"Construction des normes, entre stratégie et communication : un cas de négociation collective", Master's thesis, Systems Science, Ottawa University, 2006

When people argue, one often hears complaints that the other party is “not rational”. What does this mean? What would an ideal rational argument look like? As said elsewhere, I intend to design tools to enhance large-scale consensus processes. After having worked for many years in IT, I know something of tool building; and my mathematics education had given me an idea of what rationality looked like in its purest form. However, human decision making processes are nothing like theorem proofs; I knew I had to gain a better understanding of the multiple forces at play when a decision is taken.

I started this journey by undertaking an interdisciplinary degree, the Masters in Systems Science [] at the University of Ottawa. There, I was fortunate enough to meet Dr. Gilles Paquet [], one of the leading experts on governance, who became my adviser. He understood what I was trying to achieve, and gave me enough of a free hand so I could overcome the limits of my initial, overly rationalist perspective, and explore the social aspects of these issues at greater depth. My thesis was to be based on a case study of decision making in action, through participant observation of a process of collective bargaining in the teacher's assistant union.

My interest was chiefly to explore obstacles to rational decision making, and so my thesis starts with another case study, by Bent Flyvbjerg, of a scuttled town planning reform, where he shows how power dynamics often trump "rational" decisions. But what rationality was he referring to? Again, clearly not the rationality of formal logic; what, then?

For example, economists define (instrumental) rationality as the optimal pursuit of self-interest; this leads to the rational calculations of game theory. Interesting accounts of how power dynamics affect bargaining can be made within this model, and they owe nothing to logic. This model explains many phenomena of micro-economy, but fails experimental verification outside of this field. It can be argued that society as we know it would not exist if that were an accurate theory of how humans behave all the time.

The sociologist Max Weber distinguishes reasoning on ends from reasoning on means to ends which are fixed by social conventions. Many of our actions aim to situate us as members in good standing within a social group; there is therefore such a thing as a norm-conformative rationality, which is the basis of an analysis of society in terms of social roles and norms.

At the other end of the spectrum from the economists, the philosopher Jürgen Habermas defines rationality as a fundamentally collective process; we hardly ever think but in dialogue with other people. He argues that the very existence of language shows that the will to understand one another is a fundamental human drive. He does admit that some sections of society have developed so that the main driving force is instrumental rationality (the market) or norm-conformative rationality (the administrative sphere); but he opposes those "systemic" parts of the world to the ‘lifeworld,’ which is the social sphere where consensus-building and transmission of values occurs. He then analyses both the ideal conditions and pathologies of consensual rationality, which he calls ‘communicative action’.

Habermas’ description of a fundamental drive to consensus sounds quite idealistic, but he raises important issues about the social function of language. An alternative account is offered by political philosopher Joseph Heath, who claims that the main function of language is based in the complexity and diversity of social norms. Society depends on people putting social norms ahead of self-interest in most cases; and this is enforced by sanctions on individuals who choose self-interest ahead of social norms. However, norms cannot codify all actions, and in many cases it is possible to justify breaking a given norm to conform to another norm, in context. Such actions should not be sanctioned, and so application of sanctions must be contingent on the possibility of giving an account of the reasons for an apparent breach. In this account, the main role of rationality is giving reasons for actions; however, when the norms of two groups clash, they have to find new compatible norms; such meta-normative reasoning can involve anything from instrumental rationality (bargaining) to communicative action.

Having explored these multiple definitions of rationality, and a few others, I applied these conceptual tools to my own case study. I had observed, as had Flyvbjerg, many cases of seeming breach of rationality; however, they could not be fully explained either by a power analysis or an analysis in terms of social norms alone, but required a multi-level analysis. Heath's issue of justifiability of actions seemed to play an important role, though that analysis is made more complex by the fact that the actors are aware of playing to many audiences, and justifications have to make sense at many levels and in many spheres. (This follows yet another definition of rationality, that of dramaturgic rationality due to Erving Goffman; it also ties in with what Barbara OKeefe calls rhetoric message design.)

The most important issue is that of consensus: the parties want to agree, so they are seeking consensus, but they also have strategic reasons to maintain a tension. The main points of tension seem to have revolved around identity, which Habermas would classify as belonging to the lifeworld, where consensus-seeking should have been the main driving force. Conversely, some of the most unexpected points of agreement followed purely instrumental reasoning. My interpretation of this observation is twofold: first, the collective agreement is itself a systemic tool (what Habermas calls a steering media), and the resistance of the lifeworld to systemic regulation will naturally focus around lifeworld issues. Second, there is not one, but many lifeworlds, and internal consensus-seeking may overwhelm consensus-seeking between groups.

However, consensus between groups does exist; political theorist John Rawls proposes that pluralistic society evolved from a simple coexistence, or modus vivendi, to creating common norms to which its members may adhere for different reasons, which he calls an overlapping consensus. The evolution of such a consensus follows a process of building trust, which, in our analysis, must be developed simultaneously at instrumental, normative and dramaturgic levels before the ideal case of Habermasian communicative rationality can unfold.

So my thesis presents a much more complex picture of decision-making than the simple interplay between rationality and power proposed by Flyvbjerg. More important for me and my work, it shows that simply creating the "ideal communication situation" does not guarantee consensus. However, though I do not claim to have an answer to the difficult questions around consensus-building, the body of the thesis does point to some of the issues that I believe must be addressed to allow consensus to evolve. Among these issues: the perception that each group has of the motivations of the others (and its own members); the importance of creating contexts where the most uncharitable interpretations can be examined without derailing the process; how regulations affect the transmission of group identity; how building trust depends on multiple channels of communication... How these will find concrete expression in tool design is, of course, still a matter open to research.