“A man is always a teller of stories, he lives surrounded by his own stories and those of other people, he sees everything that happens to him in terms of these stories and he tries to live his life as if he were recounting it” Sartre, The Words (1981).

Law is full-contact storytelling. The view that the order of law is the product of actors’ storytelling contrasts with the traditional theory of legal ordering that the materials of law – commands from sovereign legislative bodies, judges’ orders, etc. – establish principles that might be interpreted at some objectively determinable level of generality that constrains legal decisions in future concrete disputes. Critiques of the traditional notion of legal ordering have failed to identify any mechanism for the apparent order in law.

This article suggests that the order we perceive as law is the product of conventions in legal storytelling on the part of legal actors. These conventions result in a limited array from which judges and other legal decision-makers choose a resolution for disputes according to their own governing narrative conventions. Legal actors make conventional narrative moves out of a need to reduce the complexity of legal disputing, a complexity that legal training renders them ill-suited to address.

The traditional, hierarchical conception of legal ordering is bottomed on linear narrative in general and a formalist conception of law, in particular. By contrast, a complexity-based view of law finds the source of order in the undirected interactions of legal actors. Legal systems persist to the extent that they provide predictable resolutions to disputes that manage to serve the interests of stakeholders while commanding perceptions of legitimacy among those affected by the idea and the reality of decision-making[1]. Others have described how agents’ storytelling both constructs legal outcomes and distorts litigants’ interpretation of the world[2]. I will concentrate on how, at least in autonomous legal systems that include iterated resolution of disputes, complexity provides a frame for understanding how order in legal systems might arise as the epiphenomenon of the interactions of storytelling by legal agents and the power relations and cultural contexts in which those stories are performed. The text explains the processes by which a predictable legal regime emerges from the interactions of legal actors; inserts employ the example of landlord-tenant law in the U.S. to demonstrate the operation of these processes and the practical implications of this view of legal order.


Complexity is a developing field of science studying how parts of a system interact to give rise to collective order and how systems sustain in the face of perpetual novelty in the absence of centralized control (Bar-Yam, 1997).

Most legal systems are complex adaptive systems (CAS) comprised of legal decision-makers acting according to each one’s own limiting heuristics. The product of their interactions is law. In other words, to find the order in law, you have to closely examine how legal actors coordinate their behaviors to navigate the big, fuzzy, indeterminate and un-bordered field of social disputing and control. What you will find, I believe, is that legal actors coordinate their actions through the medium of storytelling, the conventions of which are constrained by overarching narratives about legal systems that provide the form and substance that we see as law.

Legal systems

A sustainable legal system must meet three requirements: (1) it must produce a final decision; (2) that will be viewed as legitimate; and (3) that is predictable on some scale[3]. Where a legal system fails to provide these, parties seek out alternatives to legal disputing: moots, feuds, opt-out, ostracism or shame (Fuller, 1978). Where the results of a legal decision are so predictable as to approach certainty, the same strategies will prevail, as the party likely to be disadvantaged will seek other options – if she has any choice at all.

Order in legal systems

Traditional, formalist theories of law (now represented in natural law, legal process, law-and-economics and assorted positivistic movements) describe a highly ordered, hierarchically-directed regime.

The Realist movement arose in reaction to formalist accounts, with a focus on the individual personality of the judge as the pivotal factor in legal decision-making (Frank, 1930).

Critical Legal Studies (CLS) and its progeny of critical theories, including critical race and feminist theories, follow the Realists in undermining the determinacy of law, but see law as a vehicle for maintaining extant power relations.

Their shared indeterminacy thesis describes a legal system without a source of order. Still lawyers, some very successfully, act as if the outcomes of legal decisions are predictable and clients rely upon their predictions. To what might we attribute this perceived order, if not to some form of false consciousness?

The traditional narrative of legal ordering has been discredited[4], and empirical studies demonstrate that whole systems of law-in-action operate alongside or even in stark contravention to the order prescribed by legal materials, but no alternative narrative has arisen to explain the order that is seen in law[5]. Complexity’s narrative provides a lens through which legal agents might understand and hence challenge the system in which they are embedded[6]. Complexity science predicts that order emerges from the interactions of elements within the system; order is patterns of agent interactions over time (Goodwin, 1994; Holland, 1998). Complexity explains how the actions of dispersed agents, acting without conscious direction or central control, might produce a higher-level order that might be all the more robust for its uncoordinated redundancies and its agents’ blind complicity in accepting and reinforcing the system’s constraints.

Complex adaptive systems

CAS are known through elements that are widely shared by those systems, rather than any limiting definition. These elements include nonlinear relations from individual actions to group effects (agents acting according to simple rules produce complex collective effects); decentralized decision-making; control through feedback loops (reinserting the results of past performance); path dependence ( “you cannot unscramble an egg”); sensitivity to initial conditions ( “a butterfly flapping its wings in…”); scale independence[7] and the impenetrability of holistic functioning by looking at the parts of the system. In complexity terms, legal ordering is the emergent product of the interactions of law’s agents – lawyers, judges, litigants, court personnel – in the contexts of the legal institutions and roles they comprise through the stories they tell.

Law begins and ends in stories because that is how we, as humans, shape and give meaning to our worlds (White, 1985: 168). The legal system, as an institution devoted to giving meaning to our shared experience, organizes itself according to this ‘cognitive imperative’ (Newberg, 2001).

It is by now a commonplace that lawyers persuade by tapping into the storylines by which their audiences organize experience (Alfieri, 1989). Stories are developed in game-theoretic anticipation of others’ competing stories and undermining strategies. Lawyers are immersed in storytelling; they cannot survive without it (Alper, et al., 2005: 9).

Legal order is the play of competing stories. Lawyers engaged in full-contact storytelling likewise organize experience through their own deeply-held organizing narratives. Baskin posits that such mythic narratives act, “like the attractors of complexity science, holding people into a narrow range of behaviors that they know will be successful” (Baskin, 2004: 12). Within legal systems, however, the formalist account of law survives despite its predictive inefficacy.

The view of law as a CAS predicts that legal systems will tend toward order so long as its door-keeping agents coordinate their actions through simplifying storytelling heuristics. If so, changes in law might be more efficiently accomplished by affecting the stories through which legal agents interact and the structural constraints upon those stories.

The resolution of landlord-tenant disputes is consistent with the notion of an order constructed by agents, acting in parallel without a conscious plan, at the ground levels of legal systems.

The traditional narrative of landlord-tenant law states that, after the “tenants’ rights revolution” of the 1960s and 1970s, an implied warranty of habitability gained ascendance in every state except for Arkansas and Colorado. Under this warranty, a tenant was excused from paying all or a portion of her rent obligation if there existed on the rented premises conditions that rendered the tenancy unhealthy or unsafe. At the same time, retaliatory eviction doctrines developed to protect tenants from evictions in response to their efforts to enforce these rights.

This formalist account predicts dramatic differences in the results of legal disputes before and after the tenants’ rights revolution as well as differences in results and practices in Colorado and Arkansas as opposed to the remaining 48 states that have adopted some form of the warranty. Moreover, due to large differences in the way in which each state enacted its warranty (in terms of definitions, remedies and proofs), results should differ across jurisdictions.

A temporal and geographic comparison of landlord-tenant practice demonstrates that legal materials under-determine results in practice. Numerous studies of landlord-tenant courts demonstrate that the formalist account is wrong.

Those studies consistently find that tenants win only one to three percent of contested cases and rarely do tenants who assert valid warranty or retaliation defenses fare any better than those who stand silent or fail to show up at their hearings at all. ‘Trials’ last two minutes or less, witnesses are rare, and judges routinely ask the landlord to prove nothing or walk him through his part of the case. While courts that hear tenancy claims are notoriously overcrowded, and increasing caseloads uniformly result in speeding up the docket rather than increasing resources, similar dynamics are prevalent in non-busy, rural courts (Freidman & Perecival, 1976).

For example, a Baltimore rent court shows that it operates today exactly as it did prior to the tenants’ rights revolution of the 1970s (Bezdec, 1992). Tenants have a minute or less to tell their side of the story, and they overwhelmingly lost, in the sense they were evicted for non-payment irrespective of their claims that their units were not habitable (Bezdec, 1992). Blasi’s (2004) study of Los Angeles Housing Courts found that none of the tenants prevailed despite the author’s judgment that over half of those tenants presented claims that should have entitled the claimant to relief under the then-existing legal materials. Similar results were found in studies of Boston Housing Courts, including my own observations of only those cases in which tenants, according to legal materials, were mandated to win.

Tenants do somewhat better in the rare instances in which they are represented by lawyers, but never to a level even approaching that which a fair consideration of the legal materials might suggest. Tenants’ lawyers are most effective in securing extra time for tenants to move. In any event, tenants are so rarely represented that tenants’ lawyers have a negligible impact in absolute terms. Landlords seem to do equally well whether they are represented by a lawyer or not.

Complexity science encourages agents to embrace uncertainty because it provides a narrative by which meaning might be found, in the same way that artists and other creative people have learned to thrive under conditions of uncertainty and randomness (Naidoo, 2003).

Agent-based legal ordering

Law exists as a powerful system of social control. Legal systems exist to settle disputes, to pacify disruptive elements in society, and to allow people and firms to order their activities based upon their predictions of how the power of the state will be exercised. To satisfy these complex and often conflicting functions (Schuck, 1993) and because some indeterminacy is required to make it worthwhile for disputing parties to enter into the process, order cannot be a function of formal, deterministic rules. In any event, the inherent indeterminacy of language and the unpredictability of the forms of human disputing make any formalist model of legal ordering impracticable.

In landlord-tenant courts, practice and results continue unchanged several generations of legal agents after the “tenants’ rights revolution.” It is not only that local customs and unwritten laws prevail over written laws or that local customs fill in the gaps in laws (Seielstad, 1999). That proposition would fail to surprise scholars, judges or, for that matter, the average seven-year-old child who has learned that the rules of any standard childhood game are subject to different application at a friend’s home than at his own. The mind-numbing uniformity in descriptions of landlord-tenant courts – ranging from the ultimate results to such details as who gets to speak, with what effect and where the parties are positioned in courts – is not explained by any currently-prevailing theory of law. I have practiced in Massachusetts, a state that might have the most tenant-protective legal materials in the United States, as well as in Colorado, which likely has the least. There is a stark and predictable uniformity that prevails in practice and results in those states and descriptions of courts in Ohio, Los Angeles, New York, Chicago, Detroit and Baltimore. Indeed, in practicing in, and observing, courts throughout the United States and Europe, I am struck by the eerily similar court proxemics, discourse, practice and results.

In homogeneous societies such as ancient Athens, legal order might have been bottomed on similarities in the shared contexts by which stories are understood (Gagarin, 2003). In societies in which the consent of a heterogeneous governed is required, control is effected through the establishment of a relatively homogeneous elite who maintain control over access to justice (Posner, 1995). When the homogeneity of the bar was threatened in the United States, law school and bar requirements were instituted as substitute alternative constraints (Auerbach, 1976). By the time that the homogeneity of law school classes in turn was threatened, it became clear that the process of law school socialization by itself was adequate to the task of constraining the kinds of stories lawyers would tell to clients and in courts (Posner, 1995).

Law is constructed by lawyers in deciding whether to take on a cause, how much time to devote to it and how it will be presented in relation to contravening stories told by other litigants. Judges or juries make their determinations within the very narrow range between the stories presented. The heavy lifting of legal ordering is conducted by litigants through their lawyers (Nader, 2002).

Humans evolved to detect patterns in the world (Rescher, 1998). Where complexity overwhelms our faculties, stories bridge the gap between patterns of meaning and chaos: stories are the human mechanism to reduce complexity (Baskin, 2004).Law is a social mechanism towards that same end. Law draws clear lines through the messy complexities of social life, and lawyering is essentially the development of stories that, as Don Dellilo had Oswald imagine in Libra, tie the complexities of individual experience to the arc of the whole world. Lawyers, faced with the task of developing coherent stories to translate the meaning in irreducible complexity, similarly rely upon organizing narratives to reduce the inherent complexity of their role to a manageable level. The durable order of law rests on the stereotypical ways in which lawyers can be counted on to react to this challenge which are reinforced by the effects of randomness, routinization and similarities in legal socialization.

Order out of randomness

Complexity illuminates how legal systems thrive in the space between order and disorder. In law, as in other CAS, without order there is no opportunity for a sufficient period of stability for change to evolve; without noise there is no impetus for the system to seek peaks of greater fitness[8]. Randomness prevents long term co-evolutionary relationships from binding too tightly into a runaway spiral. A legal system resulting in rigidly predictable outcomes results in a battle of infinite regress, a recursive battle to the end which cannot be sustained. Jerome Frank recognized over half a century ago that uncertainty in law is not an unfortunate accident; it is the engine of social stability[9].

Stability is perceived in the dynamic tension between competing narratives. Ewick and Sibley document the narratives by which people understand law – an objective area of disinterested action, a landscape for strategic interaction or an arbitrary and capricious exercise of raw power (Ewick & Sibley, 1998: 28). Legality’s power relies on the unarticulated duality between visions of legal order; resistance recognizes and reveals this duality (Ewick & Silby, 1998: 84). Lawyers observed in advising clients have been shown to similarly vacillate between visions of law in a state of frozen predictability and as pure chance (Felstiner & Sarat, 1989). The centripetal forces of disorder and raw interests interact with the centrifugal forces of belief in an objective legal ideal to foster order (Schlag, 1996)[10]. If it is this tension between competing narratives that keeps law in place, then the resistance of competing voices has the potential to change law by creating new relationships among actors using a new language (Alfieri, 1989: 686-687; Bakhtin, 1986). Sharing stories of resistance may be one means by which individual encounters with power become the basis for collective action (Ewick & Silby, 1998: 220).

As the initial array of results in landlord-tenant law developed at a time when tenants had no say in the process, it is not surprising that initial results overwhelmingly favored landlords; however, that does not explain why that distribution persists despite repeated interventions to change it.

Courts have built up increasingly immense dockets without investing resources to provide full hearings (Mosier & Soble, 1973). Landlord-tenant legal systems reinforce this effect by ensuring that tenants’ stories will be drastically curtailed. A landlords’ bar has developed based upon high volume, low per-eviction costs. The limited tenants’ bar has learned to fit tenants’ stories within the constraints that this structure demands and has largely been satisfied with saving their few clients at the expense of tenants generally (Fitzgerald, 1974; Bellow, 1977). Fear of uncertainty, time limitations, their own interests as repeat players in the system and cognitive dissonance make it hard to do much else.

In landlord-tenant disputes, tenants might believe just enough in the narrative of objective neutrality and in the efficacy of court remedies that they do not develop a story-based identity as renters and so to situate their stories within that of the collectivity (c.f. Wood, 2001).

Institutionalization of stories

If judges assigned wins and losses randomly as an initial matter, that original distribution of cases would still help to organize future interaction patterns by devolving resources to initial winners. A win means more to winners in terms of information than loss means to losers, so winning strategies are more likely to be replicated (Axelrod & Cohen, 2000). Those winning strategies, over time, form the template from which the next case will be decided, and conventions of discourse leave the next party to argue their situation an exception.

So, even a purely random system for resolving disputes will evolve to provide order and predictability, so long as the parties do not realize it is random (Fuller, 1978).

Initial tendencies toward order and predictability establish themselves through the filtering processes of perception, cognition, understanding, memory and meaning. People make the world fit stock stories (Balkin, 1998; Lopez 1984; Ricoeur, 1984). Persuasion requires the understanding and manipulation of the stock stories others hold (Lopez, 1984: 3). The wear marks of stock stories can be expected to result in an emergent order (see, Kelly, 1994). A party, by involving herself in a legal dispute, is required to make use of institutional knowledge of stock stories to make interchange meaningful, and thereby the party is constrained to reinforce the legitimacy of that knowledge (Giddens, 1984: 331). Legal agents are not only likely to filter information through the lens of the previously winning stories, but cognitive dissonance might then kick in to shape experience as well as to conform experience to beliefs (Balkin, 1998: 58).

If the legal system sustains to iterate the above processes over many years, institutions will evolve around them. Repeat players have advantages in such systems – not least because they can structure relations to exploit them (Galanter, 1975)[11].

Agent socialization

The complicity of lawyers is reinforced by the iatrogenic effects of law school training. Law school training fosters habits of mind that reinforce the tendencies that sustain the legal order. Law is, to paraphrase Justice Frankfurter, what lawyers do, and lawyers are what law schools (and professional socialization) have made of them (quoted in Henderson, 2003: 79).

This is not ‘order for free’, precisely (Kauffman, 1995). Law schools were established to ritualize language, to promote conformity in social characteristics and political thought (Henderson 2003: 55) at a time when immigration threatened the homogeneity of the bar (Auerbach, 1976; Posner, 1995). The pedagogy of law school continues to instill deference to hierarchy, passivity and cold, context-free inquiry (Kissam, 2003).

Empirical evidence supports the idea that lawyers are different because of their law school experience. Law school has such a pervasive socializing effect that it causes students to become unduly paranoid, hostile and obsessive-compulsive (Daicoff, 2004).

Attorneys appear different than the general population in the way that they make decisions, what they value and respond to and, hence, the kinds of stories they tell. Law school’s preoccupation with formalism continues to stifle creativity and blind indoctrinates to contextual complexities (Stefancic & Delgado, 2005: 63, 65). Lawyers exhibit inflexibility and intolerance for change expressed through authoritarianism (Daicoff, 2004). Daicoff finds some evidence to suggest an increase in cynicism and convergence towards homogenous norms during law school, although it is questionable how long such effects last (Landsman & McNeel, 2004). It would not be surprising if it were so, however. Law school curricula are uniform, state-mandated and have changed little in the past 125 years.

While there is little doubt that distinct habits of mind are fostered in law school[12], neither data nor experience establishes that such effects are either sufficiently universal or long-lasting to explain the uniformity in practice and results that persist long after law school. Lawyers are indeed an unhappy and somewhat uniform lot, but that does not explain how they so effectively construct, without any conscious plan or direction, a legal order in contravention to the formalism to which all their training has been directed.

Law school constructs lawyers, but mostly by making them afraid of uncertainty. The neurosis exhibited by lawyers is, “the suffering of a soul whose meaning is obscured” (Hartwell, 2004). Law school training and socialization fail to provide a narrative lens by which lawyers can probe the complexity of their roles, leaving students “cognitively defenseless against the practices of institutions” (Schlag, 1996: 38). Law is a system that is built on fear.

Reduction of complexity

Lawyering demands an understanding of the complexities of the worlds within which their clients live and predicting the effects of multiple ways of translating the essence of those worlds into stories that will resonate in legal decision-making. Nothing in the training or socialization of lawyers prepares them to address the complexity of this act of translation. Indeed the core of a lawyer’s training is the abandonment of tools for contextual understandings of situations and roles in society in favor of an esoteric, allegedly objective conceit called “thinking like a lawyer” (Scheingold & Sarat, 2004: 51).

Agents in CAS develop models in order to reduce the complexity of the world to a level with which they can deal (Holland, 1998). In human systems, those simplifying models are stories; stories are how we make sense of lived experience (Holland, in Cowan, et. al., 1999: 309; Lopez, 1984: 9). When the complexities we face exceed the stories we have to deal we them[13], we limit ourselves to that limited part of the environment that we can understand (Bar-Yam, 1997: 816). Lawyers recognize that reducing the complexity of contexts requires, “patterns of practice, role identities and conceptions of clients that fit together” (Lipskey, 1980: 155).

Indeed, a primary function of law-in-action, no less than other human systems, is to reduce the complexities of the world outside of us (Ross, 1996). One strategy to do this is to act to foster system imperatives (LoPucki, 1997: 496). Another is to follow fast upon the footsteps of the one preceding, or next to, you (Axelrod & Cohen, 2000: 89). That is, when in doubt, act so as to keep the system going, and, when you have no idea how to do so, mimic others in the crowd and hope for the best. This “flocking trick” is a key to self-organizing behavior (Jost, 2005).

This is an evolutionary-successful model: it assures neither being embarrassed nor being subject to malpractice liability. Feedback loops confirm the practice: adding novel twists to familiar narratives invites accolades from peers and courts; challenging entrenched legal stories, by contrast, risks embarrassment, legal sanction and ontological insecurity (Fitzgerald, 1974).

Agents also reduce complexity by means of routinized behaviors (Holland, 1995; Simon, 1996). Organizing routines prevail over doctrinal sources of order in any legal setting in which classes of disputes or disputants are repeatedly processed, urban or rural, busy or not (Friedman & Percival, 1976)[14].

Faced with complex environments, lawyers will seek simplification to the point where the complexity does not exceed their own. The practices of lawyers and other legal agents congeal together over time into frames of reference by which all information past or future is perceived, understood, remembered and assessed, which are reinforced each time stories are told.

The narrative of complexity provides a framework to appreciate the complexity of legal contexts that might allow agents to subvert prevailing practices of legal storytelling.

In landlord-tenant courts, attorneys face a system that has grown up around voiceless tenants and represented landlords. Courts are set up for the convenience of landlords and their attorneys, who typically represent several landlords who might be facing off against tens, hundreds even thousands of tenants, all of whose cases are set to be disposed of within a few hours. There are precious few opportunities to tell tenants’ stories and few cognitive places for those stories to attach. Tenants’ attorneys, constrained to narrate their situations in terms of exceptions to entrenched stories, are forced to pick and choose among the tenants in the room; to separate the stories they tell from the collective narrative of poor tenants and to, in effect, tell the judge that, “this one is different; this time I mean it” (Amsterdam & Bruner, 2000).

Steven Lubet recounts how the context of storytelling in a Chicago housing court was changed when a well-known lawyer came to that court to represent a tenant. That lawyer’s mere presence was enough to shift the reference point by which the tenant’s story was understood. The entire court, including the judge, allowed not only that tenant, but all tenants that day to speak and attached their stories to reference points drawn from the standard story of law and scripts of how television trials progress (Lubet, 1993).

Law students in a clinical education program affected the nature of storytelling in the Boston Housing Court merely by getting to court early and taking up all the seats usually reserved for landlords’ attorneys – to the point where court personnel at first told students that they had to vacate their seats when the landlords came in.

In a low-income city north of Boston, community voices were brought into court and were allowed to speak, often from seats in the audience, on the effects of legal decision-making in their neighborhoods (see also Fitzgerald, 1974).

Where institutional structures have developed on the assumption of constraints in narrative tellings, the story may be changed by altering those structures directly. Hence tenant lawyers have been successful in changing the story by moving cases out of familiar contexts, by claiming rights to jury trials and by instituting strategic delays (Lazerson, 1981).

Other lawyers advocate change by restoring the authenticity of clients’ voices in legal storytelling (Eastman, 1993; Gordon 1995).

What all these stories have in common is that those involved acted according to a narrative understanding that legal ordering arises out of the story-based interactions of legal actors.

Stories of change

Complexity suggests that change in legal systems occurs by the process of expanding the stories that are heard by legal decision-makers, attaching stories to different reference points held by those decision makers or by changing the contexts in which contested storytelling occurs.

Good stories work by confirming felt certainties (Deutch quoted in Winter, 2001). Stories attach themselves to stories that one already knows. Time demands, the lawyer’s duty to her clients, her own interests as a repeat player in the court, consistency effects and cognitive dissonance make it extraordinarily difficult to do much else (Fitzgerald 1974).

Most stories work by confirming felt certainties, attaching to stories the audience already holds (Lopez, 1994). Stories of change work by altering resonant certainties, or by forcing one to examine already-entrenched stories through a different lens. By reframing stories to leverage vulnerable points in legal systems, lawyers can effect change in law. There is a paradoxical quality to this optimism in changing a system by recognizing the robustness of its dispersed control. Scale-free networks exhibit remarkable persistence in the face of constantly changing environments; however, such networks are also extremely vulnerable to coordinated attacks centered on network hubs or the means of coordination among them (Barabasi & Bornabeau, 2003; Fukuyama, 1999: 225-226).

Change in institutions is built upon an understanding of how storytelling interacts with institutional structures and power relations. Legal change is seldom spoken of in these terms because actors lack the language to categorize the deeply subversive nature of storytelling practices. Complexity offers an organizing narrative that attends to the processes by which the interactions of structures and agents constructs global properties which feed back to constrain the behavior of the agents who produced it (Urry, 2003: 39-49).


I propose that legal order is the residue of the interactions of legal actors, faced with radical uncertainty, reducing complexity to a sustainable level by resorting to simplifying stories, which in turn calcify into routinized behaviors around which institutions coalesce. This approach explains the ordered results and practices we see in law and offers hope for those who wish to change it.

Complexity encourages agent interactions that affect the stories that are told as well as the contexts that give stories meaning. At least, complexity science provides inspiration and the tools to take on the task of trying to understand the legal system in all its messy reality, while it encourages advocates to fill up the problem space of legal change with risk-taking, experimentation and feedback.