! "#    $%"&  "'(%% 
) * + ,    +    - "$(

 
.       /          /     0     1 
  2         /      3!4(       
            ( .            
                    2 ( .     
   ( /         +         +  
 +                     ( 5
                2  !     +     
      /  !   /     2    1      
2 (
/        +            
                 (    / 
    6       /    2      +    
   ( 7                     2 /
 /             +          / + 
 / +      /           1 +      
              ( .      0     1 / 
 2      0   1  0   1(
               + /  6  +      (
5     /          8   /         
        ( .   /             
                 /  1 / ( 7+ 
   9        2  (         / 
    +                    (
      /  !   1  2           /
    0     1( +    1        / 
               +       
     ( :               /     
     +           9 ;   (         + +
 0     1 /       +         <   (
.  +     +  +            
   (    +           /         
  <         ( .    /            
         /                  
      /    ( !  +   2       
            /     $  1 ;  
(

                  
                    
    

  $%"&
 8== ( ( = > ? 8 8 8 8;"@&#%$

) AB&;A";BBAB;'#C;"
) AB&;A";BBAB;'#B;&
) %@C$;"%AB

     

   
    + "%C A"  
CONSTRUCTIONS OF LANGUAGE COMPETENCE
Maria Rydell
Constructions of Language
Competence
Sociolinguistic Perspectives on Assessing Second Language Interactions
in Basic Adult Education

Maria Rydell
©Maria Rydell, Stockholm University 2018

ISBN print 978-91-7797-346-1

ISBN PDF 978-91-7797-347-8
ISSN 0562-1097

Printed in Sweden by Universitetsservice US-AB, Stockholm 2018

Distributor: Department of Swedish Language and Multilingualism, Stockholm University
To Frank, Elis and Siri.
Acknowledgements

Writing a thesis on perceptions and constructions of language competence

from the perspective of second language speakers has made me think about
my own journey of learning how to write and speak English in academia. This
has been a rewarding, but at times, difficult process. The words are not always
in my control. As a language user, you expose yourself to others and this is
the only section in the thesis that has not been proofread. The words, and their
flaws, are thus my own, and I trust my readers to engage constructively in
understanding me. For who, in the end, would we be without others?
This thesis would not have been possible without the support of (from?)
my colleagues. First and foremost, I would like to thank my main supervisor
Catrin Norrby. Ever since my first day as a Phd-student, Catrin has engaged
with my project and with my ideas with enthusiasm. She has been a quick and
a critical reader. My supervisors Gunlög Sundberg and Caroline Kerfoot have
read numerous drafts and provided insightful and constructive feedback.
Throughout these years, Catrin, Caroline and Gunlög have showed an unwa-
vering belief in my capacity as a researcher, for which I am deeply grateful.
I have been lucky to have Linnea Hanell as my travel companion through-
out this project. Thank you Linnea for being a critical reader and for stimulat-
ing discussions about what research is and ought to be. Likewise, David Kar-
lander and Linus Salö, thank you for reading and discussing numerous drafts
and for encouragement.
I am profoundly indebted to teachers and sfi students who generously par-
ticipated and made the thesis possible and to former colleagues at sfi-provet,
in particular Katrin Ahlgren and Sofia Engman. I would also like to express
my gratitude to Numa Markee, who generously shared his knowledge and
templates in his course Multimodal aspects of interaction as a visiting profes-
sor in 2016 and to Nigel Musk, who provided constructive feedback reviewing
the manuscript. I also wish to thank Johan Asplund for technical support and
Cecilia Falk for editing support.
At my home department, I have had the opportunity to discuss my work at
different seminars such as doktorandkollokviet, högre seminariet, inter-
aktionsseminariet och språkbrukskollokviet. These seminars have been im-
portant meeting-points. Moreover, I would like to thank colleagues who have
discussed my work at conferences, seminars and at the lunch table. In partic-
ular, I wish to thank Hannah Botsis, Mona Blåsjö, Asta Cekaite, Jessica Doug-
lah, Natalia Ganuza, Marjorie Goodwin, Christina Hedman, Kenneth Hylten-
stam, Gunilla Jansson, Martha Karrebæk, Kathrin Kaufold, Silvia Kunitz,
Inger Lindberg, Ali Reza Majlesi, Tim McNamara, Guillermo Montero Melis,
Gunnar Norrman, Lina Nyroos, Stefan Norrthon, Marcelyn Oostendorp, Josep
Soler, Elin Svahn, Christopher Stroud and Henning Årman for encouragement
and for engaging with my ideas.
Finally, to my family, thank you for love, support and for putting up with
me during the final stage of the thesis. To my parents and to my brother, thank
you for always believing in me and being there for me. To Niklas, Frank, Elis
and Siri: thank you for being in this world with me.

Stockholm August 2018

Maria Rydell
 The present thesis is based on the following studies:

I. Rydell, M. 2015. Performance and ideology in paired speaking

tests for adult migrants. Journal of Sociolinguistics 19(4). Pp. 535–
558.

II. Rydell, M. Negotiating co-participation: Embodied word search-

ing sequences in paired L2 speaking tests Submitted manuscript

III. Rydell, M. 2018. Being ‘a competent language user ‘in a world of

Others – Adult migrants’ perceptions and constructions of commu-
nicative competence. Linguistics and Education 45. Pp. 101–109.
doi.org/10.1016/j.linged.2018.04.004

Reprints have been made with permission from Wiley and Elsevier, respec-
tively.
Contents

1 Introduction ............................................................................................1
1.1 Language testing and assessment from a sociolinguistic perspective ................1
1.2 Aim and short summary of the studies................................................................4
1.3 Organization .......................................................................................................6

2 Language education for adult migrants in Sweden – the case of SFI ...7
2.1 Migration to Sweden and migration policies .......................................................8
2.2 The development of Swedish for Immigrants (SFI) .............................................9
2.2.1 SFI from 1965 to 2018: a short history ...................................................10
2.2.2 Current organization, syllabus and participants enrolled in SFI..............14
2.2.3 The national test in SFI ..........................................................................15
2.3 Intersecting discourses on adult migrants language competence.....................17
2.3.1 The duty to learn....................................................................................19
2.4 Previous research related to SFI ......................................................................21
2.5 Concluding remarks .........................................................................................23

3 Constructions and perceptions of L2 competence ..............................24

3.1 The development of the notion communicative competence ............................25
3.1.1 Constructions of language competence in SFI: a communicative and
functional approach ...............................................................................................29
3.1.2 Critiques of communicative competence and its use .............................30
3.2 Being ‘a competent language user’ and ideologies of ‘good language’ .............32
3.3 Adult language learners’ perception of language competence .........................34
3.4 Concluding remarks .........................................................................................35

4 The testing and assessment of speaking ............................................37

4.1 The sociopolitical context of language testing and assessment ........................37
4.2 The development of speaking tests ..................................................................40
4.2.1 Paired speaking tests ............................................................................41
4.2.2 Issues in testing and assessing speaking ..............................................42
4.2.3 The speaking test as a speech event.....................................................45
4.3 Concluding remarks .........................................................................................48

5 Analysing spoken discourse: data presentation, methodological

considerations and theoretical outlines.........................................................49
5.1 Overview of data ..............................................................................................49
5.1.1 The paired speaking tests......................................................................50
 5.1.2 The post-test interviews .........................................................................54
5.1.3 The focus group discussions .................................................................54
5.2 Analysing spoken discourse and communicative experiences: methodological
and theoretical outlines ...............................................................................................57
5.2.1 Transcribing and translating L2 spoken discourse .................................59
5.2.2 Notes on reflexivity ................................................................................61

6 Summary of Studies.............................................................................63
6.1 Study I – Performance and ideology in speaking tests for adult migrants .........63
6.2 Study II – Negotiating co-participation: Embodied word searching sequences in
paired L2 speaking tests .............................................................................................66
6.3 Study III – Being ‘a competent language user’ in a world of Others – adult
migrants’ perceptions and constructions of communicative competence ....................68

7 Concluding discussion .........................................................................72

8 Sammanfattning...................................................................................78

9 References...........................................................................................82
1 Introduction

The present thesis is concerned with constructions and perceptions of lan-

guage competence among adult migrants in Sweden enrolled in a language
programme called Swedish for Immigrants (SFI).1 Learning a new language
sheds light on the assumptions made on the ‘nature’ of language (Harris
1981:77–80, Heller 2007:9). A particular practice in which constructions and
perceptions of competence are brought to the fore is language testing and as-
sessments (both institutionalized assessments and those made in everyday
life). The prevailing communicative approach to language teaching and test-
ing makes it relevant to investigate both language use and reflections on com-
municative experiences. In this light, the present thesis sets out to investigate
sociolinguistic perspectives on language testing and assessment with a partic-
ular focus on oral interaction. More specifically, this is done, on the one hand,
by investigating paired speaking tests (a candidate-candidate discussion) in
the final national test in SFI2 as a speech event and, on the other hand, by
investigating SFI students’ perceptions and constructions of what it means to
be ‘a competent language user’ as manifested in paired speaking tests and fo-
cus group discussions.

1.1 Language testing and assessment from a

sociolinguistic perspective
Studying language testing from a sociolinguistic perspective is important for
a number of reasons. Testing practices shape and confirm language learner
identities, and, by passing tests, candidates can claim legitimacy as ‘competent
language users’ (Jørgensen et al. 2011). With respect to adult migrant lan-
guage learners, passing language tests has an increased symbolic value since
such tests often perform a gatekeeping function in relation to other arenas such

1
The official name of the language program has changed over time. The current official name
is Municipal adult education in Swedish for immigrants (kommunal vuxenutbildning i svenska
för invandrare). However, the commonly used term by both students and teachers, as well as in
research, is SFI (e.g. Lindberg & Sandwall 2017, Rosén 2013). Henceforth, the abbreviation
SFI will be used.
2
See section 5.1.1 for a presentation of the paired speaking test.

1
as the labour market (Carlson 2002). As a core professional activity for teach-
ers, assessment practices set out to make claims of the language abilities and
communicative competence of individuals. Conversely, test takers need to in-
terpret the task at hand and act accordingly in order to display themselves as
‘competent language users’. In this sense, testing and assessments are prime
sites for the construction and perception of knowledge (e.g. Foucault 1977,
McNamara 2001, Shohamy 2001). In this light, testing oral interaction sheds
light on tensions relating to assumptions made on the ‘nature’ of language
competence, such as being an individual ability versus a relational construct
(Chalhoub-Deville 2003, McNamara 1997, McNamara & Roever 2006; see
also Studies II and III). Another issue that is brought to the fore in oral testing
is the view of language as an embodied practice in relation to the logocentric
view or the lingual bias prevailing in linguistics and in language testing (e.g.
Block 2014, Goodwin 2004, Heughs 2004, Studies II and III). The ‘lingual
bias’ is defined as ‘the tendency to conceive of communicative practices ex-
clusively in terms of the linguistic (morphology, syntax, phonology, lexis)’,
thus neglecting embodied and multimodal aspects of language use (Block
2014:56). Moreover, this study addresses how a monolingual ideology of com-
petence is constructed and reproduced in language tests (e.g. Jenkins & Leung
2017, Shohamy 2011, Studies I and II) and how ideas of ‘being a competent
language user’ are intertwined with such ideas as being ‘a good immigrant’
and being ‘a good student’ by taking responsibility for one’s learning (Study
I).
In the wake of globalization, the symbolic value of adult migrants’ language
competence has become increasingly salient. The importance placed on mi-
grants becoming competent in Swedish concerns a contemporary ‘language
ideological consensus’ (Salö 2016:7). A language ideology can be defined as
a ‘cultural system of ideas about social and linguistic relationships, together
with their loading of moral and political interests’ (Irvine 1989:255) that cre-
ate a ‘linkage of microcultural worlds of language and discourse to macroso-
cial forces’ (Kroskrity 2000:2). Hence, language ideologies are never about
language alone but should be understood in relation broader cultural dis-
courses. As underscored by Miller, ‘[i]deologies are not floating in the
“macro” world as intangible social beliefs and values but are locally generated
and (re)constituted in “micro” interactions’ (2014:21). From this perspective,
the situated interactions studied in the present thesis provide an emic perspec-
tive with respect to language ideologies of what it means to be ‘a competent
language user’ and how ideologies of language competence are reproduced in
educational and institutional discourse. Hence, studying perceptions and con-
structions of language competence provides a window into contemporary lan-
guage ideologies of what it means to be ‘a competent language user’.

2
 What is more, language competence serves as a feature for distinguishing
between people (Stroud 2004),3 and becoming competent in the receiving
country’s language indexes belonging to the new country of residence (Black-
ledge 2005, 2009, Milani 2007). Learning the official language is often per-
ceived as a key to accessing the mainstream society and the labour market as
well as forging a future identity that belongs to part of an imagined community
(Carlson 2002, Fejes & Dahlstedt 2018, Norton B. 2013). Conversely, mi-
grants’ (perceived) lack of proficiency is often positioned as a threat to social
cohesion, associated with unemployment and sometimes even constructed as
a threat to national security (Khan 2014). From this point of view, adult mi-
grants’ language competence is given a symbolic value (e.g. Bourdieu 1991)
that is taken as an index for non-linguistic tokens such as commitment to the
new country of residence and of making an effort (cf. Allan 2013, Blackledge
2005, 2009, Milani 2007, Stroud 2004), giving adult migrants’ language com-
petence a political and moral valuation.
In the current thesis, a particular focus is given to the notion of communi-
cative competence since it provides an important framework and rationale for
communicative language testing. The advent of the communicative teaching
paradigm in the 1970s put an emphasis on oral language and performance test-
ing, where test takers are expected to produce language in an act of commu-
nication (McNamara 1996, Savignon 1983). Paired speaking tests provide a
case in point of communicative language testing and serve as a commonly
used format in both language classrooms and high and low stakes tests around
the world (Lazaraton 2014).4 The focus of oral skills have not only been nota-
ble in language teaching. Rather, living in ‘communication culture’ in global-
ized societies, ‘[c]ommunication becomes not just something workers are re-
quired to do, but something they are expected to be, or to become good at’
(Cameron 2002:73), which is of concerns to both L1 and L2 speakers.
The term ‘competence’ stems from a cognitive tradition, but has also been
used in more practice-based approaches such as in communicative competence
(e.g. Hymes 1972a), interactional competence (Kramsch 1986, Young & He
1998) or in performative competence (Canagarajah 2013). How knowledge
should be defined is rarely discussed with students (Carlson 2002, Zachrison
2014), and students’ perceptions and subjective experience of language have
been described as ‘neglected dimension in language learning’ (Kramsch
2009:27). Given the emphasis on language use in communicative approaches
to language teaching and testing, the relevance of investigating both language
use as well as reflections upon communicative experiences become apparent.
The current thesis does not set out to launch a new term or model of compe-
tence. Instead, my interest lies in the conceptual image of being ‘a competent

3
As noted by Rosén (2013:21-22), the name of the language programme itself contains a dis-
tinction and categorization of ’immigrants’.
4
A test is considered high-stakes ‘when a single test score is used as the main or sole factor in
determining significant educational decisions’ (Menken 2017:386).

3
language user’, and how SFI participants orient to this image interactionally
and metapragmatically in test discourse and in focus group discussions. Rather
than treating language competence as a fixed and absolute concept, I treat
competence as being shaped during intersubjective processes, and as a concept
learners often positon themselves in relation to and by which they are posi-
tioned by others (Jaffe 2013, Kataoka et al. 2013). From this vantage point,
language competence is treated in a broad sense as ‘what it means to know a
language well’ (cf. Park 2010).
Furthermore, as Park maintains, ‘competence is used as an extremely per-
vasive but frequently overlooked index for one’s identity and social status’
(2010:23). Given the symbolic value assigned to language competence, it is
pertinent to pause and scrutinize perceptions and constructions of competence
and how adult migrants perceive what it means to be ‘a competent language
user’ (cf. Kramsch 2009:21). Moreover, with adult education being a space
for ‘forming the citizen of tomorrow’ (Olson & Dahlstedt 2014:200), it is par-
ticularly interesting to study perceptions and constructions of competence in
adult education. The fact that Sweden has a long tradition of adult education
(e.g. Fejes & Dahlsted 2018) and that language education for adult migrants
has been institutionalized for decades, makes Sweden an interesting case to
study.

1.2 Aim and short summary of the studies

The prevailing communicative approach to language teaching and testing,
with its focus on language use in different contexts, makes it relevant to in-
vestigate both language use and reflections on communicative experiences.
Taken as a whole, the current thesis explores sociolinguistic perspectives on
language competence and assessment from the perspective of SFI students and
how perceptions of what it means to be ‘a competent language user’ are con-
structed interactionally and metapragmatically in test discourse and in focus
group discussions. Instead of treating competence as a fixed object, ‘there’ to
be studied and described, the three studies approach perceptions and construc-
tions of being ‘a competent language user’ from different perspectives. The
analytical focus is on spoken discourse (see chapter 5). This research interest
is pared down to the following overarching research interests:

1) How can the test situation be understood?

2) What ideologies of language competence and what it means to be ‘a
competent language user’ are constructed by SFI students?
3) How do communicative experiences impact SFI students’ perceptions
of what it means to be ‘a competent language user’?

4
Studies I and II build on an interactional analysis of 27 video-recorded paired
speaking tests in the final national test in SFI in the final course (course D),
whereas Study III draws on six focus group discussions with SFI students.
Question 1 is addressed in Study I through the discussion of speaking tests as
stage institutionalized performances and in Study II through the lens of inter-
actional practices in speaking tests. Question 2 is discussed in Study I and in
Study III, and to some extent in Study II, and Question 3 is addressed in Study
III.

Study I: Performance and ideology in speaking tests for adult migrants draws
on linguistic anthropological performance theory (e.g. Bauman & Briggs
1990, Bauman 2004) and sets out to discuss a framework for analysing paired
speaking tests. The article argues that paired speaking tests should be analysed
as staged institutional performances, i.e. an audience-oriented discourse that
put speaking and ideologies on display. As shown in the analysis, the SFI stu-
dents display an orientation to dominant ideological discourses on language
and integration, such as the importance of learning Swedish, speaking good
Swedish, having Swedish friends, speaking Swedish only, and being a respon-
sible and active student. Hence, ‘a competent language user’ is indexed by
being ‘a good student’ and ‘a good immigrant’.

Study II: Negotiating co-participation: Embodied word searching sequences

in L2 speaking tests pursues the theme of investigating speaking tests as a
social practice through the lens of a particular interactional practice: word
searching sequences. Word searching sequences tap into two important as-
pects of communicative language ability – vocabulary knowledge and the abil-
ity to negotiate meaning. Drawing on an interactional multimodal analysis,
Study II shows how participants prioritized the progressivity of talk rather
than pursuing linguistic precision and how they draw on different embodied
semiotic resources to negotiate co-participation or to display an avoidance to
participate. In addition, by avoiding using other languages than Swedish, the
test takers sustain a monolingual ideology in their display of being ‘competent
language users’.

Study III: Being ‘a competent language user’ in a world of Others – Adult

migrants’ perceptions and constructions of communicative competence ad-
dresses perceptions and constructions of what it means to be ‘a competent
language user’ through the lens of a group of SFI participants and their lived
experience of language (Busch 2017). Study III discusses how competence is
perceived as a relational construct shaped by intersubjective processes in com-
municative practices. A particular focus is given to the role of assessments –
both those made in institutional and everyday practices – pointing to how as-
sessments become internalized by the speaker and impact the speaker’s self-

5
perception of being or not being ‘a competent language user’. Being a ‘com-
petent language user’ is oriented to as a desired, yet vulnerable, subject posi-
tion, and the relational view of competence stands in contrast to the view of
language learning as an individual responsibility and achievement often ex-
pressed in dominant discourse.

1.3 Organization
Following this introduction, chapter 2 discusses the institutional framing of
this thesis by addressing the development of SFI from a historical and con-
temporary perspective. Chapter 3 discusses constructions and perceptions of
competence in relation to second language learning with a particular focus on
the communicative teaching and testing movement. In addition, chapter 3 ad-
dresses ideologies of ‘good language’ and adult migrants’ perceptions of com-
petence. Chapter 4 sets out to discuss issues in relation to testing and assessing
speaking, followed by chapter 5, which presents the data used in this thesis
together with methodological and theoretical considerations and assumptions
underlying the analyses in the three studies. Chapter 6 presents an extended
summary of the studies upon which this thesis builds, and chapter 7 provides
a concluding discussion on sociolinguistic perspectives on language assess-
ment and testing, followed by a summary in Swedish (chapter 8).
Perceptions and constructions of competence are not context-free. In order
to understand the discursive practices investigated in the current thesis, the
following chapters serve to contextualize and address perceptions and con-
structions of adult migrants’ language competence from a sociohistorical and
sociopolitical perspective. As stated by Bauman, social life is

discursively constituted, produced and reproduced in situated acts of speaking

and other signifying practices that are simultaneously anchored in their situa-
tional contexts of use and transcendent of them, linked by interdiscursive ties
to other situations, other acts, other utterances. (2004:2)

Hence, the following chapters on the sociohistorical context of SFI and dis-
courses on adult migrants’ language competence (chapter 2), frameworks un-
derlying communicative language testing (chapter 3) and the testing of speak-
ing (chapter 4) should not be read solely as a ‘background’, but rather consti-
tutive in the situated meaning-making practices investigated in this thesis. As
such, the situated interactions investigated in this thesis form part of an insti-
tutional, educational and societal discourse.

6
2 Language education for adult migrants in
Sweden – the case of SFI

In order to contextualize the studies in the present thesis, the following chapter
will discuss the development and the current status of the tailored language
programme for adult migrants called Swedish for Immigrants (SFI).5 The aim
of the education programme is to provide students with a basic knowledge of
Swedish by developing a functional second language. The programme pro-
motes a communicative approach and is designed to provide linguistic tools
for ‘communication and active participation in daily, societal and working life
and for continuing studies’ (National Agency for Education 2017a). As stated
in the Education Act, anyone over 16 years old, a resident of Sweden and in
need of tuition in basic Swedish is allowed to enrol in the courses free of
charge (SFS 2010:800, chapter 20 §7, §28)
The development of SFI from the 1960s until today is closely linked to mi-
gration to Sweden and the regulating of migration policies. It is regarded as
one of the first government initiatives in the development of an integration
policy and is still considered a cornerstone in Sweden’s migration and inte-
gration politics (Dahlström 2004). Demographic changes due to migration and
shifting migration policies have become palpable in SFI classrooms inter alia
in terms of an increasing number of students and changing student popula-
tions. From this outlook, this chapter will start by addressing migration to
Sweden from a historical and contemporary perspective (2.1), followed by a
description of the development of SFI and of its current organization, its syl-
labus and those who are enrolled in SFI (2.2). Section 2.3 addresses how SFI
finds itself in a ‘nexus of discourses’ (Rosén 2013:133) and how these dis-
courses impact SFI and the view of adult migrants’ language learning. The
chapter ends with a short summary of previous studies on SFI (2.4).

5
SFI is not the only educational programme for adults in Swedish as a second language in
Sweden. SFI is the first basic language programme, and after passing SFI, students become
eligible for continuing courses in municipal adult education. In addition, there are courses for
exchange students at the universities and at private schools, and there are study associations
that offer language instruction and individual tutoring. However, the focus in this thesis is on
SFI.

7
 A point to be made in this chapter is that SFI constitutes an educational
setting surrounded by political discourses that become noticeable in the recur-
rent changes characterizing the language programme and in the symbolic
value given to the language competence of adult migrants. These societal dis-
courses on national identity, integration and employment become visible in
situated classroom interactions (e.g. Karrebæk 2013, Rosén & Bagga Gupta
2015; see also Study I).

2.1 Migration to Sweden and migration policies

A frequent discursive image of Sweden is that, until recently, it has been an
ethnically and linguistically homogenous country. This stereotype often ob-
scures the fact that Sweden has historically been populated by diverse groups,
both in terms of indigenous groups like the Sami and immigrants, e.g. Ger-
mans from the Hanseatic League in the Middle Ages, Finns, Wallons and
Romani people settling in the 1600s and Jews in 1800s (Hult 2004, Milani
2008, Rosén & Bagga-Gupta 2013, Runblom 1994, Svanberg & Tydén 2005).
Still, Sweden has unquestionably undergone considerable demographic
changes since the second half of the 20th century. As pointed out by Gal
(2018), there is no neutral way of talking about migration. This section mainly
represents the perspective of state regulations and official statistics. While
Sweden at the end of the 19th century and the beginning of the 20th century
was largely a country of emigration, the country became a territory of immi-
gration in the 1940s and onwards (Nilsson 2004). Meanwhile, it is important
to note that Sweden, like many other countries, is a space for both emigration
and immigration. However, with a few exceptions in the 1970s, immigrants
have outnumbered emigrants since the 1930s (Statistics Sweden 2016). Sub-
sequently, the portion of the population born outside Sweden has continually
expanded. In 2000, 11 % of the population was born abroad, compared to 15%
in 2012 and 18 % in 2016 (Milani 2008, Statistics Sweden 2013, 2016).
The demographic changes are not only manifested in the number of mi-
grants. Since the 1950s, there has also been an increasing diversity in the char-
acteristics of the immigrated population, concerning country of origin and rea-
sons for migrating to Sweden. During the second half of the 20 th century, the
official migration policies and regulating laws have shifted, directly impacting
immigration flows. While the 1950s and 1960s were characterized by work
immigration, mainly from Germany, Greece, Turkey, the former Yugoslavia
and the neighbouring Nordic countries, immigration was, for the first time,
regulated and more restricted in the late 1960s (Borevi 2012, Nilsson 2004).
Hence, following the new immigration regulations during the 1980s and
1990s, immigration was mainly characterized by asylum seekers and family

8
reunions from regions of conflict such as Chile, Iraq, Eritrea, the former Yu-
goslavia and Somalia. In 1995, Sweden became a member of the European
Union, which granted EU citizens the right to live and work in Sweden. It is
important to note, however, that shifts in the immigration population are not
mainly due to official regulations in Sweden, but rather to circumstances in
the world (Banke 2017). During the 2000s, deliberations and regulations, both
at the EU level and at the national level, have been characterized by both a
will to facilitate need-based work immigration, as well as to control and re-
strict undocumented migration (Parkhouse 2016). Following an international
refugee crisis, a record of asylum seekers was noted in Sweden in 2015, which
led to a dramatic shift in migration policies in the autumn of 2015. Briefly,
this shift can be described as going from a generous migration policy with
permanent residence permits for those granted residency to a standard target-
ing the EU minimum level with short temporary residency and constraints in
the regulations for family reunions (Banke 2017, SFS 2016:752).
The increased mobility and diversification of immigrants does not, of
course, only hold true for Sweden. Rather, this is characteristic of the situation
in many countries. Some scholars (e.g. Blommaert & Rampton 2011, Creese
& Blackledge 2018) use the notion super-diversity (Vertovec 2006, 2007) to
stress this level of diversity, which Vertovec (2006:1) has described as a ‘di-
versification of diversity’. Characteristic of super-diverse societies is a com-
plex interplay of different features of the population concerning country of
origin, migration channel, legal status, migrants’ human capital (particularly,
educational background), access to employment, locality and transnationalism
(Vertovec 2007:1049). The notion of super-diversity has been criticized for its
emphasis on being a recent Late Modern phenomenon and by only represent-
ing a Western perspective, which obscures the fact that from a historical and
global perspective, diversity is commonplace (e.g. May 2014). However, as
underscored by Milani, what is interesting here, is ‘the way in which this in-
creasing diversity has been perceived and officially addressed’ (2008:28). In
this thesis, this is discussed in the light of constructions and perceptions of
what it means to be ‘a competent language user’ in the context of language
education for adult migrants in Sweden.

2.2 The development of Swedish for Immigrants

(SFI)
SFI can be considered a privileged language programme (Lindberg & Sand-
wall 2007). Ever since its inception, it has been provided free of charge to

9
those who need such instruction. Each municipality in Sweden is obliged, ac-
cording to the law, to provide SFI courses within three months if there is a
demand for them (SFS 2010:800 chapter 20, 29§). In addition, as of 2009,
when the Language Act (SFS 2009:600) was issued, SFI finds support in 14§,
which states that ‘[a]ll residents of Sweden are to be given the opportunity to
learn, develop and use Swedish’.
Despite its intentions, the language programme has received much criti-
cism. As pointedly indicated by the title of the National Agency for Educa-
tion’s report (1997) Who loves SFI?, SFI has, time and again, been criticized
in terms of alleged low efficiency, a lack of quality in its implementation and
insufficient individualization, which subsequently has resulted in recurrent
changes with respect to curricula, syllabi and organizational structures
(Hyltenstam & Milani 2012, Lindberg & Sandwall 2007; 2017, Rosén &
Bagga Gupta 2013). These recurrent changes have been motivated, on the one
hand, by a will to address an alleged inefficiency, and, on the other hand, to
what can be described as an institutionalization process that seek to integrate
SFI in the national education system together with other school forms. More-
over, the recurrent changes can be seen as symptomatic of how SFI has been
given a symbolic function in the political debate (see section 2.3) and how
organizational changes are perceived as a ‘quick fix’ to a complex reality
(Hyltenstam 1996).

2.2.1 SFI from 1965 to 2018: a short history

Following a government decision in 1965, SFI took shape as a provisional
pilot activity to meet the needs of demographic changes due to increased mi-
gration to Sweden and as a consequence of a new political focus on the role
of language in society (Hyltenstam & Milani 2012, National Agency for Ed-
ucation 1997). The tuition primarily targeted work immigrants (Rosén 2013)
and was ‘a joint responsibility of the state, trade unions and employers’ (Lind-
berg & Sandwall 2007:81). The Swedish debate on language and migration
pertained to an international discussion on providing language instruction for
work migrants. In 1968, the Council of Europe dictated that foreign workers
were entitled to at least 200 hours of teaching in the receiving country’s offi-
cial language, which came to serve as a bench-mark in the Swedish discussion
(National Agency for Education 1997). Initially, the tuition was organized by
private study associations through the form of study circles, with no formal
curricula and no formal requirements for teachers (Lindberg 1996). In addi-
tion, the tuition could also be organized in larger workplaces during regular
working hours. The first curriculum, manifested as recommendations to the
study associations, was issued in 1971. The curriculum prescribed the so

10
called direct method, a form of audiolingual teaching method ‘in an accord-
ance with a structural view of language and behaviouristic learning theory’
(Lindberg & Sandwall 2007:84).
Following years of deliberations discussing the need for more cohesive ed-
ucation, two government bills (SOU 1981:86, SOU 1981:87) prepared for or-
ganizational changes and a new status for SFI as part of the national education
system. Finally, in 1986, SFI became a permanent programme with a formal
curriculum and teacher qualification requirements, albeit modest ones (Lind-
berg & Sandwall 2007). The previously recommended direct method had
largely been criticized as being too mechanical (Lindberg & Sandwall 2012).
The curriculum now advocated a communicative approach, focusing inter alia
on the active participation of students and individualized and learner-centred
teaching. However, as surveys later showed, these ambitions were scarcely
implemented, and as a consequence, a new SFI reform accompanied by a new
curriculum was issued in 1991 (National Agency for Education 1997). The
curriculum of 1991 aimed to facilitate a more flexible education that mini-
mized the waiting period before enrolling in the language programme in order
to increase its efficiency.
Still, the 1991 curriculum did not last long, and in 1994 SFI underwent
considerable changes. This time change was motivated by a major national
school reform, which marked an important shift in the history of the Swedish
education system. First and foremost, the national educational system was de-
centralized, and the municipalities became the principal organizer of all
schools. In addition, for SFI, this meant that the municipalities were, accord-
ing to law, obliged to provide SFI classes within three months after applying,
leading to an open enrolment policy for learning centres. Moreover, a new
curriculum accompanied by syllabi was issued for all forms of schooling (in-
cluding SFI), reflecting an important political change and view of knowledge.
A new grading system was implemented for upper primary schools and sec-
ondary schools, shifting from a relative normative system to goal-based syl-
labi and a criterion-referenced grading system (Lindberg & Sandwall 2012).
As a step towards integrating SFI in the national education system, SFI no
longer had its own curriculum. Instead, SFI was issued a new syllabus per-
taining to the common curriculum for non-compulsory school forms.6 For the
first time, SFI was accorded a formal grading system (Pass/Pass with distinc-
tion), and the implementation of the final national exam in SFI in 1996 (see
section 2.2.3) marks an important step in the institutionalization of the pro-
gramme.

6
In the Swedish education system, there is one curriculum common for all compulsory schools
and another curriculum for non-compulsory schools. Hence, the curriculum is quite general.
Syllabi are issued for different subjects and courses. As the syllabi are issued by the National
Agency for Education, they have a formal status and cannot be changed by individual schools.

11
 Individualized and learner-centred teaching, together with a need to in-
crease SFI’s throughput, continued to be the focus of debates and surveys of
SFI, leading to a new syllabus being issued in 2002. In order to meet the dif-
ferent needs of a diverse student group and to create attainable goals, a system
of four courses with three study tracks that depended on estimated study pace
of students, their prior schooling and knowledge of Swedish was implemented
(see Figure 1). This was also the system in use during my data collection.
Track 1 targets students with no or limited prior schooling, and track 3 targets
students with greater previous experience in formal education. The courses
range from beginning level up to intermediate level, approximately equivalent
to B1/B1+ on the CEFR scale (Council of Europe 2001, National Agency for
Education 2009).

Figure 1. Study paths (1-3) and courses (A-D) in SFI implemented in the syllabus in
2002. This was the system in use during my data collection. The final course (course
D) thus targets students from all study paths.

New syllabi were issued again in 2006, 2009, 2012 and 2017. Some changes
were quite minor, such as the new syllabi in 2006 and in 2017, whereas others
were the result of larger national education reforms (1994, 2012) or due to
changes in the structure of the language programme (2002). As of 2007, the
requirement to study civics as part of SFI was no longer part of the syllabus,
thereby putting a stronger emphasis on SFI as primarily a language pro-
gramme. Moreover, the syllabus stated that students in need of basic and func-
tional literacy training should be provided teaching in order to acquire those
skills. Following a substantial national school reform in Sweden, a new sylla-
bus was put in place in 2012, introducing a new seven-step grading system
(A-F), which is currently in use. This grading system mainly addressed the
needs that had emerged in the primary and secondary school systems, but was
imposed on SFI for the sake of a homogenized school system. In 2016, SFI
was merged into the municipal adult education comprising courses up to sec-
ondary school level, and was thereby no longer considered a school form of
its own, but still had its own syllabus and national tests. Following this organ-
izational shift, a new syllabus was issued once again for SFI (National Agency

12
for Education 2017a), with slight modifications such as the addition of the
need for students to develop digital skills and source criticism. ‘Continuing
studies’ was added to the target domains ‘everyday, societal and working life’
describing the overall aim of the language programme. Moreover, the new
syllabus foregrounds that the student should be able to take all courses within
their study path:

Figure 2. Study tracks and courses in SFI as of 2018.

This characterizing instability is noticeable in the current thesis as well. From

the time I started collecting data in 2011 until the last data was collected in
2016, two different syllabi and grading systems had been in use. However, the
underpinning communicative approach has remained, as well as the aim that
students should develop functional language skills for participation in daily,
societal and working life. During the entire history of SFI, the crucial role of
language for participation in societal and working life has remained unques-
tioned, and the link between SFI and the labour market and integration policies
has always been strong (Hyltenstam & Milani 2012, National Agency for Ed-
ucation 1997 Rosén 2013). Increasingly, the link to the labour market and that
students should become ‘employable’ has been foregrounded.
Ever since its beginning, SFI has been struggled with professionalization
and recognition (Hyltenstam & Milani 2012, Lindberg & Sandwall 2007,
2017). A parallel worth mentioning is that struggles for legitimacy are not only
characteristic of SFI, but also of the primary and secondary school subject
Swedish as a second language (Hyltenstam & Milani 2012, Sahlée 2017). Re-
peatedly being questioned leads to a stigmatization of the language pro-
gramme.

13
2.2.2 Current organization, syllabus and participants enrolled in
SFI
SFI is organized at the municipality level, but it is up to the municipality to
decide whether the education should be provided by the municipal adult edu-
cation, by private tendering or by study associations (studieförbund). In addi-
tion, SFI is provided for in state prisons. SFI is regulated by three main policy
documents: the Education Act (SFS 2010:800), the regulation of adult educa-
tion (SFS 2011:1108) and the current syllabus (National Agency for Educa-
tion 2017a). According to Sandwall (2013:11), SFI is divided by policies for
integration (to take active part in society), employment (SFI as a tool for be-
coming employable and self-supporting) and educational policies (a qualified
language programme giving eligibility to further studies in the education sys-
tem). The aim of the programme is for students to be provided with ‘language
tools for communication and active participation in daily, societal and work-
ing life for continuing studies’ (National Agency for Education 2017a). The
syllabus highlights flexibility and the need to facilitate work placement with
language instruction and stresses the link between the education programme,
employability and the labour market (Rosén & Bagga-Gupta 2013). The
knowledge requirements are divided according to different skills following
the Common European Framework of References (Council of Europe 2001):
listening comprehension, reading comprehension, oral interaction, oral pro-
duction and writing ability.
One political reform that had an impact on SFI is the Swedish Procurement
Act, which has allowed for competitive tendering and corporate profit for pri-
vate schools (Lindberg & Sandwall 2017, SFS 2007:1091). The school system
has, as it were, become a market (Dahlstedt & Fejes 2018). The marketization
and the competitive system has been particularly noticeable in SFI, where an
important share is run through short-term private tendering. Due to the open
enrolment policy and a competitive system, some schools receive new stu-
dents every week. The official benchmark concerning the time allocated for
each student to the language programme is 525 hours (§ 24 SFS 2011:1108),
but the time frame is flexible depending on the individual student’s needs.
Following increased migration to Sweden, the number of students enrolled
in SFI is increasing year by year. In fact, a comparison between the years 2000
and 2013 shows that the number of students has tripled (National Agency for
Education 2014). In 2011, when I started the data collection, about 102, 400
students were enrolled in SFI, and in 2016 when I finished collecting data, the
number pf SFI students had increased to 150, 000 (National Agency for Edu-
cation 2013a, 2017b).7

7
The stricter migration policies issued in late 2015 have led to a decrease in the number of
asylum seekers and immigrants. However, as stated in a recent report by the National Agency
for Education (2018), the demand for language instruction in municipal basic education is still
strong.

14
 SFI admits students from all over the world with different linguistic, cul-
tural and educational backgrounds and of varying ages. From this vantage
point, the SFI classroom can be described as a ‘contact zone’ (Pratt 1994),
where people from all over the world meet. More than 160 first languages are
represented among the students, with the most common first languages cur-
rently being Arabic, Tigrinja, Somali and Farsi. The age distribution among
SFI students has been quite stable over the years. In 2016, the largest group,
56 % of the students, were 25˗39 years old; 30 % were older than 39; and 14
% were younger than 25. 12% studied track 1; 42 % studied track 2; and 46 %
studied track 3 (National Agency for Education 2017b). The prior formal ed-
ucation of those studying SFI varied from no formal education at all to com-
pletion of post graduate studies. However, 60 % of the students had under-
taken at least 10 years of schooling. Notwithstanding that years of prior
schooling is a rather fuzzy measurement, i.e. it is hard to say whether one year
of schooling in one country is equivalent to one year of schooling in another
country (cf. Simpson 2006:47), prior formal education and literacy are im-
portant aspects with respect to the students’ success and how far they go, i.e.
which course becomes their final course (National Agency for Education
2013a).

2.2.3 The national test in SFI

The final national test in SFI was implemented in 1996. At that time, the test
was only constructed for the final course (course D). As of 2009, national tests
for courses B and C have been provided by the National Agency for Educa-
tion. The aim of the national tests is to guarantee equal assessment nation-
wide and to support teachers in assessment and grading (e.g. Ahlgren 2016).
The national test is compulsory but is not an exit exam, i.e. students do not
need to pass the test in order to pass the course, even though in practice, it is
often used as an exit exam (National Agency for Education 2006a, Lindberg
& Sandwall 2007:85). As of 2018, the importance given to the test has been
further enhanced and institutionalized due to a change in the Education Act
stating that the test result should be decisive for the final grade of the course
(SFS 2010:800; see also section 4.1).
In accordance with the open enrolment policy and the fact that SFI has no
fixed semesters, the national test is given on a recurrent basis during the year.
How many times a year tests are administered depends on the municipality.
As an illustration, the municipality of Stockholm organized 10 test dates in
the academic year 2011/2012, which increased to 12 test dates in 2018, i.e. it
is administrated once a month (Stockholms stad 2012, 2018). The students’
regular teachers decide when the individual student is ready to take the test,
i.e. when he or she is considered to have fulfilled the course requirements
(National Agency for Education 2006a). It follows that the attainment for the

15
test is generally high (National Agency of Education 2017b). In 2016, 21, 800
students took the test (in the final course, course D) and 85 % passed.
The national test consists of different components that basically follow the
syllabus division for aims and grading criteria with respect to listening com-
prehension, reading ability, oral production, oral interaction and writing abil-
ity (Ahlgren 2016). The national test is administered at the local schools, but
since it is distributed by a national institution and thus belongs to a wider in-
stitutional context, it is likely to have a normative impact on the view of
knowledge and subsequently on the teaching (Carlson 2002:92).
In contrast to many other countries, Sweden has a tradition of assigning the
responsibility of assessing external tests (like the national test) to regular
teachers at the local schools (Lundahl 2011, Skar 2013). Subsequently, assess-
ment is a core professional activity for teachers. Hence, the administration of
the test is decentralized and the organization of the assessment of the test can
vary. A recent study on the practices of the speaking test of another final na-
tional test in the Swedish education system (the test in English) showed that
despite efforts of standardization in the instructions accompanying the exam,
the way the test is carried out varies greatly across schools (Sundquist et al.
2018).
A compilation of the results in the final national test (course D) between
2009 and 2012 show that the oral component differs in relation to the other
components (reading ability, writing and listening comprehension), in terms
of being the component with the highest attainment (Stockholm University &
National Agency for Education 2013). Moreover, years of prior formal school-
ing did not seem to have an impact on attainment in oral language. By contrast,
this was the case for the other components (the higher level of prior schooling,
the higher level of attainment). The reason the oral component differs from
the other components can be several. One reason could be the impact of the
communicative approach to teaching, i.e. that speaking is often practiced in-
side and outside the classroom. Another possible reason is that oral language
is a more ‘general’ ability that everyone is socialized to attain, whereas liter-
acy practices such as tests in reading and writing are more linked to access to
formal education. Due to the fact that it is up to the local teachers to decide
when a student is ready to take the test, the high attainment of the oral com-
ponent could indicate that most students reach requirements in oral language
ability first, whereas the other abilities take more time to develop. A more
pessimistic interpretation would be that the perceived difficulty in assessing
speaking and related problems of reliability impact the assessment (Stockholm
University & National Agency for Education 2013).

16
 2.3 Intersecting discourses on adult migrants
language competence
Discursive struggles over migration have become increasingly salient in pub-
lic and political discourse in Sweden and elsewhere (cf. Gal 2018). According
to political scientists Karin Borevi (2012), discourses on migration and migra-
tion policies in Sweden are closely associated to the development of the wel-
fare state and Sweden’s self-image as a progressive state. In general, there has
been a discursive shift from the emphasis in the 1960s on what the state should
do for newcomers to today’s focus on ‘what the newcomers can and should
do for society’ (Rosén & Bagga-Gupta 2013:82). While the 1960 and 1970s
were characterized by an emphasis on multiculturalism and pluralism8, inter
alia, granting immigrants an opportunity to learn Swedish (Hyltenstam &
Milani 2012, Runblom 1994), the beginning of a retreat from multiculturalism
was noted in the 1980s (Borevi 2012, Dahlström 2004). Ever since, there have
been competing ideologies; between discourses on rights and obligations, be-
tween Sweden’s self-image as a humanitarian and progressive country and
calls for stricter migration policies accompanied by political discourses fore-
grounding the responsibility for immigrants to learn about Swedish values, to
acquire the Swedish language and to become employable (Borevi 2012,
Milani 2008). The view of migration as a threat to social cohesion and national
security has become an increasingly common rhetorical trope in current polit-
ical and public discourse (Strömbäck et al. 2017). In line with this, there has
been a shift towards a rhetoric on requirements in relation to adult migrants
living in Sweden. Discourses on adult migrants’ language learning and com-
petence thus tap into wider discourses on migration and the nation state. His-
torically, language has played a constitutive role in the construction of nation
states and to what is perceived as a culture through the Herderian paradigm of
one nation-one language –one people (e.g. Anderson 1991, Pujolar 2007). A
case in point of how this assumption still operates is the use of language tests
in the regulation of citizenship, a current practice in most European countries
(van Avermaet 2009). Language tests for naturalization point to how testing
serves as a distinguishing feature with the ambiguous role of acting ‘as mech-
anisms of exclusion in the name of national inclusiveness’ (McNamara &
Roever 2006:182). While Sweden has resisted implementing language tests
for citizenship9, the implementation of such a testing system has recurrently

8
This is noticeable in the highly influential commission report (Invandrarutredningen) in 1975,
which outlined the major principles for the immigration politics: equality – freedom of choice
– partnership. As stated by Runblom, ‘[t]he freedom-of-choice goal supported immigrants’
right to choose whether to retain their their homeland culture, to “become Swedes”, or to blend
traits from the homeland and Swedish culture (1994:208)’.
9
As stated in a commission report on citizenship (SOU 1999:34), knowledge of Swedish was
regarded important for integration. But an official language requirement for citizenship was

17
been suggested in political debates and in government bills (such as in 1991,
2002, 2011, 2015, 2018; see also Milani 2008 for a discussion on the proposal
made in 2002 by the Liberal Party). In general, it is the language test that is
the focus of this thesis (the national test in SFI) that has been suggested as a
requirement in the regulation of naturalization.
As previous research on SFI has noted, a deficiency discourse often pre-
vails in relation to SFI, where SFI students are positioned by their lack of
abilities in Swedish and their lack of relevant work experience in relation to a
perceived Swedish norm (Carlson 2002, Rosén & Bagga Gupta 2013). In ac-
cordance with this deficiency perspective, adult education serves as a means
of remediating learners’ deficiencies and as a ‘space for citizen training’ (Ol-
son et al. 2018:95). As Hyltenstam and Milani (2012) maintain, these defi-
ciency discourses should be interpreted in light of the symbolic function of
SFI and of adult migrants’ language competences in relation to migration and
ideologies of language and the nation-state.
On a general note, language education for adult migrants ‘has come to be
viewed more as a process of integration than education’ (Khan 2014:9). In this
vein, SFI finds itself in a ’crossfire of discourses’ (Carlson 2002:220) and in
the intersection of different political agendas concerning integration, employ-
ment and national identity (e.g. Burns & Roberts 2010, Hyltenstam & Milani
2012, Lindberg & Sandwall 2012). Based on the idea that language is key to
society, there is also a discourse on inclusion with respect to language instruc-
tion for newly arrived migrants (Fejes & Dahlstedt 2018). As a consequence,
there has been a political focus on SFI as a means for solving other societal
problems like unemployment and integration (Carlson 2002, Lindberg &
Sandwall, 2012:405), by ‘fixing the language problem’ (Rosén & Bagga
Gupta 2013:77). The instability described above in section 2.2.1 can thus be
interpreted in light of the ideological linking of SFI and adult migrants’ lan-
guage competence into larger societal discourses on migration, integration,
employment, national identity and language competence (Hogan-Brun et al.
2009, Milani 2007, Rosén 2013). Similar discourses have been reported in
studies of language education for adult migrants in other countries such as
Denmark (Bjerg Petersen 2010), the United Kingdom (Khan 2014, Roberts et
al. 2007) and Australia (Burns & de Silva Joyce 2007), to mention a few. This
creates a tension between long-term goals, such as developing linguistic re-
sources, and short-term goals, such as making the students employable as
quickly as possible (Lindberg & Sandwall 2017).

considered inappropriate and unjust due to the fact that language learning and learning out-
comes are dependent on many factors of which some are not in the control of individual learn-
ers. Adult immigrants’ language learning was also considered an issue for the education system
rather than for the migration agency.

18
 All these societal discourses on national identity, integration and employ-
ment become visible in situated classroom interactions (e.g. Karrebæk 2013,
Rosén & Bagga Gupta 2015, see also Study I). Moreover, shifting societal
discourses become palpable in the different curricula and syllabi of SFI. In a
study of curricula and syllabi regulating SFI between 1971 and 2006, Rosén
and Bagga-Gupta (2013:75) note how the early days of SFI were characterized
by a worker-oriented focus dealing with rights and duties in the work place
and basic communicative skills. This worker-oriented focus has gradually
shifted to a work-oriented focus enmeshed in discourses on employability (see
also Fejes [2010] for a discussion on discourses on employability in relation
to adult education). Furthermore, the strong emphasis on SFI as a labour mar-
ket tool has led to an increasing demand on vocational language courses, lead-
ing to an instrumental view of language competence (Carlson & Jacobsson
2018, Sandwall 2013).

2.3.1 The duty to learn

The increasing rhetoric on requirements discussed above is, inter alia, palpa-
ble in discourses on language learning and learning ‘good Swedish’ and how
this is achieved. This can be illustrated by a recent quotation from Ulf Kris-
tersson, the leader of the Conservative Party:

The situation now is that many people don’t succeed with SFI. Then I think it’s
better to raise the demands and say that we expect, not only that you are there
but that you learn Swedish and actually link the results to social benefits. (Swe-
dish Public Radio, Ekot [the public radio’s news programme] 2018, my trans-
lation)

Alleged low achievement by SFI students can, according to Kristersson, be

solved by ‘raising the demands’ and by state benefits being conditional on
successful achievements in the language course. In other words, SFI students
dependent on benefits must pass SFI in order to keep their financial support.
Underpinning this quote is the view of language competence as located in the
individual and the construction of successful language learning as dependent
on the individual’s own effort (cf. Park 2010). Conversely, ‘unsuccessful’
learning can be explained by the individual’s lack of effort (Lindberg & Sand-
wall 2017).
A key element in the Swedish welfare model is free access to education
(Carlson 2002, Fejes et al. 2018). SFI courses are state-subsidized, and the
municipalities are obliged by law to provide language courses if there is a
demand. From this point of view, the quote above provides an example of
Foucauldian ‘governmentality’, where the state enables learning by providing

19
free language courses, but the responsibility for learning lies with the individ-
ual student (Fejes 2010).
Like all adult education in Sweden, SFI is a non-compulsory school form.
However, this is not true for all students. Students registered at an employment
agency and dependent on government-funded benefits are often required to
follow SFI courses as part of their so-called establishment plan, i.e. scheduled
activities by the employment agency (Abdulla 2017). These students are sub-
jugated to government control by reporting their attendance at the risk of los-
ing their benefits if they fail to do so. The fact that SFI is part of a nexus of
government agencies (National Agency for Education, the Employment
Agency and the Social Security Agency) reinforces its institutional frame and
the symbolic value of passing SFI. As noted by Carlson (2002), there seems
to be a tension between ‘freedom of choice’, which has been a political leit-
motif since the 1990s, and the requirement to study, and pass, SFI. In the law
regulating the establishment of new arrivals, reformed in 2017 (SFS
2017:584), a formal ‘learning duty’ (utbildningsplikt) was introduced target-
ing unskilled migrants registered at an employment agency. As the law stipu-
lates, members of this particular target group can lose their benefits if they do
not follow the educational programmes (of which SFI can be one) assigned in
their activity plan. In addition, the Social Democrats (the party currently in
government) suggested during the electoral campaign in spring 2018 what
they called a ‘language duty’ (språkplikt), an obligation to study Swedish for
asylum seekers and other newly arrived migrants depending on social benefits.
The alleged low efficiency of SFI has not only been addressed by or-
ganizational changes and new syllabi as those described in section 2.2.1. Other
political reforms have addressed perceived low efficiency of SFI and alleged
lack of motivation of the students. One such reform is the SFI bonus (SFS
2010:1030), a bonus system for SFI students passing the SFI courses within a
given time frame (12 months). Underpinning this reform was the idea that SFI
students are not sufficiently motivated, and that a financial bonus will moti-
vate them and thereby increase the throughput. The SFI bonus system oper-
ated only for a short time before it was abandoned, but was a reality during
the time I collected the test data. Another example is the Official Report of the
Swedish Government Time for quick flexible learning (Tid för snabb flexibel
inlärning SOU 2011:19),10 which suggested that newly arrived migrants
should start SFI within a year and that the time studying SFI should be limited
to a maximum of two years. Underpinning these proposals, and the recurrent
critique of low efficiency (i.e. that some students take too long to pass the SFI
requirements), is the idea that learning Swedish depends on the individual’s
effort, and if the student is motivated, that it should not take too long (Lindberg
& Sandwall 2017). Such a view obscures the relational construction of com-
petence (see Study III) and the highly different pre-conditions for learning in

10
This report did not become a government bill and the suggestions were never implemented.

20
the target group (Kerfoot 2009, Lindberg 2004), as well as and how language
learners are ’historically and sociologically situated active agents, not just in-
formation processing machines’ (Block 2003:109). Nor do they improve the
learning conditions from an organizational perspective (such as the
teacher/student ratio or the open enrolment system). Instead, successful and,
conversely, unsuccessful learning is located within the individual, and the
completion rate in SFI would seemingly be easily improved if students were
to be given a financial incentive or by restricting their time in SFI (Lindberg
& Sandwall 2017).
In contrast to the negative image of SFI often depicted in debates and re-
ports, surveys and interview studies show that SFI students are, by and large,
satisfied with the language programme (e.g. National Agency of Education
1997, 2013b). Moreover, interview studies have foregrounded the empower-
ing role of the education, in particular, for students who have limited or no
prior access to formal education (Ahlgren 2014, Carlson 2002).
While uncontested claims that migrants are not motivated enough or do not
make enough effort to learn Swedish ‘constitute a frequent element in the
dominant discourse on integration’ (Lindberg & Sandwall 2017:124), this
does not seem to be the case made in interview studies with adult language
learners. Rather, interview studies show how adult language learners claim to
be motivated to learn and often state the importance of learning the main-
stream language (Ahlgren 2014, Carlson 2002, Cooke 2006, Miller 2014).
From another perspective, it can be argued that adult migrants claim to be
motivated to learn partly ‘because they are “recognized” as social beings who
should do so’ (Miller 2014:26; see also Rosén & Bagga-Gupta 2013:69).

2.4 Previous research related to SFI

For a long time, SFI was a fairly under-researched educational programme

(Rosén 2013), but a notable change in this regard has been observed in recent
years. The great majority of studies relating to SFI have been qualitative stud-
ies drawing on ethnographic observations and interviews (Lundgren et al.
2017). Previous studies have addressed such issues as classroom discourse and
communicative teaching (Lindberg 1995), and have included a longitudinal
case study of a Greek woman’s second language learning process (Granberg
2001) as well as studies on constructions of knowledge and learning in SFI
from a sociological perspective (Carlson 2002), low literate SFI students and
their visual literacy (Franker 2011), the construction and negotiation of iden-
tity positions and categorizations in classroom discourse and policy docu-
ments (Rosén 2013), SFI students’ interaction and opportunities for learning

21
during their work placements (Sandwall 2013), life narratives from a longitu-
dinal perspective (Ahlgren 2014), institutional talk between SFI students and
career counsellors (Sheikhi 2013), embodied classroom interaction and the
construction of ‘learnables’ (Majlesi 2014), literacy practices in and outside
the classroom among five Kurdish SFI students (Norlund Shaswar 2014), so-
ciocultural influences on adult migrants’ learning process in basic adult edu-
cation (Zachrison 2014), and the formation of teacher identity among teachers
in initial literacy training (Colliander 2018).
Of particular interest for this thesis are the studies made by Carlson (2002),
Rosén (2013), Ahlgren (2014) and Zachrison (2014). Carlson’s (2002) socio-
logical study discusses knowledge construction and perspectives on learning
in SFI at macro-level (Sweden’s history of education), meso-level (the insti-
tutional framework for SFI) and micro-level perspectives (interviews with SFI
students, teachers, principals, career advisers, a school nurse and welfare case
workers). According to Carlson, SFI is rooted in the Swedish welfare model
and discourses on educational optimism, where education is regarded as in-
herently good and as a means of fostering ‘good citizens’ to take responsibility
for their own life and learning. The SFI participants interviewed (12 Turkish
women with limited prior schooling) emphasized how participating in SFI is
empowering, but also how they are positioned in relation to a Swedish norm
as being ‘passive, traditional and backwards’ - an image they tried to resist.
The fact that a norm of Swedishness is pervasive throughout the language pro-
gramme and how discourses of Swedishness, gender, language and migration
are reproduced in classroom interactions and in teaching materials is discussed
by both Carlson (2002), Rosén (2013) and Zachrison (2014). In particular, the
categorization of being an immigrant in relation to Swedishness influences
SFI students and leads to an ‘othering’ of them.
The categorizations and identity positions in relation to SFI are the focus
of Rosén’s study (2013). Combining classroom ethnography with discourse
analytical studies of policy documents regulating SFI from its beginning in
1965 until 2011, Rosén investigates discourses surrounding SFI both in policy
documents and how these discourses are reproduced in classroom interactions.
Her study also shows how the discursive construction of the target group for
SFI and of knowledge have shifted over time from a pluralistic caring per-
spective to a deficiency discourse and a focus on employability.
SFI students’ reflections on communicative experiences have been ad-
dressed by Carlson (2002), Ahlgren (2014) and Zachrison (2014). In particu-
lar, communicative difficulties and a perception of being regarded as less
knowledgeable due to limited access to linguistic resources in Swedish have
been anchored in the participants’ communicative experiences (Ahlgren 2014,
Carlson 2002, Zachrison 2014). Drawing on Ricoeur’s notion of ‘narrative
identity’ and a longitudinal interview study with five adult second language
speakers (who in the first data set had just finished their SFI studies), Ahlgren
(2014) foregrounds how adult language learning is a demanding and time-

22
consuming process with considerable impact on one’s self perception. This
became palpable in the participants use of metaphors for describing language
learning and use, e.g. ‘journey’ and ‘struggle’ causing ‘stage fright’.
As pointed out by Lundgren et al. (2017) in a survey of research devoted
to SFI, the national test in SFI has been given little attention despite its strong
status, both from a societal perspective and within the language programme
itself. The current thesis contributes a sociolinguistic perspective on testing
and assessment practices by investigating paired speaking tests in the final
national test in SFI as a speech event constituting a specific kind of institu-
tional and classroom discourse. Furthermore, this thesis adds to the growing
body of research devoted to SFI by examining how perceptions of what it
means to be ‘a competent language user’ are constructed both metapragmati-
cally and interactionally by SFI students in test discourse and focus group dis-
cussions.

2.5 Concluding remarks

In sum, section 2 has shown how SFI is a language programme vested in po-
litical and ideological interests, characterized by recurrent changes. Relatedly,
the symbolic value of being ‘a competent language user’ has become increas-
ingly salient in globalized societies. This has become apparent in dominant
discourses on requirements in relation to adult migrants being responsible for
becoming ‘competent language users’ in Swedish.

23
3 Constructions and perceptions of L2
competence

This chapter sets out to discuss different understandings and constructions of

what constitutes language competence and what it means to be ‘a competent
language user’. Communicative approaches to language teaching and lan-
guage competence have provided a rationale for the use and proliferation of
speaking tests. In this chapter, a particular focus is therefore given to the com-
municative teaching movement and the models used to promote and to test
communicative language ability in a second language context. The chapter
starts by addressing models and theories in relation to communicative compe-
tence and communicative language ability (3.1). In addition, different cri-
tiques of the notion of communicative competence are discussed. Given that
the communicative approach to language teaching and language learning often
stresses the ability to use language in different contexts, it becomes necessary
to discuss language as a social practice and how ideologies of ‘good language’
(3.2) are played out in social life. The chapter ends by addressing previous
studies on adult language learners’ perceptions of language competence and
what it means to be ‘a competent language user’.
As constructed conceptualizations of language competence, models of
competence are formed in their sociocultural and sociopolitical contexts, and
are subject to change. As such, models of competence should be understood
as artefacts ‘created to facilitate our conceptual thinking’ (Chalhoub Deville
2003:375) in order to define, and establish limits, in relation to what consti-
tutes language competence. Importantly, it follows that language constructs
should be ‘subjects to critical interrogations’ and not taken for given
(McNamara 2001:339). In contrast to discourses of language competence in
public and political discourses (see 2.3), where language competence is treated
in a simplified way and never really discussed in more detail, definitions on
how to define language knowledge have been extensively discussed within
linguistic and educational research. Conceptualizations of language
knowledge have changed over time and have been subject to theoretical shifts.
Accordingly, there is no consensus on how to define language knowledge, not
even within theories of communicative language ability and communicative
competence (e.g. Purpura 2008). A particular tension can be noted between
the view of language competence as an underlying ability becoming observa-
ble in language use (i.e. first and foremost being a cognitive construct) and

24
language-use-oriented theories highlighting the dynamics of language use
shaped in moment-by-moment negotiations (e.g. Chalhouab-Deville 2003,
Kataoka et al. 2013, McNamara 1997; see also chapter 4 for a discussion on
how this tension is particularly noticeable in the assessment of speaking), res-
onating with debates in second language acquisition-studies between cogni-
tive and interactionist approaches to language learning (e.g. Ortega 2014).
Another point made in this chapter is that being ‘a competent language user’
in migration contexts is an ideologically loaded image caught in discursive
struggles between a multilingual and diverse reality and a monolingual ideol-
ogy.

3.1 The development of the notion communicative

competence

One major arena for constructions of language competence is theoretical ad-

vancements in defining and ‘modelling’ language competence. Underpinning
curricula, syllabi and language testing programmes, these models are crucial
for an understanding of and, subsequently, the operationalisation of language
competence. Theoretical debates on how to define language competence have
been recurrent in applied linguistics and educational research. Dominating for
the last 40-50 years or so in language teaching and testing, are models of com-
municative language ability, and one of the major contributions to this devel-
opment is the notion of communicative competence. Originally proposed by
the American anthropologist and sociolinguist Dell Hymes (1967, 1971,
1972a), communicative competence became a buzzword in applied linguistics
in the 1970s. Its diffusion is mainly due to the advent of the communicative
teaching paradigm as a reaction to grammar-based instruction, which took
place around the same time. Hymes’ work was in opposition to the dominating
notion of competence as proposed by Chomsky, which contended that linguis-
tic theory should be based on:

an ideal speaker-listener, in a completely homogenous speech-community, who

knows its language perfectly and is unaffected by such grammatically irrelevant
conditions as memory limitations, distractions, shifts of attentions and interest,
and errors (random or characteristic) in applying his knowledge of the language
in actual performance’. (Chomsky 1965:3)

Hence, for Chomsky, the main interest was the universal and mental ability
for language that humans possess, and that competence should be studied apart
from performance. In stark contrast, Hymes argued that ‘this idealized con-
ception becomes inadequate as soon it is confronted with real children in a

25
particular environment’ (1971:4). Criticizing Chomsky’s notion of compe-
tence as too narrow and too concerned with idealized speakers and situations,
Hymes set out to develop a sociolinguistic theory of language dealing with ‘a
heterogeneous speech community, differential competence and the constitu-
tive role of sociocultural features’ (1971:9). Stating that language use was
ruled by social context, Hymes thereby criticized the prevalent dichotomy of
competence (as knowledge of an abstract system) and performance (use of the
knowledge) as being inadequate. This general point was made by adding com-
municative to competence. For Hymes (1971, 1972a), communicative compe-
tence comprised both knowledge (grammatical and sociolinguistic) and ability
for use (including volatile and affective factors). If the current linguistic the-
ory and view of language competence was too narrow, it needed to be ex-
panded. Hymes acknowledged grammar as an important part of communica-
tive competence, but it was useless if not put in relation to rules of use. As one
of the most cited quotations goes:

There are rules of use without which the rules of grammar would be useless.
Just as rules of syntax can control aspects of phonology, and just as semantic
rules perhaps control aspects of syntax, so rules of speech acts enter as a con-
trolling factor for linguistic form as a whole. (Hymes 1972a:278)

Central to Hymes’ understanding of competence to use language was the ques-

tion of appropriateness according to the context. What counts as appropriate
is learnt through participation in particular speech communities and is ac-
quired by children through socialization, along with the acquisition of gram-
mar, making communicative competence a relational construct. According to
Hymes, researchers should be interested in how communication is conducted
in particular contexts in particular speech communities through ethnographic
studies.11
Even though the notion of communicative competence is widely used, its
largest impact has been in education, particularly in language teaching and
testing. Although the communicative approach to teaching was seen as prom-
ising in the 1970s, communicative competence for language education was for

11
At the time, Hymes was not the only one engaged in departing from Chomsky’s theory and
moving towards a social approach. In 1970, the philosopher Habermas, pointed out a different
direction for the emerging notion of communicative competence. Not concerned with education
as such, Habermas considered communicative competence a promising direction for an ex-
panded theory of language. If Hymes wanted to change the direction of ethnographic work and
inform it by theory, Habermas was interested in developing a more general theory of language
and meaning. Mirroring Chomsky, communicative competence was described as relating to ‘an
ideal speech situation in the same way that linguistic competence relates to the abstract system
of linguistic rules’ (Habermas 1970:369). However, in applied linguistics, it was Hymes’ un-
derstanding of communicative competence that gained ground.

26
a long time undertheorized (Canale 1983). The work by Canale and Swain
(1980) and Canale (1983) marked a significant turn. Drawing on Hymes’ work
and others (e.g. Savignon 1972), Canale and Swain aimed to build a theory
for second language programmes, teaching and testing. In educational set-
tings, the question is not only what the students should learn and how to teach
it, but also what to assess and how. Models of communicative competence and
communicative language ability in language education and testing have there-
fore been quite specific, by dividing competence into different components.
Canale and Swain (1980:29f.) proposed a model consisting of three compo-
nents: grammatical competence (knowledge of lexical items, morphology,
syntax, sentence-grammar semantics, phonology), sociolinguistic competence
(the appropriateness of an utterance in relation to the context of use) and stra-
tegic competence (strategies to circumvent communicative problems). Canale
(1983) later added discourse competence (the ability to form coherent and co-
hesive texts and knowledge of genres) to the model. As pointed out by
McNamara (1996), one important difference between the work of Hymes and
the model proposed by Canale and Swain (1980) is the treatment of ability for
use, which was central to Hymes but deliberately omitted by Canale and
Swain, who doubted that ‘there is any theory of human action that can ade-
quately explicate ability for use’ (1980:7). Conversely, for Hymes
(1972a:282), ‘[i]n speaking of competence, it is especially important not to
separate cognitive factors from affective and volatile factors, so far as the im-
pact of theory on educational practice is concerned.’ Another difference is the
relation between competence and performance. Refuting Chomsky’s view of
performance as ‘an imperfect manifestation of an underlying system’ (Hymes
1972a:272), Hymes argued for a broader understanding of performance.
Drawing on Goffman’s (1959) early work on performance, Hymes posited
that

performance of a person is not identical with a behavioural record, or the im-

perfect or partial realization of individual competence. It takes into account the
interaction between competence (knowledge, ability for use), the competence
of others, and the cybernetic and emergent properties of events themselves. A
performance, as an event, may have properties (patterns and dynamics) not re-
ducible to terms of individual or standardized competence. Sometimes, indeed,
these properties are the point (a concert, play, party). (Hymes 1972a:283)

Whereas Hymes’ work made the case for a sociolinguistic theory addressing
both knowledge, ability for use and performance12, Canale and Swain empha-

12
Performance studies, in Hymes’ sense, later developed as a research object on its own in
linguistic anthropological work (e.g. Hymes 1975, Briggs 1988, Bauman 2004). See Study I for
a discussion on how paired speaking tests can be analysed as performances

27
sized that ‘[c]ommunicative competence is to be distinguished from commu-
nicative performance, which is the realization of this competence’ (1980:6).
In language education, it was by and large Canale and Swain’s view that
gained ground. It follows that it became a central task for language educators
and testers to create tasks to elicit the students’ underlying language ability.
Subsequent models targeting language teaching and testing (e.g. Bachman
1990, Bachman & Palmer 1996, Celce-Murcia 2007) have defined communi-
cative competence as a set of subcompetencies. These multi-componential
models, often de-contextualized and general, distinguish between knowledge
of form and structures (e.g. grammatical competence) and knowledge of use
(e.g. sociolinguistic competence, interactional competence) together with stra-
tegic competence. Bachman and Palmer’s model (1996), in particular, has had
an immense influence in the language testing field. Bachman and Palmer’s
conceptualization of components underlying language use provided a model
of both linguistic and non-linguistic components such as language knowledge,
topical knowledge, personal characteristics and affective schemata. Commu-
nicative language ability is in Bachman and Palmers’ model divided into or-
ganizational competence (grammatical competence and textual competence),
pragmatic competence (sociolinguistic competence and illocutionary compe-
tence) and strategic competence (metacognitive aspects).
By introducing a model of communicative competence with separate com-
ponents, the question arose how these components interact with each other.
Since the communicative teaching paradigm to a large extent was a reaction
to grammar-based instruction, a particular discussion was devoted to function
versus form or meaning versus grammaticalness (Canale & Swain 1980:10).
Another issue in adopting a framework of communicative competence into
language education has been how to describe different proficiency levels of
communicative competence, since the models as such do not provide any
guidance. Later on, policy makers have tried to meet this need of describing
different proficiency levels, with the transnational Common European Frame-
work of References – CEFR (Council of Europe 2001) being a case in point.
The CEFR divides communicative language ability into linguistic compe-
tence, sociolinguistic competence and pragmatic competence. Language use
is described through five skills: writing, reading, listening, oral production and
oral interaction, and language proficiency is described at six levels. The CEFR
is based on a functional and social approach to language that describes lin-
guistic resources and defines what learners can do with their language in dif-
ferent contexts. It is important to note that developments in language teaching
and testing are not only underpinned by theoretical movements. Rather, as
Spolsky argues (1990, 1995), the CEFR is a prime example of how political
and economic interests impact educational planning by defining educational
goals in precise terms such as in the notional-functional syllabus with its focus
on real-world situations, functions and can do-statements (see also McNamara

28
2011 for a discussion on the global implemention of the CEFR and how it has
affected testing practices).

3.1.1 Constructions of language competence in SFI: a

communicative and functional approach
The syllabus of SFI and the theoretical framework of the national test in SFI
provide an example of how the theoretical discussions outlined in 3.1 are im-
plemented in an educational programme. As outlined in 2.2, since the SFI re-
form in 1986, a communicative and functional approach to teaching and ex-
pected learning outcomes prevails in the documents regulating SFI. The cur-
rent syllabus states that students should acquire ‘communicative language
skills’ and develop a ‘functioning second language’. Communicative language
ability is described as being based on knowledge of a language system and
knowledge of how to use this system (National Agency for Education 2009,
2012a, 2017a). Moreover, the syllabus states that ‘[l]anguage correctness
should be related to the content and the complexity of the language’ and the
teaching should be based on the learners’ needs (National Agency for Educa-
tion 2017a). The link to the CEFR was implemented in the syllabus of 2009
(Ahlgren et al. 2013) and remained until it was removed from the latest sylla-
bus in 2017. However, the syllabus is still based on the CEFR’s division of
five skills (reading ability, writing ability, listening comprehension, oral pro-
duction and oral interaction), which is reflected in the national test (2.2.3).
The communicative approach to language is also noticeable in the theoret-
ical framework for the national test in SFI (Ahlgren et al. n.d., Ahlgren 2016).
Originally, the national test in SFI drew on Bachman and Palmer’s (1996)
framework for communicative language ability. The assessment grid accom-
panying the exam reflected Bachman and Palmer’s division between organi-
zational knowledge and pragmatic knowledge (paraphrased in the test as a
division between language system and language use) together with strategic
competence. Following the new syllabus in 2012, the theoretical framework
of the national tests was revised. The current theoretical framework is still
inspired by different models of communicative competence and communica-
tive language ability. In particular, the theoretical framework put forth a model
inspired by Celce-Murica’s model for communicative competence together
with Halliday’s model of register (see Ahlgren et al. n.d. for a fuller descrip-
tion). Celce-Murcia’s model of communicative competence includes soci-
ocultural competence, textual competence, interactional competence, linguis-
tic competence and strategic competence. These aspects are reflected in the
assessment grid accompanying the exam (see National Agency for Education
2012b). In the theoretical framework of the national test as well as in the syl-
labus there is an emphasis on the ability to use language in different contexts,

29
for different purposes and with different recipients, together with an emphasis
on the ability to adapt one’s language accordingly (Ahlgren et al. n.d., Na-
tional Agency of Education 2017a). The assessment thus requires an under-
standing of the social context of language use.
Following the functional and communicative approach in the syllabus and
the theoretical framework underpinning the test, the tasks in the test aim to,
‘as far as possible resemble authentic situations’ (Ahlgren et al. 2013:67, my
translation). As stated in the syllabus, ‘[t]he starting point for the assessment
should be the pupil’s ability to use the Swedish language in a comprehensible
way for different purposes in everyday, societal and working life and for con-
tinuing studies’ (National Agency for Education 2017a, my translation). The
target domains are thus mainly situated outside school.

3.1.2 Critiques of communicative competence and its use

Over the years, the notion of communicative competence and its use has been
critiqued by several scholars for different reasons. Lillis (2006) points out how
notions like context, speaker, speech community and appropriateness, which
are treated as quite unproblematic by Hymes, have been subject to recent reex-
aminations by different scholars. As pointed out by Leung (2005:124), the re-
contextualization of the notion of communicative competence from Hymes
ethnographically oriented work to educational contexts has resulted in an ep-
istemic transformation ‘from empirically oriented questions to an idealized
pedagogic doctrine’. Instead of investigating language use in different con-
texts, language education has been informed by idealized models and ideal-
ized speakers. The target language and its ‘rules of use’ easily get treated as
homogenous in language teaching, obscuring differences and underlying ten-
sions and foregrounding a normative view (see, for instance, Fairclough
[1992] for a discussion on the use of ‘appropriateness’ in course syllabi). An-
other central problem is that models of competence as those discussed in 3.1
tend to lead to ‘an inert and decomposed knowledge view’ (Leung 2014:135).
A central tenet in the communicative teaching paradigm is that language
learning takes place through language use. However, drawing on Bourdieu,
Norton B. (2013) argues that the right to speak and to be heard depends on
social relations and power and should not be taken for granted. Other scholars
have criticized the essentialist assumptions underlying the notion of commu-
nicative competence (e.g. Bucholtz 2003), its frequent alignment with mono-
lingual ideologies and ‘nativeness’ (Canagarajah 2013, Miller 2014), how the
logocentric view of competence obscures embodied perspectives on compe-
tence (Block 2014, Goodwin 2004, Kataoka et al. 2013), and the individual
focus on competence as misleading (McNamara 1997; see section 4.2.2). Fail-
ing to see language as a situated practice, the individual and essentialist focus
has often led to a view that communicative competence is something individ-
uals have or lack (Blommaert et al. 2005).

30
 Time and again, scholars have pointed out the need to extend communica-
tive competence by using new notions like symbolic competence (Kramsch
2009), i.e. drawing on one’s multilingual resources and semiotic choices to
create subject positions, or
interactional competence13 (Kramsch 1986, Young 2011) and performative
competence (Canagarajah 2013) in order to highlight the role of negotiation
and co-construction in meaning making and the practice-based nature of lan-
guage, while others have launched alternative notions like language expertise
(Rampton 1990), linguistic repertoires (e.g. Blommaert & Backus 2013,
Busch 2012), spatial repertoires (Canagarajah 2018) or being a resourceful
speaker (Pennycook 2012) to describe language use and language knowledge.
Busch’s (2012, 2017) extended take on linguistic repertoires also includes lan-
guage ideologies and emotions like desire to know a language. Recent work
in applied linguistics and educational research has thus moved away from the
term ‘competence’ and its cognitive roots to emphasize more practice-based
and multilingual approaches. By and large, these new notions still relate to
communicative approaches to language use and learning. However, as argued
by Shohamy, ‘[w]hile dynamic, diverse, and constructive discussions of mul-
tilingual teaching and learning are taking place within the language education
field, the phenomenon is overlooked in the testing field, which continues to
view language as monolingual, homogenous, and based on native-like con-
structs’ (2013:230). The focus in the current thesis is not to suggest new terms,
but rather to investigate perceptions and constructions of images of the ‘com-
petent language user’. While the notion of competence might not capture lan-
guage use adequately (Canagarajah 2018), the image of being ’a competent
language user’ is a powerful language ideology in both public discourse and
educational settings as well as in relation to speakers’ self-perception as lan-
guaging subjects. In this light, the present thesis sets out to explore how SFI
participants perceive and orient to images of being ‘a competent language
user’ and what it means to know a language.

13
In some models, such as the one used in the theoretical framework of the national test on SFI
(drawing in turn on Celce-Muricia’s [2007] model) interactional competence refers to interac-
tional management such as turn taking and topic initiation. Furthermore, interactional compe-
tence is often used a framework for oral interaction tests (e.g. Galaczi & Taylor 2018). I my
studies I have chosen not to use the notion of interactional competence since my interest is on
a broader image of what it means to be ‘a competent language user’.

31
 3.2 Being ‘a competent language user’ and ideologies
of ‘good language’
Since the advent of the communicative teaching paradigm there has been a
considerable amount of literature devoted to discussing communicative ap-
proaches to curricula, language teaching and testing (e.g. Brumfit & Johnson
1978, Nunan 1991, Roberts 2004, Savignon 1983). However, as stated by
Martín Rojo and Marquez Reiter, communicative competence is not only a
question for education; rather, it is ‘valorized in every social field according
to linguistic ideologies’ (2013:10). While models conceptualizing communi-
cative competence used in language education, such as those described in 3.1,
to a large extent describe communicative competence as an individual ability,
scholars interested in communicative competence in everyday and institu-
tional practices instead point to how communicative competence is formed in
social practice (e.g. Hymes 1972a, Kataoka et al. 2013). Moreover, since syl-
labi with a communicative approach often foreground language use in differ-
ent settings situated outside school, it becomes necessary to take account of a
social theory of language use. Kataoka et al. describe communicative compe-
tence as ‘the product of moment-by-moment negotiations within the act of
communication and large-scale social structures, historical dynamics, and the
actions and ideologies of institutions like the state’ (2013:349). Furthermore,
they point to how communicative competence is intersubjectively achieved
and maintain that ‘[i]t invokes dynamics of authenticity, plurality, and mobil-
ity, and links people’s ability to communicate with their moral valuation,
rights to citizenship and belonging, and position in the hierarchies and struc-
ture of inequality, both local and non-local’ (ibid.).
Hence, looking at language as a social practice makes it pertinent to take
account of speakers’ different social positions and the symbolic function of
language. The French sociologist Pierre Bourdieu, who also criticized Chom-
sky’s view of competence, aimed to develop a social theory of the role of lan-
guage (e.g 1977, 1982, 1991). The Bourdieusian perspective has contributed
to broadening the focus on language to see language as a practice, and Bour-
dieu’s contribution (e.g. 1977) makes a strong case for the link between com-
petence, practice and legitimacy. According to Bourdieu (1977, 1982), lan-
guage cannot be separated from the speaker, and legitimacy concerns the re-
lationship between linguistic resources and the speaker’s social position and
ability to speak with authority in different social spaces. As Bourdieu holds,
this broader sense of competence leads to the insight that ‘a language is worth
what those who speak it are worth’ (1977:652). Following this line of thought,
linguistic resources do not have a value per se, but acquire their value in rela-
tion to a specific market (Bourdieu 1982). Being a competent speaker is thus
not only about ‘speaking correctly’ and being understood, but also about being
believed and respected. As Bourdieu and Wacquant (1992:146) contend,

32
‘[i]nequality of linguistic competence constantly reveal themselves in the mar-
ket of daily interactions’. In this vein, being seen as a competent language user
is closely associated with being a legitimate language user, and, as under-
scored by Park, this ‘link between competence and legitimacy makes compe-
tence an effective resource for the reproduction of social distinctions’
(2010:24).
In addition to the theoretical uses of the notion of competence (section 3.1),
language competence is also a ‘folk linguistic’ concept, often understood in
terms of ‘being good at a language’ (cf. Park 2010). Ideas of ‘good language’
and ‘good speakers’ are neither neutral nor objective; rather they are histori-
cally and contextually situated. It follows that ideas of ‘good language’ are
political and perspectival, representing a point of view from a certain stand-
point (e.g. Miller 2014, Park & Wee 2008, Spolsky 1995). In second language
education, idealized conceptions of the native speaker serve as ‘objective
measures of good language’ (Miller 2014:85).14 As argued by Ortega (2014),
the ideal speaker has been conceptualized as a native speaker with a monolin-
gual upbringing, leading to an inevitable deficiency perspective on language
learners. From this follows, then, that underpinning the notion of language
competence is an ideology of monolingualism (see Studies I and II).
Furthermore, Sjögren (1996) notes how the importance of learning good
Swedish, with its touch of common sense, is a collective image aimed at cre-
ating a sense of order in a diverse and changing society. Incrementally, this
leads to the view of ‘good Swedish’ as a key to inclusion in society. As
Sjögren further notes, ‘the more the importance of language is foregrounded,
the more deviations from the standard language become delicate for those who
listen’ (1996:36, my translation). Thus, an important language ideology in re-
lation to language competence is the valorization of the standard language and
linguistic correctness. As argued by Gal, ‘[c]ontrary to the common sense
view, standardization creates not uniformity but more (and hierarchical) het-
erogeneity’ (2006:21). As a consequence, for those who do not ‘speak cor-
rectly’ or strive to achieve good speakerhood it becomes necessary to express
loyalty to the standard (Silverstein 1996; see also Study I). In this light, lin-
guistic knowledge and norms of correctness become features for positioning
people and for ‘constructing opposing social groups’ (Milani 2008:38). Dis-
courses on language competence in migration contexts pertain to what Busch
has called ‘pre-Babel fantasies’ (2012:517), where linguistic diversity is con-
structed as a problem against the imaginary of a common and unified language

14
The rating scale for Certificate of Proficiency in English (Cambridge English Proficiency) in
1975 provides an historical example of the link between the ‘competent language user’ and a
certain type of native speaker. The highest level of oral proficiency was described as follows
(Vidaković & Galaczi 2013:331):
Excellent His command of spoken English approaches that of a soundly educated native
speaker of English.

33
and universal understanding (see also Cameron 2002). In this vein, percep-
tions and constructions of competence and ‘good language’ lie at the heart of
discursive struggles between a multilingual reality and a monolingual ideol-
ogy (e.g. Silverstein 1996, Hogan-Brun et al. 2009; see also Studies I and II).
Hence, when discussing language competence, idealization of language and
speakers form part of our understanding of ‘competent language users’.

3.3 Adult language learners’ perception of language

competence
In educational settings, theories of language knowledge as a formal construct
(i.e. communicative competence and communicative language ability), inform
course syllabi and testing programmes. However, as stated by Kramsch
(2009:4), for learners, language is a ‘lived embodied reality’ rather than a for-
mal construct. In this vein, investigations of perceptions and subjective expe-
riences of language provide an important complementary perspective on lan-
guage in relation to cognitive and usage-based approaches to language (Busch
2017). Explorations of adult language learners’ perceptions of language learn-
ing and use often provide a broader perspective on language by highlighting
embodied and emotional aspects of language (Ahlgren 2014, Kramsch 2009,
Pavlenko 2006). Furthermore, language learners are often objectified in mod-
els of language competence and in syllabi. Hence, investigating language
learners’ perceptions shifts the focus from language to speakers (Kramsch
2009). Considering dominant discourses of the symbolic importance for mi-
grants to learn the receiving country’s mainstream language, it becomes im-
portant to investigate adult migrants’ perceptions of communicative compe-
tence. Even though language learning and communicative experiences have
been investigated, relative little attention has been given to adult language
learners’ perceptions of competence, based on their lived experience of lan-
guage (see Study III). Perceptions are based on embodied experiences of being
in the world and are formed in and through discursive practices (Busch 2017).
As foregrounded in Study III, language assessment practices, those made in
both educational and everyday practices, and experiences of being seen or not
seen as a competent language user, form an important part of the adult lan-
guage learners’ perception of competence (Study III). Since perceptions take
shape in social experience, they are constructed in accordance with the
speaker’s social position. Hence, investigating expressed beliefs on compe-
tence provides a window into contemporary language ideologies and construc-
tions of competence (Study I, Study III).

34
 Interview studies with adult language learners’ often broach perceptions of
‘good language’ (‘good Swedish’, ‘good English’ and so on). However, per-
ceptions of ‘good language’ often remain as a quite vague phenomenon asso-
ciated with a native like construct (Miller 2014). Furthermore, perceptions of
competence are rarely unitary among different learners. Meanwhile,
knowledge of vocabulary and grammar often holds a key position in percep-
tions of what it means to be competent in a language (Carlson 2002, Zachrison
2014). In Carlson’s interview study, both SFI teachers and students emphasise
the role of grammar, even though the teachers did so more than the students.
Carlson (2002) notes how, for many teachers and language learners, grammar
enjoys the status of corresponding to ‘proper language education’ and gram-
mar is linked to a sense of structure and security. However, closer scrutiny of
why grammar and vocabulary are needed point to the fundamentally social
aspect of language, since for many language learners it is in language use that
vocabulary and grammar become relevant (Study III; see also Hymes 1971:8
which underscores ‘a social component at the heart of grammar’).
Meanwhile, interview studies with adult language learners do not only fore-
ground the knowledge of grammar and vocabulary. As adults, these students
often need to participate in different social arenas, such as in institutional en-
counters at governmental agencies, going to see a doctor or dealing with issues
related to being a parent. This often leads to adult language learners’ fore-
grounding the importance of managing independently in social encounters, of
understanding and of being understood (Carlson 2002, Zachrison 2014, see
also Study III). More importantly, as adults, they are used to being able to
participate in different social settings in their previous country of residence.
Migration often entails a loss of symbolic, cultural and economic capital, and
migrants often need to re-evaluate their language competence. From this per-
spective, migration can represent a loss (Busch 2017, Marquez Reiter & Mar-
tín Rojo 2013), and language learning in migration contexts is driven by a
desire to regain one’s voice, a process that impacts on adult language learners’
perceptions of competence.

3.4 Concluding remarks

The focus on communicative language teaching has been to ‘promote the de-
velopment of functional language ability through learner participation in com-
municative events’ (Savignon 2006:675) and to develop a pedagogy based on
learners’ needs. Models of communicative language ability have tried to pin
down what resources are needed in order to communicate effectively and to
make meaning. However, models of language competence and of different
proficiency levels construct language competence as an ostensibly objective
and transparent construct to be achieved by individual learners. Even though

35
scholars in education like Savignon (1983) considered communicative com-
petence to be a dynamic concept, being primarily interpersonal rather than
intrapersonal, the individual focus of competence has dominated. The individ-
ual view of competence is highly institutionalized, with band descriptions and
grading criteria describing the linguistic resources of an individual and what
the individual student can do with his or hers linguistic resources, and scores
or grades given to individuals.
The communicative approach to language learning and competence, with its
focus on language use, has led to a proliferation of speaking tests, which we
now turn to.

36
4 The testing and assessment of speaking

One of the most important arenas where constructions and perceptions of com-
petence play out is in language testing practices and language assessments.
Language testing refers to a ritualized practice of collecting information about
learners’ language competence carried out by institutional representatives (i.e.
language teachers and testing bodies), whereas language assessment is more
broadly defined as conscious and unconscious judgments based on observa-
tions of a speaker’s language competence, occurring in both institutional and
everyday practices. From this perspective, language testing is part of language
assessment, but not necessarily the other way around.
Developments in language testing are not only motivated by theoretical and
methodological considerations such as those discussed in section 3.1. Rather,
testing, as an instrument of powerful institutions, can also be closely linked to
political and economic conditions in society (McNamara 2011, Spolsky 1995,
Weir 2013). Therefore, this chapter starts by addressing the sociopolitical con-
text of language testing and assessment, followed by a discussion on the de-
velopment speaking tests. A particular focus is given to previous research on
paired speaking tests and how they can be understood as a speech event.
One of the points made in this section is that testing and assessment of oral
interaction shed light on tensions between different assumptions made on the
nature of language, such as being an individual or a relational construct. More-
over, testing is a social and ideological practice, which makes the reproduction
of language ideologies in testing practices interesting to study.

4.1 The sociopolitical context of language testing and

assessment
Testing and assessment as a gatekeeping activity in order to sort and select is
a human activity that has been established for centuries (e.g. Foucault 1977,
McNamara & Roever 2006, Spolsky 1995). An often cited example of the
inherent social and political character of language assessment is provided in
the story of the shibboleth in the Bible, in the Book of Judges, where the pro-
nunciation of the Hebrew word ‘shibboleth’ was used to distinguish between
friends and enemies (Spolsky 1995, McNamara & Roever 2006; see also
McNamara 2012 for an extended discussion of language tests as shibboleths).

37
According to the biblical story, the ‘result’ was highly consequential since it
led to the killing of those pronouncing the word ‘wrong’ and thereby revealed
belonging to the outgroup. Spolsky (1995, 1997, 2017) attributes the birth of
examinations to the Imperial Chinese system of selecting civil servants, which
was imported to Europe in the 16th century, first to religious schools and, later,
to the secularized school system. Some of the early examination systems, like
the French Baccalauréat, which marks the end of the secondary school system
and was introduced by Napoleon I at the beginning of the 19th century, are still
in use today (Spolsky 2017).
In his book Discipline and Punish, the French philosopher Michel Foucault
(1977) discusses how examinations (alongside the founding of prisons) were
established in France in the 18th century in order to control and discipline. The
disciplinary power of examinations is achieved through the surveillance and
documentation of individuals. Typically, this is achieved by a ‘compulsory
visibility’ where the students are the ‘visible’ ones, which leads to a process
where ‘subjects’ become ‘objects’ to be observed and categorized. In testing
practices, there is a tradition of emphasizing individual differences. Foucault’s
writings are particularly interesting in relation to communicative language
testing, with its focus on productive skills, such as the testing of oral interac-
tion. In paired speaking tests, the candidates have the interactional floor more
or less to themselves, thus enhancing the visibility (and subsequently the vul-
nerability) of the test takers. Moreover, as pointed out by Foucault, examina-
tions provide a mechanism for linking power to the ‘formation of knowledge’
(1977:187). In this sense, testing practices construct and reproduce what
counts as valuable knowledge. Tests are powerful instruments, and in all test-
ing practices there is an inherent power imbalance between the assessor and
the assessed.
The ‘inevitable uncertainty’ in obtaining reliable assessment was pointed
out in one of the earliest articles on assessment (Edgeworth 1890:662), and
psychometric approaches to language assessment have strived to achieve valid
and reliable tests. Modern language testing, as an institutional practice based
on standardized and ‘objective’ tests, took shape in the second half of the 20th
century, within and in relation to developments in psychology and psychomet-
rics (i.e. the statistical measurement of psychological tests). Hence, language
testing, as a discipline, was rooted in psychology and psychometrics before it
became a field in applied linguistics. However, as pointed out by McNamara
and Roever, ‘through marrying itself to psychometrics, language testing has
obscured, perhaps deliberately, its social dimension’ (2006:1).
During the 21st century ‘testing has become big business’ (Spolsky
2017:375) and has ‘assumed increased importance in economic, educational
and socio-political affairs of society’ (Weir 2013:1). The use of examinations
to control the education system and teaching is one of the most pervasive
mechanisms of testing (McNamara & Roever 2006, Shohamy 2001, 2006).

38
Increasingly, test results are used as part of the political concern for account-
ability where test results are interpreted as representative of the quality of
teachers’ performances, the quality of schools and the education system as a
whole (Lundahl 2011, Spolsky 2017). McNamara and Roever see this ‘politi-
cization of assessment’ as the ‘most striking feature of current developments
in language assessment’ (2006:213). The link between language tests and po-
litical interests is most noticeable in relation to migration, such as language
requirements (and subsequently language tests) in regulations of citizenship.
This is a practice in many countries today, but closer inspection shows that the
language requirements for obtaining citizenship in different countries vary
(van Avermaet 2009). As McNamara and Roever (2006:182) rightly point out,
the different levels of requirement set for being granted citizenship in different
countries show that these requirements are first and foremost political. The
increased use of politically motivated language tests has provoked discussions
on ethics in language testing and the development of ethical codes for testers
(McNamara & Roever 2006, Saville 2009, Spolsky 2017).
Hence, throughout its history, language testing has had a social and political
side. Accordingly, language tests constitute an important site for the reproduc-
tion of dominant ideologies and for constructing images of the ‘competent
language user’ (see Study I). The educational psychologist Samuel Messick’s
seminal work on validity (e.g. 1989), stressed the need to investigate social
values imbedded in testing practices and the social consequences of testing.
This has, inter alia, been taken up by the movement known as Critical Lan-
guage Testing (Lynch 2001, Shohamy 2001, 2006, 2017). As argued by Sho-
hamy, ‘there is ample evidence that tests have an effect on curriculum change,
teaching and testing methods, redefinition of language knowledge, motivation
to teach “test language”, narrowing the linguistic knowledge’(2006:103).
Notwithstanding these important insights, it is important to note that lan-
guage testing and assessment do not only have an impact on issues in relation
to curricula, teaching and knowledge construction. They also deeply affect
learners (Shohamy 2001). Language assessments made in institutional and
everyday practices are pervasive mechanisms influencing L2 users’ self-per-
ception, both in a potentially empowering way as well as the opposite (see
Study III). Classroom-based assessment, comprising both formative (i.e. as-
sessment for learning) and summative assessment (i.e. assessment of learning,
such as tests), is a recurrent classroom practice. Appropriately used, assess-
ment can be an important instrument and a pedagogical tool for enhancing
students’ learning (e.g. Hill & McNamara 2011, Lundahl 2011).
Meanwhile, the implementation of national tests or external testing instru-
ments like PISA is decided at a political level. Sweden, as elsewhere, has seen
an increasing use of national standardized tests and a public and political belief
in testing as a way to control, evaluate and improve the education system. In
2017, a change in the Swedish Education Act shifted the status of the results
of the national tests from being advisory in the teacher’s grading to ‘be given

39
special consideration for grading’ (SFS 2010:800, chapter 20, 37a§, my trans-
lation), hence making the test scores more decisive. Around the same time,
the government, for the first time in the last 30 years or so, decided to reduce
the number of compulsory national tests in the Swedish secondary school.
Hence, national tests are not uncontested but subject to political changes.

4.2 The development of speaking tests

The youngest sub discipline in language testing is the assessment of second
language speaking (Fulcher 2003). Even though oral language ability was gen-
erally acknowledged to be important throughout the 20th century,15 speaking
skills were often omitted from language tests, or given a limited role, since it
was considered unfeasible to test speaking in a valid and reliable way (Fulcher
2003, Spolsky 1995, Vidaković & Galaczi 2013). On a general note, the de-
velopment of speaking tests has been driven by two concerns. On the one
hand, the proliferation of speaking tests has been motivated by societal needs,
i.e. in terms of employability and meeting language requirements in work-
places. On the other hand, there has been a concern for tests to reflect con-
structs of communicative language ability, and classroom teaching practices
leading to an increased use of speaking tests (e.g. Ducasse 2010). Moreover,
oral language ability is often seen as a central ability among learners (Rydell
2016).
How tests are constructed sheds light on how ‘ideal’ competence is per-
ceived. A brief look at the history of oral testing will show that the under-
standing and subsequent operationalization of language competence changes
over time. One interesting example is provided by the Certificate of Profi-
ciency in English (CPE) offered by the University of Cambridge University
Local Examinations Syndicate (and later Cambridge ESOL Assessments), one
of the first modern language tests established in 1913 and seen as increasingly
viable (see Weir, Vidaković & Galaczi 2013 for a historically overview of
Cambridge language assessments). The 1913 version of the test included an
oral component consisting of a conversation with an examiner, a dictation and
reading aloud. In addition, pronunciation was tested in an essay on phonetics.
Tasks like reading aloud were used until the mid-1980s, when the communi-
cative approach to the speaking construct gained importance and made room
for tasks such as the paired format (Vidaković & Galaczi 2013; see also sec-
tion 4.2.1 for a discussion on the paired format). Even though oral testing was
introduced early in the British tradition, for a long time it remained ‘seriously
undertheorized’ (McNamara & Roever 2006:44).

15
Of course, oral language has been of importance to language education longer than that, such
as in classical rhetoric and in pedagogical approaches such as the Reform Movement based on
spoken language in the 19th Century (Weir 2013).

40
 Another contribution to the development of modern oral tests is generally
attributed to the Oral Proficiency Interview (OPI), developed in the 1950s by
the American Army’s Foreign Service Institute’s test for staff stationed abroad
(McNamara 1996). The proliferation of the Oral Proficiency Interview from
the 1950s onwards and the growing body of research on the OPI ‘brought up
the issue of whether a single type of interaction (i.e. an interview) is sufficient
to assess oral proficiency and whether it is representative of conversation’
(Vidaković & Galaczi 2013:271). The critique of the OPI led to the use of
multiple test formats of which pair and group orals became important. As
noted by Lazaraton, ‘the research on individual candidate discourse seemed to
ebb in the 2000s, when the focus shifted to a relatively new speaking test for-
mat, what is known as the pair (or group) oral’ (2014:1382), which we now
turn to.

4.2.1 Paired speaking tests

Paired speaking tests provide a case in point for communicative language test-
ing. Underpinning the use of the paired format is the idea that a discussion
between students will elicit a wider range of their communicative competence
and entails more symmetrical and collaborative interaction and a wider range
of speech functions in comparison to teacher-student interaction, which tends
to be dominated by students answering the teachers’ questions (Ahmadi &
Sadegi 2016, Brooks 2009, Taylor 2001). It is important to note, however, that
the choice of test format does not only pertain to theoretical issues of how to
define and measure speaking ability, but also to practical issues (Vidaković &
Galaczi 2013).The fact that it is cost- and time-efficient to test in pair and
group constellations has also been put forward as an advantage (e.g. Ducasse
& Brown 2009). The paired format was introduced by the University of Cam-
bridge Examination Syndicate in the 1980s (Ducasse 2010). Following the
advancements of theories of communicative competence and communicative
language teaching, pair- and group discussions became popular in language
classrooms and as a testing format in the 1980s and 1990s (Vidaković &
Galaczi 2013), and are now used in high and low stakes tests around the world.
The fact that the CEFR expanded the traditional four skills (reading, writing,
listening, speaking) to five by dividing speaking into oral interaction and oral
production has also led to an increased use of the paired format as it is regarded
as being suitable for assessing oral interaction.
Many language tests, such as the one investigated in this thesis, use multi-
ple formats (e.g. Taylor 2003; see also section 5.1.1), such as a paired task and
an individual task. In a comparison between the paired task and an individual
task, Brooks showed that‘[w]hen the test takers interacted with other students
in the paired test, the interaction was much more complex and revealed the
co-construction of a more linguistically demanding performance than did the

41
interaction between examiners and students’ (2009:341). Galaczi’s (2008) in-
fluential study on paired speaking tests in the Cambridge First Certificate in
English illustrates how different interactional patterns could be distinguished
in the test discourse as either being collaborative (equally distributed turn tak-
ing where the test takers engaged in each other’s ideas and gave listener sup-
port), parallel (equally distributed turn taking, but where the test takers did
not engage in each other’s ideas) or asymmetrical (one of the test takers dom-
inating interactionally). The collaboratively driven speaking tests were re-
warded when scored for ‘interactional management’, leading Galaczi to the
conclusion that being able to engage in a collaborative discussion is a sign of
a higher level of interactional competence. Hence, collaborative and coopera-
tive interaction have been pointed out as important benefits of the paired for-
mat. Relatedly, positive washback (i.e. how tests impact classroom practices)
has been reported in the sense that the paired speaking tests encourage peer
interaction in language classrooms (Együd & Glover 2001, Galaczi & Taylor
2018).

4.2.2 Issues in testing and assessing speaking

Alongside the increased popularity of the paired format in language tests, a
growing body of research on interaction in speaking tests has emerged (e.g.
Brooks 2009, Ducasse 2010, Galaczi 2008, 2014, Lazaraton & Davis 2008,
Nyroos et al. 2017; see also Taylor & Wigglesworth’s [2009] special issue in
Language Testing). Research on interaction in paired speaking tests has ad-
dressed diverse issues such as discourse patterns (Brooks 2009, Galaczi 2008,
2014), divergent understandings of the task (Sandlund & Sundquist 2013) and
code-switching in repair sequences (Nyroos et al. 2017). A number of studies
have addressed the so-called ‘interlocutor-effect’, i.e. how interlocutors’ char-
acteristics such as personality (Ockey 2009), gender (Brown & McNamara
2004), proficiency level (Davis 2009, May 2009) and familiarity with one an-
other (O’Sullivan 2002) influence the discourse produced. In addition, re-
search on interaction in speaking tests has investigated the construct of inter-
action in relation to how raters (May 2011) and candidates perceive it and how
the construct of interaction can be used in designing rating scales (Ducasse
2010), as well as how the test situation impacts on the test discourse (Luk
2010, Simpson 2006; see also Studies I and II).
Testing and assessing speaking is considered to be a difficult endeavor (e.g.
Luoma 2004, Vidaković & Galaczi 2013). A main concern for testing bodies
is to provide fair, reliable and valid testing practices. Meanwhile, interaction
is difficult to control and oral assessment ‘has presented a whole array of dif-
ficult and at times controversial issues’ (Vidaković & Galaczi 2013:257). In
particular, issues of variability, co-construction and authenticity have been

42
raised in relation to the testing of speaking. Studies drawing on Conversa-
tional Analysis (see section 5.2) have brought attention to how the perfor-
mance in speaking tests is a joint accomplishment achieved in interaction (e.g.
Brown 2003, Lazaraton 2002, Galaczi 2014, Nyroos et al. 2017, Young & He
2008). This has led to tension, to say the least, between the view of perfor-
mance as a projection of an individual ability and the view of social interaction
as co-constructed (Jacoby & Ochs 1995, McNamara 1997, McNamara &
Roever 2006). Co-construction is defined as ‘the joint creation of a form, in-
terpretation, stance, action, activity, identity, institution, skill, ideology, emo-
tion, or other culturally meaningful reality’ (Jacoby & Ochs 1995:171). Re-
search, such as Brown (2003), showing that the examiner’s behaviour, i.e. in
terms of interactional support in an OPI, has an impact on the candidate’s per-
formance and the subsequent scoring of it, led to efforts in standardizing tests
and the interactional frame such as the use of scripted instructions and a focus
on examiner training. But the fact that interaction is co-constructed, i.e. dy-
namic and based on moment-to-moment negotiations (e.g. Jacoby & Ochs
1995, Swain 2001), makes the practice of standardization difficult. This be-
comes even more so in the case of paired speaking tests, where the interac-
tional floor is given to test takers. However, the view of language use as some-
thing that needs to be standardized and a fixed object with testable features
also shifts the focus ‘from the relativity of interpersonal communication’
(Kramsch 1986:370), which at worst can stand in the way of achieving and
displaying communicative competence. Hence, there is a tension between en-
suring reliability through standardizing testing procedures and collecting evi-
dence of test takers’ communicative competence through interaction.
Likewise, collaboration in paired speaking tests, which is seen as one of the
benefits of the paired format, has also been considered problematic in the test
literature. The co-construction of the performance makes it a challenge to give
individual grades (May 2009), which is generally required. Moreover, the
pairing of the test takers in relation to their proficiency level, familiarity with
each other and personality may influence the result. Investigations of such is-
sues have presented contradictive findings, and as noted by Lazaraton ‘while
test taker proficiency level, gender, familiarity, and the like do influence the
discourse produced with a partner, it is unclear if these variables impact the
outcome scores on the test’ (2014:1383). However, instead of regarding the
interlocutor as a problem, managing complexity and diversity in interaction
through negotiation strategies can be considered an integrated part of what it
means to be able to communicate in a language (Brooks 2009, Canagarajah
2013). Insights like these have provoked an ongoing discussion within the
testing field on whether the constructs in use are sufficiently rich or need to
be expended in order to capture relevant aspects of language use and commu-
nicative language ability (Galaczi & Taylor 2018, Harding 2014). Relatedly,
Norton J. (2013) makes the case for an extended definition of co-construction

43
to include how test takers and examiners interpret and interact with examina-
tion materials (i.e. interlocutor frames, topics) and how the in situ interaction
in a test situation is interdiscursively connected to the participants’ previous
experiences and expectations.
The view of language as a situated practice also leads to the problem of
making generalized claims of test takers’ communicative competence based
on observations of a performance in a specific setting like testing (Chalhoub-
Deville 2003, McNamara 1996). The need to justify the generalized claims
made in language tests has been met by argument-based approaches to valid-
ity, i.e. frameworks for justifying claims on students’ language ability based
on particular observations (i.e. tests) and the inferences of these claims to non-
test settings (see Kane [1992] and Bachman & Palmer [2010] for comprehen-
sive frameworks on argument-based approaches to validity). An important
step in this procedure is the generalization of test scores to provide an argu-
ment for why observations of a test performance say something about the can-
didate’s language ability in other contexts (such as work or study life). This is
a particular problem in performance testing, with performance being highly
situated (McNamara 1996). Meanwhile, argument-based approaches to valid-
ity do not resolve the inherent incompatible views of performance as a projec-
tion of an individual ability and performance as co-constructed and achieved
in interaction. As argued by McNamara over 20 years ago, ’[t]he focus of the
ability of the candidate in conversational approaches within second language
assessment views the candidate in strangely isolated light, it is he or she who
is held to bear the brunt of the responsibility for the performance, in this sense
the inevitable gap between a test and real life appears unusually stark’
(1997:452f). This has, to some extent, been met by including aspects of inter-
actional competence (such as turn-taking management, topic management and
interactive listening) into grading criteria and assessment grids, thus acknowl-
edging the dynamic and co-constructed nature of interaction (e.g. Ducasse
2010, Galaczi 2014, Galaczi & Taylor 2018).
Ducasse (2010) investigated how teachers and students oriented to the con-
struct of interaction through verbal protocols while watching video-recorded
paired speaking tests in Spanish. The construct of interaction was defined in
terms non-verbal interpersonal communication, interactive listening and inter-
actional management. Both students and teachers highlighted to role of em-
bodied signs such as gestures and gaze in the test taker
s’ meaning making. Furthermore, while watching the videos, the students
recalled difficulties in managing the interactional flow and interpreting the
fellow test taker’s interactional cues. Negotiating co-participation and the use
of embodied semiotic resources are investigated interactionally in Study II in
this thesis. Notwithstanding important additions on including interactional
competence and attempting to account for co-construction, the institutional
practice is still focused on individual ability, since scoring and grading, in the
end, is applied to individuals.

44
 On another note, preoccupation with the features of spoken discourse in
test talk in relation to ‘normal conversation’ led to the so-called authenticity
debate, i.e. whether the discourse produced in test settings corresponded to the
target language use domain, either by the cognitive process triggered by the
task or by the interactional pattern (Galaczi & Taylor 2018, Lazaraton 2002,
van Lier 1989, Luk 2010; see also Young & He [2008] for an overview of
discourse studies on the OPI). Interaction in speaking tests was, as it were,
accused of being ‘unauthentic’ (e.g. see Luk [2010] for a discussion on how
students and teachers perceive interaction in group orals to be ‘unnatural’).
The authenticity debate lead to an increase of empirical studies on discourse
produced in test settings. Meanwhile, as argued by McNamara and Roever,
‘the gap between the evidence obtained under test conditions and inferences
about performance in non-test conditions is a given in assessment’ (2006:48).
From another perspective, interaction in test situations can be understood as a
social practice and a speech event in its own right, rather than being ‘inauthen-
tic’ and ‘unnatural’ (Study I; see also Lazaraton [2014]). The following sec-
tion will discuss how paired speaking tests can be seen as staged institutional-
ized performances, thus representing one type of performance ‘in the commu-
nicative life of a society’ (Hymes 1972a:284).

4.2.3 The speaking test as a speech event

Tests are sometimes discussed in terms of being more or less ‘context-rich’ or
‘context-specific’(Galaczi & Taylor 2018:12) by constructing a social context
in the test (such as role play and given social settings), obscuring the fact that
the test situation is a context in itself. In order to understand the language per-
formed in a test setting, it is important to investigate it as a ‘speech event in
its own right’ (Galaczi 2008:92). A speech event in Hymes’ understanding is
‘governed by rules or norms for the use of speech’ occurring in a particular
speech situation (Hymes 1972b:56). From this perspective, the speaking test
is a speech situation, and the paired task is a speech event. The paired speaking
test is a complex interactional activity, where the test takers are both address-
ing each other and the teachers acting as examiners, who are listening and
assessing the conversation (cf. Eklund Heinonen 2009; see also Study I). The
speaking tests are framed by what Linell labels ‘a pre-interactional framing’
(2011:175), impacting the participants’ understanding and expectations of the
speech event. This is achieved though talk about the speech event as ‘a test’
or ‘an oral test’ and by deciding in advance where, at which time and between
whom the speaking test should take place. Hence, the test takers are highly
aware of the fact that it is a test. As argued by Perren (see also Stevenson
[1985] for a discussion on validity and authenticity, drawing on Perren’s quote
below):

45
 The traditional oral test has often been one of ‘conversation’ in which the stu-
dent is required to demonstrate his ability under conditions fondly supposed to
be representative of real life – in a two-way conversation. In theory this is a sort
of work-sample test in which all the elements of skill in spoken language can
be used. But, of course, it is complete nonsense to suppose that the situation of
examiner-candidate is representative of ‘real life’ or ‘social communication’;
both participants know perfectly well that this is a test and not a tea-party.
(1967:26)

The use of the paired format instead of, or in addition to, an OPI changes the
institutional asymmetry in a conversation between a candidate and examiner,
but it does not, of course, change the fact that the conversation is part of a test.
Importantly, the move towards pair and group orals has not eliminated the
examiners from the interaction - they still form part of the conversations as
addressees even though they may remain silent (see Study I). The awareness
that the situation has an impact on the interaction has mainly been discussed
in terms of test takers’ anxiety and ‘communicative stress’ (Brown & Yule
1983:34) and possible ways of reducing it. However, the test situation also
influences the discourse produced in the sense that the test situation constitutes
a speech event. Following Hymes, investigating the test situation sheds light
on how a performance is not only about an individual ability but about the
interaction between a person’s competence, the competence of others and ‘the
emergent properties of the event themselves’ (1972a:283). Hence, the co-
construction of the spoken discourse in not only based on the interaction be-
tween the participants, but also in relation to the context. The view of speaking
tests as a social practice and as a speech event thus leads to the need to con-
sider the test situation as a context. This thesis draws on understandings of
context from a dialogic point of view, where context mutually shapes the un-
folding interaction at the same time as it is shaped by it (eg. Goodwin & Du-
ranti 1992, Linell 2009). From this perspective, a closer investigation of the
discourse produced in test settings provides insights into speaking tests as a
speech event.
More generally, speaking tests can be understood as institutional talk
(Drew & Heritage 1992), informed by an institutional goal (here assessment
and grading), carried out by at least one institutional representative (here the
teachers acting as examiners) and entailing possible constraints to the interac-
tion (here the scripted instructions and the ‘rules’ for the performance [see
4.4.2] and the discussion topic given by the examiners, but decided upon by
the national test constructors). As research has shown, the test situation entails
contextual constraints on the interaction, i.e. through the avoidance of negoti-
ation of meaning (Luk 2010, Study II). Drawing on Goffman’s (1959) work
on team performances, Luk (2010) shows how students in a paired speaking
test orient towards positive impression management leading to ‘colluded talk’,

46
where the test takers team up and cooperate through agreements and avoid-
ances of negotiation of meaning. According to Goffman, ‘participants coop-
erate together as a team or are in a position where they are dependent upon
this cooperation in order to maintain a particular definition of the situation’
(1959:96). In the current thesis, contextual constraints and avoidance strate-
gies are discussed in Study II. The avoidance of the use of languages other
than Swedish constructed a monolingual orientation in the test discourse and,
at times, a possible contextual constraint on the meaning making. Investigat-
ing contextual constraints in the interaction in test settings is of central im-
portance, since it is the observations of the performance that underlie the
claims made of test takers’ language ability.
Study I draws on performance theory from linguistic anthropology in order
to analyse the speaking tests as staged institutionalized performances. Perfor-
mance, in the linguistic anthropological sense, is characterized by a responsi-
bility towards the audience for a display of communicative competence (Bau-
man 2004). Moreover, a performance ‘puts the act of speaking on display,
objectifies it’ (Bauman & Briggs 1990:73). Study I confirms Luk’s (2010)
findings, that this display can lead to a consensus-driven interaction achieved
by agreements. This can be viewed in light of the test situation as a particular
kind of speech event in a stressful situation, leading to the importance of sol-
idarity strategies (Luk 2010, Simpson 2006). It is important to note, though,
to reach an agreement can be part of the topic, thus encouraging consensus-
driven interaction (see Appendix A). As discussed in 4.1, other studies have
identified collaborative and cooperative interaction in test discourse in terms
of higher level of interactional competence. On another note, the post-test
interviews with the participants in the current thesis, showed that the test tak-
ers claimed to prefer the paired task over the individual task. One reason for
this was that they felt that they obtained support from each other. Similar find-
ings are reported by Egÿud and Glover (2001) and Fulcher (1996).
Given the fact that many adult language learners have limited experience
in formal schooling, their experience in institutionalized testing varies and can
influence their understanding of the test event. Simpson (2006) investigated
how different expectations of the test event among adult ESOL learners with
low levels of prior schooling led to divergent understandings of the tasks. In
light of this, Simpson argues for the need for explicit training before partici-
pating in tests and for questioning ‘whether it is fair to expect migrant learners
with little or no previous education experience to possess appropriate and ad-
equate frame interpretations for a speaking test. If not, other alternative as-
sessment approaches may have to be explored’ (Simpson 2006:53). Further-
more, the understanding of the speech event and the experience of participat-
ing in it does not only vary amongst students, but also amongst teachers who
act as examiners (Simpson 2006, Sundquist et al. 2018; see section 4.4.1.).
Much of the research conducted on paired and group oral tests is based on
tests in university settings and in relation to academic studies (e.g. Brooks

47
2009, Davis 2009, Ducasse 2010, May 2009, Nakatsuhara 2013, O’Sullivan
2002) or from larger commercial tests such as the Cambridge testing battery,
TOEFL and IELTS or preparation courses for the tests (Galaczi 2008, 2014,
Norton J. 2013). Few studies address language testing in groups of adult mi-
grants in educational contexts (Simpson [2006] being a notable exception).16
It is therefore of importance to investigate interaction in test settings in basic
adult education where the participants have different linguistic, educational
and cultural backgrounds.

4.3 Concluding remarks

As can be seen from the discussion above, language testing, in particular oral
testing, sheds light on the underlying tensions in language assessment with
respect to theoretical, educational and practical issues in ‘measuring language’
(McNamara 1996, Spolsky 1995). A particular tension can be noted between
the view of communicative competence as an individual or a relational con-
struct (see also Studies II and III), also pertaining to the view of language use
as objective and measurable in relation to language as a local and situated
practice. Language tests based on a communicative approach, such as the one
investigated in the present thesis, often stress the ability to adapt language use
to the social context and to the recipient in real-world settings (see section
3.1.1). This leads to the importance of investigating language use in different
settings (e.g. Elder et al. 2017) and the importance of investigating speaking
tests as a speech event (Luk 2010, Study I and Study II). The current thesis
adds new knowledge to these fields in a migration context by investigating
speaking tests as a speech event through the lens of performance theory (Study
I) and by investigating a particular kind of interactional practice in the speak-
ing tests (Study II).

16
Research conducted in relation to the so-called LADO testing (language analysis for the de-
termination of origin of asylum seekers) also provides an exception (e.g. Eades 2009,
McNamara 2012). LADO tests constitute an analysis of asylum seeker’s speech and linguistic
profile in order to determine the asylum seeker’s origin and right to asylum. It is, therefore,
quite different from an educational assessment.

48
5 Analysing spoken discourse: data
presentation, methodological considerations
and theoretical outlines

This chapter describes the data on which the three studies are based on, and
discusses methodological considerations and theoretical outlines for analysing
spoken discourse.

5.1 Overview of data

The present thesis is based on the following sets of data:

a) 27 video-recorded paired speaking tests in the final national test in

SFI (course D) 17
b) 53 audio-recorded post-test interviews with the test takers and ques-
tionnaires with background information
c) 6 audio-recorded focus group discussions with 31 SFI-participants

The participants in the paired speaking tests and the post-test interviews (a and
b) are the same, whereas the focus group discussions were conducted at a later
stage. Hence, c) are with other participants than in a) and b), who had most
likely finished their SFI studies by then. In total, 85 SFI students from seven
learning centres participated. The paired speaking tests comprise approxi-
mately 5.5 hours of recordings, the post-test interviews approximately 6 hours
and the focus group discussions 4.5 hours. Participation was voluntary and the
participants signed a consent letter following the guidelines laid out by the
Swedish Research Council (Vetenskapsrådet 2011). The participants were in-
formed that they had they right to turn off the camera and withdraw their par-
ticipation at any stage.

17
Initially, there were 30 video-recorded paired speaking tests, but three of them were removed
for different reasons; one due to the fact that one of the test takers contacted me afterwards and
wanted her test to be removed from the data set, one due to initial technical problems (two
minutes of sound missing) and one because a test taker was missing, and I was asked to step in
to ‘enact’ his part.

49
 As outlined in section 2.2.2, one of the major characteristics of SFI is diver-
sity. This is reflected in the data. Among the participants, 24 first languages
reported by the participants were represented: Arabic, Chinese, English, Ger-
man, Greek, Hindi, Indonesian, Japanese, Kurdish, Laos, Pare, Farsi/Persian,
Polish, Portuguese, Punjabi, Rumanian, Somali, Spanish, Swahili, Tagalog,
Thai, Tigrinja, Tigre and Urdu. The participants’ time in Sweden varied from
10 months to 20 years, and on average, the participants had been in Sweden
for approximately three years. Their age varied between 18 years and 48 years.
Among the participants, students from all study tracks are represented, and
their prior schooling varies from one year to completed post graduate degrees.
All grades are represented in the data.18 The large majority of the students
passed the test (45 students out of 53, i.e. 85 %).19

5.1.1 The paired speaking tests

The oral component of the national test consists of two tasks, one individual
task testing the criteria for oral production and one paired task testing the cri-
teria for oral interaction. The focus in the current thesis is on the paired task.
Even though the data reflects tests based on different syllabi (National Agency
for Education 2009, 2012a), the overarching aim of oral interaction and the
paired format has more or less remained the same. As stated in the syllabus
from 2009, the aim of oral interaction is that ‘the student should be able to
participate in both informal and more formal situations in daily, societal and
working life’, which was reformulated in the syllabus from 2012 to ‘the stu-
dent can, with some adaptation to purpose and their interlocutor, communicate
in both informal and more formal situations in daily, societal and working life’
(National Agency for Education 2012a).20 Meanwhile, these changes have not
affect the paired format in the test, which has remained the same. The fact that
the national test has a theoretical framework based on research in applied lin-
guistics (Ahlgren 2016, Ahlgren et al. n.d., see also section 3.1.1) probably
makes the test formats in use less susceptible for the recurrent changes in the
syllabi described in 2.2.1. The assessment grid accompanying the is based on
a holistic assessment, targeting features such as adaptation to purpose, content

18
Two grading systems are represented in the data: G (pass)/VG (pass with distinction) and A-
F (A excellent – F fail). The syllabus in use from July 2012 introduced the A-F system, but
students who had enrolled in SFI while the previous syllabus was in use had the right to follow
that syllabus. Each component of the test I graded. The test is then graded based on a overall
assessment of the grades given to each of the components.
19
This corresponds quite well with the attainment on the oral component (course D) of the tests
given between 2009 and 2012, which were reported to be 82 % (Stockholm University & Na-
tional Agency for Education 2013).
20
The only difference between the syllabus from 2012 and the current syllabus is that the syl-
labus from 2017 has added ‘continuing studies’ to the domains (National Agency for Education
2017a).

50
and range, comprehensibility, fluency and an analytical assessment including
aspects such as interactional management (including strategies to solve inter-
actional problems), pronunciation and variation and accuracy in vocabulary
grammar (National Agency for Education 2006b, 2012b). The focus of this
thesis is not on the teachers’ assessment but on a sociolinguistic understanding
of the test situation and of the interaction produced in the tests.
As described in section 2.2.2, due to an open enrolment policy, SFI has no
fixed semesters. Instead, students begin and complete the SFI courses contin-
uously during the year. As a consequence, the national test is given on a re-
current basis throughout Sweden with varying numbers of students on each
test date. Another consequence of the open enrolment policy is that there are
several different test versions in use. Since the national test is a recurrent ac-
tivity, administered all over Sweden regularly during the year, I wanted the
data to reflect this practice. To this end, I collected data from different learning
centres, that had administrated the test on different occasions, in different parts
of Sweden. Another reason for not collecting all data at once was the fact that
I considered the first part of the data collection a pilot study. Subsequently, I
started analysing the first data set before I continued to collect more data.
Table 1. Overview of the test data.
Learning centre Date Different test Number of video-
versions in use recorded tests
A (pilot) November 2011 a 1
B (pilot) January 2012 b 1
C (pilot) February 2012 a 4
D December 2012 c 8
C December 2012 c 4
E February 2013 a 1
F March 2013 d 2
G June 2013 e 6
Total: 7 Total: 5 Total: 27

As shown in Table 1 above, the test data consists of 27 paired speaking tests,
stemming from seven learning centres, collected between 2011 and 2013 and
comprising five different test versions and 12 topics. Hence, what brings the
data together is neither the place, the time, the participants nor the discussion
topics. Rather, the link between the different paired speaking tests is the fact
that they are understood and enacted as paired speaking tests. The number of

51
speaking tests video-recorded in each learning centre varied depending on
how many tests were scheduled on the particular test occasion and how many
students and teachers were willing to participate. At some of the learning cen-
tres, I was only able to film one test, either because only one test was sched-
uled or there were no more willing participants, whereas in another learning
centre, I could record all the speaking tests planned that day. What is more, in
some learning centres, the teachers were more willing to give me permission
to record when the number of test-taking students was quite small.
Each paired speaking test lasted between 7 and 17 minutes. The official
recommendation is for the test to last approximately 15 minutes, including the
examiner’s instruction, and to conduct the test with two examiners present
(Ahlgren et al. 2013, National Agency for Education 2012b). This was the
case at four of the learning centres (A, B, C, D), whereas in three of them (E,
F, G) there was only one examiner present.
In the paired task, the test takers are provided with a given topic to discuss.
Accompanying each test, are four to five topics from which the examiner can
choose. At the start of the exam, the examiner is supposed to read a scripted
instruction (Appendix A, my translation)21, that goes as follows:

You should discuss for about 10 minutes. It’s important that you listen to each
other, ask each other questions and ask each other if you don’t understand. Keep
in mind that the discussion is between the two of you. I will only listen. What
you are going to discuss is this question (hand out the task sheet).

Hence, the written instruction puts the emphasis on staged interaction, i.e. that
the students should give a ‘team performance’ in front of the examiners
(Goffman 1959; see also Study I). In the collected data, the instruction could
be performed differently, but all examiners emphasized that the test takers
should discuss with each other (see also Study I, pages 539-540 for an illus-
tration of how the metacommunicative frame of the paired speaking test could
be negotiated). Since students enrolled in SFI represent a highly diverse group,
the topics tend to be quite general, where the test takers are supposed to dis-
cuss a dilemma and possible solutions, evaluate different statements or to en-
gage in a decision-making task (Ahlgren 2016, National Agency of Education
2012b; see also Appendix A.). The test takers are provided with a mind map,
or other visual or written support for the topic, which the test takers can use if
they need support. In addition, the examiners are encouraged to ask follow-up
questions to support the test takers’ conversations if necessary.
The paired speaking tests were video-recorded using one camera. The cam-
era was placed dialogically behind the examiners, since the test takers often

21
This written instruction was introduced in 2012, but instructions to the examiners accompa-
nying prior tests foregrounded the same aspects. As of 2012, the role of the examiner as a silent
partner have been further accentuated. Overall, in the collected data 2011-2013, examiners
stressed their role as silent audience.

52
turned to examiners, and placing it behind the test takers may have resulted in
their turning away from the camera.

Figure 3. Illustration of a paired speaking test.

From this position, the camera’s angle represents the examiners’ perspective,
i.e. the main audience, rather than that of the test takers. The speaking tests
took place in regular classrooms or in smaller group rooms at the learning
centres. The size of the room also had an impact on how I could place the
camera. As soon as I had started the camera, I left the room, as I wanted to
minimize the effect of my presence. In most of the speaking tests, there were
two test takers and two examiners. If I had stayed in the room, my presence
would have resulted in the institutional representatives outnumbering the two
test takers. A consequence of leaving the room before the test started was that
I had no control of whether the participants moved out of the camera angle.
This happened on a few occasions, as one of the examiners in the room moved
to create distance with the test takers.
The preliminary analysis of the pilot study showed that embodied semiotic
resources like gaze direction and gestures were very important in the partici-
pants’ negotiation of the interactional frame. Thus, the main reason for choos-
ing to video record was to capture embodied and multimodal aspects of the
interaction. It is important to note, however, that a camera does not capture
everything. Rather, the video represents a certain angle and ‘gaze’, which sub-
sequently impacts the analyst’s viewpoint. Moreover, video-recording can be
problematic. The main objection to using a camera is that it creates an intru-
sion that can possibly have an impact on the participants. This is particularly
relevant in high stakes settings where decisions are made about the partici-
pants. The camera was only on during the paired task and not during the indi-
vidual task in order for the examiners and the test takers to have one task that

53
was not video-recorded. In a couple of tests, before the test officially ‘started’,
there were instances of ‘camera behaviours’ (Duranti 1997:118), where the
participants commented on the camera’s presence (cf. Åhlund 2015:39).
Meanwhile, in the post-test interviews, I asked whether the test takers had
thought of the camera during the test. With two exceptions, all participants
claimed to have forgotten about the camera when the test started.

5.1.2 The post-test interviews

The aim of the post-test interviews was to collect background information
about the students and to capture their experiences and view of the speaking
tests. The post-test interviews were conducted immediately after the paired
speaking tests, often in the hall outside the classroom where the speaking tests
took place. The national test is very time consuming, both for teachers and for
students. In general, the learning centres cancelled all other teaching during
the test days. Hence, most of the time, it was calm in the hall outside the class-
rooms, but during breaks it could be quite noisy. The interviews were con-
ducted in Swedish (see section 5.1.3 for a discussion of the use of Swedish in
interviews and focus group discussions).
Background information (e.g. time in Sweden, prior schooling and initial
study track) was collected through a questionnaire and during the interview.
The interview was conducted as a structured interview with scripted standard-
ized questions as a starting point, but there was also room for follow-up ques-
tions if necessary. The interviews addressed questions such as how the partic-
ipants had experienced the test and the discussion topic, how they had pre-
pared for the exam and what they thought of the two test formats in use – an
individual task and the paired task. All test takers participated in these inter-
views except one, who did not have time to stay after the test.
In the current thesis, the post-test interviews have mainly been used as
background information about the participants and to account for their expe-
riences with the speaking test (see section 4.3.1; see also Rydell 2016).

5.1.3 The focus group discussions

The focus group discussions were conducted in a later phase. Even though the
focus group discussions took place in a less regimented space than testing,
they still form part of an institutional and educational discourse, as they took
place at the participants’ regular learning centre. The focus group discussions
took the form of conversational group interviews with the researcher as mod-

54
erator. During the discussions, the SFI participants had several addressees, in-
cluding me as the researcher and their fellow students, and the discussion fo-
cused on both individual as well as shared experiences (e.g. Wibeck 2010).

Table 2. Overview of the focus group data.

Groupe Learning Date Number First languages Time in
centre of par- Sweden
ticipants
1 G December 5 English, Kurdish, 2–8
2015 Somali years
2 G December 6 Tigrinja 1–2
2015 years
3 G December 5 Arabic, Kurdish, 1–8
2015 Laos, Somali years
4 G December 4 Arabic, Tigrinja, 2–4
2015 Turkish years
5 A February 5 Arabic, Japanese, 8
2016 Rumanian, So- months–
mali 20 years
6 A February 6 Arabic, Tigre, Ti- 2–7
2016 grinja years

Four group discussions were conducted in December 2015 at one learning

centre in a medium-sized town in central Sweden (learning centre G), and an
additional two focus group discussions were conducted in February 2016 in
an inner-city learning centre in the capital of Sweden (learning centre A).
Both learning centres were part of the test data, but the SFI participants are
not the same. The discussions lasted between 40 to 60 minutes, with five to
seven participants in each group.22 The participants were either students at the
final course in SFI (course D) or they had recently finished course D and were
taking vocational training courses together with the SFI group. I had no pre-
vious acquaintance with the students. The group discussions took place in
group rooms, in a classroom and, for one group, in the teachers’ recreation
room (without teachers being present). The focus group discussions were con-
ducted as semi-structured group interviews, based on scripted questions as a
starting point for the discussion, but allowing for follow-up questions and for
directions initiated by the students (see Appendix B). The questions addressed
such issues as what it means to be good at a language, how one knows if one
is good at a language, how one learns a language and how they assessed their
own linguistic resources.

22
See Study III for a more detailed description of the participants who were in focus in that
study.

55
 The group discussions were conducted in Swedish, my first language and
the second language of the participants. Using a language other than the par-
ticipants’ first language is, by and large, not recommended in focus group dis-
cussions (Stewart et al. 2007). Nevertheless, in this study, Swedish was used
for both practical and methodological reasons. Swedish is the lingua franca
used in class and in the participants’ everyday lives. The use of Swedish as an
important lingua franca was particularly stressed by students who did not
speak English. The 31 participants in the focus group discussions represent a
diverse group, with nine first languages represented. Therefore, it would have
been difficult to engage enough interpreters. Furthermore, several participants
displayed a negative stance towards interpreters (see Study III). From a meth-
odological point of view, it was important to treat the participants as legitimate
speakers of Swedish. After the group discussions, several of the participants
asserted that they enjoyed the opportunity to practice and use their Swedish
(see also Ahlgren [2014] for a similar line of argument and experience). On
the whole, they had developed oral skills and were able to communicate in
Swedish. Undeniably though, the participants had differential access to lin-
guistic resources in Swedish, and some of the discussions flowed more easily
than others. Importantly, the participants would probably have expressed
themselves differently if the discussions were conducted in their first lan-
guage.
Bourdieu has argued that research interviews are characterized by an im-
portant asymmetry where:

The market of linguistic and symbolic goods, which is set up in each interview,
varies in structure according to the objective relation between the investigator
and the investigated, or which is the same thing, between capitals of all kinds –
especially linguistic capitals – with which they are endowed. (1996:19)

From this perspective, the choice of interview language has a symbolic func-
tion. Following Bourdieu (1996), it remains important to reduce what he calls
the symbolic violence inherent in interviews. For this reason, I chose to con-
duct focus group discussions (rather than individual interviews), where the
SFI students were in the majority and where they could get support from each
other (cf. Kvale & Brinkmann 2014:191) and had the opportunity to translate
for each other. The local teachers helped me arrange the groups so that there
would be, when possible, at least two speakers with the same first language.
The participants were told to help each other translate when necessary and did
on a few occasions translate to Arabic and Somali. English was also used a
couple of times by me and some of the participants.

56
 5.2 Analysing spoken discourse and communicative
experiences: methodological and theoretical
outlines
As outlined in previous sections, the data that the three studies are based on
are video recorded paired speaking tests (Studies I and II) and audio-recorded
group discussions (Study III). The starting point for the analysis was repeated
listening and watching (e.g. Cameron 2001). A second step consisted of tran-
scribing the spoken discourse. All recordings were transcribed, and the se-
lected excerpts were transcribed in more detail (see section 5.2.1 for a meth-
odological and theoretical discussion on transcription). The analytical focus is
on situated interaction and metapragmatic discourse in relation to communi-
cative experiences and perceptions and constructions of language competence.
Metapragmatic discourse is defined in the ‘ways in which people think and
speak about language and language use; about what language and communi-
cation is or ought to be; and what the reasons for this might be’ (Karlander
2017:28). The three studies have different research aims and analytical foci
(see chapter 6 for an extended summary of the studies). However, all three
studies are united by a common core of theoretical assumptions. Following
Jones (2016:25), ‘the study of spoken discourse is an interdisciplinary project
that draws on fields such as linguistics, philosophy, anthropology, sociology
and critical theory’.
This thesis is based on a dialogic view of language (Bakhtin 1981, 1986,
Linell 2009, Vološinov 1973) and phenomenological approaches to the lived
experience of language (Busch 2017, Merleau-Ponty 2012[1945]). Language
is understood as action and as a fundamentally social phenomenon where ‘lan-
guage and the world are everywhere interconnected’ (Hanks 1996:119). So-
cial reality and meaning making is created, maintained and reproduced in lan-
guage use. In this light, language is relational. Following Linell ‘[d]ialogism
assumes that others are co-constitutive of selves; others are involved at all
levels of social life and individual lived experience’ (2009:42). This resonates
with phenomenological thinking. From a phenomenological point of view, our
understanding of language is based upon our experiences and how we, through
speech, learn to think ‘according to others’ (Merleau-Ponty 2012[1945]:184,
italics removed). By contrasting the Cartesian view of private selves, Merleau
Ponty points to how being in the world entails being ‘outside ourselves in the
world’ and how ‘life-world appears between the subject and the world within
the focus of perception’ (Ihde 1973:67). Accordingly, the participants’ per-
ception of competence is formed by experiences of being a social actor in the
world and in relation to other social actors (see Study III).
For Merleau-Ponty, language is primarily a bodily phenomenon. Moreover,
contrary to the Cartesian dualism of body and mind, Merleau-Ponty conceived
of the lived body (corps vivant) as a unit, simultaneously perceiving and being

57
perceived, both an object and a subject. Placing the body at the centre of hu-
man action leads to a view of language as an embodied social action drawing
on different semiotic resources that make the action visible and subject to the
interpretation of others (e.g. Block 2014, Goodwin 2000, 2013, Mondada
2016). Language as an embodied practice is investigated to different degrees
in the three studies (see 5.2.1).
Meanwhile, the analysis of verbal interaction is still central in all three stud-
ies. It follows that the utterance is a central unit of analysis, together with
addressivity and intersubjectivity. An utterance should not be analysed in iso-
lation but rather in relation to its sequential position, i.e. what previously has
been said and done and what comes after. The importance of analysing turn-
taking within its sequential position in the unfolding interaction has been put
forth by Conversation Analysis – CA (e.g. Sacks et al. 1974, ten Have 2007).
Moreover, following a CA tradition of analysing L2 spoken discourse, the an-
alytical focus is on what the participants are able to say and do with the lin-
guistic resources at their disposal, not on their deficiencies (cf. Firth & Wagner
1997, Gardner & Wagner 2004). In this light, the studies investigate interac-
tion as co-constructed and as cooperative action (e.g. Goodwin 2000, 2013).
Intersubjectivity holds a central position in both conversational analysis
and phenomenological work. In CA work intersubjectivity refers to an estab-
lishment of a sufficiently shared understanding that allows for the progressiv-
ity of talk (e.g. Heritage 2007), whereas in phenomenology intersubjectivity
is a broader concept referring to the ‘possibility of being in the place where
the Other is’ (Duranti 2010:16). In this sense, intersubjectivity is a fundamen-
tal aspect of all human experiences (Duranti 2015). In this thesis, both under-
standings are used in the study of situated interaction (Studies I and II) and in
the analysis of the reflections of lived experiences of language (Study III).
Overall, the studies are ‘engaged but not identified with CA’ (Rampton
2017:7).
Furthermore, another concept that is important for the analysis of the utter-
ances, is the Bakhtinian notion of addressivity (1986). This notion foregrounds
how all utterances are addressed to someone and how the addressees are con-
stituitive of what is being said. In particular, this is important for the analysis
of the test discourse, where the test takers have multiple addressees, and where
the institutional representatives, i.e. the examiners, remain silent. It follows
then that silence is also important for the understanding of the speech event
(see also Hanks 1996:112).
Studies I and II set out to investigate speaking tests as a social practice and
as ‘a speech event in its own right’ (Galaczi 2008:92). The starting point for
analysing the context, i.e. the test situation, is ‘from the perspective of the
participant(s) whose behaviour is being analyzed’ (Goodwin & Duranti
1992:4), and where the context both shapes and is shaped by the interaction.
The speech event is thus analysed interactionally together ‘with reference to

58
the whole social context in which they occur’ (Cameron 2001:52), in particu-
lar in relation to language ideologies of adult migrants’ language competence
(Study I). Following Bakhtinian dialogism (Bakhtin 1981; 1986) an utterance
can never be analysed and understood as a single unit belonging to an individ-
ual, since ‘[e]very conversation is full of transmissions and interpretations of
other people’s words’ (Bakhtin 1981:338). What we have heard becomes a
part of what we say, and accordingly dialogue is both situated and transcend-
ing situations and becomes ‘constitutive of consciousness, society, and cul-
ture’ (Bauman 2004:5). This pertains to what Linell has termed double dia-
logicality, which ‘makes us see an act or utterance both in its singularity and
its wider sociocultural belongingness’ (2009:53). From this perspective,
Study I discusses how dominant discourses become reproduced in the situated
test talk, and Study III addresses how communicative experiences and the em-
bodiment of the voices of others become constitutive in the participants’ per-
ceptions of competence.

5.2.1 Transcribing and translating L2 spoken discourse

Working with recordings necessarily involves transcribing at different stages
of the research process. During this process, the act of transcribing serves dif-
ferent purposes. It is both a crucial analytical step in interpreting the spoken
discourse as well as an important means of representing the data to readers
(Bucholtz 2000, Ochs 1979). Transcribing talk is not only a technical endeav-
our, and there is no objective way of transcribing talk. Instead, transcribing is
an act pertaining to theoretical, methodological, esthetic and ideological issues
(Bucholtz 2000, Linell 2005, Ochs 1979). As stated by Duff and Roberts , ‘[i]f
talk is a social act, then so is transcription’ (1997:167). The goal, then, for the
researcher is not to try to be neutral, but to be reflexive (Bucholtz 2000).
In essence, transcription is concerned with transforming embodied spoken
language into a written medium. In this vein, a transcription is a written text
adhering to the original spoken discourse to different degrees (e.g. Linell
2011). Speech is multidimensional and multimodal, and humans generally
draw simultaneously on several embodied semiotic resources in communica-
tive practices. The richness of spoken discourse entails consequences for the
act of transcribing. First and foremost, transcription is a selective process
(Ochs 1979). Thus, a central task is to decide what to represent in the transcript
and how (Bucholtz 2000, Linell 2011). These choices should be guided by the
analytical focus. Meanwhile, most transcripts end up being a compromise

59
(Duranti 1997) between, on the one hand, ‘faithfulness’ to the original embod-
ied interaction and, on the other hand, accessibility and being reader-friendly.
Multilayered transcriptions that are too fine-grained, in particular, struggle in
terms of accessibility and readability. In addition, multimodal transcripts pre-
sent challenges with respect to temporality and selection of images. What is
more, there is no consensus on how to do multimodal transcripts (Mondada
2016).
It is important to note that there is no such thing as a final transcript (Du-
ranti 1997). One way to be reflexive about transcription is to do different kinds
of transcriptions through ‘a layered approach’ offering different readings’
(Roberts & Duff 1997:170). As argued by Linell (2005), linguistics is deeply
influenced by a written language bias, which, inter alia, is manifested in the
assumption that the written language can represent speech (Bucholtz 2000).
Another consequence, is that spoken discourse in transcriptions is given a
seemingly static and structural look, making it difficult to ‘capture the ambi-
guities and indeterminacies, which were there to be negotiated and temporar-
ily resolved in the actual communicative activity’ (Linell 2005:32). Hence,
rather than representing language as a fixed object (and being consistent
throughout the three studies), I have chosen to use somewhat different ways
of transcribing in the three studies with respect to the different research objec-
tives of the studies. In particular, the three studies draw on interactional anal-
yses represented in different types of transcriptions with respect to detail level
and representation of multimodal aspects. In Study I, the focus is mainly on
what the participants said and how they co-constructed different stances. Em-
bodied resources like nods were marked in the transcript; otherwise the tran-
scripts from Study I focus on verbal interaction. Study II addresses how em-
bodied semiotic resources like gaze, gestures and body posture are used to-
gether with speech. To this end, the transcripts used in Study II are based on
a multimodal approach (e.g. Goodwin 2013). Study III focuses on what the
participants say in the focus group discussions and, to some extent, how this
is said (in particular, how different prosodic means are used to embody and
contrast different voices). The transcripts draw on conventions developed in
CA research and multimodal CA (e.g. Goodwin 2013, ten Have 2007).
How to transcribe L2 talk has in many ways been treated as a controversial
topic, especially concerning how the speakers are represented. It is commonly
argued that representing L2 features, like deviations in grammar and idiomatic
expressions, runs the risk of stigmatizing the speakers (Bucholtz 2000, Rob-
erts & Duff 1997). From this perspective, it is tempting to standardize, or as
the term goes, ‘clean up’, the L2 speakers’ language (cf. Bucholtz 2000).
However, engaging in correcting non-standard varieties is an ideological pro-
cess as well, since it leads to the erasure of both non-standard varieties and the
linguistic diversity among the participants (Bucholtz 2000). Such erasure is
itself a language ideological process (Irvine & Gal 2000) and can be seen in
the light of ‘verbal hygiene’ (Cameron 1995). Moreover, the link between a

60
non-standard L2 variety and stigmatization is not a direct causal relation, but
is created and maintained by an indexical link between a particular way of
speaking and a social status (e.g. Jaffe 2016). In contemporary globalized so-
cieties, it remains crucial for people to challenge these kind of inferences. In
addition, studies on second language interaction have shown that deviations
in relation to the standard rarely stand in the way of meaning making and that
the participants rarely orient to ‘deviations’ (e.g. Canagarajah 2013, Gardner
& Wagner 2004, Sundberg 2004).
Although it is important to be aware of the fact that non-standard varieties
are often stigmatized, it was a methodological choice to treat the participants
and their speech as legitimate, despite not being proficient. To this end, the
analyses in the three studies are based on the participants’ own words and
ways of expressing themselves (see also Ahlgren [2014] and Zachrison [2014]
for similar lines of argumentation). From this perspective, I bring my own
ideology to the transcription (Duff & Roberts 1997). This is not to say that all
transcripts of L2 spoken discourse should be conducted in this way. But in this
particular study, with its focus on perceptions and constructions of compe-
tence, it was important to adhere as much as possible to the speakers’ own
ways of expressing themselves and not correct their speech.
One crucial problem, though, is that written language is underdeveloped
in its capacity to convey nonstandard varieties (Bucholtz 2000, Duranti 1997).
The crux of the matter lies with ‘dealing’ with L2 varieties in an ideological
context where standard language and correctness have a strong symbolic
value. I have retained L2 features like non-idiomatic expressions, incongruen-
cies and deviations in word order. I have, to a large extent, used standard or-
thography and not marked deviations in pronunciation. The translations into
English adhere as much as possible to the Swedish originals (cf. Åhlund &
Aronsson 2015). Since Study III focuses more on what is said than how, the
transcripts and the translations have been more modified than in Studies I and
II (see Study III). All three studies represent both the Swedish original as well
as the English translation. As Duff and Roberts argue (1997), representing
both the original version together with the translation is a way to capture, in
Bourdieu’s terms, the ‘whole social person’ (1991:54). Moreover, only using
the English translation would make ‘a significant ideological statement about
the power of English to represent everyone and everything’ (Duff & Roberts
1997:170).
All names are anonymized in the transcripts.

5.2.2 Notes on reflexivity

When I started this research project, I was, in several ways, close to the activ-
ity of language testing in SFI. As a former SFI teacher and a test constructor
for the final national test in SFI, my research interest at the very beginning

61
was very much practice-oriented. This interest could be paraphrased as re-
search questions concerning ‘how are speaking tests enacted at the learning
centres?’ This interest was later developed into more theoretical approaches
of understanding paired speaking tests as a social practice, how perceptions of
competence are constructed during, and in relation to, language assessment
and testing practices and the language ideological part of adult migrants’ lan-
guage competence.
Being an insider has both advantages and disadvantages. As an insider, it
is often easier to gain access to the field of interest. Moreover, being close to
the object of inquiry provides an important understanding of the practice of
interest and what is at stake for different stakeholders. However, being an in-
sider is also problematic, since it often leads to a position where practices and
underlying assumptions are taken for granted (Cameron 2001:57). As such,
the insider runs ‘the risk of not seeing the viewpoint from which it [the view-
point itself] is stated, and thereby producing an account which says exactly
what the researcher’s position in the field allows him or her to say – and noth-
ing else’ (Salö 2018:27). Therefore, a crucial step in the research process is to
create a distance in order to endorse critical thinking (Carlson 2002). This can
be achieved by reflexivity in the research process (e.g. Ahlgren 2014, Bour-
dieu 1996, Bourdieu & Wacquant 1992, Salö 2016). One way to be reflexive
in research is to describe, as explicitly as possible, the collection of data and
the choices made in the analysis (Ahlgren 2014), which I have tried to do in
this chapter. Thus, the aim is not for the researcher to achieve self-effacement
but ‘vigilant self-awareness’ (Bucholtz 2000:1461). Importantly, though, re-
flexivity is not only about an individual’s self-awareness in relation to the re-
search process. Rather, reflexivity should be understood as a process of scru-
tinizing the position from which the research object is constructed and as col-
lective, rather than individual, reflexivity. Drawing on Bourdieu’s notion of
‘epistemic reflexivity’ and Bachelard’s ideas of rupture, Salö (2016, 2018)
discusses how reflexivity implies a rupture with commonsensical understand-
ings and pre-given constructions of the research problem. By contextualizing
and historicizing SFI, perceptions and constructions of competence and lan-
guage assessments as a social practice, I have aimed to create an understand-
ing of the presuppositions made in the fields (adult education and language
education) that I was invested in when I started the research process.

62
6 Summary of Studies

Taken as a whole, the current thesis explores sociolinguistic perspectives on

language testing and assessment, on the one hand, by investigating paired
speaking tests as a particular kind of social practice and, on the other hand, by
investigating how perceptions of what it means to be ‘a competent language
user’ are constructed - interactionally and metapragmatically - in test dis-
course and in focus group discussions during reflections on communicative
experiences. Instead of treating competence as a fixed object, ‘there’ to be
studied and described, the three studies approach perceptions and construc-
tions of being ‘a competent language user’ from different perspectives, as out-
lined below. Studies I and II draw on an interactional analysis of 27 video-
recorded paired speaking tests in the final national test of SFI, whereas Study
III builds upon an analysis of focus group discussions with SFI participants.

6.1 Study I – Performance and ideology in speaking

tests for adult migrants
Study I is motivated in a sociolinguistic interest in the test situation and the
focus is on how test takers stage being ‘a competent language user’ in paired
speaking tests and how they express their beliefs with respect to language
competence, language use and language learning in an interactional context
where their language ability is being assessed. In order to create a framework
for analysing paired speaking tests, Study I draws on performance theory from
linguistic anthropology (e.g. Bauman & Briggs 1990, Bauman 2004), and the
Bakhtinian notion of addressivity (1986). As Bauman and Briggs state in their
seminal paper on performance, ‘[p]erformance puts the act of speaking on dis-
play – objectifies it’ (1990:73), and following Bauman, performance ‘resides
in the assumption of responsibility to an audience for a display of communi-
cative competence, subject to evaluation for the skill and efficacy with which
the act of expression is accomplished’ (2004:110). This is exactly what hap-
pens in language assessment, in general, and in paired speaking tests, in par-
ticular – language is objectified, and evaluating the candidates’ communica-
tive competence is part of the teachers’ professional practice. The paired
speaking test is a complex interactional setting, where the test takers have

63
multiple addressees. They are simultaneously addressing each other and the
examiners who are listening and assessing the conversation.
Performance studies have mainly focused on cultural performances like
storytelling and verbal art (e.g. Bauman 2004; Briggs 1988; Hymes 1975),
rituals (e.g. Schieffelin 1985) and artistic performances (e.g. Bucholtz &
Lopez 2011). Common to all descriptions of performance is a view of perfor-
mance as an audience-oriented discourse. What follows from this, then, is that
inherent in performance is the risk of failure and of losing face (Howe 2000).
A crucial difference in this context, though, is the audience’s (i.e. the examin-
ers’) silence and the fact the evaluation of the test takers performance is not
given directly afterwards, i.e. the test takers’ do not know whether they have
passed or not at the end of their exam. By using the notion of addressivity,
Bakhtin (1986) points to how an utterance is always addressed to someone,
and the addressee(s) are constitutive of the utterance. Thus, even though the
examiners remain silent to a great extent, they form an important part of the
conversation as addressees. The institutionality and the power imbalance be-
tween performers and audience adds a layer to this particular kind of perfor-
mance, and foregrounds the vulnerability of the performers. In this light, Study
1 adds to the body of research on performance by investigating institutional
performance.
A general challenge for the test takers is to keep the discussion going in
front of the examiners. The main argument put forth in Study I is that speaking
tests should be analysed as staged institutional performances where not only
speaking, but also ideologies, are put on display. Examining embedded ideo-
logies in language tests provides insights into testing as a social practice
(Spotti 2011). As Study I shows, language ideologies become embedded in
testing practice and are manifest in the performance, making testing an im-
portant site for the reproduction of dominant language ideologies. One im-
portant resource for the test takers to keep the discussion going and thereby
assume responsibility for the audience was to draw on dominant discourses on
language and integration. During the test, the test takers stated the importance
of learning Swedish and expressed how learning ‘good Swedish’ is achieved
through a preferred behaviour by, for instance, only speaking Swedish, mak-
ing an effort to find Swedish friends, avoiding people of the same nationality
and taking responsibility for one’s success in studying. In this sense, the ori-
entation to being ‘a competent language user’ was performed by indexing
other images of being ‘a good student’ and ‘a good immigrant’.
Moreover, the role of the standard language and the need to learn it was
given an iconic (Irvine & Gal 2000) status as something necessary in the
course of the interactions. A corollary to the claimed importance of the stand-
ard language is the celebration of linguistic correctness and, contingently, the
treatment of linguistic diversity in the student group as problematic. Hence, in
the test taker’s display of being ‘a competent language user’, a monolingual

64
ideology was indexed by expressing loyalty to the standard language and lin-
guistic correctness.
The view of successful language learning thus has moral implications when
linked to a preferred behaviour, both inside and outside the classroom. Argu-
ably, entrenched ideas of the communicative teaching paradigm and of SLA,
i.e. that a language is learned by using it and the importance of exposure to it,
need to be treated with sociolinguistic awareness. Otherwise, these pedagogi-
cal beliefs will reinforce an un-reflected strong monolingual ideology where
second language speakers should only speak Swedish and avoid using other
languages. Subsequently, then, migrants’ ‘lack of language skills get conflated
with lack of appropriate cultural and behavioural attributes’ (Allan 2013:57).
Inherent in institutional talk are constraints on what is considered to be an
appropriate contribution (Drew & Heritage 1992:22), and the orientation to
dominant discourses could be seen, as it were, as ‘safe topics’. Hence, these
beliefs do not necessarily need to be interpreted as reflecting the views of the
test takers themselves; rather, as Goffman holds, ‘the performance serves
mainly to express the characteristics of the task that is performed and not the
characteristics of the performer’(1959:83), i.e. that they reveal something
about testing as a social and institutional practice. The orientation to dominant
discourses can also be understood from a dialogic point of view. According to
Bakhtin (1981), internalizing others’ discourses and others’ words are part of
all meaning-making activities and part of ‘our ideological interrelations with
the world’ (Bakhtin 1981:342). From this vantage point, the test discourse re-
vealed an ongoing dialogue with current values and ideologies in society.
These enunciations were co-constructed by the test takers through the use of
agreements and affiliation tokens in the interaction (such as statements like I
agree and through embodied actions such as nodding). Performance is an in-
teractional accomplishment, and the analysis in Study I shows how acts of
stance-taking and interactional alignments (like agreeing with each other) con-
stitute important resources for accomplishing the performance.
Study I also has implications for testing practices. At the heart of commu-
nicative language testing lies observations and interpretations of what the test
takers are able to do with their language. Communicative language testing
struggles with ideologies of authenticity, i.e. how to create ‘authentic tasks’
that generate ‘authentic language use’. Likewise, a common critique to the
paired format is that it entails ‘unauthentic interaction’ since it is staged. As
stated by Luk, for test takers ‘staging a performance to present the best possi-
ble impression of themselves as interlocutors in front of the examiner might
have resulted in a form of ritualized and colluded talk that does not represent
an authentic replica of ordinary conversations’ (2010:49f.). However, instead
of only looking at linguistic features in the spoken discourse and treating test
talk as ‘unauthentic’, performance should be seen as an interactional accom-

65
plishment in its own right (Bauman 2004), where assuming responsibility to-
wards an audience and managing the examiners’ silence is part of what the
test takers are able to do.

6.2 Study II – Negotiating co-participation:

Embodied word searching sequences in paired L2
speaking tests
Study II pursues the theme of investigating speaking tests as a social practice
and as a speech event. In Study II, this is achieved through the lens of a par-
ticular interactional practice – the activity of searching for a word. Vocabulary
knowledge is often used as an index of competence and as a salient feature for
distinguishing between different L2 proficiency levels (Iwashita et al. 2008).
At the same time, the ability to negotiate meaning (i.e. through the use of ges-
tures, confirmation checks, clarification questions, etc.) and manage commu-
nicative breakdowns has been foregrounded in communicative language test-
ing (Canagarajah 2006, Harding 2014). Word searching sequences thus tap
into several aspects of communicative language ability: vocabulary
knowledge, the ability to negotiate meaning and to solve interactional prob-
lems. As discussed in Study I, a main challenge for the test takers in the paired
speaking test is to keep the discussion going. As a ‘performance team’
(Goffman1959), the test takers depend on each other in order to display and
maintain the image of being ‘competent language users’ in front of the audi-
ence (i.e. the examiners). Hence, a potential critical interactional moment is
when the test takers encounter troubles in finding a word or knowing how to
continue the conversation. Moreover, there is a potential tension, in this par-
ticular setting, between the display of lexical precision and ensuring the pro-
gressivity of talk. In this light, the aim of Study II is to explore how embodied
semiotic resources are used to negotiate meaning and co-participation in word
searching sequences in an interactional context where the participants’ lan-
guage ability is being assessed. The focus in Study II is on extended word
searching sequences or sequences where the speaker actively searches for the
interlocutor’s assistance.
Word searches are a common interactional practice occurring many kinds
of interaction. Word searches have mainly been studied in L1 talk (Goodwin
& Goodwin 1986, Hayashi 2003), L2 talk (predominantly L1-L2 talk, e.g.
Brouwer 2003, Gullberg 2011, Kurhila 2006) and in aphasic speech (e.g.
Helasuvio et al. 2004). Importantly, searching for a word is not only an inter-
nal process. Rather, the use of different embodied semiotic resources (like
gaze and body posture) makes word searching a visible activity manifested by
both verbal and non-verbal features, which can invite others to participate

66
(Goodwin & Goodwin 1986). It is common for word searches to contain two
phases: first, a self-directed phase (often made visible through gaze aversion),
and then, if the speaker does not find a solution, an other-directed phase invit-
ing the interlocutor to join in the word search (i.e. by looking at the interlocu-
tor and using gestures). Word search behaviour has been found to be more or
less similar across languages and between first and second language speakers
(e.g Hayashi 2003, Gullberg 2011), but with respect to L2 talk, word searches
can be more frequent and elaborated (Gullberg 2011). Moreover, the fre-
quency can be dependent on the interactional setting and the discourse pro-
duced.
Study II shows, in line with previous research (e.g. Goodwin & Goodwin
1986, Greer 2013, Gullberg 2011, Hayashi 2003), how embodied semiotic re-
sources constitute a fundamental aspect of word searching sequences that
make word searches public. The investigated word searching sequences in
Study II were performed differently. As a visible practice, word searching se-
quences are subjugated to the interpretation of others and the negotiation of
co-participation. In this light, embodied semiotic resources, in particular gaze
and body posture, were used either to signal that the speaker was holding the
turn or for inviting the interlocutor to participate in the word search. Inviting
fellow test takers to participate in the word search through embodied semiotic
resources like gestures, gaze direction, circumlocutions and word coinage did,
in several cases, allow the participants to jointly achieve the word search.
However, an invitation was not always taken up by the recipient, who could
display an avoidance or hesitancy to participate in the word search by not con-
firming, correcting or proffering a candidate solution or by delaying his or her
contribution. By avoiding further negotiation of meaning, the participants pri-
oritized the progressivity of talk. In both achieved and abandoned word
searches, the participants oriented towards manifesting sufficient intersubjec-
tive understanding, allowing the discussion to move on rather than pursuing
linguistic precision.
Negotiating meaning and co-participation in word searches build upon
closely attending to the other test takers’ bodily conduct. The difficulty of co-
participation in word searching sequences pertains to the predictability of the
missing item and whether the other test taker has access to it. Meanwhile, the
test situation also entails limitations to the negotiation of meaning, which can
explain why some test takers avoided participating in the word search even
though they were invited to do so. Following Goffman, avoidance ‘is the sur-
est way for a person to prevent threats to his face’ (1967: 15). Thus, instead of
orienting to the other test taker expressing difficulty or displaying one’s own
lack of understanding and thereby putting the image of being ‘competent lan-
guage users’ at risk, avoidance of negotiation of meaning becomes an inter-
actional resource allowing the discussion to move on. A dilemma between the
institutional task of assessing individuals in an interactional context where a
conversation is co-constructed has been put forth by language testers. As

67
shown in Study II, negotiating co-participation and co-construction also pre-
sents a challenge for the test takers (see also Ducasse 2010).
An argument put forth in Study II is that, although it has been established
that word search behaviour is more or less similar across languages, it remains
critical to treat word searches as a highly situated practice. Hence, the fact that
the speaking tests are staged (see Study I) has consequences for how the word
searches are enacted and achieved. Previous studies on L2 word searches have
pointed out how the use of L1 or other languages constitute important re-
sources and commonly used strategies for negotiating meaning and solving
word searches (Kurhila 2006, Lin 2014, Siegl 2016). However, in assessment
practices, the use of other languages is often associated with a deficiency per-
spective, and language tests are a case in point where students are expected to
behave monolingually (Shohamy 2013). As discussed in Study II, in the data
set as a whole, the use of languages other than Swedish was generally avoided.
The use of linguistic items associated with languages other than Swedish did
occur, albeit rarely. However, when this was the case, they were used in a
‘marked’ fashion (cf. Kurhila 2006:110) surrounded by different interactional
‘signs of perturbation’ (Silverman & Peräkylä 1990), such as pauses, hesita-
tion marks and repetitions. Consequently, the deficiency perspective on using
other languages was reproduced in institutional talk, enhancing a monolingual
ideology of competence (Shohamy 2011; see also Study I).
In the word searching sequences studied, the test takers sought assistance
mainly from the other test taker and not the teacher(s), who could be seen as
a more knowledgeable party. This is interesting in light of other studies on
word searches with both L1 and L2 speakers, where relying on the L1 speaker
to solve the word search is a common strategy (e.g. Gullberg 2011, Kurhila
2006). Of course, of relevance to the interactional context at hand, is the fact
that it is part of the instruction that the test takers should mainly discuss with
each other and not address the examiners. Hence, by not turning to the exam-
iners the test takers oriented to ‘managing on their own’ in their display of
being ‘competent language users’.

6.3 Study III – Being ‘a competent language user’ in

a world of Others – adult migrants’ perceptions
and constructions of communicative competence
Drawing on an analysis of six focus group discussions with SFI students,
Study III sets out to bring the voices of SFI students in to a discussion on
perceptions of competence in relation to their lived experience of language
(Busch 2017). How to define language competence is an ongoing discussion
within applied linguistics and educational research. For the past 40 to 50 years

68
or so, the notion of communicative competence has held a central position
within applied linguistics in general and in educational linguistics in particu-
lar. Following Kataoka et al. (2013), communicative competence is shaped by
language ideologies and the education system, negotiated and contingent upon
both situated interactions as well as larger social structures. From this vantage
point, Study III discusses how perceptions and constructions of competence
are shaped in intersubjective processes.
From the perspective of educational policy and language assessment, stu-
dents are often ’objectified’ through descriptions like can do-statements and
wordings in grading criteria and syllabi using third person formulations.23
However, for learners, language is instead more ‘a lived reality than a formal
construct’ (Kramsch 2009:4). In order to explore a group of SFI students’ per-
ceptions of competence, Study III draws on the concept of the lived experience
of language (Busch 2017), inspired by Merleau-Ponty’s phenomenology
(2012[1945]). For Merleau-Ponty, language is considered primarily a bodily
phenomenon, and our knowledge is constructed through our bodily experience
of being in the world. Subjects come into being in interdiscursive practices
and in interaction with others – an intrinsically dialogic and intersubjective
process (Busch 2017, Kramsch 2009, Merleau Ponty 2012[1945]). The con-
cept of lived experience of language thus emphasizes the intersubjective di-
mension of language.
The starting point for investigating the lived experience of language is the
experiencing subject. Following Busch, ‘[m]oments of lived experience of
language inscribe themselves to the linguistic memory’ (2017:343). The focus
group discussions should also be treated as an interactional event in and of
itself and as a social practice (cf. Briggs 1986, Talmy 2011). Hence, the state-
ments made in the group discussions are not to be seen as ‘objective truths’ or
a ‘mirror of reality’. Rather, what is interesting is what the participants say in
this particular setting and how they orient to and construct perceptions of com-
petence in an institutional and educational setting. The analytical focus is not
on summarizing themes, but rather on investigating the discursive construc-
tions of lived experience of language (Busch 2017) in relation to perceptions
of competence. Study III focuses on themes that emerged during the discus-
sions, for being either recurrent or being treated as significant in one way or
the other by the participants. Importantly, reported events in interview dis-
course should not be seen as a mere reflection of reality. Rather, they point to
how the reported experiences were lived and how they gain meaning in talking
about them (Hanell 2017, Ochs & Capps 1996).

23
One of the criteria for oral interaction provides an example: ‘The student chooses and uses
functional strategies to solve problems in the interaction’ (National Agency for Education
2012a/2017a).

69
 Study III points to how lived experience of language constitutes the partic-
ipants’ understanding of what it means to be ‘a competent language user’. Ac-
cordingly, the participants’ comments primarily constructed a view of compe-
tence as made relevant and being shaped in social practice. By doing so, the
participants oriented to competence as a primarily relational construct shaped
in intersubjective processes within specific ideological and social conditions.
This was manifested in different ways in the focus group discussions outlined
below.
A frequent discursive image that emerged during the discussions was man-
aging on one’s own as a sign of being a competent language user. This view
was extended in comments describing how being able to interpret for others
was a sign of competence. Corroborating this view were the negative feelings
and experiences attached to being in need of an interpreter. SFI, like many
other language programmes, is underpinned by an overarching functional ap-
proach in the syllabus stating what students can do with their language in do-
mains such as daily, societal and working life. Hence, orientations to subject
positions such as managing on one’s own in social encounters and being re-
sponsible for one’s own learning resonate both with the communicative and
functional approach to language in the SFI syllabus and with contemporary
neoliberal ideologies on ‘own responsibility’ and ‘self-improvement’ (Carlson
2002, Park 2010, Sandberg et al. 2016). However, this seemingly individual
perspective is constructed in relation to others. Thus, being ‘a competent lan-
guage user’ can be described in terms of orienting towards different subject
positions, both as a speaker as well as someone who is understood and re-
spected (cf. Bourdieu 1977). Moreover, managing on one’s own as a sign of
competence constitutes a holistic approach to what it means to know a lan-
guage (Carlson 2002).
Further corroborating the relational view of competence was the importance
given to assessments, both those made by others as well as internalized self-
assessments, made in both institutional and everyday practices. Assessments
could serve both as being empowering as well as having a negative impact on
the speaker’s self-perception and confidence, which ultimately could lead to
silencing. One of the most pervasive functions of assessments is that they be-
come internalized through the process of self-surveillance and self-censorship,
sometimes leading to silencing (cf. Bourdieu 1977, Foucault 1977). Discipli-
nary power, exercised through surveillance, is relational (Foucault 1977) and
represents a powerful mechanism that can be assigned to symbolic others like
teachers and shopkeepers or internalized by oneself and having a decisive im-
pact on the participants’ perception of themselves as competent or non-com-
petent language users. The embodied experience of internalizing the view of
the other was made relevant in relation to linguistic correctness and speaking
incorrectly. Labels like ‘grammar’ and ‘vocabulary’ came up frequently dur-
ing the discussions (cf. Carlson 2002), and became icons both of what it means

70
to know a language and what is difficult to learn. The embodiment of the sym-
bolic power of linguistic correctness is manifested in experiences of shame,
embarrassment (cf. Bourdieu 1998) or ‘being out of place’ (Salö 2015).
Hence, the feeling of shame or inadequacy is due to the transgression of a
norm, sometimes leading to silencing. Silence is thus given a relational di-
mension (Fivush 2010). Moreover, the comments made by the participants
show how the meaning attached to negative experiences in communicative
practice play a central role in the participants’ self-perception (Busch 2012).
Emotional aspects in relation to perceptions of competence were also fore-
grounded in terms of self-confidence, in the sense that being ‘a competent
language user’ implies gaining self-confidence as a speaker, which in turn is
based on an intersubjective process. The emotional aspects extend the func-
tional view of language competence used in adult education (cf. McNamara
2011).
Furthermore, the central role given to others and the relational view of lan-
guage and competence was accentuated by the frequent use of reported speech
in the participants’ comments and reflections upon their communicative expe-
riences. Using reported speech and embedding the voices of Others under-
scores the affective and relational aspect of how competence is perceived and
constructed. Importantly, inner speech or reported thought should not be con-
sidered as a mere reflection of an inner world, but rather how the ‘outer comes
in’ (Vološinov 1973:93) and how the reported utterances were received by the
speaker. Moreover, the excerpts analysed in Study III show how other-at-
tributed reported speech was more commonly used than self-attributed re-
ported speech. This is interesting in the light of discourses on individual re-
sponsibility. Reported speech is frequently used in tellings and in interview
discourse (e.g. Miller 2014), and assigning an utterance to someone else is
‘closely associated with the assignment of responsibility’ (Duranti 2015:130).
The use of other-attributed reported speech illustrates how the participants
embodied and embedded the voices of Others in their perceptions of compe-
tence and point to a perceived limited agency in becoming ‘a competent lan-
guage user’.
In sum, the lived experience of language is central to our understanding of
being social actors and languaging subjects in the world, shedding light on the
relational aspect of language and, ultimately, how being ‘a competent lan-
guage user’ is constructed as an unstable, vulnerable and ambiguous subject
position.

71
7 Concluding discussion

The communicative language movement has put an emphasis on language use

in specific contexts (e.g. Savignon 2006). By emphasizing language use, the
scope of ‘what it means to know a language’ has broadened to include how
speakers act in relation to others and in different settings. Relatedly, research-
ers in the language testing field have stressed the need for language testing to
interact with the field of sociolinguistics to enhance the understanding of lan-
guage use (Harding 2014, McNamara & Roever 2006). This thesis contributes
with a sociolinguistic perspective on language assessment by investigating
paired speaking tests as a speech event and by discussing SFI students’ per-
ceptions and constructions of what it means to be ‘a competent language user’.
Language testing and assessment are understood as social practices where per-
ceptions and constructions of language competence are constructed and regi-
mented metapragmatically as well as interactionally. Moreover, one of the
main arguments in this thesis is that language testing constitutes a situated and
specific social practice that impacts the discourse produced. This thesis has
made the case that paired speaking tests can be analysed as staged institution-
alized performances (Study I). As put forth by Norton J. (2013), how test tak-
ers act in a speaking test is dependent on their interpretation of the event itself.
A better understanding of the test situation as a speech event can thus help
students and teachers to prepare for the test.
Furthermore, taken together, the three studies can be seen in light of differ-
ent tensions in relation to the view of language competence. The three studies
have addressed ideological, embodied and relational perspectives on language
and the image of being ‘a competent language user’. Conversely, in institu-
tional practices like assessment, testing and grading, language competence is
often seen as an objective and individual ability that can be measured (such
as in the models of competence discussed in section 3.1). By contrast, this
thesis has discussed the relational view of language and language compe-
tence. The relational view of language has been discussed from the perspec-
tive of the test takers and how they attend to others – the other test taker, the
examiners listening and assessing, and the role given to others in reflections
upon communicative experiences, ultimately leading to the internalization and
embodiment of the voices of others.

72
 The focus of this thesis has been on oral interaction. On a general note,
there is an inherent dilemma between the institutional need of ‘measuring lan-
guage’ and the insight that language use is dynamic and situated (e.g.
McNamara 1996, Spolsky 1995). As Merleau-Ponty underscores, communi-
cation comes without any guarantees (2012[1945]:194). Ambiguity, or inde-
terminacy, is a fundamental aspect of language (e.g. Cameron 1995, Harris
1981, McNamara 2012). This ambiguity is further accentuated in L2 interac-
tion when words and expressions have not yet been conventionalized in the
speaker’s language use. What follows from this, then, is the importance of
negotiation of meaning and the creative use of words and expressions and
multimodal resources. As noted by Harris, the ability to do so ‘is one of the
most important aspects of linguistic creativity’ (1981:154). The ability to draw
on different semiotic resources to negotiate meaning has also been stressed in
recent work on language use in multilingual and diverse settings (e.g. Cana-
garajah 2013, Rhymes 2014, Seidlehofer 2009). Furthermore, negotiating
meaning, managing different L2 varieties and communicative breakdowns
and their solutions has been identified as a future key area for communicative
language testing (Harding 2014). The ability to solve interactional problems
is also stressed in the grading criteria for oral interaction in SFI (National
Agency for Education 2012a/2017a). Certainly, differential access to linguis-
tic resources constitutes a challenge to successful negotiation of meaning and
for achieving intersubjective understanding. Meanwhile, the possibilities of
negotiation of meaning are context-dependent. As shown in Study II, overall,
the test takers prioritized the progressivity of talk in their staged conversation,
which sometimes led to an avoidance of negotiation of meaning or meaning
making. One of the rationales for the use of the paired format is to encourage
negotiation of meaning, which it in many ways does (e.g. Galaczi 2008). Un-
deniably though, there is an idealized view in language testing of how tasks
and different test formats, like the paired format, can elicit ‘the desirable lan-
guage’. Arguably, if the test situation constrains the possibility of negotiation
of meaning, it affects the inferences and conclusions drawn about the test tak-
ers’ language ability.
Furthermore, by investigating speaking tests as a social practice, embodied
aspects of language use have been foregrounded (e.g. Block 2014, Bucholtz
& Hall 2016, Goodwin 2013, Study II). Previous studies of interaction in test
settings have highlighted the role of the examiner and how the examiner’s
interactional behaviour (i.e. through the support given) has consequences for
the co-construction of the test taker’s discourse (Brown 2003). In Study I, an-
other aspect of examiners’ interactional behaviour is foregrounded as a chal-
lenge – namely silence. Analysing the speech event, and thereby looking be-
yond linguistic features of spoken discourse (i.e. beyond the lingual bias
(Block 2014), provides an enhanced understanding of what the test takers ac-
tually manage to perform (e.g. assuming responsibility in front of an audi-
ence).

73
 Even though language testsers acknowledge non-verbal features to be im-
portant for oral interaction (e.g. Galaczi & Taylor 2018, National Agency for
Eudcation 2012b), language testing still has a tradition of evaluating ‘disem-
bodied speech’, inter alia evidenced by the fact that few studies have, to my
knowledge, addressed multimodal and embodied aspects of interaction in test
settings (Ducasse 2010, Jenkins & Parra 2003, Gan & Davidson 2011 being
exceptions) and the growing market of speaking tests delivered digitally and
scored by automatized computer programmes (Galaczi & Taylor 2018:13). In
contrast to the logocentric view dominating linguistics in general and language
assessment in particular (Goodwin 2004, Harris 1981, Heughs 2004), Study II
shows how embodied resources are crucial for the negotiation of meaning.
What is more, as argued by Kataoka et al. ‘collaborative, embodied aspects
of instruction and socialization challenge the ‘individual cognition’ assump-
tion of traditional approaches to communicative competence’ (2013:346).
This is relevant in light of Study II. Investigating word searching sequences
in speaking tests taps into issues of co-construction and individuality. Ideolo-
gies of individuality are pervasive in testing (cf. Foucault 1977:189ff.) and are
highly institutionalized, since test scores and grades are given to individuals.
Following the advent of the communicative testing movement, there has been
raised awareness of the social character of performance testing and the co-
construction of interaction (e.g. Chalhoub-Deville 2003, Ducasse 2010,
Galaczi 2014, McNamara 1997, 2001, McNamara & Roever 2006, Norton J.
2013, Swain 2001). On a general level, the question of individuality and co-
construction has been addressed from the perspective of testers by addressing
issues such as how to define a relevant construct of interaction and if individ-
ual grades should be assigned or not (e.g. May 2009) or by investigating dif-
ferent aspects of the so-called ‘interlocutor effect’ (see 4.2.2). This thesis does
not provide any answers to such issues. The individual view of knowledge is
highly institutionalized and, in many ways, necessary. Instead, this thesis has
sought to show that what we consider individual aspects are part of a larger
relational construct. In this vein, individual aspects of communicative compe-
tence always represent a limited view of what it means to be ‘a competent
language user’. Arguably then, it follows that interpretations of a person’s lan-
guage competence solely based on tests should be treated caution.
As argued in Study II, how to display individual language knowledge in a
conversation is also of concern for the test takers and may, in cases of inter-
actional troubles, present a challenge. As put forth in Study II, even when the
test takers orient to the display of vocabulary knowledge as an individual re-
sponsibility (i.e. by holding the turn extendedly and not seeking assistance
from the interlocutor), it is important to underscore that this is contingent upon
cooperative embodied interaction among all participants. Hence, following
Vološinov (1973:34), ‘every sign, even the sign of individuality, is social’.
Another aspect of individuality in relation to ideologies of language com-
petence that has been addressed in different ways is how managing on one’s

74
own is an emblem of being ‘a competent language user’. This image is con-
structed metapragmatically in the instruction to the test takers before the
paired speaking test (see section 5.1.1), enacted interactionally in the test tak-
ers’ team performance by not turning to the more knowledgeable party (i.e.
the examiner) in word searching sequences and by being discursively con-
structed in the focus group discussions. This image is also rooted in functional
and communicative of frameworks for language teaching, where aims and
proficiency level descriptions are described in terms of what the students can
do with their language in different contexts (e.g. Council of Europe 2001; see
also section 3.1). Likewise, ‘managing on one’s own’ fits with the political
discourse stating that the aim of language instruction for adult migrants is for
the migrants to become employable so they can ‘manage on their own’ in so-
ciety. From this perspective, communicative approaches to language compe-
tence both resonates with how learners’ express their needs and the politiciza-
tion of language competence, giving communicative approaches to language
programs and language tests a market value. Related to managing on one’s
own in social interaction is the orientation to being responsible for one’s own
learning. This reflects contemporary discourses on language learning for mi-
grants that often posit that it is a duty to learn Swedish, and that doing so is
mainly dependent upon the individual learner’s motivation and responsibility
(see section 2.3). However, as discussed in Study III, seemingly individual
achievements are constructed in intersubjective processes and are not depend-
ent only on the individual.
Yet another ideology that has been discussed in relation to constructions
and perceptions of language competence is monolingualism. Studies I and II
both show how a monolingual orientation to language competence is sustained
in testing practices. Language testing practices are in general based on mono-
lingual constructs and monolingual behaviour (Shohamy 2011). This is inter-
esting in light of recent debates in sociolinguistic and educational research. In
recent years, there has been a proliferation of new terms suggesting a para-
digm shift in the view of multilingual speakers, linguistic practices and edu-
cational research, shifting away from the term ‘competence’ (e.g. Blommaert
& Backus 2013, Canagarajah 2013, 2018, Pennycook 2016). Instead, in order
to foreground the multilingual and multimodal practice-based nature of lan-
guage, terms like translanguaging and polylanguaging have been suggested
(see Pennycook [2016] for an overview). At the same time, it remains im-
portant to investigate settings that sustain a more traditional view of language
and language competence, such as testing practices (see also Charalambous et
al. [2017] for a study on classroom discourse in Cyprus where the students
avoid using all the linguistic resources at their disposal since Turkish is seen
as ‘the language of the enemy’). From this perspective, Studies I and II pro-
vide insights into how a monolingual ideology of language competence is sus-
tained and reproduced both metapragmatically and interactionally in test dis-
course.

75
 Significantly, there are also differences between the studies. Comparing
the discourse produced in test settings with focus group discussions points to
the existence of competing ideologies. Where the test discourse displayed a
monolingual orientation, some participants in the focus group discussions
constructed a multilingual view of language competence by being able to
translate for others, and some participants rejected the view of the importance
of learning Swedish by claiming English to be sufficient in certain contexts.
Another difference is between what is expressed metapragmatically and
how the participants act in interaction. In both Study I and Study III, the par-
ticipants highlight the symbolic value of linguistic correctness, whereas in the
interactional practice investigated in Study II, the participants rarely oriented
to linguistic correctness (here, in terms of precision or lexical accuracy). In
many cases, the participants prioritized the progressivity of talk rather than
finding a correct form of a lexical item – either by displaying intersubjective
understanding of the linguistic and embodied resources that they found appro-
priate to use, or by avoiding negotiation of meaning and moving on with the
discussion. Hence, the symbolic value of linguistic correctness is expressed
metapragmatically as a ‘loyalty to norms of correctness’ (Gal 2006:17) rather
than oriented to in interaction. As shown in a large body of studies on L2
interaction, linguistic correctness is not necessary for meaning making (e.g.
Canagarajah 2013, Kurhila 2006, Seidlehofer 2009, Sundberg 2004). Instead,
as Bourdieu (1977) holds, the recognition of the importance of ‘correct lan-
guage’ rather than the knowledge of it constructs the symbolic value of lan-
guage competence. From this perspective, claiming the importance of linguis-
tic correctness and knowing ‘perfect Swedish’ serves as a mechanism of ar-
rangement on a linguistic market for those who do not (yet) have access to the
resources considered valuable. As argued in this thesis, the image of being ‘a
competent language user’ in relation to migration is a powerful language ide-
ology, where successful language learning has a moral valuation linking the
image of ‘a competent language user’ with being ‘a good student’ and ‘a good
immigrant’.

Testing and assessment are key elements of education. Meanwhile, language

assessment is a practice that goes beyond institutional practices taking place
in language classrooms. As shown in Study III, language assessments made
in both institutional and everyday practices form part of the participants’ lived
experience of language and impact their self-perception. As such, language
assessments should not be seen as a neutral practice. Language testers often
hold that ‘[w]hatever attributes or abilities we measure, it is important to un-
derstand that it is these attributes or abilities and not the persons themselves
that we are measuring (Bachman 1990:20; see also Erickson [2015] for a sim-
ilar line of argument). This is, of course, a professional stand-point. Yet, it
must be seen as naïve. As discussed in Study III, language assessments, both

76
those made by others as well as those internalized by the speaker, are perva-
sive practices affecting a speaker’s self-perception and agency, sometimes
leading to silencing (see also Shohamy 2001, 2017). From this perspective,
the role of language assessments point to the relational construction of lan-
guage. The relational view of language brings to the forefront the existential
experience of being positioned in relation to others and the emotions attached
to these experiences. Following Jaffe (2013), communicative competence
needs to be socially and interactionally sanctioned, leading to being ‘a com-
petent language user’ as a situated and vulnerable subject position, depending
on others. Overall, in the investigated discursive practices in the present thesis,
being a ‘competent language user’ is oriented to as a desired subject position.
Meanwhile, it is important to note that being a desired and highly valorized
subject position means vulnerability for those who strive to achieve it (cf.
Bourdieu 1998). Subsequently, experiences of not being able to ‘live up’ to
the ideal of the standard language and linguistic correctness become anchored
in speakers’ lived experience of language.
In sum, investigating language testing and assessment from a sociolinguis-
tic perspective sheds light on the assumptions made about language compe-
tence and what it means to be ‘a competent language user’. This thesis should
not be read as an argument against assessment and testing, since ‘there is no
human society where assessment will not occur’ (McNamara 2012:459).
However, since testing and assessment practices are important, they need to
be examined more closely. Language testing and assessment are both periph-
eral and central part to applied linguistics. On the one hand, being a specific
professional practice with the aim of creating valid and reliable testing prac-
tices (McNamara 2012), testing and assessment practices are often viewed as
peripheral. On the other hand, language testing and assessment are also central
to the field of applied linguistics as well as sociolinguistics, since it deals with
fundamental assumptions and constructions of ‘language competence’. Con-
sidering the fact that testing has become ‘big business’ (Spolsky 2017:375)
and are important for individuals, it remains important to investigate language
testing and assessment practices from a sociolinguistic perspective.

77
8 Sammanfattning

Föreliggande avhandling undersöker uppfattningar om vad det innebär att vara

”en kompetent språkbrukare” med fokus på vuxna andraspråksinlärare som
studerar sfi.24 Språktestning och bedömning av språkfärdighet (både institut-
ionell och sådan som sker spontant i vardagliga sammanhang) utgör särskilda
praktiker då uppfattningar om språklig kompetens aktualiseras. I denna av-
handling studeras därför testning och bedömning i ett sociolingvistiskt per-
spektiv med ett särskilt fokus på muntlig interaktion och sfi-deltagares upp-
fattningar om språklig kompetens och vad det innebär att vara ”en kompetent
språkbrukare” så som de kommer till uttryck, interaktionellt och metapragma-
tiskt, i testsamtal (parsamtal i de nationella slutproven) och i fokusgrupper.
Parsamtal (ett samtal mellan två testtagare inför bedömare) är motiverat av
och idéhistoriskt rotat i det paradigm som kallas kommunikativ språkunder-
visning (t.ex. Lindberg 2005, Savignon 2006) som växte fram under 1960- och
1970-talen som en reaktion mot den traditionella grammatikfokuserade under-
visningen. Den har bland annat inneburit ett ökat fokus på språkanvändning i
olika sociala sammanhang. I kommunikativ språktestning förväntas ofta test-
tagarna producera språk på ett eller annat sätt och parsamtal är ett typexempel
på kommunikativ språktestning. Parsamtal är intressanta att studera närmare
eftersom de belyser spänningen mellan att se språklig kompetens som en in-
dividuell färdighet och något som kommer till uttryck och samkonstrueras
med andra i interaktion.
Avhandlingens sociolingvistiska perspektiv tar sig uttryck bland annat ge-
nom att bilden av språklig kompetens ses som situerat, både till det samman-
hang det kommer till uttryck i och förankrat i sin tid. Det innebär att synen på
språklig kompetens är historiskt föränderligt samtidigt som det ger en inblick
i samtida språkliga ideologier. Kappan ägnar därför en stor del åt att kontex-
tualisera avhandlingens studier i ett sociopolitiskt och sociohistoriskt perspek-
tiv. Språklig kompetens ses inte heller som ett ”fast objekt” som finns där
färdigt att analyseras, utan avhandlingens tre studier närmar sig konstruktioner
och uppfattningar om språklig kompetens och ”den kompetente språkbruka-
ren” genom olika angreppssätt. Kappan diskuterar därför synen på språklig
kompetens i olika perspektiv, så som modeller av språklig kompetens inom

24
Det officiella namnet är kommunal vuxenutbildning i svenska för invandrare, men i avhand-
lingen används den vedertagna förkortningen sfi (se även Rosén 2013 och Sandwall 2013).

78
kommunikativ språkundervisning och testning, ideologier om en ”god
svenska” och andraspråkstalares egna uppfattningar om språklig kompetens.
Kommunikativt inriktad språkundervisning betonar vikten av att kunna an-
vända språket i olika sociala sammanhang och anpassa det efter syfte och mot-
tagare, ofta utanför undervisningen. Sfi är ett tydligt exempel där målet är att
deltagarna ska utveckla ”språkliga redskap för kommunikation och aktivt del-
tagande i vardags-, samhälls- och arbetsliv samt fortsatta studier” (SKOLFS
2017:91). Det gör det relevant att undersöka både språkbruk och reflektioner
över kommunikativa erfarenheter. På ett övergripande plan kretsar avhand-
lingen kring följande forskningsintressen:

1) Hur kan testsituationen förstås som en språkbrukssituation?

2) Vilka uppfattningar om språklig kompetens kommer till uttryck i test-
samtal och i fokusgrupperna?
3) Hur påverkar sfi-deltagarnas kommunikativa erfarenheter deras syn
på språklig kompetens?

Avhandlingen bygger på en kappa och tre separata studier (artiklar). Studie I

och II bygger på interaktionella analyser av videoinspelade parsamtal i de nat-
ionella slutproven i sfi (kurs D) medan studie III baseras på analyser av fokus-
gruppsdiskussioner med sfi-deltagare. Frågeställning 1) diskuteras i Studie I
och i Studie II. Frågeställning 2) behandlas ur olika perspektiv i alla tre studi-
erna. Studie I och studie III tar fasta på sekvenser i testsamtalen och i fokus-
grupperna då språklig kompetens, språkanvändning, kommunikativa erfaren-
heter och språkinlärning diskuteras, medan Studie II diskuterar vilka semio-
tiska resurser som aktualiseras interaktionellt i ordsökningssekvenser. Fråge-
ställning 3) behandlas slutligen i Studie III.
Studie I diskuterar hur lingvistisk antropologisk performance-teori (t.ex.
Bauman & Briggs 1990, Bauman 2004) tillsammans med Bakhtins (1986) be-
grepp addressivitet (att ett yttrande alltid är riktat till någon och den man riktar
sig till påverkar det som sägs), kan utgöra ett ramverk för att analysera inter-
aktionen i parsamtalen. Performance definieras som en publikorienterad dis-
kurs. I det här fallet utgör bedömarna publiken och blir på så vis en del av
interaktionen i egenskap av addressater. Studie I argumenterar för att parsam-
talen kan analyseras som institutionell performance där det talade språket ob-
jektifieras och ideologier synliggörs. Studie I fokuserar på tillfällen då testta-
garna tar upp synen på språklig kompetens, språkinlärning och språkanvänd-
ning. Analyserna visar att deltagarna orienterar sig mot samtida ideologier om
språk och integration, så som vikten av att kunna bra svenska, vikten av språk-
lig korrekthet och att språket lärs genom att bara prata svenska, helst ha
svenska vänner och att ansvaret för att lära sig ligger hos den studerande.
Dessa ideologier samkonstrueras av deltagarna genom interaktionell sam-
stämmighet, det vill säga att testtagarna håller med varandra. Bilden av ’den

79
kompetente språkbrukaren’ sammanfaller på så vis med ’den gode invandra-
ren’ och ’den gode studenten’.
Studie II utgår från ordsökningssekvenser som en interaktionell praktik
och diskuterar hur dessa utförs och hur provsituationen påverkar de resurser
som används. Ordsökningar ringar in flera viktiga områden inom kommuni-
kativ språkundervisning: ordkunskap, förhandling om mening och förmågan
att lösa problem i interaktionen. Ordsökningssekvenserna analyseras som en
synlig förkroppsligad praktik (t.ex. Goodwin & Goodwin 1986, Goodwin
2013). Genom användandet av olika kroppsliga resurser så som blickföring,
kroppshållning, gester, förhandlar testtagarana deltagande i ordsökningsse-
kvenserna. Analyserna visar hur dessa resurser både används för att hjälpa
varandra och gemensamt lösa ordsökningen och skapa mening eller för att
undvika att delta i den andres ordsökning. Det leder i vissa fall till ett undvi-
kande av förhandling om mening. Överlag orienterar sig deltagarna mot att
upprätthålla samtalet och gå vidare än att orientera sig mot lexikal precision.
Till skillnad från andra ordsökningsstudier med andraspråkstalare undviker
deltagarna här att använda andra språkliga resurser än svenska (t.ex. Kurhila
2006, Lin 2014, Siegl 2016). På så vis orienterar de sig mot och upprätthåller
en monolingvistisk bild av språklig kompetens.
Studie III undersöker sfi-deltagares syn på språklig kompetens och vad det
innebär att vara ”en kompetent språkbrukare” utifrån deltagarnas kommuni-
kativa erfarenheter. Känslan av att vara eller inte vara en kompetent språkbru-
kare är en känsla deltagarna förhåller sig till inte minst via erfarenheter av att
inte bli sedd som en. Att vara ”en kompetent språkbrukare” konstruerades i
fokusgruppsdiskussionerna som en åtråvärd, men sårbar position. Utifrån fe-
nomenologisk språkforskning diskuteras hur uppfattningar om språklig kom-
petens formas i intersubjektiva processer och i levda erfarenheter (Busch
2017). Studie III diskuterar således språklig kompetens i ett relationellt per-
spektiv och lyfter bland annat fram emotionella perspektiv och hur att vara
”en kompetent språkbrukare” är en subjektsposition som skapas i relation till
andra. Bedömningar av ens språkförmåga (både de som utförs av lärare i ut-
bildningssammanhang och de som yttras i vardagliga sammanhang) har bety-
delse för individer som levda erfarenheter och formar talares självuppfattning,
vilket ytterligare förstärker det relationella perspektivet på språklig kompe-
tens. Bedömningar kunde även införlivas hos talarna vilket kunde leda till en
sorts självcensur där de tystnade av rädsla för att låta ”fel”. Det relationella
perspektivet på språklig kompetens står i kontrast till hur lyckad språkinlär-
ning i offentlig debatt ofta framställs som i huvudsak beroende på individens
vilja, ansvar och ansträngningar.

I globaliseringens kölvatten har vuxna migranters språkkunskaper tillmäts ett

allt större symboliskt värde, där lyckad språkinlärning ses som tecken på loja-
litet och tillhörighet med det nya landet. Detta har lett till en ökad ”kravreto-
rik” som bland annat har tagit sig uttryck i tal om språkplikt och betoning av

80
det egna ansvaret för att lära sig svenska. I den allmänna debatten om vuxna
invandrares språkkunskaper råder ofta en ensidig och förenklad syn på språk
och språkinlärning. I den här typen av diskurser framställs ofta framgångsrik
språkinlärning som något som endast bygger på individens vilja och ansträng-
ning. I kontrast till den förenklade bilden av språk och språkinlärning som ofta
styr diskursen kring vuxnas andraspråksutveckling, har studierna i denna av-
handling undersökt den ideologiska synen på språklig kompetens, språklig
kompetens som en relationell konstrukt och språk som en situerad förkropps-
ligad praktik. Denna syn står i kontrast till ideologier om språk som en indivi-
duell färdighet, ett beskrivbart objekt som går att mäta och i huvudsak logo-
centriskt. Det skapar i förlängningen en spänning i relation till det institution-
ella behovet av att mäta språk och bedöma individer. Den relationella synen
på språk har både tagit sig uttryck i hur deltagarna förhåller sig till andra (den
andre testtagaren, lärarna som lyssnar och bedömer och i beskrivningar av
kommunikativa möten med andra) och hur de är beroende av andras bedöm-
ningar för att känna sig som kompetenta språkbrukare. Det relationella per-
spektivet tar sig även uttryck genom att deltagarna införlivade andras blick
och röster i sina utsagor.
Språktestning har inom tillämpad språkvetenskap setts som ett speciali-
serat, och i viss mån perifert, fält. Denna avhandling har dock diskuterat och
visat på hur bedömning och testning av språk blottlägger olika antaganden om
vad det innebär att kunna ett språk och att vara ”en kompetent språkbrukare”.
På så vis är bedömningspraktiker centrala inom tillämpad språkvetenskap och
för uppfattningen och konstruktionen av vad det innebär att vara ”en kompe-
tent språkbrukare”. Ett annat huvudargument har varit att testpraktiker, i det
undersökta fallet parsamtal, bör ses som en specifik språkbrukssituation vilket
påverkar interaktionen. Detta synsätt har även konsekvenser för språktestning.
Det som går att fånga och bedöma i en testsituation representerar en begränsad
del av språkanvändning. Det är därför viktigt att konsekvenserna av testning
inte är för stora.

81
9 References

Abdulla, Afra. 2017. Readiness or resistance? – Newly arrived adult migrants’ expe-
riences, meaning making, and learning in Sweden. Linköping Studies in
Behavioural Science No. 203. Linköping: Linköping University.
Ahlgren, Katrin. 2014. Narrativa identiteter och levande metaforer i ett andraspråks-
perspektiv. Doktorsavhandling i språkdidaktik 4. Stockholm: Stockholms univer-
sitet.
Ahlgren, Katrin. 2016. Teoretisk utveckling och framtida utmaningar – nationella
prov i svenska för invandrare. Kieli, koulutus ja yhteiskunta – joulukuu. Pp. 1–8.
Ahlgren, Katrin, Inger Lindberg, Mikael Olofsson & Maria Rydell. N.d. Teoretisk
modell för bedömning av vuxnas andraspråksutveckling med fokus på kommuni-
kation i vardags-, samhälls- och arbetsliv. Stockholm: Institutionen för språkdi-
daktik, Stockholms universitet.
Ahlgren, Katrin, Grazyna Bartholdson, Sofia Engman, Mikael Olofsson, Anders
Philipsson, & Ingrid Skeppstedt. 2013. Bedömning av språkfärdighet inom ut-
bildning i svenska för invandrare – med fokus på de nationella proven i sfi. In:
Monica Axelsson, Marie Carlson, Qarin Franker & Karin Sandwall (red.) Pro-
fession, Politik och Passion. Inger Lindberg som andraspråksforskare – en vän-
bok. Pp. 55–74.
Ahmadi, Alireza & Elham Sadegi. 2016. Assessing English Language Learners’ Oral
Performance: A Comparison of Monologue, Interview and Group Oral Test. Lan-
guage Assessment Quarterly 13(4). Pp. 341–35.
Allan, Kori. 2013. Skilling the Self: The Communicability of Immigrants as Flexible
Labour. In: Alexandre Duchêne, Melissa Moyer & Celia Roberts (eds) Language,
Migration and Social Inequalities. A Critical Sociolinguistic Perspective on In-
stitutions and Work. Bristol: Multilingual Matters. Pp. 56–78.
Anderson, Benedict.1991. Imagined communities: Reflections on the origin and
spread of nationalism. London: Verso.
Bachman, Lyle. 1990. Fundamental considerations in Language Testing. Oxford: Ox-
ford University Press.
Bachman, Lyle & Adrian Palmer. 1996. Language Testing in Practice. Oxford: Ox-
ford University Press.
Bachman, Lyle & Adrian Palmer. 2010. Language Assessment in Practice. Oxford:
Oxford University Press.
Bakhtin, Mikhail Mikhailovich. 1981. The Dialogic Imagination. Four Essays by
M.M. Bakthin. Edited by Michael Holquist. Austin: University of Texas Press.
Bakhtin, Mikhail Mikhailovich. 1986. Speech Genres & Other Late Essays. Edited by
Caryl Emerson & Michael Holquist. Austin, Texas: University of Texas Press.
Banke, Viktor. 2017. Andrum. Om stölden av en flyktingkris och om de bestulna.
Stockholm: Norstedts.
Bauman, Richard. 2004. A Worls of Others’ Words. Cross-Cultural Perspectives on

82
 Intertextuality. Oxford: Blackwell Publishing.
Bauman, Richard & Charles L. Briggs. 1990. Poetics and Performance as Critical Per-
spectives on Language and Social Life. Annual Review of Anthropology 19. Pp.
56–88.
Bjerg Petersen, Karen. 2010. Undervisningsfaget dansk som andetsprog for voksne:
nogle udviklingstendenser. Unpublished thesis. Århus: Department of Education,
Århus University.
Blackledge, Adrian. 2005. Discourse and Power in a Multilingual World. Amster-
dam/Philadelphia: John Benjamins.
Blackledge, Adrian. 2009. Being English, speaking English. Extension to English lan-
guage testing legislation and the future of multicultural Britain. In: Gabriella Ho-
gan-Brun, Clare Mar-Molinero & Patrick Stevenson (eds) Discourses on Lan-
guage and Integration. Amsterdam/Philadelphia: John Benjamins. Pp. 83–107.
Block, David. 2003. The Social Turn in Second Language Acquisition. Edinburgh:
Edinburgh University Press.
Block, David. 2014. Moving beyond “Lingualism”: Multilingual Embodiment and
Multimodality in SLA. In: Stephen May (ed.) The Multilingual Turn. Implications
for SLA, TESOL and Bilingual Education. London/New York: Routledge. Pp. 54–
77.
Blommaert, Jan & Ad Backus. 2013. Superdiverse Repertoires and the Individual. In:
Ingrid de Saint-Georges & Jean-Jacques Weber (eds) Multilingualism and Multi-
modality: Current Challenges for Educational Studies. Rotterdam: Sense Pub-
lishers. Pp. 11–32.
Blommaert, Jan, Jim Collins & Stef Slembrouck. 2005. Spaces for multilingualism.
Language & Communication 25. Pp. 197–216.
Blommaert, Jan & Ben Rampton. 2011. Language and Superdiversity. Diversities
13(2). Pp. 1–21.
Bourdieu, Pierre. 1977. The economics of linguistic exchanges. Social Science Infor-
mation 16(6). Pp. 645–668.
Bourdieu, Pierre. 1982. Ce que parler veut dire. L’économie des échanges linguis-
tiques. Paris: Fayard.
Bourdieu, Pierre. 1991. Language and Symbolic Power. Edited and introduced by
John B. Thompson. Cambridge: Harvard University Press.
Bourdieu, Pierre. 1996. Understanding. Theory, Culture & Society 13(2). Pp. 17–37.
Bourdieu, Pierre. 1998. Masculine domination. Cambridge: Polity.
Bourdieu, Pierre & Loïc J.D. Wacquant. 1992. An Invitation to Reflexive Sociology.
Cambridge: Polity Press.
Borevi, Karin. 2012. Sweden: The Flagship of Multiculturalism. In: Grete Brochmann
& Anniken Hagelund (eds) Immigration Policy and the Scandinavian Welfare
State. New York: Palgrave Mcmillian. Pp. 25–96.
Briggs, Charles. 1986. Learning How to Ask: A Sociolinguistic Appraisal of the Role
of the Interview in Social Science Research. Cambridge: Cambridge University
Press.
Briggs, Charles. 1988. Competence in Performance: The Creativity of Tradition in
Mexicano Verbal Art. Philadelphia: University of Philadelphia Press.
Brooks, Lindsay. 2009. Interacting in pairs in a test of oral proficiency: Co-construct-
ing a better performance. Language Testing 26 (3). Pp. 341–366.
Brown, Annie. 2003. Interviewer variation and the co-construction of speaking profi-
ciency. Language Testing 2003(1). Pp.1–25.

83
Brown, Annie & Tim McNamara. 2004. “The Devil Is in the Detail”: Researching
Gender Issues in Language Assessment. TESOL Quarterly 38(3). Pp. 524–538.
Brown, Gillian & George Yule. 1983. Teaching the spoken language. Cambridge:
Cambridge University Press.
Brouwer, Catherine. 2003. Word searches in NS-NNS conversations: Opportunities
for learning? Modern Language Journal 83(4). Pp. 534–545.
Brumfit, Cristopher J. & Keith Johnson. 1979. The communicative approach to Lan-
guage Teaching. Oxford: Oxford University Press.
Bucholtz, Mary. 2000. The politics of transcription. Journal of Pragmatics 32. Pp.
1439–1465.
Bucholtz, Mary. 2003. Sociolinguistic nostalgia and the authentication of identity.
Journal of Sociolinguistics 7. Pp. 398–416.
Bucholtz, Mary & Kira Hall. 2016. Embodied sociolinguistics. In: Nikolas Coupland
(ed.) Sociolinguistics. Theoretical debates. Cambridge: Cambridge University
Press. Pp. 173–197.
Bucholtz, Mary & Qiuana Lopez. 2011. Performing blackness, forming whiteness:
Linguistic minstrelsy in Hollywood films. Journal of Sociolinguistics 15(5). Pp.
680–706.
Burns, Anne & Celia Roberts. 2010. Migration and Adult Language Learning: Global
Flows and Local Transpositions. TESOL Quarterly 44(3). Pp. 409–419.
Burns, Anne & Helen de Silva Joyce. 2007. Adult ESL programs in Australia. Pro-
spect 2(3). Pp. 5–17.
Busch, Brigitta. 2012. The linguistic repertoire revisited. Applied Linguistics 33(5).
Pp. 503–523.
Busch, Brigitta. 2017. Expanding the notion of the linguistic repertoire: On the con-
cept of Spracherleben – The lived experience of language. Applied Linguistics
38(3). Pp. 340–358.
Cameron, Deborah. 1995. Verbal Hygiene. London/New York: Routledge.
Cameron, Deborah. 2001. Working with spoken discourse. London: Sage publica-
tions.
Cameron, Deborah. 2002. Globalization and the teaching of ‘communicative skills’.
In: David Block & Deborah Cameron (eds) Globalization and Language Teach-
ing. London/New York: Routledge. Pp. 67–82.
Canagarajah, Suresh. 2006. Changing communicative needs, revised assessment ob-
jectives: Testing English as an international language. Language Assessment
Quarterly 3(3). Pp. 229–242.
Canagarajah, Suresh. 2013. Translingual Practice. Global Englishes and Cosmopoli-
tan Relations. London/New York: Routledge.
Canagarajah, Suresh. 2018. Materializing ‘Competence’: Perspectives from Interna-
tional STEM scholars. The Modern Language Journal 102(2). Pp. 268–291.
Canale, Michael. 1983. From communicative competence to communicative language
pedagogy. In: Jack Richards & Richard Schmidt (eds) Language and Communi-
cation. London: Longman. Pp. 2–27.
Canale, Michael & Merrill Swain. 1980. Theoretical bases of communicative ap-
proaches to second language teaching and testing. Applied Linguistics 1. Pp. 1–
47.
Carlson, Marie. 2002. Svenska för invandrare – brygga eller gräns? Syn på kunskap
och lärande inom sfi-undervisningen. Göteborg Studies in Sociology 31. Göte-
borg: Göteborgs universitet.

84
Carlson, Marie & Bengt Jacobson. 2018. Om spänningar i en marknadsorienterad
SFI-utbildning. In: Magnus Dahlstedt & Andreas Fejes (red.) Skolan, markanden
och framtiden. Lund: Studentlitteratur. Pp. 93–112.
Celce-Murcia, Marianne. 2007. Rethinking the role of Communicative Competence
in Language in Language Teaching. In: Eva Alcón & Maria Pilar Safont Jordà
(eds). Intercultural Language Use and Language Learning. Berlin: Springer. Pp.
41–57.
Chalhoub-Deville, Micheline. 2003. Second language interactions: current perspec-
tives and future trends. Language Testing 20(4). Pp. 369–383.
Charalambous, Panayiota, Constadina Charalambous & Michalinos Zembylas. Limits
to translanguaging: Insights from a conflict-affected context. Working Papers in
Urban Language & Literacies paper 190. London: King’s College.
Chomsky, Noam. 1965. Aspects of a theory of syntax. Cambridge, Massachusetts:
The M.I.T. Press.
Colliander, Helena. 2018. Being and becoming a Teacher in Initial Literacy and Sec-
ond Language Education for Adults. Linköping Studies in Behavioural Science
205. Linköping: Linköpings universitet.
Cooke, Melanie. 2006. “When I wake up I dream of electricity”: The lives, aspirations
and ‘needs’ of Adult ESOL learners. Linguistics and Education 17. Pp. 56–73.
Council of Europe. 2001. The Common European Framework of References for Lan-
guages: Learning, Teaching, Assessment (CEFR).
Creese, Angela & Adrian Blackledge. 2018. The Routledge Handbook of Language
and Superdiversity. London/New York: Routledge.
Dahlstedt, Magnus & Andreas Fejes. 2018. Skolan, marknaden och framtiden. Lund:
Studentlitteratur.
Dahlström, Carl. 2004. Nästan välkomna. Invandrarpolitikens retorik och praktik.
Göteborg Studies in Politics 87. Göteborg: Göteborgs universitet.
Davis, Larry. 2009. The influence of interlocutor proficiency in a paired oral assess-
ment. Language Testing 26(3). Pp. 367–369.
Drew, Paul & John, Heritage. 1992. Analyzing talk at work: an introduction. In: Paul
Drew & John Heritage (eds) Talk at work. Interaction in institutional settings.
Studies in Interactional Sociolinguistics 8. Cambridge: Cambridge University
Press. Pp. 3-65.
Ducasse, Ana Maria. 2010. Interaction in Paired Oral Proficiency Assessment in
Spanish. Rater and Candidate Input into Evidence Based Scale Development and
Construct Definition. Language Testing and Evaluation series 20. Frankfurt am
Main: Peter Lang.
Ducasse, Ana Maria & Brown, Annie. Assessing paired orals: Raters’ orientation to
interaction. Language Testing 26(3). Pp. 423–433.
Duff, Patricia & Celia Roberts. (1997). The politics of transcription; Transcribing talk:
issues of representation. TESOL Quarterly 31(1). Pp. 167–172.
Duranti, Alessandro. 1997. Linguistic Anthropology. New York: Cambridge Univer-
sity Press.
Duranti, Alessandro. 2010. Husserl, intersubjectivity and anthropology. Anthropolog-
ical Theory 10(1-2). Pp. 16–35.
Duranti, Alessandro. 2015. The Anthropology of Intentions. Language in a World of
Others. Cambridge, U.K.: Cambridge University Press.
Eades, Diane. 2009. Testing the Claims of Asylum Seekers: The Role of Language
Analysis. Language Assessment Quarterly 6(1). Pp. 30–40.

85
Edgeworth, Francis Ysidro. 1890. The element of chance in competitive examina-
tions. Journal of the Royal Statistical Society 53. Pp. 644–663.
Együd, Györgyi. & Philip Glover. 2001. Oral testing in pairs – A secondary school
perspective. ELT Journal 55(1). Pp. 70–76.
Eklund Heinonen, Maria. 2009. Processbarhet på prov. Bedömning av muntlig språk-
färdighet hos vuxna andraspråksinlärare. Skrifter utgivna av Institutionen för
nordiska språk vid Uppsala universitet 78. Uppsala: Uppsala universitet.
Elder, Catherine, Tim McNamara, Hyejeong Kim, John Pill & Takanori Sato. 2017.
Interrogating the construct of communicative competence in language assessment
contexts: What the non-specialists can tell us. Language & Communication 57.
Pp. 14–21.
Erickson, Gudrun. 2015. Kunskapsbedömningars grund och syften. In: Gustaf Skar &
Michael Tengberg (red.) Bedömning i svenskämnet. Årskurs 7-9. Stockholm: Na-
tur & Kultur.
Fairclough, Norman. 1992. The appropriacy of ’appropriateness’. In: Norman Fair-
clough (ed.) Critical Language Awareness. Harlow: Addison Wesley Longman
Limited. Pp. 34-56
Fejes, Andreas. 2010. Discourses on employability: constituting the responsible citi-
zen. Studies in Continuing Education 32(2). Pp. 89–102.
Fejes, Andreas & Magnus Dahlstedt. 2018. Popular education, migration and a dis-
course of inclusion. Studies in the education of adults 49(2). Pp. 214–227.
Fejes, Andreas, Magnus Dahlstedt, Maria Olson & Fredrik Sandberg. 2018. Adult ed-
ucation and the formation of citizens. A critical interrogation. London/New York:
Routledge.
Fivush, Robyn. 2010. Speaking silence: The social construction of silence in autobi-
ographical and cultural narratives. Memory 18(2). Pp. 88–98.
Firth, Alan & Johannes Wagner. 1997. On discourse, communication and (some) fun-
damental concepts in SLA research. Modern Language Journal 81. Pp. 285–300.
Foucault, Michel.1977. Discipline and Punish. The birth of the prison. Translated by
Alan Sheridan. London: Penguin Books.
Franker, Qarin. 2011. Litteracitet och visuella texter. Studier om lärare och kortutbil-
dade deltagare i sfi. Doktorsavhandlingar i språkdidaktik 1. Stockholm: Stock-
holms universitet.
Fulcher, Glenn. 1996. Testing tasks: issues in task design and the group oral. Lan-
guage Testing 13(1). Pp. 23–51.
Fulcher, Glenn. 2003. Testing Second Language Speaking. London: Pearson.
Gal, Susan. 2006. Migration, Minorities and Multilingualism: Language Ideologies in
Europe. In: Clare Mar-Molinero & Patrick Stevenson (eds) Language Ideologies,
Policies and Practices. Language and the Future of Europe. New York: Palgrave
Macmillian. Pp.13–27.
Gal, Susan. 2018. Discursive struggles: a commentary. Language & Communication
59. Pp. 66–69.
Galaczi, Elevlina D. 2008. Peer-peer Interaction in a Speaking Test: The Case of the
First Certificate in English Examination. Language Assessment Quarterly 5 (2).
Pp. 89–119.
Galaczi, Evelina D. 2014. Interactional Competence across Proficiency Levels: How
do Learners Manage Interaction in Paired Speaking Tests? Applied Linguistics
35(5). Pp. 553–574.

86
Galaczi, Evelina & Lynda Taylor. 2018. Interactional competence: Conceptualisa-
tions, operationalisations, and outstanding questions. Language Assessment
Quarterly 15(3). Pp. 219–236.
Gan, Zhengdong & Chris Davidson. 2011. Gestural behavior in group oral assess-
ment: A case study of higher- and lower-scoring students. International Journal
of Applied Linguistics 21(1). Pp. 94–120.
Gardner, Rod & Johannes Wagner. 2004. Second language conversations. Lon-
don/New York: Continuum.
Goffman, Erving. 1959. The Presentation of Self in Everyday Life. London: Penguin
Books.
Goffman, Erving. 1967. Interaction Ritual: Essays on Face-to-Face Behavior. Ga-
rolen City: Anchor/Doubleday.
Goodwin, Charles. 2000. Action and embodiment within situated human interaction.
Journal of Pragmatics 32. Pp. 1489–1522.
Goodwin, Charles. 2004. A Competent Speaker Who Can’t Speak: The Social Life of
Aphasia. Journal of Linguistic Anthropology 14(2). Pp. 151–170.
Goodwin, Charles. 2013. The co-operative, transformative organization of human ac-
tion and knowledge. Journal of Pragmatics 46. Pp. 8–23.
Goodwin, Charles & Alessandro Duranti. 1992. Rethinking context. Language as an
interactive phenomenon. Cambridge: Cambridge University Press.
Goodwin, Marjorie Harness & Charles Goodwin. 1986. Gesture and coparticipation
in the activity of searching for a word. Semiotica 62. Pp. 51–75.
Granberg, Nils. 2001. The Dynamics of Second Language Learning. A longitudinal
and qualitative study of an adult’s learning of Swedish. Umeå: Umeå University.
Greer, Tim. 2013. Word search sequences in bilingual interaction: Codeswitching and
embodied orientation toward shifting participant constellations. Journal of Prag-
matics 57. Pp. 100–117.
Gullberg, Marianne. 2011. Multilingual Multimodality: Communicative Difficulties
and Their Solutions in Second-Language Use. In: Jürgen Streeck, Charles Good-
win & Curtis LeBaron (eds) Embodied Interaction. Language and Body in the
Material World. Cambridge: Cambridge University Press. Pp. 137–151.
Habermas, Jürgen. 1970. Towards a theory of communicative competence. Inquiry
13:1-4. Pp. 360375.
Hanell, Linnea. 2017. The knowledgeable Parent. Ideologies of Communication in
Swedish Health Discourse. Stockholm Studies in Scandinavian Philology. New
Series 65. Stockholm: Stockholm Univeristy.
Hanks, William. 1996. Language and communicative practices. London/New York:
Routledge.
Harding, Luke. 2014. Communicative language testing: Current issues and future re-
search. Language Assessment Quarterly 11(2). Pp. 186–197.
Harris, Roy. 1981. The Language Myth. London: Duckworth.
Have, Paul ten. 2007. Doing Conversation Analysis. A Practical Guide. Los Angeles:
Sage.
Hayashi, Makoto. 2003. Language and the body as resources for collaborative action:
A study of word searches in Japanese conversation. Research on language and
social interaction 36(2). Pp. 109–141.
Helasuvo, Marja-Liisa, Minna Laakso & Marja-Leena Sorjonen. 2004. Searching for
words: Syntactic and sequential construction of word search in conversations of

87
 Finnish speakers with aphasia. Research on Language and Social Interaction 37.
Pp. 1–37.
Heller, Monica. 2007. Bilingualism as ideology and practice. In: Monica Heller (ed.)
Bilingualism – a social approach. New York: Palgrave. Pp. 1–22.
Heritage, John. 2007 Intersubjectivity and Progressivity in Person (and Place) Refer-
ence. In: Nick J. Enfield & Tanya Stivers (eds) Person Reference in Interaction:
Linguistic, Cultural, and Social Perspectives. Cambridge: Cambridge University
Press.
Heughs, Rebecca. 2004. Testing the visible: literate biases in oral language testing.
Journal of Applied Linguistics 1(3). Pp. 295–309.
Hill, Kathryn & Tim McNamara. 2011. Developing a comprehensive, empirically
based research framework for classroom-based assessment. Language Testing
29(3). Pp. 395–420.
Hogan-Brun, Gabriella, Clare Mar-Molinero & Patrick Stevenson. 2009. Discourses
on Language and Integration. Amsterdam and Philadephia: John Benjamins.
Howe, Leo. 2000. Risk, Ritual and Performance. The Journal of the Royal Anthropo-
logical Institute 6(1). Pp. 63–79.
Hult, Francis M. 2004. Planning for multilingualism and minority language rights in
Sweden. Language Policy 3(2). Pp. 181–201.
Hyltenstam, Kenneth. 1996. Förord. In: Kenneth Hyltenstam (red.) Tvåspråkighet
med förhinder? Invandrar- och minoritesundervisning i Sverige. Lund: Student-
litteratur. Pp. 5–7.
Hyltenstam, Kenneth & Tommaso Milani. 2012. Flerspråkighetens sociopolitiska och
sociokulturella ramar. In: Kenneth Hyltenstam, Monica Axelsson & Inger Lind-
berg (red.) Flerspråkighet – en forskningsöversikt. Vetenskapsrådets rapportserie
5:2012. Pp.17–134.
Hymes, Dell H. 1967. Models of the Interaction of Language and Social Setting. Jour-
nal of Social Issues 23(2). Pp. 8–28.
Hymes, Dell H. 1971. Competence and Performance in Linguistic Theory. In: Renira
Huxely & Elisabeth Ingram (eds) Language Acquisition: Models and Methods.
Proceedings a C.A.S.D.D. Study Group on Mechanisms of Language Develop-
ment. London & New York: Academic Press. Pp. 3–28.
Hymes, Dell H. 1972a. On Communicative Competence. In: J.B. Pride & Janet
Holmes (eds) Sociolinguistics. London: Penguin Books. Pp. 269–293.
Hymes, Dell H. 1972b. Models of the Interaction of Language and Social Life. In:
John Gumperz & Dell Hymes (eds) Directions in Sociolinguistics. The Ethnog-
raphy of Speaking. New York: Holt, Rinehart and Winston. Pp. 35-71.
Hymes, Dell. 1975. Breakthrough into Performance. In: Dan Ben-Amos & Kenneth
S. Goldstein (eds) Folklore. Performance and Communication. The Hauge: De
Gruyter Mouton. Pp. 11–74.
Ihre, Don. 1973. Singing the world: Language and perception. In: Garth Gillian (ed.)
The Horizons of the Flesh. Critical Perspectives on the thought of Merleau-Ponty
Carbondale and Edwardsville: Southern Illinois University Press. Pp. 61–77.
Irvine, Judith. 1989. When talk isn’t cheap: Language and Political Economy. Amer-
ican Ethnologist 16 (2). Pp. 248–267.
Irvine, Judith & Susan Gal. 2000. Language Ideology and Linguistic differentiation.
In: Paul V. Kroskrity (ed.) Regimes of Language. Ideologies, Polities and Identi-
ties. Santa Fe: School of American Research Press. Pp. 35–83.

88
Iwashita, Noriko, Annie Brown, Tim McNamara & Sally O’Hagan. 2008. Assessed
Levels of Second Language Speaking Proficiency. How Distinct? Applied Lin-
guistics 29(1). Pp. 24–49.
Jacoby, Sally. & Elinor Ochs. 1995. Co-construction: An introduction. Research on
Language and Social Interaction 28(3). Pp. 171–183.
Jaffe, Alexandra. 2013. Minority language learning and communicative competence:
Models of identity and participation in Corsican adult language courses. Lan-
guage & Communication 33. Pp. 450–462.
Jaffe, Alexandra. 2016. Indexicality, stance and fields in sociolinguistics. In: Nikolas
Coupland (ed.) Sociolinguistics. Theoretical debates. Cambridge: Cambridge
University Press. Pp. 86–112.
Jenkins, Jennifer & Leung, Constantin. 2017. Assessing English as a Lingua Franca.
In: Elana Shohamy, Iair Or & Stephen May (eds) Language Testing and Assess-
ment. Encyclopedia of Language and Education (3rd ed.). Basel: Springer Inter-
national Publishing. Pp. 103–117.
Jenkins, Susan & Isabel Parra. 2003. Multiple Layers of Meaning in Oral Proficiency
Tests: The Complementary Roles of Nonverbal, Paralinguistic, and Verbal Be-
havoiur in Assessment Decisions. The Modern Language Journal 87 (i). Pp. 90–
107.
Jones, Rodney H. 2016. Spoken Discourse. London: Bloomsbury.
Jørgensen, Normann J., Martha S. Karrebæk, Lian M Madsen & Janus S. Møller.
2011. Polylanguaging in Superdiveristy. Diversities Vol. 13(2). Pp. 23–37.
Kane, Michael. 1992. An argument-based approach to validity. Psychological Bulle-
tin 112(3). Pp.527–535.
Karlander, David. 2017. Authentic Language. Övdalsk, metapragmatic exchange and
the margings of Sweden’s linguistic market. Dissertations in Bilingualism 28.
Stockholm: Centre for Research on Bilingualism, Stockholm University.
Karrebæk, Martha S. 2013. ‘Don’t speak like that to her!’: Linguistic minority chil-
dren’s socialization into an ideology of monolingualism. Journal of Sociolinguis-
tics 17(3). Pp. 355–375.
Kataoka, Kuniyoshi, Keiko Ikeda, & Niko Besnier. 2013. Decentering and recentering
communicative competence. Language & Communication 33. Pp. 345–350.
Keerfot, Caroline. 2009. Changing Conceptions of Literacies, Language and Devel-
opment. Implications for the provision of adult basic education in South Africa.
Dissertations in Bilingualism 18. Stockholm: Centre for Research on Bilingual-
ism, Stockholm University.
Khan, Kamran. 2014. Citizenship, securitization and suspicion in UK ESOL Policy.
Working Papers in Urban Languages and Literacies paper 130. London: King’s
College.
Kramsch, Claire. 1986. From language proficiency to interactional competence. The
Modern Language Journal 70(4). Pp. 366–372.
Kramsch, Claire. 2009. The Multilingual Subject. Oxford: Oxford University Press.
Kramsch, Claire & Anne Whiteside. 2008. Language Ecology in Multilingual Set-
tings. Towards a Theory of Symbolic Competence. Applied Linguistics 29(4). Pp.
645–671.
Kroskrity, Paul V. 2000. Regimenting Languages: Language Ideological Perspec-
tives. In: Paul V. Kroskrity (ed.) Regimes of Language. Ideologies, Polities and
Identities. Santa Fe: School of American Research Press. Pp. 1–34.

89
Kurhila, Salla. 2006. Second language interaction. Amsterdam/Philadelphia: John
Benjamins.
Kvale, Steinar & Svend Brinkmann. 2014. Den kvalitativa forskningsintervjun. Tredje
upplagan. Lund: Studentlitteratur.
Lazaraton, Anne. 2002. A Qualitative Approach to the Validation of Oral Language
Tests. Studies in Language Testing 14. Cambridge: Cambridge University Press.
Lazaraton, Anne. 2014. Spoken Discourse. The Companion to Language Assessement.
Volume III Evaluation, Methodology, and Interdisciplinary Themes. Antony John
Kunnan (ed.). Hoboken: Wiley Blackwell. Pp. 1375–1389.
Lazaraton, Anne & Larry Davies. 2008. A Microanalytic Perspective on Discourse,
Proficiency, and Identity in Paired Oral Assessment. Language Assessment Quar-
terly 5(4). Pp. 313–335.
Leung, Constantin. 2005. Convivial communication: recontextualizing communica-
tive competence. International Journal of Applied Linguistics 15(2). Pp. 119–
144.
Leung, Constantin. 2014. Communication and participatory involvement in linguisti-
cally diverse classrooms. In: Stephen May (ed.) The Multilingual Turn. Implica-
tions for SLA, TESOL and Bilingual Education. London and New York:
Routledge. Pp. 123–146.
Lillis, Theresa. 2006. Communicative competence. In: Keith Brown (ed.) Encyclope-
dia of Language and Linguistics. New York: Elsevier Science. Pp. 666–673.
Lin, Fenlan. 2014. A conversation-analytic study of word searches in EFL classrooms.
Unpublished thesis. Newcastle: University of Newcastle.
Linell, Per. 2005.The Written Language Bias in Linguistics: Its Nature, Origins and
Transformations. Routledge Advances in Communication and Linguistic Theory.
London/New York: Routledge.
Linell, Per, 2009. Rethinking Language, Mind, and Wold Dialogically. Interactional
and Contextual Theories of Human Sense-Making. Charlotte: Information Age
Publishing.
Linell, Per. 2011. Samtalskulturer. Kommunikativa verksamhetstyper i samhället.
Linköping: Institutionen för kultur och kommunikation, Linköpings universitet.
Lindberg, Inger. 1995. Second Language Discourse in and out of Classrooms. Studies
of learner discourse in the acquisition of Swedish as a second language in edu-
cational contexts. Dissertation in Bilingualism 2. Stockholm: Centre for Research
on Bilingualism, Stockholm University.
Lindberg, Inger. 1996. Svenskundervisning för vuxna invandrare (sfi). In: Kenneth
Hyltenstam (red.) Tvåspråkighet med förhinder? Invandrar- och minoritesunder-
visning i Sverige. Lund: Studentlitteratur. Pp. 224–284.
Lindberg, Inger. 2004. Andraspråksresan. Stockholm: Folkuniversitets förlag.
Lindberg, Inger. 2005. Språka samman. Om samtal och samarbete i språkundervis-
ning. Andra upplagan. Stockholm: Natur och Kultur.
Lindberg, Inger & Karin Sandwall. 2007. Nobody’s darling? Swedish for adult immi-
grants: A critical perspective. Prospect 22(3). Pp. 79–95.
Lindberg, Inger & Karin Sandwall. 2012. Samhälls- och undervisningsperspektiv på
svenska som andraspråk för vuxna invandrare. In: Kenneth Hyltenstam, Monica
Axelsson & Inger Lindberg (red.) Flerspråkighet – en forskningsöversikt. Veten-
skapsrådets rapportserie 5:2012. Pp. 368–478.
Lindberg, Inger & Karin Sandwall. 2017. Conflicting Agendas in Swedish Adult Sec-
ond Language Education. In: Caroline Kerfoot & Kenneth Hyltenstam. Entangled

90
 Discourses: South-North Orders of Visibility. London/New York: Routledge. Pp.
119–136.
Luk, Jasmine. 2010. Talking to Score: Impression Management in L2 Oral Assess-
ment and the Co-Construction of a Test Discourse Genre. Language Assessment
Quarterly 7(1). Pp. 25–53.
Lundahl, Christian. 2011. Bedömning för lärande. Stockholm: Norstedts.
Lundgren, Berit, Jenny Rosén & Anette Jahnke. 2017. 15 års forskning om sfi – en
överblick. Ifous rapportserie 2017:3. Nyköping: Ifous.
Luoma, Sari. Assessing Speaking. Cambridge Language Assessment Series. Cam-
bridge: Cambridge University Press.
Lynch, Brian. 2001. Rethinking assessment from a critical perspective. Language
Testing 18. Pp.351–372.
Majlesi, Ali Reza. 2014. Learnables in Action: The Embodied Achievement of Oppor-
tunities for Teaching and Learning in Swedish as a Second Language Class-
rooms. Studies in Language and Culture No. 24. Linköping: Linköping Univer-
sity.
Martín Rojo, Luisa & Rosina Marquez Reiter. 2013. Position paper WG2- Language
and Migration. COST Action. New Speakers in a Multilingual Europe – Oppor-
tunities and Challenges.
May, Lyn. 2009. Co-constructed interaction in a paired speaking test: The rater’s per-
spective. Language Testing 26(3). Pp. 397–421.
May, Lyn. 2011. Interactional competence in a Paired Speaking Test: Features Salient
to Raters. Language Assessment Quarterly 8(2). Pp. 127–145.
May, Stephen. 2014. Disciplinary Divides, Knowledge Construction, and the Multi-
lingual Turn. In: Stephen May (ed.) The Multilingual Turn. Implications for SLA,
TESOL and Bilingual Education. London/New York: Routledge. Pp. 7–31.
Menken, Kate. 2017. High-Stakes Tests as De Facto Language Education Policies. In:
Elana Shohamy, Iair Or & Stephen May (eds) Language Testing and Assessment.
Encyclopedia of Language and Education (3rd ed.). Basel: Springer International
Publishing. Pp. 385–396.
McNamara, Tim. 1996. Measuring second language performance. London and New
York: Longman.
McNamara, Tim. 1997. ’Interaction’ in second language performance assessment:
Whose performance? Applied Linguistics 18(4). Pp. 446–466.
McNamara, Tim. 2001. Language assessments as social practice: challenges for re-
search. Language Testing 18(4). Pp. 333–349.
McNamara, Tim. 2011. Managing learning: Authority and language assessment. Ple-
nary Speech at the Third International CELC Symposium. National University of
Singapore. Cambridge: Cambridge University Press.
McNamara, Tim. 2012. Language assessments as shibboleths: A post-structural per-
spective. Applied Linguistics 33. Pp. 564–581.
McNamara, Tim & Carsten Roever. 2006. Language Testing: The Social Dimension.
Language Learning Monograph Series. Malden: Blackwell Publishing.
Messick, Sameul. 1989. Meaning and Values in Test Validation: The Science and
Ethics of Assessment. Educational Researcher 18(2). Pp. 5–11.
Merleau-Ponty, Maurice. 2012[1945]. Phenomenology of Perception. London/New
York: Routledge.

91
Milani, Tommaso. 2007. Debating Swedish. Language Politics and Ideology in Con-
temporary Sweden. Dissertations in Bilingualism 14. Stockholm: Centre for Re-
search on Bilingualism, Stockholm University.
Milani, Tommaso. 2008. Language Testing and Citizenship: A Language Ideological
Debate in Sweden. Language in Society 37(1). Pp. 27–59.
Miller, Elizabeth R. 2014. The language of Adult Immigrants: Agency in the Making.
New Perspectives on Langauge and Education 39. Bristol: Multilingual Matters.
Mondada, Lorenza. 2016. Challenges of multimodality: Language and the body in
social interaction. Journal of Sociolinguistics 20(3). Pp. 336–366.
Nakatsuhara, Fumiyo. 2013. The Co-construction of Conversation in Group Oral
Tests. Frankfurt: Peter Lang.
National Agency for Education. 1997. Vem älskar sfi? Utvärdering av svenskunder-
visningen för invandrare – en utbildning mellan två stolar. Skolverkets rapport
nr 131.
National Agency for Education. 2006a. Det nationella provets funktion och status
inom sfi. En intervjustudie. Rapport 289.
National Agency for Education. 2006b. Samtal på gång. En vägledning för bedöm-
ning av muntliga färdigheter.
National Agency for Education. 2009. SKOLFS 2009:2. Förordning om kursplan för
svenskundervisning för invandrare.
National Agency for Education. 2012a. SKOLFS 2012:13. Förordning om ändring i
förordningen (SKOLFS 2009:2) om kursplan för svenskundervisning för invand-
rare.
National Agency for Education. 2012b. Samtal på gång. En vägledning för bedöm-
ning av muntlig produktion och interaktion. Utbildning i svenska för invandrare,
kurs D. Andra utgåvan.
National Agency for Education. 2013a. Beskrivande data. Förskola, skola och vuxen-
utbildning. Rapport nr. 383.
National Agency for Education 2013b. Attityder till vuxenutbildningen 2012. Elever
och lärares attityder till vuxenutbildningen. Rapport 389.
National Agency for Education. 2014. Elever och studieresultat i sfi 2013. Dnr:
2014:00047.
National Agency for Education. 2017a. Skolverkets föreskrifter för kommunal vux-
enutbildning i svenska för invandrare. SKOLFS 2017:91.
National Agency for Education. 2017b. PM. Elever och studieresultat i kommunal
vuxenutbildning i svenska för invandrare år 2016. Dnr 2017:00824.
National Agency for Education. 2018. Uppföljning av språkintroduktion. Beskrivande
statistik på nationell nivå och nyanlända elevers övergångar till och från språk-
introduktion. Rapport 469.
National Agency for Education. N.d. Exempel på uppgifter ur delprovet Tala.
https://www.skolverket.se/bedomning/nationella-prov/alla-nationella-prov-i-
skolan/utbildning-i-svenska-for-invandrare/exempeluppgifter/exempel-pa-
uppgifter-ur-delprovet-tala-1.203273 (Downloaded 8 march 2018).
Nilsson, Åke. 2004. Efterkrigstidens invandring och utvandring. Demografisk rapport
2004:5. Stockholm: Statistiska centralbyrån.
Nitta, Ryo & Fumiyo Nakatsuhara. 2014. A multifaceted approach to investigating
pre-task planning effects on paired oral assessments. Language Testing 31(2). Pp.
147–175.

92
Norlund Shaswar, Annika. 2014. Skriftbruk i vardagsliv och sfi-utbildning: en studie
av fem kurdiska sfi-studerandes skriftbrukssituation och skriftpraktiker. Umeå:
Umeå Universitet.
Norton, Bonny. 2013. Identity and Language Learning. Extending the Conversation.
2nd Edition. Bristol: Multilingual Matters.
Norton, Julie. 2013. Performing Identities in Speaking Tests: Co-Construction Revis-
ited. Language Assessment Quarterly 10(3). Pp. 309–330.
Nunan, David. 1991. Communicative tasks and the language curriculum. TESOL
Quarterly 25(2). Pp. 279–295.
Nyroos, Lina, Erica Sandlund & Pia Sundquist. 2017. Code-switched repair initiation:
The case of Swedish eller in L2 English test interaction. Journal of Pragmatics
120. Pp. 1–16.
Ochs, Elinor. 1979. Transcription as Theory. In: Elinor Ochs & Bambi Schieffelein,
(eds) Developmental Pragmatics. New York: Academic Pr. Pp 43–72.
Ochs, Elinor & Lisa Capps. 1996. “Narrating the Self”. Annual Review of Anthropol-
ogy 25(1). Pp. 19–43.
Ockey, Gary. 2009. The effects of group members’ personalities on a test taker’s L2
group oral discussion test scores. Language Testing 26(2). Pp. 161–186.
Olson, Maria & Magnus Dahlstedt. 2014. Citizen formation for a new millennium in
Sweden – a prognosis of our time. The Journal for Critical Education Policy
Studies 12(2). Pp. 200–224.
Olson, Maria, Magnus Dahlstedt, Andreas Fejes & Fredrik Sandberg. 2018. The for-
mation of the willing citizen – Tracing reactive nihilism in the late capitalist edu-
cation. Educational Philosophy and Theory 50(1). Pp. 95–103.
Ortega, Lourdes. 2014. Ways Forward for a Bi/Multilingual Turn in SLA. In: Stephen
May (ed.) The Multilingual Turn. Implications for SLA, TESOL and Bilingual
Education. London and New York: Routledge. Pp. 32–53.
O’Sullivan, Barry. 2002. Learner acquaintanceship and oral proficiency pair task per-
formance. Language Testing 19(3). Pp. 277–295.
Park, Joseph Sung-Yul. 2010. Naturalization of Competence and the Neoliberal Sub-
ject: Success Stories of English Language Learning in the Korean Conservative
Press. Journal of Linguistic Anthropology 20(1). Pp. 22–38.
Park, Joseph Sung-Yul & Lionel Wee. 2008. Approriating the language of the other:
Performativity in autonomous and unified markets. Language & Communication
28. Pp. 242–257.
Parkhouse, Anna. 2016. Asyl- och migrationspolitik: väktare av den svenska mo-
dellen. In: Daniel Silander & Mats Öhlén (eds) Svensk politik och EU: Hur den
svensk politik har förändrats av medlemskapet i EU. Stockholm: Santérus Förlag.
Pp. 171-192.
Pavlenko, Aneta. 2006. Emotions and Multilingualism. Cambridge: Cambridge Uni-
versity Press.
Pennycook, Alistair. 2012. Language and mobility. Unexpected Places. Bristol: Mul-
tilingual Matters.
Pennycook, Alistair. 2016. Mobile times, mobile terms: The trans-super-poly-metro
movement. In: Nikolas Coupland (ed) Sociolinguistics. Theoretical debates.
Cambridge: Cambridge University Press. Pp. 201–216.
Perren, George. 1967. Testing ability in English as a second language: 3. Spoken lan-
guage. English Language Teaching 22(1). Pp. 22–29.
Pratt, Mary Louise. 1991. Arts of the contact zone. Profession 91. Pp. 33–40.

93
Pujolar, Joan. 2007. Bilingualism and the Nation-State in the Post-National Era. In:
Monica Heller (ed.) Bilingualism – a social approach. New York: Palgrave. Pp.
71–95.
Purpura, James. 2008. Assessing communicative language ability: Models and their
components. In: Elana Shohamy & Nancy Hornberger (eds) Encyclopedia of Lan-
guage and Education. 2nd Edition, Volume 7: Language Testing and Assessment.
Berlin: Springer Science. Pp. 53–68.
Rampton, Ben. 1990. Displacing the “native speaker”: Expertise, affiliation and in-
heritance. ELT Journal 44. Pp. 97–101.
Rampton, Ben. 2017. Interactional Sociolinguistics. Working Papers in Urban Lan-
guages and Literacies paper 205. London: King’s College.
Rhymes, Betsy. 2014. Communicating Beyond Language Everyday Encounters with
Diversity. London/New York: Routledge.
Roberts, Celia, Melanie Cooke, Mike Baynham & James Simpson. 2007. Adult ESOL
in the United Kingdom: Policy and research. Prospect 2(3). Pp. 18–31.
Roberts, John. 2004. The communicative approach to language teaching. The king is
dead! Long live the king! IJES 4(1). Pp. 1–37.
Rosén, Jenny. 2013. Svenska för invandrarskap? Språk, kategorisering och identitet
inom utbildningsformen Svenska för invandrare. Örebro Studies in Education 38.
Örebro: Örebro universitet.
Rosén, Karin & Sangeeta Bagga-Gupta. 2013. Shifting identity positions in the devel-
opment of language education for immigrants: an analysis of discourses associ-
ated with ’Swedish for immigrants’. Language, Culture and Curriculum 26(1).
Pp. 68–88.
Rosén, Karin & Sangeeta Bagga-Gupta. 2015. Prata svenska vi är i Sverige! [Talk
Swedish, we are in Sweden!] A study of practiced language policy in adult language
learning. Linguistics and Education 31. Pp. 59–73.
Runblom, Harald. 1994. Swedish Multiculturalism in a Comparative European Per-
spective. Sociological Forum 9(4). Pp. 623–640.
Rydell, Maria. 2016. Konsten att föra ett samtal framåt. Om bedömning av muntlig
interaktion på sfi. In: Björn Kindenberg (ed.) Flerspråkighet som resurs. Sympo-
sium 2015. Stockholm: Liber. Pp. 236–247.
Sacks, Harvey, Emanuel Schegloff & Gail Jefferson. 1974. A Simplest Systematic for
the Organization of Turn-Taking for Conversation. Language 50(4). Pp. 696–735.
Sahlée, Anna. 2017. Språket och skolämnet svenska som andraspråk: Om elevers och
skolans språksyn. Skrifter utgivna av Institutionen för nordiska språk vid Uppsala
universitet 101. Uppsala: Uppsala universitet.
Salö, Linus. 2015. The linguistic sense of placement: Habitus and the entextualization
of translingual practices in Swedish Academia. Journal of Sociolinguistics 19(4).
Pp. 511–534.
Salö, Linus. 2016. Languages and Linguistic Exchanges in Swedish Academia. Dis-
sertations in Bilingualism 26. Stockholm: Centre for Research on Bilingualism,
Stockholm University.
Salö, Linus. 2018. Seeing the point from which you see what you see: An essay on
epistemic reflexivity in language research. Multilingual Margins 5(1). Pp. 24–39.
Sandberg, Fredrik, Andreas Fejes, Magnus Dahlstedt & Maria Olson. 2016. Adult
Education as Heterotopia of Deviation: A Dwelling for the Abnormal Citizen.
Adult Education Quarterly 66(2). Pp. 103–119.

94
Sandlund, Erica & Pia Sunquist. 2013. Divergent task orientations in L2 oral profi-
ciency tests – a conversation analytic approach to participant understandings of
pre-set discussion tasks. Nordic Journal of Modern Language Methodology 2(1).
Pp. 1–21.
Sandwall, Karin. 2013. Att hantera praktiken. Om sfi-studerandes möjligheter till in-
teraktion på praktikplatser. Göteborgsstudier i nordisk språkvetenskap 20. Göte-
borg: Göteborgs universitet.
Savignon, Sandra. 1972. Communicative competence: an experiment in foreign lan-
guage teaching. Philadelphia: Center for Curriculum Development.
Savignon, Sandra. 1983. Communicative Competence: Theory and Classroom Prac-
tice. Texts and Contexts in Second Language Learning. Reading, Massachusetts:
Addisin-Wesley Publishing Company.
Savignon, Sandra. 2006. Communicative Language Teaching. In: Keith Brown (ed.)
Encyclopedia of Language and Linguistics. New York: Elsevier Science. Pp.
673–679.
Saville, Nick. 2009. Language assessment in the management of international migra-
tion: A framework for considering the issues. Language Assessment Quarterly
6(1). Pp. 17–29.
Scieffelin, Edward. 1985. Performance and the cultural construction of reality. Amer-
ican Etnologist 12. Pp. 691–706.
Seidlhofer, Barbara. 2009. Common ground and different realities: World Englishes
and English as a lingua franca. World Englishes 28(2). Pp. 236–245.
Sheikhi, Karin. 2013. Vägar till förståelse. Andraspråkstalare i samtal med en studie-
och yrkesvägledare. Göteborgsstudier i nordisk språkvetenskap 21. Göteborg:
Göteborgs universitet.
Shohamy. Elana. 2001. The Power of Tests. A Critical Perspective on the Uses of
Language Tests. Essex: Pearson Education Limited.
Shohamy, Elana. 2006. Language policy: Hidden agendas and new approaches. Lon-
don/New York: Routledge.
Shohamy, Elana. 2011. Assessing Multilingual Competencies: Adopting Construct
Valid Assessment Policies. The Modern Language Journal, 95 iii. Pp. 418–429.
Shohamy, Elana. 2013. The discourse of language testing as a tool for shaping na-
tional, global and transnational identities. Language and Intercultural Communi-
cation. 13(2). Pp. 225–236.
Shohamy, Elana. 2017. Critical Language Testing. In: Elana Shohamy, Iair Or & Ste-
phen May (eds) Language Testing and Assessment. Encyclopedia of Language
and Education (3rd ed.). Basel: Springer International Publishing. Pp. 441–454.
Siegl, Aki. 2016. Longitudinal development of word search sequences in English as
a lingua franca interaction. Unpublished thesis. Newcastle: Newcastle Univer-
sity.
Silverman, David & Anssi Peräkylä. 1990. AIDS counselling: the interactional organ-
isation of talk about ’delicate issues’. Sociology of Health and Illness 12(3). Pp.
293–318.
Silverstein, Michael. 1996. Monoglot “Standard” in America: Standardization and
Metaphors of Linguistic Hegemony. In: Donald Brenneis & Ronald K.S. Macau-
lay (eds) The Matrix of Language. Contemporary Linguistic Anthropology. Boul-
der: Wesview Press.
Simpson, James. 2006. Differing expectations in the assessment of the speaking skills
of ESOL learners. Linguistics and Education 17. Pp. 40–55.

95
Sjögren, Annick. 1996. Språket, nykomlingens nyckel till samhället men också en
svensk försvarsmekanism. In: Annik Sjögren, Ann Runfors & Inrgid Ramberg
(red.) En ”bra ” svenska? Om språk, kultur och makt. Stockholm: Mångkulturellt
centrum. Pp. 19–40.
Skar, Gustaf. 2013. Skrivbedömning och validitet. Fallstudier av skrivbedömning i
svenskundervisning på gymnasiet. Doktorsavhandlingar i språkdidaktik 2. Stock-
holm: Stockholms universitet.
SKOLFS 2017:91. Skolverkets föreskrifter för kommunal vuxenutbildning i svenska
för invandrare..
Spolsky, Bernard. 1990. Oral examinations: an historical note. Language Testing 7(2).
Pp. 158–173.
Spolsky, Bernard. 1995. Measured Words. Oxford: Oxford University Press.
Spolsky, Bernard. 1997. The etics of gatekeeping tests: What have we learned in a
hundred years? Language testing 14(3). Pp. 242–247.
Spolsky, Bernard. 2017. History of Language Testing. In: Elana Shohamy, Iair Or &
Stephen May (eds) Language Testing and Assessment. Encyclopedia of Language
and Education (3rd ed.). Basel: Springer International Publishing. Pp. 375–384.
Spotti, Massimiliano. 2011. Ideologies of Success for Superdiverse Citizens: the
Dutch Testing Regime for integration and the Online Private Sector. Diversities
13. Pp. 39–52.
Statistics Sweden. 2013. Utrikes födda 2012. Fortsatt ökning av utrikes föda i Sverige.
https://www.scb.se/sv_/Hitta-statistik/Artiklar/Fortsatt-okning-av-utrikes-fodda-
i-Sverige/ (Downloaded October 2017).
Statistics Sweden. 2016. Befolkningsstatistik. http://www.sverigeisiffror.scb.se/hitta-
statistik/sverige-i-siffror/manniskorna-i-sverige/befolkningsutveck-
ling/#story_7dc62dbd-e2db-4555-a31f-bdc6a58436e8 (Downloaded October
2017).
Stevenson, Douglas. 1985. Authenticity, validity and a tea party. Language Testing
2(1). Pp. 41–47.
Stewart, D. W., Prem Shamdasani & Dennis Rook. (2007). Focus Groups: Theory
and Practice, 2nd edition. Applied Social Research Method, Vol. 29. Thousand
Oaks, CA: Sage Publications.
Stockholm stad. 2012. Svenskundervisning för invandrare.
http://www.stockholm.se/ForskolaSkola/Svenskundervisning-for-invandrare-sfi/SFI-
bonus/Att-gora-SFI-test-/ (downloaded 21 March 2012)
Stockholms stad. 2018. Prov och betyg sfi. http://www.stockholm.se/Fristaende-
webbplatser/Fackforvaltningssajter/Arbetsmarknadsforvaltningen/utbildning-
sanordnare/Anvisningar-riktlinjer-och-rapporteringsmallar/Prov-och-betyg-sfi/
(downloaded 14 June 2018).
Stockholm University & National Agency for Education. 2013. Rapport och uppfölj-
ning av nationella slutprov. Utbildning i svenska för invandrare. Stockholm: Stock-
holm University.
Stroud, Christopher. 2004. Rinkeby Swedish and semilingualism in language ideolog-
ical debates: A Bourdieuean perspective. Journal of Sociolinguistics 8(2). Pp.
196–214
Strömbäck, Jesper, Felicia Andersson & Evelina Nedlund. 2017. Invandring i medi-
erna – Hur rapporterade svenska tidningar åren 2010-2015? Delmi rapport
2017:6. Stockholm: Delegationen för migrationsstudier.
Sundberg, Gunlög. 2004. Asymmetrier och samförstånd i rekryteringssamtal med

96
 andraspråkstalare. Stockholm studies in Scandinavian philology. New Series
38. Stockholm: Stockholms universitet.
Sundquist, Pia, Peter Wikström, Erica Sandlund & Lina Nyroos. 2018. The teacher as
examiner of L2 oral tests: A challenge to standardization. Language Testing
35(2). Pp. 217–238.
Svanberg, Ingvar & Mattis Tydén. Tusen år av invandring. En svensk kulturhistoria.
Stockholm: Dialogos.
Swain, Merrill. 2001. Examining dialogue: another approach to content specification
and to validating inferences drawn from test scores. Language Testing 18(3). Pp.
275–302.
Swedish Public Radio. 2018. ”Skärp kraven på svenska för invandrare” http://sveri-
gesradio.se/sida/artikel.aspx?programid=83&artikel=6863550 (Downloaded
January 17 2018)
Talmy, Steven. 2011. The Interview as Collaborative Achievement: Interaction, Iden-
tity and Ideology in a Speech Event. Applied Linguistics 32(1). Pp. 25-42.
Taylor, Lynda, 2001. The paired speaking test format: recent studies. Research Notes
6. Cambridge University ESOL Examinations. Cambridge: Cambridge Univer-
sity. Pp.15–17.
Taylor, Lynda 2003. The Cambridge Approach to speaking assessment. Research
Notes 13. Cambridge University ESOL Examinations. Cambridge: Cambridge
University. Pp. 2–4.
Taylor, Lynda & Gillian wigglesworth. 2009. Special Issue: Pair Work in L2 Assess-
met. Language Testing 26(3). Pp. 325–466.
van Avermaet, Piet. 2009. Fortress Europe? Language policy regimes for immigration
and citizenship. In: Gabriella Hogan-Brun, Clare Mar-Molinero & Patrick Ste-
venson (eds) Discourses on Language and Integration. Amsterdam/Philadelphia:
John Benjamins.
van Lier, Leo. 1989. Reeling, writhing, drawing, stretching, and fainting in coils: Oral
proficiency interviews as conversation. TESOL Quarterly 23/3. Pp. 489–508.
Vertovec, Steven. 2006. The Emergence of Super-Diversity in Britain. COMPAS
Working Paper 25. Oxford: Centre on Migration, Policy and Society.
Vertovec, Steven. 2007. Super-diversity and its implications. Ethnic and Racial Stud-
ies 30(6). Pp. 1024-1054.
Vetenskapsrådet, 2011. God forskningssed. Vetenskapsrådets rapportserie 1:2011
http://www.vr.se/dowload/18.3a36c20d133af0c12958000491/1321864357049/
God+forskningssed+2011.1.pdf (downloaded Jauary 15 2012)
Vidaković, Ivana & Evelina Galaczi. 2013. The measurement of speaking ability
1913–2012. In: Cyril J Weir, Ivana Vidaković & Evelina D Galaczi (eds) Meas-
ured Constucts. A history of Cambridge English language examinations 1913-
2012. Studies in Language Testing 37. Cambridge: Cambridge University Press.
Pp. 257–346.
Vološinov. Valentin Nikolaevic 1973[1929]. Marxism and the Philosophy of Lan-
guage. Translated by Ladislav Matejka & I.R. Titunik. Harvard: Harvard Univer-
sity Press.
Weir, Cyril. 2013. An overview of the influences on English language testing in the
United Kingdom 1913–2012. In: Cyril Weir, Ivana Vidaković &
Evelina Galaczi (eds) A histoty of Cambridge English language examinations
1913-2012. Studies in Language Testing 37. Cambridge: Cambridge University
Press. Pp. 1–102.

97
Weir, Cyril, Ivana Vidaković & Evelina Galaczi. 2013. A histoty of Cambridge Eng-
lish language examinations 1913-2012. Studies in Language Testing 37. Cam-
bridge: Cambridge University Press.
Wibeck, Victoria. 2010. Fokusgrupper. Om fokuserade gruppintervjuer som under-
sökningsmetod. Andra upplagan. Lund: Studentlitteratur.
Young, Richard. 2011. Interactional Competence in Language Learning, Teaching
and, Testing. In: Eli Hinkel (ed.) Handbook of research in second language teach-
ing and learning. London/New York: Routledge.
Young, Richard & Agnes Weiyun He. 1998. Talking and Testing. Discourse ap-
proaches to the assessment of oral proficiency. Studies in Bilingualism 14. Am-
sterdam/Philadelphia: John Benjamins
Zachrison, Mozhgan. 2014. Invisible voices. Understanding the sociocultural influ-
ences on adult migrants’ second language learning and communicative interac-
tion. Malmö Studies International Migration and Ethnic Relations No. 12. and
Linköping Studies in Arts and Sciences No. 626 Department of Social and Wel-
fare studies. Linköping: Linköping University.
Åhlund, Anna. 2015. Swedish as multiparty work. Tailoring talk in a second language
classroom. Doctoral dissertation. Stockholm: Department of Child and Youth
studies, Stockholm University.
Åhlund, Anna & Karin Aronsson. 2015. Corrections as multiparty accomplishment in
L2 classroom conversations. Linguistics and Education 30. Pp. 66–80.

Laws and regulations

SFS 2007:1091 Lag om offentlig upphandling [The Procurement Act]
SFS 2009:600 Språklag [The Language Act]
SFS 2010:800 Skollag [The Education Act]
SFS 2010:1030 Förordning om prestationsbaserad stimulansersättning inom
svenskundervisning för invandrare. [Regulation of perfor
mance-based incentive compensation in Swedish for immi-
grants]
SFS 2011:1108 Förordning om vuxenutbildning [Regulation of Adult
Education]
SFS 2016:752 Lag om tillfälliga begränsningar av möjligheten att få uppe
hållstillstånd i Sverige. [Temporary law on restrictions on
residenship in Sweden]
SFS 2017:584 Lag om ansvar för etableringsinsatser för vissa nyanlända
invandrare. [Law on establishment contributions for newly
arrived immigrants]

Commission reports
SOU 1981:86 Svenskundervisning för vuxna invandrare, del 1. Övervä-
ganden och betänkanden. [Swedish tuition for adult immi-
grants, part 1]

98
SOU 1981:87 Svenskundervisning för vuxna invandrare, del 2. Kartlägg-
ning av nuläget. [Swedish tuition for adult immigrants, part
2]
SOU 1999:34 Medborgarskap [Citizenship]
SOU 2011:19 Tid för snabb flexibel inlärning. Utredning om tidsbegräns-
ning i svenskundervisning för invandrare. [Time for flexible
learning. Investigation of time restrictions in Swedish tuition
for adult immigrants]

99
APPENDIX A. Example of a paired task in the final national exam (National Agency of Education
n.d.).

[My translation]: Paired task: Discussion (Course D)

The aim of the paired task is for the test takers to discuss an issue based on a mind-map and come to
an agreement on which factors they believe are the most important. The task also aims to test the test
takers ability to take the discussion forward by putting forward and asking for information and
expressing and receiving views using simple arguments.

Instructions to the test takers. The entire instruction should be read by the examiner:
You should discuss for about 10 minutes.
It’s important that you listen to each other, ask each other questions and ask each other if you don’t
understand.
Keep in mind that the discussion is between the two of you, I will only listen.
What you are going to discuss is this question (hand out the task sheet):
APPENDIX B. Questions used as stating points for the group discussions.

Vilka språk kan ni?

Varför läser ni sfi?
Är det viktigt att kunna svenska? Varför?
Vilka språk använder ni? I vilka sammanhang?
Vad kan man när man kan ett språk bra? Hur vet man det?
Vad är viktigt att kunna?
Hur lär man sig ett språk?
Hur bra är ni på svenska?
Vad är lätt att lära sig? Vad är svårt?
Vad är kommunikativ kompetens?
Vad vill ni göra i framtiden?

What languages do you know?

Why do you study SFI?
Is it important to know Swedish? Why?
What languages do you use? In which contexts?
What is it that you know when you know a language well? How do you know if you know a
language well?
What is important to know?
How do you learn a language?
How good are you in Swedish?
What is easy to learn when learning Swedish? What is difficult?
What is communicative competence?
What do you want to do in the future?