Avhandlingens övergripande syfte är att bidra till förståelsen av hur studenter med svenska som andraspråk möter de krav på språkanvändning som formuleras i lärandemål och i skriftliga exa­mina­tioner.     Avhandlingens forskningsfrågor rör dels vilka språkfunktioner och skribent­roller som efterfrågas i hemtentor under första årets högskoleutbildning, dels vilka språkliga strategier studenter med svenska som andraspråk använder för att möta de krav som ställs. En delfråga är om dessa skiljer sig från de strategier som används av studenter med svenska som första språk.      Materialet utgörs av fyra hemtentor och omfattar dels uppgiftsinstruktioner och tentamens­frågor, dels skriftliga provsvar på dessa frågor skrivna av 19 andraspråksstudenter  och 12 förstaspråksstudenter. Av dessa har 5 andraspråks­studen­ters texter valts ut för näranalys. Analysen av språkfunktioner i både uppgiftsinstruktioner och provsvar relateras till Blooms reviderade taxonomi och de skribentroller som efterfrågas i uppgifterna, liksom de skribent­roller studenterna intar i sina provsvar har undersökts utifrån Hoods identifierade skri­bent­roller (voice roles): iakttagare, utredare och kritiker. I analysen kombineras Hoods skribentroller med Ivaničs uttryck för identitet i skrift; dessa är självbiografiskt jag, diskursivt jag och skribentjag.  Genom att kombinera skri­bentroller och identitetsuttryck kan interimsspråksdrag förklaras med att efterfrågade roller intas i en annan identitet än förväntat, vilket resulterar i ytterligare skribentroller. Utöver Hoods roller identifieras i mitt material även rollerna inlärare, upplevare och debattör. De två första är interimspråksroller som visar att studenten är medveten om dis­kursiva krav men ännu inte har strategier att göra det som förväntas. Debattören är en skri­bentroll som snarare efterfrågas i argumenterande texter på gymnasialnivå än på akademisk. Av analysen framgår att frågeformuleringarna ibland leder studenten fel, eftersom de ännu inte erövrat innebörden av språkfunktioner som att resonera och diskutera inom den akademiska diskursen. Frågeformuleringar tolkas därför alltför bokstavligt.     Analysen visar dock att andraspråksstudenternas språkanvändning är funktionell i den be­mär­kelsen att studenterna oftast möter kraven för betyget godkänd. Där­emot når de inte de högre betygen i samma utsträckning som förstaspråks­stu­­dent­­erna. Skillnaden kan förklaras med graden av genrean­passning och andra­språks­stu­­denter­­nas avsaknad av strategier för att språkligt demonstrera förmågan att inta de roller och tillämpa de språkfunktioner som är förknippad med de hög­re nivåerna i Blooms taxonomi. Skillnaden mellan nybörjarstudenter och studen­t­er med tidigare erfarenhet av högskolestudier är dock större än skillnaden mellan förstaspråksstudenter respektive andraspråksstudenter      Resultaten diskuteras i relation till högskolornas uppdrag om breddad rekrytering och utbildningspolitiska diskurser. / The number of students who use Swedish as a second language for their studies is increasing in Swedish Higher Education. This is a consequence of the fact that more upper secondary school students come from a foreign background and of the task of the universities to work for widening participation in Higher Education. In spite of the increasing number of students using Swedish as a second language there is not much research on the use of L2 in Higher Education or the success of L2-students. One reason can be that the group is difficult to identify since language background is not a factor which is registered in the national statistics of students. L2-students are to be found within the wider group “students with a foreign background”. As a contribution to a better understanding of the use of language of L2-students in a Swedish context, I have looked at how students in their first year in a social science undergraduate programme meet an academic discourse and its demands on academic writing. The study is a text analysis which looks at the use of language in take-home examinations consisting of the instructions and questions of the teachers, the answers of the students and the results. The questions of the study are: What language functions and what voice roles are embedded in questions in take-home examination for first year students on a social science undergraduate programme? This question is answered by analysing the wording of the questions and the criteria for assessment in take-home examinations. The results of this analysis constitute the basis for my second research question: How do first-year students with Swedish as their first or second language adapt their answers to the language functions and voice roles requested?  I analyse the answers written by students who define Swedish as their first or second language, and I discuss the similarities and the differences in their strategies to meet the requirements. In order to study how L2-students meet the requirements of adaptation to the language of the discourse, my third research question is: What voice roles do L2-students use in their examination answers? I answer the question by a deeper analysis of some texts written by students with Swedish as their second language. The theoretical framework is based on sociocultural theory, systemic functional linguistics, Academic Literacies and critical discourse analysis. The main conceptions of the study are language func­tions and voice roles. The language functions are the active verbs required in the instructions for the tasks and questions. The verbs are related to the cognitive levels of Bloom’s taxonomy. The analysis of which voice roles are used and required in the instructions, and used in the answers, draws upon Hood’s (2004) description of voice roles. Following Hood I study the use of the following voice roles: the Observer voice, the Investigator voice and the Critic voice. The latter two are discursive roles and construct the Researcher role.  In the analysis I study the language functions and voice roles in the texts of the students in the light of Ivanič’s (1998) expressions of identity: autobiographical self, discursive self and self as author.  The group consisted of 31 students, 12 L1-students and 19 L2-students and 5 of the L2-students formed a group for close analysis of their answers to two essay questions. The material was collected in the academic year of 2010/2011 and consists of the instructions and questions of four take-home examinations and the answers of the students and their results. I have also analysed documents like the programme syllabus and course syllabi as background material. Two of the four take-home examination were chosen for close analysis. The answers of students to four of the questions in these two take-home examinations formed the basis for the analysis of the language use of the students. The collected texts of the students consist of 96 answers in total, out of which 50 are included in the overall analysis and 10 of these are also included in the close analysis. The studied take-home examinations consist of general instructions for how to answer the exam and also further instructions for each question. The questions are of three different types: short questions, longer questions and essay questions. In the instructions for the take-home examination, there are clear demands for an adaptation to an academic discourse. In these general descriptions, an expectation of all the three voice roles of Hood is included but it is the Researcher voices as Investigator voice and Critic voice, which are explicitly asked for. The three voice roles are also asked for in the different phrasing of the questions but they are not explicitly asked for. Not all questions have to be answered in all the roles. It is the type of question that implies which voice role or voice roles the students are expected to use in their answers. The short questions can be answered in the voice role of the Observer and the longer questions implicate the voice roles of the Observer and the Investigator but to a certain extent also that of the Critic, whereas the essay questions require the students to use all the three voice roles in their answers. This means that the demand for the Researcher voice in the general instruction does not apply to all the different parts of the take-home exam, which can be confusing for first year students. The language functions asked for in the exams represent all the six cognitive levels of Bloom’s revised taxonomy. However, each separate exam does not include all levels. Gradually there is a progression from the lower towards the higher levels. It is also possible to see a connection between the type of questions and language functions and thus a cognitive level. The short questions are used for the lower levels, essay questions are used for all levels. The language functions that occur most frequently and in different types of questions are to explain or describe something, and also to reason about something. Based on the results of the analysis of the instructions and questions, a clear picture appears of what is asked for in the exam and what the students are expected to do to pass the exam. On the basis of these results I analyse the students’ answers, especially the language func­tions: to explain, to define, to refer to, to reason about and to reflect on. An analysis of language functions is made on the answers to one short question, one longer question and one essay question in the first take-home exam and, finally, one essay question in the fourth take-home exam. The questions are chosen because the five language functions which I have identified as particularly relevant are asked for in them. I can see that the answers of the two different groups of students show greater similarities than differences, overall. The answers are more colloquial and personal than academic. However, the texts demonstrate more interim language traits which illustrate that the students are aware of the academic language and strive to acquire the language of the discourse. The results are comparable with the results of studies of L2-students in the undergraduate English-speaking academic discourse (Hyland 2004, Ivanič 1998, Schleppegrell). Not surprisingly, I found that the language functions that are higher in Bloom’s taxonomy are more difficult for the students. Level 4 – analysis – seems to be a critical level for many students but particularly for L2-students. The level of analysis is also the level that Rienecker & Stray Jørgensen regard as the lowest level for a good academic essay (2013: 48). Thus, the level is central to the academic language.  In addition to the three voice roles defined by Hood, the Observer voice, the Investigator voice and the Critic voice, I was also able to identify the voice roles of the debater, the experiencer and the learner. Crucial to the voice roles is whether they are taken by the student’s autobiographical self or discursive self. The experiencer and the learner are interim roles, which lie between the autobiographical and the discursive self and are not explicitly asked for in the instructions. For the student who is not familiar with the discourse, instructions to “reflect”, “reason about” and “give examples of your own” can imply that it is the voice roles as experiencer and learner which are required. The voice roles differ from the required ones by being taken from the students’ autobiographical self rather from their discursive self, as expected. When a voice role is taken by the students’ autobiographical self instead of the expected discursive self, new voice roles emerge and the text becomes less functional. The same effect arises when the wrong strategy, for example the choice of language function, is used to demonstrate a certain voice role. A further observation is that the phrasing of the questions is crucial to which language functions and voice roles appear in the answers of the students. Some phrasings of questions lead the students in the wrong direction. For example, yes/no questions result in answers that resemble argumentative texts on lower levels in the educational system rather than the expository texts that are expected in the academic take-home examinations. The students have not understood the difference between arguing on an upper secondary school level and arguing in an academic context, nor have they understood the academic meaning of reasoning. The differences between L1- and L2-students’ adaptations to the demands of the discourse can be linked to which voice roles they take in the text and how they use their own voices in the texts. The results also show that even if both groups of students use a more colloquial language than expected, the L2-students, as a group, are closer to colloquial language and to the autobiographical self in their use of language than L1-students. I have noticed that the use of language of the students in the form of voice roles and the adaptation to the required language functions are crucial to how well the texts function in their context. This probably influences how the texts are assessed and, thereby, how well the students succeed in their studies. A challenge for teachers is to give both new and experienced students adequate conditions and realistic opportunities to develop their use of language towards the academic target language. However, it is important that such teaching does not focus only on formal demands but goes below the surface of the text and creates an understanding of the reasons for the demands and expectations of the discourse in order to secure a common understanding of what is required in the examinations. The results in this study imply that teaching based on the levels of the taxonomies, expressed as language functions and the voice roles, can be a way to support the students in their acquisition of the academic discourse.   Translated by Anna Maria Staaf Wernheden