Legal Argumentation and Justice in Luhmann’s System Theory of Law

The paper reconstructs Luhmann’s conception of legal argumentation and justice especially focussing on the aspects of contingency and self-referring operative closure. The aim of his conception is to describe/explain in a disenchanted way—from an external, of “second order” point of view—the work on adjudication, which, rather idealistically, lawyers and judges present as being a matter of reason. As a consequence of some surface similarities with Derrida’s deconstructive philosophy of justice, Teubner proposes integrating the supposed reductive image of formal justice described by Luhmann with the ideal conception of justice presented by Derrida. Here this kind of attempt is rejected as epistemologically wrong. In addition, Luhmann’s theory is argued to have other shortcomings, namely: the failure to understand the pragmatic function of principles, and the incapacity to describe the current legal questions linked with cultures and legal pluralism, which characterise our society.

Luhmann makes a very strong claim: For him the sociological, external observation of the legal system is capable of describing and therefore explaining in an disenchanted way more, more freely and better than the internal self-description of the system provided by legal dogmatics, legal theory and judiciary [39, pp. 18-19, 26; 37, p. 293]. But to observe meaningfully from the outside we need to know the self-reference of the system, i.e., its self-description, and how the system works [36, pp. 17-18, 75]. Thus, an external observation can describe, in particular, what must remain invisible to the internal observation and operations. For example, the fact that the system is grounded on paradoxes like, e.g., its (operative) unity which is given neither by a ''fundamental norm'' or a constitution, nor by justice. This happens because its unity depends on both an operational difference, that of the code legal/illegal, and justice, defined as the constancy of equality of treatment. In this way the ''norm justice'' hides the paradox of treating equal cases equally and different cases differently [26, pp. 374-418]. So the unity of the system is a ''unity of differences.'' If openly faced by an internal observer this kind of paradoxes would bring the system into a cognitive dissonance and an operative blockage. This kind of paradoxes cannot be discussed except from an external position. 2 Anyway, Luhmann's claim is there and we shall try to see if it is justified with reference to my thematic.

2
In a differentiated and contingent social reality the function of social systems is to reduce the complexity of the possibility of actions and-therefore-the uncertainty of living conditions for human beings, making probable what is improbable, i.e., meaningful interaction [27,Ch. 3]. Accordingly, the stabilizing function of the legal system is to distinguish people's actions that conform to the (normative behavioural expectations selected by the) law from those which do not, using legal rules as Footnote 1 continued to externality, please do not waste your time reading the paper. My serious and at the same time ironical statement means-among other things-that we have to avoid making the mistake of asking a theory questions for which it is not fit: see, as an example, Rosenfeld [48, p. 89]. If we consider ''justice'' as a legal concept-and not as a moral, political or religious one-as Luhmann does theorising the legal system, what does it mean for a descriptive theory of the autopoietic legal system to contribute to a theory of ''justice according to law,'' if not a truism? And what does it mean to contribute to a theory of ''justice against law,'' if not pure nonsense? To adopt an internal point of view, on the contrary, does not mean to avoid a critical stance: see Michael Walzer's concept of ''subversiveness of immanence'' [60,Ch. 3;cf. 9,Ch. 4,§4]. A functionally equivalent concept would be Jullien's ''potential of situation'' [20]; with regard to its critical meaning see [63,Ch. 2,§5]. I am an example for being a Luhmannian without being proud of autopoietic semantic [5]. In this sense-according to the Luhmannian terminology-with respect to Luhmann's legal theory I am a second order observer that tries to describe his theory and see where his ''blind spot'' is through which he cannot see what he cannot see. 2 Although Luhmann supports the external observation he does not claim the priority of this perspective. There is-so to speak-a kind of labour division between internal and external observation and both have their proper function (see [36, pp. 16-18] and, as an example, [36, p. 356]). But a sure performance of Luhmann's heterodox description is the ''deconstruction'' and demythologisation of the work of the legal system and of the activities of legislators, judges and lawyers. conditional programmes for the application of the system-specific functional code legal/illegal, or lawful/unlawful. Nevertheless, elements of the (legal) system are neither human beings nor actions, but the communications which inform us about the (legal or illegal) meaning of human actions. In a functional differentiated worldsociety as ours is [25,36,Ch. 12,§5;39,Ch. 10], only systems can produce and reproduce such communications and only the system of law can properly and autonomously (i.e., self-referring) produce legal meanings on what is defined as legal or illegal. This is what is meant by ''autopoiesis'' [36, pp. 43, 57-65; see also 42].
The autopoiesis of the system has important consequences: first of all, the system must distinguish itself from other systems-including that of the whole societyidentifying itself as a unitary unit, while all those systems that exist outside its boundaries constitute-together with the natural environment itself-its environment. Secondly, communication occurs-with few exceptions 3 -only within a system. Normally there is no direct communication between different systems: There is no direct communication with the environment. Notwithstanding, the system operates in a condition of ''cognitive openness.'' Only simple irritations come from outside into a system. The single system selects these as relevant or irrelevant. The former are transformed into information to be elaborated according to its proper language, its conditional programmes and its self-observations and descriptions. To return to our subject, the legal system and only the legal system elaborates legal significant information distinguishing them as lawful or unlawful (''operative closure'') [36,Ch. 2,. 4 3 According to Luhmann, law is a peculiar unitary system. Its unity is symbolized by two concepts: that of ''validity'' and that of ''justice.'' Validity symbolizes the formal unity of the legal system in the multiplicity of its operations (''variety'') [36, pp. 358-359] that in their recurrence concern the true reality of the legal system. It has an ''operative function'' that permits the system to change from one distinguishing operation to others (e.g.: from the passing of a statute to its application in a decision, from the production of a written rule to its interpretation, or from a judicial decision to its enforcement). Validity makes the recursive selfreferences possible that-so to speak-keep the system alive and preserve its identity [36, pp. 75, 98]. 5 4 There is another important consequence, this time from an epistemological point of view: If autopoiesis means self-reference and self-reference means that a system implicate only itself, then autopoiesis means tautology. Indeed: only the legal system determines what law and justice are. In other words, the observer-of any order-observes only what he observes [cf. 36, pp. 31, 493-494]. What is at stake, then, is the de-tautologising tautologies through their development, but in this way the result are paradoxes: ''A rose is not a rose if…,'' or if you prefer: ''Law is the distinction of lawful and unlawful'' (see also [27,esp. Ch. 11,§ §3,8]). 5 Therefore, there cannot be a ''communication breakdown'' within the system. Legal Argumentation and Justice 343 because the ''formal symbol of legal validity […] mediates the reference to the system without any concern for its content'' [36, p. 215]. The unity of the system can be represented also on a substantial level, concerning the quality not of the code lawful/unlawful, but that of the conditional programmes (statutes, norms, principles) which give the system the criteria to apply the code to the behaviours at stake. For the system, the distinction of a human action as legal or illegal is-intrinsically-just (and must also be claimed in this way!): ''Independently from the facts, the legal system wants itself as just'' [36,217], and so in principle also the judicial decision is a just one. Paradoxes cannot be overtly exposed, but ought to (or must?) be concealed. For instance, it is better not to ask if it is just that the legal system distinguishes actions as legal or illegal in the way it does [36, pp. 170, 176, 309-310]. Therefore, no legislator and no judge may say: ''Obey my decision, which is wrong!'' 6 4 Justice refers itself to the difference of ''what is undeterminable and what is determinable.'' Its function as a ''contingency formula'' is to cross the boundaries of these categories making more case solutions that conform with the law plausible. This means that justice keeps open the possibility to treat facts and cases in different ways, i.e., contingently [36, p. 220].
On the other hand, the function of justice as a symbol is to represent the legal system within the legal system [36, p. 221]. It is a kind of metarule, ''a programme for all the programmes.'' Through justice the legal system gives a sign of its substantial unity and the correctness of its operations, which ought to be congruent with it. Nonetheless, the consideration of justice as a norm raises some problems. The ''norm justice'' is a vague formula that cannot be better specified. Indeed justice is a criterion of self-control for the legal system, but if its meaning would be better defined the-paradoxical-consequence would be that some ''system's own operations would be qualified as not belonging to it'' [36, p. 218]. Therefore justice is a criterion for the inclusion or exclusion of norms and decisions, but as a norm it remains under the conditions of the positivity of law. As an ideal, a principle or a value, justice is an undetermined rule that does not let us ''foresee what decision will follow and what interests will be supported'' [36, p. 221]. On the other hand, in reference to a case, justice serves to define it [36, p. 220], not searching for the only ''right answer'' [14, esp. Ch. 13 and Appendix, §3], but allowing more possibilities to reach a plausible, correct (and/or fair) decision. Under the condition of contingency and positivity every decision-however just-could be taken in another way.
To define cases is possible when justice is considered as formal equality in the operative sense that ''equal cases are treated equally and different cases differently'' [36, pp. 223, 227 and 272]. 7 This means that justice must be understood as ''consistency'' [36, pp. 18-19, 223-226], but not regarded as congruent with values (positive law is not natural law, after all). This would make the system inadequate with regard to the complexity of society and the multiplicity of different cases [36,Ch. 6;and 29]. On the contrary, justice must be compatible with an ''adequate complexity of the consistent deciding'' [36, p. 225] and, therefore, the legal system must have at its disposal a ''variety'' of operations to manage the cases [36, p. 358]. Nevertheless, equality as consistency also means coherence in the way the judges take their decisions (stare decisis). This is why, Luhmann claims, ''justice is redundancy'' [36, p. 356]. Redundancy reduces the quantity of information that the system needs to process the cases and, by contrast to it, the system selects new significant information to decide new and/or hard cases: so the ''reduction of complexity serves to increase complexity'' [36, p. 354; see also 38, p. 292].

5
In everyday life we call the core operativity of the legal system not merely application of law but using the item ''administration of justice.'' And what do we require from the administration of justice? First of all, we require that the judges decide also when they cannot decide, according to the principle of ''prohibition of denial of justice'' [36, pp. 310-319]; secondly, we require that they decide in the right way, or at least correctly (or fairly); finally, we require that they demonstrate that their way is right both grounding on a correct comprehension of the true meaning of applied-or ''invented'' [36, p. 317]-norms (through interpretation and analogy), and giving good reasons for their decisions (justification through argumentation).
Obviously, there are just and unjust decisions, so that judicial decisions may be sustained or criticised from points of view that are different from the legal one, e.g., from a moral, political, or religious one. Or they may be sustained or criticised from the legal perspective itself by either proposing a different-and supposedly betterinterpretation of the legal rules concerned, or using different-and supposedly better-arguments than those used to justify the decision at stake.
When the legal sociologist describes the legal system applying his code to (qualify the meaning of) behaviours, he poses himself outside the system and makes 7 A second order, external, observer could ask: Is this reasonable formula a just or a fair one? The true problem with it is that it is full of contingency and arbitrariness: Actually, the judge, who wants to decide a case justly (or fairly), how and according to what criteria and standards does he distinguish between equal and different cases? Luhmann marginalises fairness (equitas, Billigkeit) placing it only in the premodern law and coupling it with the political power of the prince. But, in my opinion, especially under the conditions of our pluralist society fairness plays an important role. On fairness in a system theoretical perspective, see [11].
Legal Argumentation and Justice 345 a ''second order observation,'' but when he describes a judge or a lawyer respectively justifying or criticising the decision at stake, he is operating on a different, higher level. Namely, he makes a third order observation, because he observes how the judge or the lawyer works on the ground of their personal observation of the way in which the legal system functions. Moreover, at this level the sociologist no longer observes the legal system itself (as a whole), but one of his organisations: the court [36, pp. 328-333]. 8 ''Courts interpret […] in an argumentative way: to mean the ratio of their own decisions'' [36, p. 327]. Now we must remember a fundamental and radical circumstance: The law on the grounds of which the legal system operates its basic distinctions is only the positive law.
Describing positive law-and its functioning in a legal system-under the sociolegal perspective means to consider it no longer as being in force (valid) on the grounds of a final justification, as internal observers of the first order do (e.g., lawyer, judges, legal theorists). From this external point of view, positive law is in force only on the grounds of the contingent constitutional, legislative, administrative and judicial decision that poses, abrogates or changes it. Contingency is the very nature of decisions: They have improbably landed as they are, but they could occur probably in another way (and this is true also with recourse to legally transformed moral values and principles). From a sociological, external point of view, the consequence of this way of observing the system of law is that the system is seen as really being contingently consistent of improbable decisions which are in/ congruently, recurrently and seriously taken. Now the main point can be stated as follows. Under the conditions of our complex and functionally differentiated society, legal decisions ought to-and must-be justified [40, p. 35]. 9 And justification happens through rule interpretation and argumentation. 10 Luhmann even reverses this order of causality: ''We may define argumentation as production of a context of arguments and-mediated through arguments-decisions'' [40, p. 36]. Moreover, without justifications ''it is often impossible to ascertain on the basis of the tenor of the judgement what precisely has been decided'' [40, p. 35]. Therefore, the legal system needs (interpretation and) argumentation.
But notice that the Luhmannian view on argumentation does not try to argue in favour of the ''normative suggestion'' (Sollsuggestion) of arguments. His aim is to highlight the contradiction between the normative validity (Sollgeltung) and the logic of the system, and dissolve it in empirical facts [40, p. 25]. 11 The observer can see facts and not normativity, 12 behavioural expectations crystallised in rules and not ideal meanings, paradoxes and not foundations when he sees what he sees in the way it happens and functions.

6
The first order (internal) observer of the legal system is concerned with the quality of his operations, i.e., with the correctness, or-at least-appropriateness (see [18]), justification and legitimation of his distinctions and decisions, that consequently must fit with standards such as rules, principles and values. On the contrary, the second order observer-especially when he is also external to the system-is concerned with the ''conditions of possibility'' of the legal operations and with their function, because for him good or bad, right or wrong, operations are always and only operations. Moreover, from his standpoint the external observer sees that those operations, which for the judges and lawyers are almost implicitly right and justjust because referring to standards! [36, pp. 343-344]-do not enjoy a true, final justification: ''The ultimate grounds are always the penultimate'' [36, p. 406]. Especially those justifying argumentations that make recourse to (legal or legally transformed) principles and values as ''formulas of consistency'' actually hide the embarrassing fundamental (i.e., grounding!) paradox of the contingency of all the law: ''Who understands grounding as referring to grounds will have the need to ground the grounds, too'' [36, p. 347].
Notwithstanding their static and consistent appearance, the (correct, just, objective, true, valid) meaning of values and principles is and remains undetermined and uncertain: It cannot but give a vague orientation and does not determine properly any decision. Those standards ''sell inconsistency as consistency'' [36, p. 348]. 13 So, every legal decision could be taken differently and every legal justification and legitimation is not ultimate and-although grounding on the same values or principles-could be argued not only differently, but also contrarily. Therefore, Luhmann proposes to define ''argumentation'' avoiding any reference to justification, foundation, reason and reasons as instead theorists of legal argumentation usually do [36, pp. 366-367, 38, pp. 285-286]. 14 According to them, argumentation obviously means to search for mistake-free good, i.e., rational grounds, which sustain the interpretation of normative texts to solve a case: 11 Here there is clearly a significant difference from Habermas, who keeps two worlds distinct: that of facts and that of norms [19]. 12 Therefore, according to Luhmann ''normativity'' means that ''determined expectations can be confirmed also when they are disappointed'' [36, p. 61]. Bertea [7, pp. 29-48], is of another opinion, but he argues as a legal theorist. According to him, the fact is that we follow legal rules under the suggestion that law makes an implicit ''normative claim'' on us by so giving us reasons for actions that we accept. 13 Already the young Luhmann knew that the recourse to values served to ''embellish arguments'' [24, p. 44, n. 12]. On the (operative) meaning of values and principles, see [35,34]. In general, see [28, §5]. 14 As an example, see [1].

Legal Argumentation and Justice 347
Argumentation means just grounding. 15  . This is the circumstance in which ''somebody knows a topos, a concept, or a knowledge that is considered in the system as true, [so that] he can guess with more or less probable/ improbable success how it will be argued in the future'' [40, p. 37]. Therefore, the function of redundancy is to explain the conditions under which it is possible to claim the validity of a decision to be taken: either that its justification coheres with one or more of the grounds already known, or that it selects new information which is able to change the state of the legal system [36, p. 354, cf. 38, pp. 291-292]. The function of grounding, then, is to resist moderately the increase in the variety of the system caused by its reaction against external irritations, and ''restore adequate redundancy, using redundancies already present.
[Argumentation] overwhelmingly reactivates known grounds, but in the practice of distinguishing and overruling occasionally also invents new ones'' [38, p. 292]. This is an example of ''continuity in discontinuity'' [38, p. 297]. Therefore, legal argumentation serves ''the production of further reproduction of the system'' and involves directly ''the autopoiesis of the system'' [38, p. 292]. 16 Argumentation is needed not so much to justify simple, routine cases where the distinction of what is legal and what is not depends on the mere application of the law in force. According to Luhmann, the need to argue arises in ''cases in which strict legal/illegal coding fails as a guide so that escape routes must be found, making it possible for someone who has acted lawfully to be treated in special circumstances as if he had acted unlawfully'' [38, p. 294]. Then judges must find decisional criteria-so to speak-outside the law.
What we know is that the system operates using the code strictly self-referentially (operative closure), but it is stimulated by irritations coming from the social environment out there (cognitive openness). When the system resorts to considerations about factconsequences, interests, values and principles it observes and tries to give an account of the environmental claims to justice, i.e., of something that is not law, but facts. This is only possible through ''the re-entry of a distinction into what is distinguished'' [38, p. 296]: the re-entry of the environment into the legal system, being the law ''what is distinguished'' through the basic difference between (legal) system and environment [27, Ch. 5, esp. §1]. On the other hand, the environment corresponds to the ''other side'' of the system/ environment form: to that ''negative value'' which is excluded as principally not important to observe and describe. In this way, external, ''other-referential observations are also operations internal to the legal system itself, and can enter argumentation only in that way'' [38, p. 297]. Consequently, interests, principles and values are no longer external standards, but ''eigenvalues'' [29, p. 27], it means that these are stable ''values'' of the system: These are legal standards and as such their meaning is exposed to contingency. 15 On this topic, see [38, pp. 287-290]. 16  Hence, justice is equal treatment, is consistency, is redundancy, is argumentation. At the end argumentation is justice. We may think that: ''If everything is justice, nothing is justice!'' This is not the case. Within this conceptual chain there is a substantial (and functional) shift occurring in the passage from consistency to redundancy. Justice as consistency is placed on the normative level of the legal system concerning its operative closure: What is at stake here is the recursive application of norms and code. Justice as redundancy, on the other hand, is placed on the cognitive level of the system concerning its mediate, cognitive openness: What is at stake here is its (internal) orientation to the outside, the social environment through the production and selection of information relevant for the cases.
Now we can say that legal justice is the result of a communicative process through which a meeting point, i.e., an ''operative coupling,'' a loose congruency, between normativity and factual information is stabilised (cf. [36, pp. 359-360]). Argumentation is the bridge contingently connecting norms with the reluctant redundancy (reluctant because resisting to variety, 17 which is necessary, too!) of decisions.

7
In order to further develop Luhmann's theory of justice, which ''grasps only half the problem'' insofar as it only foresees a concept of formal justice and not a substantial one [56, p. 15], 18 Gunther Teubner puts the question of the ''ecological justice,'' i.e., of the ''environmental adequacy of the law.'' This means to deal with not only issues of internal justification, but also issues of social legitimacy and acceptance of legal decisions. According to Teubner the legal sociological theory must acquire a better, ''deeper comprehension of justice […] understood as subversive practices of self-transcending the law […] As a final consequence justice would be then a legal self-description, which sabotages itself because its realisation produces ever new injustice'' [56, p. 11]. 19 The ''responsiveness'' of the legal system with respect to its ''social, human and natural environment'' would be possible through the ''re-entry'' of the original distinction law/not law (which is different from that of legal/illegal, because ''not law'' means ''external to the law'') within the legal system itself. This distinction is not an operative distinction, but an observational one. 20 According to Teubner, that means 17 The court is bound ''to its decisions and will change them only with prudence'' [36, p. 97]: only with juris prudentia! (see also [36, pp. 271-272]). 18 For a slightly different version of the issue see [58, p. 197]. 19 The only thing that is really sabotaging here is Teubner's declared attempt to turn Luhmann into a ''quasi-natural law'' scholar [54, pp. 204-205]. This move has been anticipated in [53, §3, and 58]. 20 As Renner [47, p. 68], commenting on Teubner correctly remarks, according to Luhmann re-entry is possible only for a distinction that returns ''in what was already distinguished by it,'' (see [36, p. 76]): Therefore, the re-entry makes the distinction itself visible (and that is the legal system) and does not indicate that the distinction was revoked. So, the re-entry proposed by Teubner is incorrect (see also [36, p. 205]), but this is not the main issue that is at stake here. Anyway: The return to-at the end-a ''value orientation of the legal system is always in danger of abandoning the specificity of the legal concept of the system theory, i.e., the emphasis put on self-reference and operative closure, in favour of normative demands'' [47, p. 70]. According to Luhmann, the (external) ''value orientation'' of the system is impeded not only by its operative closure (as far as constitution is concerned, see [31, p. 212]), but also by the that at the level of self-observation, the legal system can give an account of its environments, although internally reconstructed [56, pp. 18-19].
In this way he can go on stating that legal justice can ''transcend'' the operative boundaries of the system. That is, the contingency formula of justice ''operates at the boundary between law and its external environment and aims for the historical variability of justice and, at the same time, for its dependency on the environment. The call for justice-this is the nucleus of the contingency formula-claims to derive consequences from the dependency of the law […] from its social, human and natural environment.
[…] The intention of justice orientates […] itself […] to respond sensitively to the extremely divergent external claims aiming for the highest consistency. The contingency formula does not aim for a law-immanent justice, but at one transcending the law. The double formula of legal justice means internal consistency plus responsiveness to ecological claims.'' [56, p. 17; my emphasis] 21 From this position, what follows is a kind of epistemologically confused romantic afflatus: ''Legal justice […] appears as a structured social dynamic within the law. Legal justice could be an obstinate process of self-description within the law, which blocks, sabotages, undermines, the routine recurrence of the legal operations, which therefore compels the law to transcend itself beyond any meaning, which, at the same time, subjects itself to the compulsion to go on to produce further legal operations and so sabotages itself precisely because in this way it creates new injustices.'' [56, p. 22] The result of these assertions would mean to transform the conception of autopoiesis of the legal system into that of its ''allopoiesis.'' 22 In my view, Teubner's [56, esp. pp. 25-27] formulations, which are in line with Jacques Derrida's deconstructive philosophy, are epistemologically questionable. 23 Notwithstanding certain assonances (e.g., both speak about paradoxes 24 and improbability, or today's impossibility of groundings), it is not possible to integrate Luhmann's theory of autopoietic systems with the deconstructive philosophy of Footnote 20 continued redundancy implied by the consistency of decisions [36, p. 353]. Here the concept at stake is ''indifference.'' 21 On the whole issue it is inescapable to read [36, Ch. 2, §6]. 22 The term is used first by Neves [45, pp. 258, 260-261] criticising the conception of ''reflexive law'' proposed by Teubner in the 1980s as a kind of law which ''has overcome the limits of formal-rational law and of the substantial legal rationality''. Neves [45, pp. 262-268] also proposes considering the allopoiesis of the law, but in an epistemologically correct way, which concerns the under-development of the functional differentiation of a social system. With respect to the legal system it means ''weakness, irrelevance or absence of the constitution understood as the structural coupling of politics and law'' [45, p. 264]. 23 See also the criticism of Ladeur [22, pp. 120-122].
Derrida. 25 First of all, Luhmann's [32,Ch. 7, §7] account is a constructivist theory and not a deconstructive theory, since it is based on the construction and functional description/explanation of socially important facts starting with the information that a social system obtains from the environment and processes within itself. Derrida's deconstructive philosophy, instead, moves principally on the level of linguistic analysis of concepts and of the criticism of their common sense, ideally trying to reach the depth of their origin, which proves itself as impossible to be founded and inexpressible, because it is without real just, legitimate and true grounds. As far as law and justice are concerned, the breakdown of (legal) common sense is evident: Law and justice are fundamentally related to force and violence [12, pp. 925-929, 941-943], 26 and concerned not with a real person, but with the ideal presence of the other [12, pp. 959, 965] through the mediation of a decision. However, that decision is undecidable/impossible [12, pp. 963-965], namely it is neither authentic nor just or true because of the constitutive paradox resulting from the legal requirement that a universal rule be applied to an actual and individual situation which demands to be treated in its specific aspects that are incalculable [12, pp. 947, 961-963]. In this case, the-paradoxical-result of deconstruction is that of constructing the absolute mystic meaning of Justice, 27 that is obviously unattainable and, therefore, impossible. 28 Law is not justice. Law is the element of calculation, and it is just that there be law, but justice is incalculable, it requires us to calculate with the incalculable; and aporetic experiences are the experiences, as improbable as they are necessary, of justice, that is to say of moments in which the decision between just and unjust is never insured by a rule. [12, p. 947] 29 Luhmann aims for something completely different: With his theory he does not purport to deconstruct in an ontological way; it rather deconstructs in the sense of demythicising, i.e., dismantling the self-narratives (the ''self-descriptions of the first order'') on the social world and, in our case, on the law, for which, for instance, the work on law and justice is a matter of reason. Moreover, for him (legal) justice is not impossible, but only improbable, because it is not absolute, but rather as contingent as the decision concretising it.
Teubner himself seems to be aware of the ''diametrical opposition'' of the perspectives of the two authors [55, pp. 3-4]. Notwithstanding, he tries to elaborate 25 For a Luhmannian criticism of Derrida's deconstructive philosophy see [33]. 26 On the instrumental relationship between law (i.e., the legal system) and force/violence in Luhmann's legal theory see [36,Ch. 6,§4]. Here force plays a role not in the foundation of law, but in the evolution of the legal system. 27 From his ''alternative deconstructivist conception,'' Rosenfeld speaks of the ''imaginary realm of perfect justice where comprehensive justice becomes possible'' [48, pp. 20, 60]. Justice is defined as ''comprehensive'' because-according to Rosenfeld-Derrida claims that the ''ultimate realisation of justice depends on full consideration of all similarities and all differences that come into play in the course of intersubjective dealings between self and other'' [48, p. 58]. 28 For a similar deconstructive thought strategy of impossibility see [13]. a ''social theory of justice,'' combining what epistemologically is uncombinable, i.e., self-reference and transcendence. His central question becomes whether ''it is [correspondent with] law to apply the distinction lawful/unlawful to the world,'' so that this were ''the real question of justice.'' [56, p. 27] And in fact this is ''the real question of justice,'' but we ought to ask the terms in which we understand justice: in those of Luhmann, or in those of Derrida? Or in those of the common man here personified by Teubner? Moreover, if it is ''the question of justice,'' it is erroneously formulated. The real question could be: Is it right to apply the distinction of lawful/unlawful to the world? 30 Firstly, this issue indicates that in our complex and functionally differentiated society the law (i.e., the legal system) is ultimately ungroundable. Secondly, this issue shows thataccording to Luhmann's legal theory-it is not really a legal question, but a political question about the legitimacy of the political power which produces law in order to make uncertain normative expectations certain [36,Ch. 3, § §2, 4, Ch. 9, §2], and a moral question that refers to the issue of giving a right (not a legal) answer to conflicting claims. But thirdly, posing this question of justice Teubner is not integrating the autopoietic legal theory, but simply trying to ''transcend'' it through an epistemologically non-sustainable eclecticism. In our contingent reality it is not necessary to adopt the autopoietic system theory to describe (scientifically or, at least, sociologically) our social context: In order to describe society there are alternative theories enough. But if we move on a scientific level we must operate congruently with the theory we adopt, or want to criticise.
For this reason, according to Luhmann, it is possible to speak neither about the transcendence of justice with respect to law nor about the self-transcendence of the law. 31 For, from the point of view of the autopoiesis and autonomy of the legal system, those speeches are simply meaningless or, to use a phrase famously coined by Jeremy Bentham, ''nonsense upon stilts.''

8
In the end the question is: What does Luhmann not see, or what can he not see through his system theoretical observation? Because, no observer is perfect, everyone has a ''blind spot'' [32, pp. 85, 95, 133], and his observation and description are anyway contingent. In my view the limit of Luhmann's theory of self-referential legal system is not the impossibility that the law transcends itself.
What Luhmann does not see-or perhaps: does not want to see?-is that there exists a substantial difference between a legal system provided with a liberal constitution and a legal system without constitution 32 -and, consequently, there is a 30 For a correct formulation of the question in these terms see [54, p. 202]. 31 Similarly Clam is critical towards the ''reception of transcendence'' in the law [10, pp. 50-51]. 32 See, for example [40,30,31,213,36,Ch. 10,§4]. significant difference between the Rechtsstaat or rule of law (''state of law'': D.N. MacCormick) and the constitutional state [61,Ch. 2,62,Ch. 7]. In the constitutions of liberal-democratic states, principles like liberty, equality, solidarity, toleration, etc. do not concern simply external references that the legal system has just recognised and internalised to adapt his complexity to the social environment. Rather, today such principles are true self-referential ''eigenvalues'' of the constitutional law and of the legal system itself. This is not a minor difference, because lawyers-at least some constitutional lawyers-and judges-at least some constitutional judges-through their first order self-observation are conscious of the fact that in making (hard case) decisions by recourse to principles they do not produce certain, necessary and just decisions and argumentations, but adequate, plausible and contingent decisions and arguments. But notice that even if those products are not just-that is, not absolutely just!-still they are (or tend to be) at least more just than they were without referring to principles.
Contrary to what Dworkin had assumed once, i.e., that there is a right answer also-or especially-for hard cases [14], (some) lawyers are absolutely aware of operating in a state of uncertainty, although certainty remains a ''regulative ideal'' of the judicial practice. This corresponds to the spirit of principles, the meaning of which has to be specified in connection to the cases at stake. From the undefined aspect of principles comes the uncertainty, but also their pragmatic value. As Gustavo Zagrebelsky-a former president of the Italian Constitutional Courtstates: ''Principles give us criteria to take a stand with respect to a priori undetermined situations when they effectively occur […] They demand that we take a stand […] facing reality when it expects our reaction'' [61, pp. 149-150, emphasis in original; for an example see Ch. 7, §4].
So in adjudicating (hard) cases the result to be pursued is-although contingent-a correct, adequate, fair decision [61, pp. 203-204], that could be justified through good, i.e., principled reasons.
There is also another explanation of the fact that the contemporary law of a constitutional legal system is an uncertain law, which orientates itself to constitutional principles. It is that our society is not only a complex, functionally differentiated society, but it is also a pluralistic and multicultural one, and principles, with their open, undetermined meaning are fit to respond to the demands of parties with different moral, political, religious, ideal values: because principles admit the ''co-existence'' of diversities [61, pp. 11-14]. The liberal state (of law) of the nineteenth and twentieth century was conceived of as governing a national homogeneous society so that the general and abstract norms were seen as the appropriate instrument to regulate the behaviours of (also culturally) equal citizens [61, pp. 33-38, 200-201]. This is no longer the situation in which we live. In a pluralist/multicultural society a common legal ideology of the judges that permits a uniform legal interpretation and adjudication does not exist anymore. And the culturally different parties often demand to be treated differently, i.e., according to their own culture. Principles are-or at least can-be the instruments that tolerate a reasonable differentiation of treatments (cf. [61, pp. 168, 203-204]).
Principles are carriers of substantial justice. 33 But, as a matter of fact, even the supposed formal principle of equality, that-following Luhmann-instantiates the ''normative form of justice,'' has a core of substantial justice, too. The criterion, according to which equal cases ought to be equally treated and different cases differently, is complementary to the positive version of the prohibition of discrimination, which can be stated as the Dworkinian [14, pp. 226-228, 272-274] principle of ''equal concern and respect.'' In fact, we may distinguish equal cases from different ones to decide them appropriately only if we give equal concern and respect to the parties concerned and do not want to discriminate between or against them.
In a liberal constitutional legal system, forms and procedures are not to be confused with formalism. Forms and procedures are not completely self-referential, they do not exist for their own sake. As far as they are selective and cogent, they have justificatory reasons related to the realisation of some principles (see [59]) such as certainty, equality, impartiality, participation, etc. that imply ways of treating people correctly.
But to recognise equal concern and respect for people, the state must have a liberal democratic legal system [41] and not all states of the world society have such a system. It follows that the true pluralist/multicultural society forms only a part of the world society and only this part has a fully functional differentiated legal system. We cannot think, e.g., of the Islamic law or the Canon law, where the will of Allah or of God constitutes the ''fundamental norm,'' in terms of a differentiated system. But we cannot think in the same terms also of the customary law of native peoples or of immigrant minorities that live in our society, who demand multicultural recognition and often ask for cases not be decided by courts but rather to be treated by means of mediation or negotiation processes.
So the world legal system-contrary to what Luhmann thinks-is not a unitary system or-better-it simply does not exist at all, while the legal context of our societies is not a unitary, but a pluralist one. In it problems arise that Luhmann's theory [36, pp. 22, 279-280] cannot see, 34 or for which it could give us only the unsatisfactory, obvious explanation that from ''values derive no consequences for the real cases,'' because they do not give certain orientation either for the action to be held or for the decision to be taken, or that in the end also cultural, value oriented 33 About which Luhmann argues that its meaning must remain unspecified, because otherwise some operations of the legal system could be unrecognised as belonging to it: but this is precisely what happens through the judicial review!. 34 For a contrary opinion see Nobles and Schiff, but they arrive at the not so satisfying conclusion that ''Systems theory, with its identification of modernity with functional differentiation within separate social subsystems […] offers a basis for understanding why, in the modern world, it is possible to distinguish law from other subsystems of society, and why that distinction is not dependent on any essential set of institutions or structures. Systems theory neither limits its recognition of the legal to state law nor ignores state law and its coercive potential. Rather, systems theory's explanation of state law in terms of the structural coupling and coevolution of law and politics provides a basis for exploring the relationships between law and other subsystems that generate and sustain nonstate law. In other words, through its understanding of state law it stimulates the study of forms of coevolution in which patterns of legal pluralism arise'' [46, p. 293]. cases are decided within the legal system through its operative closure. 35 Multicultural issues concern not only ''simple'' questions like food rules, festivities, the wearing of special clothes, but also health care, patriarchal relationships, coerced marriage, separate jurisdictions for minorities, offences implying the socalled ''cultural defence'' like female genital mutilation (FGM), honour homicide/ suicide, etc. If the theory cannot describe the existential meaning that culture has for human beings, it can not even adequately describe/explain these kinds of issues. 36 According to Luhmann, ''culture'' cannot constitute the positive value of a code as it is for ''legal'' for the legal code ''legal/illegal.'' This is not possible since it is difficult to find the counter value for culture. A candidate may be ''nature.'' But both culture and nature are too general and universal; so they do not allow for meaningful distinctions for observing actions. Culture, for instance, cannot serve ''to distinguish cultural objects from other kinds of objects'' [37, pp. 32, 53-54].
That notwithstanding, in our society there are culture-related (legal) problems which require not only that social complexity is reduced, but also that the plurality of culturally equivalent, normative institutions is made possible according to a global-local dialectic. And if the Luhmannian theory of autopoietic systems cannot find suitable distinctions for the pluralist and multicultural social reality, this is not a problem for the reality itself, but a fault of the theory.