It is not yet clear to me why certain topics become part of my research agenda, although I suspect a mix of political and personal interest, serendipity, and social influence is at play. I do have, however, a good sense of what drives me while doing research, meaning what I look for in the data, how I “treat” my subjects, and how I interpret their observed and unobserved behavior. Here I discuss two seemingly contradictory principles: a naïve commitment to the search for the truth, and a strong belief in the generative role of ambiguity and multivocality in the unfolding of social life.

First, I am driven by some sort of naïve commitment to the search for the truth. Not that I believe in its epistemological existence, but I find it useful to act as if there is something “definitive” to discover that, once revealed, other people would be able to see too. This naïve commitment affects the way I do research, in a way that is best captured in one of Italo Calvino’s short stories: The Count of Monte Cristo (Calvino, 1967). In it, both Edmond Dantes and Abbé Faria linger in the dungeons of the prison of Monte Cristo/Château d’If. They both speculate upon the possibility of escaping; Dantes describes their strategies as follows:

Faria proceeds in this way: he becomes aware of a difficulty, he studies a solution, he tries out the solution, encounters a new difficulty, plans a new solution, and so on and on. For him, once all possible errors and unforeseen elements are eliminated, his escape can only be successful: it all lies in planning and carrying out the perfect escape. I set out from the opposite premise: there exists a perfect fortress, from which one cannot escape; escape is possible only if in the planning or building of the fortress some error or oversight was made. While Faria continues taking the fortress apart, sounding out its weak points, I continue putting it together, conjecturing more and more insuperable barriers. The images of the fortress that Faria and I create are becoming more and more different: Faria, beginning with a simple figure, is complicating it extremely to include in it each of the single unforeseen elements he encounters in his path; I setting out from the jumble of these data, see in each isolated obstacle the clue to a system of obstacles, I develop each segment into a regular figure, I fit these figures together as the sides of a solid, polyhedron or hyper-polyhedron, I inscribe these polyhedrons in spheres or hyper-spheres, and so the more I enclose the form of the fortress the more I simplify it, defining it in a numerical relation or in an algebraic formula. (Calvino, 1967, pp. 143–144)

Consider the two strategies. Faria and Dantes look for respectively, the realizable, perfect escape or a perfect, inescapable fortress and they adjust their everyday search practices accordingly. They implement two different approaches to the same goal: breaking out. Two approaches, nonetheless, that lead to different outcomes. While Faria directly searches for a breakout, he generates a more complicated fortress. By pursuing “the multiplicity of possible things” (ibidem, p. 147), he discovers an infinity of possible errors, and thus escaping becomes impossible. In contrast, Dantes searches for the perfect fortress, in which break out is impossible. While Faria dismantles the fortress piece by piece, Dantes constructs more and more impeding barriers. In so doing, he captures the geometric essence of the fortress, and thus simplifies it. Such simplification, far for being a pure analytical exercise, is instead the product of a continuous try out/examination of the real fortress.

But to conceive a fortress in this way I need the Abbé Faria constantly combating landslides of rubble, steel, bolts, sewers, sentry boxes, leaps into nothingness, recesses in the sustaining walls, because the only way to reinforce the imagined fortress is to put the real one continuously to the test. (Calvino 1967, p. 144)

Paradoxically, Faria’s spasmodic attempt at removing all possible errors and unforeseen elements does not lead to an escape, but to a more complicated fortress, while, in contrast, Dantes’ conjecture of the inescapable fortress allows him to find the possible discrepancies between the real fortress and the imaginary one and thus plan his actions accordingly.

And so we go on dealing with the fortress, Faria sounding out the weak points of the wall and coming up against new obstacles, I reflecting on his unsuccessful attempts in order to conjecture new outlines of walls to add to the plan of my fortress-conjecture. If I succeed in mentally constructing a fortress from which it is impossible to escape, this conceived fortress either will be the same as the real one — and in this case it is certain we shall never escape from here, but at least we will achieve the serenity of one who knows he is here because he could be nowhere else — or it will be a fortress from which escape is even more impossible than from here — and this, then, is a sign that here an opportunity of escape exists: we have only to identify the point where the imagined fortress does not coincide with the real one and then find it. (Calvino, 1967, pp. 151–152)

What does breakout stand for? In theory, it may symbolize different things, either social achievements (such as freedom, emancipation, or salvation) or cognitive fulfillments (i.e., knowledge, self-consciousness). In general, we can think of it as the search/call for truth. Indeed, in the Christian tradition there is an intimate relationship between freedom and knowledge:

You will know the truth, and the truth will set you free (John, 8, 32).

And this relation has maintained its prominence during the Enlightenment, even though a rhetoric of salvation was substituted by a rhetoric of emancipation and knowledge (partly) replaced faith.

What Dantes’ and Faria’s story suggests is that beliefs about the possibility of a breakout strongly affect research practices and eventually their outcomes. In other words, the type of empirical evidence one looks for and digs up depends on his/her orientation to knowledge: diverse images of reality will emerge depending on whether one proceeds by making the fortress increasingly more complex, or by simplifying it, capturing its mathematical form.

There is a general tendency in sociology, more than in any other social science, at capturing social reality in its complexity and at uncovering multiple layers of meaning. And this seems a sensible way to proceed, given that the social phenomena we are interested in are usually quite complicated. However, maybe because I am a bit lazy, my instinct is often to tell the simplest story possible. Not because I do not believe that reality is complex — I indeed witness many Farias around me taking the fortress apart — but because I think our only chance at walking away from the fortress is to capture its analytical essence and thus reduce its complexity. My reliance on analytical categories, (simple) formal models, behavioral games, and field experiments can be understood in this framework. Sketching the bare bones of a formal model (Manzo & Baldassarri, 2014; Baldassarri & Bearman, 2007), studying human interdependence within the minimalist constraints of a behavioral game (Baldassarri, 2015), as well as embedding unobtrusive field experiments in everyday life (Baldassarri & Abascal, 2017) are all ways in which I try to capture geometric complexity through simplification. What many would see as arid, ahistorical, empty of context and meaning, is for me the final test in which the real fortress gets compared to the hypothetical one — the one from which it is impossible to escape —, and only if there is a difference, then a breakout is possible.

This predilection for analytical simplicity however combines with a second, almost orthogonal tendency to consider the generative role of ambiguity and multiplexity in shaping social relationships.

To give you an example of what I mean by the positive role of ambiguity, let's consider the (mostly American) concept of going on a “date”. To me, this was a cultural shock, because in the culture I come from there is no equivalent attempt at predefining the content of a relationship of this kind, the first solo-encounter between two individuals. Or, at least, there is no explicit attempt at drawing a line between romantic and non-romantic relationships. This is an event defined by its potential, in which fluidity is maximal, and defining ex ante the boundaries of the relationship does inexorably mean to constrain it, and therefore limit its possibilities. To label it a “date” kills the romance! A romantic relationship unfolds from the ambiguity that underlines courtship, the multiple and sometimes contrasting signals, i.e. dressing up, but remaining casual, choosing a romantic restaurant but avoiding the candlelight cliché, showing interest, but not falling in love, etc.

Of course a romantic relationship is a very personal, peculiar event, but the same positive role of ambiguity can be seen in many other patterns of relations, involving individuals as well as institutions. In general terms, we should think of ambiguity as a property of the situation, rather than of the actors themselves. Ambiguity can generate from many factors, including the fact that actors have imperfect information, multiple and even contrasting desires, an inadequate or limited understanding of the situation, of their own means and of other people’s intentions, sometimes even scarce control over their own will, thus limited control over their course of action.

I’m here less concerned with the ambiguity that derives from imperfect information, misperceptions, cognitive biases, and other individual limitations. These have already been object of extended investigation in the social sciences. What I want to highlight is the relational dimension of ambiguity, the fact that the meaning of social action is contingent on how it is perceived by other individuals, and how it cumulates into a sequence of actions. Moreover, social actions might be multivocal, sometimes even deliberately oriented at providing noisy signals.

I have the sense that in the social sciences we tend to think of ambiguity as a limitation, either as a distortion from the model of perfect information (and rational action) or a situation in which individuals and organizations are under stress, in which actors experience multiple and counteracting tendencies, a potential source of misunderstanding, wrong calculation, or a possible cause of unattended consequences.

In contrast, I tend to think about the ambiguity that permeates social actions and interactions as a necessary precondition for the unfolding of social relations, a buffer space that allows shifts in roles, fluid understanding, partial disclosure, and might contribute to preventing direct confrontation, and conflict.¹

Ambiguity conceived in these terms is complementary to the concept of social role, it captures what is left out, what we miss when we constrain individuals to categories, expected behaviors, and obligations that are defined by their socio-demographic profile and status. Ambiguity should not be conceived as an alternative to social roles, but instead as the fertile terrain that provides the preconditions for roles to be defined, negotiated and, more importantly, for contrasting roles (and identities) to be carried out simultaneously. The roots of this complementarity can be easily found in Merton’s role-set theory (Merton, 1957).

So, what does this attention to ambiguity add to our research? The most important thing is that it prevents the researcher from the risk of role reification, in which actors are stuck, crystallized into a set of categories, labeled as working-class, Evangelicals, mothers, Latinos, Southerners, Democrats, and then expected to follow courses of action that are coherent with the labels we give them.

Here’s three pieces of scholarship that well exemplify how ambiguity and multivocality can have a generative role in the explanation of social action. Eric Leifer, in “Interaction Preludes to Role Settings: Exploratory Local Action” (1988), argues that in settings in which roles are not “given”, in which roles are not yet established, actors have to strategically interact to acquire desirable roles. During the interaction that precedes the establishment of roles and social hierarchy, individuals should avoid claiming a coveted role until there is evidence that it will be conferred. Otherwise they will reveal their intention and desires, thus giving an advantage to other people competing for that role. In general, explicit status claims are dangerous, because people are exposed to the possibility that the status won't be conferred: “The same actions that confer status can also take it away, depending on the responses they elicit.” As an alternative, Leifer discusses a distinct action ideal, called “local action”, that is intended to suppress role differentiation. It is a non role-specific action, “that, ex ante, leaves open a range of roles, and ex-post, does not prove inconsistent with any role that might be claimed later” (ibidem, p. 868).² The relative status of individuals remains fluid until ambiguity failures occur and roles emerge from local interactions. In these terms, interaction preludes role setting. Roger Gould built his theory of status hierarchies on similar principles.

In addition to being a strategic tool that individuals deploy in the competition to achieve desirable roles, however, ambiguity is also a mean for individuals to juggle their different, and sometimes alternative identities. In his Insurgent Identities (1995), a study of the Parisians protests of the Nineteenth Century, Gould asks how the collective identity of workers emerges as such and becomes a mobilizing identity. Instead of categorizing individuals as workers, he considered the multiple identities that were available to the actors, from their gild affiliation to the neighborhood where they lived, and shows how patterns of social interactions — the web of interpersonal relationship in which people are embedded — are fundamental in eliciting certain identities and neglecting others, thus producing an alignment of interest and identity. Class in se becomes class per se: the initial ambiguity of multiple potential identities is resolved through social interaction.

Finally, Padgett and Ansell in their “Robust Action and the Rise of the Medici” (1993) describe the birth of the Renaissance state in Florence deploying Leifer's concept of local action. According to them, Medici’s political power emerged from their capacity to span across the network disjunctures within the élite. The Medici's party was a “mixture of contradictory interests and crosscutting networks” (ibidem, p. 1262). At the basis of Cosimo de Medici's action there is the notion of multivocal identity, where multivocality is intended as “the fact that single actions can be interpreted coherently from multiple perspectives simultaneously” (ibidem, p. 1263). It is through this multivocal action that interest remains opaque, hidden. In more general terms, they conclude: “Ambiguity and heterogeneity, not planning and self-interest are the raw materials of which powerful states and persons are constituted” (ibidem, p. 1259).

As it should be clear from this last example, this perspective suggests a theory of action that is alternative not only to classical rational choice, but also to many means-goal explanations. The main difference between Leifer's theory of local action and these other theories of action, however, doesn’t have to do with the rationality of the actors or their level of agency. In fact, as it is easy to deduce from the cases described before — namely the strategic ambiguity that preludes role settings, the emergence of collective identities through the alignment of identity and interest, and the robust action of Cosimo de Medici — entail different conceptions of agency and are not necessarily orthogonal to rational action.

The real difference between rational action theory and a theory of local action has to do with the focus of observation, namely rational action theory is centered on the individuals and their motivations, while local action starts from relationships. Consistently, scholars working on this vein have found in social network analysis a very powerful analytical instrument. In fact, acknowledging the primacy of relationships is a way to maintain the composite and contradictory nature of individuals, and therefore keep alive the source of their ambiguity.

Some readers might wonder whether this focus on ambiguity and multivocality is simply another way of saying that individuals are complex creatures, and thus cannot be categorized and quantified. It is not. What I have in mind is instead the basic dualism that, according to George Simmel,

[…] pervades the fundamental form of all sociation. The dualism consists in the fact that a relation, which is a fluctuating, constantly developing life-process, nevertheless receives a relatively stable external form. (Simmel, 1971, p. 351)

Simmel builds on the analytic distinction, and dialectic, between forms and life-process: according to him, the stream of life, its potentiality and energy, in order to realize itself, inevitably has to be constrained by forms:

our inner life, which we perceive as a stream […] becomes crystallized, even for ourselves, in formulas and fixed directions often merely by the fact that we verbalize this life. Even if this leads only rarely to specific inadequacies […] there still remains the fundamental, formal contrast between the essential flux and movement of the subjective psychic life and the limitations of its forms. These forms, after all, do not express or shape an ideal, a contrast with life’s reality, but this life itself. (Simmel, 1971, p. 352)

Thus, forms are the realization of the life-process, their crystallization. Life-process is always more than forms, but can only express itself by becoming forms. Focusing on the network of social relations in which individual and institutions are embedded is a powerful way to capture these forms. We need, however, to remember the enduring discrepancy between life contents and their objectification in relational forms, and keep alive the tension between individual possibilities and their realization. Only this way we would avoid the risk of form and network reification. Some hard core structuralists have run into this problem in the past. Now, the same problem is visible in most research based on large-scale networks, in which big data comes at the expense of information about the nodes and the nature of their relationships.

To avoid social structural reification, we should recognize individual possibilities and agency, as well as the historical and social constraints actors face, and the ways in which they overcome them. Let us consider a simple example: the episode from the Odyssey of Ulysses and the sirens. As the story goes, Ulysses is traveling back home and has to pass by the sirens’ island. Sirens have the power of charming by their song everyone who heard them, so that mariners were impelled to cast themselves into the sea to destruction. Ulysses needs to pass by them, but simultaneously wants to listen to the sirens voice. His solution is “to stop the ears of his seamen with wax, so that they should not hear the strain; he had himself bound to the mast, and instructed his people, whatever he might say or do, by no means to release him till they have passed the Sirens’ island.”

Now, many of you might be familiar with the use that Jon Elster does of this episode (1979). Elster treats Ulysses as a rational decision maker and the episode is an example of pre-commitment: a decision is taken at time 0 in order to restrict, to bind the decisions one can make at time 1. However, this episode has attracted the attention of many scholars dealing with the problematic of self and order, from Dante to Joyce. Differently from Elster, Ulysses and the sirens has been used to reflect on the meaning of boundary crossing, overcoming of one’s limitation and the pursue of knowledge.

For instance, Max Horkheimer and Theodor Adorno, in their Dialectic of Enlightenment (1944) use the same episode, but put Ulysses behavior in relation to the historical context and broader social order in which he is acting. First of all, Ulysses in their reading has multiple and contrasting goals: he wants to get home, but also wants to satisfy his own curiosity, thus listen to the sirens. Moreover, he has a specific position in the social structure, he has power over his subordinates and exercises it, in order to prevent potential weaknesses of his own will, and therefore maintain his social status. The entire journey, and this episode in particular, can be read as an allegory of modernity, and the growing control that the bourgeois individual has over myth.

In sum, while through Elster’s lens we are presented a mythology of individual efficacy, Adorno and Horkheimer's dialectic reveals the historical, relational, and contextual preconditions that enable Ulysses to cross boundaries, to overcome his own limitations in a new and creative way. This is not by chance: a dialectic method is, first and foremost, the rejection of all reifications, and therefore more likely to capture new realizations of life-process.

Finally, this episode is deployed by Adorno and Horkheimer in the broader context of their work to show how positivist and scientistic models of science turn to myth “by failing to reflect on their own practices” (1944). This leads me to a final question: if we agree that ambiguity and multivocality have a positive role in determining the unfolding of social life, should they have a role in our research activity?

I believe that keeping alive the sources of contradiction in our research practices can be potentially useful. The important thing is to be aware of them, although it might be sometimes painful. And it is in this spirit that I decided to title this little exercise in self-reflection “Keeping One's Distance”, after a beautiful two-page fragment by Theodor Adorno in Minima Moralia in which he shows how “distance and self-criticism is not a safety-zone but a field of tension. It is manifested not in relating the claim of ideas to truth, but in delicacy and fragility of thinking” (Adorno, 1951, p. 127). In other words, in the “concrete awareness of the conditionality of human knowledge” (ibidem, p. 128). If there is something that I should reveal about my way of doing research, is this intellectual tension.