This thesis takes its starting point in an initial questioning of the social work paradoxy: Why implement a form of action that involves an expectation of an active, responsible user, when granting this action requires a reduced ability? Does this mean that some users are not getting the help they actually need?

The initial query is linked to the definition of social work, which includes social work as actual interventions against social problems. In general terms, these interventions could be termed rehabilitative efforts. Over time, these rehabilitative interventions have changed in form, and the general understanding of the term has moved from an understanding of traditional rehabilitation to a modernly termed recovery. Concerning the responsibility of recovery, the change in terminology has brought about a paradigmatic shift - composed by a neo-liberal meta-discourse with an associated accountability discourse - in which the responsibility in today's social work is placed upon the users themselves. Hence, the initial paradoxy is related to this paradigmatic shift. A shift, I continuously problematize during the introductory chapters, based on an assumption that there are users with an actual pressing need for someone else to take responsibility for the situation. In relation to this, I have developed an analytical caring discourse, which brings about a desire to further explore how the two given discourses play out in real life, and thus how social work practice is both constituted and constituent, and how then issues associated with everyday practice can be explained.

In my assessment of the thesis problem, I supplement Fairclough's critical discourse analysis with Smith’s institutional ethnography. I equate the social practice dimension of Fairclough with Smith's trans-local conditions and Fairclough's discursive practice dimension with Smith's ethnographic level. Thereby, gaining the opportunity to examine the relationship between the particular and the general.

My analyses show that hegemony is challenged in a discursive struggle between the discourses of responsibility and care - and that the struggle is associated with ambivalence, which can be regarded as an expression of the dominance of the responsibility discourse. The same is true, when looking at the problems social workers experience in practice - the responsibility discourse is so dominant that its constructive effects overrule the social workers application of the subjectification forms of the care discourse. Thus, my analyses produce an alternative and necessary perspective into social work, and within the context, they advocate that care in social work must be rehabilitated because the downside to recovery is neglect.