LEGAL AUTONOMY AND

REFLEXIVE RATIONALITY IN
COMPLEX SOCIETIES
PATRICK CAPPS AND HENRIK PALMER OLSEN
University of Bristol, UK and University of Copenhagen, Denmark

ABSTRACT

Proponents of the idea that law has assumed an endemically closed character consider
that there are severe limits to the possibility of successful legal regulation in highly
complex and socially differentiated societies. This idea reaches its most radical form
in the theory of law as an autopoietic system which suggests that the autonomous
individual at the core of most traditional legal theory should be replaced by the self-
creating and self-maintaining legal system. Advocates of this theory suggest that legal
scientists and practitioners should reject traditional rationales of legal regulation and
adopt a new, and more modest, mode of legal rationality called reflexive law. In this
article, we argue, first, that the idea that specialized autopoietic subsystems have epi-
stemic primacy is flawed and, second, that this initial argument coupled to a concep-
tion of the function of law which is held in both autopoietic theory and more
traditional conceptions of law reveals an insight into the possibility of effective legal
regulation in complex societies. Specifically, effective legal regulation in complex
societies requires a reformulation of reflexive law so that basic principles of political
morality are incorporated within its autonomous domain.

INTRODUCTION

R
OSCOE POUND wrote in 1954: ‘I am content to see in legal history
the record of a continually wider recognising and satisfying of human
wants or claims or desires through social control; a more embracing
and more effective securing of social interests . . . in short, a continually more
efficacious social engineering’ (1954: 47). Pound would have been greatly
dismayed to have learned that his characterization of the relationship
between law and social interests was to become the antithesis of the un-
deniably influential view introduced by the proponents of autopoiesis. The
theory of autopoietic law as developed by Gunther Teubner and Niklas

SOCIAL & LEGAL STUDIES 0964 6639 (200212) 11:4 Copyright © 2002
SAGE Publications, London, Thousand Oaks, CA and New Delhi,
Vol. 11(4), 547–567; 029253
548 SOCIAL & LEGAL STUDIES 11(4)

Luhmann suggests that the days of law as a mechanism of social engineering

are past. Specifically, they consider that as social complexity increases, the
various elements of society – such as law, economics or politics – will fracture
and close themselves off from one another in various subsystems. This leaves
the legal system as one among a number of subsystems which is unable to
effectively achieve fixed, substantive goals in society via regulatory processes.
In what follows, reflexive law – which is seen as a plausible alternative to
more traditional forms of legal regulation – is examined. Through a critique
of the epistemological and ontological assumptions made by proponents of
autopoiesis we reconceive the relationship between substantive and reflexive
forms of legal rationality so that legal dysfunction, in light of increasing social
complexity, can be avoided.

LAW’S AUTONOMY AND AUTOPOIETIC LAW

At the core of Teubner’s and Luhmann’s theories (e.g. Luhmann, 1981,

1995a, 1995b; Teubner, 1985, 1986, 1988, 1993) are two key presuppositions:
(a) that the production and enforcement of legal norms represent an attempt
to solve problems of coordination in complex societies and (b) that law is a
social subsystem, which is isolated from other subsystems in the sense that
it is normatively closed.1 These two presuppositions are common in legal
theory and are in fact shared by a variety of characterizations of law which
Gerald Postema has usefully drawn together under the label of ‘the
autonomy thesis’ (see Postema, 1996).2 Postema considers that a key presup-
position of the autonomy thesis is that while ‘. . . law’s ultimate aspiration
may be justice, its proximate aim and defining task is to supply a framework
of practical reasoning designed to unify public political judgment and co-
ordinate social interaction’ (1996: 80). Furthermore, it is this fundamental
function that holds the key to explaining the nature of law. Moreover, for
Postema, the autonomy thesis contains three logically indispensable and
logically interrelated features which purport to solve the problem of social
coordination. First, law forms a distinct sphere of practical reason which
offers reasons for action to officials and citizens alike. Second, the norms
within the limited domain are preemptive in the sense that they offer exclu-
sionary reasons for action which override non-legal reasons for action.
Third, legal norms must be readily identifiable at source without recourse
to moral argument or political evaluations outside of the limited domain.
All these features point to the idea of law as an autonomous normative
system. What, then, is the difference between an autonomous system and an
autopoietic system?
Autopoietic theory derives many of its concepts from systems theory in
evolutionary biology (Maturana and Varela, 1980).3 At the core of auto-
poietic theory lies the assumption that society is comprised of a number of
subsystems.4 These systems operate in society independently of each other
and create the conditions for their continued existence and evolution by
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 549

themselves. Thus the legal system produces law, the economic system
produces payments, the psychic system produces consciousness and so on.
According to Teubner (1983), among these subsystems there is an increas-
ing potential for social conflict which traditional legal rationales5 are unable
to solve. Following Luhmann, he says:
The functional differentiation of society, so Luhmann tells us, induces highly
specialised sub-systems to develop their own specific rationality to such a
degree that radical system conflicts are inevitable. . . . These clashes include
those between universal social structures (economy, science) and territorially-
bound political and legal structures, between scientific planning and economic
production control, and between the temporal requirements of social inter-
dependencies and the slow-developing processes of education and insti-
tutionalisation. (Teubner, 1983: 270–1)

The problem that is alleged to arise from this situation is that the conceptual
apparatus of the legal order is not sufficiently sophisticated to deal with such
a high level of social fragmentation. But the proposition that legal regulation
has become more difficult as social differentiation has increased is neither
new nor particularly radical. Postema says:
. . . as modern Western societies have become increasingly pluralistic, these
problems of principle [that is, solving the problem of social co-operation] may
have become more serious and more difficult to solve. Our societies embrace
a number of different communities whose histories and practices yield funda-
mentally different and potentially deeply conflicting moral points of view.
While any one such moral point of view might prove a sufficiently rich and
stable set of principles for social interaction within a portion of the society, no
single one of them is widely accepted enough to structure interaction in society
as a whole. (1996: 91)

While the similarity between the two theories is readily apparent, the theor-
etical frameworks that are employed to analyse this problem are different in
two key ways. First, while both theories consider the law to be an autonomous
form of practical reasoning, the orthodox, agent-centred approach of the
autonomy thesis is replaced in the highly abstract systemic approach of
autopoietic law by an explanation which holds that conflict resolution is
achieved, somehow, by subsystems rather than by social engineers or lawyers
working within them. Second, in the orthodox autonomy thesis, autonomy
(from political morality) is considered to be a rational response to the problem
of social conflict, and hence it becomes an issue for legislators and officials
to pursue and develop law’s procedural and substantive autonomy as a foun-
dational aspect of its legitimacy. Autonomy in the theory of autopoietic law is
presented as an endemic feature of increasingly differentiated societies – a mere
system effect. Luhmann says: ‘. . . autonomy is not a desired goal but a fateful
necessity. Given the functional differentiation of society no sub-system can
avoid autonomy’ (1986: 112; see also Rottleuthner, 1989a: 788–91).
Proponents of autopoietic theory do not consider that law – as a subsystem
– is completely ‘closed’, but rather, that it operates with a dialectic between
550 SOCIAL & LEGAL STUDIES 11(4)

normative closure and cognitive openness. This distinction is not entirely

unambiguous, but perhaps the idea of normative closure might be caught in
the phrase that ‘only the law can change the law’ (Luhmann, 1986: 113). We
take this to mean that valid law can be altered only via processes that are
prescribed by other valid laws (Luhmann, 1995b: 289–91). This creates a
circularity, whereby the various normative elements in the legal system legiti-
mate each other and form the basis of the system’s self-creation, hence,
autopoiesis (Rottleuthner, 1989a: 791–4).
Of the notion of cognitive openness, Luhmann tells us that the legal system
is open to cognitive information about its environment (other subsystems)
and that this information is used to adapt to changing circumstances. We take
this to mean that the legal system will adapt, for example, to changing forms
of economic transaction or to new informal ways of family life, by creating
new rules that are capable of solving problems that might arise under these
new circumstances, or by adapting already existing rules to a new situation.
Neither the concept of normative closure nor of cognitive openness,
despite the novel terminology, is radically divergent from other, more
traditional theories. Again, when reading Postema on the concept of law’s
autonomy, it is striking that the interplay between the formal isolation of
legal reasoning and the more open character of judicial reasoning (see
Postema 1996: 87–8; also Raz, 1979: 37–53) makes it possible to conceptual-
ize what appear to be the same features of law as those which the concepts
of normative closure and cognitive openness address.6
Luhmann and Teubner do not identify this similarity but rather focus on
the problems that can develop out of the interplay between openness and
closure. Inputs into the legal system become transformed into legal
categories, which can then be used as a raw material for the performance of
legal operations. Thus the cognitive openness of the legal system makes it
possible for it to ‘learn’ and thereby adapt to the regulatory needs of a
changing social landscape. The assumption is that if the legal system does not
‘learn’, it produces crisis conditions which result in a lack of social stability
and compliance to the law. Within the modern welfare state, Teubner claims
that ‘. . . social processes and economic arrangements are simply too dense,
complex and potentially contradictory to be adequately accounted for in the
kinds of interventionist control mechanisms that have been created’ (1983:
268). Thus if the legal system is to maintain its role as social coordinator, it
must develop new and more sensitive ways of absorbing information from
the environment (i.e. from other subsystems). In a changing and diverse
social environment, it is crucial that the legal system develops a regulatory
strategy that is based on a form of legal rationality which has a greater poten-
tial to learn (i.e. which is more cognitively open). Against this background,
Teubner envisages a new form of legal regulation which is capable of over-
coming the problems attached to the more traditional forms. This is what he
calls ‘reflexive’ law (Teubner, 1983: 246).
An effective way of explaining the nature of reflexive law is to compare
it with substantive law.7 Substantive regulation imposes specific values on
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 551

other subsystems when attempting to solve social conflict and is character-

istic of the regulatory structures associated with the welfare state. By incor-
porating specific, substantive social aims within its normative structure, and
by trying to enforce these aims within other subsystems, the legal system
becomes insensitive to the normative autonomy of other subsystems. To
borrow a term from Habermas, law ‘juridifies’ other subsystems by
imposing upon them substantive norms that are alien to their internal values.
(Habermas, 1987: 356). The central problem here is that norms are justified
from the point of view of the legal system, but are not necessarily so justified
from the point of view of other subsystems. This leads to functional stagna-
tion for the legal system because social conflicts will remain unsolved, or –
in so far as this is possible – conflicts will be resolved outside the legal
system. Disputants in subsystems external to law might then question
whether the law is worth following generally, or whether legal decisions
actually represent plausible or justifiable (adequate, reasonable or useful)
solutions to actual conflict.8 Moreover, companies, interest groups or private
citizens, whenever they are parties to a conflict, might not want to seek a
remedy to their conflict from the legal process as they will have to employ
lawyers to translate their dispute into one that is compatible with the legal
construction of the social world, but perhaps incompatible with the way
these potential litigants view the problem. The uprising of various forms of
alternative conflict resolution can be seen as a result of this development, and
it raises the question whether law can, in its substantive formulations, really
solve the social problems that it was designed to solve in the first place. If
law’s general purposive orientation must be geared towards solving
problems of social conflict – and if it is this which must be the leading
rationale for the normative orientation for the theory of autopoietic law –
then it would appear that the legal subsystem must adapt. Reflexive law,
then, should be seen as an attempt to pursue the goal of regulation without
the unintended and undesirable side-effects associated with substantive law
(Brownsword, 1993: 255). Teubner says:

A reflexive orientation does not ask whether there are social problems to which
the law must be responsive. Instead it seeks to identify opportunity structures
that allow legal regulation to cope with social problems without, at the same
time, irrevocably destroying valued patterns of life. (1983: 274)

Law, therefore, operates as an ‘external constitution’ which can promote

‘discursive decision making processes and consensus-orientated procedures
of negotiation and decision’ (Teubner, 1983: 275; Habermas, 1981: 544, 1986)
and so the problem of non-congruence between the legal system and other
subsystems is solved by adopting a reflexive approach to law.9 Furthermore,
in the face of increasing social differentiation, the development of a reflex-
ive legal rationality becomes necessary in order to muster an adequate
response to the problems associated with other, dysfunctional, forms of legal
regulation.
552 SOCIAL & LEGAL STUDIES 11(4)

ROLES, LANGUAGE AND REASON IN LEGAL THEORY

On the premises of the theory of autopoietic law, legal operations are beyond
the control not only of those who are subject to the system, but also of those
who are agents of the system. The theory of autopoietic law construes legal
norms and legal decisions as properties of, or operations by, ‘the system’,
rather than as products of practical deliberation performed by people who
act on behalf of public institutions. This claim is founded upon an ontological
perspective which refuses to prioritize critical human reason as a theoretical
starting point. From the viewpoint of this particular social theory, the indi-
vidual is relatively unimportant, and based upon a ‘falsely selected anthro-
pocentrism’ (Luhmann, 1984: 141). Even if there are some ‘co-evolutionary’
or ‘symbiotic’ links between the systems contained within the individual and
the legal system, the latter seems to take on a life of its own (see Paterson,
1995: 217). Thus, the claim is made that the legal system is able, at least in
principle, to operate without agents. The legal system is running itself and
there are no genuine and deliberative roles for individual actors such as
lawyer, judge, police or official within the system. Teubner (1992) explains
this point by creating an analogy to an episode in a fictional novel by
Marquez (1982). Marquez describes a murder in a north Colombian village,
and Teubner uses this description as a metaphor for his theory of autopoietic
systems. For Teubner, the activities leading to the murder display: ‘. . . a total
separation of inner motives and external action, with the dissociation of
psychic and social processes, with the reciprocal closure of the two auto-
poietic systems’. He continues: ‘Against the will of all the participants, even
of the murderers themselves, a murder is committed – or, rather, the murder
commits itself’ (1982: 454).10
Metaphorically we might make some sense out of this analogy. But given
the high philosophical stakes involved, it is hardly compelling at the most
theoretical or justificatory level. We must remember that the fundamental
issue is whether or not it is ontologically rational to dispense with the role
of agency in characterizing the phenomenon of legality. A systems approach
is perfectly coherent in a general approach to social phenomena, and there is
no objection to it per se. Durkheim’s theory of suicide, for example, takes
the most isolated, singular and poignant aspect of human experience and
shows how, causally, it might be related to the large-scale morphologies of
the social structure. But we should also note that Durkheim’s great contri-
bution to social theory is to show the inadequacy of a purely voluntaristic
and individualistic attempt to render the causal complexity of history and
society susceptible to explanation: it is not to assert the non-existence of
agency, or the causal irrelevance of individual and collective practical reason
in social development. Durkheim shows that human consciousness develops
and operates in a structural context, and that it is theoretically indispensable
to take account of system-imperatives in relation to agent and group inten-
tions and mores if we are to hope to achieve a comprehensive and systematic
grasp of the nature of social action. There is no paradox here and the point
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 553

might be accurately and mundanely expressed in Marx’s aphorism that, ‘men

make their own history, but not in conditions of their own choosing’. This
axiom would undeniably unite the two great ‘structuralists’ of modern social
theory – even if it left them a great deal to argue about in respect of precisely
how, and to what extent, we might characterize the influence of ‘the social
system’. But the bluntly undialectical assertion of ontology that we are being
asked to accept in autopoietic theory is entirely insensitive to the profoundly
important issue of just what it is reasonable to assume a priori about the
causal processes operating in society and history at this crucial methodo-
logical point of departure into sociological analysis. For after this step,
empirically and ‘theoretically’, there is no room for argument. There is
merely the decision to characterize and explain the legal process ‘autopoiet-
ically’ – or to proceed with an agential analysis. The two approaches are,
quite genuinely, incommensurable once the decision to exclude agency has
been taken. The ontological Rubicon has been crossed, and no ‘empirical’
discovery can ever alter our course in this regard.
Insensitivity to this issue, and a failure to grasp just what sort of an
assumption underpins the ontology of autopoiesis is, in our view, an obstacle
to a reasonable discussion of the approach taken by Teubner and Luhmann.
Paterson, for instance, suggests that we should not take Teubner too seri-
ously in his use of analogy, and rather should see his theory as merely an
explanation of ‘why’ the crime was committed rather than by ‘whom’
(Paterson, 1995: 218). But this does answer any of the following questions:
In what way does the legal system exist independently of the actions of
human beings? How does the system perform its functions distinct from
individuals? Who or what keeps the law isolated in its own autonomous
domain? And perhaps most importantly, does not the idea of the basis of
individual obligation appear to be central to the very attempt to distinguish
law from other systems of coercion; and can this issue be plausibly sidelined
to allow a systems theory to proceed unhampered by serious questions of
individual, reasoned claims to rights and duties ranged against the claims of
authority and competence of officialdom? These questions not only remain
unanswered by the proponents of autopoietic theory, they simply do not
arise within the purity of the systems framework. This, as we have attempted
to explain, is an indication of unsolved problems at the ontological level, and
thus shows that the ‘first order’ expression of the theory is not entirely defen-
sible. To elaborate on this argument we shall pursue two lines of critique: the
first Habermasian, the other related to a more traditional defence of socio-
logical actionism.
In the theory of both Luhmann and Teubner, it is possible to identify the
idea that events can occur in several subsystems simultaneously (Luhmann,
1988: 342; Teubner, 1989: 745), and that subsystems, although they are
normatively closed, can nevertheless, through the dimension of what they
call cognitive openness, communicate with each other.11 Habermas, address-
ing these two issues, believes that there is an inconsistency in autopoietic law
that is not easily resolved. He says:
554 SOCIAL & LEGAL STUDIES 11(4)

On the one hand, legal discourse is supposed to be trapped in its self-

reproduction, constructing only its own internal image of the external world;
on the other hand, it is supposed to use ‘general social communication’ so that
it can ‘influence’ general social constructions of reality, and in this way influ-
ence those of other discursive worlds as well. It is difficult to reconcile these
two statements. If the first statement is correct, then one and the same act of
communication can belong to two or more different discourses, but the identity
of the corresponding utterances in these discourses would be recognizable only
from an objective standpoint, and not from the perspectives of the involved
discourses. Otherwise one must postulate the possibility of a translation
between them – a hermeneutic relation that would explode the closure of each
of the mutually impenetrable circuits of communication. (Habermas, 1996: 53)

In other words, either the extreme social differentiation that is assumed in

the theory of autopoietic systems is not so extreme after all because it takes
place in a context of mutually shared meaning, that is, in Habermas’ words
‘within the boundaries of a multifunctional language’ (1996: 55), or else the
postulated normative closure is not as radical as it is presented, in that it must
be possible to break the code from the outside and extract information from
the system – which again presupposes some form of shared basic rationality.
Habermas thinks this basic form of rationality is embedded in ordinary
language. Thus, for Habermas, the biggest problem with autopoietic theory
is the simultaneous occurrence of the following three suppositions: initially,
that the system-specific interpretations of the communicative events have
epistemic primacy; second, that some form of communication between the
systems is possible; and finally, the denial of the existence of a general
communication medium through which the function-specific language can
be translated into a more simple, but shared language.12 Habermas takes
exactly the opposite position from proponents of autopoietic theory and
proposes the epistemic primacy of ordinary language:

Like that other anthropological monopoly, the hand, ordinary language – with
its grammatical complexity, propositional structure, and reflexivity – possesses
the merit of multifunctionality. With its practically unlimited capacity for
interpretation and range of circulation, it is superior to special codes in that it
provides a sounding board for the external costs of differentiated subsystems
and thus remains sensitive to problems affecting the whole of society. The ways
of defining and processing problems in ordinary language remain more diffuse,
are less differentiated, and are less clearly operationalized than under the
code-specific, unidimensional, and one-sided aspects of cost/benefit, command/
obedience, and so on. In return, however, ordinary language is not tied down
to just a single code but is inherently multilingual. It does not need to pay the
price of specialization, namely, deafness to problems formulated in a foreign
language. (1996: 55)

This point made by Habermas takes us to the second line of critique which
emphasizes the notion of the self as a bearer of social roles. Proponents of
autopoietic theory assert that human beings are in some sense the ‘bearers’
of systems in that they are necessary for the continuing existence of these
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 555

systems. At the same time, however, the theory makes it clear that human
beings do not participate in the system, in the sense of having an influence
on how the system behaves. For this to be true, the point must be that human
beings serve the function of executing the autopoietic operations of the
system, rather than interpreting what the system requires of them. Even if
we accept (as in Paterson, 1995: 216) that there is some kind of ‘causal’
connection between the system and individuals, there are several problems.
Initially, human beings are themselves made up of a number of symbiotically
connected autopoietic systems (1995: 217). But this does not square well with
the idea that human beings are bearers of the legal system, for whenever there
is a conflict between the legal system and, say, the psychic system, the psychic
system, according to autopoietic theory, will not simply accede to, and
execute, the decision of what appears to it as an ‘alien’ system rationality. In
other words, the theory, on its own premises, is unable to explain precisely
how the system can run itself.
Furthermore, even if we accept that human beings are ‘bearers’ of the legal
system, then the individual who is acting within the system seems to be
determined by the system, and hence, free from responsibility. It is not s/he
as a person who acts, but the system. Certainly, the actual physical events
are performed by identifiable human beings, but that is irrelevant because
they do not act. Rather, they are consumed by the system in its process of
autopoiesis. On this account individual responsibility disappears into system
responsibility. If anything goes wrong, the system is to blame.13 The problem
with this is not so much the account of how the system works, for there is
no problem with the claim that the legal system might, in some sense, be seen
as a social or structural entity, or an institution with imperatives and a ‘ration-
ality’ of its own. Rather, the problem lies with the account of how the indi-
vidual works. The individual must be seen as extremely fragmented: made up
of, and connected to, many different subsystems. But each subsystem has its
own system rationality, and thus the individual has to apply different ration-
alities depending on the system to which s/he is presently connected. The
person is never really a self, but always the bearer of a social role created by
one of the many subsystems. All the actions that are being performed by the
individual can be ascribed to some social role. It is never the personally
embodied and individually responsible self that acts, but always the social
role. But at this point, the autopoietic theory trades heavily on a simplistic
(undialectical) tradition of social and structural determinism which has a
notorious gap in its explanatory scheme. Martin Hollis, discussing the
relationship between personal identity and social identity in his Models of
Man says:
The strange lacuna is the one just where we hope to learn who plays the roles.
Admirers of Erving Goffman, for instance, are well served with nuanced, coolly
sardonic tales of how actors in the life-world play their parts with varying styles
and skill, at varying degrees of distance and for varying ends. Interactions are
far from mechanical and the actor can keep control of them by means of second-
ary adjustments, distancing rituals and elusive negotiations. If he succeeds, he
556 SOCIAL & LEGAL STUDIES 11(4)

has an identity not merely defined for him or thrust upon him; and this identity
is crucial in understanding and explaining his conduct. Although ‘identity’ here
is in part a set of attributes, it refers also to a subject or substance who manipu-
lates his attributes and his Umwelt. (Hollis, 1977: 88)

Any social theory that purports to explain social phenomena, no matter how
abstract, must at some point account for the relationship between individuals
and their social environment. Thus, autopoietic theory, despite its emphasis
on communications as the basic epistemic (and ontological) units, must at
some level conceptualize the role of the bearers of the system, and hence
account for the relationship between the system-communications and the
subjects that are somehow related to the system. In other words, the theory
must unveil what lies behind the mask of communications. Bankowski
(1994) presents this objection in a more concrete fashion by asking how
autopoietic theory would respond to the debate between Detmold and
Susskind concerning the statement, ‘I sentence you to death, but I think the
death penalty is wrong’ (Detmold, 1984; Susskind, 1986). Detmold thinks
that a judge contradicts himself when he says he makes such a statement.
Susskind replies by arguing that there are many normative spheres within
which the individual is operating. Hence, acting as a judge, within the norma-
tive sphere of law, the death penalty is the normatively justified course of
action. In another normative sphere, the same judge (now as the father, for
instance) may consider the death penalty to be wrong. But Bankowski chal-
lenges this view on the grounds that even though the individual might be
filling out a social role – in this case the role of a judge – it is implausible to
claim that it is an isolated part of the individual that makes the decision,
rather, it is the person as a whole. Thus Bankowski says:

It is he as judge, father, moralist etc. [i.e. as a whole person] that takes the
decision and not a bit of him. That would be to accept the metaphysical illusion
and lose oneself in a particular normative sphere. Responsibility is evaded
because he does not decide but the self constructed by the theory of the norma-
tive sphere that he has chosen does. At the same time he forgets that he is also
dealing with some particular individual. For what happens is that he deals with
that person as constructed by the theory, as an offender and not as a person.
(1994: 263)

We would endorse this view as a tenet of methodological rationality, and

reject the implausibility of the ‘metaphysical illusion’ that Paterson’s rever-
sion to a symbiotic relationship between the psychic and social system seeks
to perpetuate in defence of the systems approach.14 And this rejection of the
fragmented and role-determined self can be extended: when a role is being
played, the person, as Hollis and Goffman suggests, might play the role in a
number of different ways. Just as different players lay emphasis on different
character traits in Hamlet or Macbeth, so too the role of the judge might be
played differently by different actors. The point here is that because the self
embodies several roles (normative spheres) within it, when interaction takes
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 557

place between role and self, this interaction simultaneously constitutes an

interaction between the various roles. The effect of this is that the self effect-
ively serves as a medium through which communication between the various
subsystems is made possible. This point can be pressed further.
According to autopoietic theory, the individual within the system acts
according to a system-constructed self when s/he makes decisions. If we
assume that the roles within a subsystem might be run by different indi-
viduals, we must question whether these individuals all respond in the same
way when there is a requirement that the subsystem perform one of its tasks.
That is, if there only exist system-constructed ‘selves’, it seems that all the
different ‘selves’ that belong to a given system – for example, all the judges
within the legal system – must be functionally identical, and must respond
in an identical way to identical inputs. If they do not, then how does auto-
poietic theory account for the existence of competing versions of what is an
appropriate system response to a given input? Quite obviously, institutions
or subsystems are not couched in a clear and unequivocal form of rationality,
to which all its bearers uniformly adapt. Rather, the identity of an institution
is always open to questioning as is the continuing unfolding of debate about
the rationale of the institution – a debate that shapes the institution and
thereby the roles that exist within it. Thus the practical reasoning of system-
constructed ‘selves’ cannot be seen to be entirely separated from the system,
and nor can it be seen to be entirely separated from the selves; hence, it seems
much more plausible to suggest that the individuals within the system are
making choices on behalf of the system (see also the lucid critique by Beck,
1994).
Reconsidering law in light of this critique, it follows that if the legal system
was genuinely normatively closed, then presumably all persons within the
system would act in the same manner. That they do not means that there is
some input within their decision making that cannot be accounted for
directly by the normatively closed legal system. In order to account for
divergences in the decisions of legal officials, we must acknowledge that such
officials are exhibiting traits associated with cognitive openness and when
officials are being cognitively open, they are somehow using extralegal sources
when engaging in legal processes. But what are these extralegal sources and
what is their exact relationship to the rationality of the legal system?
We shall revert to this question presently in our discussion of the poten-
tial of reflexive law for solving disputes in a hypercomplex society. But at this
point we think that enough has been said at a foundational level to question
the validity of the assumptions of autopoietic theory, and to establish the
conclusion that there are serious difficulties with its attempt to characterize
the phenomena of law. The idea of normative closure and the metaphysical
apparatus intended to explain the autopoietic event and reproduction of the
subsystem cause the most theoretical concern. And it seems, in our view, not
only possible, but infinitely more plausible to consider subsystems as being
inhabited by individuals who employ their own practical reason to make
decisions on behalf of the system, rather than the reverse. This conclusion,
558 SOCIAL & LEGAL STUDIES 11(4)

however, can be taken further if it is combined with our initial theme of legal
autonomy and the functional role law has in maintaining the stability of the
total social system. In what follows, we argue that legal autonomy must be
understood to be legitimate if law is to be effective in maintaining stability
throughout the system. To achieve legitimacy, we argue that law’s autonomy
must consist in the production of substantive values commensurable with,
and communicable to, other subsystems, and it is this which informs the kind
of rationality that the law should be operating with.

NORMATIVITY AND AUTONOMY IN REFLEXIVE LAW

One of the key practical consequences of the legal theory advanced by

Teubner and Luhmann is the idea that subsystems should engage in a degree
of self-regulation and that law should take a step back from direct inter-
ference in the determination of substantive values that guide subsystemic
decision making. But the problem of ascertaining how the legal system
performs its social function of ensuring social stability and resolving social
conflict remains.
Focusing on this conception of the social function of law it is clear that,
whether society is complex or not, the legal system must be able to deter-
mine and fix its normative supremacy over other forms of normativity found
in other subsystems if it is to resolve social conflict. So when complexity in
a society increases, the legal system must adapt to this development. It is
plausible that the inability of law to be responsive can produce what is
referred to by Habermas as a legitimation crisis (Habermas, 1976). From the
viewpoint of Habermas’s distinction between the ‘system’ and ‘lifeworld’,
this crisis occurs when groups of individuals think that their shared values in
a ‘lifeworld’ are normatively overridden by legal regulation incompatible
with what are considered to be essential patterns of justifiable or acceptable
conduct. Albeit from a different perspective, proponents of autopoietic
theory identify the same problem concerning the unresponsiveness of the
law.
The study by the Danish anthropologist Thomas Højrup of small
communities in West-Jutland is an effective illustration of legal unrespon-
siveness (Højrup, 1989). He shows how these communities have, over
centuries, developed a way of life in which fishing is mixed with other occu-
pations. Danish law was modified in 1978 so that professional fishing was
only allowed if it was fully professional, and this made the semi-professional
fishing practices in the West-Jutland communities illegal. From the point of
view of the local population, there is no justification for such regulation,
which they perceive as unnecessarily rigid and unjust, and therefore there is
some justification in characterizing this situation as one that established a
legitimation crisis.
If this use of example is correct, it is necessary to consider the circumstances
under which individuals, acting upon the basis of the varied normative criteria
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 559

they hold, may accept or reject different forms of legal regulation. This
should then indicate a rationale for law, which can be applied when there is
a risk that traditional forms of regulation cannot ensure the necessary legit-
imacy and social effectiveness.
To develop this point, it is initially important to examine why reflexivity
may provide a solution to problems concerning the legitimation of the law.
Reflexive law, as far as proponents of autopoiesis are concerned, may
overcome problems with legitimation because it focuses on the adaptation of
legal rules to a changing normative environment. Specifically, it is character-
ized by an orientation towards the establishment of organizational and
procedural frameworks for dispute resolution at a local level and by allowing
private or semi-official organizations to participate in various ways in regu-
latory programmes. Substantive law, as has been seen, is characterized by its
orientation towards certain specific goals which are defined by the law and
thus settled before the law is put into practice. While substantive law operates
with a relatively fixed set of values, reflexive law absorbs values from the
disputants into the normative basis of legal decisions.
Perhaps the relaxation of law to allow previously excluded groups locus
standi in various parts of the judicial review process is an empirical reflection
of a move towards a more reflexive legal rationality. In the United Kingdom,
for example, environmental groups such as Greenpeace can now challenge
the validity of executive and administrative decisions even when their rights
are not affected directly and when it is in the public interest.15 This may
reflect a willingness from legal decision makers to include a broader number
of value systems within the review process and to make the legal process
more inclusive and pluralistic (see Feldman, 1992). A more mundane explan-
ation, however, is that giving such groups locus standi is merely to ensure that
the rules that govern executive actions are correctly applied.16
Either way, Danish law has developed further with regard to environ-
mental regulation and perhaps can be seen as a more authentic version of
reflexive law. Since the middle of the 1980s, Greenpeace has been given
competence to demand information about certain decisions made by local
authorities which may affect the environment. Furthermore, Greenpeace has
competence to appeal these decisions to a ministerial level if they should find
that the decisions violate the obligation to consider environmental issues in
the decision-making process. The effect of this is that Greenpeace is given
responsibility to review decisions made by local authorities and the latter are
forced to take environmental issues seriously.
If we assume that this development in Danish and English law is, to a
greater or lesser extent, an example of reflexivity, then it may demonstrate
how we can avoid legitimation crises. For example, the problems with regard
to legitimation that resulted from the regulation of fishing practices could
have been solved if the values of those who were likely to be affected by the
regulation were taken into account and balanced against other competing
value systems. But assuming that the judiciary remains the primary insti-
tution for dispute resolution, two interrelated problems remain.
560 SOCIAL & LEGAL STUDIES 11(4)

The first problem concerns the issue of legitimacy. From the point of view
of social coordination, the decisions that emerge from the adjudicative process
must normatively override the inputs that enter it. If the legal decision does
not do this, then the dispute is not solved. At some point, the adjudicative
process must normatively transform the inputs that are fed into it. So, for
example, in the context of standing in environmental disputes, each partici-
pant who resorts to adjudication will have its own set of divergent values and
therefore each participant will have to consider a particular decision or
compromise fair to the extent that they will be prepared to give up substan-
tive values that they brought to the law. The legitimacy of a decision must
depend upon the way in which the decision is made or, rather, the legal
process. This, in turn, means that the legal process has to yield such powerful
normative reasons that the parties to the dispute are willing to follow the
legal decision and give up their prior normative commitments at least for
the purposes of orientating their action in the particular case in question.
But the presumption of the validity of proceduralism as a rationale for
law is unlikely to ensure this level of normative commitment. A generic
argument is required to demonstrate why, regardless of the values or inter-
ests held by disputants, all parties must concur with the decision made in
line with the legal procedure.
Second, while one can explore the possibility and problems of providing
such a generic argument (see Pettit, 1982), we can foresee a greater practical
problem with a purely procedural form of law. To explain: imagine a dispute
over two highly divergent values which are brought to court in a reflexively
rational legal order. If – because the legal system is reflexive – the judge
cannot take a distinct substantive stance on these two values, then the only
possible solution is one to which both disputants are disposed to agree. This,
in fact, would not be a legal decision at all, at least, not as is commonly under-
stood. But even if it was, and even if sometimes procedures alone can compel
disputants to agree, it is not difficult to imagine conflicts from which neither
party is likely to compromise its values. The facts of the recent Refah case
which was decided by the European Court of Human Rights illustrate this
point.17 Refah Partisi was a political party which gained 22 percent of the
vote in the Turkish general election in 1995. This made it Turkey’s largest
single political party. In 1997 the Turkish Constitutional Court dissolved
Refah Partisi because it threatened to abolish the secular constitution in
Turkey and to replace it, forcibly if necessary, by Shariah law.
In this particular case, it is not hard to see that if left to legal procedure,
the two sets of individuals holding these conflicting value systems would not
be able to reach a compromise. This is essentially because the sets of values
that they hold are incommensurable. Further, as the values held by indi-
viduals in society diverge (that is, as society becomes more complex), reflex-
ive law would be less likely to be able to solve social conflict.
Is it possible for reflexive law to overcome these problems? Initially, a way
out of the problem which concerns the inability to solve social conflict would
be to judge competing claims or values brought by disputants to the law
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 561

according to a set of substantive values, which would function as a form of

meta-law. The problem-resolving potential of this solution would be brought
out in so far as judges (and others who make legal decisions) would be able
to decide that certain values ought to normatively override others as a matter
of substantive principle and hence resolve disputes between highly divergent
value systems. In the Refah case, for example, it could be decided that the
general protection of human rights incorporated in the secular constitution
is more important for the maintenance of a democracy than the acceptance
of a particular political party that is committed to abolishing secularism in
favour of religious law. This is exactly what the Turkish Supreme Court and
the European Court of Human Rights decided. But, once again, this decision,
and the set of substantive norms upon which it is based, are (presumably)
not accepted by a relatively large number of people in Turkey and therefore
reproduce the circumstances under which a legitimation crisis may emerge.
With these comments in mind, the problem of legality in complex societies
can be stated as follows: substantive law lacks the legitimacy to solve conflict,
and reflexive law appears instrumentally ineffective as a means of solving
disputes concerning radically divergent opinions. If we reject substantive
law in favour of reflexive law, we are replacing the debilitating process of
‘juridification’ with a form of law that cannot effectively ensure the stability
of society. If law is relatively ineffective in resolving social conflict in either
of its substantive or reflexive rationales, then we have a society that is highly
complex, but which is – from the point of view of the present state of legal
regulation – in a quasi state-of-nature. We can call this a problem of legal
inertia.
The solution to the envisaged problem requires that we go beyond the
impasse of the legitimation problem associated with substantive law, and
the potential inertia and reversal of the rule of law that ensue from reflexive
law. Our first step in this regard is to establish a platform of analysis upon
the distinction between what we shall call passive reflexive law and active
reflexive law. Passive reflexivity refers to a legal rationale that is solely
procedural, and in fact presupposes some form of self-regulatory processes.
Active reflexivity refers to a legal rationale that, like passive reflexive law, is
open to values that are brought to the law by disputants, but which simul-
taneously has a pre-fixed normative apparatus that makes it possible to weigh
and balance competing claims according to substantive criteria.18
Active reflexive rationality operates, at the judicial level, by absorbing and
applying values accepted by those individuals or groups that are in dispute.
However, instead of merely acting procedurally and thereby generating the
potential for legal inertia, active reflexive law works by attempting to rank,
weigh and balance the values that are being absorbed according to substan-
tive criteria which are fixed prior to the dispute taking place. On this approach
legal norms take on a kind of meta-status vis-à-vis other social norms. While
this kind of theory, prima facie, might allow us to cut through the problem of
legal inertia, it could also possibly be the worst of all worlds: it may be that
active reflexive law will reproduce the conditions by which legitimation crisis
562 SOCIAL & LEGAL STUDIES 11(4)

will occur but with the attendant problems of reflexive law. This argument
only follows because with substantive law, pre-fixed substantive norms in the
legal system are isolated from other value systems. However, if (at least) the
judicial element of the legal process, when its substantive rationale is
considered, attempts to apply norms that are accepted within the social system
generally, then the problems with legitimation crisis may be averted. This is
because the judge is applying generally held values which have been incor-
porated into the law rather than specifically legal substantive values which are
incongruent with values held in other subsystems. From this point of view,
active reflexive law must not only be sensitive to the specific values of the
disputants, but must be sensitive to those meta-norms that are accepted in
society; for it is these norms that help judges to make decisions that are legiti-
mate and effective. Legitimacy is promoted because the norms that are applied
in the legal decision-making process are ones which are considered justifiable
for society as a whole rather than for a particular isolated group. Effective
decision making is promoted because substantive norms can be employed by
judges to evaluate the relative normative merits of competing claims.
But how far does this proposal get us? Perhaps meta-law resembles
Dworkin’s (1994) principles. But can the assumption be made that judges can,
through a Herculean effort, find such principles? The facts of the Refah case,
to some extent, reaffirm the point made by proponents of reflexive law: that
different groups in a society accept normative values that are incommensu-
rable and therefore it is implausible to build legal rationality on a general-
ization about the correct substantive content of legal norms. This brings us
back to the question that was raised above: how do we handle a situation in
which a relatively large number of people do not accept the principles of
balancing and priority that are contained in the meta-law? This is the heart
of the matter and two further elaborative points need to be made.
First, it must be stressed that for law to be legitimate it is not necessary
that all who participate in the legal discourse consent to the meta-law. It is,
fairly obviously, a condition of a complex society that no such consensus can
be brought about. Thus the lack of consensus poses the problem of legiti-
macy, but this problem cannot be solved by removing the cause. We must
think harder about this problem and must somehow create the tools that
make it possible for us to continually handle this problem in the best possible
way. As lawyers, the best we can do under these circumstances is to try to
determine, logically, what principles must be in place in a society for it to
survive and, more generally, to flourish, and it is these values that form the
meta-norms in our concept of reflexive rationality. Although this may sound
impossibly difficult, we should not overstate the case against the possibility
of developing such a framework of principles and this leads directly to the
second, and final, point.19
Attempts to ground these norms of meta-law in the face of an exponential
growth of social complexity are not a contemporary problem. In fact this
issue has been central to moral and political philosophy throughout history.
Rawls says:
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 563

There are periods, sometimes long periods, in the history of any society during
which certain fundamental questions give rise to sharp and divisive political
controversy, and it seems difficult, if not impossible, to find any shared basis
of political agreement. Indeed, certain questions may prove intractable and may
never be fully settled. One task of political philosophy in a democratic society
is to focus on such questions and to examine whether some underlying basis
of agreement can be uncovered and a mutually acceptable way of resolving
these questions publicly established. (Rawls, 1985: 226)

Reflexive law requires a search for a typology of goods to provide the criteria
required to weigh and balance the normative merits of competing claims in
the process of legal decision making. This search is identical to the search for
‘some underlying basis of agreement’ that Rawls refers to. Now, full
agreement is not possible in a modern complex society, but nor is such
agreement necessary for the law to be legitimate. What is important is that
the very process of searching for higher principles (i.e. meta-law) – a process
which is necessary for the weighting and balancing of competing claims to
come about – generates legitimacy. Entering into this dialectical discursive
process forces the litigants to relax their prior commitments and employ
higher principles which are embedded in society and incorporated into the
law. These principles can then be employed to support such litigants’ claims.
This solution relieves the problems associated with incommensurability of
values because, while claimants may not agree with the decision, they at least
will be able to perceive the reasonableness of the decision against such higher
principles. It is implausible to suggest that there will be agreement with all
legal decisions and, in fact, the adjudicative process of decision making
presupposes that there will not be. After all, such decisions are made in
favour of someone and against someone else. The point is that the principles
that are employed by the law are justifiable and transparent because they
accord with basic principles of practical reasonableness embedded in society.
This is, perhaps, the most we can expect of a legal system when considering
the issues of effectiveness and legitimacy in complex societies.
In conclusion, active reflexive law could be charged with being nothing
more than a rationalization of what is already present in modern legal systems.
This, to some extent, is true.20 But what has been argued here is a concep-
tion of legal rationality which may help to increase understanding of how it
might be possible to tackle some of the problems concerning the functional-
ity of law as moral diversity increases between individual agents in society.
Some may decline to agree with our positing of agent-centred practical
reason at the heart of deliberations on the rationality of law. However, we
are certain that the rejection of such reasoning by autopoietic theory is, in
effect, a reductio ad absurdum of the aspiration to realize a reflexive law
appropriate to a modern complex society.
564 SOCIAL & LEGAL STUDIES 11(4)

NOTES

We would both like to thank Stanley Paulson, Aileen McHarg, Synne Sæther Mæhle,
Stuart Toddington and Christine Wilmore for their invaluable help and advice.

1. We treat Teubner and Luhmann as proponents of the same theory, although we

are aware that there are differences between the two. Our critique is based upon
what we consider to be common strands in a theory of law as an autopoietic
system, or autopoietic law.
2. For responses by exclusive legal positivists to this conception of legal
autonomy, see Raz (1998) and Marmor (2001: 15–17, 19–21 et passim).
3. While this, perhaps with the exception of the idea of self-creation, is fairly
straightforward, it is difficult to present a non-controversial description of its
main features. See Patterson’s (1995) reply to Bankowski’s (1994) criticism of
autopoietic theory for an example of the potential disagreement.
4. Teubner considers that economics, politics, religion and law are examples of
subsystems. However, it is clear that the family, sports or social clubs and
networks of friends might be considered subsystems.
5. The key examples are formal and substantive law. Formal law is applied equally
to everyone. It is concerned with the setting up of a legal institution which does
not differentiate in its applicability according to the social status of its subjects.
Laws ought to be neutral with regards to the distribution of opportunities for
free and successful social action. However, the crisis of legitimacy that occurs
with regard to this type of law, so the argument goes, concerns the inability of
law to deal with the unequal position of parties in disputes or to be instrumental
in the attainment of social goals in general. Substantive law attempts to correct
these deficiencies by being orientated towards the achievement of certain
‘socially engineered’ goals (see Brownsword, 1993: 241).
6. Luhmann (1995b: 296ff) notes that normative self-reference and cognitive
other-reference seem to correspond to the distinction between concepts (which
are internal to the law) and interests (which are external to the law), both of
which are part of legal argumentation.
7. See Note 5.
8. For example, in contract law, disputes may be commonly solved via negotiation
and arbitration rather than seeking recourse to the law. The disputants may
consider that legal norms do not provide an adequate solution to a problem
from the point of view of the values they, as contractors, hold. See Collins
(1999: 5 et passim) and Vincent-Jones (2000).
9. See Teubner (1983: 276) for a clear example of how this might work.
10. If Teubner’s claim, as suggested by Bankowki (1994), is equivalent to the non-
existence of human beings then, perhaps from the point of view of autopoietic
theory, a murder, at least as it is commonly understood, cannot take place.
Merely, a physical and (possibly) psychic system has ceased to exist. However,
we would note that the epistemic primacy of system over agency seems to cause
immense confusion in relation to the notion of intentions, and murder requires
as a necessary condition, intention. It is the implausibility, in the absence of
agent-centred ontology, of recharacterizing – or ignoring as irrelevant to the
coherence of the concept of law – the entire contents of the category culpabil-
ity, that seems most difficult to reconcile with a defensible approach to concep-
tualizing legal phenomena.
11. See pp. 549–50 above.
12. Luhmann says, by way of a response to Zolo (1987), that autopoietic theory
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 565

does not claim for itself a privileged access to ‘external reality’. Rather, ‘[t]he
statement that “autopoietic systems exist” . . . means nothing other than that
the reality construction of the theory is based upon this assumption’ (1987:
348).
13. It should be noted, however, that in the literature we have studied on this
subject we have not once encountered the idea that the legal system – or any
other systems for that matter – might act wrongly.
14. On the assumption that Paterson is genuinely engaging with Bankowski, we
can read his reply as meaning that there is a symbiosis between the psychic and
the social within the individual. But this presupposes that it is possible for the
individual to distinguish between the two systems within him or her ‘self’ qua
a bearer of systems or roles (see Hollis, 1977). This contradicts the determin-
ism which is at the heart of autopoietic theory.
15. See, for example, R. v Her Majesty’s Inspectorate on Pollution, ex p. Greenpeace
(No. 2) [1994] 4 All E.R. 329. A similar relaxation of the juridical review
process has taken place in Denmark. For example, in 1994, Greenpeace was
allowed locus standi to challenge a decision by the Danish government to
construct a bridge between Copenhagen and Malmö on the grounds that the
planning of the construction of the bridge did not follow the rules laid down
in EU Directive 85/337. See Ugeskrift for Retsvæsen [1994] at 780. Note that
this development has not occurred in the same way at the level of European
Law (MacRory, 1995–6).
16. See R v Secretary of State for Foreign Affairs, ex parte World Development
Movement [1995] 1 All ER 611 at 617.
17. Case of Refah Partisi (The Welfare party) and Others v Turkey. Judgment
delivered 31 July 2001.
18. It should be noted that our argument in favour of an active reflexive law bears
a close resemblance to Nonet and Selznick’s (1978) idea of responsive law. Their
study encompasses a whole range of features pertaining to legal phenomena,
whereas our discussion of active reflexive law specifically aims at the problems
of legitimacy and effectiveness. Considering the characteristic features of
responsive law, we generally share Nonet and Selznick’s views with regards to
the dimensions of ‘legitimacy’, ‘rules’, ‘reasoning’, ‘discretion’ and ‘partici-
pation’ when characterizing reflexive law (1978: 16). But it would be impos-
sible, within the limits of the present article, to further compare active reflexive
law and responsive law and there are clear areas of divergence (for example, see
Nonet and Selznick’s idea of ‘expectations of obedience’, 1978: 92). However,
our main point is that Teubner’s conception of reflexive law, by purifying the
procedural elements of responsive law by eliminating the substantive elements
has lost more problem-solving potential than it has gained.
19. See Rottleuthner (1989b: 274–5 and 280–3) and Jansen (1998: 6).
20. See, for example, Alexy (2002).

REFERENCES

Alexy, Robert (2002) A Theory of Constitutional Rights. Oxford: Oxford University

Press.
Bankowski, Zenon (1994) ‘How Does It Feel to Be on Your Own? The Person in the
Sight of Autopoiesis’, Ratio Juris 7: 254–66.
Beck, Anthony (1994) ‘Is Law an Autopoietic System?’, Oxford Journal of Legal
Studies 14: 401–18.
566 SOCIAL & LEGAL STUDIES 11(4)

Brownsword, Roger (1993) ‘Towards a Rational Law of Contract’, pp. 241–71 in

Thomas Wilhelmsson (ed.) Perspectives on Critical Contract Law. Aldershot:
Dartmouth Press.
Collins, Hugh (1999) Regulating Contracts. Oxford: Oxford University Press.
Detmold, Michael (1984) The Unity of Law and Morality. London: Routledge and
Kegan Paul.
Dworkin, Ronald (1994) Taking Rights Seriously. London: Duckworth.
Feldman, David (1992) ‘Public Interest Litigation and Constitutional Theory in
Comparative Perspective’, Modern Law Review 55: 44–72.
Habermas, Jürgen (1976) Legitimation Crisis. London: Heinemann Educational.
Habermas, Jürgen (1981) Theorie des kommunikativen Handelns. Frankfurt am
Main: Suhrkamp.
Habermas, Jürgen (1986) The Theory of Communicative Action Vol. 1: Reason and
the Rationalisation of Society. Cambridge: Polity Press.
Habermas, Jürgen (1987) The Theory of Communicative Action, Vol. 2: The Critique
of Functionalist Reason. Cambridge: Polity Press
Habermas, Jürgen (1996) Between Facts and Norms. Cambridge, MA: MIT Press.
Højrup, Thomas (1989) Det glemte folk – Livsformer og centraldirigering. Copenhagen:
Institut for Europæisk Folkelivsforskning. Statens Byggeforskningsinstitut.
Hollis, Martin (1977) Models of Man: Philosophical Thoughts on Social Action.
Cambridge: Cambridge University Press.
Jansen, Nils (1998) ‘The Validity of Public Morality’, Archiv für Rechts- und Sozial-
philosophie: 1–15.
Luhmann, Niklas (1981) Ausdifferenzierung des Rechts: Beiträge zur Rechtssoziolo-
gie und Rechtstheorie. Frankfurt am Main: Suhrkamp.
Luhmann, Niklas (1984) Sozial Systeme. Frankfurt am Main: Suhrkamp.
Luhmann, Niklas (1986) ‘The Self-Reproduction of Law and its Limits’, pp. 111–27
in Gunther Teubner (ed.) Dilemmas of Law in the Welfare State. Berlin: Walter
de Gruyter.
Luhmann, Niklas (1988) ‘Closure and Openness: On Reality in the World of Law’,
pp. 335–48 in Gunther Teubner (ed.) Autopoietic Law: A New Approach to Law
and Society. Berlin: Walter de Gruyter.
Luhmann, Niklas (1995a) Social Systems. Stanford, CA: Stanford University Press.
Luhmann, Niklas (1995b) ‘Legal Argumentation: An Analysis of its Form’, Modern
Law Review 58: 285–98.
MacRory, Richard (1995–6) ‘Environmental Citizenship and the Law: Repairing the
European Road’, Journal of Environmental Law 8: 219–35.
Marmor, Andrei (2001) Positive Law and Objective Values. Oxford: Oxford University
Press.
Marquez, Gabriel Garcia (1982) Chronicle of a Death Foretold. London: Cape.
Maturana, Humberto and Francisco Varela (1980) Autopoiesis and Cognition: The
Realization of the Living. Dordrecht: Reidel.
Nonet, Philippe and Philip Selznick (1978) Law and Society in Transition: Towards a
Responsive Law. New York: Harper.
Paterson, John (1995) ‘Who Is Zenon Bankowski Talking To? The Person in the Sight
of Autopoiesis’, Ratio Juris 8: 212–29.
Pettit, Phillip (1982) ‘Habermas on Truth and Justice’, pp. 207–29 in G. Parkinson
(ed.) Marx and Marxisms. Cambridge: Cambridge University Press.
Postema, Gerald (1996) ‘Law’s Autonomy and Public Practical Reasoning’, pp. 79–119
in Robert George (ed.) The Autonomy of Law. Oxford: Oxford University
Press.
Pound, Roscoe (1954) Introduction to the Philosophy of Law (Revised edn). New
Haven, CT: Yale University Press.
 CAPPS & OLSEN: LEGAL AUTONOMY AND REFLEXIVE RATIONALITY 567

Rawls, John (1985) ‘Justice as Fairness: Political not Metaphysical’, Philosophy and
Public Affairs 14: 223–51.
Raz, Joseph (1979) The Authority of Law: Essays on Law and Morality. Oxford:
Clarendon Press.
Raz, Joseph (1998) ‘Postema on Law’s Autonomy and Public Practical Reasons: A
Critical Comment’, Legal Theory 4: 1–20.
Rottleuthner, Hubert (1989a) ‘A Purified Sociology of Law: Niklas Luhmann on the
Autonomy of the Legal System’, Law and Society Review 23: 779–97.
Rottleuthner, Hubert (1989b) ‘The Limits of Law – The Myth of a Regulatory
Crisis?’, International Journal of the Sociology of Law 17: 273–85.
Susskind, Richard (1986) ‘Detmold’s Refutation of Positivism and the Computer
Judge’, Modern Law Review 49: 125–38.
Teubner, Gunther (1983) ‘Substantive and Reflexive Elements in Modern Law’, Law
and Society Review 17: 239–85.
Teubner, Gunther (1985) ‘Autopoiesis in Law and Society: A Rejoinder to Blanken-
burg’, Law and Society Review 18: 291–301.
Teubner, Gunther (1986) Dilemmas of Law in the Welfare State. Berlin: Walter de
Gruyter.
Teubner, Gunther (1988) Autopoietic Law: A New Approach to Law and Society.
Berlin: Walter de Gruyter.
Teubner, Gunther (1989) ‘How the Law Thinks’, Law and Society Review 23: 727–57.
Teubner, Gunther (1992) ‘Regulatory Law: Chronicle of a Death Foretold’, Social &
Legal Studies 1: 451–75.
Teubner, Gunther (1993) Law as an Autopoietic System. Oxford: Blackwell.
Vincent-Jones, Peter (2000) ‘Contractual Governance: Institutional and Organisa-
tional Analysis’, Oxford Journal of Legal Studies 20: 317–51.
Zolo, Danilo (1987) ‘The Epistemological Status of the Theory of Autopoiesis and its
Application to the Social Sciences’, pp. 67–124 in Gunther Teubner (ed.) State,
Law, Economy as Autopoietic Systems. Berlin: Walter de Gruyter.