from the primary facts but flawed by legal error. That error had been to draw a conclusion ‘without any evidence or upon a view of the facts which could not reasonably be entertained’.[59]

[26] Lord Radcliffe also introduced ‘reasonableness’ into an area hitherto dominated by starker (and presumably more stringent) formulas:

[27] The critical interpretative issue was whether their Lordships had intended to push more than marginally beyond the boundaries of the ‘no evidence’ and ‘not open’ categories, which are already legal errors in their own right in the context of challenges to primary fact-finding.

[28] There is more debate about the importation into Australian law of a ‘reasonableness’ standard, or more precisely, of a reasonable margin for disagreement. One might start with Chief Justice Jordan’s ‘classic’[61] exposition in Australian Gas Light Company v Valuer-General (‘Australian Gas’):

[29] Some of these propositions relate to challenging primary fact-finding, which has already been discussed. The remaining questions focus on when ‘meaning’ is a question of law (when it is not ‘ordinary’), and with the application of the law to the primary facts. His Honour distinguished between situations in which only one result was legally possible, and those in which the application of the law to the primary facts could produce a result either way. This article challenges the workability of the distinction between ordinary and special meanings, and asks whether in any event, the ‘either way’ analysis should be limited to situations where the legal term in question carries its ordinary English meaning. It is argued that it also applies where the legal term itself has a broad range of possible applications, and is thus one of ‘degree’ or ‘evaluation’. Finally, the article considers whether the ‘either way’ analysis is qualified by some sort of reasonableness standard in the context of applying law to the primary facts.

[30] Arguments over the ‘true’ meaning of Edwards v Bairstow and Australian Gas are ultimately bound to be inconclusive, because whatever their framers’ intent, these have in fact become cases for all seasons. The current House of Lords regards Edwards v Bairstow as having been a generous extension of the old tests,[64] and the Australian High Court has now referred with approval to both cases. To treat Edwards v Bairstow as an extension, one’s starting point must be that the application of the law to the primary facts is usually a factual exercise.

[31] Some people start from the opposite premise, that all applications of legal standards to the facts as found must logically raise questions of law. Hayes v Federal Commissioner of Taxation (‘Hayes’) [65] is both a prominent Australian example and an interesting counterpoint to Edwards v Bairstow. As in Edwards v Bairstow, the question was whether a tax tribunal had made an error of law because the High Court’s appellate jurisdiction was limited to such circumstances. Also, as in Edwards v Bairstow, the Hayes facts were fully found, and the only question was whether a certain receipt of shares was income. Justice Fullagar said that this had to be a legal question, not because (as in Edwards v Bairstow) the tribunal’s conclusion was beyond any reasonable line drawn in the sand, but simply because it involved applying law to facts:

The views from time to time expressed in this House have been far from unanimous, but in my humble judgment where all the material facts are fully found, and the only question is whether the facts are such as to bring the case within the provisions properly construed of some statutory enactment, the question is one of law only.[67]

[32] Academic proponents of this viewpoint have attempted to give it an intellectual edge by calling it the ‘analytic’ definition of error of law, as opposed to the so-called ‘pragmatic’ definition.[69] It is submitted that the so-called analytic approach has no greater logical appeal than its pragmatic contenders. It is a matter of choice, not logic, whether one characterises the application of law to facts as raising questions of law, fact, or both.

[33] The Migration Act 1958 (Cth) was amended in the early 1990s to restrict the scope of Federal Court judicial review of migration decisions. The amendments were poorly drafted,[70] and had limited success. Section 476(1)(e) stated the error of law ground of review thus:

[34] That was probably intended to be narrower than the common law’s definition of error of law, because it encompassed only two ways of committing the error. The Federal Court, however, held that it was wider than the common law in one respect. It allowed no margin (whether expressed in terms of reasonableness or otherwise) for the Federal Court to agree to differ with any particular application of a legal standard to the primary facts. It did not matter whether that standard was expressed in words carrying ordinary or technical meanings. That is, the section was read as accepting the analytic theory without any modification.[71]

[35] Aside from statutory definitions, those who insist upon their analytical starting point usually come fairly quickly to some common ground with their contenders, simply because their logic too often includes too much within the concept of error of law. If every application of law to facts were necessarily a legal question, then juries would be deciding law as well as fact each time they found someone guilty.[72] In the law of judicial review, many more decisions would become subject to review for error of law if all applications of law to facts raised questions of law. In judicial review of decisions of the Refugee Review Tribunal, for example, whether the facts as fully found gave rise to a well-founded fear of persecution would inevitably raise a legal question. Chief Justice Barwick specifically rejected the analytic approach in a workers’ compensation case, where the question was whether the deceased’s daughter had been wholly dependent upon him at the time of his work-related death.[73] His Honour said that such an approach would make appellable every conclusion as to dependency, and said that ‘there is much to be said against the’ analytic view. The relative autonomy of the tribunal of fact would in each of these cases be seriously compromised.

[36] The academic proponents of the analytical approach stick to their principles, but in doing so, end up either criticising all lawyers for getting it wrong, or calling for a more articulate legislative restatement of the types of legal errors which should be reviewable.[74] It is as if there need be no empirical relationship between their theories and the ways the law-fact distinction plays out in practice.

[37] Most of the judicial proponents of the analytical approach are, ironically enough, more pragmatic. Whilst Justice Fullagar’s judgment in Hayes, thought it likely that all applications of law to the facts as found would raise a legal question, most judges of the analytic orientation have insisted upon ill-defined exceptions. Justice Mason’s judgment in Hope v Bathurst City Council (‘Hope’)[75] is a good example. His Honour acknowledged the force of Justice Fullagar’s approach, but added:[76]

[38] Justice Mason’s pragmatic blurring of the hard edges of the analytical approach treated Brutus v Cozens as a ‘special case’ because the statutory language at issue had used ordinary words intended to be understood without any special legal gloss. Brutus v Cozens is not an easy case to apply, however, and it met with fierce criticism in England, where the academic[78] and judicial[79] commentaries suggest that it is more honoured in the breach than the observance. Certainly, if the application to facts of ‘ordinary words’ did constitute a special case, then one might well expect an increasing incidence of special cases. Parliamentary counsel have sought for some time to draft their statutes in plain English, particularly where their direct readership goes beyond legal practitioners.[80] The criticisms are varied.

[39] It will be recalled that the Court in Australian Gas treated a legal term’s meaning and its application as raising only questions of fact if an ordinary English meaning applied. Justice Phillips tentatively disagreed.[81] His Honour thought that the correct construction of a legal term may always be a question of law. Where that involves simply adopting ordinary English usage, its application to the primary facts will so often raise no legal questions as to invite an elision of the processes of construction (law) and application (fact). Justice Davies held in another case that misunderstanding an ordinary meaning had the effect of departing from the ordinary meaning, which was an error of law.[82] If it is error of law to stray beyond the boundaries of an ordinary meaning, then fixing the ordinary meaning must itself raise a question of law.

[40] Although this reasoning does contradict a large body of case law, it is submitted that the conclusion is inevitable. Of course, there is no error of law if the decision-maker has not strayed beyond the ordinary in determining meaning. One can thereby achieve the desired outcome of Australian Gas’ first two propositions without attempting a distinction between a statutory term’s legal and factual meaning. The distinction between ordinary and special meanings is the result of determining a legal term’s proper meaning or meanings. It is not a test for deciding when the court need not determine that meaning or those meanings.

[41] Justice Santow came to this conclusion in Anderson Stuart v Treleaven.[83] His Honour accepted that one must sometimes make factual inquiries as to a word’s usage before fixing upon its legal meaning. The occasional need for a prior factual inquiry, however, does not alter the overall principle that statutory construction is always a question of law.

[42] Glanville Williams thought that juries could often do with assistance, even with ordinary words.[84] Étienne Mureinik argued that ‘No word is born technical; every term of legal art becomes such by acquiring a technical legal meaning’.[85] His argument was that it takes an Act or a judge to give it that technical meaning. In the latter case, the meaning of even ordinary words is set by reference to statutory and common law context, and the structure and purpose of the Act. In other words, there are no ordinary meanings in the law, in the sense of meanings unaffected by their authorised interpreters. This accords with those jurisprudential theorists who treat interpretation as an exercise inevitably affected by the interpreter’s perspective. Interpretation always makes the words legal words, and meaning and interpretation are one and the same thing.[86]

[43] If allowing a margin for the application of ‘ordinary meanings’ were indeed special, one might have difficulty explaining why the same margin appears to be conceded where the legal term has a special meaning. Tax appeals have established that ‘income’ sometimes[87] has a special, rather than an ordinary meaning for income tax purposes.[88] Even there, however, the same margin seems to be allowed for applications of the legal criteria as if the term bore an ordinary meaning.[89]

[44] It is difficult to account for the precedent value of a judicial decision on a question of fact, if the application of ordinary meanings must always be a question of fact. The issue in Re Minister for Immigration and Multicultural Affairs; Ex parte Cohen[90] was whether a visa applicant was a ‘special need relative’ of his son, who was an Australian citizen or permanent resident. That turned on whether the son’s age was enough in itself to put him into the categories of ‘disability’ or ‘other serious circumstance’. Justice McHugh thought that those categories were defined by words carrying ordinary meanings, and that their application was therefore a question of fact. It remained a question of fact, even though the administrative decision-maker had applied a Federal Court precedent establishing that a child’s age could never be sufficient to place him or her into either category. Right or wrong, the precedent was simply a view of the facts. The Full Court of the Federal Court, however, had assumed that it was deciding a question of law.[91] Its interpretation of the words in question turned much more on their context alongside other visa categories than upon the meaning to be attributed to words considered in isolation. Chief Justice Barwick had earlier made the same point as McHugh J, saying that the question of dependency in workers’ compensation claims was a question of fact, which ‘cannot be turned into a question of law by the citation of authorities’.[92]

[45] The ‘ordinary words’ exception to the strict analytic approach was clearly intended to reduce the scope of what would otherwise be an overly inclusive definition of error of law. Its similarity to the ‘reasonable application’ approach of the Court in Edwards v Bairstow is striking, although its motivation is entirely different. In Edwards v Bairstow, the Court wanted to extend the scope of error of law to wholly unreasonable applications of law to the facts, which makes sense largely on the background assumption that such applications are usually questions of fact. The ‘ordinary words’ approach takes the opposite (ie, analytical) assumption. The cases largely refuse to endorse either assumption as a proposition of universal validity. Two recent examples suffice.

[46] The first example is found in the case of Collector of Customs v Pozzolanic Enterprises Pty Ltd (‘Pozzolanic’),[93] in which the Federal Court attempted a synthesis of some of the case law on the law-fact distinction. That synthesis acknowledged its debt to Australian Gas. It propounded five enumerated general propositions, and added one rider. The High Court quoted the five propositions and summarised the rider in Collector of Customs v Agfa-Gevaert Ltd (‘Agfa’).[94] The Court then suggested reducing the number of propositions by one, commenting adversely on a distinction the Federal Court took to underlie its second and fourth enumerated propositions. However, the High Court prefaced its discussion by noting the elusiveness of an all-encompassing definition, and felt able to avoid coming to any settled conclusions of its own. It is necessary to reproduce the High Court’s quotation from Pozzolanic:[95]

[47] The High Court made no direct comment on propositions one, three and five, nor on the rider to the fifth proposition, inspired by Brutus v Cozens. However, it did take issue with the assumption underlying the separateness of propositions two and four, namely, that there is a relevant or workable distinction between ‘meaning’ on the one hand, and ‘effect or construction’ on the other. The Court thought that if the second and fourth propositions are to coexist, then the Federal Court’s understanding of ‘construction’ required adjustment. It also indicated in fairly clear terms that it thought the distinction between meaning and construction to be ‘illusory’.[102] In other words, it appears likely that the High Court intended to discard part or all of Pozzolanic’s fourth proposition, but not its second. Whilst acknowledging the step forward in shortening the list of indicative criteria for resolving the law-fact distinction, one might have wished that the High Court had discarded both the second and fourth propositions.

[48] The main burden of High Court’s criticism of Pozzolanic in Agfa, however, might be characterised as something of a red herring. The Court in Agfa went to some lengths to demonstrate that the meaning of a string of words might be different than the sum of the meanings of each separate word in the string. It is true, of course, that a single word will take some colour from those around it, even if the single word is very ordinary. It is difficult to believe that Pozzolanic intended to deny that. An alternative reading of Pozzolanic’s fourth proposition would be that ‘effect or construction’ referred not to the arithmetical addition of the meaning of each single word, but rather to the application to the facts of an unambiguous[103] law, even a law composed wholly of ordinary words intended to be given their ordinary meaning.

[49] That still leaves the puzzle of why the High Court refrained from commenting on Pozzolanic’s general adherence (by its fifth proposition) to the analytic approach.[104] One might conclude from Agfa that ordinary meanings are facts, but their application to the facts as found is a legal question unless that application is reasonably open.

[50] The second case is the more recent High Court decision in Vetter v Lake Macquarie City Council (‘Vetter’),[105] which appears to accommodate the analytic and pragmatic stances. The issue in this case was whether a worker was covered by workers’ compensation when she ran her car off the road. Because she was not working at the time, that depended on whether her driving was in the course of a ‘journey’ home from work.[106] She was driving home, but had first made her regular, fortnightly detour to her grandmother’s home for an evening meal. The trial judge had thought she was on a ‘journey’ from work to home, despite her detour being substantial in terms of direction, time and distance. A majority of the Court of Appeal had thought the detour so substantial as to make two journeys out of her trip, which meant that her crash did not occur whilst on a journey from work to home.[107] The High Court acknowledged the indeterminacy of the word ‘journey’,[108] but thought that the Court of Appeal’s test of ‘one journey or two’ was no substitute for the statutory term.[109] The High Court thought that the trial judge’s application of the term ‘journey’ to the facts as found was within the bounds of reasonableness. Justice Handley had applied Lord Radcliffe’s test in Edwards v Bairstow of whether the trial judge’s conclusion contradicted ‘the true and only reasonable conclusion’.[110] Justice Kirby called this ‘a somewhat sterile criterion’,[111] as it begged the question. His Honour nevertheless cited Lord Radcliffe’s speech in Edwards v Bairstow in support of the following proposition:

[51] This is classic Wednesbury language, although its difference from Lord Radcliffe’s ‘true and only reasonable conclusion’ is elusive.

[52] The joint judgment of Gleeson CJ, Gummow, Hayne and Callinan JJ cited Edwards v Bairstow with apparent approval.[113] At the same time, however, their Honours quoted with apparently equal approval from Justice Mason’s judgment in Hope,[114] in which, it will be recalled, the notion of a reasonable margin was a special exception to the analytic approach, applying only where words with ordinary meanings were being applied to the facts. The issue in Vetter did not require the court to indicate whether it was endorsing Edwards v Bairstow only because the relevant legal test was whether the worker had been on a ‘journey’, and because that was an ordinary word to be given its ordinary meaning. Referring to Hope rather than Edwards v Bairstow, Justice Hayne’s use of the ‘either way’ analysis was based squarely on his characterisation of ‘journey’ as an ordinary word carrying its ordinary meaning.[115] The joint judgment, however, seems more likely to have explained its use of the ‘either way’ analysis not on the basis that ‘journey’ is an ordinary word, but because, as the joint judgment said: ‘The term “journey”, used in the relevant sense, has an indeterminate meaning or meanings’.[116]

[53] Their Honours had earlier quoted with approval from a judgment of Mason JA in Williams v Bill Williams Pty Ltd.[117] The Court of Appeal had held in this case that workers’ compensation covered an employee’s injuries received when shot by a jealous husband who had pursued him out of his work-place onto the footpath. Justice Mason said that a contrary conclusion was an error of law, either because the facts had been fully found and it only remained to apply the law,[118] or because on those facts, it was not ‘reasonably possible’ to reach a different conclusion. It would have been an error of fact if the question had been ‘largely one of degree upon which different minds may take different views’.[119]

[54] Justice Mason had said in Hope[120] that ‘special considerations’ applied where a legal term carried its ordinary English meaning. There has long been a parallel line of cases, however, allowing the same margin to reasonable applications of law to the primary facts where the legal term has a range of acceptable meanings.[121] That was the position in Vetter, where ‘journey’ was a term of indeterminate meaning. The same approach has long been evident where the legal term itself requires a degree of evaluation or judgment,[122] as in the case of applying the ‘reasonable care’ standard to the primary facts in a negligence case. Applications of the law to the primary facts in these cases are often characterised as questions of degree, or of mixed law and fact. Questions of degree and mixed questions are often classified as questions of fact.[123] It suffices in some statutory contexts to say that a question is mixed if the facts are not fully found.[124] The concept of a mixed question is usually less well-defined, and has been condemned on that score as far back as 1842.[125] Nevertheless, it is frequently used.

[55] It appears that answers to mixed questions, and questions of degree, turn into errors of law under exactly the same conditions as in Australian Gas, namely, where the answers are necessarily wrong.[126] That imports a margin,[127] and the only question is whether to expand on the Australian Gas margin by introducing a reasonableness standard.[128] Chief Justice Spigelman acknowledged the introduction of tests such as that ‘no other conclusion was reasonably open’, but insisted that this was no different in essence from the Australian Gas question of whether the impugned application of the legal standard to the primary facts was necessarily wrong.[129] After criticising ‘reasonable’ as a ‘weasel word’,[130] his Honour concluded that the test ‘has a stringency equivalent to that of Wednesbury unreasonableness’.[131] Justice Phillips characterised ‘reasonable’ as a distracting intrusion, which effects no alteration to tests such as those in Australian Gas if properly construed:

[56] It is submitted that it would be a mistake to try to confine the ‘either way’ formulation of Edwards v Bairstow to cases where the law to be applied carries its ordinary English meaning. Although their case-law canon is not identical, but merely overlaps, the ‘ordinary words’ cases are in reality no different to the cases dealing with questions of degree and mixed questions of law and fact. If that proposition is accepted, then the ‘either way’ margin is not as ‘special’ as Mason J had supposed in Hope.[133] Indeed, it would become the norm, with the consequence that one’s starting point would be that applications of law to the primary facts are usually questions of fact. Justice Phillips reached that conclusion in S v Crimes Compensation Tribunal, although his formulation was that the application of law to facts ‘is essentially a question of fact’.[134] The Full Court of the Federal Court had reached the same conclusion in an earlier tax case.[135]

[57] The exceptions to that proposition would be twofold. The first exception would apply where the application has produced a result beyond any margin for reasonable minds to agree to differ. Justice Phillips opposed the ‘intrusion’ of ‘reasonable’ into this test, but saw a good many cases (including Edwards v Bairstow and Azzopardi) as proposing essentially the same margin for essentially the same reasons.[136]

[58] The second qualification to applications of law being factual exercises is not so much an exception as a separate proposition. To return to Pozzolanic, whether a term’s meaning is ordinary or special is itself a question of law, and the meaning of a technical legal term is also a question of law. Each of those propositions is well-established. However, neither of them turns on the application of the legal term (technical or otherwise) to the primary facts. Misunderstanding the governing law has always been an error of law in its own right, and that should include misunderstanding the legal meaning of a statutory term, ordinary or special. Misunderstanding is the error, and that can occur in relation to ordinary as well as technical terms. In other words, the proper meaning of any legal term should itself be a question of law.[137]

[59] Indeed, this article has discussed a word’s ordinary meaning in the singular, whereas ordinary and special words alike can be more than indeterminate. They can have quite different meanings, in contexts which require the court to choose the most appropriate meaning. The necessity to choose between different ordinary meanings presents a question of law, if one of those meanings would undermine the statutory objectives.[138] The issue in Hope[139] was whether small-scale use of land for an agistment constituted the ‘carrying on of a grazing business’. ‘Business’ was used in its ordinary meaning, but the Shorter Oxford Dictionary gave nineteen meanings. Justice Mason thought that the choice of meanings presented a question of fact, but that the trial judge had nevertheless erred in law in two respects. He had impermissibly required the business activity to be ‘significant’. He had also impermissibly treated the word in isolation from its attachment to the notion of something being ‘carried on’, thereby overlooking the need to adopt a meaning which accommodated the need for a repetitive and ongoing activity. That was an error of construction which was an error of law. Even working with the distinction between ordinary and special meanings, it seemed fairly easy to make the move from fact to law. Overlooking the restrictive effect of context upon an ordinary word is a constructional, and therefore legal error.

[60] Of course, if the impugned decision-maker properly understood the term, and if its construction were not complicated, there would be less likelihood of it being so misapplied to the primary facts as to exceed the permissible margin, however expressed.

[61] What emerges from the analysis so far? The High Court has implied that the search for a universal set of principles for distinguishing legal from factual questions is probably doomed.[140] However, it is clear that it is extremely difficult to discern an error of law in a finding of primary facts, unless, of course, the finding was guided by an incorrect legal test or used impermissible materials. Those possibilities aside, primary fact-finding can be criticised for error of law only if there was literally no evidence, or if there was so small a factual basis as to render the finding ‘simply not open’. The margin in the latter case is usually expressed without adjectives such as ‘unreasonable’ or ‘perverse’, although those qualifications have their proponents.

[62] The second stage of decision-making is determining the applicable law. Many cases say that any mistakes here are errors of law, unless the legal term in question carries an ordinary, non-technical meaning. This article argues that the distinction between ordinary and special meanings is unworkable, and announces a result rather than a test. The purpose of the distinction seems to be to allow more latitude for some terms, but that is a matter of statutory interpretation, not an inherent, qualitative distinction between words.

[63] Moving to the third stage of decision-making, the application of the law to the facts can produce a question of law, but the cases conflict as to when and why. They even conflict as to whether it is the norm or the exception, and the High Court has endorsed both viewpoints. Whether the norm or the exception, each approach has a counterpoint, whose bottom line is to treat applications of the law to the facts as involving an error of law where a certain margin has been exceeded. The coverage and extent of that margin might depend on the viability of the ‘ordinary words’ category, although this article argues that this would be a mistake. The High Court has recently used the word ‘reasonable’ as part of its account of the margin, although it has its opponents on two grounds, namely, that the use of the word ‘reasonable’ either changes nothing or should change nothing in the reasoning to be applied by the courts. The proponents of some sort of reasonableness test, however strict, seem to have left behind the old strategy of arguing from an implausible outcome to an inference that the law must have been misunderstood. That shift was probably inevitable as decisions are increasingly accompanied or followed by detailed reasons.[141]

[64] It is therefore not surprising to find statements of despair or even cynicism littered through the law reports. The law-fact distinction has been called ‘slippery’,[142] ‘elusive’,[143] ‘too easily manipulated’,[144] ‘sterile and technical’,[145] and something which can generate ‘artificial, if not illusory’ distinctions.[146] Some have wondered whether it might not be meaningless.[147] In the absence of an all-encompassing set of rules, the inevitable result is the idea that the distinction is made differently in different contexts.[148] Justice Windeyer seemed to take that approach in Da Costa v R,[149] and Spigelman CJ took it unequivocally in Attorney General v X.[150] Justice Windeyer acknowledged the opportunities for judicial manipulation of the distinction according to whether the judge wanted to intervene. His Honour said that the distinction’s manipulability ‘may engender cynicism or stimulate self-examination or merely show that the distinction is not capable of precise formulation’.[151] His Honour felt able to avoid coming to any firm conclusion because the statute he had to construe talked of ‘a question of law alone’. Chief Justice Spigelman felt able to read down ‘questions of law’ to ‘questions of law alone’ because of the context in which it appeared.[152]

[65] It is submitted that cynicism is simply not an available option, if only because so many statutes stipulate appeal rights or limitations according to whether fact or law is involved. It is not open to take the cynical view that the distinction is meaningless or infinitely manipulable.[153] It is almost as unconvincing to be wholly cynical about the law-fact distinction in light of its pervasiveness in common law doctrine. It is critical, for example, in the law of judicial review, which generally leaves administrative fact-finding reasonably well alone.[154] The common law allows judicial review for error of law only if the error is jurisdictional, or (failing that), if it is apparent on the face of the record. In deciding where to draw the line between jurisdictional and non-jurisdictional errors of law, the Australian trend has been to allow a generous margin to inferior court judges, but not to administrators. There is room for debate, however, as to the strength of that trend.[155] English common law has converted almost all errors of law into jurisdictional errors,[156] and Kirby J has indicated that he would like to follow suit.[157] Statutory judicial review schemes have overtaken the common law in this respect, either by allowing review for any error of law,[158] or by stipulating a more generous definition of the record.[159]

[66] It is submitted that the original intent of the distinction between jurisdictional and non-jurisdictional errors of law was to accord decision-makers a small measure of independence. The erosion of the distinction at common law and under statutory judicial review schemes has prompted English commentators to call for a more flexible, context-dependent approach to defining error of law,[160] for fear that this ground of review might otherwise prove to be a monster.[161]

[67] Timothy Endicott[162] concluded that the law-fact distinction has in fact always been context-dependent. He argued that it cannot have been otherwise. This argument must be correct, because the distinction turns up in such diverse contexts, and is clearly meant to serve different purposes. Examples include the rules about pleading material facts, not law; the division of functions between judge and jury; some aspects of the law of misrepresentation in business dealings; criminal law’s distinctions between mistakes of fact and law; the division of functions between judges and witnesses or counsel; and the division of functions between administrator and judicial review court. By contrast, the High Court rejected the law-fact distinction as an irrelevancy in restitutionary claims for moneys paid under a mistake, and was therefore free to add that the distinction was ‘difficult’, ‘illogical’ and ‘artificial’.[163] It can only have been the context which earned the distinction those criticisms.

[68] Analysis divorced from context is indeed half-baked, and Endicott lays claim to having the truly ‘analytical’ approach. His main conclusion is an important one which this article gratefully adopts. It is that good analysis asks why the question is being asked. It does not ask what result might be convenient. There is room for pragmatism in answering the first question, but an answer based wholly on convenience is pure cynicism. When the law-fact distinction determines eligibility to judicial review or appeal, one should pay regard to the relationships between the decision-maker and the court, and to the functions sought to be served by appellate or review mechanisms.

[69] Chief Justice Spigelman said in Attorney General (NSW) v X[164] that the considerations of context include ‘the scope, nature and subject matter of the statute, including the nature of the body making the relevant decision’.[165] The line might well be drawn differently, for example, depending on whether the decision-maker was a judge or an administrator.

[70] Lord Denning once propounded a functional test of sorts for the law-fact distinction, which was to ask whether the question needed a lawyer’s skills to give it a proper answer.[166] That begged the question, and was never accepted. In Edwards v Bairstow Lord Radcliffe refused to draw the line according to the relative tax skills of the tax tribunal and generalist court, although the court would in any event have fared better in comparison with the woodenly mechanical tax tribunal in that case.[167]

[71] One of the contextual factors is the stage in the decision-making process in which the error of law allegedly arises. Justice Glass had distinguished three stages in Azzopardi,[168] namely, fact-finding, rule-stating and rule application. Error of law is defined extremely tightly as regards the fact-finding stage, and with some degree of flexibility as regards the application of law to the facts. There is no flexibility, however, as regards the decision-maker’s misunderstanding of the applicable legal tests. This reflects the relative need for superior court intervention. Error of law looks not only to the past, but to the future. Its central concern is precedent.

[72] Whether the primary facts be found by an administrator, tribunal or inferior court, the finding has little if any lasting precedent effect beyond its effects on the parties themselves. The same can often be said as regards the third stage, where the correct legal test is applied to the primary facts; this usually affects only the parties, although there will inevitably be some contexts where the decision will become a precedent or represent a policy decision. That might explain the relative rarity of error of law in assessing, say, degrees of contributory negligence or impairment, as compared to some tax, welfare and migration decisions. Principles from the latter areas are more likely to be discerned from a range of specific applications over time. It also explains the tension occasionally apparent in appeal cases between elevating earlier applications of the law to similar facts as precedents, and other more functional concerns than the need for consistency. One of the reasons that Barwick CJ did not want applications of the test for ‘dependency’ to become precedents was the need in workers’ compensation for speed and finality of decision-making.[169] Justice Mason recognised in the same case that the trade-off was to allow a greater level of inconsistent outcomes in workers’ compensation, where ‘the decisions are not notorious for their uniformity’.[170]

[73] The County Court judges in Pearlman v Harrow School[171] produced two lines of authority on whether a tenant’s installation of a common central heating unit qualified for a rent reduction as a ‘structural addition’. Lord Denning thought that he had to classify this as a question of law, because there would otherwise have been no way of resolving an inconsistency he regarded as ‘intolerable’.[172] There were concerns about precedent in Edwards v Bairstow, one more obvious than the other. At the surface level, their Lordships felt obliged to settle a perceived inconsistency between the Scottish and English courts about when they could overturn a tax tribunal’s application of the law to a commonly recurring fact situation. At a deeper level, they felt obliged to downgrade the precedent status of an earlier decision,[173] which they thought had been applied too mechanistically.

[74] It is submitted that the precedent concerns in Pearlman v Harrow School might have been as adequately addressed by focusing not on the third stage of decision-making (the application of the law to the facts), but on the second stage. A declaration that the installation of that type of central heating installation could constitute a structural alteration was all that was needed. The courts had no choice but to progress from stage two to three in Edwards v Bairstow, because the tax tribunal had ignored the earlier direction from Upjohn J that an isolated transaction could constitute an adventure in the nature of trade. That is, a declaration as to a legal term’s meaning usually carries more precedent force than an individual instance of its application to specific facts. Furthermore, it will in fact carry that precedent force regardless of whether the declared meaning is of a word bearing an ordinary meaning.[174]

[75] Even in cases where only the third stage of the decision-making process is involved, there are clear indications in the English cases that inconsistent administrative treatment of like cases might be a ground for judicial review.[175] A handful of Australian cases have treated inconsistency as either an unreasonableness issue or a natural justice issue.[176] Furthermore, there is a well-established line of authority invalidating a preconceived administrative policy on how to treat commonly recurring facts if that policy serves to confine the proper legal test.[177] These interventions are not on the ground of error of law, but they do proceed from a concern as to the proper relationship between administrative precedent and the meaning of the governing law.

[76] The Americans accord considerable deference to the administrative agencies’ interpretations of their own statutes and rules, provided that those interpretations carry the force of agency authority and precedent.[178] The High Court emphatically disagrees with that approach,[179] but the ‘either way’ latitude it allows to applications of the law to the primary facts as found serves to confine the differences between the two countries to the interpretative or second stage of the decision-making process.

[77] Conclusions are necessarily few and generalised. The law-fact distinction is not meaningless, but it can be difficult to draw. This is partly because some aspects of decision-making might be ‘essentially’ factual, but they can at the same time yield errors of law. There are several contexts in which there are no bright lines between law and fact. The distinction is also difficult because its resolution varies according to context, and the contextual factors include a workable relationship between court and decision-maker. This article has confined its discussion of the law-fact distinction to two related contexts, namely, appeals limited to questions or errors of law, and judicial review for error of law. Whilst the cases are by no means consistent, it is argued that the most workable distinctions are between three stages of the decision-making process, namely, fact-finding, stating the applicable legal rule, and applying the rule to the facts.

[78] Inferential error in the process of finding the facts is the least likely to produce an error of law, perhaps because it is the least likely to have a precedent effect. Misinterpreting the applicable legal rule will always be regarded as an error of law, although it may not always be material. Applying the law to the facts should usually be regarded as a factual exercise, producing errors of law only where the result reached is beyond any tolerable margin. Whether that margin is usefully defined by its reasonableness is still disputed by those who fear that ‘unreasonable’ is too flexible (and therefore too manipulable) an adjective for this area. However, whatever the adjective, there is general agreement that the application of law to fully found facts produces an error of law only where the outcome is clearly not open. It is suggested that there is no harm in expressing the requisite clarity in terms of unreasonableness, provided one understands that the level of unreasonableness is equivalent to that required by the Wednesbury test. Applications of the law to the facts as found are usually treated as factual exercises because of the need to allow the decision-maker a margin for his or her own judgment. The margin’s existence and rationales are reasonably clear, despite its formulation having yet to be authoritatively pronounced.

[79] Several leading judgments have listed a number of general propositions, whilst taking care to disown any attempt at stating a universally applicable formula. The High Court has in more recent times attempted a modest reduction in the number of those general propositions. It has suggested rejection of a distinction previously drawn between meaning and construction or effect, but it has repeated a number of distinctions turning on whether Parliament intended its terminology to bear an ordinary or special meaning. This article argues that the court could go further. The greater the number of categories or prima facie rules, the greater the difficulties in distinguishing between them. That leads to greater manipulability, which in turn leads to a sense of cynicism in the whole exercise of seeking some meaningful distinctions between questions of law and fact.

[80] This article argues that the margin allowed to decision-makers, involved in applying the law to the facts of a case as found, is the same, regardless of whether the legal terms being applied bear an ordinary or a special meaning. Australian case law contains two oft-repeated distinctions which should be discarded. The first is between words with ordinary or special meanings, and the second is between applications of ordinary or special meanings to the facts as found. All legal words have legal meanings, although some might be easier to grasp, wider, or less determinate than others. Their application to the facts calls for correction whenever they are clearly wrong.

[*] Faculty of Law, University of New South Wales. The author gratefully acknowledges the comments and assistance of Sir Gerard Brennan, Arthur Glass, Justice Handley, John McMillan, Robert Shelly and Neil Williams.
[1] See, eg, Leon Green, Judge and Jury (1930) 270-1: ‘No two terms of legal science have rendered better service than “law” and “fact”. They are basic assumptions; irreducible minimums and the most comprehensive maximums at the same instant. They readily accommodate themselves to any meaning we desire to give them. In them and their kind a science of law finds its strength and durability. They are the creations of centuries. What judge has not found refuge in them? The man who could succeed in defining them would be a public enemy.’
[2] Migration Act 1958 (Cth) s 476(1)(e), repealed by the Migration Legislation Amendment (Judicial Review) Act 2001 (Cth).
[3] Collector of Customs v Agfa-Gevaert Ltd [1996] HCA 36; (1996) 186 CLR 389, 394.
[4] See Da Costa v R [1968] HCA 51; (1968) 118 CLR 186, 194 (Windeyer J): ‘When the distinction [between questions of fact and questions of law] determines whether or not in a particular case an appeal lies, there is room for questioning whether it has in philosophy or logic an essential and abstract and universal character.’
[5] Louis Jaffe, Judicial Control of Administrative Action (1965) 548. Justice Finkelstein quoted Jaffe with approval in SRL v Minister for Immigration and Multicultural Affairs [2001] FCA 765, but only in the context of dissenting from Justice Katz’s view that a predicted fact is not a ‘particular fact’ for the purposes of s 476(4)(b) of the Migration Act 1958 (Cth). That paragraph was the same as s 5(3)(b) of the Administrative Decisions (Judicial Review) Act 1977 (Cth). It provided for judicial review where the impugned decision was based upon a particular fact which did not exist, and was recently explained in Jegatheeswaran v Minister for Immigration and Multicultural Affairs [2001] FCA 865 (‘Jegatheeswaran’). Justice Finkelstein explored the different characterisations of ‘fact’ (namely, positive and negative; past, present and future; primary and secondary; perceived and inferred; evidential and ultimate, operative, dispositive, material or constitutive; proven and intuited) in Jegatheeswaran [2001] FCA 865 [52]-[59].
[6] Jegatheeswaran [2001] FCA 865 [58]-[59].
[7] See R D Friedman, ‘Standards of Persuasion and the Distinction Between Fact and Law’ (1992) 86 Northwestern Law Review 916.
[8] See A A S Zuckerman, ‘Law, Fact or Justice?’ (1986) 66 Boston University Law Review 487.
[9] See Burnie Port Authority v General Jones Pty Ltd [1994] HCA 13; (1994) 179 CLR 520, 540.
[10] Cf Donald Nicholson, ‘Truth, Reason and Justice: Epistemology and Politics in Evidence Discourse’ (1994) 57 Modern Law Review 726.
[11] Cf Kenneth Vinson, ‘Artificial World of Law and Fact’ (1987) 11 Legal Studies Forum 311; and Kenneth Vinson, ‘Disentangling Law and Fact: Echoes of Proximate Cause in the Workers’ Compensation Coverage Formula’ (1996) 47 Alabama Law Review 723. Both articles argue that the distinction between errors of law and fact is made entirely to produce the desired outcome, and that it should therefore be discarded. See also Bibi Sangha, ‘The Law/Fact Distinction in Contract: a Lawyer’s Plaything?’ (1994) 7 Journal of Contract Law 113.
[12] March v Stramare (E & MH) Pty Ltd [1991] HCA 12; (1991) 171 CLR 506; and New South Wales v Taylor [2001] HCA 15; (2001) 178 ALR 32, 36.
[13] See AMP General Insurance Ltd v Roads and Traffic Authority of NSW [2001] NSWCA 186 [27] (Spigelman CJ): ‘As emphasised by the High Court in March v Stramare [(E & MH) Pty Ltd [1991] HCA 12; (1991) 171 CLR 506 at 516 and 524], causation is not only a factual question, it is also a normative one.’
[14] Smith New Court Securities Ltd v Scrimgeour Vickers (Asset Management) Ltd [1996] UKHL 3; [1997] AC 254, 285.
[15] See Chen v Minister for Immigration and Multicultural Affairs [2000] HCA 19; (2000) 201 CLR 293, 314-15 (Kirby J).
[16] See, eg, Woolfe v Tasmania [2001] TASSC 66; and Minister for Immigration and Multicultural Affairs v Sarrazola [No 2] [2001] FCA 263; (2001) 107 FCR 184, 199. See also Nicola Padfield, ‘Clean Water and Muddy Causation: Is Causation a Question of Law or Fact, or Just a Way of Allocating Blame?’ [1995] Criminal Law Review 683, where it is argued that causation is a question of fact in English homicide cases, but a question of law in English pollution prosecutions. (It is also a question of fact in Australian murder cases: Royall v R [1991] HCA 27; (1991) 172 CLR 378, 387-8.)
[17] Brodie v Singleton Shire Council [2001] HCA 29; (2001) 180 ALR 145.
[18] Australian Safeway Stores Pty Ltd v Zaluzna (1987) 162 CLR 479.
[19] Burnie Port Authority v General Jones Pty Ltd [1994] HCA 13; (1994) 179 CLR 520, holding that Rylands v Fletcher (1868) LR 3 HL 330 should no longer be seen as generating a liability regime separate from negligence.
[20] Rogers v Whitaker [1992] HCA 58; (1992) 175 CLR 479.
[21] Associated Provincial Picture Houses Ltd v Wednesbury Corporation [1947] EWCA Civ 1; [1948] 1 KB 223 (‘Wednesbury’).
[22] See, eg, Minister for Immigration and Multicultural Affairs v Eshetu [1999] HCA 21; (1999) 197 CLR 611, 626, in particular (Gleeson CJ and McHugh J):

Someone who disagrees strongly with someone else’s process of reasoning on an issue of fact may express such disagreement by describing the reasoning as ‘illogical’ or ‘unreasonable’, or even ‘so unreasonable that no reasonable person could adopt it’. If these are merely emphatic ways of saying that the reasoning is wrong, then they may have no particular legal consequence.

[23] Minister for Immigration and Ethnic Affairs v Wu (1996) 185 CLR 259 requires a benign reading of an administrator’s ambiguous statement of reasons. It does not require a brief reading of the reasons.
[24] Wednesbury [1947] EWCA Civ 1; [1948] 1 KB 223, 229-30 (Lord Greene MR).
[25] Mark Aronson and Bruce Dyer, Judicial Review of Administrative Action (2^nd ed, 2000) 283.
[26] Council of Civil Service Unions v Minister for the Civil Service [1984] UKPC 17; [1985] AC 1 AC 374, 410 (Lord Diplock).
[27] Wednesbury certainly requires rationality, but it may not require strict logic. The courts may nevertheless be more prepared than in the past to imply a statutory requirement for logic for particular regulatory circumstances. See: Geoffrey Airo-Farulla, ‘Rationality and Judicial Review of Administrative Action’[2000] MelbULawRw 23; , (2000) 24 Melbourne University Law Review 543; Minister for Immigration and Multicultural Affairs v Epeabaka [1999] FCA 1; (1999) 84 FCR 411; Hill v Green [1999] NSWCA 477; (1999) 48 NSWLR 161; Gamaethige v Minister for Immigration and Multicultural Affairs [2001] FCA 565; Cujba v Minister for Immigration and Multicultural Affairs [2001] FCA 699; and Bakhtyar v Minister for Immigration and Multicultural Affairs [2001] FCA 947 [35]. See also Minister for Immigration and Multicultural Affairs v Al-Miahi [2001] FCA 744 [34]; and Minister for Immigration and Multicultural Affairs v Perera [2001] FCA 1212.
[28] Abebe v Commonwealth [1999] HCA 14; (1999) 197 CLR 510; and Minister for Immigration and Ethnic Affairs v Eshetu [1999] HCA 21; (1999) 197 CLR 611.
[29] As Mason J contemplated in Minister for Aboriginal Affairs v Peko-Wallsend Ltd (1986) 162 CLR 24, 41.
[30] The unreasonableness standard seems easier to establish in England than Australia, especially in the field of human rights and fundamental freedoms. See: R v Ministry of Defence; Ex parte Smith [1995] EWCA Civ 22; [1996] QB 517, 554; R v Secretary of State for the Home Department; Ex parte Simms [1999] UKHL 33; [2000] 2 AC 115, 130, 144; R v Secretary of State for the Home Department; Ex parte Daly [2001] UKHL 26; [2001] 3 All ER 433 (‘Daly’); and Gamaethige v Minister for Immigration and Multicultural Affairs [2001] FCA 565 [18]-[28]. Lord Cooke was a critic of Wednesbury’s tolerance of all but the totally capricious or absurd. See Robin Cooke, ‘The Struggle for Simplicity in Administrative Law’, in Michael Taggart (ed), Judicial Review of Administrative Action in the 1980s: Problems and Prospects (1986) 13-16; R v Chief Constable of Sussex; Ex parte International Trader's Ferry Ltd [1998] UKHL 40; [1999] 2 AC 418, 452; and Daly [2001] UKHL 26; [2001] 3 All ER 433 [32]. Lord Steyn suggested in Daly that the unreasonableness standard in human rights cases may by now have become completely distinct from its Wednesbury past: Daly [2001] UKHL 26; [2001] 3 All ER 433 [25]-[26].
There is also some uncertainty as to whether the reasonableness requirement is an implied statutory duty or an underlying common law rule imputed to Parliament. See Abebe v Commonwealth [1999] HCA 14; (1999) 197 CLR 510, 554; Minister for Immigration and Multicultural Affairs v Eshetu [1999] HCA 21; (1999) 197 CLR 611, 650; and R v Refugee Review Tribunal; Ex parte Aala [2000] HCA 57; (2000) 176 ALR 219, 230-1. Unreasonableness might be an error of law: Hymix Industries Pty Ltd v Alberton Investments Pty Ltd [2001] QCA 334.
[31] J M Evans, De Smith's Judicial Review of Administrative Action (4^th ed, 1980) 348; Aronson and Dyer, above n 25, 279-80.
[32] Parramatta City Council v Pestell [1972] HCA 59; (1972) 128 CLR 305; and Minister for Immigration and Multicultural Affairs v Eshetu [1999] HCA 21; (1999) 197 CLR 611.
[33] [1955] UKHL 3; [1956] AC 14.
[34] [1985] 1 AC 374, 410-11.
[35] Gamaethige v Minister for Immigration and Multicultural Affairs [2001] FCA 565 [18].
[36] Namely, illegality, irrationality and procedural impropriety.
[37] See, eg, Yates Property Corporation Pty Ltd (In Liq) v Darling Harbour Authority (1991) 24 NSWLR 156, 186, where a mistake of fact as to the scope of a dispute led to a breach of natural justice, which was in its turn characterised as an error of law.
[38] See Parisienne Basket Shoes Pty Ltd v Whyte [1938] HCA 7; (1938) 59 CLR 369, 391; Australian Heritage Commission v Mount Isa Mines Ltd [1997] HCA 10; (1997) 187 CLR 297; Timbarra Protection Coalition Inc v Ross Mining NL [1999] NSWCA 8; (1999) 46 NSWLR 55; and Enfield Corporation v Development Assessment Commission [2000] HCA 5; (2000) 199 CLR 135. It is acknowledged that the interpretative presumption against jurisdictional facts is sometimes claimed to be stronger in the case of judicial decision-makers than in the case of tribunals or bureaucratic decision-makers.
[39] See Mark Aronson, ‘The Resurgence of Jurisdictional Facts’ (2001) 12 Public Law Review 17; Buck v Comcare [1996] FCA 1485; (1996) 66 FCR 359; Timbarra Protection Coalition Inc v Ross Mining NL [1999] NSWCA 8; (1999) 46 NSWLR 55; Canberra Tradesmen’s Union Club Inc v Minister for the Environment, Land and Planning [2000] FCA 203; (2000) 107 LGERA 164; Enfield Corporation v Development Assessment Commission [2000] HCA 5; (2000) 199 CLR 135; and Cabal v Attorney General (Cth) [2001] FCA 583.
[40] Curragh Queensland Mining Ltd v Daniel [1992] FCA 44; (1992) 34 FCR 212; Aronson and Dyer, above n 25, 218-19.
[41] For criminal appeals, see: Doney v R [1990] HCA 51; (1990) 171 CLR 207; M v R [1994] HCA 63; (1994) 181 CLR 487; Jones v R (1997) 191 CLR 439; and Gipp v R (1998) 194 CLR 106. For civil appeals, see: Warren v Coombes [1979] HCA 9; (1979) 142 CLR 531; State Rail Authority (NSW) v Earthline Constructions Pty Ltd (In Liq) [1999] HCA 3; (1999) 160 ALR 588; and Coal and Allied Operations Pty Ltd v Australian Industrial Relations Commission [2000] HCA 47; (2000) 174 ALR 585.
[42] Doney v R [1990] HCA 51; (1990) 171 CLR 207; and Naxakis v Western General Hospital (1999) 197 CLR 269, 282, 294, 309-11.
[43] Naxakis v Western General Hospital (1999) 197 CLR 269, 310 (Callinan J).
[44] Attorney General (NSW) v X [2000] NSWCA 199; (2000) 49 NSWLR 653, 667. His Honour was referring to Professor Craig’s account of Master of the Rolls, Lord Denning’s reading (in Instrumatic Ltd v Supabrase Ltd [1969] 1 WLR 519) of Edwards v Bairstow [1955] UKHL 3; [1956] AC 14. See P P Craig, Administrative Law (4^th ed, 1999) 268, which accuses Lord Denning of turning a test of ‘reasonably open’ into a test of whether the judge thought the application of the law to the primary facts was simply wrong. See also Timothy Endicott, ‘Questions of Law’ (1998) 114 Law Quarterly Review 292, 308: ‘Lord Denning ... followed an unswerving rule of calling a question a “question of law” when he wanted to.’
[45] S v Crimes Compensation Tribunal [1998] 1 VR 83, 90-1.
[46] [1990] HCA 33; (1990) 170 CLR 321, 355-60.
[47] The English cases included Edwards v Bairstow [1955] UKHL 3; [1956] AC 14, 36; see also Bond [1990] HCA 33; (1990) 170 CLR 321, 356.
[48] (1985) 4 NSWLR 139.
[49] Azzopardi (1985) 4 NSWLR 139, 151.
[50] Ibid 156 (emphasis added).
[51] Note that Kirby J argues that Azzopardi has not finalised the debate as to whether to accept perversity of primary fact-finding as an error of law. See Aronson and Dyer, above n 25, 211-12; IW v Perth [1997] HCA 30; (1997) 191 CLR 1, 71; X v Commonwealth (1999) 200 CLR 177, 218-19; and Vetter v Lake Macquarie City Council (2001) 178 ALR 1, 20. See also Yu Feng Pty Ltd v Maroochy Shire Council (1996) 92 LGERA 41, 69, where Fitzgerald P apparently rejected Azzopardi, in favour of a test that there was error of law if the application of ordinary words in an Act to the primary facts produced a result which the evidence did not ‘reasonably admit’; and Wallaby Grip Ltd v Peirce [2000] NSWCA 299 [40], where Fitzgerald JA indicated that the High Court might have to resolve an argument as to whether Azzopardi was entirely consistent with Australian Gas Light Company v Valuer-General [1940] NSWStRp 9; (1940) 40 SR (NSW) 126. The Federal Court regards Azzopardi as ‘problematic’: Minister for Immigration and Multicultural Affairs v Applicant C [2001] FCA 1332 [72].
[52] Namely, the Commissioners for General Purposes of the Income Tax.
[53] [1955] UKHL 3; [1956] AC 14.
[54] Ibid 31.
[55] See Federal Commissioner of Taxation v Whitfords Beach Pty Ltd [1982] HCA 8; (1982) 150 CLR 355, 376. Justice Mason there noted a certain irony in the Edwards v Bairstow conclusion that the facts admitted of only one legal outcome, when an earlier House of Lords decision (namely, Jones v Leeming [1930] AC 415) seemed to have come to the opposite conclusion on facts which his Honour regarded as being substantially the same. Jones v Leeming prompted the inclusion into the Income Tax Assessment Act 1936 (Cth) of what became (for some considerable time) s 26(a). The joint majority judgment in Commissioner of Taxation v Montgomery [1999] HCA 34; (1999) 198 CLR 639, 673-4 felt able to reconcile Jones v Leeming and Edwards v Bairstow.
[56] Edwards v Bairstow [1955] UKHL 3; [1956] AC 14, 30, 36.
[57] Ibid 31.
[58] Ibid 32.
[59] Ibid 29.
[60] Ibid 36.
[61] Chief Justice Jordan’s exposition was called ‘classic’ in Environment Protection Authority v Daracon Engineering Pty Ltd (1998) 97 LGERA 415, 416; and Singh Gill v District Court of New South Wales [2001] NSWSC 386 [9]. It was called ‘authoritative’ in Wallaby Grip Ltd v Peirce [2000] NSWCA 299 [7]. A recent check on the Australasian Legal Information Institute website (www.austlii.edu.au>) produced a list of over 60 cases (including some in the High Court) which have used it. See also Attorney General (NSW) v X [2000] NSWCA 199; (2000) 49 NSWLR 653, 677, where Spigelman CJ said that Jordan CJ had expounded a ‘few, old sturdy and serviceable tools’ in an area cluttered with distracting formulas. His Honour was using language from Endicott, above n 44, 297. Endicott used the image of a tool box, cluttered with mostly useless and superfluous tools, but containing some that were useful. The doctrinal task was to sort them out.
[62] See now Pepsi Seven-up Bottlers Perth Pty Ltd v Commissioner of Taxation [1995] FCA 1655; (1995) 62 FCR 289; Dyson v Pharmacy Board [2000] NSWSC 981; (2000) 50 NSWLR 523; and Freight Rail Corporation v Chief Executive Officer of Customs [2000] FCA 1796; (2000) 107 FCR 15.
[63] Australian Gas [1940] NSWStRp 9; (1940) 40 SR (NSW) 126, 137-8 (original references omitted).
[64] R (Alconbury Developments Ltd) v Secretary of State for the Environment [2001] UKHL 23; [2001] 2 WLR 1389 [62]. See Hanlon v McKay Investments Pty Ltd [2001] TASSC 37.
[65] [1956] HCA 21; (1956) 96 CLR 47.
[66] [1915] UKHL TC_6_590; [1915] AC 922, 932.
[67] [1946] HCA 23; (1946) 73 CLR 93, 97.
[68] Hayes [1956] HCA 21; (1956) 96 CLR 47, 51. The final emphasis is added, because it has guided analysis in many cases; eg, Re Monger; Ex parte Dutch [2001] WASCA 220 [11].
[69] See, eg, Étienne Mureinik, ‘The Application of Rules: Law or Fact?’ (1982) 98 Law Quarterly Review 587. Christopher Enright, ‘Distinguishing Law and Fact’, in Chris Finn (ed), Sunrise or Sunset? Administrative Law for the New Millenium (2000) 301 claims both logic and policy in support of the proposition that all applications of law to fact raise questions of law. His policy argument is that any misapplication frustrates the legislative purpose.
[70] The Court in Yusuf v Minister for Immigration and Multicultural Affairs [2001] HCA 30; (2001) 180 ALR 1 held that a number of judicial review grounds made unavailable in the Federal Court could be raised in that court under other grounds without much difficulty.
[71] Minister for Immigration and Multicultural Affairs v Hu (1997) 79 FCR 309 (‘Hu’). The Full Court of the Federal Court was differently constituted in Gamaethige v Minister for Immigration and Multicultural Affairs [2001] FCA 565, which did not mention Hu. Justice Hill’s approach seems to be inconsistent with Hu, but his Honour’s reasoning seems to have been overtaken by Yusuf v Minister for Immigration and Multicultural Affairs [2001] HCA 30; (2001) 180 ALR 1. The Migration Legislation Amendment (Judicial Review) Act 2001 (Cth) repealed this version of s 476.
[72] The example comes from Edwards v Bairstow [1955] UKHL 3; [1956] AC 14, 31, where Viscount Simonds said that a jury properly instructed as to the legal meaning of murder makes an inference of fact when it finds the accused guilty.
[73] Aafjes v Kearney (1976) 180 CLR 199 at 204. His Honour did not specify whether his opposition to the analytic view (as propounded by Lord Parker in Farmer v Cotton's Trustees [1915] UKHL TC_6_590; [1915] AC 922, 932) was limited to determinations of dependency. However, such a limitation seems unlikely from the context.
[74] Mureinik, above n 69, lamented the lack of judicial attention to academic analyses, and blamed Parliament for drafting appeal statutes too widely. Enright, above n 69, 301 blamed both the lawyers and Parliament.
[75] [1980] HCA 16; (1980) 144 CLR 1.
[76] Ibid 7.
[77] [1972] UKHL 6; [1973] AC 854.
[78] Glanville Williams, ‘Law and Fact’ [1976] Criminal Law Review 472, 472, 532; Mureinik, above n 69, 611-13; and D W Elliott, ‘Brutus v Cozens; Decline and Fall’ [1989] Criminal Law Review 322.
[79] See Pearlman v Harrow School [1978] EWCA Civ 5; [1979] 1 QB 56, 66-7; and Sir John Laws, ‘Law and Fact’ [1999] British Tax Review 159.
[80] Alan White, Misleading Cases (1991) 2.
[81] S v Crimes Compensation Tribunal [1998] 1 VR 83, 88.
[82] Universal Magazines Pty Ltd v Comptroller-General of Customs [1990] FCA 244; (1990) 21 ALD 502. See also Soboleva v Minister for Immigration and Multicultural Affairs [2001] FCA 528, and Vassilieva v Minister for Immigration and Multicultural Affairs [2001] FCA 733, where the Federal Court rejected an attempt to confer quasi-precedent status on Refugee Review Tribunal findings as to a particular country’s conditions. It had been argued that if a party relied on a previous Tribunal finding in a case to which he or she was not a party, the Tribunal would have to deal with the argument in its written reasons. It would have to accept, reject or distinguish the previous finding. The argument failed, because the High Court gave a more restricted reading of the matters which had to be covered in the written reasons; see Yusuf v Minister for Immigration and Multicultural Affairs [2001] HCA 30; (2001) 180 ALR 1.
[83] [2000] NSWSC 283; (2000) 49 NSWLR 88, 97-9.
[84] For Australia, see R v Howes [2000] VSCA 159; [2000] 2 VR 141, which held it appropriate for a trial judge to assist the jury by giving them dictionary definitions of words bearing ordinary meanings in a context where only one result was open.
[85] Mureinik, above n 69, 612.
[86] Arthur Glass, ‘Interpretation/Application/Decision-making’ (2000) 25 Australian Journal of Legal Philosophy 97.
[87] But see Commissioner of Taxation v Montgomery [1999] HCA 34; (1999) 198 CLR 639, 661-3.
[88] Hayes v Federal Commissioner of Taxation [1956] HCA 21; (1956) 96 CLR 47, 51; XCO Pty Ltd v Federal Commissioner of Taxation [1971] HCA 37; (1971) 124 CLR 343, 348; Lombardo v Federal Commissioner of Taxation (1979) 40 FLR 208, 218; and TNT Skypack International (Aust) Pty Ltd v Federal Commissioner of Taxation (1988) 82 ALR 175, 182-3.
[89] See TNT Skypack International (Aust) Pty Ltd v Federal Commissioner of Taxation (1988) 82 ALR 175; and Commissioner of Taxation v Roberts [1992] FCA 363; (1992) 37 FCR 246, 251-2.
[90] [2001] HCA 23; (2001) 177 ALR 473.
[91] Huang v Minister for Immigration and Ethnic Affairs (1996) 71 FCR 95.
[92] Aafjes v Kearney (1976) 180 CLR 199, 202. See also Narayan v Minister for Immigration and Multicultural Affairs [2001] FCA 789 [46]-[49], where Sackville J adopted Justice McHugh’s approach.
[93] [1993] FCA 356; (1993) 43 FCR 280, 287.
[94] [1996] HCA 36; (1996) 186 CLR 389.
[95] Ibid 395.
[96] Brutus v Cozens [1972] UKHL 6; [1973] AC 854; and Jedko Game Co Pty Ltd v Collector of Customs (NSW) (1987) 12 ALD 491.
[97] Life Insurance Co of Australia Ltd v Phillips [1925] HCA 18; (1925) 36 CLR 60, 78; NSW Associated Blue-Metal Quarries Ltd v Federal Commissioner of Taxation [1956] HCA 80; (1956) 94 CLR 509, 512; Neal v Department of Transport [1980] FCA 45; (1980) 3 ALD 97, 107-8; and Jedko Game Co Pty Ltd v Collector of Customs (NSW) (1987) 12 ALD 491.
[98] Australian Gas [1940] NSWStRp 9; (1940) 40 SR (NSW) 126, 137-8; and Lombardo v Federal Commissioner of Taxation (1979) 40 FLR 208, 215.
[99] Life Insurance Co of Australia Ltd v Phillips [1925] HCA 18; (1925) 36 CLR 60, 79.
[100] Hope v Bathurst City Council [1980] HCA 16; (1980) 144 CLR 1, 7 (Mason J), with Gibbs, Stephen, Murphy and Aickin JJ agreeing; Australian National Railways Commission v Collector of Customs (SA) [1985] FCA 312; (1985) 8 FCR 264, 277 (Sheppard and Burchett JJ).
[101] Pozzolanic [1993] FCA 356; (1993) 43 FCR 280, 288, citing Hope v Bathurst City Council [1980] HCA 16; (1980) 144 CLR 1, 8.
[102] Yu Feng Pty Ltd v Maroochy Shire Council (1996) 92 LGERA 41, 68, 75 contains the same criticisms.
[103] It might be unambiguous per se, or because prior authority has resolved any ambiguity.
[104] See Yu Feng Pty Ltd v Maroochy Shire Council (1996) 92 LGERA 41, 76, where Pincus JA thought the correctness of rule five to be ‘somewhat doubtful’. It was sufficient for his Honour’s purposes, however, to note that it did not apply where the legal term in question carried its ordinary English meaning.
[105] (2001) 178 ALR 1.
[106] Even if it were such a journey, there were other barriers to compensation, turning on whether the crash was partly or wholly her fault, and whether the risk of a crash had been materially increased by reason of any interruption or deviation in the journey.
[107] Lake Macquarie City Council v Vetter [1999] NSWCA 98; (1999) 18 NSWCCR 34.
[108] Vetter (2001) 178 ALR 1, 9-10.
[109] Ibid 22: Justice Kirby agreed with the joint judgment in this respect, but added that the Court of Appeal may not have intended its ‘one journey or two’ test to be the complete explanation of ‘journey’.
[110] Edwards v Bairstow [1955] UKHL 3; [1956] AC 14, 36.
[111] Vetter (2001) 178 ALR 1, 21. See Soulemezis v Dudley (Holdings) Pty Ltd (1987) 10 NSWLR 247, 254, where his Honour seemed to favour the Edwards v Bairstow test, which he called ‘much wider’ than Australian tests.
[112] Ibid 22.
[113] Ibid 8-9.
[114] Hope v Bathurst City Council [1980] HCA 16; (1980) 144 CLR 1, 7. Their Honours also quoted an earlier judgment of Mason JA in Williams v Bill Williams Pty Ltd [1971] 1 NSWLR 547, 557.
[115] Vetter (2001) 178 ALR 1, 30. Justice Kirby also noted that ‘journey’ carried its ordinary meaning, but his Honour did not indicate whether that was critical to his reasoning: Vetter (2001) 178 ALR 1, 22.
[116] Ibid 9-10 (Gleeson CJ, Gummow and Callinan JJ), quoting from Mahoney JA in Minchinton v Homfray (1994) 10 NSWCCR 778, 785. Justice Hayne had used the same reference.
[117] [1971] 1 NSWLR 547, 557.
[118] His Honour applied what is here called the analytical approach.
[119] Ibid.
[120] Hope [1980] HCA 16; (1980) 144 CLR 1, 7.
[121] See, eg, Blackwood Hodge (Australia) Pty Ltd v Collector of Customs (NSW) [1980] FCA 96; (1980) 47 FLR 131, 145-6.
[122] See, eg, Universal Magazines Pty Ltd v Comptroller-General of Customs [1990] FCA 244; (1990) 21 ALD 502; Australian Football League v Carlton Football Club Ltd [1998] 2 VR 546, 553, 556, 559, 564 (but note that instead of expressing the margin as in Edwards v Bairstow, the court seemed to prefer a test in terms of whether the application of an evaluative legal term to the facts was so aberrant as to be irrational); Larsen v Vile [1999] NSWCA 397 [38]-[39]; Edgley v Federal Capital Press of Australia Pty Ltd [2001] FCA 379 [85]; and Saha v Minister for Immigration and Multicultural Affairs [2001] FCA 520 [57]-[59]. See also DPP v Jones [1999] UKHL 5; [1999] 2 AC 240 (whether a demonstration was a ‘reasonable use’ of a highway was a question of degree, and therefore, fact).
[123] See Commissioner of Taxation v Cooper [1991] FCA 190; (1991) 29 FCR 177, 194; Commissioner of Taxation v Roberts [1992] FCA 363; (1992) 37 FCR 246, 251-2; S v Crimes Compensation Tribunal [1998] 1 VR 83, 89; and Edgley v Federal Capital Press of Australia Pty Ltd [2001] FCA 379 [85]-[86].
[124] Justice Windeyer took that approach in Da Costa v R [1968] HCA 51; (1968) 118 CLR 186, where leave to appeal was needed if the question was mixed, but not if it involved ‘a question of law alone’. Justice Gibbs said in Aafjes v Kearney (1976) 180 CLR 199, 207 that facts were not ‘fully found’ if one must draw ‘an inference of fact [in addition to the facts found] before the ultimate facts can be determined’.
[125] See Da Costa v R [1968] HCA 51; (1968) 118 CLR 186, 195, where Windeyer J quoted a critical passage from Thomas Starkie, Evidence (3^rd ed, 1842) vol 1, 519-20.
[126] See, eg, TNT Skypack International (Aust) Pty Ltd v Federal Commissioner of Taxation (1988) 82 ALR 175, 182; and Commissioner of Taxation v Roberts [1992] FCA 363; (1992) 37 FCR 246, 251-2.
[127] Matthew Hall Pty Ltd v Smart [2000] NSWCA 284 [48].
[128] See Blackwood Hodge (Australia) Pty Ltd v Collector of Customs (NSW) [1980] FCA 96; (1980) 47 FLR 131, 145-6 (asking whether no reasonable person could have come to the particular conclusion); and Matthew Hall Pty Ltd v Smart [2000] NSWCA 284 (rejecting the term ‘perverse’).
[129] Attorney General (NSW) v X [2000] NSWCA 199; (2000) 49 NSWLR 653, 677.
[130] Ibid.
[131] Ibid 677.
[132] S v Crimes Compensation Tribunal [1998] 1 VR 83, 90-1.
[133] Hope v Bathurst City Council [1980] HCA 16; (1980) 144 CLR 1, 7.
[134] S v Crimes Compensation Tribunal [1998] 1 VR 83, 88-9.
[135] Lombardo v Federal Commissioner of Taxation (1979) 40 FLR 208.
[136] S v Crimes Compensation Tribunal [1998] 1 VR 83, 91.
[137] Ibid 88.
[138] See, eg, Industry Research and Development Board v Bridgestone Australia Ltd [2001] FCA 954.
[139] Hope v Bathurst City Council [1980] HCA 16; (1980) 144 CLR 1, 8-10.
[140] Agfa [1996] HCA 36; (1996) 186 CLR 389, 394.
[141] The decline in jury usage accounts for that shift in the judicial arena, whilst the spread of statutory duties to give reasons (at least on request) accounts for the shift in the administrative arena. It is difficult (although strictly not impossible) to infer misunderstanding of the governing law when the reasons for decision stated the law correctly.
[142] Inn Leisure Industries Pty Ltd (Prov Liq) v DF McCloy Pty Ltd (No 1) (1991) 28 FCR 151, 168.
[143] Hu (1997) 79 FCR 309, 324.
[144] Blackwood Hodge (Aust) Pty Ltd v Collector of Customs (NSW) [1980] FCA 96; (1980) 47 FLR 131, 145-6; R v Hull University Visitor; Ex parte Page [1993] 2 AC 682, 694 (Lord Griffiths); Craig v South Australia [1995] HCA 58; (1995) 184 CLR 163, 186 (Brennan CJ, Deane, Toohey, Gaudron and McHugh JJ); and Kleinwort Benson Ltd v Lincoln City Council [1998] UKHL 38; [1999] 2 AC 349, 372 (Lord Goff).
[145] Aafjes v Kearney (1976) 180 CLR 199, 206 (Gibbs J). Justice Hill called it sterile in Commissioner of Taxation v Roberts [1992] FCA 363; (1992) 37 FCR 246, 252.
[146] As stated by Kirby J in Vetter (2001) 178 ALR 1, 20, quoting from the joint judgment in Agfa [1996] HCA 36; (1996) 186 CLR 389, 396.
[147] Nizich v Federal Commissioner of Taxation (1991) 91 ATC 4 747, 4 752; and Commissioner of Taxation v Roberts [1992] FCA 363; (1992) 37 FCR 246, 252. By contrast, Pincus JA in Friends of Stradbroke Island Association Inc v Sandunes Pty Ltd (1998) 101 LGERA 161, 163 hailed Agfa [1996] HCA 36; (1996) 186 CLR 389 as removing the earnest inquirer from the Slough of Despond.
[148] Hanlon v McKay Investments Pty Ltd [2001] TASSC 37 [9].
[149] [1968] HCA 51; (1968) 118 CLR 186, 194.
[150] [2000] NSWCA 199; (2000) 49 NSWLR 653.
[151] Da Costa v R [1968] HCA 51; (1968) 118 CLR 186, 194.
[152] Namely, an Act empowering the Attorney to refer a question of law to the Court of Appeal, whose answer would not affect the trial judge’s decision.
[153] However, the significance of the law-fact distinction is often diminished to a threshold question; eg, an appeal may need to identify a question or error of law, but the appellate court’s function might then extend to errors of fact because that court’s more general appellate powers have been triggered. The possibilities are discussed in: Federal Commissioner of Taxation v Walker (1984) 2 FCR 283; TNT Skypack International (Aust) Pty Ltd v Federal Commissioner of Taxation (1988) 82 ALR 175; De Domenico v Marshall [2001] ACTSC 52; Maurici v Chief Commissioner of State Revenue [2001] NSWCA 78 [53]-[57]; and Vetter (2001) 178 ALR 1, 5-6, 20.
[154] See Bond [1990] HCA 33; (1990) 170 CLR 321; Aronson and Dyer, above n 25, ch 5; and Aronson, above n 39.
[155] See Craig v South Australia [1995] HCA 58; (1995) 184 CLR 163; Re Bennett-Borlase; Ex parte Commissioner of Police (Unreported, WA Supreme Court, Full Court, 20 June 1997); Returned & Services League of Australia (Vic Branch) Inc v Liquor Licensing Commission [1999] VSCA 37; [1999] 2 VR 203, 214; Hartley v O’Loughlin [1999] VSC 138 [28]; Re Robins; Ex parte West Australian Newspapers Ltd [1999] WASCA 16; (1999) 20 WAR 511; Re Real Estate and Business Agents Supervisory Board; Ex parte Cohen [1999] WASCA 47; (1999) 21 WAR 158, 165-6; Re Calder; Ex parte St Barbara Mines Ltd [1999] WASCA 25 [39]-[44]; Re Plutonic Operations Ltd; Ex parte Roberts [1999] WASCA 133 [47]; Construction, Forestry, Mining and Energy Union v Australian Industrial Relations Commission [1999] FCA 847; (1999) 93 FCR 317, 339-41; Edwards v Guidice [1999] FCA 1836; (1999) 94 FCR 561, 567, 590-2; Absolon v NSW TAFE Commission [1999] NSWCA 311 [146]; Custom Credit Corporation Ltd (In Liq) v Commercial Tribunal (NSW) [2000] ASC 155-041, [101]-[102]; Sophiadakis v Reljic [1999] VSC 147; Greek Orthodox Community of SA Inc v Ermogenous [2000] SASC 329 [66]; Police Association (NT) Inc v Police Arbitral Tribunal [2000] NTSC 32; Coal and Allied Operations Pty Ltd v Australian Industrial Relations Commission [2000] HCA 47; (2000) 174 ALR 585; Re Ruddock; Ex parte Reyes (2000) 177 ALR 484, 487; Minister for Immigration and Multicultural Affairs v Yusuf [2001] HCA 30; (2001) 180 ALR 1, 32; Berry v Melbourne Magistrates’ Court [2001] VSC 228; and Matson v Racing Appeals Tribunal [2001] VSC 264.
[156] R v Hull University Visitor; Ex parte Page [1992] UKHL 12; [1993] AC 682; and Boddington v British Transport Police [1998] UKHL 13; [1999] 2 AC 143, 154.
[157] See Re Minister for Immigration and Multicultural Affairs; Ex parte Miah (2001) 179 ALR 238, 290-7. See also Aronson, above n 39, 18-21.
[158] Administrative Decisions (Judicial Review) Act 1977 (Cth) s 5(1)(f); Administrative Decisions (Judicial Review) Act 1989 (ACT) s 5(1)(f); and Judicial Review Act 1991 (Qld) s 20(2)(f).
[159] Administrative Law Act 1978 (Vic) s 10; Supreme Court Act 1970 (NSW) s 69(4); and ASIC v Farley [2001] NSWSC 326.
[160] Endicott, above n 44.
[161] Jack Beatson, ‘The Scope of Judicial Review for Error of Law’ (1984) 4 Oxford Journal of Legal Studies 22.
[162] Endicott, above n 44.
[163] David Securities Pty Ltd v Commonwealth Bank of Australia [1992] HCA 48; (1992) 175 CLR 353, 374.
[164] [2000] NSWCA 199; (2000) 49 NSWLR 653.
[165] Ibid 660-1.
[166] British Launderers’ Research Association v Hendon Borough Rating Authority [1949] 1 KB 464, 472.
[167] Edwards v Bairstow [1955] UKHL 3; [1956] AC 14, 38. See also TNT Skypack International (Aust) Pty Ltd v Federal Commissioner of Taxation (1988) 82 ALR 175, 179, where Gummow J referred to that part of his Lordship’s speech, and seemed to imply that at the Commonwealth level in Australia, constitutional principle forbids judicial deference to administrative applications of legal standards.
[168] Azzopardi (1985) 4 NSWLR 139.
[169] Aafjes v Kearney (1976) 180 CLR 199, 204.
[170] Ibid 210.
[171] [1978] EWCA Civ 5; [1979] 1 QB 56.
[172] Ibid 66-7.
[173] Namely, Jones v Leeming [1930] AC 415.
[174] Cf McHugh J in Re Minister for Immigration and Multicultural Affairs; Ex parte Cohen [2001] HCA 23; (2001) 177 ALR 473, discussed above.
[175] See R v Inland Revenue Commissioners; Ex parte National Federation of Self-Employed and Small Businesses Ltd [1981] UKHL 2; [1982] AC 617, 651; and Boddington v British Transport Police [1998] UKHL 13; [1999] 2 AC 143, 170. See also Kruse v Johnson [1898] UKLawRpKQB 101; [1898] 2 QB 91, 99-100.
[176] See Sunshine Coast Broadcasters Ltd v Duncan (1988) 83 ALR 121, 132; Fares Rural Meat and Livestock Co Pty Ltd v Australian Meat and Livestock Corporation [1990] FCA 139; (1990) 96 ALR 153, 167-8; David Jones Finance and Investments Pty Ltd v Commissioner of Taxation (1991) 28 FCR 484, 488, 504; Premalal v Minister for Immigration, Local Government and Ethnic Affairs (1993) 41 FCR 117, 140; Hamilton v Minister for Immigration, Local Government and Ethnic Affairs [1993] FCA 657; (1993) 48 FCR 20, 36-7 (the appeal did not touch this point, Hamilton v Minister for Immigration and Ethnic Affairs [1994] FCA 1424; (1994) 53 FCR 349); NSW Aboriginal Land Council v Aboriginal and Torres Strait Island Commission (1995) 59 FCR 369, 387; Pickering v Deputy Commissioner of Taxation (1997) 37 ATR 41, 48-50; Pickering v Federal Commissioner of Taxation (1998) 98 ATC 4 977; Bellinz Pty Ltd v Federal Commissioner of Taxation [1998] FCA 615; (1998) 84 FCR 154, 167; and Daihatsu Australia Pty Ltd v Deputy Commissioner of Taxation (2000) 46 ATR 129, 143-8. See also Aronson and Dyer, above n 25, 292-4.
[177] Aronson and Dyer, above n 25, 234-9.
[178] Chevron USA Inc v Natural Resources Defence Council Inc [1984] USSC 140; 467 US 837 (1984). This case has generated mountains of analysis and case law. Its ‘repositioning’ in United States v Mead Corp 530 US 1296 (2001) will produce similarly sized mountains, as its vehement dissenter predicted. Canada also has a ‘deference’ standard for applications of law to fact where the context is so particularised that no new law is being made; see Canada (Director of Investigation and Research) v Southam Inc [1997] 1 SCR 748.
[179] Enfield Corporation v Development Assessment Commission [2000] HCA 5; (2000) 199 CLR 135, 151-4. See also Justice W M C Gummow, ‘The Permanent Legacy’ (2000) 28 Federal Law Review 177, 183-4.

University of New South Wales Law Journal

Aronson, Mark --- "Unreasonableness and Error of Law" [2001] UNSWLawJl 26; (2001) 24(2) UNSW Law Journal 315

Unreasonableness And Error Of Law

I INTRODUCTION: MIXING FACT AND LAW

II UNREASONABLENESS AS LAW AND FACT

III CHALLENGING PRIMARY FACT-FINDING

IV CHALLENGING THE APPLICATION OF LAW TO THE PRIMARY FACTS

V ANALYTICAL VS PRAGMATIC APPROACHES

VI DETERMINING AND APPLYING ORDINARY MEANINGS

VII SHORTENING THE LISTS OF CRITERIA

VIII UNIVERSAL PRINCIPLES, CYNICAL MANIPULATION, OR VARIABLE DEFINITIONS?

IX THE PRECEDENT FACTOR

X CONCLUSIONS