Prepublication Draft 2009

Overruling as a speech act: performativity and normative discourse

Ross CHARNOCK Université de Paris

Abstract: In the common law system, judges are said to be bound by precedents decided in courts of the same level or above. However, in higher courts, under certain conditions, they have the right to overrule. Overruling declarations may be analysed as performative speech acts, having the effect of changing the law. This analysis raises both linguistic and legal problems, discussed with reference to English and American law and language. As the judges are reluctant to be seen to be assuming a legislative function, they tend to use indirect rather than explicit language, especially in the most significant cases. Alternatively, they present their overruling decisions not as new legislation, but rather as declarations of the true state of the unchanging common law. However, this view implies increased illocutionary force, as it may involve retrospective application. Secondly, the legal validity of overruling declarations depends to a large extent on their perlocutionary effects. Even after successful performance, these effects may be cancelled by later decisions in higher courts. Finally, the legal effect of overruling decisions suggests a close relation between performativity and normativity. However, this relation does not in itself provide a satisfactory explanation of the normativity of judicial discourse.

1. Introduction

It is trivially true that legal procedure involves performative utterances. In criminal trials, in the common law system at least, the accused is required to plead guilty or not guilty. During the interrogation of witnesses, legal representatives may declare an objection to a particular line of questioning, using either the explicit performative form (“I object”), or, perhaps more typically through the conventional use of the single word “objection”. The judges may sustain or reject such objections, using such phrases as “objection sustained” or “objection overruled”. They are also required to make an official declaration of the result of each case, with such phrases as “appeal allowed”, “judgment affirmed” or “so ordered”. Equally clearly, legal documents (also known, significantly, as “instruments”, precisely because of the legal acts which they are used to represent) also have a performative function. A will is not a simple description of how property will be distributed at some time in the future, or a mere statement of preferences. On the contrary, such documents have an ascriptive function (Hart 1951). Similarly, contracts are not simply descriptions of what is envisaged, but prototypical speech acts involving reciprocal promises, which create legal obligations. Like ordinary language performatives, which may fail (or misfire) in infelicitous circumstances, putative contracts may be invalid or void if certain necessary features are not present. It may be observed that such legal speech acts do not necessarily correspond to Austin’s (1962) canonic performative form; they do not make systematic use of explicit performative verbs, and they are often expressed in the passive. Nor are they necessarily expressed in the first person, as the judge is not speaking in his personal capacity, but rather in the name of the court. While wills need not be expressed orally, certain contracts need not be expressed verbally at all.1 Nevertheless, to the extent that they create new rights and duties, they may still be said to have performative effect. Furthermore, the legal system as a whole is neither descriptive or constative (Austin 1962), but rather prescriptive. In his early discussion of positivism and the normative nature of law, Austin (1832) considered the law as based on a fundamental speech act. In his theory, rules are seen as general commands, emanating from a metaphorical ‘sovereign’ or political superior: “Every law or rule [...] is a command. Or rather, laws or rules, properly so-called, are a species of command” (1832: 21). The law is therefore fundamentally performative in nature. This approach now appears somewhat simplistic, as it does not account for power- conferring or for constitutive rules, which cannot be said to create obligatory duties. English Statutes (also known as “Acts” of Parliament) are clearly presented as “speech acts”, being founded on a conventional introductory formula, expressed by means of the jussive subjunctive:

“Be it enacted by the Queen’s most excellent Majesty, by and with the consent of the Lords Spiritual, and

1 Some contracts may be accepted through performance; others, for example contracts for the sale of real property, are only valid when reduced to written form. Commons, in this present Parliament assembled, and by the authority of the same, as follows: [...].”2

Successful performance of the act of legislation requires appropriate felicity conditions, in particular the fact that (in the English system) the Bill has been passed by both Houses of Parliament and received the Royal assent.3 Austin (1962) points out that, while performative speech acts may be described as inappropriate or insincere, they cannot be said to be false. In the same way, a legal statute may not be considered right or wrong, although it may be thought good or bad. Perhaps this is what he (1962: 4, n 2) had in mind when he described the idea of the law as a statement of fact as a “timorous fiction”.

“Of all people, jurists should be best aware of the true state of affairs. Perhaps some now are. Yet they will succumb to their own timorous fiction, that a statement of ‘the law’ is a statement of fact.” 4

It is clear that legislation has the effect of creating, rather than simply describing law.

1.1 Judge-made law In all legal systems, judges are required to declare how the law should be interpreted and applied in circumstances which are not covered by existing law and which may not have been envisaged by the legislator. In this sense, the judges may be said to have an “interstitial” legislative function (Holmes, 1881). In the common law system, according to the rule of precedent (or stare decisis), the judges are bound by earlier decisions given in courts of the same level or above. In theory, such decisions are said to constitute binding precedents, having force of law Thus judicial as well as legislative declarations have the function of creating new legal rules.5 Where the common law precedents appear to lead to injustice, they may be avoided in various ways. Like statutes, they may be reinterpreted in order to obtain a more acceptable result. Alternatively, the judge may claim that the rule stated in the earlier case was not part of the ratio decidendi, but only obiter dicta (or passing remarks), in which case it will be considered not as binding but as merely persuasive. In addition, he may “distinguish” the earlier case, by showing that the (new) facts are different in some material way. This latter technique allows the judges to develop exceptions to the general rule, without necessarily rejecting the underlying principle. In this way, they are able to preserve the previously existing law, while restricting the conditions required for its valid application. Where the principle itself is no longer found acceptable, higher courts have the right, under certain conditions, to overrule the earlier decision, whether in order to correct regrettable decisions or to adapt the common law to changing social needs. To overrule an established precedent, which may have been followed in a series of cases, amounts to a declaration that the law is no longer as it was hitherto understood to be. Such decisions are often seen in retrospect as constituting historical turning points in the law. This judicial function is often stated to be a necessary feature of the legal system:

“The whole of the common law is judge-made and only by judicial change in the law is the common law kept relevant in a changing world.” (Kleinwort Benson v Lincoln CC 1998, per Lord Browne-Wilkinson)

As common law judgments are given by the judges orally, in open court, the performative nature of such declarations may seem to appear even more clearly than is the case with primary legislation, in which the written form predominates, the utterance act being merely symbolised by an official declaration.6 However, this view of the common law is in contradiction with the view conventionally attributed to the general public, according to which the function of judges is simply to state and to apply pre-existing law.

2 The typical US (Federal) equivalent is “Be it enacted by the Senate and the House of Representatives of America in Congress assembled. 3 The Royal assent is given in the third person, and in old French, in the form: “La reine le veult”. 4 See also Austin (1962: 19): “Only the still widespread obsession that the utterances of the law, and utterances used in, say, ‘acts in the law’, must somehow be statements true or false, has prevented many lawyers from getting this whole matter much straighter than we are likely to.” 5 In practice the “rule of precedent” can be applied in different ways, and thus functions as a guiding principle, rather than as an absolute rule. 6 In practice, as this procedure now appears antiquated, it tends to be preserved only in the higher courts. In the lower courts (especially in the US), the parties are more likely to receive the opinions by mail. Commentators often speak of American judges as “writing that ...”, rather than “declaring that ...” In common law jurisdictions, even where the judgments are given orally, the written reports are still referred to as authority. However, this is because the report functions as an authoritative record of what was said For them to do otherwise would be to assume an unacceptable level of discretion. Judges are therefore reluctant to be seen to overrule, and avoid doing so where it is not strictly necessary. Even when they are obliged to do so in order to do justice in new, unforeseen situations, they often claim, contrary to the evidence, that they are simply applying the established law.7 Judicial attitudes to overruling therefore remain equivocal, and this is reflected in the language used to signal changes in the law. Before discussing the legal effect of overruling declarations, and the linguistic problems raised, it is therefore appropriate to discuss the language of overruling as observed in a number of celebrated cases in English and American law.8

2. The language of overruling

Judges are reluctant to be seen to be changing the law, even where they are convinced that change would be beneficial. Since the availability of law reports,9 they have preferred to follow established precedent. Until the Practice Statement’ of 1966, the House of Lords did not even have the right to overrule its own precedents. Lord Halsbury LC made it clear in 1898 that “a decision of this House once given upon a point of law is conclusive upon this House afterwards”, so that “it is impossible to raise that question again as if it was res integra and could be reargued, and so the House be asked to reverse its own decision.” In confirming the principle according to which, once a decision was given, the rule could only be overturned by an ,. he observed that it had been established for “some centuries” (London Street Tramways v LCC HL 1898). As a judge of the enlightenment years, Lord Mansfield is remembered as an important figure in the development of the English common law, in part because he refused to follow precedent blindly.10 Nevertheless, he often found it necessary or expedient to follow precedents which he considered unreasonable and illogical. In R v Pedley (1782), he found himself reluctantly obliged to declare that it was not a felony at common law for a tenant to set fire to a house of which he was in possession. He considered that the earlier case was “based on wretched reasoning”, but observed:

“[I]f the point were now originally before the court, I could hardly have assented. But the question has been submitted to the consideration of the judges, and nine of them (all who attended) were unanimous [...] the legislature alone can therefore now supply the remedy.” R v Pedley 1782, per Mansfield CJ)

In R v Wilkes (1770), under public pressure to reverse a declaration of outlawry, which would have led to injustice if allowed to stand, he relied on old authorities requiring technical forms of words:

“The authorities I have stated stand, to this day, uncontradicted. They are many; and have prevailed above a century. I think, they begun against law and reason. The former authorities were otherwise [...]. There is no reason for requiring these words: there is sufficient certainty, without them. [...]. But, this is a criminal case, highly penal.” (R v Wilkes 1770, per Mansfield CJ)

The rule of precedent is often a source of difficulty, especially where the judge considers that the earlier case was wrongly decided. However, it has the advantage of ensuring a certain stability in the law. In this context, Lord Mansfield often stated that legal certainty was of greater importance than the result in particular cases:

“In all mercantile transactions the great object should be certainty: and therefore, it is of more consequence that a rule should be certain, than whether the rule is established one way or the other. Because speculators in trade then know what ground to go upon.” (Vallejo v Wheeler 1774, per Mansfield CJ)

The need for stability explains why respect is paid to decisions which have been followed for many years, in spite of inconvenient consequences. A well-known example is Pinnel’s case (1602), involving payment of a lesser sum in satisfaction for a debt. A basic rule in English law is that contracts are normally unenforceable in the absence of “consideration”, given in return for the promise. It follows that a simple, unilateral promise does not constitute a contract. Further, it has long been established that a promise to fulfil a pre-existing 7 For Shapiro (1994) this amounts to a lie. 8 Although those trained in the English or US law will be familiar with these cases, brief indications of the facts and arguments may be useful to linguists. 9 The English Reports (W. Green, Edinburgh 1932, 180 volumes) is a collection of early reports, including cases 1120-1865. 10 He was often justified in questioning the accuracy of the report. obligation cannot be accepted as legally sufficient consideration. Following these rules, it was stated by Lord Coke in Pinnel that a promise to accept part payment in settlement of a debt could not be enforced. Eminently logical rules were laid down, according to which payment of a lesser sum before the due date at the request of the creditor, or on the due date in a different place, or accompanied by “an object of any value”, may be allowed as good consideration. However, payment of less than the full sum cannot otherwise be legal satisfaction. In consequence, even after payment has been made as agreed, the creditor cannot be prevented from suing for the rest. Predictably enough, the strict application of this rule led to injustice when creditors failed to keep their promises. Exceptions were made where possible, for example where the payment was made by a third party. However, the multiplication of exceptions to the general rule led to incoherence. In Couldery v Bartrum (1881), Sir George Jessel complained that, because the rule decided in Pinnel’s case referred to “an object of any value”, the courts were entitled, in spite of the examples given in the original case,11 to accept “canary- birds, tom-tits, or rubbish of that kind” as valid consideration. Yet a substantial sum of money could never be sufficient. In Foakes v Beer (1884) the principle was stretched to its fullest extent. A debtor was given time to pay and eventually paid in full. However, the creditor then sued for payment of interest (attracted automatically by a legal debt). It was generally thought that this was morally wrong, and when the case reached the House of Lords, it was assumed that the judges would take the opportunity to overrule Pinnel. However, in a celebrated speech, the Lord Chancellor expressed his reluctance to overrule:

“ The question, therefore, is nakedly raised by this appeal, whether your Lordships are now prepared, not only to overrule, as contrary to law, the doctrine stated by Sir Edward Coke to have been laid down by all the judges of the Common Pleas in Pinnel’s Case, but to treat a prospective agreement by a series of payments on account to a total amount less than the whole debt, as binding in law, provided those payments are regularly made. [...] The doctrine itself, as laid [sic] by Sir Edward Coke, may have been criticised, as questionable in principle, by some persons whose opinions are entitled to respect, but it has never been judicially overruled; on the contrary I think it has always, since the sixteenth century, been accepted as law. If so, I cannot think that your Lordships would do right, if you were now to reverse, as erroneous, a judgment of the Court of Appeal, proceeding upon a doctrine which has been accepted as part of the law of England for 280 years.” (Foakes v Beer HL 1884, per Earl of Selbourne LC)

He seems to have considered that the older the case, the greater the authority. Lord Blackburn, sitting in the same case, considered it absurd that the inclusion of a “pepper-corn” in addition to the agreed sum, could make an agreement legally enforceable. However, he accepted the recommendation of the Lord Chancellor not to overrule. Thus the judges in Foakes agreed unanimously to follow the longstanding rule. The result was that Pinnel’s case, now upheld by the House of Lords, had renewed authority. English judges continue to express their reluctance to overrule. In Rickards v Oppenhaim (1950), for example, Lord Simonds remarked on the unwisdom of casting doubt on “decisions of respectable antiquity in order to introduce a greater harmony into the law of charity as a whole”. In the US, as in England, the certainty of an unchanging law is also stated to be more important than justice in particular cases:

“Adhering to precedent is usually the wise policy, because in most matters it is more important that the applicable rule of law be settled than it be settled right.” (Burnet v Coronado Oil & Gas Co 1932, per Brandeis J, dissenting)

Eminent judges like Cardozo J preferred not to reject the rules laid down, even where they appeared to have no logical justification. In Allegheny College v National Chautauqua County Bank (1927), he decided not to overrule established precedents requiring consideration in contract. He justified his decision in this way:

“ Decisions which have stood so long, and which are supported by so many considerations of public policy and reason, will not be overruled to save the symmetry of a concept which itself came into our law, not so much from any reasoned conviction of its justice, as from historical accidents of practice and procedure. The concept survives as one of the distinctive features of our legal system. We have no thought to suggest that it is obsolete or on the way to be abandoned.” (Allegheny College v National Chautauqua County Bank 1927, per Cardozo J)12

11 The examples given were: “horse, hawk or robe”. 12 Although he explicitly refused to overrule, Cardozo was in fact drastically reinterpreting the applicable case law in order to obtain More recently, in Planned Parenthood of Southeastern PA v Casey (1992), a leading case on the rule of precedent, Rehnquist CJ wished to take the opportunity of overruling Roe v Wade (1973), in which it had been decided that state laws restricting the right to abortion were unconstitutional. However, it was held in an opinion written by a plurality of judges in the name of the Court, that such a step would be a negation of the rule of stare decisis, and would weaken the rule of law:

“Overruling Roe’s central holding would not only reach an unjustifiable result under stare decisis principles, but would seriously weaken the court’s capacity to exercise the judicial power and to function as the Supreme Court of a Nation dedicated to the rule of law.” (Planned Parenthood of Southeastern PA v Casey 1992)

It was considered that a decision to overrule should rest on some special reason “over and above the belief that a prior case was wrongly decided”.

2.1 Necessity When they do overrule, judges feel obliged to state their reasoning explicitly. In order to minimise any possible weakening of judicial authority, they attempt to demonstrate that no other decision was possible, that the earlier decision was clearly wrong and can no longer be followed without causing injustice, and that the new statement of the law will not have adverse consequences. In the UK, in BR v Herrington (1972), the House of Lords gave detailed arguments to justify the decision to overrule Addie (Collieries) v Dumbreck (1929). The earlier case had attracted criticism, but was nevertheless still accepted as law. According to Addie, a landowner owed no duty of care to trespassers until he became aware of their presence on his land. In Herrington, it was considered on the contrary that the presence of child trespassers was reasonably foreseeable, and that if the landowner failed to take precautions, he should be liable for any harm or damage. Lord Wilberforce pointed out that it was not possible to create exceptions to the Addie rule in order to achieve the desired result. To the objection that a change in the law was a matter for Parliament, he answered that the earlier rule was judge-made, and that, although the legislature had taken no legislative action on the subject, it could not be presumed to have intended to prevent all new developments:

“There might be some force in an argument that for this House to depart from (i.e. overrule) Addie’s case would, in effect, be to legislate where Parliament has abstained, but I can see no sense in supposing that when Parliament left the law alone as regards trespassers the intention was to freeze the law as it was taken to be, in 1929.” (BR v Herrington 1972, per Lord Wilberforce)

Although equally reluctant to overrule, Lord Reid also thought it necessary to reject the earlier decision:

“I dislike usurping the functions of Parliament. But it appears to me that we are confronted with the choice of following Addie and putting the clock back or drastically modifying the Addie rules.” (BR v Herrington 1972, per Lord Reid)

Similar arguments are given in more recent cases. In Mannai Investment v Eagle Star Insurance (1997), a tenant mistakenly gave notice to “determine” (terminate) his lease on 12 January, whereas the correct date should have been 13 January 1995. The court could only preserve the rights of the tenant by correcting the mistake. According to a precedent which had remained unchallenged for fifty years, the court had the right to correct slips (or misprints), but not to substitute one date for another. Lord Hoffmann rejected the argument that judicial modification of the law would lead to uncertainty, and considered that the advantages of a common sense approach outweighed the disadvantages of abandoning authority:

“Nor do I think that a decision overruling the old cases will create uncertainty as to what the law is. In fact I think that the present law is uncertain and that only a decision of this House, either adopting or rejecting the Hankey v Clavering rule of construction, will make it certain.” (Mannai Investment Co Ltd v Eagle Star Assurance HL 1997, per Lord Hoffmann)

The American courts sometimes show less respect for earlier decisions. In Payne v Tennessee (1991), the United States Supreme Court exerted its authority in overruling a series of cases which had followed Booth v Maryland (1987), pouring scorn on the original decision: an acceptable result “If there was ever a case that defied reason, it was Booth v Maryland, imposing a constitutional rule that had absolutely no basis in constitutional text, in historical practice, or in logic.” (Payne v Tennessee 1991, per Justice Scalia)

Because overruling judgments may be based on rhetoric and personal preference, rather than on objective legal argumentation, they are not always unanimous. Dissenting judges often protest in no uncertain terms against the decision of the majority. In Mapp v Ohio (1961), for example, the US Supreme Court held that evidence obtained by searches and seizures, in violation of the Federal Constitution, should be considered inadmissible, not just in Federal but also in State courts. This was contrary to the rule declared earlier in Wolf v Colorado (1949). Justice Harlan dissented:

“In overruling the Wolf case the Court, in my opinion, has forgotten the sense of judicial restraint which, with due regard for stare decisis, is one element that should enter into deciding whether a past decision of this Court should be overruled.” (Mapp v Ohio 1961, per Justice Harlan, dissenting)

In National League of Cities v Usery (1976), the Supreme Court declared that the Fair Labor Standards Act could be enforced on the separate States so as to apply to state employees. This was contrary to the decision taken in Maryland v Wirtz (1968). Justice Brennan protested because he considered this overruling decision inadequately justified:

“ I cannot recall another instance in the Court’s history when the reasoning of so many decisions covering so long a span of time has been discarded in such a roughshod manner. That this is done without any justification not already often advanced and consistently rejected, clearly renders today’s decision an ipse dixit reflecting nothing but displeasure with a congressional judgment. [...] Without even a passing reference to the doctrine of stare decisis, Wirtz - regarded as controlling only last Term [...] is by exercise of raw judicial power overruled.” (National League of Cities v Usery 1976, per Justice Brennan, dissenting) 13

Justice Marshall dissented in Payne v Tennessee (1991):

“In dispatching Booth and Gathers to their graves, today’s majority ominously suggests that an even more extensive upheaval of this Court’s precedents may be in store. [...] The implications of this radical new exception to the doctrine of stare decisis are staggering. The majority today sends a clear signal that scores of established constitutional liberties are now ripe for reconsideration, thereby inviting the very type of open defiance of our precedents that the majority rewards in this case.” (Payne v Tennessee 1991)

2.2 Prior overruling Both English and American judges prefer to avoid taking personal responsibility for changes in the law. They often claim instead that inconvenient precedents have already been considered and overruled by earlier courts. However, the decisions in those cases are rarely as clear as claimed. Mutual Life v Evatt PC (1971) involved a claim for compensation for financial loss following negligent investment advice. The defendants relied on a longstanding rule derived from Le Lievre v Gould (1893) and affirmed in Candler v Crane Christmas (1951), according to which purely financial loss is not recoverable in cases of negligence. In Mutual Life it was claimed that this rule had been overruled in Hedley Byrne (1964):

“In Hedley Byrne 1964, Le Lievre v Gould 1893, which concerned a practising surveyor, and Candler v Crane, Christmas & Co 1951, which concerned practising accountants, were expressly overruled.” (Mutual Life v Evatt PC 1971, per Lord Diplock)

However, in Hedley Byrne itself, the decision is far from being as explicit as suggested. Indeed, in that case, Lord Devlin carefully avoided any clear pronouncement on the subject, simply presenting the dissenting

13 Usery itself was overruled in Garcia v San Antonio Metropolitan Transit (1985). In the latter case, Justice Blackmun appeared to pay more respect to “recent” than to long established precedent: “We do not lightly overrule recent precedent. We have not hesitated, however, when it has become apparent that a prior decision has departed from a proper understanding of congressional power under the Commerce Clause (per Blackmun J). judgment of Lord Denning in Candler as a more acceptable statement of the law:

“I am prepared to adopt any one of your lordships’ statements as showing the general rule; and I pay the same respect to the statement by Denning LJ in his dissenting judgment in Candler v Crane, Christmas & Co about the circumstances in which he says a duty to use care in making a statement exists.” (Hedley Byrne v Heller Partners 1964, per Lord Devlin)

There is occasional debate as to whether inconvenient common law precedents have been overtaken by statute. In Beswick v Beswick (1966), Danckwerts LJ admitted that common law rules preventing third party rights in contract had never been judicially overruled; however, the Law of Property Act (1925), section 56(1), dealt with third party rights in “right of entry, covenant or agreement over or respecting land or other property”. Given that ‘property’ was defined in the statute as including a “thing in action”, it may be understood as including contractual rights. This wide reading would allow the extension of contractual rights to third parties, notwithstanding the longstanding common law rules stated in Tweddle v Atkinson (1861). He admitted that the change in the law was not made in express terms, but considered that it was none the less real:

“It may be that the change in the law was not done with a fanfare of trumpets, but the words of the section, in my opinion, are clear and cannot be ignored. If the views of Lord Denning MR and myself are correct, as I think they are, Tweddle v Atkinson has received the mortal blow which it well deserved.” (Beswick v Beswick CA 1966, per Danckwerts LJ)

In support of this opinion, Lord Denning insisted that a third person may have rights under contract, stating baldly that “The observations to the contrary in Re Miller’s agreement [...] are in my opinion erroneous”. Unsurprisingly, this ingenious and original interpretation was later found unacceptable by the Lords (1968) on the grounds that the Law of Property Act was a consolidating, not a codifying measure, and that Parliament could not have intended “surreptitiously to change a longstanding rule of law in another field”.14 A similar strategy is observed in American cases. In Benton v Maryland (1969), after having first been acquitted, the accused was found guilty of burglary after a second trial in a state court. The Supreme Court found this unjust. It had, however, been established in Palko v Connecticut (1937) that federal double jeopardy standards were not enforceable against the States except in cases of “acute and shocking hardship”. In that case, Palko had been found guilty of first degree murder at a second trial, having originally been convicted of second degree murder only. In Benton, the Supreme Court did not explicitly overrule Palko, but claimed instead that the precedent no longer represented the law. However, apart from vague references to “fundamental fairness” in “recent cases”, no specific case was mentioned to justify this claim:

“Our recent cases have thoroughly rejected the Palko notion that basic constitutional rights can be denied by the States as long as the totality of the circumstances does not disclose a denial of ‘fundamental fairness’.” (Benton v Maryland 1969, per Justice Marshall)

In College Savings v Florida Prepaid Education (1999), the court considered that an inconvenient precedent should no longer be taken as part of the law, as various earlier decisions could be seen “as ‘handwriting on the wall’, which even an inept cryptologist would recognize as spelling out the caption of today’s opinion” (per Justice Scalia, n.2 ).

2.3 Implicit overruling Because of the general reluctance to overrule authoritative precedents, English judges usually prefer to avoid stating explicitly that a case has been overruled; they prefer a variety of more indirect, often picturesque expressions having essentially the same function. Thus, rather than rejecting, or explicitly overruling earlier rules, for example, they commonly admit only to “departing from” them and minimise the scope of the decision by claiming to overrule only particular aspects of the inconvenient precedent:

“The proper exercise of the judicial function requires this House now to depart from Anns in so far as it affirmed a private law duty of care.” (Murphy v Brentwood 1990, per Mackay LC) [my underlining]

14 Lord Hodson asked rhetorically: “Contained, as it is, in a consolidation Act, an Act, moreover, dealing with real property, is it to be believed that by a side wind, as it were, Parliament has slipped in a provision which has revolutionised the law of contract?” Parliament finally changed the law regarding the rights of third parties in the Contracts (Third Parties) Act 1999. and:

“For all these reasons, I can see no good reason to depart from the law on this topic as established in the authorities. I would therefore hold that Khorasandjian v Bush must be overruled in so far as it holds that a mere licensee can sue in private nuisance.” (Hunter v London Docklands Developments HL 1997, per Lord Goff) [my underlining]

Even in the official Practice Statement (1966), which finally allowed the House of Lords to overrule its own previous decisions (thus rejecting London Street Tramways 1898, above), the word itself is not used. Instead of “overruling”, the House is entitled to “depart from” its own previous judgments “when it appears right to do so”, and where “too rigid adherence to precedent may lead to injustice in a particular case and also unduly restrict the proper development of the law”. [my underlining] Such indirect speech acts are not necessarily unclear, and do not necessarily depend on implicit communication. Indeed, many of the expressions used are so common as to be considered conventional. To this extent they usually function unproblematically and are commonly reported in later cases as instances of clear overruling. However, as it is often unclear which elements of the earlier precedent have been abandoned and which are still good law, this strategy frequently leads to problems later. For similar reasons, involving perhaps respect for their distinguished predecessors, English judges tend to avoid crude statements to the effect that an inconvenient precedent was simply wrong. In Jobling v Associated Dairies (1981), owing to his employer’s negligence, the plaintiff suffered a back injury which reduced his earning capacity by fifty per cent. However, before the case was heard, he was diagnosed as suffering from an unrelated disease which would soon render him totally unfit for work. The employer was reluctant to go on indefinitely paying compensation for loss of earnings, given that the earnings would soon have been reduced to zero in any case. The relevant precedent was Baker v Willoughby (1970). There, the plaintiff had suffered injury to his left leg at work, again owing to his employer’s negligence. However, in another non work-related incident, he was later shot in the same leg while witnessing an armed robbery. The leg now had to be amputated. It was held that the employer should continue to pay compensation for the injury to the leg even though it was no longer attached to his employee. The decision was justified by the fact that the armed robbery had neither reduced the victim’s loss of earnings nor his suffering. Further, and perhaps more significantly, he was unlikely to obtain any compensation from the armed robbers. Despite this authority, the judges in Jobling held the employer liable only for the loss of earnings up to the onset of the new disease. Although this was contrary to Baker, the earlier case was not expressly overruled. Indeed, in view of the particular circumstances, that decision could not be considered wrong. It was therefore said to be restricted to its own facts:

“ I can formulate no convincing juristic or logical principles supportive of the decision in this House in Baker v Willoughby 1970 and none were there propounded. [...] it may therefore be that Baker v Willoughby is acceptable on its own facts.” (Jobling v Associated Dairies 1981, per Lord Edmund-Davis)15

Junior Books v Veitchi (1982) was an equally extreme case. The defendants were subcontractors, who had been engaged to lay the floor of a new factory, and did the job negligently, so that the floor had to be re-laid. The Lords held that although there was no contract between the parties, there was a sufficiently close relationship to justify holding the defendants directly liable for the damage. This decision led to difficulties in the lower courts as it appeared to ignore the longstanding rules regarding purely economic loss. The problem was avoided by presenting the case, not as wrongly decided, but as “unique”:

“It would be intellectually dishonest in this case to attempt to distinguish Junior Books and I do not do so. I simply decline to apply it on the basis that it is unique and that it depends on the Hedley Byrne doctrine of reliance.” (Nitrigin Eareann v Inco Alloys 1992, per May J) [My underlining]

Numerous other expressions are observed. In Simaan Contracting v Pilkington Glass (1988), Dillon LJ said that “it is difficult to see that any future citation from Junior Books can ever serve any useful purpose”. In BR v Herrington, Lord Pearson said that Addie had been “rendered obsolete”. In Mannai Investment, Lord Steyn stated that Hankey “no longer represents the law”.16

15 This wording has the effect of “distinguishing the case out of existence”. 16 Dunn (2003:512) gives examples of equally picturesque expressions used in the American courts. Cases have been said to have Lord Diplock pointed out in BR v Herrington that, whether or not the overruling decision is made explicitly or goes under another name, the same result is achieved, In his view, since the Practice Statement of (1966), it was no longer necessary to use technical devices in order to avoid inconvenient precedents:

“My Lords, this House has since 1966 abandoned its former practice of adhering rigidly to the ratio decidendi of its previous decisions. There is no longer any need to discuss whether to discard the fiction of a so-called ‘licence’ to enter granted by the occupier of land to the person who suffers personal injury on it, should be characterised as over-ruling Addie’s case or as doing no more than explaining its reasoning in terms which are in harmony with the general development of legal concepts since 1929 [...].” (BR v Herrington 1972, per Lord Diplock)

It should be noted that certain judges are more willing than others to state clearly that they are changing the law. In Candler v Crane Christmas (1951), in which he was in the minority, Denning LJ described those judges who were reluctant to overrule as “timorous”:

“On the one side there were the timorous souls who were fearful of allowing a new cause of action. On the other side there were the bold spirits who were ready to allow it if justice so required. It was fortunate for the common law that the progressive view prevailed.” Candler v Crane Christmas 1951, per Denning LJ)17

In Conway v Rimmer CA (1967), Denning LJ was again in the minority in proposing to overruling Duncan v Cammell Laird (1942), concerning crown privilege. He claimed that the earlier case was based on a misunderstanding of the applicable (Scottish) law, that it had been rejected in the commonwealth countries, and that following the Practice Statement (1966), the doctrine of precedent had been relaxed so that the court was no longer bound:

“My brethren today feel that we are still bound by the observations of the House of Lords in Duncan v Cammell Laird [...]. I do not agree. The doctrine of precedent has been transformed by the recent statement of Lord Gardiner LC. This is the very case in which to throw off the fetters.” (Conway v Rimmer CA 1967, per Lord Denning) 18

In Murphy v Brentwood (1990), Lord Keith was exceptionally vigorous in overruling Anns v Merton LDC (1978):

“[Anns] did not proceed on any basis of principle at all but constituted a remarkable example of judicial legislation [...] There can be no doubt that to depart from the decision would re-establish a degree of certainty in this field of law which it has done a remarkable amount to upset.” (Murphy v Brentwood 1990, per Lord Keith)

Yet even in such a clear statement as this, Lord Keith expressed himself in the conditional. The reason for this may be purely practical. In the English system, judgments are given individually. As no single judge can overrule without the agreement of the majority, their decisions tend to be phrased as suggestions. In the American courts, on the other hand, the majority opinion is normally agreed in advance. This may be a partial explanation of the fact that American judges are often more willing to state explicitly that a precedent has been overruled, as is seen for example in Lawrence (2003):

“Bowers was not correct when it was decided, and it is not correct today. It ought not to remain binding precedent. Bowers v Hardwick should be and now is overruled.” (Lawrence v Texas 2003, per Justice Kennedy)

The official summary of this case (the syllabus) is even more explicit, to the extent of using the adverb ‘hereby’ as an explicit performative marker (“Bowers was not correct when it was decided, is not correct today, and is hereby overruled”). In view of the intense publicity aroused by Lawrence, it may be that the judges felt obliged to abandon their usual tactic of evasiveness. However, the explicit language was still reinforced by rhetorical devices. It was said notably that “Bowers had caused uncertainty” and that it had already been weakened by subsequent case law. The dissenting opinion of Justice Stevens in Bowers was

“outlived their usefulness”. Another case was overruled on the basis that it had earlier been “greeted with less than warmth”. 17 Asquith LJ replied to this accusation in the course of his judgment in the same case: “If this relegates me to the company of ‘timorous souls’, I must face that consequence with such fortitude as I can command.” 18 It was stated explicitly in the Practice Statement (1966) that it was “not intended to affect the use of precedent elsewhere than in this House”. The majority on the Court of Appeal therefore correctly refused to question the “conclusive authority” of Duncan, which was, however, “not followed” when the case reached the Lords the following year. approved. Furthermore, even in American jurisprudence, there are many equally significant cases in which the overruling decision remains implicit.

2.4 Significance It is natural to assume, as a preliminary hypothesis, that the more significant the decision, the greater the need for clarity. However, the opposite is usually observed. Indeed, in the most important decisions, where a significant line of authority is rejected, the tendency is not to match the force of the speech act to the significance of the legal declaration, but rather to rely on implicit expressions. Lochner v NY, decided by the USSC in 1905, was a particularly important precedent, repeatedly confirmed during a period of great social upheaval. The case concerned a New York statute limiting working hours to 10 hours a day or 60 hours a week. It was held that such regulations violated the constitutional guarantee of freedom of contract. This view of the Constitution was in direct contradiction to the “new deal” later proposed by President Roosevelt. Indeed, it appeared that Roosevelt’s new policy would inevitably be struck down.19 Lochner was followed in numerous cases, including Adkins v Children’s Hospital (1923), in which a minimum wage for women was held to be unconstitutional. However, when the principle was finally rejected in West Coast Hotel v Parrish (1937), Hughes CJ claimed, against all reason, that the law remained unchanged. Far from being explicitly overruled, Lochner was not even mentioned in the opinions, but simply cited in a footnote. Instead, Adkins was rejected as a “departure” from the true principles of the law:

“We think that the views thus expressed are sound, and that the decision in the Adkins case was a departure from the true application of the principles governing the regulation by the State of the relation of employer and employed. [...] Our conclusion is that the case of Adkins v Children’s Hospital, supra, should be, and it is, overruled.“ (West Coast Hotel v Parrish 1937, per Hughes CJ)

Recognising the significance of the decision, the majority judgment in Casey (1992), pointed out ironically that West Coast Hotel had “signalled the demise of Lochner by overruling Adkins”. Because of its profound social repercussions, Brown v Board of Education of Topeka (1954) was perhaps the most important case decided by the US Supreme Court during the 20th century. In a unanimous opinion, the court held that racial segregation was unconstitutional. This was in direct contradiction to the “separate but equal” doctrine, accepted as constitutional in Plessy v Ferguson as long ago as 1896. Through his use of indirect language, Warren CJ seems to have minimised the scope of the overruling decision, perhaps because he was concerned to ensure unanimity:

“We conclude that, in the field of public education, the doctrine of “separate but equal” has no place. [...] Any language in Plessy v Ferguson contrary to this finding is rejected.” (Brown v Board of Education of Topeka 1954, per Warren CJ)

Yet the entire purpose of the finding was to reject the “separate but equal” doctrine as defined in Plessy. The new principle was not limited to education but was immediately applied in all fields. It is still considered as an emblematic case for the civil rights movement.

2.5 Uncertainty The reliance on implicit language in overruling decisions frequently leads to disagreement as to whether a particular precedent is still authoritative. For this reason, legal argument is often concerned with the question of whether and to what extent a particular case has been overruled. Examples are numerous, both in the US and in English law. Before Lochner was (surreptitiously) overruled in the West Coast Hotel case, it had been frequently distinguished and avoided, so that its status as authority had become unclear. Many thought it was no longer part of the law. Hence the evident surprise of Taft CJ, dissenting, when it was followed by the majority in Adkins:

“I have always supposed that the Lochner Case was thus overruled sub silentio. Yet the opinion of the court herein in support of its conclusion quotes from the opinion in the Lochner Case as one which has been sometimes distinguished but never overruled.” (Adkins v Children’s Hospital (1923) per Taft CJ, dissenting)

Similarly for Justice Holmes in the same case:

19 In order to ensure a majority in favour of his policy, President Roosevelt notoriously threatened to “pack the court” by naming extra judges to the Supreme Court. “It will need more than the Nineteenth Amendment to convince me that there are no differences between men and women, or that legislation cannot take those differences into account. I should not hesitate to take them into account if I thought it necessary to sustain this Act. [...] I had supposed that it was not necessary, and that Lochner 0v New York would be allowed a deserved repose.” (Adkins v Children’s Hospital (1923) per Justice Holmes, dissenting)

The status of the rule decided in Gitlow v NY (1925) was considered during the McCarthyist period, in Dennis v US (1951). The accused were found guilty of organising an American communist party and of advocating the overthrow of the government. Their defence was based on the right to free speech contained in the 1st Amendment to the Constitution. Gitlow had established that exceptions to the 1st amendment regarding free speech could only be made where the incriminated words represented a “clear and present danger” of substantive harm.20 In Dennis, the Chief Justice considered the status of the Gitlow rule without coming to a clear conclusion as to whether it should still be seen as a valid precedent:

“Although no case subsequent to Whitney and Gitlow has expressly overruled the majority opinions in those cases, there is little doubt that subsequent opinions have inclined toward the Holmes-Brandeis rationale. [...] It has not been explicitly overruled. But it would be disingenuous to deny that the dissent in Gitlow has been treated with the respect usually accorded to a decision.” (Dennis v United States (1951) per Vinson CJ)

In the event, the point was purely academic, given that the majority in Dennis considered that the advocation of such policies by the American Communist Party did in any case represent a “clear and present danger” for the State. Perhaps more significantly, although most people considered that the “separate but equal” doctrine established in Plessy v Ferguson 1896 had been conclusively rejected in Brown v Board of Education 1954, this opinion was not universally shared. While ice Black, for example, held in Oregon v Mitchell(1970) that Plessyhad been "overruled", Justice Douglas continued to cite the earlier case. In Bryson v US(1969), discussing the status of a related case, he said:

"Whatever may be said technically about any remaining vitality of the Douds case, it obviously belongs to a discredited regime, though, like Plessy v Ferguson, 163 U.S. 537, it has never been officially overruled." (Bryson v US (1969) per Justice Douglas)

Similar difficulties are common in English law. Donoghue v Stevenson (1932) is now celebrated as the source of the modern English law of negligence. Rose Donoghue suffered distress after finding the remains of a decomposed snail in the ginger beer she had been served in a local tea shop. Legal debate centred on the status of two earlier cases. In Winterbottom v Wright (1842), a coach driver had been injured because of defects in the construction of the coach. It was held that because the fault lay with the company which had supplied the coach, no action lay against his employer. Further, as the driver had no contract with the supplier of the coach, he had no right to compensation from that company. On this point, Lord Abinger confirmed that there could be no liability to third parties, justifying his conclusion with a classic, slippery slope argument:

“[I]f the plaintiff can sue, every passenger, or even any person passing along the road, who was injured by the upsetting of the coach, might bring a similar action. Unless we confine the operation of such contracts as this to the parties who entered into them, the most absurd and outrageous consequences, to which I can see no limit, would ensue.” (Winterbottom v Wright (1842) per Lord Abinger)

This unashamedly rhetorical argument had been ignored by the House of Lords in George v Skivington (1869), in which their Lordships had come to a different conclusion. George involved the sale by a chemist of some hair wash which was not used by the purchaser himself but by his wife, who was harmed because it had been negligently compounded. The question, as stated by the Chief Baron, was:

“ [w]hether the defendant, a chemist, compounding the article sold for a particular purpose and knowing of the purpose for which it was bought, is liable in an action on the case for unskilfulness and negligence in the manufacturer of it whereby the person who used it was injured.” (George v Skivington (1869) per Baron Alderson)21

20 This phrase was originally introduced by Justice Holmes in Schenck v US (1919). 21 “Action on the case” (often contrasted with “trespass on the case”) was a contemporary procedure used to deal with cases involving The Lords found in George that, on the given facts, the chemist should be liable for the damage to the wife, in spite of the fact that she was a third party. In view of the contradictory conclusions, Winterbottom and George could not both represent the correct state of the law. If Winterbottom was still good law, then the Donoghue court could not impose a duty of care on the owner of the tea shop, who had served the ginger beer as supplied to him by the manufacturers. However, if George was a valid precedent, then the court could impose liability for negligence, independently of contractual duties. Lord Buckmaster preferred not to treat George as a precedent, considering it mistaken in law, as it had failed to follow the rule decided in Winterbottom. He pointed out that it had been rejected in so many earlier cases that it could safely be ignored:

“I do not propose to follow the fortunes of George v Skivington; few cases can have lived so dangerously and lived so long.” (Donoghue v Stevenson (1932) per Lord Buckmaster, dissenting)

In his view, the court should therefore refuse to countenance any suggestion of liability to third parties. However, the majority preferred to “distinguish” Winterbottom and follow George, noting that the latter case had never been expressly overruled:

“It is true that George v Skivington has been the subject of some criticism and was said by Hamilton J [...] to have been in later cases as nearly disaffirmed as is possible without being expressly overruled. I am not sure that it has been so severely handled as that. At any rate I do not think that it deserved to be and certainly, so far as I am aware, it has never been disapproved in this House.” (Donoghue v Stevenson (1932) per Lord Macmillan)

This paved the way for Lord Atkin’s celebrated “neighbour principle”, according to which the courts could impose a duty of care on “persons who are so closely and directly affected by my act that I ought reasonably to have them in contemplation as being so affected when I am directing my mind to the acts or omissions which are called in question”. It is worth mentioning in this context that a more positive view was later taken of George, said in retrospect to have been “vindicated” in Donoghue:

“George v Skivington, a decision “battered but unbowed”, was in the end vindicated by the House of Lords in Donoghue’s case.” (Candler v Crane Christmas (1951) per Asquith LJ)

Lord Macmillan’s original judgment in Donoghue (above) seems somewhat less clear-cut than as reported by Lord Asquith in Candler. Similar problems are observed in more recent cases. The status of Foakes (1884) (confirming Pinnel (1602) was recently the subject of some doubt. In Williams v Roffey Bros (1990), a pre-existing obligation was accepted as valid consideration allowing the enforcement of a promise of further payment. As Foakes had been decided in the Lords, it could not be overruled by the Court of Appeal, and it was not mentioned in Williams. However, Purchas LJ, albeit reluctantly, seemed to reject Stilk v Myrick (1809), a case which had been decided on similar principles:

“ Counsel for the plaintiff was bold enough to submit that Stilk v Myrick, albeit a judgement of no less a judge than Lord Ellenborough CJ, was a judgment at nisi prius and that therefore this court was not bound by it. I feel I must say at once that, for my part, I would not be prepared to overrule a case of such distinction involving a judgment of a judge of such distinction except on the strongest possible grounds since its forms a cornerstone of the law of contract, which as has been observed over the years, is still recognised in principle in recent authority.” (Williams v Roffey Bros (1990) per Purchas LJ)22

Although Stilk was not explicitly overruled, it was thus decided in Williams, that the performance of an existing obligation could nevertheless be considered as valid consideration, provided that it amounted to a practical benefit to the other party.23 Following that decision it seemed that the principle underlying Foakes negligence. 22 The court of Nisi Prius no longer exists. 23 Glidewell LJ considered that this did not contravene the principle relied on in Stylk v Myrick, but only refined and limited its application, while leaving the principle unscathed: “It is not in my view surprising that a principle enunciated in relation to the rigours of seafaring life during the Napoleonic wars should be subjected during the succeeding 180 years to a process of refinement and limitation in its application in the present day.” may have been rejected, so that the authority of that case was no longer clear. The status of that principle was not decided until Selectmove (1995) (see below).

3. The legal effects of overruling

Common law judges are not simply reluctant to overrule explicitly; they often go to the extent of denying, contrary to the evidence, that their overruling decisions imply a change in the law at all. They claim instead that these decisions are mere declarations of the true state of the law, in the face of misapprehensions derived from mistaken decisions in earlier cases.24 This approach is a modern version of the old “declaratory theory” of law, according to which, in the words of Blackstone (1765, vol 1: 69-70), judicial power is “not delegated to pronounce a new law, but to maintain and expound the old one.[...] Thus the judges do not “pretend to make a new law, but to vindicate the old one from misrepresentation [...] For if it be found that the former decision is manifestly absurd or unjust, it is declared, not that such a sentence was bad law, but that it was not law”. In defence of this theory, it is often said that the decisions of courts are not strictly speaking part of the law; they are no more than evidence of what the common law is. As was said by Lord Mansfield in R v Bembridge (1783): “The law does not consist of particular cases but of general principles, which are illustrated and explained by these cases.” On this view, individual decisions do not disturb the eternal principles on which the law is based.25 This theory of law would allow the legal rules to be modified according to circumstance, without affecting the unchanging principles of justice. Such general principles would no doubt be useful as an aid to adjudication in cases of ambiguity or indeterminacy, as they may serve as a guide in determining the objectives and purposes of particular rules. However, problems arise when attempts are made to define precisely the difference between a principle and a rule. If principles can be stated, then they could presumably be reinterpreted in particular cases, just like rules. If, on the other hand, they cannot be stated, then it is difficult to see how they could condition judicial decision-making. The declaratory approach is illustrated in practice by the decision in R v R (1992), in which Lane CJ adopted his “radical solution” to the old rule of immunity for marital rape, according to which “the husband cannot be guilty of a rape committed by himself upon his lawful wife, for by their mutual matrimonial consent and contract the wife hath given up herself in this kind unto her husband which she cannot retract”.26 Although this rule had been consistently applied since the 17th century, Lord Lane declared that it had never truly represented the law. He could therefore make the ingenuous claim that he was not changing the law but simply correcting a longstanding misapprehension:

“This is not the creation of a new offence, it is the removal of a common law fiction which has become anachronistic and offensive and we consider that it is our duty having reached that conclusion to act upon it.” R v R (1992) per Lane CJ)

3.1 The fairy tale In the declaratory theory, the common law rules seem to be presented as having some kind of abstract, metaphysical existence. It may be thought that this is simply a theoretical device, designed to preserve the illusion of strict objectivity. However, the point is not merely a matter of philosophical speculation. On the contrary, it can lead to serious practical problems. Indeed, the claim that law has always been as it is now said to be implies retrospectivity, as is seen from the following statement by Lord Reid:

“We cannot say that the law was one thing yesterday but is to be something different tomorrow. If we decide that [the existing rule] is wrong we must decide that it always has been wrong [...]”.(West Midland Baptist (Trust) Association v Birmingham Corp HL (1970) per Lord Reid)

This problem was implicitly recognised in the Lords’ Practice Statement (1966), which warned against “the

24 Like Kripke’s bizarre sceptic (1982: 8), common law judges thus claim to be following the rule they have always followed, even while obtaining an expected result. Even on the individual level, the content of the rule thus appears radically indeterminate. On this point, see Charnock (2006). 25 See also Rust v Cooper (1777): “the law does not consist in particular cases; but in general principles, which run through the cases, and govern the decision of them.” Readers of Dworkin will recognise a clear case of what is known in Parisian literary circles (in particular, the Oulipo group) as “plagiarism by anticipation”. 26 This rule was stated by Sir Matthew Hale (1609-1675) in his History of the Pleas of the Crown (1736), vol. 1, ch. 58, p. 629. danger of disturbing retrospectively the basis on which contracts, settlements of property and fiscal arrangements have been entered into and also the especial need for certainty as to the criminal law”. There is a sense in which new interpretations of the law always have retrospective effect, in that the new decision can only be taken when the case has reached the higher courts, long after the facts have occurred. Indeed, this was the ground of an appeal to the European Court after the conviction for marital rape in R v R. The accused claimed that at the time of the offence, Hale’s rule was still part of English law, and that he was now accused of an offence which did not then exist. The European Court avoided the problem, confirming the conviction on the grounds that the change in the law was “reasonably foreseeable”:

“The decisions of the Court of Appeal and then the House of Lords in R v R did no more than continue a perceptible line of case law development, dismantling the immunity of a husband. This evolution had reached a stage where judicial recognition of the absence of immunity had become a reasonably foreseeable development of the law.” (SW v United Kingdom, ECHR, 1996)

The retrospectivity of common law decisions does not normally entail new hearings of earlier cases, as the rule of res judicata applies to achieve finality. However, there are situations in which previous decisions have ongoing effect. In Kleinwort Benson v Lincoln CC (1988), for instance, it was decided, contrary to what had previously been supposed, that interest rate swaps between local authorities were illegal. In consequence, on the supposition that the law had always been as it was now declared to be, certain payments could now be said to have been made under a mistake of law. According to an old common law rule, such payments should be recoverable. Nevertheless, as the parties in Kleinwort Benson had correctly followed the law as it stood at the time the payments were made, they could not be considered as having been mistaken. The Court was reluctant to apply the new decision retrospectively. Lord Lloyd described the declaratory theory as a fairy tale, and considered that it should be abandoned:

“It follows that in such a case the House of Lords is doing more than develop the law. It is changing the law, as common sense suggests [...]. If this view of what happens is inconsistent with the declaratory theory of the Court’s function, then it is time we said so. It always was a fairy tale.” (Kleinwort Benson v Lincoln City Council, HL 1998, per Lord Lloyd)

Lord Browne-Wilkinson also rejected the declaratory theory:

“ The theoretical position has been that judges do not make or change law: they discover and declare the law, which is throughout the same. According to this theory, when an earlier decision is overruled the law is not changed: its true nature is disclosed, having existed in that form all along. This theoretical position is, as Lord Reid said, a fairy tale in which no-one any longer believes.” (Kleinwort Benson v Lincoln City Council, HL 1998, per Lord Browne-Wilkinson).27

However, he pointed out that, although the declaratory theory had been rejected, “its progeny - the retrospective effect of a change made by judicial decision - remains”. A related problem occurred more recently in the area of penal law in R v Governor of HM Prison Brockhill (Ex Parte Evans) (2000). This case concerned the calculation of the time to be served in prison where the prisoner is sentenced on several counts, the sentences to run concurrently. There was disagreement about the correct interpretation of the Criminal Justice Act (1991), in particular on the question of whether the time spent on remand should be subtracted from each of the sentences, or whether it should be subtracted only once. The Divisional Court concluded in 1996 that, according to the new construction of the relevant provisions, a convicted prisoner, Michelle Evans, had been kept in prison for 59 days too long. Her legal representatives accused the Prison Governor of wrongful detention. It was said, in answer to this charge, that he could not have done otherwise, as, if he had not detained the prisoner, he would have been acting in defiance of the law existing at the time. However, this was not accepted as a valid defence. Although he considered the result “highly artificial”, the Master of the Rolls felt obliged to accept the declaratory theory:

“Until the approach to the doctrine of precedent is changed, the practical consequence is that once the later decision has been given there is no right to rely on the earlier decision as correctly representing the law. [...] This is an undoubtedly highly artificial result. It involves a fairytale. However [...] it is not open to this court to

27 He was referring to an academic article by Lord Reid (1972). abandon the fairytale.” R v Governor of Her Majesty’s Prison Brockhill Ex Parte Evans CA 2000, per Lord Woolf MR)

The Court of Appeal therefore found in favour of the prisoner.28 Similar views are proposed in American cases. In spite of the Judiciary Act (1789) ch. 20, Justice Story considered, in Swift v Tyson (1842), that: “In the ordinary use of language, it will hardly be contended, that the decisions of courts constitute laws. They are, at most, only evidence of what the laws are, and are not, of themselves, laws. They are often re-examined, reversed and qualified by the courts themselves, whenever they are found to be either defective, or ill-founded, or otherwise incorrect.” Similarly in West Coast v Parrish (1937) (above), in which Hughes CJ described the Adkin’s ruling as a departure from the true application of the principles of the law.

3.2 Prospective overruling In order to avoid problems associated with the retrospective application of common law decisions, the American courts have developed a system of prospective overruling, in which it is clearly stated that any changes in the law will apply to future cases only. In Great Northern Ry v Sunburst Oil (1932), it appeared that a Railroad Commission had approved a schedule of payments, then later, after the payments had been made, held that the rates it had itself approved were in fact excessive and unreasonable. The respondent sought to recover what were now seen as overcharges. Justice Cardozo stated that the Supreme Court had no objection in principle if the individual states chose to abandon the metaphysical view of the common law:

“[Montana] may say that decisions of its highest court, though later overruled, are law none the less for intermediate transactions [...] On the other hand, it may hold to the ancient dogma that the law declared by its courts had a Platonic or ideal existence before the act of declaration, in which event the discredited declaration will be viewed as if it had never been, and the reconsidered declaration as law from the beginning.” (Great Northern Ry v Sunburst Oil 1932, per Cardozo J)

In Griffin v Illinois (1956), Griffin was unable to pay for a certified report of his trial, which he needed in order to appeal against his conviction. The State refused to supply the report as the current policy was to supply free reports only to “indigent defendants sentenced to death” (Griffin had merely been sentenced to a term of imprisonment). The Supreme Court considered that the State should be obliged to supply the report, but was reluctant to allow all previously convicted prisoners to avail themselves of this right in order to lodge new appeals. In deciding that the new decision should apply prospectively only, Justice Frankfurter described the declarative theory of law as a fiction:

“We should not indulge in the fiction that the law now announced has always been the law and, therefore, that those who did not avail themselves of it waived their rights. It is much more conducive to law’s self-respect to recognize candidly the considerations that give prospective content to a new pronouncement of law.” (Griffin v Illinois 1956, per Justice Frankfurter)

In Beam Distilling v Georgia (1991), a plaintiff sought a refund of taxes on imported liquor paid under a Georgia law, after a similar Hawaii law was declared to be in violation of the Commerce Clause. The court followed the new precedent to the extent of declaring the Georgia statute unconstitutional, but nevertheless refused to apply the ruling retroactively, Justice Scalia attempting to defend a modified version of the declaratory theory. While admitting that judges do occasionally appear to make law, he denied that this could amount to actual legislation:

“ I am not so naive (nor do I think our forebearers [sic] were) as to be unaware that judges in a real sense “make” law. But they make it as judges make it, which is to say as though they were “finding” it - discerning what the law is, rather than decreeing what it is today changed to, or what it will tomorrow be. [...] I would find both ‘selective prospectivity’ and ‘pure prospectivity’ beyond our power.” (Beam Distilling 1991, per Justice Scalia)29

28 In view of recent divorce cases in which the ex-wives of rich men have been granted generous settlements, far in excess of their financial needs (Miller v Miller / McFarlane v McFarlane HL, 2006), it is likely that similar problems of retrospectivity will arise when previously divorced couples come to apply for re-evaluation. 29 This corresponds to the “interstitial” legislation, recognised (unhesitatingly) by Justice Holmes (1881: 32): “I recognise without hesitation that judges do and must legislate, but they can do so only interstitially.” See also Holmes (1881: 269) where he recognises that judges do make “some profound interstitial change in the very tissue of the law”. Prospective decision-making is thus useful, insofar as it can be used to avoid injustice, but it is considered unconstitutional by many commentators, precisely because the judiciary thereby assumes a quasi-legislative role. This was the basis of criticism in Harper v VA Taxation (1993). After revision of a statute, tax-payers had sought a refund of taxes on retirement benefits which now appeared to have been wrongly assessed. Justice Scalia stated that:

“Prospective decision making is the handmaid of judicial activism, and the born enemy of stare decisis. It was formulated in the heyday of legal realism and promoted as a “technique of judicial lawmaking” in general, and more specifically as a means of making it easier to overrule prior precedent.” (Harper v VA Taxation 1993, per Justice Scalia)

For similar reasons, prospective overruling has not been adopted in English law. In Kleinwort Benson (1999), Lord Goff said that the system of prospective overruling “has no place in our legal system”. This view was recently confirmed in the House of Lords in National Westminster Bank v Spectrum (2005). Lord Nicholls there distinguished two kinds of overruling. In some cases, it is simply a matter of correcting an earlier mistake; in others a new decision is taken in response to social change:

“So, it is said, when your Lordships’ House rules that a previous decision on the interpretation of a statutory provision was wrong, there is no question of the House changing the law. The House is doing no more than correct an error of interpretation [...] These cases are to be contrasted with those where the later decision represents a response to changes in social conditions and expectations. Then, on any view, the declaratory approach is inapt. In this context this approach has long been discarded. It is at odds with reality.” (National Westminster Bank v Spectrum 2005, per Lord Nichols)

However, even those cases where the declaratory theory was clearly unrealistic, he was reluctant to allow prospective overruling:

“[Prospective overruling] would amount to the judicial usurpation of the legislative function. Power to make rulings having only prospective effect, it is said, is not inherent in the judicial role. A ruling having only prospective effect cannot be characterised as merely a less extensive form of overruling than overruling with both retrospective and prospective effect. Prospective overruling robs a ruling of its essential authenticity as a judicial act.” (National Westminster Bank v Spectrum 2005, per Lord Nichols)

He rejects prospective overruling, precisely because it means explicitly recognising the power of the judiciary to modify the law. However, the rejection of this system in favour of the declaratory theory cannot have the effect of limiting the power of the judges. On the contrary, the practical consequence is to extend judicial power, so as to include retrospective application. In Beam Distilling, Justice O’Connor confirmed that the judges have the power to say what the law is, and drew the inevitable conclusion:

“ I reiterate, however, that precisely because this Court has “the power ‘to say what the law is’ when the Court changes its mind, the law changes with it.” (Beam Distilling 1991, per Justice O’Connor, dissenting)30

If this is the correct view, then the illocutionary force of judicial declarations appears greater than is generally supposed.

4. The linguistics of overruling

Once it is admitted that overruling judgments do have the practical effect of changing the law, it appears natural to analyse the relevant decisions as speech acts, having normative effect. However, the performative analysis of legal judgments raises several problems for speech act theory. First, not all judgments are intended to change the law. For theoretical reasons, it would be unfortunate if overruling decisions were seen as performative, while other judgments were analysed differently. However, where the law remains unchanged, it is unclear what the performative effect may be. Secondly, although a declaration is necessary before the overruling decision can take effect, the utterance may not in itself be sufficient. Lastly, although overruling

30 She was referring to Marbury v Madison (1803), in which Marshall CJ declared that “It is emphatically the province and duty of the judicial department to say what the law is.” decisions appear to function as legal performatives, it must be admitted that in ordinary discourse performative utterances do not usually have normative effect. The relations between performativity and normativity must therefore be examined.

4.1 Illocutionary effects Austin31 points out (1962: 88) that “if you are a judge and say ‘I hold that ...’, then to say you hold is to hold; with less official persons it is not so clearly so”. What the judge says must simply be accepted as the legal decision. However, this question is complicated by the fact that, as with a referee’s decisions in a football match, debate and discussion is still possible as to whether a particular judgment was justified on the available evidence. Austin analyses official judicial declarations as a particular type of performative, which he calls “verdictives” (1962: 88). These are distinguished (1962: 153) from exercitives, which are legislative acts. An exercitive is (1962: 155) a decision “that something is to be so, rather than a judgment that it is so.” Yet Austin admits (1962: 154) that “a judge’s ruling makes law, just as a jury’s finding makes a convicted felon”. Further (1962:154): “Verdictives have an effect, in the law, on ourselves and on others.” Thus (1962: 153) “some judicial acts really are exercitives”. If legal judgments can affect the law, then the distinction between verdictives and exercitives needs to be refined.32 Austin (1962: 61-62) suggests that it may be a criterion of true performative utterances that they should be “reducible, or expandable, or analysable into a form, or reproducible in a form, with a verb in the first person singular present indicative active”. He considers this necessary because, where the performative act cannot be expressed in such an explicit form, “it will regularly be possible to take it in a non-performative way” (id). Yet overruling statements are not normally given in explicit form. Instead, as has been seen, numerous indirect and unconventional expressions are used. There is no purely linguistic criterion to distinguish successful from failed overruling decisions. This is one reason for the frequent confusion and disagreement as to their legal effect. Note, however, that even explicit performatives can be (mis)understood constatively. Further, even clearly performative expressions may be interpreted in different ways in particular contexts. What is presented literally as a promise may in fact be intended as a threat (“Is that a threat or a promise?”). In some contexts, literal warnings may also be taken as threats, as in “I’m warning you”. And it is sometimes unclear whether requests or orders should be understood as warnings (“Make my day, punk!”). In addition, although overruling decisions may be categorised as speech acts, it will never be possible to analyse any judgment as performative in its entirety. Only certain passages, constituting the ratio decidendi, will be binding in later cases. Other passages may consist of simple descriptive statements of the facts of the case, or passing remarks which are better analysed as obiter dicta. As Austin (1962: 8) points out, the same is true of legal instruments like contracts or wills, which naturally include, in addition to their operative clauses, constative recitals of the circumstances in which the transaction is to be effected. Only the “operative” passages have performative effect. In the same way, statutes may include preambles reciting the aims and objectives of the legislation. Although these may occasionally help to resolve possible ambiguities in the text, they do not have legal effect in themselves.33 Moreover, in the great majority of (easy) cases, judges make no attempt to overrule, but simply apply the law as it stands. In these cases, their judgments are not intended to have performative effect. They do not modify, but simply restate the existing law. Yet it would be unacceptable to make the linguistic status of the judgment depend on the result of the case. As Lord Hope pointed out in Kleinwort Benson (1998), there is a sense in which simple declarations can create law:

“But it would be equally wrong to say that the judges never declare the law. It may simply be that there was a gap which needed to be filled, or that there was a defect in thinking which needed to be revealed so that a point could be clarified. And to overturn an established line of authority is one thing. It is quite another where there was no previous decision on a point which no-one had sought to bring before the court previously.” (Kleinwort Benson (1998) per Lord Hope)

31 Later scholars have proposed numerous modifications to Austin’s (1962) presentation of performativity (see below). However, the original model remains a useful basis for discussion. 32 This problem is well-known to lawyers. See, e.g. Kelsen (1934: 70): “The contrast between making or creating the law and carrying out or applying the law does not by any means have the absolute character accorded to it by traditional legal theory”. 33 Exhortative prologues formed an important part of classical Greek law, but these were largely avoided in Roman law, reduced as far as possible to the operative passages (hence references to lex ipsa or ipsa verba). Austin insisted (1962: 134): “Surely to state is every bit as much to perform an illocutionary act as, say, to warn or to pronounce.” He gives (1962: 7) the example of declaring war, which, far from being a constative declaration, may have far-reaching political effects. In the same way, even where the content is purely factual, a legal declaration remains a speech act, so that a statement of the law may have legal consequences. Thus, although it may be less newsworthy, the confirmation that the law remains as it was previously assumed to be may, in certain circumstances, be more significant than an actual overruling decision. In Foakes (1884), for example, it was precisely the failure to overrule which was surprising. The same is true of Selectmove (1995) in which it was declared that the Pinnel principle, as restated in Foakes, remained valid notwithstanding the decision in Williams v Roffey (1990). A company was faced with substantial tax liabilities and allowed time to pay. Notwithstanding this agreement, the authorities then suddenly changed their mind and demanded immediate payment of the remainder of the debt, thus putting the company into liquidation. The legal effect of the promise to allow time to pay was considered. Gibson LJ, while noting the decision in Williams, did not feel entitled to overrule Foakes:

“Foakes v Beer was not even referred to in Williams v Roffey Bros, and it is in my judgment impossible, consistently with the doctrine of precedent, for this court to extend the principle in Williams to any circumstances governed by the principle in Foakes. If that extension is to be made, it must be made by the House of Lords or, perhaps even more appropriately, by Parliament after consideration by the Law Commission.” (Re Selectmove 1995, per Gibson LJ)

The judicial confirmation that the law remained unchanged thus acquired illocutionary significance. It is worth noting a parallel here with ordinary language speech acts, where the failure to give an expected order to fire, or the refusal to declare loyalty to a political leader, or to pronounce the words “I will” at the appropriate time during a marriage ceremony, may have greater social consequences than would have been the case if the conventional speech acts had been successfully performed. It should also be mentioned that legal speech acts, contrary to ordinary language speech acts, may have effects which go beyond the discursive intentions of the participants. A promise may be taken as evidence of intention to create a contract by which the speaker may be legally bound. A threat may be taken by a court as evidence of harassment, while a warning may have the effect of removing the speaker’s liability in negligence. In another context, a warning to an employee may have legal significance if the employer is later accused of unfair dismissal.

4.2 Perlocutionary effects In the classic examples of conventional performatives, the utterance is usually enough for the successful performance of the act. Given the appropriate felicity conditions, to say that you promise, order or apologise is to promise, order or apologise (or, for a judge, to hold). However, perlocutionary acts require more. Austin (1962: 109) gives the examples of ‘convincing’, ‘alarming’ or ‘persuading’, which are brought about “by” saying, rather than “in” saying. The success or otherwise of these acts thus depends not just on the words used, but also, crucially, on the consequential behaviour of the interlocutor. For similar reasons, Austin (1962: 122) points out that “a judge should be able to decide, by hearing what was said, what locutionary and illocutionary acts were performed but not what perlocutionary effects were achieved”. In the same way, an overruling judgment requires more than the simple communication of the intention on the part of the judge. It is necessary for him not just to perform the relevant speech act, but also to convince a majority of the judges sitting with him of the acceptability of the decision. Further, if the decision is to remain valid, it must later be confirmed on appeal. The judicial confirmation is not automatic, and cannot be analysed as a causal consequence of the utterance.34 Although perlocutionary acts are brought about by speech, they cannot normally be expressed in generally accepted, conventional terms. In contrast to explicit performative expressions (“I warn you that...), which often correspond to the names of the relevant illocutionary acts (warning), there is no performative verb directly associated with perlocutionary acts. It is not possible, for example, to persuade someone simply by stating explicitly that this is the intention, for example with “*I hereby convince you”. For similar reasons, it is impossible to convince someone that a particular argument is false, simply by stating “*I refute that argument”.35 With verbs like ‘insinuating’, which imply some form of implicit communication, the use of explicit performative expressions would lead to contradiction.

34 Davis (1979: 226-267) suggests a causal analysis of perlocutionary acts. This cannot apply in the legal context. 35 This supposedly performative expression is nevertheless frequently observed in English parliamentary debates. Similarly, as has been seen, overruling decisions are not normally given in explicit form. However, although it is rarely used, the explicit performative verb ‘to overrule’ does exist, and is occasionally observed, especially in the United States, where the majority opinion is given in the name of the court. Conventional phrases having the same function are often used, both in England and in America, and these may be assimilated to performative verbs. In ordinary language, what is said cannot be unsaid. Neither insults or promises can be easily withdrawn, though the speaker may later be forgiven, or released from his promise. In contrast, even where an interlocutor has been convinced, persuaded, surprised or deterred by particular language or behaviour, he may later change his mind. The apparently successful perlocutionary act then becomes inoperative. Similarly with overruling decisions, which may be reversed or superseded at a later date, if and when the legal institution comes to take a different view. Conversely, the dissent in an earlier case may be approved, as was the case in Candler v Crane Christmas (1951) or Bowers v Hardwick (above), so that the judgment acquires overruling status post facto. Unlike true performatives, the validity of legal judgments is thus seen to change over time. From this point of view, overruling may again be classed as perlocutionary, rather than illocutionary. Austin accepts (1962: 118) that some of his examples of perlocutionary acts, for example ‘alarming’, ‘surprising’ or ‘deterring’, can be achieved by other means, perhaps without the need for speech at all. Perlocutionary acts are therefore ruled out (Austin 1962: 110) as “irrelevant to the sense in which an utterance is said to be performative”. However, this is not the case with overruling, which cannot be achieved without a formal, judicial statement. Overruling declarations may thus be said to combine features of illocutionary and perlocutionary acts. The fact that they imply change in the law also suggests a close relationship between performativity and normativity.

4.3 Normative effects In the linguistic field, “norms” are not created by outside forces in the form of legislation; they depend rather on what is considered “normal”.36 Grammatical rules, for example, may be seen as social facts which only exist to the extent that they are recognised and followed by the members of the linguistic community. Linguistic norms are thus created by consensus regarding the linguistic behaviour and understanding of the collectivity. In the legal field, on the other hand, norms are imposed by external authority. They have the function of regulating behaviour by giving authoritative statements of what is acceptable. Legal rules are not mere declarative statements of what is generally expected, or of what is normal or ideal. They are essentiallt prescriptive They correspond not to any objective or social reality (as rules may always be violated), but to what is seen as ideal. Legal language is only said to be normative when it fulfills the function of creating or confirming norms of behaviour.37 Whether in the form of legislation or in common law judgments, normative discourse does not simply provide descriptive statements of existing law. Indeed, insofar as it creates new rights and obligations, it is constitutive of the law. Athough the primary function of common law judgments is to apply existing law, these judgments often entail changes in the legal rules, most obviously where earlier authority is overruled. In such cases, the judgment has a specific, normative function. Although overruling may be analysed as the result of performative speech acts, the notions of performativity and normativity cannot be taken as equivalent. Clearly, not all normative discourse is intended to have performative effect. In particular, legal judgments remain prototypically normative even where they are intended not as performative but as simple restatements of existing law.38 Even where they do imply a change in the law, they can never be taken as performative in their entirety. Conversely, in ordinary language, performatives do not usually have normative effect. Although, Austin is often assumed to have presented performative utterances as creating new social realities or states of affairs,39 Laugier (2004) points out that he never spoke in this context of the creation of a new extralinguistic reality. His ordinary language performative utterances do have certain effects, insofar as they naturally invite

36 In the medical field, ‘normal’ is applied to individuals in perfect health, a very rare condition. 37 Laws may be cited in extenso, for example in a solicitor’s office, or in an academic seminar. However, although the new statement may be linguistically identical, it will have no normative function. Legal norms may thus be distinguished from proverbs, which often take the form of doxastic statements based on common experience. Although they may affect behaviour, this is accomplished through the expression of persuasive recommendations rather than narmative obligations 38 The same is true of “consolidating statutes”, in which the legislature is presumed not to have intended to modify the law. 39 Recanati speaks of relations between speaker and hearer (1981: 19), and of social realities, or states of affairs (1981: 82). The English version of this publication no longer speaks of relations; instead, the performer of a speech act is said to take on a certain role and to assign a corresponding role to the hearer (1981: 9). appropriate responses. However, these effects are purely illocutionary; they correspond neither to the consequential perlocutionary effects, nor to the possible normative effects of the discourse. In themselves, they cannot therefore account for changes in the law. In ordinary language, for the successful performance of the act, the performative intention must be communicated. In Austin’s words (1962: 117) “the performance of an illocutionary act involves the securing of uptake”. This explains why warnings must be heard and understood before they can take effect. Similarly, a bet must be accepted by a taker (1962: 9). Austin insists (1962: 117) that orders can only be obeyed if they are properly communicated, conveyed and understood.40 In such cases, therefore, the effects are necessarily limited to the immediate context of discourse. Ordinary language performatives must therefore be distinguished from positive laws, which are of general application.41 Promises are described by Austin (1962: 9) as “one of the more awe-inspiring performatives”. They may be said to involve the acceptance of an obligation, for the benefit of the promisee. However, for Austin (1962: 9) this did not involve the constitution of a new reality. The obligation derived rather from the pre- existing social duty to speak sincerely and to keep your word. Austin could not accept the alternative view of the promise as a public manifestation of an intention. Such a view would be theoretically unacceptable as, contrary to the theory of speech acts, it would reduce the act of promising to a true or false description of an internal disposition. It would also be unacceptable on the practical level, as insincere promises would fail to have any effect.42. Austin went so far as to describe this view as immoral (1962: 10). Austin’s ordinary language analysis of promises may be contrasted with the earlier analysis presented by Reinach (1913), as part of his phenomenological theory of social acts. Although largely ignored by linguists, Reinach’s account is better known to (continental) philosophers of law.43 On the descriptive level, as indicated by Laugier (2004, 2005), Reinach’s social acts have a lot in common with Austin’s speech acts. Both philosophers claim to have “discovered” the philosophical significance of “something everyone knows” (Austin, 1962: 1; Reinach, 1983: 8).44 Reinach’s social acts are also presented, like Austin’s performatives, as being expressed in the act of speaking itself (Reinach 1983: 18). They correspond to acts like promising or commanding, for example, which require linguistic utterance, as opposed to ‘deciding’, which does not. However, for Reinach, these social acts do not correspond simply to utterance events; they also serve to bring something about within the psychic sphere.45 Reinach’s basic example is again the act of promising. Like Austin, Reinach (1983: 26) refuses to accept the idea of a promise as depending on a psychological intention, of which the promise may be a true or false description. Just as no simple intention could give rise to an obligation, no obligation could be the consequence of an intention. However, contrary to Austin, he claimed (1983: 9) that the social act of promising has the effect of creating something new in the world, a juridical “bond” between promiser and promisee, with a particular ontological existence. This “curious entity” may, however, cease to exist if and when the obligation is fulfilled or otherwise discharged. Reinach’s ontological world therefore contains what may be called “promisings”. This concept corresponds well to legal contracts, usually analysed as bilateral promises intended to give rise to obligations. In the language of lawyers, contracts are frequently presented as having been formed or brought into existence by the parties. Where disagreements arise concerning liability for breach, legal debate typically focuses on the existence or non-existence of the contract (if there is no enforceable contract, then the question of breach cannot arise.) However, while Reinach’s ontological explanation of contractual obligations may appear plausible in the legal context, his theory of social acts does not provide an acceptable account of ordinary language promises, which do not normally have contractual effect. Nor, more significantly, can it provide an explanation of normativity, as contracts have no specifically

40 It is left an open question whether a gift must be accepted. 41 There are exceptions. By the Act 1714 (July 8, Queen Anne), Parliament offered a prize for a practical solution to the problem of determining longitude at sea. It was amended in 1765 (The Longitude Act, 5 George III) to include stipulations that applied specifically to the watchmaker , whose (successful) technique was considered unacceptable. Harrison was actually named in the opening language (see Sobel 1995). In another example, certain clauses inserted in the Serious Organised Crime Act (2005, 132-138) were clearly intended to apply specifically to Brian Haw, who had been demonstrating continuously against war (to the embarrassment of the New Labour government) in Parliament Square for over five years. However, Haw was not actually named in the statute. 42 Reinach (1983: 28) spoke in this context of “pseudo-promising”. 43 The English translation (1983) of Reinach (1913) is very difficult to find. The text has been reprinted in German (1989), and has been trasnlated into French (2004). 44 Page references are to the English translation (1983) of Reinach (1913). 45 Austin was not unaware of the phenomenologial significance of his own work. However he preferred to describe his field as “ordinary language philosophy” as that he found the expression “linguistic phenomenology” “rather a mouthful” (1956: 182) normative effect. Contractual rights and duties are normally restricted to the parties directly involved, who are aware of the obligations they have accepted. Clauses which have not been properly communicated at the time of the agreement are ineffective, and third party rights are notoriously problematic. Laws, on the other hand, apply to the entire jurisdiction. On this level of analysis, there is no clearly definable interlocutor to whom the overruling declaration must be made. Regarding legal norms, therefore, communication may appear not only unnecessary (as “ignorance is no excuse”), but in practice impossible. Even the judges themselves may fail to realise the significance of their decisions until new facts arise in a later case. For this reason, even if it is accepted that promises may create legal obligations, this account still fails to explain the normative effects of judicial speech acts. On the contrary, Reinach insists (1983: 134) that what he calls “essential laws” are not derived from speech acts, but rather from apriori principles closely related to natural law. In his model (1983: 138), these exist independently of the enactments of the positive law and of the acts performed by individuals:

“It is true that we know only of social acts performed by men, only of rights and obligations held by men. But the essential laws which we understand with certainty are not grounded in the fact that these men or that some men or other perform the acts and hold the rights and obligations. but are rather grounded in the essence of the acts and in the essence of the relations of right, no matter when and where they are realized.”

This approach raises new problems for normativity in general, and fails in particular to account for the normativity of overruling declarations. It appears, therefore, that neither Reinach’s theory of social acts, nor Austin’s theory of speech acts can account for the general application which is a defining feature of normative discourse. Although norms are brought into existence by performative utterances, their normativity cannot be derived from the illocutionary effects of ordinary language performatives. However, the idea of performativity as a constituent element of normativity should not necessarily be abandoned. It may be observed that several of Austin’s own examples of conventional performatives do seem to have general consequences beyond the immediate discourse situation, insofar as they have the effect of creating new social and institutional situations. Obvious examples are marriage (“I will”)46 and baptism (“I name this child”), both of which have ongoing, extralinguistic effects. These may be explained in terms of the “institutional facts”, introduced into the literature by Searle (1969). Recanati (1981: 92), suggests that Austin may have omitted to distinguish explicit performative acts from institutional acts, also performed by speech.47 This distinction may help to provide a description of certain aspects of particular judicial performatives. Unfortunately, as an account of normativity, this approach remains dangerously circular. By definition, institutional speech acts can only function in the presence of institutions, with clear linguistic conventions. As individual utterances cannot create or modify these conventions; it is unclear how a judicial declaration could modify the institutional norm.

5. Conclusion - judicial felicity conditions

Whether in the form of legislation or as stated in common law judgments, norms are created and modified through language. The theory of speech acts may therefore provide at least a partial explanation of their normative effects. However, since the law includes rights and powers as well as constitutive rules, legal norms cannot be seen as equivalent to orders or commands. Further, not all prescriptive statements correspond to legal norms. Old-fashioned grammars were intended in principle to be prescriptive, and even modern, descriptive dictionaries are often used prescriptively. However, such publications cannot be described as normative. Legal norms are created through authorised discourse, and imposed on the community. In relation to overruling, the problem is one of authorisation. While the judges have the delegated authority to decide cases, they do not, as a matter of principle, have the authority required to create new law.48 However, as their judgments have the effect of laying down rules for the adjudication of later cases, the two functions cannot be clearly dissociated.

46 However, the marriage itself cannot take place unless the partner is aware of and agreeable to the project. 47 Recanati is here following Warnock (1973). See, alternatively Recanati (1987: 72). 48 In England, the laws are promulgated in the name of the King or Queen; yet since the Bill of Rights 1688, not even the Monarch has the authority to make law extrajudicially. Section 1 provides that “[…] the pretended power of suspending of laws or the execution of laws by regall authority without consent of Parliament is illegall.” A significant statement on this question is found in Entick v Carrington (1765), which dealt with the legality of general warrants allowing searches of private homes. In that case, officers of the crown were engaged on a search for evidence against John Wilkes, who had been accused of sedition (R v Wilkes, 1763- 70). In the earlier “Trial of Harris for a libel”, a similar warrant had been granted by “all the judges of England”. However, Sir G. Jefferies, Recorder, refused to accept this as a valid precedent:

“These are the opinions of all the twelve judges of England; a great and reverend authority. Can the twelve judges extrajudicially make a thing law to bind the kingdom by a declaration, that such is their opinion? I say no. It is a matter of impeachment for any judge to affirm it.” (Entick v Carrington 1765, per Jefferies, Recorder)

The implication here is that the judges do have the authority to make law by declaration, but only when they are sitting judicially. It is tempting to consider this as part of the felicity conditions required to validate the overruling declaration. This (simplistic) requirement of procedural validity corresponds well to basic legal theory, as only laws passed according to the correct procedure can be enforced. Such valid laws will remain generally applicable even if they lead to inconvenience and injustice. 49 The claim that performative discourse becomes normative in an appropriately normative context would thus allow a convenient correspondence between the legal and linguistic conditions for successful overruling. However, on the theoretical level, this approach is clearly circular. It fails to account for instances of failed attempts to overrule, which are the inevitable consequence of the perlocutionary aspects of such speech acts. It would also involve an infinite regress, as the mechanism would itself have to be included as part of the legal system. The declaratory theory simply avoids the problem by denying the fact of legal change. Unfortunately, the claim that the law is as it is because it has been so “since time immemorial” lacks explanatory power. Various alternative explanations of normativity have been proposed. Kelsen (1934: 57) proposes a hypothetical basic norm, which is presented not as an observed socio-legal fact but as a “transcendental condition”. As the basic norm cannot account for its own validity, it must simply be presupposed:

“What is to be valid as norm is whatever the framers of the first constitution have expressed as their will - this is the basic presupposition of all cognition of the legal system resting on the system resting on the constitution.”

While this fundamental presupposition may justify the enforcement of existing law under the present constitution, it offers no solution to the recurring problem of how the legal system provides for its own modification. As neither judges nor legislators have any authority to act outside the law, it remains unclear how change could be brought about without recourse to revolution. Kelsen’s jurisprudential system is justified in practice (1934: 60-61) by the requirements of legal efficacy and by the observation of “the actual behaviour of the human beings addressed by the system”. However, as not all human beings abide by the law at all times, Kelsen is forced to admit that complete correspondence between the legal system and actual behaviour is not necessary.50 Hart (1961), on the other hand, assumes a social “rule of recognition” allowing responsible persons the right to decide the law. It is notoriously unclear, however, whether this is presented as a genuine jurisprudential theory, or whether it corresponds merely to sociological observation. It has been suggested more recently that performative speech acts can only have normative value in virtue of independent, institutional rules. Thus, while admitting the existence of norm-creating speech acts, MacCormick / Bankowski insist (1986: 186) that these can only be valid in virtue of superior, pre-existing norms. They conclude that normativity is logically prior to the performative function and must as a matter of principle be explained independently of speech acts. In MacCormick (1978: 100), this problem is presented as one of “second-order justification”. However, the recognition of the need for independent explanation does not in itself amount to a satisfactory account of the mechanics of norm creation. It is a fundamental axiom of linguistic pragmatics that language does not function independently of extra-linguistic reality. Nor can the judicial system function efficiently without reference to wider society. It is therefore unlikely that the words of individual judges, considered independently of institutional power and influence, could suffice in themselves to change and modify the law. For this reason, it may not be possible to provide a purely linguistic explanation of normative discourse in terms of performative speech acts.51 Nevertheless, the attempt to state the problem in linguistic terms remains useful. In this connection, contributions from linguistics and from legal theory may be found reciprocally illuminating.

49 Academic criticism of unsatisfactory laws has no normative effect. 50 For MacCormick (1978: 104), “legal rules do not present a model of the world but a model for it”. 51 Searle (1989: 546) abandoned the attempt to derive the effect of explicit performatives from the purely linguistic, assertive function of the utterance itself. References

Amselek, Paul (Ed.), 1986. Théorie des Actes de Langage, Éthique et Droit, PUF, Paris. Austin, John, 1832. The Province of Jurisprudence Determined. [CUP, Cambridge (Rumble (Ed.), 1995.] Austin, John L. 1956. A plea for excuses. In: Philosophical Papers, OUP, Oxford, 175-204. Austin, John L. 1962. How to do Things with Words, OUP, Oxford. Blackstone, William. 1762. Commentaries on the Law of England [University of Chicago Press 1979]. Charnock, Ross, 2006, Lexical indeterminacy: contextualism and rule-following in common law adjudication. In: Wagner, Anne / Wouter, Werner / Cao, Deborah (Eds.), Interpretation, Law and the Construction of Meaning, Springer, Berlin, 20-46. Davis, Steven, 1979. Perlocutions, Linguistics and Philosophy 3/2, 225-43. Dunn, Pintip H. 2003. How judges Overrule -speech act theory and the doctrine of stare decisis, Yale Law Journal 113/2, 493-531. Flew, A. (Ed.), 1951. Logic and language (First series), Blackwell, Oxford. Holmes, Oliver W. 1881. The Common Law, Little Brown, Boston [1963]. Hart, Herbert L.A. 1951. The ascription of responsibilities and rights. In: Flew, A. (Ed.), 145-66. Hart, Herbert L.A. 1961. 1961. The concept of law. OUP, Oxford. Kelson, Hans, 1934. Reine Rechtslehre, Franz Deutsche, Vienna. English translation: 1996. Introduction to the problems of Legal Theory, Clarendon, Oxford. Kripke, Saul, 1982. Wittgensteinon rules and private anguage. Blackwell, Oxford. Laugier, Sandra, 2004. Performativité, normativité et droit, Archives de Philosophie 67, 505-24. Laugier, Sandra, 2005. Actes de langage et états des choses, Etudes Philosophiques, 73-98 MacCormick, Neil 1978. Legal Reasoning and Legal Theory, CUP, Cambridge. MacCormick, Neil / Zenon Bankowski, 1986. La théorie des actes de langage et la théorie des actes juridiques. In Amselek (Ed.), 195-209. Recanati, François, 1981. Les Énoncés Performatifs, Minuit, Paris. English translation: 1987. Meaning and Force, CUP, Cambridge. Reid (Lord), 1972. Declarative theory of common law, Journal of the Society of Public Teachers of Law, 22. Reinach, Adolph, 1913. Zur phänomelogie das Rechts: Die apriorischen Grundlagen des bürgerlichen Rechts. In: Husserl (Ed.), Jahrbuch für Philosophie und phänomenologische Forschung 1, 685-847 (Munich). Republished: 1989,Philosophia Verlag, Munich. Englished translation: 1983. The Apriori Foundations of the Civil Law. In: Aletheia, an International Journal of Philosophy 3 (Philosophy of law), International Academy of Philosophy Press, Irving, Texas, 1-142. French translation: 2004. Les fondements a priori du droit civil, Vrin, Paris. Searle, John R. 1969. Speech Acts, CUP, Cambridge. Searle, John R. 1989. How performatives work, Linguistics and Philosophy 12, 533-58. Shapiro, Martin, 1994. Judges as liars, Harvard Journal of Law and Public Policy 17, 155-56. Sobel, Dava, 1995. Longitude, Walker, NY. Warnock, Geoffroy.1973. Some types of performative utterance. In Berlin, I. et al. Essays on J.L. Austin, Clarendon Press, Oxford, 69-89. Cases cited

UK Addie (Collieries) v Dumbreck [1929] AC 358 Mannai Investment Co Ltd v Eagle Star Assurance HL Anns v Merton LDC [1978] AC 728 1997 AC 749 Baker v Willoughby [1970] 3 All ER 1528 Murphy v Brentwood [1990] HL 2 All 908 Beswick v Beswick CA [1966] Ch 538 Mutual Life v Evatt PC [1971] AC 793 Beswick v Beswick HL [1968] AC 58 NatWestBank v Spectrum [2005] AC 41 BR v Herrington [1972] AC 877 Nitrigin Eareann v Inco Alloys [1992] 1 All ER 854 Candler v Crane, Christmas [1951] 1 All ER 426 Ch. Rickards v Oppenhaim CA [1950] AC 616 Conway v Rimmer CA [1967] AC 1031 Pinnel’s case (1602) 5 Coke Rep. 117 Conway v Rimmer HL [1968] AC 910 Practice Statement (Judicial Precedent) HL [1966] 3 Couldery v Bartrum [1881] 19 ChD 394 All E.R. 77 Duncan v Cammell, Laird Co [1942] AC 624 R v Bembridge (1783) 3 Doug KB 32 Donoghue v Stevenson [1932] AC 562 R v Governor of HM Prison Brockhill Ex Parte Evans Dutton v Bognor Regis [1972] 1 All ER 462 CA No 2 [1999] 11 Admin LR 6 Foakes v Beer [1884] AC 605 R v Pedley (1782) Cald 218 Entick v Carrington (1765) 19 ST 1029 R v R (Marital Exemption) [1991] 4 All ER 481 George v Skivington (1869) LR 5 Exch 1 R v Wilkes (1770) 4 Burr 2527 Hankey v Clavering [1942] 2 KB 326 Rooks v Barnard [1964] ! All ER 367 Hallett’s Estate 1880 13 Ch. D. 696 Rust v Cooper (1777) 2 Cowp 629 Hedley Byrne [1964] AC 465 Re Selectmove [1995] 2 All ER 531 Hunter v London Docklands HL [1997] AC 655 Simaan Contracting v Pilkington Glass No 2 [1988] QB Jobling v Associated Dairies [1981] 2 All ER 752 758 Junior Books v Veitchi [1982] 3 All ER 201 Stilk v Myrick (1809) 2 Camp 317 Kleinwort Benson v Lincoln CC HL [1998] 4 All ER SW v United Kingdom ECHR (1996) 1 FLR 43 513 Tweddle v Atkinson (1861) 1 B. & S. 393 Khorasandjian v Bush [1993] 3 All ER 669 Vallejo v Wheeler (1774) 1 Cowp 143 Le Lievre v Gould [1893] 1 QB 491 West Midlands Baptist Trust v Birmingham Corp 1970 London Street Tramways v London CC 1898 AC 375 AC 874, 898 Miller v Miller / McFarlane v McFarlane [2006] Williams v Roffey Bros [1990] 1 All ER 512 UKHL 24 Winterbottom v Wright (1842) 10 M&W 109

US Adkins v Children’s Hospital US 795 (1923) Harper v VA Taxation 509 US 86 (1993) Allegheny College v National Chautauqua County Bank Lawrence v Texas 539 US 558 (2003) 246 NY 369 (1927) Lochner v New York 198 US 45 (1905) Beam Distilling 501 US 529 (1991) Mapp v Ohio 367 US 643 (1961) Benton v Maryland 395 US 78 (1969) Maryland v Wirtz 329 US 183 (1968) Booth v Maryland 482 US 496 (1987) Marbury v Madison 1 Cranch 137 (1803) Bowers v Hardwick 478 US 186 (1986) National League of Cities v Usery 426 US 833 (1976) Brown v Board of Education of Topeka 347 US 483 Oregon v Mitchell 400 US 112 (1970) (1954) Palko v Connecticut 302 US 319 (1937) Bryson v US 396 US 64 (1969) Payne v Tennessee 501 US 808 (1991) Burnet v Coronado Oil & Gas Co 285 US 808 (1932) Planned Parenthood of Southeastern PA v Casey 505 College Savings v Florida Prepaid Education 527 US US 833 (1992) 666 (1999) Plessy v Ferguson 163 US 537 (1896) Dennis v United States 341 US 394 (1951) Roe v Wade 410 US 113 (1973) Garcia v San Antonio Metropolitan Transit 469 US 528 Schenck v US 249 US 47 (1919) (1985) Swift v Tyson 41 US 1 (1842) Gideon v Wainwright 372 US 335 (1963) National League of Cities v Usery 426 US 833 (1976) Gitlow v NY 268 US 652 (1925) West Coast Hotel Co v Parrish 300 US 379 (1937) Great Northern Ry v Sunburst Oil 287 US 358 (1932) Wolf v Colorado 338 US 25 (1949) Griffin v Illinois 351 US 12 (1956)