Though on the whole theories of care and pragmatism have ignored each other, we think that by assuming that whatever they are, people are vulnerable and interdependent, care joins the concern of pragmatism to redefine subjects and objects by their relationships. If one considers the ability to decide and act mainly depends upon one's attachments to others, care practices aspire not so much to restore autonomy as to improve the quality of attachments. This implies that care cannot be updated through the application of general rules or principles of justice, which risk leading to the imposition of social and legal norms; it is performed in situ, in the attention to the unique needs of others (Gilligan 1982, Tronto 1993, Nussbaum 2000, Paperman & Laugier 2011, Garrau & Le Goff 2012). This invites us to substitute a pragmatist definition of autonomy to an absolute one, only seeing autonomy as the inverse of constraint, especially in the many cases when the necessity of protecting people faces their inability to evaluate what is best for them.

In practice, care work requires extensive arrangements and lots of tinkering (Mol & al. 2010). To avoid both constraint, which violates the principle of autonomy, and laissez-faire, which waives protection and safety, ambivalent behaviours proliferate to get things done: little lies, tact, trickery, misleading actions, ways to say, to say without saying exactly, to manage to be left alone to do without being seen, etc. Moreover, as the action cannot be divided into sequences, but stretches in time and space into a myriad of gestures and acts for which the ethical issues are not immediately perceptible, the principles of the choices made are never completely formulated. Such conduct, which contains an element of duplicity, submits carers to ethical tests. By what right can we go around deceiving people "for their own good"? On what grounds can we make them do something that they have not asked to do, according to a conception of their wellbeing (Callon & Rabeharisoa 2004), or conversely let them take a risk they can realistically neither measure nor assume?

The solution consisting of, on the one hand, keeping the ideal principle of rejection of constraint intact and on the other, leaving the practice to work itself out on a daily basis with the difficulties of its implementation is unsatisfactory. Care theorists have themselves met central concerns of pragmatism when they pointed out that it is in the activity itself that care is deployed; it cannot be reduced to abstract principles. It is here that, to be attentive to the attachments of individuals (Hennion 2010, Hennion & Vidal-Naquet 2015), associating the interrogations raised by the theories of care and a pragmatic approach to the aid relationship is fertile ground: it is both essential not to start from a philosophical statement of ethical principles but from a real investigation of the activity itself, as the social sciences demand; and inversely, on the side of the sociologist, to convince him/her to redefine his/her discipline: s/he must be convinced of the need to reintroduce an ethical interrogation to the analysis by learning to consider the choices and commitments made during the course of action, rather than being content with raising the values and norms of the actors from outside, as if this data were fixed and could be detached from practice and defined from outside by their content.

Leaning on minute case studies carried out at disable people's homes, we aim at describing the forms that such an ordinary, silent constraint, ceaselessly present in care activities may take. To parody Mol & al.'s "care in practice", we deal with "constraint in practice", indeterminate in its nature, uncertain in its effects, diffuse in its application. Far from being settled by a yes or a no, the issue leaves the actors in doubt: it is never possible to be sure to have done something well. In other words, it is the maintenance of trouble surrounding care itself which most respects its at once realistic and normative aims. The word ‘trouble' not only refers to some individual disorder, experienced within carers, although this aspect certainly exists: it can be expressed and discussed collectively, if not institutionalized (Dewey 1927). Far from doing away with the ethical and political burden that bears the question of constraint in care, it is embodied in the very acceptance of this uncertainty, in the concern for the consequences of acts for which it is impossible to assuredly judge the exact extent. We will address such a possible compatibility of constraint with care, and conclude by drawing the contours of a situational ethics: care itself needs care!

Callon, M., & V. Rabeharisoa (2004). Gino's lesson on humanity. Economy and Society, 33(1): 1-27.

Dewey, J. (1927). The Public and its Problems. New York: Holt.

Garrau, M., & A. Le Goff (Dir.) (2012). Politiser le care ? Lormont: Le Bord de l'eau.

Gilligan, C. (1982). In A Different Voice: Psychological Theory and Women's Development. Cambridge MA: Harvard University Press.

Hennion, A. (2010). Vous avez dit attachements ? In Akrich, M. & al. (Éds.). Débordements. Mélanges en l'honneur de Michel Callon. Paris: PEM: 179-90.

Hennion, A., & P. A. Vidal-Naquet (2015). « Enfermer maman ! » Épreuves et arrangements : le care comme éthique de situation. Sciences Sociales et Sociétés, 33/3 (in press).

Mol, A., Moser, I., & J. Pols (2010). Care in Practice. On Tinkering in Clinics, Homes and Farms. Bielefeld: [transcript] Verlag.

Nussbaum, M. (2000). Women and Human Development: The Capabilities Approach. New York: Cambridge University Press.

Paperman, P., & S. Laugier (Dir.) (2011 [2006]). Le Souci des autres, éthique et politique du care. Paris: Éditions de l'EHESS.

Tronto, J.C. (1993). Moral Boundaries: A Political Argument for An Ethic of Care. London: Routledge.