The ethnography of ‘particularly sensitive’ activities: How ‘social expectations of ethnography’ may reduce sociological and anthropological scope

Reconsidering a research on ‘ethics in action’ in neonatal intensive care

The research work I wish to reconsider is an instance of what I like to call a sociology of ethics in action. The idea behind the work was to study those aspects of social interaction that cause individuals to ponder the values underlying the activities they witness or take part in: to define what is legitimate and what is not; to position themselves in relation to such definitions; to express differing emotions and points of view; and to take part in the management of all of these. The call to consider ethics as one of sociology’s possible objects came nearly a century ago with sociologists such as Émile Durkheim (1920 [1917]) and Max Weber (1992 [1904–1905]). Since then, much research has been undertaken to study social values, the obstacles to their construction, their evolution, and the variations that may exist between different groups and individuals (Pharo, 2004). There are, however, still only a relatively small number of sociological studies that closely analyze the processes of production and mobilization of values and the relationships that people have with values in real contexts where such processes and relationships are serious issues.

The research published on medical environments during the 1980s and 1990s focused on ethical discourses, on the history of bioethics and on the way medical activity was represented in the media. For years, research seemed to avoid the real-life situations that placed patients and healthcare professionals face-to-face with dilemmas and difficult decisions (Bateman, 1998; Fox, 1989; Weisz, 1990). In fact, it was only halfway through the 1990s that work was effectively undertaken to try to capture in situ the ethical dimensions of medical activity. Methodologically, these works broke with studies of ‘ethical opinions’ using questionnaires or interviews without observation of activity (Crane, 1975; Hoffmaster et al., 1991; Jecker and Berg, 1992; Lally and Barber, 1974; Wershow et al., 1981). They considered that such publications posed as many problems as many opinion polls (Gaxie, 1990). More logically, these new studies focused on fieldwork and, notably, direct observation: they first came from the United States (Anspach, 1993; Bosk, 1979, 1992; Casper, 1998; Fox and Swazey, 1992; Strauss et al., 1985; Timmermans, 1995; Zussman, 1992), where they were labeled ethnography of bioethics (Crigger, 1995; De Vries, 1995); gradually, such research was also produced in France (Baszanger, 1995, 2002; Dodier, 1993; Memmi, 2003; Novaes, 1994).

I, too, decided to adopt a similar approach and to center my attention on the ethnographic investigation of the French world of neonatal intensive care (Paillet, 2007). Neonatal intensive care units are in charge of those newborn babies who, for instance, are under respiratory distress following a premature birth and need to be put on a respirator. A first, exploratory inquiry into the world of neonatal pediatrics allowed me to identify the fact that physicians and nurses within such units are regularly confronted with what they call ‘the hardest’, ‘the most unpleasant’, ‘the most difficult kinds of decisions’ – decisions concerning the continuation or cessation of resuscitation procedures. When the future of a newborn child is uncertain, life-or-death decisions must sometimes be made. Should resuscitation be undertaken when a newborn child is in critical condition and could perhaps develop a severe neurological handicap? Or is it better to cease all resuscitation procedures and allow the child to die? Whichever option is finally chosen, there is the risk of having made the wrong decision. In some cases, the wrong decision is to continue resuscitation, thereby allowing the child to survive, but with severe disabilities. In other cases, the wrong decision is to cease resuscitation and to kill a child who could otherwise have survived in more or less acceptable conditions.

I decided to directly observe true-life situations in an intensive care nursery. For a year, dressed in the same paper scrubs as the physicians and nurses around me, I observed the activity. I was present five or six days or nights per week during a first intensive phase. I then punctually returned, alternating my visits between weekdays or weekends. I posted myself in various locations: hallways, rooms, staff rooms, offices, and meeting rooms. I also tried to diversify my observation tactics. Sometimes, I would focus on particular cases, which I would follow continuously for several hours or several days or even weeks. Other times, I would follow one of the physicians or nurses throughout their day or night shift or attend the staff meetings (visits, tours, transmissions) held daily in their unit. On other occasions, I would share informal moments of gossip, during coffee breaks and meals with various members of staff, or sit in the ‘jar’ (the central room of the unit where all kinds of interactions constantly take place between all sorts of people). The targets of my observation were diverse. I had to identify who took part in the discussions and under which circumstances? How? Which arguments and which norms were used? Who made the final decisions? What were those decisions? Who had to carry them out? What, exactly, were the actions to be carried out? I also observed the life of the unit, from a more distanced angle. I looked at the organization, the division of labor, the working conditions of the different people concerned, the links between unit members, and their relationships with parents and with other hospital staff. And I had to not limit my observation to actions, gestures, postures, and interactions; I needed to take note of discussions taking place in the heat of the action, which are part and parcel of hospital activity.

Beyond direct observation, I also made use of individual discussions and formal interviews with the senior and junior physicians, nurses and psychologists observed. (This was usually done during the period of observation, after several months of presence in the unit.) I also interviewed members of other units and worked on the history of pediatric medicine and neonatal intensive care. I studied the literature concerning decisions of cessation or continuation of resuscitation procedures produced in France since the 1960s by pediatricians, nurses, or psychologists or by legal spheres, bioethicists, or the media, which is today ever-present in the regulation of medical activity.

The principle of hierarchy between ‘preliminary’ and ‘core’ questions

Though not the sole means of my investigation, observation was thus the linchpin of my methodology. Such an approach, as Émile Durkheim wrote when defining the project of an empirical sociology of ‘moral facts’, was consistent with the idea that ‘moral facts’ should not be broached from an essentialist or a normative angle. Durkheim’s idea was to study ‘ethics as it stands’ and not ‘ethics as it should be’; because ‘no matter one’s conception of the moral ideal, morality is an existing reality, open to observation’ (1920 [1917]: 318, 327). Moreover, it is only through observation that conscience issues, raised by real-life situations, can be adequately contextualized. Research addressing ethical issues should not be hatched from the philosophical or moral categories of the thought and language the sociologist interiorized before his fieldwork. Nor should they be derived from the categories that the actors of the studied universe (such as senior physicians and psychologists) tend to impose upon the investigator from the very start of the investigations.

However, though observation did play a prominent role in my work, I did not consider capturing actual decision-making activity as the main purpose of my research. Identifying hospital and decision-making real-life was a necessary first step in my work, but it was not my ultimate objective. Capturing concrete decision-making activity was necessary to answer what I call a preliminary question, whereas I progressively defined what I propose to call my core questions based on the relationships that physicians and nurses had with decision-making processes, their differing moral approaches, the social processes of production of these various approaches, and the management of these divergences.

To understand what I mean by this distinction and hierarchy between preliminary and core questions, let us first consider the type of facts I needed to establish during the so-called preliminary stage. For instance, it was necessary to identify the fact that, in France, most parents are kept in ignorance of the decisions of cessation or continuation of resuscitation procedures; the fact that such decisions are taken by senior physicians, generally without any consultation with younger physicians, nurses or parents; the fact that a decision of cessation does not simply mean stopping the respirator but often signifies the prior injection of death-causing medication; and the fact that these life-or-death decisions frequently give rise to tensions within teams. Of course, establishing such facts did not come from the instantaneous recording and reporting of what was observed: it was not a capture of ‘primitive facts’ (Geertz, 2000 [1973]: 16). My descriptions of these kinds of facts already necessitated an interpretative dimension: they had to be as ‘thick’ (Geertz, 2000 [1973]: 3–10) as any ethnographic description, and more particularly, as any ethnographic description of ‘moral life’ (Kleinman, 2006). In other words, the distinction I propose between preliminary and core questions does not establish an (artificial) distinction between description and interpretation. In fact, my hierarchical distinction between preliminary and core questions was a guide for organizing my research work, from fieldwork to writing and publication. To attribute a different status to these various research questions, considering descriptions of decision-making activity as (‘simple’) preliminary information was a way to rank my various tasks and to plan my heaviest work investments for my core questions, in order to try to go further in my sociology of ethics in action.

Let us consider just one example of these core questions and the work investment they necessitated. Wishing to grasp the relationships physicians and nurses had with everyday decision-making, I considered the tensions that regularly occur within teams when it comes to the cessation or continuation of resuscitation procedures. I came to decode that these tensions revealed disagreement concerning the very definition of legitimate activity. To understand this, I first had to see beyond the psychological discourse that usually presented such tensions as emotional overflow (during staff meetings, psychologists regularly said that tensions about life-or-death decisions corresponded to the symptoms of ‘stress’, or ‘burn out’, or an ‘unavoidable mental fatigue’, or an ‘inevitable psychic suffering’). I also had to see beyond the medical discourse that presented the decision-making activity as fundamentally consensual. During daily interactions, every physician made comments, not only to the ethnographer, but also to the team, such as ‘It’s very consensual here’, ‘Fortunately, we’re unanimous about the principles’, or ‘We always agree on difficult decisions, don’t we?’

To see beyond these discourses and to objectify the various moral opinions of the different members of the team, it was necessary to have already been present on the job for many months and to have followed many decisions. In the first weeks of my fieldwork, my impression was that all the team members equally feared having to cease the resuscitation of a child who could have survived without an important neuro-motor disability or to resuscitate a child who would ultimately live in conditions considered to be ‘catastrophic’. It was only at a later stage of the research that I was able to develop empirical indicators capable of capturing the presence of several hierarchies between these two main risks. It was also only later that I identified the roots of these different ways of prioritizing risks within a professional geography. For example, I paid a growing attention to the chronology of the discussions concerning the most worrisome cases. Nurses were usually the members of the team who first tried to open a discussion by displaying their skepticism in the presence of physicians through their tired bodies, their sad smiles, their long sighs, and pronouncing crucial statements like ‘There will be a difficult decision for this baby.’ The variable usages of the notion of ‘very difficult decision’ were also useful, in establishing the professional variations of risk hierarchies. I gradually noticed that senior physicians used this expression with a usually implicit, but sometimes explicit, normative meaning: life-or-death decisions must be ‘very difficult’. One of them used the most instructive words on this point (during an interview): ‘I think it’s important not to change that. It’s important these decisions stay very difficult to make. We must keep these decisions difficult. Because if they became easy … I don’t know. If they became easy, I would feel very worried. I prefer they stay difficult!’ For most senior physicians, difficulty constitutes a sort of safeguard not to decide cessation treatment and death because of human drives without rational reasoning. When nurses use the expression ‘the very difficult decision’ with an explicit normative meaning, their purpose is very different: ‘Decisions are often too difficult in our team … It’s too difficult to come to [cessation] decisions. They aren’t taken. Or they are too long to be taken.’ All of these empirical indicators, combined with other ones, showed that most senior physicians usually considered a wrongful cessation as the most difficult burden to carry, while most nurses felt that the situation they wanted to see avoided at all costs was that of an improper continuation.

I could give other examples. For instance, I needed a combination of several such empirical indicators to identify the fact that physicians usually tried to consider what would be in the child’s best interest regardless of parental interference, while nurses usually wanted to protect equally the interests of the children and of their parents and considered that parents should have their say in all decisions concerning their children. Another example of the treatment of my core questions was the necessity to invest a very precise effort in the capture of daily and concrete division of labor, which constituted one of the most crucial factors in the production of the professional geography of moral convictions.

What I want to stress here is that compared to my first series of investigation operations linked to my preliminary questions, this second series corresponding to my core questions required the understanding of more subtle processes, of longer chains of mediations, and of principles of variation based on a greater number of dimensions. Even if the boundaries between the analytical processes required by the different types of questions were often rather porous, establishing the existence of the variable moral approaches and determining correlations with the various professional positions required a more complex chain of analytical operations than those linked to processing my preliminary questions. This was also the case when it came to analyzing the social factors that led each particular profession to build its specific moral outlook (an aspect I cannot develop further in this article but which is an important core question in my work). Usually, among sociologists, little is said about research work in quantitative terms. Though the degrees of investments and costs corresponding to our different tasks are varied, these are rarely quantified and compared. In the case of this research, the treatment of my core questions finally called for the more important investments and costs during my research work.

‘Particularly sensitive activities’ and ‘expected descriptions’

Ranking my questions was thus a means for me to remain in line with my overall project and proved useful when it came to making certain choices in the field or during the writing process. However, it was not easy to keep this hierarchy between my research questions. This is one particular point I would like to stress. In fact, it was difficult to contain the descriptions of decision-making activities within the scope of simple preliminary information. Indeed, several factors made the decision-making activities I was studying ‘particularly sensitive’ for society and their descriptions particularly expected. These factors included:

– The important political stakes behind the studied activities.

These factors, which could probably be applied to other research objects, were present in the case of the French world of neonatal intensive care. Indeed, at the time of my field investigations (the end of the 1990s), neonatal intensive care in France was at a delicate moment in its history. The legitimacy of the profession’s decision-making activities was being challenged. The situation was never quite as critical as that of the United States where, since the 1970s, and even more so since 1983 and the case of ‘Baby Doe’, there has been constant media coverage, public debates, politicization, and lawsuits (Anspach, 1993; Guillemin and Holmstrom, 1986; Orfali, 2002). Over the past 20 years, however, French pediatric resuscitators have been the object of increasingly tougher criticism.

Attacks against pediatric resuscitators concern two central aspects of their activity. First, it concerns their policy of systematic initial resuscitations – that is, the fact that resuscitation procedures are applied to all newborn children with respiratory difficulties, regardless primarily of how premature they may be or how little they may weigh at birth. Second, it has to do with their policies concerning parents and the fact that, in the name of what French physicians call ‘psychological protection’, parents are kept away from decisions concerning cessation or continuation and even kept in the dark about many details concerning the existence, the issues, and the way crucial decisions are made. The challenges to the profession come in all forms and from several different directions. One challenge is the recurrent tension within teams themselves. Another concerns the difficult relationships with certain parents holding sufficient social resources to make decision-making issues explicit. Another challenge is the inquiries made by television, radio or newspaper journalists (Favereau, 1996; Nau, 2001). The calls from the French National Consultative Ethics Committee wishing to curb activities at odds with several of its bioethical principles, such as avoiding excessive care or to offer honest and clear information to patients and guardians, are also major concerns (Comité Consultatif National d’Éthique, 1998, 2000).

Moreover, the fact that these decision-making activities are particularly sensitive, and that their descriptions are expected, is linked to their precarious legal status. So far, no French pediatric resuscitator has yet been prosecuted for a life-or-death decision. From a purely legal point of view, however, it is quite clear that when a decision to stop resuscitation leads to the injection of a death-causing medication, it becomes an activity punishable by the same laws as those that punish homicide. Under the French Criminal Code, it is even possible to qualify such activity as first-degree murder with double aggravating circumstances, because the action could be considered to be premeditated, and affecting a minor. As to the unhappy decisions concerning the continuation of resuscitation procedures leading to the survival of heavily disabled children, increasing legal action against physicians in neighboring specialties (obstetrics, ultrasound, prenatal genetic counseling) has brought pediatric resuscitators to fear future lawsuits. Legal disputes concerning the withholding of patient information or the distancing of parents (Cayla and Thomas, 2002; Iacub, 2002) could soon affect neonatal pediatrics and not just its neighboring specialties. Today, French pediatricians in intensive care units are fully aware of the growing number of lawsuits filed since the 1980s against American physicians for injuries related to the birth of a severely disabled child (Some have been sued for wrongful birth by lawyers acting on behalf of stricken parents; others for wrongful life by those acting on behalf of severely disabled children).

In such a context, it is easy to understand why the control of the discourses representing the reality of decision-making activity is a major issue for French teams of neonatal intensive care. Since the beginning of the period of defiance, the profession has published many articles, centered on the description of activity. Since the mid-1980s, in all of the main pediatric and neonatology journals, articles have been devoted to resuscitation decisions or to the ethical issues of resuscitation. By presenting what they say are faithful descriptions of what really goes on in their units, pediatric resuscitators have tried to justify present policies. In the mid-1990s, it was for similar reasons that medical statistical surveys were also published. These were presented by pediatricians who had asked their colleagues either to provide the files stating the causes of death in their units or to fill out questionnaires about their decision-making activities (Gold et al., 1996; Martinot et al., 1995; Willard, 1992). Such studies systematically presented themselves as truthful descriptions of decision-making activity, even putting on a show of open-hearted discussion by unveiling certain illegal dimensions of death decisions.

Ethnographic observation reveals, however, that while claiming to be transparent, these publications carry a discrete but sure censorship. Many allegedly descriptive medical reports appear as watered-down or distorted versions of observed activity, and ultimately, carry only the misleading appearance of being explicit. For example, the many descriptions of the consensus, which is said to prevail within teams, appear far less homogeneous during observation. Or, all of those texts that, regardless of an obvious policy distancing parents and even encouraging staff in the line of duty to lie to them, describe the daily efforts of intensive care personnel to uphold trust – ‘a trust achieved through an open, consistent and clear medical attitude. Parents can admit error, but never lies nor silence’ (Dehan, 1989, p.14).

Until quite recently, the medical control of outside representations of activity has been extremely efficient. Most French lawyers, bioethicists, or journalists who have written about the way that decisions concerning continuation or cessation decisions in neonatal intensive care are made have tamely recycled set medical descriptions. Sometimes not a single word of rhetoric has been altered. Even today, medical control of legal, bioethical, and media descriptions is still strong. But, over the past 15 years or so, the medical monopoly of the representations of such activities has begun to waver. Control has been a little less efficient. A few autonomous discourses and sometimes even so-called counter-descriptions have been forwarded.

Some parents, such as one couple who managed to mobilize a film-maker around the case of their child, are among the first producers of such counter-descriptions. In a documentary meaningfully titled Births and Deaths under Influence (Naissances et morts sous influence), which was first screened in a few cinemas in France before being broadcasted in 1996 by a major TV station (Crèvecœur, 1994), a woman and her husband detail how they immediately spoke to the physicians caring for their very premature child because they ‘knew all about the problem of disability’. As social workers in a center for the disabled, they ‘made it perfectly clear [to the physicians] that [they] did not want the baby to be given any excessive care’: ‘We knew precisely what that meant and it was not something we were ready for.’ In spite of their request, the physicians decided to resuscitate their child, who grew to be heavily disabled and totally dependent. It is thus that such parents speak very angrily of the activity of ‘those who decided in [their] stead and forced [them] to be parents’. Another example of a counter-description of activity is that of a pediatrician who recently spoke in a new way about neonatal intensive care. In his book (Germain, 1999), he calls for public discussion on life-or-death decisions, breaking with the usual medical discourse, which claims that there is no issue for debate anyway. The role of the National Consultative Ethics Committee must also be underlined. Though this instance does not produce its own descriptions of the activity in neonatal intensive care, it has manifested, over the past several years, a growing interest for non-medical descriptions of decision-making activity, such as sociological and anthropological descriptions. These are often quoted, among other bibliographic references, in official publications concerning the ethics of neonatal intensive care.

How expectations concerning activity descriptions may affect sociological and anthropological projects

In order to grasp a major way by which social expectations may influence a sociologist's or anthropologist’s work, all one needs to do is reason in terms of investment and return. Context, for instance, can make evidencing data really difficult, even to just document the so-called preliminary stage of research questions. In the specific context of neonatal intensive care in France, I was forced to provide longer descriptions of certain decision processes and to go into far greater detail than I had originally planned. My work was also burdened by the heavy constraints of anonymity, as it is for every ethnographic work in small or closed communities (Béliard and Eideliman, 2008; Van Den Hoonaard, 2003), but particularly strongly because of serious illegal dimensions of certain practices I observed (Van Maanen, 1982). The general context was such that it was also preferable to increase my own observations and interviews and to compare and combine my results to and with those of other ethnographic studies in similar fields (Anspach, 1993; Frohock, 1986; Garel et al., 1997; Gisquet, 2008; Guillemin and Holmstrom, 1986; Heimer and Staffen, 1998; Levin, 1986; Rostain, 1986; Wiener et al., 1979) or with epidemiological statistics (Larroque et al., 2004; Rebagliato et al., 2000). In other words, the context pushed me to adopt, among my ‘rhetorical strategies’ (Hammersley, 1993), a sort of (quantitative) strategy of data accumulation and diversification, even if fundamentally, my ethnography practice is more linked to a logic of immersion deepening into the studied word rather than a logic of data-hoarding.

This extra work needed to convincingly document what were supposed to be preliminary descriptions might have brought me to spend less time and energy addressing my core questions, through a whole series of ‘investment transfers’. The symbolic value of the sole descriptions of decision-making activity – because of social expectations and the political significance of these descriptions in such a context – could have been enough. Indeed, managing to study activities barely visible to the outside world, and thus touching upon major power issues, can be enough to give meaning to sociological or anthropological work.

Such reflections on investments and returns, or on how tempting it could be to simply concentrate my work on reporting observed activities, were not totally abstract reflections. For example, at several different stages of my work, I was given the opportunity to present my results to various people in various circumstances: to social scientists during academic research seminars, to physicians and nurses undergoing training, and even to bioethics think-tanks bringing together healthcare professionals, philosophers, lawyers specializing in bioethics and associations from outside the medical world. I came to regard such events as being, in themselves, an observable part of my overall research project. I therefore recorded in my field journal the questions I was asked and the reactions or comments brought about by my answers. Most of the time, my descriptions of the difficult decision-making activities took center stage, and sometimes, even the whole stage, regardless of my efforts to try to present decision-making activity reports as the simple preliminary information necessary for deeper discussion on the analysis of the relationships between the various moral convictions and their social origins.

This occurred not only when I spoke to people from the hospital or the bioethics world, but also during sociological research seminars, be they centered on the sociology of health, work, ethics, or even sociological methodology! Whether the particular circumstances were those of medical denial or of academic skepticism, the result was nearly always that I was forced to detail and multiply data in order to overcome resistance concerning the reality of the decision-making activities I reported. I was also often asked to further feed a debate: discussions soon emerged on the legitimacy of current medical activity, and I was expected to provide hard facts concerning decisions.

In short, I was caught in a specific field configuration, in a sort of niche for the sociological production of descriptive assertions concerning activities within a specific environment. During my doctoral thesis, it was easier to resist the temptation to concentrate my sociological work on such issues. From this point of view, a doctoral thesis remains a relatively protected institutional framework. However, the question was inevitably broached anew when my book was being edited (Paillet, 2007). By choosing an editor who gave me free reign to decide for myself the orientation of my work, I think I could avoid making the description of on-the-job decisions the center of my work. I was able to maintain a rather clear hierarchy between those preliminary questions, which enabled the description of decision-making activities, and the core questions about the relationships that individuals have with decision-making activity and on the social origins of moral convictions. Still, the well-identified social expectations did affect the design of a book aimed at an audience (slightly) larger than that of a thesis. This can be read in the very title and subtitle of the book as they were finally published. The title juxtaposes two expressions used in the field by the physicians themselves – To Save Lives, To Give Death (Sauver la vie, donner la mort) – and explicitly describes from the outset what makes neonatal pediatric decisions socially sensitive: the fact that they can cause death. The subtitle, A Sociology of the Ethics in Neonatal Intensive Care (Une sociologie de l'éthique en réanimation néonatale), concedes less to the sensational description of neonatal pediatric activity and remains closer to the general definition of my sociological project. It is also the general context of my research object that explains certain other differences between the thesis and the book, ‘as the guiding notes given to the reader’. Indeed, in order to make very clear the fact that there was more to come after the description of the decision-making activities, I had to consolidate in my book the divide between a ‘descriptive part’ (the first part in the book, reporting actual decision-making processes) and an ‘explanatory part’ (the second part, dealing with the social genesis of moral convictions). In doing so, I offhandedly avoided the whole question of the forms of porosity and continuity between descriptive and explanatory dimensions of sociological or anthropological work.

Within and beyond the sociology and anthropology of bioethics

I think this reconsideration of my past research experience provides a means of seeing from a broader perspective the links that do exist between social expectations and research stances. These links become obvious when a sociological research explores, thanks to ethnographic methodology, certain very sensitive subjects, the descriptions of which are expected by society, but can probably be unveiled in other terrains and in other fields of research. This article is thus an invitation to broaden thought beyond the scope of the decision-making activity I studied and even beyond the study of bioethics.

But let us return for one more moment to the work accomplished in matters of bioethics and to the feeling of unease I expressed at the beginning of this article. I believe that looking back in this way on my research experience provides me with the material to think beyond the questions that were mine, and to reconsider certain studies undertaken on the ethical stakes of medical activity. Of course, to fully examine the reasons behind the affinities of the field with a research focused on descriptions of hospital activities, a socio-history of the sociology and anthropology of bioethics would be necessary. To achieve this would mean analyzing the trajectories of the sociologists and anthropologists concerned, their theoretical and institutional affiliations, of their position in the social sciences field, and so on. But already, we are touching upon what I believe is one of the channels that have contributed to this structural affinity; though the social conditions of work may vary according to individual researchers and to their research projects, certain aspects of the social expectations of ethnography are common to many ethnographic studies on difficult decisions in hospitals.

Finally, it is the very definition of this type of terrain, almost regardless of the scientific project having brought about the inquiry, which locks sociologists within a similar configuration. Because of the political importance of the activities concerned, the efficiency of medical control over social representations of medical activities, the profound changes taking place in the social regulation of these practices (Baszanger et al., 2002), a niche is created for the sole description of decision-making practices.

Of course, I do not wish here to underestimate the value of such descriptions. Their contribution to knowledge often surpasses the descriptions of activity proposed by most professional publications, documentaries or even news programs. It is also important not to underestimate the difficulty in producing such descriptions. The discussion that took place in the United States in the mid-1990s between several ethnographers interested in medicine and bioethics makes this difficulty obvious. At the heart of the discussion, which took place following a series of reviews on several important investigations published at the time, was ‘the unease of the ethnographer’ (Bosk, 1995; Crigger, 1995; De Vries, 1995). This was a discomfort born from the social expectations of ethnography – the risk of a sociology formatted by the interests of one party or another, and the difficulties of having non-normative descriptions recognized. The discussion ended with the invocation of a classic line of the sociology of medicine, namely, one that called to produce a ‘sociology of medicine’ (a sociology taking medicine as its object) rather than a ‘sociology in medicine’ (a sociology within the medical context). In parallel, a similar line was also being drawn in the ethnography of bioethics – that of an ‘ethnography of bioethics’ rather than an ‘ethnography in bioethics’.

However, is it not a pity to thus limit the scope of ethnography? To adapt one’s questions to the characteristics of the studied field, without attempting to maintain, at all costs, preconceived and perhaps too well-defined objectives is one thing (a thing that gives meaning to research, especially ethnographic research); but allowing the social expectations of ethnography to restrict the definition of the task of research is another, which the injunction to conduct an ‘ethnography of’ and not an ‘ethnography in’ does not seem sufficient to resolve.

If this issue has now become an extremely important one in the field of the sociology and anthropology of bioethics, it is because the social expectations mentioned above have today come to explicitly request a detailed ethnography of medical activities. Certain American bioethicists even use the word ‘bioethnography’ to significantly express the ‘specific case studies’ they are asking sociology and anthropology to provide, so that they can adequately ‘think them out’ (Chambliss, 1993; Conrad, 1994; Hoffmaster, 1992; Kaye, 1999; Sankar, 1997). In France, the situation has not yet come to this, but there is reason to believe that a (quasi) market of this kind will open in the years to come. It is therefore essential to initiate, as soon as possible, discussion on the links between social expectations of ethnography and the positioning of sociological and anthropological objects.