The Language Myth and the Law

1.Integrationists on the law.

In the short history of Integrational linguistics, there have been a number of discussions of aspects of legal reasoning and jurisprudential practice: one is that of Roy Harris, in The Language Myth, concerning the situated grounds on which judges may have to decide upon meanings or descriptions in particular cases: whether a drunkard wheeling a bicycle is ‘in charge of a carriage’, whether to sell a Bugs Bunny T-shirt in particular circumstances amounts to selling a souvenir, and whether for the purposes of an insurance claim, the accidental scorching of trousers by an electric iron is justifiably described as a fire. In all these cases, Harris’s comments chime with the view of Hart, in The Concept of Law [127-8], who famously observed that legal rules are necessarily, at best, open in texture. To quote just one sentence from Hart:

It is… important to appreciate why, apart from this [i.e., law’s] dependence on language as it actually is, with its characteristics of open texture, we should not cherish, even as an ideal, the conception of a rule so detailed that the question whether it applied or not to a particular case was always settled in advance, and never involved, at the point of actual application, a fresh choice between alternatives.

More recently Christopher Hutton, reviewing contributions to a conference on Law and Linguistics at Washington University, has found not proven the mainstream linguists’ claim that linguistics, in its scientificity, had important things to say to lawyers. Why so? Essentially, because the legal profession is of the world, and applies on a case by case basis, so that there is far more common ground between the literary critic, entering a plea or making a judgement on a particular unique poem, than with the linguist, hypothetico-deductively elaborating a maximally generalized and abstract system underlying performance in one or all natural languages.

As good a place as any to enter upon a discussion of the continuing influence of language-mythological assumptions upon the law is with obiter dicta in Harris’s judgement on Saussure’s Cours, delivered towards the close of his Reading Saussure. There he suggests that for all the brilliance of Saussure’s theoretical account, it nevertheless starts ‘from the wrong end’:

Saussurean linguistics begins by focussing upon the properties of the individual sign in the abstract, and hoping that somehow at the social end, where signs are put to everyday use, everything will work out satisfactorily in terms of communicational corollaries. Unfortunately, it does not work out at all. But it could have worked out if only Saussure had grasped the full implications of the economic analogy and seen that values are subordinate to transactions, and not the other way round. (1987: 230-231)

In a parallel way, we might suggest, metalinguistics must be subordinate to communication. At any rate, Harris goes on to suggest that Saussure’s failure to give priority to transactions led him to the theoretical impasse of attempting to reconcile the implicit synchronic fixity and uniformity across a speech community of la langue, with the inherent fluidity and variability of values. But these forces are utterly incompatible: “Just as in the economic case, it is only if we postulate that the economy is a closed system that it makes sense to speak of fixed currency values and equivalences” (234). Overwhelmingly, human experience suggests that in both our economic and linguistic practices, dynamism and change, in signal, signification, and value, are permanent potentials. Hence a vase bought for £500 yesterday can entirely reasonably be sold for a higher sum next week, even though, relatively independently, the general purchasing power of the pound may have increased, and even though the bought object may by then have been re-classified as a bowl rather than a vase. Here Harris makes the crucial point that it is not simply that with the new transaction, and contingent upon it, a revision of multiple values may occur; in addition, as a corollary of this, it makes no sense to privilege any one price paid, or purchasing power of the currency, or naming of the object traded, as the correct price, the correct currency value, or the correct name. Correctness is so confined of application, to each self-same transaction, as to be hardly worth invoking. (See also Harris’s rejection of the notion of context-neutral correct translation, and contrasting emphasis on communicationally-relevant efficacy of reformulation, in The Language Myth (1981: 148-9)). Rather different checks on ‘utter fluidity of revaluation’ are (a) biomechanical errors where, e.g., a price of £500 is agreed but the purchaser writes a cheque for £900: here it is not so much that in writing ‘£900’ the purchaser has failed to write the correct price, rather they have failed to write the agreed price; and (b) deceptions and unconscionable dealings, where a seller knowing an item is worth £500 induces an ‘unequal’ contractual partner to pay, say, £5,000 for it–see especially the observations of Lord Denning, on economic duress and ‘inequality of bargaining power’, in Lloyds Bank Ltd v Bundy [1975] QB 326. Again we find that the law might protect the victim here not on the grounds that they have paid an incorrect sum, but on the grounds that they have been lured into paying a wholly unreasonable and unconscionable sum. So even in these extreme cases, it seems, economics and contract law have no time for doctrines of ‘correctness’ of values. Nor, in ordinary circumstances, should linguistics.

We are approaching the question of the language myth and the law by way of the principle of correctness, then, notwithstanding the fact that in theory there is no place for a system of values or local correctnesses that is not contingent upon particular transactions.

2.The literal rule in statutory interpretation and the fixed-code theory of language.

But the most directly pertinent of previous discussions, for my purposes, is Nigel Love’s commentary on customary procedures in the interpretation of statutes. For, Love suggests, these “enshrine the fixed-code theory in a number of respects” (1998: 57), which he proceeds to itemize. They include the assumption that legislators’ communicational intentions can be precisely captured, without delimitation of the ruling or provision to a single situation, provided a sufficiently careful and general wording is adopted; this reflects a faith in words as carriers of meanings, and of meanings as stable and undetachable from those words.

In that same spirit, as Love notes, where doubts arise as to a statute’s intent and as to whether a particular set of circumstances was intended to be ‘caught’ by the statute or not, English legal procedure has traditionally firmly disbarred direct consultation of Parliament. Thus appellate judges, grappling with a hard case, and notwithstanding the constitutional sovereignty of Parliament, have not traditionally been allowed to ask Parliament what its members, or the government, or the relevant secretary of state, intended by the statute they may have only recently enacted. Among the arguments against such ‘secondary’ consultation is the prospect of endless second thoughts and revisionism, to say nothing of the complications if a new Labour-dominated Parliament is called on to explain legislation passed by an older Conservative one. Besides, how would one arbitrate as a judge, with constitutional probity, between an ambiguous statutory instrument at time>i, and its unambiguously radical reinterpretation, by the present secretary of state, at timeii? In all such concessions to assumed lack of fixity and autonomy the legal system has traditionally seen intolerable affronts to that most cherished of principles, certainty. Without certainty the lawyer complains, or without a pretty close approximation to it, no-one knows where they stand, whether she be lawyer, judge, claimant, defendant, or police officer, and the law’s vital roots in open and fair dealing are severed. Far better, then, if a new Parliament wishes to revise the laws of an older Parliament, that new laws (and, by extension, new revisions, repeals, and amendments) be passed, creating a new, different, free-standing, meaning-encapsulating, statute or provision. In that way, it is argued, a new certainty abruptly replaces an old certainty, like the handing on of kingship, and uncertainty is minimized.

Besides, it might be countered, should there really be a need for secondary and tertiary interpretations of statutes? Insofar as such a practice concedes that texts, even immensely carefully crafted texts, cannot stand alone and make autonomous communicative sense without benefit of supporting documentation, it may raise questions about the ethical and logical justification for ever resorting to the enactment of free-standing Acts. And at that point you may find, in addition to the big guns of the legal profession trained on you, the little guns of the world of literature, literary criticism and publishing. Far better then, it has been traditionally held, and more congruent with other prevailing assumptions about the place of the written law in relation to society, Parliament, and the judiciary, for statutes to have their autonomous meaningfulness reaffirmed rather than put in doubt.

Even where the reasonable meaning of the words of a statute clash with legislators’ declared intentions, the meaning ‘in the Act’ will prevail, because, it is implicitly assumed, the meaning in the wording of the Act is the meaning of the Act. [It is particularly interesting to note that the traditional qualification to this principle states that, where the literal meaning ‘in the Act’ would result in absurdity, then the golden rule can be invoked, which allows an alternative but still possible interpretation of the words, overriding the ‘ordinary signification’ of the words, to be applied: that is, a less proper signification which the Court thinks the words will bear (as per Lord Blackburn in River Wear Commissioners). What is most striking in that manouevre is the way that this preserves the principle the meaning is still ‘in’ the words, and proper to them, even if less proper than their ordinary or literal meaning.] We may be tempted to conclude that, in light of other declared jurisprudential desiderata such as certainty, consistency, and equal protection, subscription to a fixed-code picture of language is implicit in the very idea of resorting to written statute.

At this point I should disclose that I myself am almost as convinced by these arguments as traditionalist adherents of the literal rule seem to be. There is an immensely powerful pragmatic attraction in the idea of devising a form of words that will encapsulate a rule or provision authoritatively and definitively, that stands the test of time, is proof against the picker of loopholes, the sceptic and the revisionist latecomer. In short, transcendental sentences. Are such things so unthinkable, or so unreasonable an aspiration? Do not the memorable texts of religions of the book, and great literature, aspire to this condition? Why then should such fixed and permanent meaningfulness be looked for or projected onto literature and not legal statute also? Conversely, how can you question the very idea of determinate meaning instantiated in a determinate text, fully accessible by all and every fluent speaker-reader of a language by means of fixed-code telementation, without also questioning the validity of religious and literary texts?

For all such reasons, a fixed-code telementational approach remains immensely attractive, so that if it is set aside as ultimately unworkable, it is not without a sense of loss, and a sense of clarity and the promise of perdurable insight being superseded by complexity and passing illumination.

In the more complex integrationist picture that only with effort displaces myth-based thinking, fixed-code literalist telementation must not be allowed to dictate all the terms within which written statutes, or common law precepts, are viewed. That is the mistake made in suggesting that resort to and reliance upon written codification necessarily entails subscription to a myth-based picture of language: we need to distinguish different kinds of resort and reliance. The appearance of permanence, certainty, fixity of meaning, of meaning ‘in’ the sentences and fully replicated in the head of one competent legal practicioner after another, may be no more than appearances, and not tenets uncritically subscribed to by the entire legal profession. It is perfectly possible that lawyers regard statutes and case law as a network of signposts, pointing out broad general categories and principles even in the most specific-seeming of domains, so that the potential for new and revised meanings is what chiefly adheres to the wordings of the texts and judgements consulted if anything does; so that instead of fixed meaning subsisting ‘in’ the words of the statute, it is rather a matter of variable meanings, in one situation or the next, accruing to the words of the statute, on a necessarily temporary basis.

If there is a tension, in the legal system, between autonomous literalist assumptions and contextualist or integrationist assumptions, one soon encounters it. To begin with, we can set beside the statute, and its intimations of dispensing with the services of common law judging, the appellate recourse to judges, always in the plural: a telling supplement, a frank acknowledgement that the nominally shared and invariant code, like a fragile plumbing system, is prone to breakdowns. It is not a system that runs on its own, without constant maintenance.

The standard story of statutory interpretation is that the mischief or purposive approach was for many decades displaced by the literal rule, this bolstered by a ‘golden’ rule allowing application of a possible but secondary meaning where the literal interpretation would result in absurdity. Thus, through much of the twentieth century, the literal rule has reigned supreme. And that assumption in turn seems to put great trust in the principle of stable, invariant forms and meanings, generally known and shared. As Love remarks, “it is hard to see how such procedures could become established in a society which did not take it for granted that what a language basically has to offer its users is a context-neutral code for the expression and extraction of determinate meanings.” (58) On the other hard for anyone who believes that a language’s forms and meanings are generally stable and invariant, it must be a wonder why that stability is not used as a resource to stabilize the remainder of the language, where indeterminacy and difference seem to lurk; conversely, how is it that those indeterminacies do not spread like a virus, rendering the entire system open or fluid? Is it really possible for a language to be fixed in parts, motile in other parts, and how can you tell the parts apart? However one answers those questions, there is a large body of practice that we have to set beside the fixed-code assumptions that are arguably still powerfully at work in everyday statutory drafting and interpretation: this body of practice is all the work of the appellate courts, the Court of Appeal and the House of Lords. For in the very idea or countenancing of an appeal on a question of legal interpretation, which is what the appeal judges routinely hear, there seems to be an acknowledgement that fixed-codism is fallible if not fallacious.

As has often been noted the demise of the purposive or mischief approach coincided with the firm establishment of the sovereignty of Parliament and, one might suggest, the growing democratic legitimacy of Parliament. Insofar as the mischief approach is oriented towards the framers’ presumed intentions, rather than the sense of their words, it affords judges an enlarged scope, actively to postulate and interpret propositions–the legislators’ intentions–which are rather less in the public domain than the words of the statute. But as Parliament grew to predominate in lawmaking, this brand of judicial activism declined and out of the shift in the balance of power from judges to legislators a paradox emerges: where previously judges might have claimed to be ’empowering’ legislators by attending to the latter’s intentions rather than merely their potentially imprecise statutory expression, a transfer of some judicial power to Parliamentarians carried with it an injunction on judges to cease contemplating legislators’ purposes and to look no futher than the statutory text to uncover meaning.

3. Telementation and legal process.

More will be said in support of the contention that a significant part of legal culture adopts a fixed-code view of language, but in relation to the language myth, code-fixity is only one of the two founding assumptions that needs to be considered; the other assumption is that linguistic communication is telementational, an iterative transfer of one language user’s thoughts to another language user, such that ideational identity is preserved: in linguistic communication, what you get are ideas, and the ideas you get are the very same that your interlocutors have formulated and sent to you. So linguistic communication is essentially, ideational replication without change or deterioration. As Harris’s account of the language myth emphasizes, code-fixity and telementation are, like horse and cart, distinct but mutually enabling. Each, in its own domain, entails a ruthless and near-absolute act of reduction or exclusion, so as to bring our picture (and thus our theory) of language close to a ‘degree-zero’ point of variability: code-fixity asserts or promises an absolute and monosemic matching of forms and meanings, without variation or exception, uncertainty or lacuna, to the point where to know the form is to know, absolutely, the meaning, with the related requirement that linguistic forms or expressions are to be understood as linked of necessity only to well-formed thoughts. Telementation similarly reduces speculations about all the things you might ‘get’, or ‘conceive’, or come to appreciate differently in the course of an interaction involving language; it reduces this range of creative possibility to a virtual zero, a scenario in which you get only what you are given, so that you are the most passive and subordinate of interactional partners.

So, is the law’s working picture of language telementational as well as fixed-code-assuming? This is perhaps harder to demonstrate directly, in a cultural practice which involves language but does not often explicitly theorize about language. But perhaps one small reflex of telementationism is the way judges talk about statutes–as when, near the beginning of his controversial Court of Appeal judgement in Davis v Johnston, Lord Denning declares “To my mind the Act is perfectly clear.” And telementation is arguably implicit in the law at every turn, for without assuming something akin to it, it would be hard for a society to justify imposing the same penalties and sanctions on the diversity of people that come before it. Rather it is assumed that, outwith the case of minors (and what counts as a minor is itself redefined over the years) and the mentally impaired, everyone is effectively equal, equally free, equally cognizant, and–it is reasonable and necessary to assume–equally able to understand ‘the same injunctions’ in the same way, whatever their circumstances. All people can thus be judged by the standard to be expected of ‘the reasonable person’.

Another way in which one might suggest that telementationism is so much the commonsense background of judicial proceedings as to be imperceptible would be to point to the broad assumption that ‘anyone who is English-speaking’ ought to be able to stand up in court, answer questions, understand the charges, and so on, in an English court–and that a speaker who claimed to know and speak a kind of English quite removed from that of the court would be unlikely to receive a sympathetic hearing. That is, the general right to a fair hearing in practice also carries the assumption that the language heard in court will approximate educated Standard English in pronunciation and lexicogrammar; the courts proceed as if defendants and claimants and their representatives approximate that norm, and are disinclined to consider the extent that prosecution and defendant might not share one and the same language.

Telementationism is also fairly directly apparent in the matter of mens rea, the so-called ‘mental element’ which is required in most serious crime prosecutions, but that underlies a great many legal proceedings. The mens rea requirement in serious crimes means, for example, that the prosecution must establish that the defendant acted intentionally, or at least recklessly. At the same time a version of telementationism extends to at least one category of civil law cases, namely defamation cases. As far as the twin pillars of the language myth are concerned, that is, fixed-codeism and telementation, there seems a tendency for statute-controlled law to lean more heavily on the fixed-code notion, while common-law-controlled law leans more heavily on telementation–that is, on the assumption that, for prosecution or claimant to succeed, it must be the case, or is beyond a reasonable doubt the case, that the meanings that judicial participants (judges, barristers, jury) derive from defendant X’s utterances are identical to the meanings that X attached to them.

To highlight the telementational thread running through the legal system, it is important to emphasize the fact that quite systematically the law reduces situated acts to propositions–and in doing so is proceeding in a way that is perhaps analogous to Saussure’s privileging of values over transactions at the opening of this paper. In legal proceedings, a complex and ofen desperate or tragic nexus of situated acts–like those of Mrs Ahluwalia and her husband culminating in his death–are reduced to propositional form, to an idea cast as an assertion or a question: In the defence to a murder charge of provocation, must the defendant show that their sudden and temporary loss of control followed immediately upon the provocation itself? Juries and sometimes judges must decide on the facts of the case before them, but as any law student knows, legal training and practice is ultimately ill-disposed towards mere facts, and would prefer them reduced to the bare and strictly relevant minimum, when making an argument. Because in all the higher courts, cases ultimately turn on single propositions, questions and their answers, purporting to allow of general application and application to the instant case on the balance of probabilities or beyond a reasonable doubt. Those probabilistic caveats might appear to be an overt qualification of the telementational assumption, but they are a qualification that only arises at the point of referential application. Thus the entire apparatus of the law is predicated upon all parties accepting–as common ground–that the crime of theft is the dishonest appropriation of property belonging to another with intention to permanetly deprive, and is equally predicated upon all parties sharing a common idea of what dishonest means here, what appropriation means here, and so on. So like the heads in Saussure’s speech circuit, lawyers, judge, jury–even defendant–are assumed to have a shared ideational matching of the word ‘appropriation’ and a particular meaning, just as each of Saussure’s has the same sign for tree as their partner. And just as in Saussure the chief area of uncertainty is whether the sign should apply to a particular tree-like object in the world, heterogeneity and interactional difference is treated as emerging only at the stage of application of agreed ideas to a particular contested or variously-represented act or omission. I have suggested that telementationism is more germane and assumed in the higher circles of the law, where complex problems are reduced to propositional ideation; by contrast in the everyday law of Citizens’ Advice Centres, Magistrates’ Courts and high-street solicitors, where advice and warnings and settlements predominate, telementation is far less prominent. This I submit reverses the pattern with regard to fixed-codeism, where it is in everyday routine legal practice that code-fixity is presumed while in the ‘higher’ reaches of the law–in the two English courts of appeal, in the work of leading Queen’s Counsels, and in academic legal circles–it is distrusted. To support the latter contention I need to return to the principles of statutory interpretation, and comment on some important changes that this has undergone since Nigel Love’s 1985 paper.

4. From the literal back to the purposive.

From a language-mythological perspective, the literal rule must appear unequivocably as the default option, akin to Lawrence’s injunction ‘never trust the teller, trust the tale’. For the literal rule trusts that, giving the words of a statute their ordinary and reasonable meanings, a coherent account of the Act can be derived. Zander (1994: 110) suggests that the ‘most rigorous’ expression of the literal approach came in Lord Halsbury’s remarks in Hilder v Dexter [1902] A.C. 474, where he argued that the draftsman of a statute was the worst person to interpret it, since he would be the most powerfully influenced by what he meant rather than what he had said. On those grounds, because he had himself drafted the statute relevant to the case, he refused to give judgement. While this ingenious line of reasoning invokes the familiar contrast between ‘what was meant’ and ‘what was said’, it is worth emphasizing that it really entails the postulation of two potentially distinct meanings: ‘the meaning of what was meant’ and ‘the meaning of what was said’. It then asserts that the meaning of what was said is there, in the literal objective wording of the statute, autonomous and fixed, but that a clear view of that meaning is vulnerable to obscuring or contamination by parti-pris interpreters, especially the author with his exceptional stake in affairs. To proceed otherwise–according to this line of reasoning–that is, to seek to lay bare the meaning of what was meant, is clearly to seek to be guided by intentions–the draftsman’s intentions, or the government’s intentions, or Parliament’s. And to do so, particularly in the spirit of Lord Denning in Magor and St Mellens, might involve ‘filling in the gaps and making sense of the enactment’ in ways which are then criticized as an ultra vires enlargement of the judicial function, where ‘making sense of a law’ may appear to be not very different from roundly ‘making law’. So Lord Simonds, reasserting the norm, censures Denning and reminds that ‘the duty of the court is to interpret the words that the legislature has used’. Similarly Denning was in the dissenting minority in the notorious will case, Re Rowland [1963] 1 Ch. 1 (CA), where all hinged on the meaning of coinciding as applied to the deaths of the testator and his spouse. Speaking for the majority, Lord Russell argued that two contrasting interpretations or definitions of coinciding were at stake, rather like alternative dictionary senses: (i) was ‘simultaneous’ and (ii) was ‘on the same occasion and by the same cause’. There being no evidence that the deaths coincided in what Lord Russell, upping the ante, called ‘the natural sense of simultaneous’ in the mind of the ordinary man, he held that the deaths could not be deemed to have coincided. He drew attention also to the co-text, which refers to one death ‘preceding or coinciding with’ the other.

Lord Denning complained that such a construal proceeded from fallaciously assuming that in construing a will ‘It is not what the testator meant, but what is the meaning of his words’ that matters, and he disapproved of this as a nineteenth-century view. But that view, that ‘the meaning of the words’ was sovereign, prevailed in that 1963 case. By contrast, there is little fixed-codism–or at least, a heavily qualified if not locally contradicted fixed-codeism–in Lord Denning’s dissenting judgement:

You must put on [the testator’s] words the meaning which they bore to him. If his words are capable of more than one meaning, or of a wide meaning and a narrow meaning, as they often are, then you must put on them the meaning which he intended them to convey, and not the meaning which the philologist would put on them. In order to discover the meaning which he intended, you will not get much help from a dictionary. It is very unlikely that he used a dictionary, and even less likely that he used the same one as you. What you should do is place yourself as far as possible in his position, taking note of the facts and circumstances known to him at the time, and then say what he meant by his words… I decline, therefore, to ask myself: what do the words mean to a grammarian? I prefer to ask: What did Dr Rowland and his wife mean by the word ‘coincide’ in their wills?

Lord Denning proceeds to a vivid dramatisation of the Rowlands’ communicational situation at the time of making their wills; “it is not difficult to piece together the thoughts that ran through their minds”, he avers, before proposing a conjugal dialogue, in the course of which the wife says “Yes, but what if we both die together?”. This helps Lord Denning to conclude that ‘coinciding with’ means here ‘on the same occasion and by the same cause’–although as Lord Scarman noted, there was no evidence either way concerning a range of possible intervening acts; by the Denning standard, it is unclear whether or not, on the evidence available to those who found their bodies, the deaths of Romeo and Juliet would be judged to have coincided.

Lord Denning, then, long before most of his fellow judges, was sometimes prepared to favour a purposive approach over literalism; very often, however, as in Rowland, he was outvoted by his fellow judges. Accordingly we can say that the literal rule has in the recent past remained predominant, as is only confirmed by the traditional judicial disbarring from consulting Parliamentary debates and the like. This was very much the situation when Love published his article fifteen years ago.

But it is of great interest that this is no longer quite the case, particularly at the appellate level where law is not merely dispensed, but made or revised. The literal rule has lost some of its former dominance, under pressure from the purposive approach–not unrelated to the formerly marginalized mischief rule–which has taken hold over the last quarter-century. And following Pepper v Hart, there is no longer the absolute exclusion of reference to Hansard and similar Parliamentary records, as there formerly was. But the ruling in Pepper v Hart is only one of many indications that, in the last couple of decades, English law has qualified its preference for statutory interpretation to be literalist and its treatment of the text of a statute as free-standing. Other indicators include Lord Scarman’s two Interpretation of Legislation Bills of 1979-1981, both unsuccessful, but indicative of a growing willingness to contemplate a degree of intercourse between the courts and Parliament that would have been hitherto unthinkable, least of all from a leading law lord. Yet another, of profound importance, has been the growing enmeshment of English law within European Union law, the latter characteristically drafted so as to be applied by means of a purposive rather than a literal interpretation (see Denning in Bulmer v Bollinger, and again in Buchanan v Babco Forwarding). It should be no surprise that the judicial activism of the European Court of Justice was congenial to Lord Denning:

The judges do not go by the literal meaning of the words or by the grammatical structure of the sentence. They go by the design or purpose which lies behind it. When they come upon a situation which is to their minds within the spirit–but not the letter–of the legislation, they solve the problem by looking at the design and purpose of the legislature–at the effect which it was sought to achieve. They then interpret the legislation so as to produce the desired effect. This means that they fill in gaps, quite unashamedly, without hesitation. They ask simply: what is the sensible way of dealing with this situation so as to give effect to the presumed purpose of the legislation? They lay down the law accordingly. If you study the decisions of the European Court, you will see that they do it every day. To our eyes–shortsighted by tradition–it is legislation, pure and simple. But, to their eyes, it is fulfilling the true role of the courts. (Denning in Buchanan v Babco, [1977] 2 WLR 107)

Once condemned for advocating such an approach in much earlier domestic cases (Lord Simonds, in Magor and St Mellons RDC , a House of Lords case, called it a “naked usurpation of the legislative power”), Lord Denning evidently embraced this changed judicial role with satisfaction.

But as noted, most influential of all has been the ruling in Pepper v Hart (1993), where in his leading judgement Lord Browne-Wilkinson proposed that in questions of statutory interpretation, reference to parliamentary materials be allowed where the legislation was ambiguous or obscure or threatening absurdity, the material to be referred to came from the relevant minister or promoter of the Bill, and material itself was unambiguous. Despite this limited relaxation, subsequent cases suggest that references to parliamentary documents for clarification or contestation of legislative intent could become quite extensive.

5. A racial group is a group of persons defined by reference to ethnic origins.

The bulk of Love’s 1985 article is taken up with close scrutiny of the appellate court judgements in the important racial discrimination case of Mandla v Dowell Lee. The case involved the exclusion of a Sikh boy from a school on the grounds that his turban and long hair violated the school dress and appearance code. On behalf of the boy’s father, the Commission for Racial Equality claimed that the private school’s rejection of the boy was a discriminatory contravention of the 1976 Race Relations Act, a crucial part of the claim asserting that the Sikhs constituted, within the meaning of the act, a racial group; this argument was controversially dismissed by the Court of Appeal (as it had been by Birmingham Crown Court), and not uncontroversially upheld on appeal by the House of Lords.

How, Love asks, did the two appeal courts go about scrutinizing and interpreting the language of the Act, especially s.3(1), so as to determine whether Sikhs were or were not, in the Act’s terms, a racial group? His premise is straightforward: by the way in which judges scrutinize the words of an act shall be disclosed their implicit theory of language: what it is, what it does, and how it relates to people and the world.

Love demonstrates the lack of cogency of both the CA and HL attempts to pin down whether the Sikhs are a racial group, where the Act provides, as the grounds for defining a group as a racial group, that it is a group of persons “defined by reference to colour, race, nationality, or ethnic or national origins”. He shows the incoherencies that the judges fell into in pursuing the hare of ethnic origins, and the clear dissonance between declaring that Sikhs were a racial group by virtue of being “defined by reference to ethnic origins” when it was also common ground that some people, of disparate ethnicities, become Sikhs by conversion. Since no particular ethnic origin is, in fact, an essential criterion of being a Sikh, Sikhs evidently are not defined by reference to ethnic origins. The law lords did not distinguish defining characteristics from typical characteristics; or, to put things more positively, chose to interpret ‘defined by reference to’ as meaning ‘typically recognizable by distinctiveness of’ rather than ‘necessarily having a distinctiveness of’. In this they might be said to have taken a purposive rather than literalist approach to the statute. It is important to note that s.3(1) of the Act also implies that a person can belong to more than one racial group. On this occasion, the earlier judgement of the Court of Appeal and of Lord Denning in particular, was not signally more perspicuous. For while Denning was prepared to categorize the Jews ‘as a whole’ as a racial group sharing a common characteristic, even though particular Jews may be converts, he with little explanation declines to apply the same reasoning to Sikhs.

More careful was the literalist point made by Oliver LJ in the Court of Appeal: that s.3(1) does not require simply that a protected racial group be an ethnic group but specifically a group defined by reference to ethnic origins. Even if every member of a group, without exception, shared an exclusive ethnicity, it would not be–in the act’s terms–a racial group unless it was defined by reference to those ethnic origins. Similarly an Irish chess club is not a racial group even if all its members enjoy a long shared ancestry, if club membership is defined by interest in chess, payment of dues, and so on, and not by ethnic origins. Having an attribute is not the same thing as being defined by reference to that attribute.

Love contends that if their lordships had dwelt upon the term Sikh, and had decided what characteristics or conditions were the necessary ones for someone to be defined as a Sikh, then a degree of clarity would have been established. Whatever things Sikhs were deemed to be defined by (and not merely ‘ordinarily associated with’), if those things included reference to ethnic origins then the case for saying Sikhs are a racial group would be made; if not, not. That way of putting things allows that Sikhhood might be defined by multiple criteria, whereas Love asks for a single decisive criterion, and even offers one, namely ‘adherence to the tenets of Sikhism’. He suggests that the judges erred in attempting to define racial group when they should have striven to define Sikhs, so as to reveal whether the latter was a group defined by ethnic origins.

If Parliament had intended Sikhs to be covered by the Act, it is hard not to conclude that the wording of the Act is inadequate to the purposes aimed at. The leading judgement of Lord Fraser, as Love notes, simply ignores the wording ‘defined by reference to’, while the concurring judgement of Lord Templeman interprets it very liberally. If Parliament meant what those words mean, Love says, then the judgement of the House of Lords in Mandla seems scarcely tenable.

Indeed as soon as one begins delving into this episode of near-contemporary socio-legal incident, the sense of mixed motives and social friction is pervasive. The bare bones of the interpretive issues outlined above does not alert us to the fact that the private school involved had a significant number of pupils from ethnic minorities attending, and that five Sikh boys were already in the school, wearing their hair short; that the school had a distinctly Christian ethos, including Christian religious classes, and that on that basis he would not have subsequently sent his son to the school even if the headmaster had allowed the boy to wear his turban; that the headmaster lacked funds to hire lawyers to make his defence for him, and therefore defended himself in the Crown court, and thereafter in both appeal courts; that the father of plaintiff was a solicitor practising in Birmingham, and that the case was pursued out of the public purse by the recently formed Commission for Racial Equality, whose attorneys included Geoffrey Bindman; that Lord Denning’s Court of Appeal judgement came on his penultimate day in office as Master of the Rolls before retirement, and that in saluting him the second leader in The Times commented “he was in fine clear form yesterday… on the subject of the ethnicity of Sikhs”; that in reaching their decision the Court of Appeal were not overly mindful of two recent industrial tribunal decisions, which had decided that Sikhs were an ethnic group; that the Jewish Employment Action Group responded rapidly to the Court of Appeal ruling, well before the Lords’ appeal was heard, warning that if Lord Denning’s actual criteria were embraced, then even the Jews that he believed to be protected by the Act might on closer scrutiny be held not to be; that both Lord Denning and Lord Oliver roundly criticized the CRE for relentless pursuit of the headmaster, suggesting that it had fanned the flames of racial conflict, Lord Oliver going so far as to call the Commission “an agent of oppression”; that the law lords equally sharply rebuked the appeal court judges for this unwarranted attack on a recently created statutory body and further intimated that such remarks bordered on interference with the sovereignty of the legislature; that in response to disquiet in and out of the House following the Court of Appeal decision, on November 18 the Home Office minister reported that leave to appeal to the House of Lords had been granted (notwithstanding the Court of Appeal’s earlier denial of this), and he went on to say “The Government … will be able to consider whether legislative action is necessary in the light of the Lords’ decision.”; and that the White paper on racial discrimination which preceded the 1976 Act specifically mentioned the wearing of turbans as an area which would be covered by the legislation; that in the Birmingham Post reporting of the original crown court hearing the claimant was consistently misnamed as Mangla; and so on.

Where should their lordships have gone to discover what Sikhs are defined by? Their knowledge of the world, a good encyclopedia, a dictionary aimed at furnishing definitions? Love objected to Lord Templeman looking up ethnic in a dictionary, on the grounds that whether the word ethnic applies to Sikhs would be a matter of the word’s extension (while the dictionary would only supply its intension). Arguably, had Lord Templeman looked up Sikhs in a dictionary, he would have been less misguided; whether a dictionary would say enough to give grounds for resolving the question is, however, doubtful, since dictionaries tend to deal in standard or established usage and definitions, rather than contentious, borderline or emergent ones that become critical in legal actions. Love’s integrationist-minded conclusion was that both the glosses ‘typically recognizable by’ and ‘necessarily having’ are competing interpretations of defined by reference to, and that an argument in support of the latter, over that preferred by the law lords, is an opinion rather a definitive account. The meaning of defined by reference to is not fixed, so that Lord Templeman’s opinion that it means ‘recognizable by’ is at best rebuttable rather than refutable. Interestingly, in more recent cases appeal judges have moved from ‘recognizable by’ to ‘referable to’: in the Dawkins Court of Appeal–1993, IRLR, 284–it was held that Rastafarians ‘do not have a separate identity referable to ethnic origins’. It is chiefly on grounds of coherence and consistency with other interpretations of statutes that the second, criterial approach is preferable (that defined by reference to should be interpreted as meaning ‘necessarily having’). For, Love says, in truth there is nothing ‘on the face’ of the actual wording of the Race Relations Act that can unambiguously guide us to an answer to the question whether Sikhs are supposed to count as a racial group:

Those who want to know whether Parliament would like the Sikhs to enjoy the protection of the 1976 Act will have to look elsewhere (for example, to Parliament) for an answer. (66).

If judges do that, one line of resistance counters, why should they–or Members of Parliament–strive unduly for definitive accuracy in the drafting (or even the passing) of Acts at all? That rejoinder however is more recalcitrant than reasonable: is it raised so as to protect certainty, or simply to minimize inconvenience? And if the chief objection to supplementing the interpretation of ambiguous statutory language by making reference to disambiguating Parliamentary explanation is simply that it is time-consuming, should this be given any more respect than one might extend to a complaint that a particular search was ‘too thorough’?

But perhaps the House of Lords have “violated the rules of the legal language-game” (Love, 1985: 65), and in particular the rule that says literal interpretation of statutory language, where it does not yield absurd results, can and should be relied upon to determine correct interpretation. At the same time Love points out that the law lords’ ‘going through the motions’ of interpreting by means of fixed-code literalism is important for anyone who wishes to displace the language myth. For it means that the fixed-code thinking “cannot simply be dismissed as a sort of naïve, empirically refutable mistake made by linguists” about language: rather, fixed-codism, and telementationism, are tacit assumptions in the thinking of ordinary people, the mythical reasonable man, legislators and judges.

To re-cap, when the law lords hand down a ruling that Sikhs are, or are not, ‘a racial group’ within the terms of the 1976 Race Relations Act, as in countless other legal ‘hard cases’ that hinge on a nomenclatural issue (what is ‘loose soil’, or ‘provocation’, or ‘torture’, or ‘member of the family’, or ‘mental capacity’, or ‘direct professional reliance’, or ‘dishonesty’, etc. and do the circumstances of the present case involve an instance of this?), a codificatory and classificatory act entailing both inclusion and exclusion is required from the court. Sikhs must be determined either to be, or not to be, a racial group. And so on. There lies the reflection of a ‘fixed-code’ view of language with, in light of the principle of equality before the law, the related assumption that the fixed code is the same fixed code for all citizens within the jurisdiction. The law equally, via binding precedent and the hierarchy of the courts, seeks to enforce a homogeneity, in principle, across the entire judicial system, from magistrates up: what the law lords have fixed as code prevails for an indefinite period, until such time as the lords overrule themselves (itself a relatively new notion, only explicitly declared in 1966). This is surely a Saussurean solution to the difficulty of subscribing to homogeneous code-fixity while recognizing that at another time, an equally fixed code–not intrinsically open to ‘development’ or change–may apply.

The Race Relations Act s.3(1) characterizes ‘racial group’ as follows:

‘racial group’ means a group of persons defined by reference to colour, race, nationality, or ethnic or national origins, and reference to a person’s racial group refer to any racial group into which he falls.

The difficulties that arise in interpreting and applying this section relate not merely to the fact that race is here supplemented, inter alia, by the idea of ethnicity: in addition, reference is made to ethnic origins, requiring some clarification of what can be meant by ‘origins’. It has been argued that concerning the conjoint phrase, ‘national origins’, what needed ascertaining were “identifiable, historical and geographical elements, which at least at some time revealed the existence of a nation” (NJPB v Power, 1997); but this still leaves unclear the precise basis on which, given the acknowledged existence of a nation in those terms–e.g., the Breton nation–an individual can be determined to be a member of a group defined by that nationhood: must the individual have been born in Britanny, or speak Breton, or have at least one Breton grandparent, or simply claim to be Breton? Origins would appear to have something to do with ‘status of the individual through or at birth’, but what precisely? What are these kinds of origins, that implicitly apply equally to national and to ethnic characteristics? A host of judges and academics have argued that the wording of s.3 confirms that Parliament wanted ‘racial group’ to apply on bases quite other than contentious biological or genetic grounds, and that the Act is robust in its treatment of race as a sociocultural category. That being so, is reference to ‘origins’ misleading, or a compromise with science and genetics? Caught somewhere along the continuum that ranges from nature to culture, those drafting the race relations legislation knew that the discrimination they wished to catch was directed at other than biologically distinct races (even assuming those could be established), but were evidently directed not to characterize racial group so loosely that it might include uninherited cultural group affiliations, including religious affiliation. One of the reasons given for the displacement of a problematic Race Relations Act by a more comprehensive Cultural Discrimination Act, including religions, is that such an Act would have made immense difficulties for the position of England’s Established Church, a constitutionally-privileged group. As things stand, religious discrimination in itself is not currently unlawful in England and Wales (unlike Northern Ireland), but the position is likely to shift shortly, when the Human Rights Act comes into force in October, 2000: article 9 protects freedom of religion, while article 14 disallows, inter alia, discrimination on grounds of religion. Thus new balances will likely need to be struck, reconciling these entitlements.

In the mean time, since Mandla, there have been a few important cases that have claimed to apply that ruling. Perhaps most notable has been Dawkins, which hinged on whether Rastafarians were a racial group defined by reference to ethnic origins, for the purposes of the act. (The plaintiff, refused a job as a van-driver because he refused to cut off his dreadlocks, claimed racial discrimination.) In this case counsel for the plaintiff sought to highlight the inconsistencies and uncertainties that arise in any attempt to reconcile Lord Fraser’s nomination of two essential characteristics (long shared history, and culture and customs) with Lord Templeman’s three criteria (group descent, shared geography, group history). Giving the leading judgement, LJ Neil suggested that counsel had fallen into “the error of equating the language used in speeches and judgements with that of a statute” and besides, he confessed he was unable to detect “any real difference in substance” between what the two law lords had said. Interestingly, the prosecution also attempted to make reference to the comments of the Home Secretary at the time of the passage of the Race Relations Bill, on the new Pepper v Hart grounds that the statutory term ‘ethnic origins’ remained obscure and ambiguous. But LJ Neil rejected this proposal also, on the grounds that the law lords had effectively disambiguated the contentious term. As Parpworth has noted (Parpworth, 1993: 612), it will be interesting to see if the courts will always favour prior judicial interpretations–generally felt to be inconclusive–of ambiguous terms. Applying the guidelines of Lords Fraser and Templeman as constructively as possible, it might be concluded that definition of a group defined by ethnic origins, further delimited so as to require more than group religious identity, reduces essentially to the following:

A self-identified and other-recognized group with a long and distinct culture, where that culture entails significantly more than purely religious observances.

On this basis, for the purposes of the act, Jews would be a group of ethnic origin while Catholics are not (but what about Polish Catholics, or Irish Catholics?). Whether Rastafarians or Brummies or Woodcraft folk are ethnic groups becomes a matter of largely subjective determination. Lord Neil decided that the sixty years since the founding of Rastafarianism was not ‘long enough’ to count as a long history, and this alone seems to have been the basis on which the appeal failed.

6. Hard cases, decisions, and results.

Essentially, the law would like ‘theft’, ‘capacity’, ‘dishonesty’, ‘a building’, ‘in control of’, ‘entry without permission or licence’, ‘a member of the family’, etc. etc. to carry, indefinitely, context-neutral invariant meanings: if they were truly context-neutral invariants, why ever — or how ever–would a need to change their meanings arise? But the evident reality is that meanings, applications, and rulings, turn out to be extensively non-identical, and responsive to change and innovation.

One of the things that can be said about the Mandla case is that it is incontestably an example of what lawyers call a hard case. And, as is well known, to a great extent the important reported cases heard by the two Appeal courts are hard cases. But these do have to be decided–even if, thereafter, practising lawyers, legal academics, and fellow judges continue to dispute over quite what the ratio decidendi of the decision was. In the circumstances it might be better to say that, even in hard cases, judgements finally get delivered. Whether that judgement is to be followed, or is an Appeal Court decision promptly reversed by the Law Lords, or whether there are special circumstances that make its general applicability suspect, or whether the legislature is provoked by the ruling rapidly to enact corrective legislation, are all open questions. And a related impression that disputes and crimes are dealt with on a case by case basis is thus affirmed.

But that impression has to be set beside the reality that the law works in a rather different way: in the vast majority of instances, the specificities of cases and the context-sensitive distinctivenesses of meaning are elided and framed, by literally summary means (Atiyah, 1993: 216-7, for example, reports that no more than 1.5% of the 300,000 tort claims that arise in one year actually reach judges). Most civil disputes do not go to court, or if they do so, are not heard in extenso; and most criminal cases are dealt with summarily, or upon indictment by means of a guity plea, so that most situations are concluded without anything remotely resembling a full hearing, but rather by means of a standardizing assigning of situations–as tokens–to established types. In composing this book of the law–the law as an already-completed book–the role of compositor is taken by lawyers, magistrates, and judges of the crown and county courts. By contrast the hard cases project the law as a text some parts of which are only available in draft form and subject to revision. Hard cases are decided in an active unscripted improvisatory process by means of which competing and conflicting interactionally-relevant considerations are weighed up and a balance is struck. The balance may be fair and just, or we may feel that, for example, the thumb of policy or public outrage is tilting it to one side, but decisions are made and handed down, and here much more important than ‘correctness of decision’ is the fact that a decision, a situated ending, has been reached. Similarly, even in a World Cup Final, more important than the fact that team A’s solitary winning goal is clearly adjudged offside by everyone in the stadium save the referee is the fact that by that goal team A win and team B lose. Sports contests, like disputes, must come to an end, even if this is not ‘the end’, otherwise we could impose no episodic structure on experience at all. In short, hard cases are often decided in a quite integrationist spirit, but most cases are not–indeed they may not travel far enough into the legal system to become cases at all, but are summarily determined, very much along lines laid down by our native language myth, in accordance with which it is thought reasonable to detach a particular act or omission from as many of its surrounding circumstances as possible, and assign it to a collectively agreed determinate category and impose a relatively fixed penalty.

I began this essay with Roy Harris’s important observation about how Saussure subordinates transactions to values, when the latter is nearer the truth. My conclusions are that different strata of the legal system embrace or resist the language myth to different degrees: high street law is fixed-codeist but not particularly telementational, academic and appeal court law is implicitly telementational but, perhaps increasingly, resistant to code-fixity. That is to say, the most authoritative legal circles, the appeal courts and the law commissions, are the ones where–as reflected by Pepper v Hart, European interpretive purposeness, and elsewhere–the law seems most receptive to the business of re-setting values in light of a new transaction. But if the law is approaching language differently and in a more integrational spirit, and if this gradually filters down to everyday practice at the high-street or magistrate court level, then we should expect a number of equally interesting developments: more improvisatory judicial activism, more ‘politicised’ jurisprudence, less certainty, and more anxiety about the survival of the rule of law.


  • Lloyds Bank Ltd v Bundy [1975] QB 326
  • Mandla (Sewa Singh) v. Dowell Lee [1983] C.L.Y. 1163
  • Pepper v Hart [1993] 1 All E.R. 42
  • Dawkins v Crown Suppliers (PSA)[1993] I.R.L.R. 284
  • Re Rowland [1963] 1 Ch. 1


  • Atiyah, P.S. 1993. Accidents, Compensation, and the Law. Ed. P. Cane. London: Butterworths.
  • Bell, J. and G. Engle QC. 1995. Cross on Statutory Interpretation. London: Butterworths.
  • Beynon, H. and N. Love. 1984. Mandla and the meaning of racial group. The Law Quarterly Review, 100, 120-136.
  • Bix, B. 1995. Law, Language and Legal Determinacy. OUP
  • Evans, J. 1989. Statutory Interpretation: Problems of Interpretation. Oxford: OUP
  • Goodrich, P. Legal Discourse. London: Macmillan.
  • Harris, R. 1981. The Language Myth. London: Duckworth
  • Harris, R. 1987. Reading Saussure. London: Duckworth.
  • Hart, H. L. 1961. The Concept of Law. Oxford: Clarendon Press.
  • Hurst, D. J. 1983. The problem of the elderly statute. Legal Studies. 3:1.
  • Hutton, C. 1998. Law lessons for linguists? Accountability and acts of professional classification. Language & Communication, 16:3, 205-214. Reprinted in Integrational Linguistics: A First Reader, eds. R. Harris and G. Wolf, Oxford: Pergamon, 1998, 294-304.
  • Love, N. 1985. The fixed-code theory. Language & Communication, 5:1, 1-17.| Reprinted in Integrational Linguistics: A First Reader, eds. R. Harris and G. Wolf, Oxford: Pergamon, 1998, 49-67.
  • McKenna, I.B. 1983. Racial discrimination. Modern Law Review, vol.46, 759-770.
  • Parpworth, N. Defining ethnic origins. New Law Journal, vol.143 (30.4.1993), 610-612.
  • Twining and Miers. 1999. How to do things with rules. London: Butterworths.
  • Zander, M. 1994 . The Law-Making Process. 4th edn. London: Butterworths.