The fundamental question of course that underpins any attempt to examine fiction theory is one of definition. What does one actually mean by the expression ‘legal fiction’?⁶ If one adopts a narrow definition the term could be said to apply only to situations where a statement asserts something to be true that is patently untrue. In fact what a narrow definition entails is the distinguishing of ‘fiction’ from a range of other terms like ‘presumption’,⁷ ‘concept’, ‘classificatory device’, ‘name’, ‘juristic creation’ and so on. In this situation the notion of a legal fiction becomes little more than an island in what might be called a conceptual landscape of legal thought with the result that it is rendered incapable of acting as the basis of a theory about law itself. Perhaps, then, the history of legal fictions is not just one of escape from this seemingly primitive device but also of the isolation and confinement of the very term ‘fiction’.⁸

Yet even if the notion of a legal fiction is isolated and confined by definition it still has the capacity to provoke fundamental epistemological questions as Jones indicates.⁹ The debate around the question of whether a corporation is a fictional legal person (persona ficta) or a thing that actually exists (res) relates directly to the nominalist revolution of the late Middle Ages.¹⁰ Do forests exist or are there only trees? In the eyes of the nominalist a generic expression such as forest exists only in the mind and not outside of it and thus is only a name (nomen) or sign (signum) and not an actual thing (res).¹¹ This debate had, and has, profound implications for legal theory in that nominalism challenged the idea that the mere existence of society implied the existence, and need for, law: ubi societas ibi ius.¹² If ‘society’ is merely a name—a fiction?—and what exist are only individual men and women ¹³ (although Margaret Thatcher added families as well)¹⁴ then the question arises as to why one should be obliged politically and legally. This led to another fiction, that of social contract theory.¹⁵ “If there was in truth no contract”, notes Jolowicz, “then it is merely useless fiction to base state and law upon it.”¹⁶

It is not difficult, then, to move from a specific debate about fiction theory—namely the one surrounding the nature of a legal corporation—to the more general idea that the institutional system of legal thought as devised by Gaius is itself a fictional structure.¹⁷ Just as the development of the legal person (persona) is a fiction (persona ficta) so the extension of a ‘thing’ (res) to encompass intangible property is equally a fiction.¹⁸ Gaius says that such property—namely the res incorporalis—is a creation of the law (ius), but as Jones explains the Post-Glossators tended to merge the idea of law and fiction through expressions such as fictio iuris.¹⁹ Is not therefore the whole institutional system of persona, res and actio nothing but a fictio iuris?²⁰ Certainly Michel Villey saw the system as un vaste théâtre qui est le monde du droit in which there are actors (personae), props (res) and acting (actiones).²¹ And Jones notes that there “are indeed not a few writers who frankly accept the conclusion that if, as they hold, the legal personality of the corporation is a fiction, so is also the notion of a legal right, a legal duty, and of obligation generally”.²² Yet such a conclusion does have some serious epistemological implications not just for law but also for epistemology in general. Are all conceptual structures—including those employed in the natural sciences—nothing but fictional creations?

One response to this question is to focus on the function and the validation of the conceptual model in issue. In the natural sciences the conceptual constructions for the most part are developed to model the behaviour of an external object which in its turn can act as the means of validating the model itself. If the model both explains and predicts the behaviour of the object, then the functional efficacy of the conceptual construction can be verified through correspondence between model and object. Scientists for the most part are applying a causal scheme of intelligibility to an inert object and the credibility—some would say ‘truth’ value —of the scheme is achieved through rigorous testing. In short, the function of a scientific model is to explain a phenomenon. Social science models are different because the object of the conceptual scheme are humans whose free will behaviour makes causal analysis extremely difficult and controversial. Causation becomes entangled with interpretation resulting in a scheme of intelligibility that emphasises comprehension or understanding of a phenomenon rather than an explanation of it.²³ Thus the historian who analyses an event and its effects as a ‘revolution’ is interposing between the event and its analysis a concept that is as much interpretative as causal. What qualifies an event as being a ‘revolution’? Just as important is the level of observation. As Desjeux observes, “social reality cannot be reduced to a single equation” and so when “the level of operation changes the reality being observed changes”.²⁴ Accordingly the historian who works on a history de longue durée is not denying that there are no individual heroes or great men. For if one focuses on the individual it is normal not to see social classes or institutions but this does not mean that they do not exist.²⁵

Between the natural and the social sciences there is, then, a difference of epistemological representation. The schemes of intelligibility employed by the social sciences are not, it might be said, direct representations of the object under examination nor are they based on logical relations. They are an argumentative process founded on “imaginative forms of inference ”.²⁶ The model is thus a “figurative” representation of an object “which cannot be represented in its properties in directly reducing them to a conceptual determination”.²⁷ This is what endows the regime with its fictional character.²⁸ The model can never bring together representation and reality; it can only provide a linguistic or textual treatment necessitating a hermeneutical, and not a causal, scheme of intelligibility. Thus the reasoning method associated with this ‘fictional’ model is neither deductive nor inductive. It is abductive in that concepts are regarded ‘as if’ they were ‘true’ so as to provide a means of comprehension of an object that could not otherwise be comprehended.²⁹

Does this mean that social science concepts are inferior to those employed in the natural sciences? Certainly Granger asserts that a theory dealing with human facts is constantly threatened, unless one is careful, with turning into an ideology whereby myths are substituted for concepts and prescriptions for descriptions.³⁰ But one response is to say that social science facts do not permit themselves to be conceptualised within specific contextualised constructions capable of predicting the behaviour of the phenomenon through the manipulation of logical (often mathematical) relations. In other words there is not some objective ‘reality’ waiting out there to be seized by an accurate model of concepts. There are only realities existing within the models themselves and thus the ‘reality’ or the ‘truth’ is located as much in the model as in some objective reality of social facts. Another response is to say that this is just as true of models in the hard sciences as in the social sciences. In other words that all concepts are fictiones rationis and that once this is admitted all that matters is the “practical value in facilitating the play and speed of our mental processes.”³¹

The leading exponent of this theory that all concepts, including those employed in the natural sciences, are fictiones rationis is Hans Vaihinger.³² Jones regarded this theory as “somewhat gratifying to the self-esteem of lawyers [in] that in the hands of Vaihinger the legal fiction has become the starting point of a new system of philosophy.” The human mind, according to this philosophy of ‘as if’ “is so constituted that it cannot dispense with fictions.”³³ In fact Vaihinger’s research question was much more sophisticated than perhaps Jones’ description might suggest. How is it, to borrow Fuller’s translation , that with consciously false ideas we are nevertheless able to reach conclusions which are ‘right’ or ‘true’?³⁴ Or, as Christophe Bouriau , puts it: how can one reach truth through falsity ?³⁵ In mathematics for example one can deduce from apparently absurd notions such as infinity and negative numbers results that are anything but absurd.³⁶ In the other sciences, on closer examination, the fundamental concepts such as energy and matter turn out to be tautological,³⁷ if not metaphorical.³⁸ Yet these models seem not only to explain a natural phenomenon or object but to predict its behaviour. Indeed, as we have seen, the validity of the scientific model lies in its refutability (or potential refutability), that is to say in its relation to the actual behaviour or existence of the object or phenomenon. But what is this reality—these objects and phenomena? The problem, as Fuller notes, is to be found in the notion of a ‘thing’ which is defined only in terms of its properties.³⁹ One defines sugar in terms of its sweetness and other properties thus making a distinction between a thing (sugar) and its properties (sweetness etc); yet the “properties are the thing”.⁴⁰ Even reality becomes fictional, as if ‘things’ exist independently of their ‘properties’. As Fuller goes on to say, the difficulty is not answering Vaihinger’s question; the difficulty lies in “the assumptions that were made in asking the question.”⁴¹

When one moves from social science theorising in general to theorising about law a number of particular problems emerge that are perhaps less of an issue in other social science disciplines . What actually is the object of a legal theory? Is it ‘law’ or is it ‘social reality’ ? Or, again, is it the actors within the legal world? This issue has emerged with regard to the notion of a legal ‘science’ which, thanks to the expression scientia iuris to be found in the Roman sources, has something of a long history.⁴² Is one talking of law as a science or is one talking of a science of law? The issue is one both of meta-language and of epistemology .⁴³ If one is talking of a science of law then the object of the science (or theory) is ‘law’. What is law as a ‘fact’, or at least as an object of scientific inquiry ? If, however, one is talking of law as a science the object is different: how does law conceptualise social reality ? These two different levels are evident when, for example, one compares the writings of Ronald Dworkin with those of Peter Birks . Dworkin’s enquiries were aimed at explaining the notion of ‘law’ itself while Birks was concerned with how legal systematics relates to social practices.⁴⁴

As has been mentioned, another object of investigation is what might be called legal actors , for example judges, lawyers, legal academics, policemen, bailiffs and so on. The meta-language to be employed here depends upon the nature of the enquiry: is one looking at what these officials do or is one theorising about how they think—or indeed about how they ought to think? These different levels of operation are not always clearly discernible in the literature on legal theory—in courses on ‘jurisprudence’—because the emphasis tends to be on describing, analysing and contrasting the theories of different theorists or schools of theorists. Thus for example Dworkin is often presented as the antagonist of Herbert Hart and while Dworkin did indeed build his theories on a critique of an aspect of Hart’s positivism it does have to be remembered that the two authors had two rather different objects in mind. Hart was primarily concerned with fashioning a theory of law in which his object was a model of rules. Dworkin, in contrast, had as his object one of the key actors in the legal world , namely the judge and judicial decision-making . For the most part were they really talking about the same thing?

These distinctions are important when looking at legal theories in the context of fictions. Hart, in claiming that his theory was sociological in its orientation (“an essay in descriptive sociology”),⁴⁵ would no doubt have been disturbed by the idea that his thesis was a work of fiction, although as we have seen fiction theory has a respectable place in social science epistemology.⁴⁶ Moreover Hart actually uses a fictional character to help explain his concept of law.⁴⁷ Dworkin , however, openly made use of fiction in two principal ways: he created a fictional superhuman judge and he compared the writing of judgments with the production of a chain novel . Had, then, Jones been updating his book today it would be interesting to see under which chapter headings these two legal theorists would have been discussed. In fact this issue is more interesting than it might first appear given that Jones did not employ the kind of headings with which legal theorists are today familiar. Positivism is located primarily, although not exclusively, in the chapter on metaphysical theories and while this is understandable in respect of a positivist such as Hans Kelsen—actually given his own chapter by Jones under the heading of Pure Theory—it might at first sight seem that such a label would be inappropriate for the Hart’s ‘sociological’ version of positivism .

However Hart’s key notion underpinning his thesis—the notion that provides a system of rules with its legal authority—is the ‘rule of recognition’ . As he says, this may take a variety of forms,⁴⁸ but “unlike other rules of the system” the assertion that the rule of recognition exists “can only be an external statement of fact.” It exists as the “practice of the courts, officials, and private persons in identifying the law by reference to certain criteria.”⁴⁹ In insisting on its “existence as a matter of fact”⁵⁰ is not Hart actually asserting that the courts, officials and private persons act ‘as if’ there is a rule of recognition? And even if the rule were to be reduced to a tangible text of some sort would this actually alter its ‘as if’—its fictional—quality? The point that needs stressing is that to say that the rule ‘exists’ because one can observe legal actors apparently observing it is to provide a constructive model that is not actually directly representing the object in question. The empirical object in question is the totality of practices of the actors which means that the rule of recognition is unobservable as an empirical reality. The rule of recognition can only be envisaged; it can only be presented ‘as if’ it is there in the practices .⁵¹ Dworkin’s ‘as if’ construction—his analogy between judging and writing chain novels —might be further removed as a schematic model from the actual object (judging) being described than is the case with respect to the rule of recognition and the practices of legal actors . But this does not make the Hart model any less of a fictional (‘as if’) construction. Just as the practices to be observed in churches and monasteries are of religious actors acting ‘as if’ there is a God, so legal actors act ‘as if’ there is a rule of recognition .

Another way of approaching this rule of recognition issue is to compare Hart’s model of primary and secondary rules with Hans Kelsen’s theory of norms.⁵² In order to avoid infinite regression and to provide a foundation for his pyramid of norms Kelsen proposed the fundamental norm (Grundnorm) as a ‘necessary supposition’ or hypothesis. Later, however, and influenced by Vaihinger , he abandons hypothesis for the word fiction on the basis that hypothesis is a notion that applies only to something that has existed.⁵³ A hypothesis claims to correspond with facts .⁵⁴ The Grundnorm , in contrast, not only does not correspond with any fact but also contains an internal contradiction which renders it an ‘authentic fiction’.⁵⁵ It is contradictory, notes Christophe Bouriau in his work on Hans Vaihinger, because it conflicts with the definition of a legal norm which asserts that such a norm has as its source an effective and identifiable will. The fundamental norm has no such source. Moreover it sets itself up as being above any original constitution and is thus a fiction in as much as it functions only ‘as if’ it is valid.⁵⁶ Hart’s rule of recognition appears to avoid this ‘as if’ function in as much as it is a hypothesis and not a supposition ; the rule of recognition is a term that applies to existing fact. The rule is there, within Hart’s hierarchy of rules, as an endlessly living fact which continually validates, from the viewpoint of descriptive epistemology , the concept of law. Hart would no doubt have argued that there is nothing ‘as if’ about the rule of recognition since it is not a supposition.

Moreover this continuity of the rule of recognition would prevent it, following Vaihinger, from being a pure or authentic fiction. An authentic fiction is one which is annulled the moment that it has served its purpose, that is to say the moment it has achieved the correct result through its use.⁵⁷ Bouriau gives an example. A father decides to leave his camels to his three sons. The oldest is to get at least half of them; the middle son is to inherit at least one third; and the youngest is to be left at least one ninth. The problem is that the father leaves them only seventeen camels. The problem can be solved, notes Bouriau , by acting as if there are 18 camels: nine can be left to the oldest son; six to the middle son; and two to the youngest. The total, of course, comes to 17, but the problem of the inheritance is solved by the fiction that there are 18, a fiction that is immediately annulled the moment that the distribution figures are determined.⁵⁸ One difficulty, then, with Kelsen’s use of the term fiction is that it does not fit the Vaihinger test since the Grundnorm is not annulled the moment that Kelsen’s pyramid is established. In answer to this problem, Bouriau suggests that a distinction needs to be made between fictions employed for knowledge of reality—in which the Vaihinger test applies—and fictions used in transcendental constructions where the test is not vital.⁵⁹ The Vaihinger requirement “does not apply to the fundamental norm since it does not attach to the thing itself, but to the very process of thinking (pensabilité) about a coherent legal practice.”⁶⁰

According to this logic, Hart’s rule of recognition cannot be an authentic fiction because it attaches to reality and is not annulled the moment the model of rules is constructed. Yet, when one compares the rule of recognition with the Grundnorm , the question arises as to whether it is sufficient for Hart simply to declare that it attaches to reality. Hart certainly offers no serious empirical research to support his assertion and seems, subsequently, to admit that it is something of a presupposition.⁶¹ Hart would no doubt point to his claim that ‘assuming the validity’ (Kelsen) is to be distinguished from ‘presupposing the existence’ (of the rule of recognition) . But simply to say that a given notion exists empirically is, as already suggested, not enough to lift it out of the ‘as if’ category. What Hart seems to be saying is that one should proceed on the basis as if there is a rule of recognition existing as a social fact and that it is enough that those acting within the legal model act as if (that is presuppose) its existence. If challenged, then one would have to establish its existence “by appeal to facts, i.e. to the actual practice of the courts and officials of the system when identifying the law which they are to apply.”⁶² Yet can this ‘thing’ be found to exist as a reality in the same way that black-holes in space, once presupposed, can be found to exist? Might not the facts reveal only a group of actors acting ‘as if’ there is a rule of recognition? What if the empirical research indicates that the actors—say French or German public lawyers—are acting ‘as if’ they are operating within a Kelsenian pyramid of norms whose validity is founded on some basic constitutional norm? Does the Grundnorm then become transformed into a rule of recognition? It seems arguable that the moment a legal theorist appeals to the idea of presupposition, he or she is entering the world of ‘as if’—of fiction theory.

Kelsen, Hart and Dworkin are not the only legal theorists of course. What of the legal realists : can their theories be viewed in terms of fiction theory? At first glance American Realism appears as a clear reaction against ‘transcendental nonsense’ —that is to say against transcendental formalism in law.⁶³ Yet the whole realist movement was to an extent triggered by a fictional creation, namely Holmes’ Bad Man , who does not care about conceptual and transcendental constructions but wishes only to know what the courts are likely to do as a matter of fact.⁶⁴ Of course Holmes was not asserting this fictional figure as some kind of validating supposition for a theoretical construct. In fact the figure is employed as a means of deconstruction—a revolt against formalism, as has been noted. But in order to be effective the Realist school, and those schools that followed it such as Critical Legal Studies, need to create a formalist image of law in order that they have an object that can be criticised. For example Duncan Kennedy’s essay on the ideological content of legal education devotes much space to describing what he feels is the typical activity in an American law school.⁶⁵ Can this description, undoubtedly set out in good faith and arguably creating an image that many law graduates would describe as completely accurate or true, nevertheless be described as fiction? Is Kennedy proceeding on an ‘as if’ basis? The point to be made here is not whether Kennedy’s model of the law school is inaccurate; indeed it may be the model that most closely corresponds to the typical American law school. The point is that Kennedy’s model is just one amongst several possible constructions.⁶⁶ Kennedy is creating a fictional (as if) construction of a typical law school in order to achieve a particular practical purpose, namely a critique of law schools. All discourse on legal theory, in short, is about trading in fictions in the sense that to be critical it is necessary to create an object—an intellectual construction—that can act as the thing to be criticised. Movement within any school of jurisprudence is, it might be said, entirely dependent on an ‘as if’ point de départ.

One of the principal contributions made by the realist and critical schools is the way in which they deconstruct traditional legal discourse in turn provoking questions about the nature and reliability of legal knowledge. Nevertheless at the abstract level of legal theory itself such schools tend to be presented as being independent of the other jurisprudential schools such as positivism, natural law , sociological jurisprudence and so on. Realism is just one school of thought amongst many, each, perhaps, having equal validity in its own right. At lower levels of abstraction the position becomes more complex because such schools of thought are often eclipsed behind the concepts and categories employed both by practitioners and by many academics. Of course there are critical contract courses, philosophy of tort or property modules and the like. But a glance at standard textbooks suggests that a largely positivistic approach—often called a ‘black-letter’ approach—to legal knowledge remains dominant.⁶⁷ In fairness , on closer inspection, the position may prove to be more complex. A seemingly black letter work on contract may actually reveal influences from the law and economics theorists while a work on tort might display much realist learning. After all, as William Twining once observed, “we are all realists now”.⁶⁸

Given what might be called this lower level of legal thought—the level of concepts, categories and principles—the question arises as to whether, and if so how, fiction theory is relevant. Again, of course, one is forced to ask what one means by fiction, but if an ‘as if’ approach is adopted as an ‘epistemic attitude’ ⁶⁹ it will at least permit one to think about legal concepts and categories and the extent to which they accord with socially real objects such as humans and physical things. Rather than ask, then, whether conceptual constructions and categorisations are fictions in themselves—a question admittedly not to be avoided—one can ask instead the extent to which fictions intervene in the legal perception of social reality . Two seemingly obvious fictions are of course the legal person and the intangible thing, the former certainly attracting the label from the late medieval jurist Baldus (persona ficta)⁷⁰ and from the modern English judiciary.⁷¹ They intervene in what has been labelled the institutional system of legal thought⁷² by way of extending the notion of a person and a thing to intangible forms. They are ‘as if’ institutions which, unlike the human person or the item of physical property, have no tangible counterpart in the real world.⁷³

The idea that corporations and intangible property were fictions does not seem to find direct expression in Roman law itself. However the Romans made great use of an ‘as if’ attitude.⁷⁴ Gaius appears to suggest that towns are sometimes treated as if they are a private person,⁷⁵ while another jurist states that an inheritance is an ‘as if’ (fungitur) person in the same way as (sicuti) a municipium is treated as if a human.⁷⁶ Acts and deeds by an imperial procurator are treated as if they were the emperor’s own acts;⁷⁷ adoption is based on the notion that the child is ‘as if’ (quasi) it was a natural born son;⁷⁸ a person who is not an heir might be treated ‘as if’ (quasi) he was an heir in certain circumstances;⁷⁹ and so on and so forth.⁸⁰ This ‘as if’ (quasi) attitude is also found at a more formal level in the two quasi categories that act as sources of obligations , namely quasi-contracts and quasi-delicts . The quasi-contracts were based on situations ‘as if’ a contract had been transacted and so, for example, one who receives a mistaken payment will be liable ‘as if’ the money was received under a contract for a loan.⁸¹ Similarly quasi-delicts occurred in situations where there was no delict as such but the facts dictated that an obligation ought to be imposed as if there had been an actionable wrong. Thus a judge who made a case his own, while not having committed any actual wrong, was to be treated ‘as if’ he had acted wrongfully.⁸²

Can these ‘as if’ situations be regarded as fictions? Peter Birks thought that the quasi categories were “a response to the challenge of the residual miscellany” and that the “elegant symmetry” of the fourfold classification of obligations into contract, delict, quasi-contract and quasi-delict “deceives”.⁸³ His argument was based on the assertion that legal rights have “causative events” which bring them into being.⁸⁴ The categories of contract and delict were directly based upon the causative events of contractual agreements and wrongs but the two ‘as if’ categories had no such underpinning. Thus, said Peter Birks, they were “only a variation upon the theme that all birds must be either pigeons or sparrows” and “that all those which are neither pigeons nor sparrows must be counted either as though they were pigeons or as though they were sparrows.”⁸⁵ Professor Birks no doubt thought that he was asserting an epistemological truth, but his analogy between causative events and bird species displays as much of an ‘as if’ attitude as the quasi categories themselves. Causative events are being treated as if they are an external reality like birds which in fact they are not. Causation, as Christophe Bouriau points out, was not a reality at all but an “analogical fiction” based on our experience of a voluntary act in which we ourselves appear as the author (that is to say cause) of our actions on the world (effect). What we have done is to extend this relation to things outside of us in conceiving certain of these as the ‘cause’ of certain events.⁸⁶ The idea that cause is to be found in things themselves and that these things have a “mysterious power” to produce such and such effect is an error. For in reality an indefinite number of factors contribute to the production of the smallest event. In short, in isolating just one cause for an event we are abusively simplifying the complexity of reality.⁸⁷

Even if one were to take issue with this general assertion about cause, there is no denying that it certainly applies to the notion of a ‘causative event’ . As one critic of Birks’ thesis says, classification “by causative event presupposes … that we can identify from [a] list of necessary factual elements one, or perhaps some combination, which can be regarded as the causative event, with the other facts to be considered in some sense ancillary, or as forming the background to, the ‘true’ causative event.”⁸⁸ And of course in reality events “do not come neatly packaged into pre-established categories.”⁸⁹ The ‘quasi’ categories of Roman law might well, then, be ‘as if’ notions but as classificatory devices they are no more and no less of a fiction than Professor Birks’ idea of a causative event. Indeed to claim that a causative event in law is an object to be compared with a specific bird displays a misunderstanding of classification in law as compared to classification in zoological science. The latter is founded on the identification of specific and identifiable traits in an existing tangible object while the former is a scheme that largely comprises of objects created by the classificatory scheme itself. Now whether or not fiction theory applies to scientific classification is a question, interesting as it is, that cannot be determined here.⁹⁰ However with respect to the correlation between causative event and legal category it is difficult to escape from the conclusion that both these notions are the result of a powerful imagination rather than of an accurate depiction of social reality . Indeed another commentator has recounted how Professor Birks was very fond of using storytelling—fiction—in his legal writings and how these stories have helped construct a law of restitution.⁹¹

It is arguable that all discourse about taxonomy in law is little more than a form of storytelling if one includes within the idea of storytelling the notion of ‘as if’. Take the category of ‘tort’ or ‘delict’ founded in turn on the idea of a wrong. Such a wrong has been defined “as the violation of a right, and therefore a breach of duty.”⁹² Now the author who asserts this definition goes on to define each of these three terms, as one might expect, and produces a model in which the three terms are interdependent. Thus a right is defined entirely in relation to a duty and a wrong is defined entirely in relation to a right.⁹³