International Journal for the Semiotics of Law W / 1 2 [1991]
EPISTEMOLOGY AND LEGAL INSTITUTIONS
by GEOFFREY SAMUEL Department of Law, Lancaster University
The notion of a legal institution cannot be understood without the provision of an epistemological context within which the notion can gain its meaning. It is the purpose of this article, after a short introduction (Part I), first to discuss this epistemological context (Part II) and then to propose a model of legal institutions which will allow them to be seen in a context that is free from the traditional assumption that legal knowledge is rule-based (Part III). This discussion will be followed by an examination of the relationship between science, epistemology and ideology (Part W): and by w a y of conclusion (Part V) it will be proposed that in order to understand the notion of a legal institution it is necessary to have an awareness not only of the intellectual history of legal science and legal reasoning~ but also of the history of scientific thought in general. It is in the tension between a science and the object of a science that the notion of a legal institution can be properly understood.
Various writers have m a d e the point that a history of Western legal theory invariably ends up at the Corpus Iuris Civilis, 1 a n d such an observation is of particular relevance when it comes to a theory of legal institutions. For not only did the Romans supply the w o r d itself ( i n s t i t u t i o n e s ) - used in respect of systematically arranged student t e x t b o o k s - but they furnished an epistemological context which still acts as the basis for legal rationality in
1 See, e.g., J.W. Jones, Historical Introduction to the Theory of Law (Oxford: Oxford UP, 1940), vii (Preface): M. Villey, Le droit romain (Paris: Presses Universitaires de France, 7e ~l., 1979~, 5-6.
310
GEOFFREY SAMUEL
Civilian, if not C o m m o n law, 2 families. "'Omne autem ius quo utimur," said Gaius in his Institutiones, "vel ad personas pertinet vel ad res vel ad actiones. "3 The importance of this three-fold division goes b e y o n d mere convenience in the classification of private law rules. 4 The institutional plan devised b y Gaius p r o v e d to be an epistemological construct in that it had the ability to transcend the Roman legal system and to act as a set of rational relations from which other collections of norms and principles could be subsumed and controlled. 5 Canon law, feudal law, mercantile law were all transformed from a mass of descriptive texts and propositions into rational systems formulated in terms of relationships between personae, res and actiones, these three focal points acting both as elements in a system and as concrete institutions in a social world. And it is in this ability to function at one and the same time as abstract elements and concrete institutions that the epistemologist will find the key to legal knowledge: Gaius had provided a theoretical structure, whether he was aware of it or not, 6 that was sufficiently distanced from the facts of society to be able to function as a scientific rationalisation. 7 Law was no longer about descriptive subjects and objects, about specific people and specific kinds of things; it was about institutional relationships (iura) between legal institutions. And this in effect meant that the ontological basis of law had shifted from the physical world itself to a world of legal institutions and institutional relations whose existence was quite independent both of social reality and the positive rules of law that go to make up any parti-
2 Cf. G. Samuel, "Roman Law and Modern Capitalism", Legal Studies 4 (1984), 185. 3 All law that we use relates to persons, to things or to actions: Gaius, Inst., 1.8. 4 H, F. Jolowicz, Roman Foundations of Modern Law (Oxford: Oxford University Press, 1957), 61-81; P. Stein, Legal Institutions (London: Butterworths, 1984), 125-129. 5 ViUey, supra n.1, at 44-45. 6 Cf. Stein, supra n.4, at 127-128. 7 Cf. G. Bachelard, La formation de l'esprit scientifique (Paris: Vrin, 1938), 40-44.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
311
cular legal system at any particular moment in time. 8 The institutional system was, at one and the same time, an ontology in the sense that its institutions acted as the "social realities" around which rules could affix themselves and a science in the sense that the system could be used to interpret social facts. It is at this epistemological level that Roman law has made its most enduring contribution to Western legal science. It may be that the nation states of Europe ended up with their o w n national codes, but the institutional system u p o n which these codes are based is universal in its rationality. Moreover this rationality is b y no means static in its internal structure; the institutions and the relations between them are capable of evolving and of reorganising in w a y s that allow not only for changes in social reality, but for creative developments in the abstract relations themselves. N e w forms of property such as intellectual property are as much due to the creative possibilities of the institutional system itself as to changes in the economic and social relations. Moreover the institutional m o d e l had, and has, the ability to reorganise its internal relations in such a w a y that it could, particularly during the second life of Roman law from the late Middle Ages onwards, pass from one degree of rationality to another. 9 Roman law started its existence as a descriptive and inductive set of interpretative techniques and ended up as a deductive and, later, axiomatic science. I° Of course this axiomatisation, apart from its element of wishful thinking, was, and is, not necessarily a benefit. For just as intellectual immobilism can result from a too close adhesion between a science and its object, between facts and reason, so also can too great a distance give rise to an epistemological obstacle, n Accordingly the axiomatisation of legal science in the Civilian family has brought its o w n problems when it comes to an understanding of legal knowledge. ~2 8 J. Ellul, Histoire des institutions: 3 - Le Moyen Age (Paris: Presses Universitaires de France, 9" cScl.,1982), 27. 9 Cf. R. Blanch(~, L'induction scientifique et les lois naturelles (Paris: Presses Universitaires de France, 1975), 152. 10 R. Blanche, Le raisonnement (Paris: Presses Universitaires de France, 1973), 219-220. 11 Bachelard, supra n.7, at 44, 55-72. 12 See generally C. Perelman, Le raisonnable et le ddraisonnable en droit
312
GEOFFREY SAMUEL
Axiomatised knowledge is propositional knowledge; and the logical temptation that arises is to see legal institutions as being the product of such propositional knowledge. It is this logic, a logic that functions essentially in a two-dimensional world, that is now the epistemological obstacle. II When one turns to English law, the device of a legal institution, in the Continental sense of the term, is something that is largely foreign to the C o m m o n Lawyer and, given the absence of both a Roman law and a university tradition underpinning English legal scholarship, this lack of familiarity ought not to surprise. 13 Nevertheless even if one were to accept that the Institutes have played a relatively minor role in the long development of English legal thought, 14 it does not actually follow from this historical premiss that the epistemological lessons to emerge from Roman legal scholarship have no relevance whatsoever for C o m m o n Lawyers. Indeed the vagaries of history m a y be such that it is only within the Common Law that the full epistemological significance of the institutions of law can be fully appreciated: for the Civil Law has, in terms of a scientific discourse, moved to a stage that is beyond both the Classical Roman law and the m o d e r n common law. is Consequently the Common law might be seen as being an example of a reasoning system, if not actually paralleling that of Classical Rome, or the medieval Roman lawyers, at least existing on a similar plane of rationality. The C o m m o n law seems to share the same interpretative distance from the facts as the model to be found in the writings of Ulpian, Paul, Bartolus and the like in that both models (Paris: LGDJ, 1984). 13 For a discussion of legal institutions from the position of a Civil Lawyer see J.L Bergel, ThSorie g~n&ale du droit (Paris: Dalloz, 2e &i., 1989), §§164179. 14 Cf. P. Birks & G. McLeod, Justinian's Institutes (London: Duckworth, 1987), 23-26. 15 K. Zweigert & H. Kotz (trans. Tony Weir), An Introduction to Comparative Law" Volume I (Oxford: Oxford University Press, 2nd ed., 1987), 193194; E. Agostini, Droit compar~ (Paris: Presses Universitaires de France. 1988), ~119-120.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
313
seem w a r y of the idea that knowledge of law is just a matter of learning propositions. 16 And although the Common law is now more ambiguous on this question, 17 it is unlikely that the English jurist would subscribe to the view that the technique and method in the caselaw is nothing more than the application of a pre-existing and formalised system of rules capable of responding) in advance, to any question likely to arise. Perhaps the main difference, then, between the ancient and the modern Civil Lawyer is that the latter would see legal institutions as the product of legal propositions, whereas the former would see legal p r o p o s i t i o n s - that is to say rules and p r i n c i p l e s - - as the product of the interplay between legal institutions. N o w the point of this observation is that if one takes a legal institution to be a social reality around which rules group themselves, 18 the question of the relationship between the social and legal worlds hinges not just upon whether the notion of a legal institution is seen as having its foundation either in the discourse of law or in some empirical social science system, but whether the institution is something separate, epistemologically, from the rules which surround and affix themselves to it. In other words, is knowledge of law simply a matter of learning and applying, via the syllogism, a system of propositions? Or does it consist of a learning process that is much more complex in the cognitive sense? If legal knowledge is propositional, then this raises some a w k w a r d difficulties w h e n it comes to the application of legal rules because these rules have to have i n s t i t u t i o n s - legal subjects, legal objects or legal a c t i o n s - - to which they can attach themselves. And if these institutions are equally the product of legal propositions, one is constructing a system in which the elements and the communication relations become confused. One w a y around this might, seemingly, be to say that rules operate at two levels: those that describe the institutions and those that describe the r e l a t i o n s - or, following Hart, a system of pri16 Digest 50.17.1; P. Stein, Regulae Iuris (Edinburgh: Edinburgh University Press. 1966), 154) M. Villey, La formation de la pens~e juridique moderne (Paris: Monchrestien, 4e e~i., 1975), 523-524, 700. 17 See, e.g., Lord Simon in F A & A B Ltd v. Lupton [1972] A.C. 634, 658-659. 18 Bergel, op.cit., §§166-170; P. Roubier, Th~orie g~ndrale du droit (Paris: Sirey, 2e &i., 1951), 17-23.
314
GEOFFREY SAMUEL
mary and secondary rules perhapsJ 9 But in truth the reduction of law to propositions, even to sets of propositions operating at different levels, still produces a two-dimensional knowledge system that fails to explain h o w the actual interplay of the institutions themselves is an extraordinarily c o m p l e x - and c r e a t i v e - feature of the law. To find an example of this creativity one need look no further than the notion of debt. This is not only an institution in two quite different w a y s - that is to say it is an actio and a r e s (or a chose in action) m but also an obligatio in the sense that at one and the same time it functions at the primary and secondary levels of the law of contract; 2° each time therefore that a court orders one legal subject to p a y another a contractual debt, the interplay b e t w e e n legal subjects, legal objects and legal actions is such that the complexity cannot be reduced to a set of logically consistent rule propositions because the institutional aspects of debt cuts across the legal relational aspects. Debt becomes both a "thing" to which can be attached an in rem relation and an in personam relation in itself. 21 It becomes, in other words, a cause of action and a part of the c l a i m a n t ' s patrimony at one and the same time. 22 In turn, this complexity can then be drawn upon to construct sets of creative institutional and relational reasoning systems which might actually defy the logic of the standard dictotomy of real and personal rights, 23 a logic that is in fact of fundamental importance in a world where the difference between real and personal security is paran~unt. 24 N o w C o m m o n lawyers often pride themselves on their lack of
19 H.L.A. Hart, The Concept of Law (Oxford: Oxford University Press, 1961), 77-96. 20 On which see Lord Diplock in Moschi v. Lep Air Services Ltd [1973] A.C. 331,347 and Photo Production Ltd v. Securicor [1980] A.C. 827, 848-849. 21 See generally on this point S. Ginossar, Droit r~el, propri~t~ et crdance (Paris: LGDJ, 1960); see also Lipkin Gorman v. Karpnale Ltd. [1991] 3 W.L.R. 10, 29. 22 See Overstone Ltd v. Shipway [1962] 1 W.L.R. 117 and Sinclair v. Brougham [1914] A.C. 398. 23 For example the remedy of tracing: see Aluminium Industrie v . Romalpa Aluminium [1976] 2 All E.R. 552. 24 Ginossar, supra n.21, at 182.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
315
logic and there are legal philosophers from the C o m m o n law tradition who fight hard for the rejection of the Hartian view of legal positivism. However, for the epistemologist, the issue of law as a set of propositions remains a live one because it lies at the heart of any understanding about the methods and techniques of jurists. In this respect the system of rules controversy completely transcends the one about sources of law in that some of the anti-positivists are themselves subscribing to an epistemology that describes law primarily as a matter of propositions. Thus Dworkin's rules and principles foundation from which the good judge derives a person's rights 2s is no less a set of propositions than Hart's union of primary and secondary rules; and even if rules are supplemented by propositions which do not dictate, but only argue in the direction of, a particular decision, 26 these latter remain nevertheless attempts to present knowledge in propositional form. When lawyers reason, they are applying, as Dworkin himself says, propositions at a high level of abstraction which embraces "principles, policies and other sorts of standards"; 27 and while these m a y not be "rules" as such, they remain, from the point of view of the epistemologist, knowledge in a rule-based form. 28 Indeed even legal philosophers argue that Dworkin provides what is in effect a rule-based account of reasoning. 29 In appreciating the role of a legal institution there is therefore not necessarily much help to be found in the writings of those who reject legal positivism as an epistemology. Instead it m a y be more profitable to focus upon the institution as an element in a rather different system than the one envisaged by the positivists. And it is here that legal h i s t o r y - the history of Roman legal scholars h i p - has a central role to play. The role is not, however, one of legal history as an end in itself; rather it is a question of using legal 25 R. Dworkin, Taking Rights Seriously (London: Duckworth, 1977), 14-45, 81-130. 26 Ibid., p.26. 27 Ibid., p.22. 28 R. Susskind, Expert Systems in Law (Oxford: Oxford University Press, 1987), 169-174. 29 N. MacCormick, Legal Right and Social Democracy (Oxford: Oxford University Press, 1982), 140.
316
GEOFFREY SAMUEL
history as a means of discerning not only the diverse elements which have gradually contributed to the construction of a legal science,3° but the effects on this science of the w a y this knowledge has been represented. Thus it is a question of going back to the sources and to earlier forms of rationality utilising both la m~thode r~currente and la m~thode progressive 31 in order to show h o w c o n t e m p o r a r y accounts of the cognitive performance of lawyers are limited. The advantage of these epistemological methods is that they can show that Roman law is equally explainable in terms of a network of interrelating systems as in terms of a two-dimensional hierarchy of positive rules 32 and this, in turn, might provide a w a y around the "epistemological obstacle", to use Bachelard's notion, 33 of m o d e m positivism. That positivism is an obstacle to an understanding of the cognitive performance of lawyers seems to be evidenced by the "brittleness" of m a n y expert systems founded on rule-based epistemologies; such systems can capture the details of legal phenomena only a w k w a r d l y and usually by having a separate rule for each exception to the rule. 34 In other words, rule-based systems of legal knowledge seem to "lack commonsense knowledge and reasoning they do not ' k n o w ' their own limitations"; they "are insensitive to context and are likely to give quite incorrect responses to queries that are slightly outside the domains for which they were programmed."3s In addition to the problem of brittleness, a rule-based approach
30 R. Blanche, LMpist~mologie (Paris: Presses Universitaires de France, 3e &i., 1983), 36-39. 31 H. Barreau, L'dpistdmologie (Paris: Presses Universitaires de France, 1990), 14-15. 32 G. Samuel, "The Challenge of Artificial Intelligence: Can Roman Law Help Us Discover Whether Law Is a System of Rules?", Legal Studies 11 (1991), 24. 33 Bachelard, supra n.7, at 13-22. 34 W. Bechtel & A. Abrahamsen, Connectionism and the Human Mind: An Introduction to Parallel Processing in Networks (Oxford and Cambridge: Blackwell, 1991), 17, 227, 261. 35 G. N. Reeke Jr. & G. M. Edelman, "Real Brains and Artificial Intelligence", in S. R. Graubard, ed., The Artificial Intelligence Debate (Cambridge MA & London: MIT Press, 1989), 149.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
317
to legal expertise fails to take proper account of the role of analogy in legal reasoning. It is all very well to talk of extending an existing rule by analogy, 36 or even to assert that analogy itself is based on principle; 37 but this does not offer any insight into the actual process of analogy in law. Take the case of Southport Corporation v. Esso Petroleum. 38 In order to protect the lives of the crew of a stranded Esso tanker, the captain ordered the discharge of its oil so as to lighten the ship; the oil ended up on the Corporation's beach and, having spent much money on clearing it up, the Corporation brought against Esso an action for damages founded on the torts of trespass, nuisance and negligence of the captain. At the trial it was found that the captain had not been negligent and Devlin J., whose decision was restored by the House of Lords, held, on an analogy with d a m a g e caused to adjacent property by road vehicles, that the plaintiffs could only recover damages if fault was established. As there was no fault proved there could be no liability. N o w the institutional question that arises here centres on the role of analogy. Why did Devlin J. choose to draw an analogy between the facts of the case before him and the traffic accident cases rather than with a nuisance case such as Benjamin v. Storr. 39 In this latter case the owner of some horse-drawn vehicles was held liable without proof of fault to an adjacent caf~ owner for economic loss caused by pollution emanating from the horses. And surely this latter decision provided a much closer relationship to the pollution events in Esso in terms of visual image, as indeed the Court of Appeal in Esso appeared to accept? Certainly, rules go some way in explaining part of the process of reasoning and decision. Thus it is possible to say that Devlin J. derived his analogy from some obiter dicta of Blackburn J. But to argue that the decision of the courts in the Esso case is simply a matter of deriving a conclusion from a set of propositions is surely to simplify the interplay of institutions within the facts of the authorities and the case itself. 36 See, e.g., Lord Simon in F A & A B Ltd v. Lupton [1972] A.C. 634, 658-659. 37 N. MacCormick, Legal Reasoning and Legal Theory (Oxford: Oxford University Press, 1978), 186. 38 [1952]2 All E.R. 1204 (Devlin J.); [1954] 2 Q.B. 182 (Court of Appeal); [1956] A.C. 218 (House of Lords). 39 (1874)L.R. 9 C.P. 400.
318
GEOFFREY SAMUEL
In order to arrive at the decision, it was first necessary to translate the facts of the previous authorities and of the case in hand into sets of institutional relationships between legal subjects (Esso, ship's captain, Southport Corporation), legal objects (ship, oil, beach) and legal actions (trespass, negligence, public nuisance) and then apply notions of causation, control, blame, possession, ownership and the like to particular patterns of these institutions. Once this is done it soon becomes evident that different patterns can produce different rules. Thus, for example, emphasising the proprietary relationship between Esso (legal subject) and ship (legal object) could well have given rise to a set of rules quite different from, say, emphasising the behaviour of the captain in relation to his crew (legal subjects) and stricken ship (legal object). Equally, emphasising an institutional relationship between Esso (legal subject) and oil (legal object) may have shifted attention off the actual behaviour of the individuals involved and onto a risk liability founded u p o n the commercial status of Esso vis-a-vis the public interest status of the local authority. In other words, institutional changes of emphasis may lead to quite different decisions as a result, not just of the application of a different rule, but of changes in cognitive patterns producing quite different models of potential liability. One might add that even in the area of criminal law w perhaps the most rule-orientated in an ideological, if not epistemological, sense (nulla poena sine l e g e ) ~ the institutional pattern can on occasions be fundamental to the question of liability. In one case involving a large supermarket company accused of giving inaccurate or inadequate information to a consumer, the company managed to convince the House of Lords that their employee who had allowed this misdescription to happen was "another person" from themselves and that, accordingly, his act was not "their" act. 4° In doing this, the supermarket company escaped conviction under the Trade Descriptions Act 1968 because they could take advantage of the "act or default of another person" defence provided by section 24. Now no doubt it could be said that all that the House of Lords did was to interpret and to apply a rule. But in truth it is the institutional pattern accepted by the judges that pre-determines whether section 40 Tesco Supermarkets Ltd v. Nattrass [1972] A.C. 153~
EPISTEMOLOGY AND LEGAL INSTITUTIONS
319
24 will be relevant or not. In accepting a scenario involving two legal subjects rather than a single persona the judges had effectively decided the outcome of the criminal prosecution through the manipulation not just of the facts and the personae, but of the legal remedy (criminal proceedings) as well. The institutional relationship between "'actio" (criminal proceedings) and employee eclipsed all the other institutions and institutional relationships (iura) such as those between employee and company (usually important in vicarious liability cases), consumer and supplier (consumer protection via private law) and state and citizen (consumer protection via public law). Was section 24 ever in itself capable of containing this knowledge? It is in the understanding of this cognitive process that the role of the legal institution becomes of central importance. Rules can be neither described nor applied without such institutions to which they can affix themselves and once it is appreciated that given configurations of institutions can produce different patterns of legal relationships, which in turn can attract different kinds of rules, then the easier it becomes to see that the cognitive processes of legal reasoning are, at least partly, to be found in the institutional analysis of factual situations. Take another example. A agrees with B that B shall provide a security patrol to help to protect A's factory against thieves and arsonists; B's employee sent to patrol the factory turns out to be an arsonist and succeeds in burning down A's factory.41 Can B claim (like the supermarket company discussed earlier) that this act was not "their" act but the act of a third person, the patrolman? Now it is of course possible to find a "rule" to govern this situation; 42 yet if one were to apply the course of employment test of vicarious liability to this set of facts, this would be to assume an institutional pattern that is not itself ruledetermined, although it might be rule described. What is it in the facts that determines that they should be seen in terms of three rather than two legal subjects? No doubt one could point to a rule that attempts to distinguish between the status of different kinds of employees within the hierarchy of a legal person; 43 but such a rule 41 See Photo Production Ltd v. Securicor [1980] A.C. 827. 42 See Morris v. C W Martin & Sons Ltd [1966] 1 Q.B. 716. 43 See Tesco Supermarkets v. Nattrass [1972] A.C. 153.
320
GEOFFREY SAMUEL
does not, and cannot, of itself explain w h y the security company, leaving aside the exclusion clause, ought not to be liable as a legal subject in itself for a breach of duty to the owner of the factory. The determination of liability in such a problem involves something more than the application of a rule. It involves the translation of the facts into an institutional i m a g e - into three rather than two legal s u b j e c t s - and there is no evidence that this translation is a cognitive process determined simply by the application of a major to a minor premiss. Indeed, given that the process involves a mental image of a factual situation and given that facts function in a three rather than a two-dimensional world, it would seem logical that any mental representation of legal institutions ought to be one that involves a three rather than two-dimensional representational model. That is to say, it should involve relations which not only move vertically and horizontally, but move cross-laterally as well. III The moment one moves from the idea of a two dimensional model to one that operates in three dimensions, one is entering a world where the complexities of the social facts can be represented with a much greater degree of precision and sophistication. Moreover, a three-dimensional model allows for the incorporation of the legal action as a legal institution. This third institution, lost w h e n legal science moved during the Enlightenment from the inductive to the deductive stage, 44 not only can generate its own particular concepts in relation to legal subjects, but can act, as we have seen with respect to debt, as a legal object in itself. This means that it is often possible to move relatively freely through different kinds of right without, s e e m i n g l y , giving rise to logical contradiction, at least not at the micro-level. Thus in English law the legal action as an institution can achieve the objectives of a law of restitution (unjust enrichment) simply by aiming and attaching itself to legal objects (a sum of m o n e y for instance) without having to concern itself with the possibility of contradiction between real and personal rights or,
44 H.F. Jolowicz, supra n.4, at 75-81; G. Samuel, "Historical Jurisprudence and the New Rationalism", in T O'Hagan, ed., Revolution and Englightenment in Europe (Aberdeen: Aberdeen University Press, 1991), 139-150.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
321
perhaps, between various personal rights. 45 Exactly the same might be said of the problem of good faith (bona tides) or abuse of rights: by functioning at the level of injunctions, estoppel, debt, damages or some other remedy, it becomes possible to operate "in one bit of the l a w " free from having to construct "some aesthetic or didactic structure "46 of rights and their abuse. Sometimes, it must be said, certain judges can go too far, as has happened in the case of the injunction which the Court of Appeal thought could operate freely within the law of actions without reference to the law of things. 4z But what is interesting about this saga of the injunction is that it arose only because of the drafting of section 37(1) of the Supreme Court Act 1981 : this section uses neither the legal subject nor a legal object as the institution around which the rule is framed. The section is an excellent example of a rule specifically attaching to a legal action. The section 37 injunction cases also illustrate the great advantage of a three-dimensional model to represent legal knowledge. When is a person entitled to an injunction? Before 1981 it was generally thought that the equitable r e m e d y would be available only w h e n the plaintiff's rights had been infringed or were threatened. 48 In other words the remedy was dependent upon legal relations flowing between legal subject and legal subject (obligations or iura in personam) or between legal subject and legal object (iura in rem). Once, however, a rule had been formulated which used the r e m e d y itself as an institutional focal point, the whole emphasis changed: it became possible to view problems just in terms of the relationship between legal subject and legal action and this had the effect of reducing the protection of "rights" to the protection of "interests". 49 All that a plaintiff had to do in order to obtain the r e m e d y was to show that he had a " s u f f i c i e n t " or "legitimate interest" in the r e m e d y being issued. The House of Lords has now stepped in to
45 See, e.g., Sinclair v. Brougham [1914] A.C. 398. 46 B.S. Jackson, Semiotics and Legal Theory (London: Routledge and Kegan Paul, 1985), 189. 47 Chief Constable of Kent v. V [1983] Q.B. 34. 48 The Siskina [1979] A.C. 210. 49 p v. Liverpool Post plc [1990] 2 W.L.R. 494, 506-507 (C.A.).
322
GEOFFREY SAMUEL
restore the "rights" position; 5° however the section serves to remind lawyers not just about the importance of a legal action as an institution in the construction and drafting of legal rules, but about the ability of this institution to organise facts and to formulate its o w n network of legal connections and relations between the various institutions. In a two-dimensional world, these connections can often eclipse each other with the result that concepts such as rights and interests become increasingly confused. 51 However the organising and connecting potential of persona, res and actiones is thrown into a m u c h sharper relief and the actual structural basis of certain concepts becomes very much clearer once one moves into a third dimension. Thus the notion of a "right", an "interest" and a "'duty" are put into an epistemological context which can show, historically, how they are concepts that have roots in quite different institutional relations. The notion of a right is founded in the relationship between legal subject and legal object (a right to something;); d u t y owes its existence to the iuris vinculum between legal subject and legal subject (a d u t y to someone); and interest arises largely out of the relationship between legal subject and legal action ("legitimate interest" in a legal action, s2 interests protected by an award of damages). These different historical relations can be described by language propositions but they cannot be manipulated by language because language as a system does not in itself function in this kind of three-dimensional world (although it can describe such a world). The linguistic proposition is not in itself a cognitive m e d i u m capable of envisaging the three-dimensional structural possibilites that can exist at one and the same time between all the legal institution possibilities in a single set of facts; and in turn this means that legal institutions are not and cannot be rule-governed as such because their function is to provide a systems model in which certain mechanisms interrelate in order to create mental abstractions and mental relations capable of modelling the real social world. All that language can do is to 50 p v. Liverpool Post plc [1991] 2 W.L.R. 513, 524-525 (H.L.).
51 See, e.g., IRC v. National Federation of Self-Employed and Small Businesses Ltd [1982] A.C. 617. 52 The Nordglimt [1988] 2 W.L.R. 338, 354; Nouveau code de procddure civile, art. 31; Supreme Court Act 1981, s.31(3).
EPISTEMOLOGY AND LEGAL INSTITUTIONS
323
p r o v i d e the rules and the principles within the m o d e l and, if n e c e s s a r y , to describe the model itself. In other w o r d s the legal epistemologist must distinguish between a rule-governed and a ruledescribed cognitive model of legal reasoning. 53 It is from this epistemological position that Neil MacCormick's theory of legal institutions presents some problems. Certainly one could agree with his view that the idea of an "institution" can be used "to signify the conceptual framework within which particular arrangements can be set up b y particular persons on particular occasions to last through particular periods of time. ''54 H o w e v e r to say that this conceptual framework is "structured by sets of rules ''55 rather drains the conceptual framework of its creative possibilities in that it reduces it to a two-dimensional representational model which is incapable of mirroring the complexity of the real world. It suggests that ownership and c o n t r a c t - two "institutions" used by M a c C o r m i c k 5 a - are all or nothing devices; either a particular transaction is a contract or it is not. In truth, legal reasoning is very much more subtle and involves not just two-dimensional relations between people and things (ownership) or between person and person (contract), but a complex set of interrelations between a variety of institutions. Thus the courts, when solving say a contract problem, m a y well have recourse to institutions and institutional relations like "right", "interest" and "expectations", and it is the complex interplay of all of these institutions and relations that go to make up the pattern or structure of the model. Moreover if this pattern is represented only in two-dimensional hierarchy form, there will be created, as we have already indicated, logical contradictions between the various relations and institutions in that the actual model
53 Bechtel & Abrahamsen, supra n.34, at 227-228. 54 N. MacCormick, "Institutions, Arrangements and Practical Information", Ratio Juris 1 (1988), 73, 76. 55 Ibid. 56 It should be evident that MacCormick's idea of an institution is wider than the one being advanced by the present author; the latter sees ownership and contract more as institutional relations, leaving the term "institution" to cover legal subjects (including the state), legal objects (including rights which may be separate "objects" from corporeal property) and legal actions.
324
GEOFFREY SAMUEL
used to decide any one case need not always conform hierarchically to a pattern used to decide some other case. Indeed the pattern used to decide a particular case might not even conform to a logical theory of law itself and this is w h y debt can be an obligation, a piece of property and a legal r e m e d y at one and the same time. Neil MacCormick seems to be of the view that the given facts are fitted into a rule-described institutional structure. Yet the truth is that the institutional structure is already determined and described within the facts themselves in that legal facts are nothing more than sets of interrelating institutions. Thus in a case like Southport Corporation v, Esso the jurist is not fitting captains, ships, oil and beaches into a rule-determined model: for the jurist is already operating with facts translated into a world of institutions (persons, things and actions) and institutional relations (contract, ownership and the like). This distinction between the physical world of individual people and specific things, and legal world of institutions is of the utmost importance when it comes to understanding legal problemsolving and legal reasoning because it holds the key to how legal reasoning actually functions at the levels of "fact" and "law". In MacCormick's theory it is a question of "our interpreting acts and events in the physical world in terms of rules about the setting up and about the termination of instances of various legal institutions. "sT Yet this overlooks the way in which legal decision-making and legal reasoning actually functions at the level of, or perhaps one should say within, the facts themselves. Take for example a recent contract case. Blackpool Borough Council invited tenders on a form containing clauses stating that the tenders had to be received by a certain date and that the council did not bind themselves to accept any tender. The plaintiff submitted a tender by hand by the specified date, but owing to the council's negligence the postbox was not cleared until after the closing date. The council accepted a tender that was lower than the plaintiff's and the latter successfully brought an action for damages against the former, the Court of Appeal holding that the council was in breach of an implied contractual obligation arising out of the tendering arrangements, ss Now if one approaches this case only from the position of a two-dimen57 MacCormick, supra n.54, at 77.
58 Blackpool & Fylde Aero Club Ltd v. Blackpool BC [1990] 1 W.L.R. 1195.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
325
sional rule model, the decision of the Court of Appeal does not make much sense because the actio cannot be justified by the structure of positive contract rules as they then stood. However, having come close to admitting this, Bingham L.J., who delivered the main judgment, suddenly descended into the facts themselves in order to ask a series of hypothetical questions designed to show that a strict application of the existing positive rules would lead to an abusive result. He then returned to the law to arrive at the conclusion that "if [the plaintiff] submits a conforming tender before the deadline he is entitled, not as a matter of mere expectation but of contractual right, to be sure that his tender will after the deadline be opened and considered. "59 This conclusion had little to do with the positive rules of contract and more to do with an institutional reinterpretation of the facts by the judge. After his conclusion, Bingham L.J. returned to the facts to pose another hypothetical question suggesting that if the club had enquired at the time of the tender whether it could rely upon the tender being properly considered the answer would, according to the judge, have been "of course", One can see here, not a rule-structure, but a fictional conversation between two fictional personae. Two epistemological points need to be emphasised here. The first is that the case is not just a question of subsuming a static set of facts to a static set of rules and principles attaching to the institutional framework of contract. The facts themselves were manipulated and reinterpreted from within so as to produce an image of a legal subject (persona) being entitled (dominium) to something more than a "mere expectation"; the entitlement was to a " r i g h t " - - that is to say, to a legal object (res) in the nature of an abstract thing. Secondly, it needs to be emphasised that this change of institutional image within the facts was mirrored in the institutional model by a change of emphasis, or a restructuring, of the institutional relations within the model itself. As we have just stressed, the plaintiff was entitled to his remedy (actio), not so much because of the obligatio or iuris vinculum, 6° but because of a "property" relation between legal subject and legal object; it was a question of entitlement rather than 59 /b/d at p.1202. 60 The Romans saw an obligation as a iuris vinculum (legal chain) between two legal persons: Justinian, Institutes, 3.13pr.
326
GEOFFREY SAMUEL
duty. This brings us back to the cognitive issue involved in the role of institutions. In a two-dimensional model, it is simply impossible to envisage the plaintiff's action in the Blackpool case as being founded on an institutional relationship which owes more to the law of property than the law of obligations because the plaintiff could not, with regard to the facts in issue, be said to o w n or possess anything. Yet this does not mean that an abstracted form of the relationship between persona and res cannot be used to set up an institutional pattern with respect both to the relationship b e t w e e n person and "thing" and to the relationship between person and actio. By turning a "mere expectation" into a "right" the Court of Appeal established at one and the same time both a substantive relationship analogous to a property entitlement and a remedy relationship analogous to a judicial review action founded on an "interest" (and it should not pass unnoticed that the defendant was a public body). In a three-dimensional model these extra " d i m e n s i o n s " to the legal reasoning can be revealed and be seen to be working alongside any rules and principles from the law of contract. In other w o r d s in a three-dimensional world the institution of a "right" can simply cutacross the traditional logical divide between property and obligations. Indeed, the whole success of the notion of a right has been based on its ability to carry into non-property areas the institutional pattern of ownership; 61 the problem for lawyers has been that they have never had an adequate epistemological framework to explain this phenomenon. IV It is tempting to think that much of what has been said so far is an epistemological truism that was recognised b y the Romans themselves. Of course one has to be careful here because it was, after all, the Romans w h o established the strict d i c h o t o m y b e t w e e n property and obligations. 62 But that said, they certainly appeared to distrust rules and definitions as the means by which one acquired 61 See further G. Samuel, "Epistemology, Propaganda and Roman Law: Some Reflections on the History of the Subjective Right", Journal of Legal History 10 (1989), 161. 62 See Digest 44.7.3pr.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
327
legal knowledge, 63 and this suggests that the classical jurists did not think that the key to legal reasoning was to be found just in the syllogism. But whatever the epistemological sophistication of the jurists themselves, the Corpus Iuris Civilis illustrates very clearly how the legal institutions, and the legal relations between them, function as a genuine system: the institutions and relations interact in such a w a y as to define, preserve and enrich themselves through the play of internal transformations without ever having to go outside the boundaries of the system or to appeal to exterior elements. 64 Thus the interrelation between two legal subjects (contracO could in itself create legal objects (debts) capable of existing as independent institutions in the system; 65 and the interaction of legal action and legal subject could create on occasions another legal subject in that, for example, once towns could bring or face a legal action they were transformed into legal subjects. 66 Now the m o d e r n inheritance of this Roman thinking is that the whole of the today's ius civile of legal personality (persona), patrimony, property (res) and rights (droits subjectifs) can be understood only in terms of an interrelating system. 67 Yet what is important to notice here is that one is not just talking in terms of the Institutes: the creative nature of legal institutions is just as evident in the casuistry of the Digest as it is in the works of Gaius, Justinian and their Enlightenment successors. Thus a close examination of the texts in the Classical sources will soon reveal that however empirical or concrete each problem may at first sight appear, the solution of the problem still involves a cognitive process that translates the facts into a world of personae, res and actiones between which is then constructed abstract relations described by nouns (iura, dominium, possessio and the like) and verbs (debere) which actually make sense only in an a priori model. Consequenfly it is important to realise that however "empirical" the 63 See Digest 50.17.202. 64 Cf. J. Piaget, Le structuralisme (Paris: Presses Universitaires de France, 9~6d., 1987), 6-7. 65 Gaius, Institutes, II.14. 66 See generally P. W. Duff, Personality in Roman Private Law (Cambridge: Cambridge University Press, 1938), 62-94. 67 C. Arias, Th~orie contre arbitraire (Paris: Presses Universitaires de France, 1987), 211.
328
GEOFFREY SAMUEL
science in the Digest, it remains, nevertheless, an abstract science. "Le concret, disait Langevin, c'est de l'abstrait rendu familier par l"usage."68 A n y talk, then, of reducing law to fact is simply to misunderstand law and legal science, For the role of legal institutions is to act as elements in a system which merely mirrors rather than describes the social world of fact.69 Legal subjects and legal objects are not people and things as such, anymore than numbers in mathematics are the things they count; and this remains true even of Gaius' Institutes which, apparently, saw notions such as dorninium as being directly applicable to physical objects themselves. 7° Legal institutions are, as Bernard Jackson observes, elements which need have no relation whatsoever with "brute facts" because "fact and law, behaviour and rule, are essentially distinct categories. "71 Where one might take issue with Jackson is with regard to the shape of the legal system itself. His references to Kelsen imply a systems pattern that is hierarchical or pyramidal in shape and while it is understandable, historically and ideologically, w h y this pattern should be so d o m i n a n t , the great value of the Digest is that it suggests that this kind of pattern is by no means inevitable from an epistemological position. 72 In fact, the hierarchical pattern is itself a result of reactions within the institutional system. It resulted from the gradual development of what today we would call the "state" (princeps) as a legal subject capable of generating its own legal relation (imperium) which could in its turn be used to subjugate the other sources of legal power such as dominium. Omnia principis esse intelligantur. 73 The importance of these internal developments lies in the fact that one begins to see a different pattern emerging from the Corpus Iuris than simply the one of a dichotomy between the teaching and 68 Quoted in R. Blanchd, L'axiomatique (Paris: Presses Universitaires de France, 6~6d., 1980), 104. 69 EUul,supra n.8. 70 A.M. Patault, Introduction historique au droit des biens (Paris: Presses Universitaires de France, 1989), 17. 71 Jackson, supra n.46, at 180. 72 G. Samuel, supra n.32. 73 Code 7.37.3.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
329
practice of law. 74 Doubtless the distinction is as important for legal science as it is for the natural sciences. 75 However, the need for universalist models of legal science is one not just for the teacher: it is also one for the politician keen to capture the institutional model for ideological ends and once this point is understood, the easier it becomes to appreciate that the debate between Dworkin and the positivists is about institutional patterns for political purpose. From the position of the legal institutions themselves, the contribution of Roman law to this debate is to show that, as far as strict legal science is concerned, both sides are equally "correct" because the model of Roman law institutions (persona, res, actiones, princeps, respublica, etc.) can be used to support either thesis depending upon the particular historical pattern chosen. 76 In science itself, one does not always have such a luxury of choice because the role of scientific discourse, traditionally at least, is to describe an object of science. But with legal discourse, the ontological basis of its institutions and relations are no more or less than the discourse itself. The Corpus Iuris, in other words, is both the legal science and the object of the science 77 and as such its o w n h i s t o r y - including the political history reflected in the institutional t r a n s f o r m a t i o n s - is itself part of Roman legal science. Once one moves from the realm of epistemology to that of ideology, the question arises as to how this move affects the perception and the role of legal institutions. The epistemologist wants to theorize the role of institutions in order to discover h o w this science acts as a basis of a b o d y of knowledge, whereas the ideologist is more concerned with the role of institutions in the construction, consolidation and justification of a political system. When it comes to law, h o w e v e r , the dichotomy between epistemology and ideology is particularly difficult because law as a discourse claims, at one and 74 Cf. Jackson, supra n.46, at 188-190. 75 j.p. Astolfi & M. Develay, La didactique des sciences (Paris: Presses Universitaires de France, 1989), 42-56, 114. 76 See further G. Samuel, "Science, Law and History: Historical Jurisprudence and Modern Legal Theory", Northern Ireland Legal Quarterly 41 (1990), 1. 77 On the confusion of law and science of law, see C. Arias, Epistdmologie juridique (Paris: Presses Universitaires de France, 1985), 30-42.
330
GEOFFREY SAMUEL
the same time, to be part of, yet rise above, political discourse. Thus law as an historical phenomenon is both an epistemological and ideological discourse since the history, or histories, of public and private law m and indeed legal t h e o r y - lie at the heart of the science of law both as an object of knowledge and as an ideological discourse capable of supporting certain kinds of political systems. 78 No doubt some w o u l d argue that, as far as law is concerned, even to distinguish between epistemology and ideology is unrealistic. Yet whatever the political role of the courts, and of particular notions such as the subjective right, the fact remains that law is supposed to function as a problem solving discourse and this suggests that it acts as a contextual discourse both for a set of political institutions and for a mass of reasoning techniques. In truth, the two functions become amalgamated once one sees legal discourse as the creation of a rational world of interrelating and interacting institutions quite separate from the real world: for both the political organisation and the social objects which form the basis of disputes (property, patrimonial loss and the like) become elements in a single system. The persona, to put it another way, is at one and the same time a social and a political institution. However this does not mean that the institutional system is merely an ideological construct making epistemological claims. What it means is that the notions of epistemology and of ideology may well have to adjust themselves to a history, and to a philosophy, of la technique.79 That is to say both m a y well have to see themselves as rooted in a particular history that has devoted itself to the production of discourses that think in terms of a continual interplay b e t w e e n dogmatisation, transformation and results, a0 No doubt all technics and all science can be reduced to an ideology or, alternatively, they can be reduced to an autonomous phenomenon that is interested only in results. 81 But once it is perceived 78 On this see now D.R. Kelley, The Human Measure: Social Thought in the Western Legal Tradition (Cambridge MA & London: Harvard University Press, 1990). 79 j.y. Goffi, La philosophie de la technique (Paris: Presses Universitaires de France, 1988), 13-30. 80 Astolfi & Develay, supra n.75, at 44-46. 81 j.y. Goffi, supra n.79, at 105, 119.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
331
that law is an activity in which there is a technique and that it is possible to reduce the notion of technique to a level where it is a question of constructing universal relations between phenomena, 82 the possibility arises of it being feasible to separate, in a relative way, epistemology from ideology. One must not, perhaps, exaggerate the separation; yet the adaption and use of la technique du droit in political systems that have opposed ideologies does suggest that a failure to distinguish epistemology from ideology will of itself encourage a view that legal technics are simply the application in an "if ... then" fashion of rule-based propositions. And this in turn will eclipse the cognitive and communicative techniques of different configurations of legal institutions. There is, as Bernard Jackson states, nothing inevitable about the conceptual jurisprudence of the G e r m a n Pandektenrecht and it m a y be that the approach will be seen as not being inevitable by both teachers and practitioners alike. 83 But if the institution of the legal action (actio) is reintroduced into the cognitive model of legal reasoning it is possible to see, as this article has hopefully indicated, that the technicalities of problem-solving in law take on perceptible shifts in orientation. Legal institutions, in other words, are fundamental to the way that certain social problems are perceived and organised so as to be not just rule susceptible, but susceptible to particular configurations of rules. V It is, then, in the structural configurations of legal problems that legal institutions have their central role and it is these configurations which in turn underpin the role of analogy as a central technique in legal analysis. In legal systems that still operate in the descriptive and inductive modes this role of analogy is far easier to perceive since the institutional model is one that functions much closer to the concrete and thus uses the model to create more complex and thus more visual images. 84 Institutions are a matter of parti-
82 Ibid., at 21, 26-27. 83 Jackson, supra n.46, at 189-190. 84 C£ Blanche, L'axiomatique, supra n.68, at 84.
332
GEOFFREY SAMUEL
cular kinds of named objects around which rules form. Thus one can readily understand how defence council in a product liability case might, in a descriptive system, wish to distinguish between a defective pair of underpants and a defective bottle of ginger beer, for the two objects might attract different rules. 85 In the deductive and axiomatic systems, the institutions tend to be more depersonalised and dehistoricised and much further removed from the technical problems to be resolved; 86 the configurations, in other words, have become "scientific truths" represented as "laws" which, needless to say, imply the epistemological assumption of propositional, and thus two-dimensional, knowledge representation. In an axiomatic system, therefore, it becomes perfectly feasible to transfer the risk liability relationship between person and animal to person and motor car or person and dangerous explosives. Such transfer proved impossible in the Common law system because it seemed to involve the conscious descriptive jump from two particular factual situations, 87 but in the Civil law it was quite easy;88 and this helps show how institutions working at a level where they gain their meaning from the system, rather than the world of social fact, can become free of analogical reasoning simply because they have become part of the isomorphic structure itself. Hopefully this contextualisation of legal institutions within an epistemological model will go some way both in explaining the notion of a legal institution and in indicating that the history of legal technique is not a history of deontic logic. Moreover, none of this is to d e n y the ideological role of institutions; it is to suggest that alongside the h u m a n sciences there is also a role for the history of the natural sciences if only to show h o w the philosophy of la technique is as vital to the understanding of legal thought as is political and social theory. In a way, a rule-based theory of law has the great advantage of seemingly being able to reduce the ideological and epistemological aspects of legal science to a single discourse in as much as legal institutions become rule-described elements in a system which, at one and the same time, is describing a political 85 Grant v. Australian Knitting Mills Ltd [1936] A.C. 85. 86 Astolfi & Develay, supra n.75, at 44. 87 Read v. J Lyons & Co [1947] A.C. 156. 88 Cass. Ch. r~unies 13.2.1930 (Jand'heur) S.1930.1.121.
EPISTEMOLOGY AND LEGAL INSTITUTIONS
333
and a social world. However, the historical problem here is that this reduction of, so to speak, public and private interests to a positivist system was itself an ideological development and so care must be taken to ensure that the object of legal science, the Corpus Iuris and the legal science it bequeathed, 89 does not itself become modified b y its o w n ideological history. This danger can to some extent be avoided if one moves a w a y from a rule-based epistemology, in that the ideological implications of an institutional view of legal knowledge become much more ambiguous the moment that one is operating in a three-dimensional world with a model which is very much more flexible. Dominium w the relationship b e t w e e n persona and res - - can, for example, be emphasised at the expense of imperium, just as imperium can be used to suppress obligationes. C o m m u n i s m or capitalism can, in other words, make the same rational use of la m~rne technique in order to achieve the ideological result required. All of this can be described in terms of propositional knowledge and nothing in this article should be taken as implying otherwise. What this paper has tried to emphasise is that to say that legal institutions are rule-governed is quite a different matter.
89 The aim here is not to suppress the great differences between the Civil and Common law traditions by blandly assuming that English legal science is ultimately founded on Roman law; yet if legal science is here understood as the process towards abstraction and axiomatization in legal thought, then the role of the Corpus iuris is fundamental to Western thought in general: see generally Kelley, supra n.78, at 165-186, P. Stein, Legal Evolution (Cambridge: Cambridge University Press. 1980), 78-98.