1Contrary to what can generally be observed in quantitative studies, for which (apart from a few publications, such as Peneff, 1988 and Caveng, 2012) the field is still usually thought to be a black box whose closure is a condition for the pursuit of scientific work (Latour, 1988), reflexive and critical revisits of ethnographic studies have become established conditions of their validity (Schwartz, 1993; Beaud and Weber, 1997).

2The copious literature produced on this theme has led researchers to question the effects of their presence on the groups they study. Would their practices and their discourses be the same if they were not there? What disruptions do they introduce in the socially interconnected setting that they are studying? Can such disruptions be put to use? Might they serve to reveal and analyze findings that would have otherwise remained in obscurity?

3Some ethnographers have taken a slightly less classic approach that is consistent with the foundational and now established works of Jeanne Favret-Saada (1980[1977]) and Georges Devereux (1967), reversing the perspective by acknowledging that ethnography also inevitably disturbs the person doing it. As François Laplantine wrote, “what the researcher experiences in his relationship with his interlocutors (which he suppresses or sublimates, which he despises or cherishes) is an integral part of his research” (Laplantine, 1996: 23).

4Critical and reflexive revisits of ethnographic studies have accordingly become a sort of requirement, contributing significantly to their rising legitimacy through the 1990s and developing significantly ever since (Cefaï 2010). It has led to an indexation of neutrality and the idea of methodological purity that supposedly accompanies it (Naudier and Simonet, 2011). A consensus has thus emerged that the scholar cannot be abstracted from the research context, and that the subject and object of research are mutually constructed, in that the research relationship is part of the object (Elias, 1983). Despite all that, declarations of intention generally fail to deliver, and it is still far from the norm for the specific conditions of scientific production to be exposed in detail. They are rarely put in the spotlight (Chapoulie, 2000) so as to shelter the work from the criticism that would inevitably result from their revelation (Bourdieu, 2000[1997]) or, more prosaically, due to an inability to consider them objectively. It should also be said that even when there is a manifested desire to divulge research conditions, it can be stifled by the limitations of academic publication formats that grant precious little space to their presentation (Datchary and Jeanjean, 2013). This has been confirmed especially for the relationships researchers maintain with their field sites and their investment (emotional in particular) in their informants, even in cases where the investment is deep and the relationship is a condition for the validity of their empirical work and resulting theoretical developments (Devereux, 1967; Memmi, 1999; Molinier, 2005).

5To put it somewhat melodramatically, researchers lie (to each other), consciously or unconsciously, by omission or otherwise, but with sincerity, about how reality was observed, or in other words in how some elements struck the observer and led her to choose them from the constant flow of social activity, retain them as significant, and put them in her field notes to ultimately pull them out for interpretation during analysis. This also means that they reveal very little about the processes through which otherelements were not selected, thought to be significant, or a fortiori perceived (Caveng and Darbus, 2016), precisely because of how they struck the observer, in another way. We could say that some subjectively marking elements do not make it to the rank of scientifically remarkable elements, due to infra-conscious and infra-linguistic processes.

7In a previous article on research residuals (Caveng and Darbus, 2016) –the inexplicable portion of data that we do not try to interpret, accepted as such by ethnographers and practitioners of quantitative analysis before them in the name of “irreducible empiricism” (Schwartz, 1993)– we offered some explanations of elements that are set aside without justification. The idea was that scholars expunge or do not notice actors, groups, practices, discourses, or relationships that threaten them, in several analytically identifiable ways. We then hypothesized that one fraction of what researchers did not see or deemed residual and unfit for interpretation consists of materials that are “uncomfortable” because they upset the established theoretical frameworks to which they are attached. As a result, such research disruptions generate cognitive, symbolic, institutional, and material discomfort. The residuals are connected to ineffective cases for which the theory does not work (Hamidi, 2012), which we tend to set aside from interpretation. On this point we agree with Daniel Bizeul (2007), who stressed that current practices lead researchers to avoid danger, in risky field sites as well as in the academic field (which is not the least dangerous of places for researchers!). Risky practice in academia could, for instance, take the form of publicly revealing practices that are otherwise condemned by the scientific community, or minimally adhering to the mandatory expression of a given methodological orthodoxy (neutrality, detachment, absence of value judgments, the rational or [worse yet] rationalized selection of informants, and so on).

8Bizeul used the concept of social confrontation to describe situations in which researchers find themselves seriously out of step with their field site and the people they are studying (from the perspective of social properties or ways of doing, thinking, or feeling).

When he frequents a milieu to which he is a stranger, the researcher is in a way put in his place in the social world. He sees the ordinary boundaries of collective life go against his spirit of research. His research endeavor in fact involves its own forms of social confrontation or culture shock to make him understand his externality to the studied milieu. This is one dimension of social confrontation. It is re-enforced by a second dimension due to the fact that the researcher finds himself granted one or more identities according to the mutual perceptions of the various groups present. (Bizeul, 2007: 75)

11But this reintroduction can only make a heuristic contribution so long as it remains rooted in the social sciences, or in other terms, provided that it takes account of the fact that the observer is affected or disturbed to the depth of his social composition. In other words, embarrassment, disgust, rejection, fascination, attraction, and so on should be understood as reactions to circumstances putting them at risk in dispositional terms, such as something that goes violently and deeply counter to the observer’s ways of doing, thinking, seeing, or feeling. Phenomena of concealment, denial, and detachment can then be interpreted as protective mechanisms making it possible for observers to avoid doubting who they are socially. This way, they can continue as a social being, sheltered from a fracture in the relationship of mutual obviousness that each maintains with himself. But the anthropological tradition only demonstrates that knowledge of social worlds comes through a confrontation with foreignness and what it disturbs in us, due to geographical or psychological distance. It is thus important not only to let ourselves be disturbed, but to make use of such emotional and moral disturbances by objectifying their root causes, rather than burying them in the recesses of our consciousness or thinking they are insignificant or irrelevant.

12When undertaken this way, the reflexive approach assumes a methodologically hygienic function that frees us from a vain quest for epistemological purity and the illusion of the researcher as an omniscient thinker who observes situations with remove (Bensa, 2008). This is obviously a complicated exercise, since, as we have already indicated, these phenomena are infra-conscious and infra-linguistic, meaning they occur at the level of the body itself, as exemplified by the sensorial disgust produced by sights or odors (Memmi, Raveneau, Taïeb, 2011, 2016). The epistemological break no longer happens exclusively by detaching the researchers’ common sense from that of the studied people, but also by objectifying and accounting for the causes of sensorial reactions that should first find legitimate expression in ordinary language (e.g. “that’s disgusting,” “I’m shocked,” “they should be ashamed,” etc.) before finding a way to translate it into more analytical language and using it as a tool for knowledge production (“that disgusts me because it goes against who I am socially, and this understanding provides me with access to an essential dimension of my informants’ practices”).

13Other than the articles in this special issue, which we will present shortly, an excellent example of these rationales is provided in Valérie Rolle’s article from an issue of SociologieS devoted to “research failures” (Rolle, 2017). In it she writes about her withdrawal from field research with a group of bikers after rejecting the practices and discourse that dominated within. At first the author identified gender relations as the source of this emotional rejection: it seemed that she was unable to stay in the field because of its cult of virility and a sexual expression that relegated women’s bodies to the rank of object of consumption. But by reflexively reviewing the situation and what lay behind it, she got to the point of shifting the object of this rejection (she speaks of a “return of the repressed”) to class relations. It was not only about gender relations: the author’s own ascending social trajectory out of the working classes had led her to reject the practices of this social milieu, which was also the background of most of her informants. She was not simply rejecting a virile mindset, but its expression in the working class. This is clearly a case of a class-based disgust manifesting itself emotionally and morally in visceral perceptions, therefore at an infra-linguistic level. Objectifying the causes of this disgust thus led the author to enhance her analysis of bikers’ practices, now understood as being class-related as well as gendered.

14The first article in this collection examines how the knowledge that sociologists can produce on the everyday operations of organizations is determined by the conditions of access to those organizations. In particular, author Mathilde Bourrier probes the relationships that sociologists build with actors from various international organizations (focusing on the World Health Organization) and their consequences on the concealment of certain facts and relationships between actors. Re-reading her field notes revealed that she had minimized or not used a significant share of information she had gathered on internal tensions and power games. Such silencing comes in part from a particular rationale concerned with conditions of access and staying in the field, and from a sort of institutional consideration bringing her to handle her own research carefully by carefully handling the organization that took her in. It is also related to the conceptual filters of the sociology of organizations. But leaving the classic mold, Mathilde Bourrier demonstrates that it may above all be her feelings toward people in the organization that determined the concealment of certain elements. Her article shares her fear and embarrassment relative to certain pieces of information (from observations, things people said, etc.) that were likely to upset actors for whom she felt empathy. The rationales tending to make these facts invisible and unspeakable thus respond to a moral requirement to protect actors caught up in organizational power games. In a way, some analytical possibilities were not pursued because the author saw them as potentially betraying people, as if analysis could be a moral failing. Under cover of an objectifying posture holding that the actors’ word should not be taken at face value, and by reducing its reach, the author protected her own moral integrity and in so doing concealed a significant proportion of acts taking place within the organization as well as the acts of the organization.

15The second article is based on a reflexive revisit of an ethnographic experience in a Mapuche organization in Santiago, Chile. In it, Anna Pomaro demonstrates that researchers’ thoughts are subconsciously inhabited by a collection of assumptions and blockages that influence their view of things, and they are often erased from rational analysis. In most cases such elements are eliminated from the final text for being non-scientific, suspect traces of the researcher’s irrational reflexes. She shows that researchers make decisions based not only on transparent reasoning, but also on elements that were unimaginable or repressed in response to especially anxiogenic situations. She thus reviews the main dead ends encountered during her research to unveil the avoidance mechanisms they set off, their effects on the production of data, and the kind of knowledge they can lead to later. She thus addresses the epistemological potential of upsetting things and their consequences for anthropological knowledge, starting with an aspect of her field site that is both uncomfortable and inescapable: reference to tradition. She connects her main methodological choices to situations that were sources of uneasiness and confusion, which in turn triggered avoidance reflexes and the suppression of some potential directions for research. Uneasiness, insecurity, and shock mask problematic aspects of the research experience that merit investigation because they could shed new light on certain positions or lead to novel lines of thought.

17Revisiting processes both practical (the research process, choice of field site and informants) and intellectual (classification, interpretation, and theorization processes) through which facts, events, individuals, and groups were overlooked or were not clearly stated makes it possible to rediscover two collections of “data” with dissimilar properties: one of elements that seem too troublesome or trivial, and another of elements that seem insignificant because they are too singular. In either case, they are characterized by an emotional and moral charge that often goes through affections of the body. As is often asserted without entirely reckoning with its consequences, data is thus never just given. In French, this phrase (les ‘données’ ne sont donc jamais données) is a sort of pun with a troubling connotation arising from the usual terminology, as the word for “data” (“les données) and the word for “given” (“données”) are the same. To remedy this misleading ambiguity, Bruno Latour suggests that data be designated as “obtenues,” “the obtained” (Latour, Miranda, 2015). Perhaps it would be best to call them “captées” (“the caught”), since the rationales for compiling materials go so much further than the mere intention to obtain, and the compilation is so dependent on researchers’ dispositions to receive and perceive. Knowledge of the social world is generated according to the same fundamental processes governing the production of ordinary knowledge, so the difference between them is not found there, but rather in elucidating and taking account of the conditions under which materials were compiled.