UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The virtuous circle of discourse : why Habermasian critical theory is blind to social traps Goenaga Orrego, Agustín Alonso 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2009_fall_goenaga_agustin.pdf [ 286.84kB ]
Metadata
JSON: 24-1.0067709.json
JSON-LD: 24-1.0067709-ld.json
RDF/XML (Pretty): 24-1.0067709-rdf.xml
RDF/JSON: 24-1.0067709-rdf.json
Turtle: 24-1.0067709-turtle.txt
N-Triples: 24-1.0067709-rdf-ntriples.txt
Original Record: 24-1.0067709-source.json
Full Text
24-1.0067709-fulltext.txt
Citation
24-1.0067709.ris

Full Text

THE VIRTUOUS CIRCLE OF DISCOURSE: WHY HABERMASIAN CRITICAL THEORY IS BLIND TO SOCIAL TRAPS by Agustín Alonso Goenaga Orrego B. A., Instituto Tecnológico y de Estudios Superiores de Occidente, 2007 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in The Faculty of Graduate Studies (Political Science) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) September 2009 © Agustín Alonso Goenaga Orrego, 2009 Abstract This paper raises a critique concerning the limitations that Habermas’s Theory of Communicative Action (TCA) faces to engage with topics such as social traps. The main argument is that the developmentalist explanation of ego ontogenesis that supports the possibility of the “discourse principle” reduces the TCA’s effectiveness  to  explain  social  transformation  in  less-than-ideal  situations.  The  paper  introduces  the concept of social traps through Rothstein’s account of the recursive dynamics between individual agents and social structures in non-cooperative scenarios, and follows his criticisms of rationalist and culturalist approaches. The same problems that these strands of the literature suffer are present, although for different reasons, in the TCA: wrong assumptions about human action, deterministic views of social reality, and a narrow understanding of social transformation. Along these lines, the paper explores the implications that the Habermasian notion of praxis as discursive competences and autonomy has for a wider conception of agency in cases where discourse is inhibited or disrupted. Moreover, this becomes a real problem due to the  circular  relationship  between,  on  the  one  hand,  the  development  of  the  cognitive  and  moral competences necessary to participate in discursive practices and, on the other, how these practices foster the  very  same  competences  that  they  require  to  prevail  (“the  virtuous  circle  of  discourse”).  The combination  of  these  elements  raises  a  number  of  challenges  for  the  TCA  to  provide  convincing explanations about the way in which social traps operate and, especially, how social transformation can be generated in those situations from the micro-level. Finally, the author concludes with some suggestions to be considered in the development of a critical methodology to observe social traps. ii Table of Contents Abstract.......................................................................................................................................................ii Table of Contents.......................................................................................................................................iii Acknowledgements....................................................................................................................................iv Introduction.................................................................................................................................................1 Introducing Social Traps..............................................................................................................................6 Rationalist and culturalist approaches to social traps....................................................................6 Bo Rothstein’s institutionalist argument.......................................................................................8 Two ways of understanding subjective rationality: North vs. Swidler..........................................9 The recursive interaction of structure and agency in social traps................................................11 Praxis Is Only Discourse? Only Discourse Is Praxis?................................................................................14 The lifeworld and the Theory of Communicative Action...........................................................14 The uncoupling of system and lifeworld: its consequences for an explanation of social transformation................................................................................19 The “Virtuous Circle of Discourse”...........................................................................................................24 Normativity in post-conventional legal orders............................................................................24 Self-transformation of the individual..........................................................................................26 Praxis, the Virtuous Circle of Discourse, and Social Traps.......................................................................36 Conclusions...............................................................................................................................................45 References.................................................................................................................................................48 iii Acknowledgements I want to thank my supervisors, Dr. Mark E. Warren and Dr. Maxwell A. Cameron, for their guidance, encouragement, and teachings. In the process of developing the project, I received insightful and critical observations from Dr. Bruce Baum, Priya Bala-Miller, Aubin Calvert, Sean Gray, and Kelsey Seymour. Any errors are solely my responsibility. This thesis was written with the financial support of the Canadian Bureau for International Education (CBIE) as part of a Canada Government Award. Finally, my permanent gratefulness goes to my parents for their unending generosity and love. iv Introduction Many  of  the  most  pressing  challenges  facing  contemporary  developing  democracies  are  related  to problems of cooperation and action coordination through democratic procedures: pervasive corruption, mafia organizations, oligarchic enclaves that impinge upon elections and popular representation, lack of trust between citizens of different social groups and between citizens and official authorities, inefficient fiscal  systems,  free-riders,  black  markets,  and  so  on.  The  bottom-line  question  is  how  to  build  in contemporary  societies  a  democratic  rule  of  law ―conflict-resolution  and  collective  decision-making through democratically  institutionalized  legal  procedures― when social  dynamics are  hostile  to  it.  In cases like these, actors are usually entrenched within their private interests and unwilling to cooperate with one another, either because they do not recognize the legitimacy of the ongoing collective regulatory instruments (the current legal and judicial institutions) or because they do not trust that their counterparts will live up to the norms embedded in those agreements. So  far,  Critical  Theory  has  focused  on  discussing  the  first  of  these  two  options,  that  is,  the consequences of challenging hegemonic normative frameworks and the way in which power relations intervene in the construction and perpetuation of those norms and values. Jürgen Habermas in particular has advanced a complex theory about the role of normativity in law and how it can be constructed through discourse in an ideal speech situation (1984; 1987; 1996). The Habermasian project for Critical Theory provides  a  normative  grounding,  as  inclusive as  possible,  for  the  instruments  of  collective  regulation embodied in law in post-conventional societies. This contribution represents an ambitious and promising attempt to address the first of the reasons for which the democratic rule of law might be challenged, stimulating for the past thirty years a burgeoning literature on the topic (Bohman, 1994; 2007; Dryzek, 1994; 2002; Elster, 1998; Goodin & Niemeyer, 2003; Gutmann & Thompson, 2004; Habermas, 1996; Held, 2006; Niemeyer & Dryzek, 2007; Warren, 1992; 1993; 1995). Nevertheless, Critical Theory has until  now given little  attention to the second of the aforementioned obstacles to coordinate  collective action, and, instead, other approaches have taken the lead in the study of “social traps”. 1 Indeed, problems of action coordination and cooperation are usually intertwined with conflicts about norms, knowledge, and interests tilted by asymmetrical power relations. Therefore, these problems need to be taken into account if Critical Theory is to maintain its emancipatory objectives. In this paper I will refer to Critical Theory, mostly through Habermas’s Theory of Communicative Action (TCA), not as a general theory but as “a method of analysis deriving from a nonpositivist epistemology” (Antonio, 1981:332). The question that I will try to raise concerns the limitations that the TCA, as the mainstream thread of Critical Theory, faces to engage with topics such as social traps. Moreover, what are the implications that these limitations have for the broader project of Critical Theory? Even if the critique that Horkheimer, Adorno, and Marcuse started in the 1930s and 1940s was focused on challenging the reified rationality of advanced capitalist societies (Couzens Hoy and McCarthy, 1994; Horkheimer and Adorno, [1944] 2002; Marcuse, [1964] 1991; Nagel,  2008; Wiggershaus, 1994), to ignore  situations such as social traps because they mostly happen in developing countries would be today, under the global interconnectedness of the world, a fallacious excuse. Furthermore, the dynamics inherent of social traps are present even at the core of the advanced capitalist societies once we observe how stratified layers of social integration, along the lines of gender, race, ethnicity, and class, become recurrent obstacles for action coordination between different demographic groups. The path that Critical Theory took with Habermas has increasingly moved away from these concerns. His attempt to escape from the iron-cage of instrumental rationality by developing a critique of ideology based on communicative action has paid the price of turning itself blind to forms of self-imposed  coercion  and  “self-frustration  of  conscious  human  action”  other  than  technocratic domination.1 Under these premises, to argue that Habermasian Critical Theory is blind to social traps is 1 It would certainly be naïve to expect a theory to solve every problem and be effective in every imaginable situation. However,  the  central  premise  of  Critical  Theory  has  been  to  push  knowledge  towards  emancipatory  purposes. According to Raymond Geuss (1981), “the Frankfurt account of the essential distinguishing features of a ‘critical theory’ consists of three theses: 1. Critical theories have special standing as guides for human action in that: a. They are aimed at producing enlightenment in the agents who hold them, i.e. at enabling those agents to determine what their true interests are; b. They are inherently emancipatory,  i.e. they free agents from a kind of coercion which is at least partly self-imposed, from self-frustration of conscious human action. 2. Critical theories have cognitive content, i.e. they are forms of knowledge. 3. Critical  theories  differ  epistemologically  in  essential  ways  from  theories  in  the  natural  sciences. Theories in natural science are ‘objectifying’; critical theories are ‘reflective’” (Geuss, 1981:2). 2 not just an unfair attempt to raise the bar too high in measuring the explanatory (and emancipatory) power of the TCA as a critical methodology for social research (Antonio, 1981; Prasad, 2005; Zanetti, 1997). Social traps represent one of the most common factors that hinder agents from determining, in the words of Raymond Geuss (1981), “what their true interests are”. If this is the case, the TCA’s inability to engage with action coordination problems in general, and particularly in the developing world, narrows its scope as a Critical Theory and makes it vulnerable to the critiques of subaltern (i.e., Nagel, 2008; Willet, 2001) and feminist scholars (i.e., Fraser, 1985). David Held pointed out in 1981 that Habermas had failed so far to adequately respond: “To whom is critical theory addressed? How, in any concrete situation, can critical theory be applied” (Held in Livesay, 1985:68)? I sustain that an unacknowledged ethnocentrism would become evident if we try to answer those questions from Habermas’s own writings. As Nancy Fraser puts it, this strategy “necessitates that one  read  the  work  in  question  from the  standpoint  of  an  absence;  that  one  extrapolate  from things Habermas does say to things he does not” (1985:98-99). Therefore, I suggest following Niklas Luhmann in order to find a starting point from which to analyze the Habermasian project: [T]he key formula [the Habermasian discourse principle] states: “Those norms for action are valid, to which all  potentially  affected persons could agree  as participants  in a rational  discourse.”  Every concept of this maxim is carefully explained with the exception of the word “could,” through which Habermas hides the problem. […] The master and the invisible hand will not be replaced. But who determines, and how does he do so, what  could find rational  agreement? How does this decisive operation, on which everything in the postmetaphysical age depends, become juridified? As a result, it also remains unclear, on all levels of the argument, how the conjunctive becomes an indicative, how the potential  becomes a reality, or, for example, how power “comes forth” out of the freely discussing civil society, which does not of course exist (Luhmann, 1998:164-165). By referring to the unexplained possibility  that the word  could places in Habermas’s account  of law, Luhmann  hints  that  his  theory  might  be  built  upon  unacknowledged  assumptions  which  once  made explicit can show its limitations. In this paper I want flesh-out some of the assumptions concealed by the conditional  form  in  Habermas’s  discourse  principle  and  explain  how  they  interfere  in  the  TCA’s 3 explanatory leverage of social traps. If the “ideal speech situation” 2 provides a normative cornerstone for the TCA and the theory of law based on deliberative procedures, it is at the cost of having little to say about less-than-ideal situations such as social traps. In order to advance this argument,  I will first  introduce the concept of social traps as it has been treated  by  cognitive  rational  choice,  historical  institutionalism,  and  particularly  Bo  Rothstein’s institutionalist approach (2005). Rothstein’s critique of rationalist and culturalist accounts of cooperation problems will help me draw ―in the fourth section of this paper― a parallel with some of Habermas’s limitations to explain social transformation in less-than-ideal situations,  making the TCA face similar shortcomings when observing social traps. In the following section I will introduce the TCA and then turn to Anthony Giddens’s critique, where he argues that Habermas’s normative idealism makes his account of Critical Theory too detached from praxis. The absence of a strong theory of praxis as human’s “conscious creative activity”  3 hinders the ability of the TCA to accurately observe situations like social traps, since it tends to conceal them under a deterministic explanation based on structural conditions in which individual action has little relevance. From Habermas’s perspective, negative structural conditions hamper motivational,  cognitive and moral development;  they  reproduce patterns  of  distorted  communication  (communication biased by  unequal power relations), and inhibit the materialization of agency as communicative action. This determinism is thus incapable of explaining the role of human action in cases of social transformation. The question that is then raised is whether the TCA would need to consider other forms of agency, which not necessarily 2 Habermas  described the characteristics  of  the “ideal  speech situation”  in the following  terms:  “Participants  in argumentation have to presuppose in general that the structure of their communication, by virtue of features that can be  described  in purely  formal  terms,  excludes  all  force―whether  it  arises  from within  the  process  of  reaching understanding itself or influences it from the outside―except the force of the better argument (and thus that it also excludes,  on  their  part,  all  motives  except  that  of  a  cooperative  search  for  the  truth).  From  this  perspective argumentation can be conceived as a reflective continuation, with different means,  of action, oriented to reaching understanding” (Habermas, 1984:25). 3 I use the term praxis to refer to the individual ability of human beings to actively modify their life conditions. The term praxis comes from Marx’s Theses on Feuerbach (Marx, [1888] in Tucker, 1978), where he insisted that thinking was intertwined with every other human activity:  “For Marx,  what made human beings different  from any other species of animate life was their capacity for conscious creative activity ―for practice or praxis as he called it― a concept which he used to embrace both thought and life” (Kitching, 1988:26). 4 occur through discursive practices, to be able to provide a better explanation of social change in less-than- ideal scenarios. Thirdly, I will discuss what I have called the virtuous circle of discourse in Habermas’s work, that is, the  circular  argument  claiming  that  discourse  requires  specific  moral  and  cognitive  competencies  in individuals for communication to remain undistorted, and, at the same time, that those skills can only be developed through participation in discourse. At this point I will argue that this circularity reproduces a similar  logic  to  the  one  operating  within  social  traps,  where  trust  and  cooperation  are  recursively constituted or hindered. In  the  fourth  section,  I  will  recapitulate  the  whole  argument  about  why  the  TCA faces  crucial obstacles to observe social traps. I claim that the developmentalist approach behind the discourse principle has three consequences on the TCA: (1) The evolutionary theory that explains the achievement of moral autonomy and a universal ethics of speech ends up offering a deterministic view of cognitive, moral, and interactive  development  that  occurs  independently  from individual  action.  As a  result,  the  TCA only recognizes  human praxis  as  discursive  practices.  (2)  The strong distinction between instrumental  and communicative  action does  not  consider  a  concept  like  “practical  consciousness”  (Giddens,  1984)  or “subjective rationality” (Rothstein, 2005). Such a concept would allow for a more inclusive conception of praxis  and  a  better  description  of  social  reproduction  and  especially social  transformation.  (3)  The circularity in Habermas’s account of the two functions of deliberation reproduces the reciprocal influence between social  structures  and  individual  action.  However,  by  grounding  his  theory  of  law upon this circular argument based on ideal situations, Habermas obscures the fact that the same recursive dynamics operate  in an opposite  direction  in  less-than-ideal  cases  (such as  social  traps),  where  self-reinforcing conditions systematically hinder action coordination through consensus. Finally, in my conclusions, I will use the previous discussions to present a few issues that a critical methodology should consider in the study of social traps. 5 Introducing Social Traps In  this  first  section,  I  will  introduce  social  traps  as  situations  where  the  recursive  dynamics between structures and agents are particularly significant. Bo Rothstein’s (2005) approach to the topic will be useful for these purposes since his argument is based on two main components: (1) an explanation of how social traps are perpetuated through the reciprocal influence between social structures and individual actions, and (2) the role of political institutions as linking mechanisms between them. I will concentrate on the  first  one  of  these  elements  since  it  speaks  directly  to  some of  the  problems  that  are  present  in Habermas’s TCA (which I discuss in the following sections). Along these lines, I will begin by addressing the shortcomings that rationalist and culturalist approaches face in their explanations of social traps. These limitations are central to my argument and I will return to them in my conclusions, since they are also present ―although for different reasons― in the TCA. The second and third parts of this section will focus on Rothstein’s argument and the recursive dynamics of agents and structures. Rationalist and culturalist approaches to social traps The psychologist John Platt (1973) coined the term “social traps” to describe the kind of situations where cooperation among self-interested, utility maximizing actors is hindered. Nevertheless, the topic has been central  among political  scientists  for  the  last  forty  years,  and  has been addressed under  several different  names:  “Provision of Public Goods,  Problem of Collective Action,  Tragedy of the Commons, Prisoners’ Dilemma, and  Social Dilemma are but a few” (Ostrom in Rothstein, 2005). All these labels indicate circumstances where cooperation among self-interested actors is hindered: According to the logic of  the social trap,  even people with clear preferences for “fair play” will continue their disloyal behavior because they believe, and for good reason, that almost “all other people” are going to keep playing dirty. And, again, this is not because most other people are actually evil and fundamentally disloyal, but because they expect that everyone else will cheat. Changing the situation is thus a matter of changing the worldview of large groups of citizens about the kind of society they live in and how people might conceivably act in that society (Rothstein, 2005:8). 6 Social traps have frequently been approached through either rationalist or culturalist standpoints that face three major limitations to provide a convincing explanation of social traps. Rothstein (2005:29-30) lucidly points at these interrelated problems: a) both traditions are based upon unrealistic assumptions of human action, b) they are deterministic in their conclusions, being incapable of offering a convincing theory  of  praxis,  and  c)  precisely  because  of  their  deterministic  imprint  they  cannot  explain  social transformation. “Rationalists begin with assumptions about actors who act deliberately to maximize their advantage... Analysis begins at the level of the individual and culminates in questions about collective actions, choices, and  institutions”  (Lichbach  &  Zuckerman,  1997:6).  However,  this  emphasis  on  utility-maximization adopts an “anemic or thin version of intentionality and interests” (Lichbach in Rothstein, 2005:30). The consequence of such an assumption is that this approach overlooks important variables regarding the way in which individuals deal with uncertainty and incomplete information, how norms and social contexts influence their decision-making processes, and the ways in which strategic calculations about each other’s future actions determine their social behaviour in daily life. On the  other  hand,  culturalists  tend  to  see agents  as  puppets  whose lives are  determined  by the overwhelming tide of cultural and structural conditions. From this standpoint, culture would be the only explanation for individual action: Culture, from this perspective, “is a worldview that explains why and how individuals and groups behave as they do” […]. Obviously, there is hardly any room for such things as intentions, strategic action, not to say deliberative choice, within this perspective. Once the world has been “culturally constructed” for them, individual agents are no longer “agents” in any meaningful sense of the word (Rothstein, 2005:30). In both approaches, the reductionist view of human action (individual agency only as purposeful- rational action or as an inconsequential pawn of cultural frameworks) leads to a deterministic view of social phenomena. In the context of social traps, the end of the road for both trends would certainly point at the impossibility to escape once a group falls into those dynamics. Although empirical evidence seems to support  such a pessimistic prospect showing that  social traps tend to endure  and have long-lasting 7 effects (North, 1990; North, Summerhill & Weingast, 1999; Putnam, 1993), Rothstein’s case study of the labour conflict in Sweden in the 1920s (2005) ―along with others (i.e., Ostrom, 1990)― debunks those approaches by looking at a case in which social transformation occurred and opened a way out of the social trap. For rationalist and culturalist approaches, there is very little room for agency, strategy, and choice  once  a  system  of  incentives  or  a  belief  system,  respectively,  is  in  place,  therefore  such  an assumption makes them incapable of explaining cases of social transformation. Bo Rothstein’s institutionalist argument Aware of these criticisms against the dominant literature on social traps and following the tradition of historical institutionalism (North, 1990; Putnam, 1993), Rothstein (2005) has built an argument that solves many  of  the  shortcomings  of  the  previous  approaches,  particularly  because  it  takes  into  account  the recursive interaction between social structures and individual actors, refusing to give primacy to either rationalism or culturalism (Rothstein, 2005; Ostrom, 2008). On the contrary, he underlines the necessity of combining individual and collective historical perspectives to be able to advance a powerful explanation of social traps. According to him: the theory about social traps allows us to link two approaches in the social sciences that are usually widely disparate:  those which stress the importance of  historically established social and cultural institutions and norms, and those which emphasize the importance of  human strategic actions and choices (Rothstein, 2005:14). Rothstein’s  argument  is  grounded  on  the  combination  between,  first,  the  role  of  culture  that individuals pragmatically use as a “tool box” through subjective rationality and, second, an institutional component embodied in an in-depth study of how the universal welfare state affected labour politics in Sweden in the 1920s (Ostrom, 2008:137). He sees this approach as a commitment to a methodological individualism capable of providing a descriptive account of subjective rationality by incorporating the agents’ limited computational capacity, the influence of emotions on their behaviour, and the impact of cultural and social contexts on preferences (2005:36). This methodological strategy should then be able to simultaneously observe the circular dynamics of social traps and the possibility for social change. 8 New political actions emerge from cognitive mechanisms (although very often not as intended by actors). In other words, political change is actor based and hence occurs only when agents take notice and act upon structural changes, other people’s actions or new opportunities. However, mechanisms need not  be coupled only to instrumental  rationality.  On the  contrary,  how mechanisms work in different  social  settings  should  be  an  open  empirical  question.  They  may  equally  be  based  on emotions,  problems  of  information,  or  ideology.  The  implication  is  that  a  focus  on  causal mechanisms makes it necessary to specify how we see the relation between rationality and culture and how they are related at the individual (micro-) level (Rothstein, 2005:34-45). From  this  point  of  view,  actors  orient  their  behaviour  in  relation  to  the  images,  beliefs  and expectations about how others will behave. In order to understand this kind of strategic behaviour it is necessary to take into account how those perceptions enter into human consciousness and shape political action (Rothstein, 2005:14). An understanding of rationality as a simple instrumental calculation of utility- maximization becomes untenable once we accept that self-interest and rationality are context dependent (Mantzavinos, North & Shariq, 2003; North, 1990). Moreover, empirical research has shown that such an assumption has little power to make predictions on future human conduct (Rothstein, 2005:35). Social traps then cannot be observed as an example of irrational actors undermining their own self-interests due to their non-cooperative behaviour, since what counts as rational action in those situations is determined by the individual perceptions of who the other is and whether he can be trusted or not. “This assessment may come from many different sources, such as personal knowledge about the individuals in question, culturally determined stereotypes, or memories of how the actors have acted in similar situations in the past” (Rothstein, 2005:15). If those factors affect individual action, it is impossible to determine whether cooperating or cheating is the rational option until a political anthropology manages to grasp the context in which individual action is embedded (Bates, de Figueiredo & Weingast, 1998). Two ways of understanding subjective rationality: North vs. Swidler There are two ways  in which the notion of subjective rationality can be understood. The first one, following North and the cognitive rational choice tradition, sees it as the result of “incomplete processing of information” by individuals when trying to decipher the environment.  The incomplete  feedback of information after individuals interact with the prevailing institutions of society leads “to ideas, ideologies 9 and dogmas” which play a major role in human beings’ choices (North, 1990:22-23). Therefore actors usually  make  the  most  rational  choice  for  their  perceived  interests  under  conditions  of  incomplete information. Moreover, the solutions to cooperation or exchange problems achieved in the past carry over into the present (1990:37) and become recurrent in how people interact. In short, previous institutions, perpetuated  through  formal  or  informal  incentives  and  constraints,  determine  what  counts  as  rational action through the formation of “subjective models of reality” (1990:112).4 This is the reason why, for North,  social  transformation  can  only  occur  incrementally  and  is  tenuously  linked  to  human  action. Change rather consists of “marginal adjustments to the complex of rules, norms, and enforcement that constitute the institutional framework (1990:83).” The second account of subjective rationality,  the one embraced by Rothstein, understands it as the result of knowledgeable agents that possess the practical skills to operate pragmatically in specific social contexts. These actors face the problem of fragmentary information that reinforces beliefs and perceptions about other actors based on past experiences, but also have a margin of action for strategic maneuvering and social transformation. Rothstein borrows this view of “culture as a tool-box” from Ann Swidler: A starting point in this approach is that people  know much more culture  (signals, stories, symbols, rituals, etc.) than they actually use. Secondly, there is variation in how they make use of the cultural repertoire  that  is  available  to  them  and  they  “select  within  that  repertoire  what  works  at  the moment” […]. Thirdly, people also differ in “how seriously they take their culture and how richly they deploy it” […]. Lastly, Swidler argues that people sustain a lot of contradictory or uncoordinated cultural codes in their repertoires (Rothstein, 2005:37-38). I will elaborate on the consequences of these two conceptualizations of subjective rationality in the fourth section of this paper, when I will argue that the Habermasian distinction between instrumental and discursive  rationality  falls  into  a  similar  determinism as  the  one  present  in  North’s  work,  where  the incomplete processing of information (which is not directly translatable to Habermas’s notion of distorted communication but operates in an equivalent way) highly reduces the possibility of individual agency. 4 Although North uses the term “subjective  models  of  reality”,  the oxymoron “subjective  models  of  rationality” might be more accurate to portray the challenge to a univocal  rationality that ignores how the context conditions responses and decisions. 10 The recursive interaction of structure and agency in social traps From the explanation of social traps based on the strategic behaviour of actors guided by subjective rationality, Rothstein claims that we can flesh-out four fundamental characteristics of the social world: (1) that actors behave strategically ―“what people do depends on what they believe others are going to do” (Rothstein, 2005:13)―; (2) that “individual rationality may very well be collective irrationality” (2005:13), that is, individual  actors  often  cannot  or  choose  not  to  take  into  account  the  unintended consequences that will occur in the societal or group levels as a result of their private self-interested actions; (3) that “whether or not an action is rational cannot in these types of situations be determined solely by reference to one’s individual preferences, but is rather determined by the social context” (2005:13), especially the “collective memory” (2005:37) that determines from past experiences whether the others are trustworthy or not; (4) that  “we cannot  rationally  forget”  past  betrayals  and deceitful  behaviour  even if  that would increase everyone’s will to cooperate (2005:13). Rothstein is recognizing here the recursive mechanisms that connect individual actions with social structures, arguing  that  “social  and political  structures  are  the  result  of  the  aggregation  of  individual behaviors” while, at the same time, “the particular structural position (or configuration) of agents vis-à-vis each other affects the resulting individual behavior” (2005:39). The idea that he tries to advance is that “structural  explanations  can  be  reduced  to  problems  of  aggregation  of  individual-level explanations” (2005:39). However, it is essential to clarify that individual behaviour in groups cannot be equated to the behaviour of isolated individuals: The fact that actors are in a group is what fundamentally changes the structure of their interaction, their perception of the chances of success, their perception of their safety, their perceptions of gains and losses and their distribution, their emotions, and their concerns of fairness, etc. Most importantly, what changes is their view (their belief system) about what they can expect of the other agents in that group. Will they cooperate or are they more likely to cheat? Can they be trusted or not (Rothstein, 2005:39)? 11 In other words, the connection between structure and agency manifests itself in social traps in the following way: social structures ―such as belief systems or shared perceptions and assumptions of the others― shape individual action by modifying the expectations about future cooperation between actors: will  they  cooperate  or  will  they  cheat?  This  in  turn  determines  individual  action,  reinforcing  the perceptions of each other that are built into and sustain those belief systems. Social structures are then reproduced every time an actor decides to cooperate or to cheat according to the expectations created by those prevailing structures. The circularity  that  links  the social structures  with individual agency is  the same circularity  that explains why social traps are situations from which it is extremely difficult to escape. Social trust cannot be built overnight. On the contrary, the collective memory of untrustworthy and deceitful behaviour is hard  to  erase  and  tends  to  reinforce  those  negative  assumptions  and  expectations  about  the  others. Moreover,  social  traps  become  an  example  of  “stable  but  inefficient  equilibria”,  where  although everybody suffers from the lack of cooperation,  past experiences make non-cooperative behaviour the most rational option for individual actors. Empirical evidence shows that perceptions of  the others  are highly stable and difficult to change. These cognitive or mental maps are often included in enduring cultural socialization processes, where they strongly  characterize  the worldviews of individual  actors  with respect  to  things such as the honesty  and  competence  of  state  institutions  or  whether  it  is  reasonable  to  trust  other  people  in general or specific groups of people (Rothstein, 2005:21). According to Rothstein, in order to gain a better explanatory leverage of social traps, we need then to consider a broad repertoire of factors that drive human action (2005:36). This perspective must go beyond the  utility-maximizing  paradigm of  instrumental  rationality  and  include  those factors  that  result  from social and political structures ―i.e., cultural traits, systems of beliefs, institutional (formal and informal) incentives  and  constraints.  A  full  understanding  of  how  institutions  and  subjective  rationality  link particular  social  structures  with  certain  patterns  of  individual  action  would  provide  a  more  accurate explanation of why actors choose to perpetuate social traps by entrenching themselves in non-cooperative 12 behaviour. Such an account requires a consideration of agency that goes beyond functionalist notions of purposeful action: “Agents are not perfectly rational and fully informed about the world in which they live. They base their decisions on fragmentary information,  they have incomplete models of the process they are engaged in, and they may not be especially forward looking. Still, they are not completely irrational: they  adjust  their  behavior  based  on  what  they  think  other  agents  are  going  to  do,  and  these expectations are generated endogenously by information about what other agents have done in the past” (Peyton-Young in Rothstein, 2005:37). What  people  “think  other  agents  are  going  to  do”  is  of  course  something  they  learn  (or  make inferences about) from the culture in which they live. Similarly, “what other agents have done in the past” must be seen as agents in a society sharing some sort of “collective memory” about each other as individuals and groups (who are the Serbs, the politicians, the police, the Catholics… and to what extent can they be trusted?) (Rothstein, 2005:37). Finally,  the  second  component  of  Rothstein’s  argument  addresses  the  role  of  institutions  as connecting mechanisms between structures and agents. Institutions are important for Rothstein since “they present incentives, they induce strategy because they make it plausible to calculate what the other agents are likely to do, and, in some cases, they influence ethics and norms” (2005:42). It is at this point where, for our purposes, the most important controversy in Rothstein’s work appears. “Are we to understand political institutions as any kind of repetitive behavior that influences political processes or outcomes? Or should we reserve the  term ‘political  institutions’  for  formal  rules  that  have been decided upon in  a political process” (2005:40)?  The question about whether to provide a wide or a narrow definition of institutions certainly shapes the trajectory of Rothstein’s research and produces the shortcomings that a critical method might overcome. This is not the place to explore such an issue since this paper’s objective is  not  to  outline  a  critical  method  for  social  traps,  nonetheless  I  will  briefly  return  to  this  in  my conclusions to suggest some features that a critical observation of social traps would need to consider. 13 Praxis is only discourse? Only discourse is praxis? In this section I will first introduce the TCA, emphasizing the concept of the lifeworld and its role in Habermas’s  explanation  of  social  reproduction.  I  then  follow  Giddens  in  his  critique  of  two  key distinctions  in Habermas’s  work:  first,  the  one between instrumental  and  communicative  action,  and, second, the one between the lifeworld and the system. I will devote the second part of this section to explain why the lifeworld as an intangible set of rules, resources and expectations cannot be empirically distinguished from the systems that contain it,  not even if we recognize the emergence of a formally institutionalized public sphere. This, I will argue later on, is one of the central obstacles that Habermasian Critical Theory faces to address social traps, since it fails to incorporate human agency as a day-to-day factor that intervenes, most of the time in an undiscursive fashion, in processes of social change. The lifeworld and the Theory of Communicative Action Since his writings of the late 1960s (Habermas, 1971 [German edition in 1968]; Outhwaite in Ritzer, 2003:229),  Habermas tried  to  provide a  normative  grounding  for  Critical  Theory  that  would avoid a substantive foundation (such as Marcuse’s distinction between false and real needs, for example) or some kind of “first philosophy” (Marcuse, [1964] 1991; Livesay, 1985:74). In order to achieve this, Habermas developed  the  notion  of  the  ideal  speech  situation  as  the  normative  cornerstone  of  the  Theory  of Communicative Action. By grounding communication upon the ideal speech situation and a “universal pragmatics”5,  Habermas  offered  a  conception  of  “truth”  that  looked at  itself  as  “one  among  several ‘validity  claims’  that  can  be  redeemed  in  discourse”  (Giddens,  1977:143).  These  concepts  became essential to his theory of communicative rationality, since they seemed to help him avoid a transcendental basis of knowledge. Communicative rationality would, in turn, contribute to the construction of a theory 5 “The task of universal pragmatics is to identify and reconstruct universal conditions of possible understanding […]. In other contexts one also speaks of ‘general  presuppositions  of communication’  but I prefer to speak of general presuppositions of communicative action because I take the type of action aimed at reaching understanding to be fundamental. Thus I start from the assumption […] that other forms of social action […] are derivatives of action oriented to reaching understanding “(Habermas, 1979:1). 14 of human action that “does not posit a self-sufficient subject, confronting an object world, but instead begins  from  the  notion  of  a  symbolically-structural  life-world,  in  which  human  reflexivity  is constituted” (Giddens, 1987:236). Through this strategy, Habermas tried to provide an alternative to the positivist-functionalist emphasis on purposive-rationality which had generated a loss of moral meaning and a diminution of freedom in advanced capitalist societies. Critical Theory, through the TCA, could then pursue  its  objective  of  expanding  the  ability  of  self-reflection  among  human  subjects  to  transform asymmetrical conditions of power and domination. During more than fifty years as a philosopher, Habermas has explored two different tactics to provide a foundation for Critical Theory. The “turn to language” in the 1970s marked the shift from one argument to the other, distinguishing two phases in Habermas’s career (Giddens, 1987:226; Craib, 1992:232-234; Outhwaite in Ritzer, 2003): first, an epistemological approach that culminated in Knowledge and Human Interests in 1968 (the English version appeared in 1971); and, from then on, the attempt to ground Critical Theory  upon  language  and  communication.  The  epistemological  argument  was  based  on  drawing classifications of rationality, knowledge, and science that would correspond to the “elements of the human self-formative  process”:  labour,  interaction  and  authority/power  (Habermas,  1971:196),  that  is,  “the specific  fundamental  conditions  of  the  possible  reproduction  and  self-constitution  of  the  human species”  (Giddens,  1977:138).  The  purpose  of  this  discussion  was  to  follow  the  Frankfurt  School’s critique  of  positivism by  showing  how  its  view  of  science  had  detached  knowledge  from  interests: “‘Scientism’ means science’s belief in itself:  that is, the conviction that we can no longer understand science as one form of possible knowledge, but must rather identify knowledge with science” (Habermas, 1971:4). This exclusionary view of knowledge conformed “to only one type of knowledge-constitutive interest,  that  in  the  prediction  and  control  of  occurrences,  or  ‘technically  exploitable’ knowledge” (Giddens, 1977:139). Along these lines, if this kind of knowledge was associated with labour, instrumental  action and the empirical-analytic sciences,  interaction would correspond to the search of mutual understanding characteristic of the historical-hermeneutic sciences. For him, both of them claim to separate knowledge from interests and share “the methodological consciousness of describing a structured 15 reality within the horizon of the theoretical attitude” (Habermas, 1971:303). This is the reason why a critical social science should “determine when a theoretical statement grasp invariant regularities of social action as such and when they express ideologically frozen relations of dependence that can in principle be transformed”  (Habermas,  1971:310).  This  critique  of  ideology should  then  be  concerned  with power relations and be able to self-reflect about its own interests. As Outhwaite says, “Habermas […] then came to feel that the trichotomy of empirical, hermeneutic, and critical sciences was too simplistic, especially in that reflection in the philosophical sense did not necessarily mean emancipation in practice” (Outhwaite in Ritzer,  2003:232).  Therefore,  in his  second period, Habermas took a linguistic turn to pursue the same objectives but through a different strategy: a theory of communicative action based on the analysis of linguistic communication. Only  a  rational  agreement  which excluded no  one  and  no  relevant  evidence or argument  would provide, in the last  resort,  a justification of the claims we routinely make and presuppose in our assertions. […] Moreover, if Habermas is right that moral judgments also have cognitive content and are not mere expressions of taste or disguised prescriptions, it also provides a theory of truth for issues of morality and of legitimate political authority. Moral norms are justified if they are what we would still uphold at the end of an ideal process of argumentation. […] The analysis of language-use can thus, Habermas believes, be expanded into a broader theory of communicative action, defined as action oriented by and toward mutual agreement.  […] The theory of communicative action, then, underpins a communication theory of morality, law, and democracy, and it is these aspects which have dominated Habermas’s most recent work (Outhwaite in Ritzer, 2003:229-230). As I will discuss in detail in the following section, Habermas used Jean Piaget’s stages of cognitive development and Lawrence Kohlberg’s stages of moral development to build a universal argument for the theory  of  communicative  action.  Through  this  evolutionary  account  of  moral  and  interactive competencies,  Habermas  justified  his  use  of  the  “ideal  speech  situation”  and  “universal pragmatics”  (Habermas,  1979)  as  species-wide,  ideal  foundations  for  communicative  rationality  and communicative  action.6 In the  first  volume of the  Theory of Communicative Action (1984),  Habermas discussed Evan-Pritchards investigations on the Azande tribe’s beliefs on witchcraft, as part of his attempt 6 Although Habermas has always recognized the inherent idealism of this concepts, he has also argued that the ideal speech situation is “counterfactually presupposed […] by our everyday practice of communication, which is made meaningful by the real or hypothetical prospect of ultimate agreement” (Outhwaite in Ritzer, 2003:229). Moreover, the point  of including Piaget’s cognitive psychology in his argument is to bring empirical  evidence in favour of universal pragmatics and the ideal speech situation as real possibilities for human communication and action. 16 to defend the universality of the evolutionary argument about the rationalization of wordviews through learning processes. He insisted that the idea of rational worldviews behind the notion of communicative rationality should not be understood in any substantial sense. On the contrary, Habermas followed Piaget arguing that “cognitive development signifies in general  the decentration of an egocentric understanding of  the  world”  (Habermas,  1984:69,  his  emphasis),  and  explained  this  by  making  reference  to  the differentiation of subjective, objective and social worlds. According to Piaget, the growing child draws a “demarcation  through the  construction  of  the  universe  of  objects  and  of  the  internal  universe  of  the subject”  (Piaget  cited in Habermas,  1984:68).  Furthermore,  the  external  universe  is  split  between the world of objects, on the one hand, and the world of “normatively regulated interpersonal relations on the other”  (Habermas,  1984:68),  that  is,  the  social  world.  In  sum,  the  individual’s  subjective  world  also becomes aware  of  other  individuals’  subjective  worlds,  making  it  possible  “to  adopt  in  common the perspective  of  a  third  person  or  a  non-participant”  (1984:69).  It  is  in  this  context  where  Habermas introduces the concept of the Lebenswelt or lifeworld: Subjects acting communicatively always come to an understanding in the horizon of a lifeworld. Their lifeworld is formed from more or less diffuse, always unproblematic, background convictions. This lifeworld background serves as a source of situation definitions that are presupposed by participants as unproblematic. In their interpretive accomplishments the members of a communication community demarcate the one objective world and their intersubjectively shared social world from the subjective worlds  of  individuals  and  (other)  collectives.  The  world-concepts  and  the  corresponding  validity claims provide the formal scaffolding with which those acting communicatively order problematic contexts of situations, that is, those requiring agreement, in their lifeworld, which is presupposed as unproblematic (Habermas, 1984:70). Habermas  then  goes  on  to  identify  four  formal  properties  that  cultural  traditions  must  have  for communicative rationality to operate: a) the cultural tradition must provide differentiated formal concepts, validity claims and basic attitudes for each of the three worlds; b) it “must permit a reflective relation to itself”, it should not be fixed as dogma but allow critical revisions; c) it must be open to feedback “in specialized forms of argumentation”, fostering the emergence of cultural subsystems “in which traditions take shape that are supported by arguments rendered fluid through permanent criticism but at the same time professionally secured”; and d) the cultural tradition “must interpret the lifeworld in such a way that 17 action oriented to success [...] can be uncoupled from action oriented to reaching understanding”, making possible “a societal institutionalization of purposive-rational action for generalized goals, for example, the formation  of  subsystems,  controlled  through  money  and  power,  for  rational  economics  and  rational administration” (Habermas, 1984:71-72). This last condition thus establishes the distinction between the lifeworld  and the  subsystems through which society  operates  in a daily  basis  or,  in other words,  the distinction between instrumental and communicative action. Through this line of reasoning, Habermas provides a universalist argument for a procedural concept of communicative  rationality  without  falling  into  the  trap  of  building  itself  upon  a  transcendental cornerstone, that is, a substantive notion of rationality. Such an argument would then maintain the idea of a progression in human societies from pre-conventional to post-conventional forms of morality associated with the development of higher forms of rationality.7  However, Habermas’s critics claim that he does not succeed, and that the developmentalist imprint in Habermas’s account of discursive rationality ends up refurbishing  a  transcendental  standpoint  that  limits  human  agency  to  individuals  situated  in  post- conventional societies. The consideration of  praxis recedes from the center of Habermas’ thought not because of specific theoretical intentions, but rather as an unintended consequence of his decisions about other theoretical issues:  in particular,  his  effort  to  avoid lapsing  into any  form of technocratic  instrumentalism by doggedly maintaining the absolute distinction between instrumental and communicative action, and his attempt to establish a “quasi-transcendental” normative foundation for critical theory through his model of an ideal speech situation and his rational reconstruction of the evolution of forms of social integration (Livesay, 1985:69). 7 The expression “post-conventional societies” comes from Lawrence Kohlberg’s explanation of a post-conventional morality: “He described development in terms of conventional  moral  thinking (the morality of maintaining social norms because they are the way we do things) shifting to postconventional thinking (the morality that rules, roles, laws,  and  institutions  must  serve  some  shareable  ideal  of  cooperation)”  (Rest  et  al.,  1999:2).  Along  these lines, Habermas equated the individual development of moral autonomy to the system’s morality at the collective level. Societies  would  have  followed  a  similar  trajectory  travelling  from  pre-conventional  to  conventional  and  post- conventional  forms  of  morality:  “The  disenchantment  of  religious  worldviews  not  only  has  the  destructive consequence  of  undermining  the  ‘two  kingdoms’  of  sacred  and  secular  law,  and  with  this  the  hierarchical subordination to a higher law. It also leads to a reorganization of legal validity, in that it simultaneously transposes the  basic  concepts  of  morality  and  law  to  a  postconventional  level.  With  the  distinction  between  norms  and principles  of  action,  with  the  idea that  norms  should  be  generated  from principles  and  by voluntary  agreement (Vereinbarung), with the concept of the lawmaking power of privately autonomous legal persons, and so on, there develops a notion of norms as positively enacted and hence changeable, yet at the same time criticizable and in need of justification” (Habermas, 1996:72). 18 Along  these  lines,  Anthony  Giddens’s  most  poignant  argument  against  Habermas  is  the  one concerning how he reduces the scope of human agency by not taking into account how lay actors can be knowledgeable and reproduce or transform structures in ways other than discourse. A second criticism (closely  related  to  the  former)  that  Giddens  raised against  Habermas had to  do with his  “taxonomic fervour” and the “puritanical formalism” of his writing (1987:242). This criticism goes beyond nitpicking his writing style but has to do with a constant theoretical strategy in Habermas’s works, where he recurs to classifications that might be useful in a conceptual level but push his theory further away from empirical evidence. The problems associated with the clear-cut differentiation between, on the one hand, ‘labour’ and  ‘interaction’,  and,  on  the  other,  purposive-rational  action  and  communication,  in  Habermas’s epistemological  argument  (1971)  were  once  again  reproduced  in  the  TCA  through  the  subsequent differentiation of, first, instrumental and communicative action and, later on, the economic and political subsystems and the lifeworld. In Giddens’s words: I am unhappy with your distinction between system and life-world ―as I was with the differentiation between ‘labour’ and ‘interaction’ which appeared prominently in your earlier work.  If, as you say, the  separation  between  system  and  life-world  is  methodological,  how  can  it  also  operate  as  a substantive  distinction  within  modernized  societies?  Moreover,  your  use  of  systems  theory,  of notions such as ‘steering mechanisms’ and so on, seems to do scant justice to the active struggles of individuals and groups out of which history is made (Giddens, 1987:250). The  uncoupling  of  system  and  lifeworld:  its  consequences  for  an  explanation  of  social transformation In the second volume of The Theory of Communicative Action (1987), Habermas developed a three- part  model  of  modern  societies,  where  they  appear  divided  into  political,  economic  and  societal subsystems or arenas. According to Cohen and Arato (1992), the latter could ―generally and with some conceptual restrictions― be equivalent to the concepts of lifeworld, civil society, or the Parsonian notion of social community. They insist that such translation cannot be made without distortion, and propose to differentiate  two  distinct  levels  within  the  concept  of  lifeworld:  one  that  refers  to  the  “reservoir  of implicitly  known traditions” and background assumptions (1992:428),  and another that recognizes  the 19 “reproductive processes of cultural transmission, social integration,  and socialization” (1992:428).  The main point in their argument is that this differentiation of the lifeworld is the result of the modernization process  that  fostered  the  emergence  of  institutions  specialized  in  the  aforementioned  reproductive processes.  Habermas pointed  out  that  the  differentiation  of  the  economy from the  state  in  bourgeois societies had been mirrored by a similar differentiation in the lifeworld, “taking the shape of private and public spheres, which stand in a complementary relation to one another” (Habermas, 1987:319-320). In this way, the three-part model turns into a four-part model of society, in which the lifeworld stands as an analytical  category  ―with  some  minor  caveats―  at  the  same  level  as  the  economic  and  political subsystems. Habermas’s notion of the  lifeworld comes from Schutz’s and Luckman’s definition of this concept: “the unquestioned ground of everything given in my experience and the unquestionable frame in which all the problems I have to deal with are located” (Schutz and Luckman cited in Habermas, 1987:131). The lifeworld  thus  represents  a  taken-for-granted,  inter-subjectively  shared  background,  which  includes cultural  experiences and communicative interactions,  temporarily  stabilizing certain presuppositions in order to allow mutual understanding through meaningful utterances (Habermas, 1987:131). Individuals use  the  lifeworld  as  the  unproblematized  ―“until  further  notice”,  as  Luckman  says  (in  Habermas, 1987:130) ― basis  from which life  experiences  are  interpreted  and  conceptualized.  Therefore,  in the words of Simone Chambers: Communication  is  the  way  we  transmit  and  reproduce  our  lifeworld.  More  particularly,  we  can identify three activities that function as transmitters of the lifeworld: cultural reproduction, through which  traditions  and  cultural  meanings  are  passed  down;  social  integration,  through  which  we recognize norms of cooperation and interaction; and finally, socialization, through which we acquire identities both as collectives and as individuals. These three functions are symbolically mediated. We pass on cultural understandings, learn to live together under certain rules, and form our identities, by talking and communicating with one another (1995:241-242). Chambers (1995) has analyzed how this shared background is constantly regenerated and transformed. She speaks about situations when cultural justifications are no longer convincing to support a specific normative  claim,  and  thus  the  claim needs to  be regenerated or replaced through the  communicative 20 practice of providing reasons to support or dismiss that norm (Chambers, 1995:242). For our purposes, this issue is significant because it illuminates the process through which the lifeworld is constructed and regenerated in a day-to-day basis: According to Habermas this process often takes place unreflectively, in what he calls the “negotiation of  a  new situation definition.”  The negotiation is  informal  and  partial,  and is  characterized  by  a “diffuse,  fragile,  continuously  revised  and  only  momentarily  successful  communication  in  which participants  rely  on  problematic  and  unclarified  presuppositions  and  feel  their  way  from  one occasional commonality to the next” (Chambers, 1995:242). Once  again,  we  see  how  the  lifeworld  is  mostly  constituted  through  microscopic  bargaining  and discussion.  Partial  understandings  are  created through symbolic interaction in an attempt  to solve the normative questions that arise around specific issues. As Habermas and Chambers indicate, these are not discursive practices, but only partially reflective forms of everyday communication where the “force of the  better  argument”  is  constantly  overridden  by  threats,  bribes,  manipulation,  lies,  etc.  (Chambers, 1995:243). However, the negotiation of new situation definitions reproduces the validity of the norms that support the symbolic system of the lifeworld. Certainly, cultural reproduction, social integration and socialization, happen in every realm of social life and every subsystem of society, especially in the interactions that individuals experience in their daily routines. These transmitters of the lifeworld get into play and are reinforced every time two people meet and interact with each other, regardless if it is in a private household, a church meeting, the stock market, or the parliament house. Therefore, the lifeworld cannot be understood as a synonym for civil society or the societal sphere (not even an imperfect one), since it is an undercurrent of non-concretized rules, norms, practices, conventions and shared meanings, that cuts across all the subsystems despite the specialized languages that in some of them allow the unburdening of communicative practices. This is the reason why the distinctions between labour and interaction, and between instrumental and communicative action, are conceptual  categories  that  do  not  correspond  to  reality.  If  the  lifeworld  is  reproduced  within  the subsystems,  the  problem  is  not  just  that  the  subsystems  (or  the  specialized  languages  of  particular subsystems) have colonized the lifeworld (giving primacy to purposive-rational action, for example) but 21 that they are also the grounds in which the lifeworld is reproduced through daily interactions creating the incentives to perpetuate specific behaviour. These spaces foster networks of routinized social interaction where the regeneration of the lifeworld unfolds through situations of co-presence and all sorts of face-to- face interaction, including forms of partial communication. Despite differences in the overarching forms of morality, in the degree of industrial development, in the depth of the separation between the economic and the political spheres, or, if we accept the Habermasian evolutionary model, in their transition from pre-conventional  to post-conventional  forms of normative justification,  all  societies provide routinized situations of interaction through which the lifeworld is reproduced. In sum, Giddens’s second critique of Habermas points towards certain limitations that the distinction between the lifeworld and the system suffers when grasping the way in which cultural reproduction, social integration and socialization occur in a daily basis. The inaccuracy of the distinction cannot be solved by acknowledging an overlap among the subsystems or by referring to the colonization of the lifeworld by the economic or political spheres. The lifeworld and the subsystems cannot be seen as different categories at  the  same level  of  analysis,  rather  the  subsystems are  part  of  the  scenario,  of  the  space where  the lifeworld  operates.  In  a  way,  their  relationship  might  be  more  accurately  described  as  one  between continent  (subsystems) and content  (lifeworld).  Even if  the specialized subsystems of the market  and public  administration  develop  particular  languages  to  unburden  the  communicative  tasks  and  allow purposive-rational action for generalized goals, the lifeworld is still being reproduced through the daily interactions  that  occur  within  them.  This  not  only  challenges  Habermas’s  explanation  of  social reproduction in different stages of the evolution of societies, but fails to consider how human actions have an  effect  on  social  transformation.  In  other  words,  the  two  parallel  distinctions  (instrumental  / communicative action and lifeworld / system) make human agency recede from view in less-than-ideal societies. His account of the uncoupling of lifeworld and system ends up presenting social evolution as a unidirectional path in which every new level of system differentiation is achieved regardless of human action. It is precisely for these reasons that the TCA faces important obstacles to explain the complex dynamics  occurring  in  social  traps.  Conversely,  a  more  accurate  explanation  needs  a  more  inclusive 22 conception of praxis than individual agency as action coordination through discourse, and incorporate something similar to what Giddens called “practical consciousness” through the “reflexive monitoring of action” (1984:xxiii). I will come back to this point in the fourth section of this paper. 23 The “virtuous circle of discourse” Before moving forward and explaining how a reductionist  view of praxis as discourse  affects the TCA’s explanatory leverage of social traps, I want to address a second feature in Habermas’s work that also undermines its capacity to engage with less-than-ideal situations. Habermas’s model of discursive democracy seems to work in a self-referencing way, where legitimate norms, laws and institutions derive from deliberative processes and, at the same time, those deliberative processes nurture the cognitive and affective  capacities  that  the  citizens  require  to  engage  willingly  in  deliberation.  This  tautological argumentation  refers  to  what  I  have  called  the  virtuous  circle  of  discourse or  the  two functions  of deliberation: (1) providing legitimacy to norms in post-conventional societies, and (2) encouraging the self-transformation of the individual towards more autonomous forms of moral reasoning. Normativity in post-conventional legal orders In  Between Facts and Norms,  Habermas (1996) put forward the idea that  deliberation provides a source of legitimacy for the construction of legal norms in post-conventional societies. He reversed “Max Weber’s query  about  how political  legitimacy can result  from legality”,  asking  instead “how we can justify  the  legitimacy of  legality” (Benhabib,  1996:725).  From this  standpoint,  Habermas saw human beings as embedded in structures of communication through which they could build together the systems of norms to rule their actions. For the results of those processes of deliberation to be normatively valid, they should have been achieved through the discussion of validity claims that were accepted or rejected depending on “the force of the better argument” (Habermas, 1984). In other words, “just those norms deserve to be valid that could meet with the approval of those potentially affected, insofar as the latter participate in rational discourses” (Habermas, 1996:127). Essential to his argument is the question:  “do we obey the law because of the threat  of sanction backed by the state or because we see the law as reasonable and legitimate?” (Dodd, 1997:329). Habermas insists that individuals should be simultaneously subjects and participants in the construction of the law 24 that rules their societies. From that discursive participation would then arise the normative component of law, which, combined with the facticity element (the capacity of the state for coercion), would create a situation where law and politics precede and influence each other. Habermas follows Talcott Parsons in the belief that actors have behavioural expectations about each other’s  actions.  Therefore,  any  social  order  needs  to  rely  on  mechanisms  of  action  coordination (1996:139) to stabilize those expectations and prevent anomic action. Those action coordination problems involve either issues of interpersonal conflicts or the pursuit of collective goals, and in both cases that coordination  can  by  driven  by  value-oriented  or  interest-governed  motivations  (1996:139-140).  This constructs a two-by-two matrix that offers four types of conflict resolution strategies and collective will formation: consensus and arbitration, as forms of regulation of interpersonal conflicts; and decision by authority and compromise, as strategies for collective goal setting. This allows Habermas to assemble an almost social-evolutionary account on the “the co-original constitution of binding law and political power” (1996:141) that operates in two stages. First, in pre-modern societies a leader who enjoys some kind of transcendental normative authority displays the functions of a judge-king and concentrates the exclusive attributions to interpret the norms of the community, as well as the resources to factually enforce those laws. In the second stage that normative legitimacy and the factual political power are institutionalized into the administrative system: the state. “Not only does law now legitimate political power, power can make use of law as a means of organizing political rule” (1996:142). Nonetheless, in post-conventional societies,  there  is  no longer  the  transcendental  source of  authority  that  had originally  legitimized the judge-king to regulate interpersonal conflicts through arbitration or to decide on collective goals through authoritarian decisions. Habermas insists here that modern societies need then the motivating force that comes from consensus, from “discursively produced and intersubjectively shared beliefs” to provide law with  the  normative  content  that  had  once  been  invested  through  transcendental  sources  (Habermas, 1996:149).  For Habermas,  only  through discourse,  through communicative power,  the transition from sacred law to positive law can maintain that normative component, allowing action coordination to be more than simple compromise and to integrate value-oriented motivations in its daily operation. 25 Self-transformation of the individual Consensus through discourse becomes the motivational force that upholds norms and laws in  post- conventional societies. “At the same time, Habermas argues that only in a discursive context can moral capacities  develop  fully”  (Warren,  1993:218).  It  is  widely  assumed  by  participatory  democrats  that engaging in democratic participation is a social experience “likely to develop [in the individual] just those values  and  capacities  that  democracy  needs  to  be  a  viable,  thriving,  and  vibrant  system  of government”  (Warren,  1993:210).  In the following,  I  will  present  the second function of deliberation through  one  of  Habermas’s  early  attempts  to  explain  how discursive practices  operate  as  sources  of individual self-transformation. This explanation was later refined, but not substantially changed, in his Theory of Communicative Action (Habermas, 1984; 1987). In “Moral Development and Ego Identity”, Habermas (1979) correlated the ontogenesis of the ego with the notion of interactive competence, trying to explore the cognitive capacities required to engage in communicative  action.  As Thomas McCarthy mentions  in his  introduction to  Communication and the Evolution  of  Society (Habermas,  1979),  that  essay  is  part  of  a  project  to  distinguish  the  different dimensions  of  human development  (linguistic,  cognitive,  interactive,  and  ego),  in  order  to  provide  a general  explanation  of  the  interdependence  of  personality  structures  ―in  the  form  of  species-wide competences  achieved  and  ordained  in  a  hierarchical  sequence  of  stages―  and  social  structures (McCarthy in Habermas, 1979:xx). First  of  all,  Habermas  recognized  three  theoretical  traditions  that  had  treated  the  issue  of  the development  of  ego  identity:  “in  analytic  ego  psychology  (H.  S.  Sullivan,  Erikson),  in  cognitive developmental psychology (Piaget, Kohlberg), and in the symbolic interactionist theory of action (Mead, Blumer, Goffman, et al.)” (Habermas, 1979:73). Although these intellectual trends differ from each other in important aspects ―and for Habermas none of them “has as yet led to an explanatorily powerful theory of development” (1979:75)―, they share important commonalities that represent the departure point to eventually equate full moral autonomy with the capacities to function within a universal ethics of speech. 26 Among the commonalities that Habermas identified, three are especially relevant for the purposes of this paper: 4.  The  developmental  direction  of  the  formative  process  is  characterized  by  increasing autonomy. By that I mean the independence that the ego acquires through successful problemsolving, and through growing capabilities for problemsolving in dealing with― a) The reality of external nature and of a society that can be controlled from strategic points of view; b) The nonobjectified symbolic structure of a partly internalized culture and society; and c)  The internal  nature  of  culturally  interpreted  needs,  of  drives  that  are  not  amenable  to communication, and of the body. 5. The identity of the ego signifies the competence of a speaking and acting subject to satisfy certain consistency requirements.  […] It  consists  rather in a competence that is formed in social interactions. Identity is produced through socialization, that is, through the fact that the growing child first of all integrates itself into a specific social system by appropriating symbolic generalities; it is later secured and developed through individuation, that is, precisely, through a growing independence in relation to social systems. 6.  The  transposition  of  external  structures  into  internal  structures  is  an  important  learning mechanism. […] With this mechanism is connected the further principle of achieving independence ―whether from external objects, reference persons, or one’s own impulses― by actively repeating what one has at first passively experienced or undergone (Habermas, 1979:74). These three assumptions are in the core of the idea that a universal ethics of speech can only occur among individuals that have reached the highest stage of moral autonomy and viceversa. Moreover, they already point at the importance of social interactions: first through the integration of external structures and symbolic generalities; and later on with the differentiation of the self from those social systems. The next couple of steps in Habermas’s  account of the formation of ego identity are, first, to equate Kohlberg’s stages of moral consciousness with an action-theoretic framework (Habermas, 1979:82-83), and, secondly, to integrate in that framework the “reciprocity requirement” that bridges action structures with the stages of moral consciousness (1979:88-89). By doing this, he tried to prove that the cognitive mastery of “general levels of communication” and the “ability to give one’s own needs their due in these communication structures” (1979:78) are essential in the development of ego identity. Let me elaborate on each of these moves separately: 27 Habermas defines moral consciousness as one aspect of ego development,  the cognitive ability to make  moral  judgments  (1979:78).  Only  action  conflicts  that  are  capable  of  consensual  resolution (excluding force and “cheap” compromises) classify as morally relevant under this account. Therefore, following  Kohlberg,  there  are  six  moral  stages  divided  in  three  hierarchical  levels  (preconventional, conventional and post-conventional). Each one of these stages is defined by the consequences of action that  determine  the  outcome  of  moral  judgments,  going  from  physical  stimuli  (avoiding  physical punishment  and  looking  for  hedonistic  rewards)  in  the  first  stage,  to  the  capacity  of  making  moral judgments based on abstract ethical principles in the sixth and final stage. The second move in the argument involves comparing the stages of moral development with general qualifications for role behaviour (1979:82). In this context, behavioural expectations also evolve in three levels, travelling from individual and concrete expectations and actions, to the emergence of social roles adopted from the surrounding symbolic universe, and finally to the capacity to question those social roles. The purpose here is to emphasize how reaching the third level of role behaviour entails the possibility for the individual to distance herself from her own cultural tradition, abandon a role-dependent position, and observe roles and systems of norms as objects of discursive will-formation. Habermas adds to the symbolic universes of role behaviour and moral judgments the need for specific abilities (interactive competence) to move within those structures (1979:86). Moral consciousness will then be “the ability to make use of interactive competence for consciously processing morally relevant conflicts  of  action”  (1979:88).  Besides  the  degree  of  moral  autonomy  achieved  in  Kohlberg’s  post- conventional  level,  agents  in  the  last  stages  of  interactive  competence  still  require  certain  structural conditions to be able to solve those conflicts of action through consensus. At this point, Habermas argues that reciprocity embodies the structural conditions capable of shaping possible interactions (1979:88). The requirement of reciprocity (whether it is complete or incomplete, that is, whether two actors expect the same thing from each other or hold different expectations based on unequal power-relations) leads to each one of the stages of moral consciousness when it is applied to the levels of role competence. 28 Habermas’s argument has hitherto only focused on cognitive aspects of the ontogenesis of the ego. Nonetheless, he was careful to warn that a general theory of ego development should include affective and motivational development (1979:91). In the last part of his essay, he briefly addressed the motivational side of moral consciousness, the psychodynamics of superego formation and defence mechanisms that can explain “the discrepancies between moral judgment and moral action” (McCarthy in Habermas, 1979:xxi). The correlation between levels of interactive competence and stages of moral consciousness […] means that someone who possesses interactive competence at a particular stage will develop a moral consciousness  at  the  same stage,  insofar  as  his  motivational  structure  does  not  hinder  him from maintaining,  even under  stress,  the  structures  of  everyday action in the  consensual  regulation of action conflicts” (Habermas, 1979:91). I  have  described  thus  far  how  Habermas  draws  a  logical  connection  between  higher  stages  of interactive  competence  and  the  development  of  moral  autonomy  (and  thus  capacities  to  engage  in discursive participation). However, it remains unclear how the self-transformation function of discourse operates. How exactly do discursive practices, particularly participation in discourse in political settings, develop capacities of autonomy? Warren (1993) unpacks the answer to this question by distinguishing three different levels of analysis in Habermas’s argument: (1) “the potentials within social relations for the development  of  autonomy”;  (2)  “the  specific  moral  competencies  required  by  situations  of  political conflict”; and (3) “the motivational force of speech in the direction of autonomy” (Warren, 1993:216). The first  of  these levels of  analysis  relates  to  the common assumption in all  the aforementioned traditions of ego psychology on the social constitution of identity. The individual develops interactive competencies  and  autonomy  through  socialization,  by  her  integration  into  a  social  system  and  the appropriation  of  its  symbolic  generalities;  and  then  through  individuation,  by  adopting  the  ability  to distance  herself  from  those  social  systems  and  reconstruct  her  ego  identity  by  discursive  means (Habermas, 1979:74). These two processes can only be fully undertaken through social relations. The second level of analysis is grounded on Habermas’s claim that moral development in Kohlberg’s theory is already embedded in social relations and “general structures of interaction” (Warren, 1993:218). At this point, reciprocity plays an important role in the construction of the self, since it develops out of 29 mutual recognitions  which, depending on whether they are complete  or incomplete,  push forward the ontogenesis  of  the  ego through the different  stages of  moral  consciousness.  Warren argues that  since individuals  are  unlikely  to  challenge  themselves,  they  must  be  exposed to  others  and  to  the  need  of discursively justifying their needs and interests (1993:219). This is why, in the context of a discursive model of democracy, moral capacities have to go past Kohlberg’s six stages into a seventh moment that Habermas calls  “discourse ethics”, where generalizable norms of action can only be built discursively (Habermas, 1979:90; Warren, 1993:218-129). Finally, the third level of analysis poses the question about why individuals should resolve conflicts by means of democratic discourse. Warren follows Habermas in arguing that speech frames cognitive capabilities in the context of the social relationships (that is, the levels of role behaviour) in which the individual is embedded. [S]peech  intrinsically  relates  cognitive  competence  and  motivation.  To  the  extent  that  we  deal cognitively with the relations that situate us in the world, we do so through the medium of language. But since language is not  private,  since it  is learned and sustained intersubjectively,  we are also motivated  to  come  to  understandings  with  others  about  the  validity  of  our  claims  about  these relations. Cognitive veracity depends on intersubjective validity (Warren, 1993:220). Further on, Warren elaborates on the gaps in Habermas’s attempt to flesh-out the assumptions of the self-transformation thesis, addressing the necessity to include the affective dimensions of the self in any theory that tries to explain the psychological impact of discursive democracy on the individual (Warren, 1993:221). He begins by bringing up the premise in Habermas’s argument that individuals, besides being morally autonomous, would also be willing to solve conflicts through discourse even when other means “might be more satisfying from the perspective of maintaining a psychodynamic balance of desires and impulses” (1993:221).  In politics,  ideologies might be examples where such an assumption is hard to uphold.  Warren,  commenting  on  Habermas’s  Knowledge  and  Human  Interests  (1971),  compares ideologies  with  the  effects  that  neurosis  has  on  individuals,  where  unconscious  resistances  hinder discourse because it would challenge the identity of the self. Under those circumstances, the cognitive approach  to  the  self-transformation  thesis  seems  to  be  insufficient,  and  thus  Warren  introduces  the 30 therapeutic  models  of  communication  as  a  complementary  element  to  Habermas’s  account  on  the influence of discursive forms of democracy on the ontogenesis of the ego (Warren, 1993:223). Described politically, therapeutic critique involves dispelling the kind of false consciousness that is a matter of cognitive incapacity. Political enlightenment, even in Habermas’s terms, will mean not just developing  cognitive  capacities  but  also  transforming  character  structures  in  ways  that  allow individuals to engage in discourse that is not ‘distorted’ by self-defeating psychodynamics. Discursive democracy, if it is to be workable in a world where character structures are not ideal, will of necessity involve a therapeutic dimension (Warren, 1993:225). The question is whether therapeutic communication can be directly included in democratic theory since politics are not very similar to a psychoanalyst’s office. Discursive democracy is only workable under  at  least  minimum  symmetric  conditions  of  complete  reciprocity  between  morally  autonomous individuals. These conditions would then provide enough incentives to engage in deliberation. The logical conclusion suggests that the pre-requisites to engage in discourse need to be developed before democratic participation can reach undistorted levels of communication. It is essential to mention here that Habermas is aware of these obstacles and therefore regards therapeutic critique not as a form of discourse but as its “presupposition”,  as  an  antecedent  condition  to  solve  problems  of  distorted  communication  before autonomous individuals can engage in discursive forms of normative justification (Warren, 1993:225). At this  point  appears an inevitable  question for any theory concerning false consciousness and cognitive barriers.  Rephrasing  this  question in  his  on terminology:  how can a society move from structures  of distorted communication to discourse as the source of communicative power? The previous paragraphs have attempted to describe the reciprocal effects that the two functions of deliberation have upon each other. Furthermore, low levels of distortion in communication are a necessary prerequisite for societies to gain access into this self-reinforcing form of democratic practices, where: (1) discursive procedures are the source of legitimacy for laws and institutions, and (2) autonomous citizens feel  comfortable  without  transcendental  justifications  for  norms  of  action,  for  whom  instrumental 31 calculation  is  not  necessarily  the  prevalent  motivation  to  live  up  to  their  collective  agreements  (the element of coercion and the facticity of law), and who are willing to engage in discourse to define those norms of action without perceiving communication as a threat to their particular identities. I want to re-describe now the self-reinforcing model of discursive democracy in a different  way. Quoting  Simone  Chambers’s  remarks  on  the  circularity  of  the  Habermasian  account  of  discursive legitimacy: The institutional  form of democratic  will  formation must  itself  meet  with standards  of  discursive validity. This appears to lead to a circle: The institutional arrangements that make discourse possible must be justified by a discourse. If the mandate to set up a discourse can only be conferred in a discourse, we are left with no means of justifying the initial establishment of discourse (1995:241). Indeed,  I  have  introduced  a  further  complication  by  analyzing  in  parallel  the  two  functions  of deliberation.  It  is  not  just  that  “the  institutional  arrangements  that  make  discourse  possible  must  be justified by a discourse”, but also that the moral and interactive competences that make discourse possible must be developed through discourse. In a way, what this implies is that structural arrangements under which  social  interaction  occurs  constrain  the  range  and  form  that  those  interactions  can  adopt. Socialization processes that  depend on how those social  interactions  happen are thus tilted in certain directions that may or may not stimulate the development of ego identities compatible with discursive justification. In  other  words,  the  circular  relationship  between  individual’s  identities  and  competences  and discursively constructed norms and institutions can also be seen as the ideal relationship between agents and structures that are reciprocally constituted through discourse. The Habermasian understanding of ego ontogenesis suggests the appearance of individual agency only in the latter stages of moral development, when the individual differentiates herself from those constraining structures and analyzes them in a self- reflective  manner.  This  necessarily  poses  the  question  about  the  point  of  entry  to  those  dynamics favourable to the establishment of a (discursively) democratic rule of law. Once we observe the negative cases,  those where  discursive dynamics  are  not  present,  we face a  typical  example  of  a  social  trap: structures  that  promote  forms  of  asymmetrical  communication  would  tend  to  hinder  the  progression 32 towards higher stages of moral  consciousness among those involved.  The scenario that would ideally predate the development of the virtuous circle of deliberation actually seems to move in the opposite direction than that expected by the Habermasian evolutionary argument. It displays a path-dependence where asymmetrical structures, mostly caused by vertical power relations and fragmented social groups, develop constraints  that  reinforce  cognitive barriers  and non-discursive habits  for action coordination. These dynamics in turn inhibit the competences and motivations that agents would require to discursively modify those structures. Habermas’s ideal explanation of law is challenged when someone tries to observe less-than-ideal societies that have not developed a discursive form of social organization and coordination for collective action. On the contrary, many of these societies rely on the factual element of law (strong coercive power by the state) or have improvised non-democratic mechanisms to provide some sort of normative content to precarious forms of order, mostly based on asymmetrical power relations and authority: i.e., narco-culture in Latin America,  patronalism among Mafia organizations,  or  factional  identities  in opposition to  the antagonistic  other in situations of ethnic, religious or class conflict. For this reason, in order to observe how  the  virtuous  circle  of  discourse  reproduces  the  deadlock  of  social  traps,  we  need  to  follow Habermas’s argument, but instead of looking at the ideal cases where discursive democracy is solid and self-reinforcing, we need to focus on the negative situations where this assumption does not hold. The purpose of Habermas’s theory of law is to show how discursive democracy can perpetuate itself once it is established, but the very same reasons that allow this to happen are the ones that inhibit the development of a democratic rule of law in less-than-ideal societies. The interdependence of personality and social structures that Habermas explains through his thesis of ego  ontogenesis  starts  from  the  commonalities  shared  by  analytic  ego  psychology,  cognitive developmental psychology, and the symbolic interactionist theory of action. The three points8 mentioned before refer to one of the central conclusions that Rothstein draws from the logic of social traps (2005:8): 8 1) Moral  autonomy as growing  capabilities for problemsolving when engaging with the object-world,  with the symbolic  structures  of  society,  and with the internal  nature  of  interpreted needs;  2) the cognitive  developmental process travelling from socialization to individuation; and 3) the separation of external and internal structures. 33 that individuals base their actions on the expectations about others’ behaviour. Nevertheless, this kind of strategic  behaviour  does  not  necessarily  imply  the  kind  of  discursive  consciousness  that  Habermas discovers in the last stages of ego development, where the individuation process has occurred and external structures have been internalized and challenged. However, strategic behaviour does entail a degree of self-reflexivity that manifests itself in a different way, under something that would be closer to the idea of subjective rationality (Rothstein, 2005:35) or “practical consciousness” (Giddens, 1984:xxiii). If we include in our critique the motivational and affective sides of the ego besides the cognitive one, we  also  observe  that  strategic  behaviour  under  non-discursive  conditions  will  actually  prevent  the development of discourse, enhancing the social trap in which nobody is willing to cooperate with the rest. Warren’s point on the “general structures of interaction” and the reciprocity requirement through which they  shape  social  relations  is  particularly  useful  for  this  task (Warren,  1993:218).  The  conditions  of reciprocity  will  affect  the  trust  in  the  behaviour  of  others,  to  the  extent  that  negative  or  incomplete reciprocity  (tainted  by  power  relations,  by  past  experiences  of  deceit  and  betrayal  or  by  negative perceptions of the  other) might push the individual to withdraw from discourse and action coordination and pursue her own self-interests. In the long run, if these reciprocity requirements are not fulfilled, the transition  to  latter  stages  of  moral  development  and  the  acquisition  of  discursive  capacities  will  be hindered, reinforcing the individual’s “taste” for self-interested non-cooperative behaviour. Lastly, Warren (1993) also asked why individuals would prefer to go through the costly process of discursive forms of conflict resolution, even if they have achieved the higher levels of moral autonomy, and  not  other  strategies  that  might  be  less  threatening  to  their  personal  identities:  i.e.,  bargaining, intimidating,  accommodating,  or  walking  out  of  the  discussion.  Indeed,  the  structural  conditions  that determine  social  interaction  might  not  provide  the  incentives  to  engage  in  discursive  practices. Furthermore, in a social trap situation they would tend to reinforce instrumental calculation as part of strategic action, turning normative questions irrelevant and maintaining order only through the factual element  of  law.  In these situations,  order  becomes particularly  precarious  and oscillates  between this calculation of costs and benefits regarding the authority’s capacity for coercion and the constant attempt to 34 overrun those constraints to maximize private interests. In other words, when we apply Habermas’s line of reasoning to less-than-ideal societies, we discover that the same elements that interconnect the virtuous circle of discourse are the ones that make social traps so difficult to solve. Before the virtuous circle of discourse can be achieved, those damaging circularities need to be transformed. 35 Praxis, the virtuous circle of  discourse, and social traps In the two preceding sections I have tried to point at some features of the TCA that make it incapable of engaging successfully with topics such as social traps. Moreover,  these features are all  interrelated because they grow from the basic assumptions without which the discourse principle cannot survive. In this section I will tie up my argument by recapitulating how those assumptions lead to three characteristics of the TCA that make it blind to the complexities of social traps. I began this essay by citing Luhmann’s critique of the unexplained conditionality in the discourse principle:  “Those norms for action are valid,  to which all  potentially  affected persons could agree as participants in a rational discourse” (Habermas, 1996:107). The conditional introduced through the word could is  explained  by  Habermas through the  developmentalist  process  of  ego  ontogenesis  and  moral autonomy. Individuals can participate in rational discourse once they achieve the last stages of moral and cognitive  development  and  are  capable  to  situate  themselves  within  a  universal  ethics  of  speech (Habermas, 1979:89). I argue that by grounding the possibility of communicative action upon cognitive, moral  and  interactive  competences  that  are  achieved  in  a  unidirectional  and  univocal  path  the  TCA acquires three characteristics that unfold sequentially: First, the evolutionary imprint in Habermas’s attempt to establish discourse as a normative goal and as  the  last  step  of  cognitive  development  produces  a  deterministic  view  of  social  transformation. Habermas’s attempts to universalize his model fell again into a quasi-transcendental first philosophy, in which societal change occurs despite human agents. “Instead of finding progressive structures in history, [Habermas]  tries  to  find  them  in  pre-history,  in  ‘anthropologically  deep-seated  general structures’” (Couzens Hoy in Couzens Hoy & McCarthy, 1994:158).  If there is a path in the form of cognitive development towards collective systems of post-conventional morality, we should expect a long- term evolutionary pattern in which actors don’t really have a say. The determinism in Habermas’s theory is  also symptomatic of  the  transcendentalism that  lingers underneath his adaptation of Piaget’s distinction between competences and interactions: 36  Since  language  precedes  humans,  the  structure  of  language  conditions  human  interaction. Historiography can chart these interactions, but only theory, according to Habermas’s adaptation of Piaget, can reach more deeply and describe the development of universal competences that underlie these interactions (Cozens Hoy in Couzens Hoy & McCarthy; 1994:159). This  shift  towards  competences  seems to  help  Habermas move away from a  transcendentalist  set  of speculations of universal history and at the same time maintain a cornerstone for his TCA. Nonetheless, this theoretical move only conceals the way in which he still relies upon some kind of univocal notion of “truth” that anyone could agree upon under the right conditions. The implications that this move towards competences has for the deterministic imprint of the TCA is one of the first reasons why it is not capable to provide a convincing explanation of social traps. An interesting way of developing this point further is by drawing parallels with Douglass North’s account of social traps and its inbuilt determinism (1990). In the first section of this paper I mentioned how the notion of subjective rationality can be understood in two different senses: one, along North’s argument,  that sees it as the result  of incomplete information resulting from imperfect feedback processes and human’s limited computational capacity; or, secondly, as the strategic manoeuvring of knowledgeable agents within certain structural constraints. If we follow the logical implications of North’s line of reasoning, we discover that Habermas falls into similar problems when trying  to  avoid  the  determinism of  instrumental  views of  human action.  For  North,  subjective models of rationality are the result of long-term processes of repeated interaction between actors.  North argues that: institutions basically alter the price individuals pay and hence lead to ideas, ideologies, and dogmas frequently playing a major role in the choices individuals make. […] [T]he subjective and incomplete processing of information plays a critical role in decision making. It accounts for ideology, based upon subjective perceptions of reality, playing a major part in human beings’ choices (1990:22-23). For North, the difference between informal and formal institutions is one of degree; they stretch along a continuum where societies, as they become more complex, formalize the rules of the game (1990:46). He argues that informal constraints reduce uncertainty as compared to a world of no institutions, and are the immediate source of choice in daily interactions,  even if there  are underlying formal rules  (1990:36). These informal constraints are part of the cultural heritage (maybe the result of previous formal rules that 37 have been internalized). The solutions they offered to exchange problems in the past carry over into the present (1990:37) and are recurrent in how people interact. To sum it in a nutshell: “actors make choices based  on  subjectively  derived  models  that  diverge  among  individuals  and  the  information  the  actors receive  is  so  incomplete  that  in  most  cases  these  divergent  subjective  models  show no  tendency  to converge” (North 1990: 17). A  one-dimensional  understanding  of  human  motivation  and  action  remains  in  North’s  work, regardless  how culturally-sensitive  it  attempts  to  be.  From this  perspective  of  “subjective  models  of reality” or rationality, cultural differences and ideologies are only the result of divergent past experiences that  contaminate  with  wrong  information  purely  instrumental  calculations.  The  assumption  is  that  if feedback processes were complete and actors received full information, their subjective models would converge completely because they would also replicate an objective law-like universe. Habermas’s whole project as a critical theorist has indeed taken the shape of a strong critique against a one-dimensional account of human action: communicative rationality is built precisely as a description of human motivation and action that  includes  normative  and  aesthetic  principles  in an attempt  to  go beyond mere instrumental motives. However, the developmentalist argument and the distinction between instrumental  and  communicative  rationality  seem  to  operate  in  a  similar  way  as  North’s  notion  of incomplete and complete information. If North upholds a conventional understanding of rationality and objective truth that overlap under perfect (ideal) conditions, the concepts of universal pragmatics and the ideal speech situation place Habermas in a similar position. For North, the difference between incomplete and complete information and the resulting models of reality is one of degree, while Habermas insists that the difference between instrumental and communicative rationality is one of kind. Habermas’s hesitations to adopt a full-fledged post-foundationalist  view of truth and rationality  makes him ground his theory upon the ideal speech situation and the notion of “the force of the better argument”, which in turn are supported by the developmental explanation of ego ontogenesis. Under this framework, strategic action guided by instrumental rationality is the result of intermediate stages of cognitive, moral and interactive development,  or of adverse motivational structures, whereas communicative action successfully guides 38 collective will-formation and conflict resolution in the last stage of a universal ethics of speech and under ideal  conditions  of  undistorted  communication.  This  is  where  Habermas,  despite  all  the  obvious differences,  is closer to North than he would be willing to acknowledge:  in the end of the day, how different is a claim “to which all potentially affected persons could agree as participants in a rational discourse” from the idea of increasingly convergent subjective models of reality that result from more complete  processes  of  feedback  information?  When  North  speaks  about  ideal  conditions  (complete information) does he mean something substantively different from Habermas when he speaks about ideal conditions  (undistorted  communication)?  Are  not  both  talking  about  the  cognitive  capacities  and  the access to information / communication that actors require to be able to self-reflect upon their interests and preferences? An important  difference along these lines needs to be raised:  North believes only in incremental change and does not recognize it as the result of human agency. This could only happen if individual actors  had  access  to  complete  information  and  an  extraordinary  computational  capacity  to  use  that information to predict every possible unintended consequence. This information should then be used to self-reflectively observe how previous institutions can be tainting their lines of reasoning. Needless to say, such an expectation is unrealistic and thus only an ideal category for North. Habermas, instead, believes that human actors do have a say once they engage in discursive practices. For him, such a computational capacity is not necessary, because there is no objective information from the real world that needs to be sorted out, but utterances that need to be interpreted, modified, dismissed or adopted only through the force of the better argument. Instead, certain moral and cognitive competences are the prerequisites for agency. Habermas insists that the ideal conditions for communicative action, and thus human agency, are achievable or, at least, can be approximated. North doesn’t. Nevertheless,  Habermas  still  keeps  a  strong  deterministic  imprint  when  he  establishes  those  as necessary  conditions  for  individual  agency  by  understanding  human  praxis  only  as  discursive participation. In other words, for him, individuals’ capacity to transform structural processes is obstructed until  they reach the last  levels of moral autonomy and interactive  competence,  when they can finally 39 intervene self-reflectively upon those structural conditions. The question, once we talk about social traps, remains unresolved: how do we build those conditions of undistorted communication in the first place? If there is so little room for human agency when motivational structures hinder discursive practices, then how can actors change those motivational structures? Moreover, such a poor consideration of praxis and individual agency makes Habermasian Critical Theory blind to situations like social traps, where less- than-ideal societies that are not deeply embedded in communicative practices utilize different mechanisms of collective decision-making and conflict-resolution. Under these circumstances, individual actors make use of their agency without recurring to discourse but entrenching themselves in secured positions that are perpetuated  in  their  daily  interactions.  At  the  same time,  social  transformation  occurs  as  part  of  the incremental changes produced by the unintended consequences of actors that strategically appeal to the interlocked systems of incentives and beliefs to guide their action. David Couzens Hoy has discussed how Habermas’s evolutionary theory fails to move away from the transcendentalism of universal history and the implications that this has for the TCA (in Couzens Hoy & McCarthy,  1994:158-1964).  By following  Piaget  in  the  distinction  between interactions  (contingently charted by history) and competences (as statable rules that predate interaction), Habermas’s theory tries to reconstruct those rules that intervene in ethical reasoning. I have already discussed how, by making this move and recurring to the developmentalist argument of ego ontogenesis, Habermas abandons “one set of empty speculations, those of universal history, for another,  those of his evolutionary vision” (Couzens Hoy in Couzens  Hoy and McCarthy,  1994:158-159).  However,  Couzens  Hoy adds in his  critique an interesting point to explain why this transcendentalist influence makes Habermas miss the mark of how ethical judgments actually occur. Following Bernard Williams, Couzens Hoy argues that moral judgment does not operate like a linguistic rule that precedes human interaction, but involves something like the Aristotelian concept of phronesis or practical wisdom. This means that judgments depend more on social and  historical  factors  rather  than  universal  principles,  since  ethical  action  requires  more  than  the knowledge of moral principles and demands “something like the skill of seeing what is called for in a 40 practical situation”, which can only be obtained by the individual’s situatedness in a specific community (1994:160). The absence of a relevant consideration of phronesis as individual action in the TCA leads us to the second problem that it faces when observing social traps. The dichotomous view of human rationality and action ―where instrumental rationality is linked with false-consciousness and the unreflective pursue of perceived  self-interests,  and  communicative  rationality  is  linked  with  discursive  competence  and individual  agency―  pushes  Habermas  to  draw  the  distinction  between  lifeworld  and  systems. Furthermore, the idealized alignment of economic and political subsystems with instrumental action, and of the public sphere with communicative action, results  from the absence of a concept like “practical consciousness” or “subjective rationality”. A better  consideration of  phronesis would explain how the reproduction and transformation of the intangible rules and resources that guide individual action operate across all the different subsystems of society. Although  Habermas  abandoned  the  distinction  between labour  and  interaction  of  his  first  phase because it was too simplistic and incapable to grasp practical situations (Outhwaite in Ritzer, 2003:232), the influence of that line of reasoning remained palpable after his turn to language (Livesay, 1985:67; Giddens,  1987).  This  is  particularly  important  in  Habermas’s  distinction  between  instrumental  and communicative action, since this dichotomy is still based upon conceptual distinctions that face limitations to  explain  how  social  reproduction  and  transformation  work.  The  three  activities  that  function  as transmitters of the lifeworld (cultural reproduction, social integration, and socialization) are put in practice constantly every time two individuals interact with each other. Even those interactions that occur within the economic and political subsystems carry on these tasks through processes of partial communication (Chambers, 1995:242). An attempt to distinguish motivational structures between systems of incentives that guide instrumental action and systems of beliefs that would be contested within the public sphere does not  manage to integrate both as complementary factors that shape human action and determine social reproduction  and  transformation.  Not even pointing  at  the  colonization  of  the  lifeworld  is  enough to achieve this  task,  since it  still  assumes that  individuals constantly  have to choose between these two 41 systems of motivations in their daily lives as two separate sets of structural conditions. In less-than-ideal situations  (and  I  would  argue that  even in  those societies  that  lead their  action coordination  through discourse)  these  two systems  are  intricately  intertwined  and  operate  conjunctively  in  every  realm of society through micro-processes of interaction and daily ethical judgments.  If we accept the idea that human action is guided by expectations  about others’  behaviour,  we can then follow Peyton-Young’s definition of praxis, which Rothstein borrows (2005:37), to show how systems of incentives and systems of beliefs combine to shape those expectations. This is a fundamental issue that must be considered in any observation of self-reinforcing phenomena such as social traps, since it is the result of that combination what hinders the possibility of cooperation and collective action. A general description of the economic or political subsystems in advanced capitalist societies as reified spheres dominated by purposeful-rational action can only be accurate at the collective level. Individuals that act within those systems are always challenging  and  bending  them  through  strategic  actions  that  require  an  understanding  of  subjective rationality through which people use the cultural repertoire as a tool-box in their benefit (Giddens, 1984; Rothstein, 2005). Of course, the realms of work and interaction are not absolutely separable since: a) human labor ordinarily takes place through the medium of social relations which are normatively structured, b) technical  rules  themselves  have  a  conventional  quality,  and  c)  communicative  action  ordinarily involves  not  just  formal attempts  to  understand  and  be understood by others,  but  also efforts  to influence or control their subsequent action (Livesay, 1985:67). Regardless of how we call this competence of human agents to guide their lives in specific social and historical contexts ―practical consciousness,  phronesis, subjective models of reality, culture as a tool- box―, we need to acknowledge this level of agency if we want to provide a consistent explanation of social reproduction and transformation, which in turn has to be at the core of any approach to social traps.  The claim that civil society (Benhabib, 2002) or the public sphere (Habermas, 1996) are the ideal places where systems of beliefs can be contested, and where, for example, “the resolution of multicultural dilemmas [can be undertaken] through processes of will- and opinion-formation” (Benhabib, 2002:106), should  not  be  justified  by  drawing  strong  distinctions  between the  kind  of  competences  and  action- 42 motives  that  guide  behaviour  in  different  realms  of  social  action.  An  illuminating  example  of  the consequences that this has for the empirical applicability of the TCA has been advanced by the feminist critiques of Habermas: Now what are the critical insights and blindspots of this model? Let us attend first to the question of its empirical adequacy. And let us focus, for the time being, on the contrast between “the private sphere of the lifeworld” and the (official) economic system. Consider that this aspect of Habermas’ categorical  divide  between  system  and  lifeworld  institutions  faithfully  mirrors  the  institutional separation  of  family  and  official  economy,  household  and  paid  workplace,  in  male-dominated, capitalist societies. It thus has some prima facie purchase on empirical social reality. But consider, too, that the characterization of the family as a socially-integrated, symbolic reproduction domain and of the paid workplace, on the other hand, as a system-integrated material reproduction domain tends to exaggerate the differences and occlude the similarities between them (Fraser, 1985:106-107). Finally, the determinism in the TCA is further enhanced by the circularity of the two functions of deliberation. From this standpoint, there seems to be an insurmountable abyss between, on the one hand, post-conventional societies that successfully establish their forms of collective regulation and engage in conflict-resolution and collective  will-formation through discursive practices,  and,  on the other,  those societies that operate through a stable but inefficient equilibrium in which actors are entrenched in self- interested, non-cooperative behaviour. As I pointed out before, the virtuous circle of discourse provides an explanation  of  how discursive practices  can  be  perpetuated  once  they  are  established,  at  the  cost  of providing little insight into how situations of distorted communication can be transformed. The recursive influence  between  social  structures  and  individual  agents  that  permits,  in  certain  cases,  the  circular relationship between consensus  based on discourse  and a  citizenry  composed of  morally  autonomous individuals,  also  closes  the  trap  between  generalized  social  distrust  and  individual  non-cooperative behaviour in less-than-ideal situations. It is for this reason that Critical Theory requires a more coherent and unified theory of  praxis if it wants to address social traps and explain social transformation. Going back to Rothstein (2005), the circular dynamics of social traps, through which individual actors determine their behaviour in relation to their expectations about others’ actions, demand a detailed explanation of the recursive mechanisms that  connect  social structures  and individual actions.  Why is it  that “individual rationality can be collective irrationality”? Under which conditions are expectations stabilized to grant 43 social  order?  How  exactly  does  collective  memory  shape  individual  preferences  and  resources  to determine what will count as an individual rational action in specific contexts? Most importantly: In the cases that have successfully escaped from a social trap, which individual practices and through which unintended consequences pushed social transformation in that direction? 44 Conclusions Rothstein’s  critique of  culturalist  and  rationalist  approaches  to  social  traps  had to  do with three common problems of both traditions: a) their unrealistic assumptions of human action, b) the determinism in their conclusions, and c) their inability to explain social transformation. From my previous discussion of the  TCA,  I  argue that  these problems are  also present  in Habermas’s  work although for different reasons. The first one, the unrealistic assumptions of human action, appears in Habermas’s reduction of human agency to discursive practices and the dichotomous conceptualization of human behaviour as either instrumental or communicative action, which forces something like practical consciousness to recede from view. The absence of such a concept does not become a problem for Habermas to address the recursive dynamics between social structures and individual action in ideal situations through the virtuous circle of discourse. However it does become a problem if he wants to explain the transitions from pre-discursive scenarios  to  post-conventional  societies  with  moral  and  legal  systems  grounded  on  discourse. Interestingly, Habermas’s TCA builds a very strong idealization of human agency through the notion of individual  autonomy,  but  fails  to  consider  how  human  actions  can  push  a  society  through  the developmental trajectory from pre-conventional to post-conventional forms of morality. The TCA’s (a) unrealistic  assumption  of  human  action  (not  considering  individuals  as  knowledgeable  agents  that strategically deploy culture as a tool-box) leads to its (b) underlying determinism. Such a deterministic imprint in the TCA can adopt two possibilities: one, closer to Habermas’s expectations, would suggest that the  “anthropologically  deep-seated  general  structures”  of  language  and  cognition  that  predate  the individual  would push societies  towards  post-conventional  forms of  morality  regardless  of  individual action. However, a second and contradictory possibility is raised by his account of the recursive dynamics between agents and structures. This circularity, combined with the absence of a thorough consideration of practical  consciousness,  seems  to  generate  a  path-dependence  in  the  opposite  direction,  where  non- discursive practices and strategic action would be self-reinforcing, closing an inescapable social trap. This 45 is where the third problem, (c) the TCA’s inability to explain social transformation, becomes evident and undermines its emancipatory potential as a critical theory. Stephen Bronner has pointed at the problems that Habermasian fundamental categories face once questions about social or historical specificity are raised, as well as the implications that this has for the project of Critical Theory as a whole: A  real  “postmetaphysical”  philosophy  of  a  critical  sort  would  concern  itself  not  with  the deconstruction or reconstruction of “truth” but with the specification of those material constraints on its  pursuit.  Neither the legal nor the linguistic theory of Habermas can link the prerequisites for communicative  competence  or  the  stages  of  moral  evolution  with  the  reality  of  compromise, violence, and the structural imbalance of power (2002:209). An implicit objective of this paper has been to suggest some paths other than the TCA that Critical Theory can follow to achieve a closer connection between its theoretical categories and their applicability in non-ideal situations. If we follow the critique of the TCA that I have presented so far, two questions remain for further research: 1) What can a critical approach contribute to the literature on social traps? And, 2) What should a critical approach to social traps consider in order to offer a better explanatory leverage and hold a stronger emancipatory potential? I will leave the first of these questions unaddressed, only suggesting the importance of maintaining on the one hand, praxis and social transformation, and on the other, power relations and non-idealistic accounts of human action and social dynamics, in the center of concern. The self-reflexivity that rests at the core of Critical Theory should be its  major distinction from the current  literature  on social traps. Regarding the second of these questions, the previous discussions should have provided a few points to take into account if one wants to develop a critical methodology for the study of social traps: 1. Avoiding  quasi-transcendental  or  evolutionary  arguments,  which  might  produce  analytical categories that lose their explanatory power once they are contrasted with social and historical specificity. 46 2. Providing a thorough explanation of the recursive interplay between structure and agency to avoid fallacious assumptions of human action, whether cultural determinism, voluntarism, or atomistic descriptions of unencumbered selves. 3. Building  a  wide  understanding  of  praxis  that  includes  instrumental,  practical  and  discursive rationality. This is essential to be able to provide an explanation for social transformation in non- ideal situations, or, in other words, to explicate the immanent processes that may or may not lead towards those ideal scenarios. 4. Developing a typology of institutions that includes informal cultural practices and worldviews as determinant factors for human action. This is a major point of discussion in Rothstein’s work (2005:40). Although he is aware of the need of providing a definition of institutions as inclusive as  possible  to  be  able  to  incorporate  the  impact  of  informal  practices  in  shaping  political outcomes, he still believes that maintaining categorical distinctions between political and other kinds of institutions is necessary. At this point, his emphasis on formal institutions (the welfare state  in  Sweden,  for  example)  makes  him  overlook  other  kinds  of  repetitive  behaviour  that through  daily  interactions  reproduce  or  transform  social  and  political  structures.  A  critical approach,  not  only  for  purposes  of  explanatory  leverage,  but  also  to  be  as  self-reflexive  as possible, needs to develop theoretical strategies to cope with those institutionalized patterns of behaviour without conflating them with formal political institutions. 47 References Antonio, R. J. (1981). “Immanent critique as the core of critical theory: its origins and developments in Hegel, Marx and contemporary thought” in British Journal of Sociology, Vol. 32, No. 3 (Sept., 1981), pp. 330-345. Bates,  R.  H.,  De  Figueiredo,  R.  J.  P.,  and  Weingast,  B.  R.  (1998).  “The  Politics  of  Interpretation: Rationality, Culture, and Transition” in Politics & Society, Vol. 26, pp. 221-256. Baxter, H. (1987). “System and Life-World in Habermas’s ‘Theory of Communicative Action’” in Theory and Society, Vol. 16, No. 1 (Jan., 1987), pp. 39-86. Benhabib, S. (1996). “Review of Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy” in The American Political Science Review, Vol. 91, No. 3 (Sep., 1997), pp. 725-726. Benhabib, S. (2002). The Claims of Culture: Equality and Diversity in the Global Era. Princeton: Princeton University Press. Bohman, J. (1994). “Complexity, Pluralism, and the Constitutional State: On Habermas’s Faktizität und Geltung” in Law & Society Review, Vol. 28, No. 4 (1994), pp. 897-930. Bohman, J. (2007). Democracy across Borders: From Dêmos to Dêmoi. Cambridge: The MIT Press. Brandom, R. B. (2000). Rorty and His Critics. Malden: Blackwell. Bronner, S. (2002). Of Critical Theory and Its Theorists. New York: Routledge. Chambers,  S.  (1995).  “Discourse  and  democratic  practices”  in  White,  S.  K.  (ed.),  The  Cambridge Companion to Habermas. Cambridge: Cambridge University Press, pp. 233-260. Chriss, J. J. (1998). “Review of Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy” in Theory and Society, Vol. 27, No. 3 (Jun., 1998), pp. 417-425. Cohen, J. L. & Arato, A. (1992). Civil Society and Political Theory. Cambridge: The MIT Press. Couzens Hoy, D. and McCarthy, T. (1994). Critical Theory. Oxford: Blackwell. Craib, I. (1992). Modern Social Theory: From Parsons to Habermas. New York: Harvester / Wheatsheaf. Dodd, N. (1997). “Review of Between Facts and Norms: Contributions to a Discourse Theory of Law and Democracy” in The British Journal of Sociology, Vol. 48, No. 2 (Jun., 1997), pp. 329-330. Dryzek,  J.  S.  (1994).  Discursive  Democracy:  Politics,  Policy,  and  Political  Science.  Cambridge: Cambridge University Press. Dryzek, J. S. (2002). Deliberative Democracy and Beyond. Liberals, Critics, Contestations. Oxford: Oxford University Press. Elster, J., (ed.). (1998). Deliberative Democracy. Cambridge: Cambridge University Press. Fraser, N. (1985). “What’s Critical about Critical Theory? The Case of Habermas and Gender” in  New German Critique, No. 38 Special Issue on Jürgen Habermas (Spring-Summer, 1985), pp. 97-131. 48 Geuss,  R.  (1981).  The  Idea  of  a  Critical  Theory.  Habermas  and  the  Frankfurt  School.  Cambridge: Cambridge University Press. Giddens, A. (1977). Studies in Social and Political Theory. New York: Basic Books. Giddens, A. (1984). The Constitution of Society. Berkeley and Los Angeles: The University of California Press. Giddens, A. (1987). Social Theory and Modern Sociology. Oxford: Polity Press. Goodin,  R. E., and Niemeyer, S. J.  (2003). “When does deliberation begin? Internal  reflection versus public discussion in deliberative democracy” in Political Studies Vol. 51, No. 4. Gutmann, A. and Thompson, D. (2004).  Why Deliberative Democracy? Princeton: Princeton University Press. Giddens, A. (1987). Social Theory and Modern Sociology. Stanford: Stanford University Press. Habermas, J. (1971). Knowledge and Human Interests. Boston: Beacon Press. Habermas, J. (1979). Communication and the Evolution of Society. Boston: Beacon Press. Habermas, J. (1984). The Theory of Communicative Action, vol. 1. Boston: Beacon Press. Habermas, J. (1987). The Theory of Communicative Action, vol. 2. Boston: Beacon Press. Habermas, J. (1989). Jürgen Habermas on Society and Politics: A Reader. Boston: Beacon Press. Habermas,  J.  (1996).  Between  Facts  and  Norms.  Contributions  to  a  Discourse  Theory  of  Law  and Democracy. Cambridge: The MIT Press. Habermas, J. (1998). The Inclusion of the Other: Studies in Political Theory. Cambridge:MIT Press. Held, D. (2006). Models of Democracy. Stanford: Stanford University Press. Horkheimer, M. and Adorno, T. (2002). Dialectic of Enlightenment. Stanford: Stanford University Press. Kitching, G. N. (1988). Karl Marx and the Philosophy of Praxis. London:Routledge. Lichbach,  M.I.  &  Zuckerman  A.S.  (2006).  Comparative  Politics:  Rationality,  Culture  and  Structure. Cambridge: Cambridge University Press Livesay, J. (1985). “Normative Grounding and Praxis: Habermas, Giddens, and a Contradiction within Critical Theory” in Sociological Theory, Vol. 3, No. 2 (Autumn, 1985), pp. 66-76. Luhmann, N. (1998). “Quod Omnes Tangit: Remarks on Jürgen Habermas’s Legal Theory” in Rosenfeld, M., and Arato, A. (eds),  Habermas on law and democracy: Critical exchanges.  Berkeley and Los Angeles: University of California Press. Mantzavinos, C., North, D. C., and Shariq, S. (2003). “Learning, Institutions and Economic Performance”. Preprints of the Max Planck Institute for Research on Collective Goods, Bonn, 2003/13. Marcuse, H. (1991) One-Dimensional Man: Studies in the Ideology of Advanced Industrial Society. Boston: Beacon Press. May, T. (1996). Situating Social Theory. Buckingham: Open University Press. 49 Nagel, M. (2008). “What if Habermas Went Native?” in Peace Studies Journal, Vol. 1, No. 1 (Fall, 2008), pp. 1-12. Niemeyer, S. J., and Dryzek, J. S. (2007) "The Ends of Deliberation: Metaconsensus and Intersubjective Rationality as Deliberative Ideals" in Swiss Political Science Review, Vol. 13, no. 4, pp. 497–526. North, D. (1990).  Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press. North, D. C., Summerhill, W.,  and Weingast, B. (2000). “Order, Disorder and Economic Change: Latin America  vs.  North  America”  in  Bueno  de  Mesquita,  B.,  and  Root,  H.  (eds.),  Governing  for Prosperity. New Haven: Yale University Press. Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. New York: Cambridge University Press Ostrom, E. (2008). “Review of Social Traps and the Problem of Trust” in Political Psychology, Vol. 29, No. 1 (Feb., 2008). Platt, J. (1973). “Social Traps” in American Psychologist, Vol. 28, pp. 641-651. Prasad, P. (2005).  Crafting Qualitative Research. Working in the Postpositivist Traditions. New York: M. E. Sharpe. Putnam,  R.  (1993).  Making Democracy Work.  Civic  Traditions  in  Modern Italy.  Princeton:  Princeton University Press. Rest,  J.  et  al. (1999).  Postconventional  Moral  Thinking.  A  Neo-Kohlbergian Approach.  Mahwah  and London: Lawrence Erlbaum Associates. Ritzer, G. (2003). The Blackwell Companion to Major Contemporary Social Theorists. London: Blackwell Publishing. Rothstein, B. (2005). Social Traps and the Problem of Trust. Cambridge: Cambridge University Press. Seidman, S. (ed.). (1989). Jürgen Habermas on Politics and Society: A Reader.  Boston: Beacon Press. Tucker, R. C. (ed.). (1978). The Marx-Engels Reader. New York: W. W. Norton & Company. Van Evera, Stephen. (1997) Guide to methods for students of political science. Ithaca: Cornell University Press. Warren, M. E. (1992). “Democratic Theory and Self-Transformation” in The American Political Science Review, 86:8-23. Warren, M. E. (1993). “Can Participatory Democracy Produce Better Selves? Psychological Dimensions of Habermas’s Discursive Model of Democracy” in Political Psychology, Vol.14, No. 2 (Jun. 1993), pp. 209-234. Warren,  M.  E.  (1995).  “The  self  in  discursive  democracy”  in  White,  S.  K.  (ed.),  The  Cambridge Companion to Habermas. Cambridge: Cambridge University Press, pp. 167-200. 50 Warren, M. E. (2000). Democracy and Association. Princeton: Princeton University Press. Willet, C. (2001). Soul of justice : social bonds and racial hubris. Ithaca: Cornell University Press. Wiggershaus,  R.  (1994).  The  Frankfurt  School.  Its  History,  Theories,  and  Political  Significance. Cambridge: The MIT Press. Wren, T.E. (1990).  The Moral Domain. Essays in the Ongoing Discussion between Philosophy and the Social Sciences. Cambridge: The MIT Press. Zanetti,  L.  A.  (1997).  “Advancing  Praxis:  Connecting  Critical  Theory  with  Practice  in  Public Administration” in The American Review of Public Administration, Vol. 27, No. 2 (June, 1997), pp. 145-167. 51

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0067709/manifest

Comment

Related Items