7 Semi-Structured Interviews

Petros Karatsareas

7.1 Introduction

Interviews are among the most widely used methods of data elicitation
in the social sciences. In the earlier stages of their use, up to the 1930s, they were
distinguished from questionnaires (see Chapters 9 and 10) based on whether
participants completed written questionnaires themselves or whether this was
facilitated by an interviewer. As questionnaires became a more established way
of collecting quantifiable data in the 1940s and 1950s, interviews came to be more
firmly grounded in the qualitative realm of research. The 1960s and 1970s saw
researchers engaging in methodological discussions about different aspects of
interviews as contextually and socially situated speech events, including
researcher positionality and the power dynamics between interviewer(s) and
interviewee(s) (see Platt 2012 for an insightful overview of the history of the
interview, albeit with a focus on social research undertaken in the US).
The interview was first used as a method for eliciting linguistic data by Labov
(1966) in his study of English in New York. However, the sociolinguistic
interview, as it came to be known, is different to the type of interview that this
chapter is concerned with. In the Labovian paradigm, the aim of the interview is
to collect multiple and varied occurrences of a linguistic variable in a range of
speech styles, which are later analysed quantitatively to identify patterns of
distribution across different groups of speakers in terms of broad social factors
such as age, gender, and socioeconomic status. In that, the Labovian interviewer
is not as much interested in the content of the participants’ responses to their
questions (or, rather, prompts) as in their productions. In language attitudes
research, by contrast, the focus is primarily the content of participants’ responses
(but see also Chapter 6 for a discussion of instances where production is of
interest). Interviews are thus used in ways similar to other disciplines within the
social sciences, such as linguistic ethnography and anthropology. Their aim is to
elicit information in a direct way about what people believe, think, and feel about
language – and why.
In their most basic form, interviews involve the elicitation of information from
a participant by a researcher in a speech event that resembles a one-to-one
conversation. (For focus group interviews, in which information is elicited from
multiple participants at the same time, see Chapter 8.) Different types of inter-
views can be identified depending on the degree of structuredness of the

99

[Link] Published online by Cambridge University Press

 100 petros karatsa reas

structured unstructured
(yes/no questions, multiple- (open-ended questions,
choice questions, pre-set order) no pre-set order)

semi-structured
(mix of closed and open-ended
-
questions, monitored order)

Figure 7.1 Continuum of structuredness

exchange between the researcher and the participant. At one end of the con-
tinuum (Figure 7.1), there are structured interviews, whereby the researcher asks
all participants in their sample the same questions, each time following the same
pre-set order. Structured interviews include a high number of yes/no and mul-
tiple-choice questions and provide little space for open-ended questions. If open-
ended questions are included, participants’ responses are generally expected to
be brief. At the other end of the continuum are unstructured interviews.
Revolving mostly around open-ended questions, which participants are given
the freedom to answer in any way and to any extent they want to, and jumping
from one topic to another in no particular order or controlled fashion, they are
much like casual conversations.
Semi-structured interviews lie between these two types. While they do not
necessarily follow a pre-set order in covering the topics in which the researcher is
interested, they are designed to ensure that these topics – or, at least, as many
topics as possible – are covered. To this end, the researcher monitors how the
interview is going, using a pre-prepared interview protocol so that, when the
participant’s responses veer away from the agenda, the researcher can employ
discursive strategies to refocus the interview. The researcher can also spontan-
eously ask questions that are not in the protocol, if new and interesting dimen-
sions are revealed during the interview. Semi-structured interviews are mostly
based on open-ended questions that prompt participants to develop their thoughts
and ideas in depth, expressing their views on the subject matter from their own
personal perspectives, talking about their experiences, and using their own
words. Yes/no questions may be asked, but they must almost always be followed
up with open-ended questions that seek to obtain further information in the form
of clarifications, justifications, elaborations, exemplifications, or illustrations.

7.2 Strengths and Limitations

7.2.1 Strengths
Semi-structured interviews can be used at the earliest stages of the
research design to put together the research background, formulate and refine

[Link] Published online by Cambridge University Press

 Semi-Structured Interviews 101

research questions, and generate hypotheses. Interview data can then form the
basis for further investigation, both qualitative and quantitative; for example,
they can inform the design of questionnaire and survey items. Interviews can also
be used at the final stages of research to collect feedback from participants or
assess the impact of a research-based intervention on a group of people (see
Chapter 21 for a more detailed discussion of interviews and other methods as part
of mixed-methods approaches to the study of language attitudes). Many projects,
however, employ interviews as their sole data collection method.
Interviews can elicit information about both status and solidarity as the two
main evaluative dimensions of language attitudes (Woolard 1989; see also
Chapter 1). They can give the researcher direct access to participants’ (own
accounts of their) affect and cognition but only indirect access to conation as
participants may well self-report on their behavioural intentions, but their actual
behaviour in real-life circumstances is not directly observed as part of the
interview (Rosenberg and Hovland 1960; Hilgard 1980; see also Chapter 1).
They can bring to the fore both commonalities and shared understandings across
participants, and different viewpoints around a central theme. Thanks to their
flexibility and relative spontaneity, interviews can not only corroborate or chal-
lenge knowledge that the researcher has formed about the topic at hand prior to
the interview, but they can also bring to light entirely new information, new
topics, or new dimensions to established knowledge. Owing to their direct
nature, researchers can use interviews to build substantial sets of rich and in-
depth data relatively quickly while at the same time having the opportunity to
establish good rapport with participants. Interviews create a space for participants
to talk about their views on language in their own words and to construct their
own narratives around their lived experiences of issues such as the stigmatisation
of their or other people’s language, language-based discrimination, or language
racism. This is certainly another strength of the method. Edley and Litosseliti
(2010) discuss the sense of empowerment that interviews generate among par-
ticipants, and the ability they give them to shape the research agenda by steering
it towards directions that they deem worthwhile. This becomes particularly
important in the case of participants from minoritised, underrepresented, or
disadvantaged groups.

7.2.2 Limitations
Garrett et al. (2003: 24) see in interviews ‘a high degree of obtrusive-
ness’ in that ‘it is the informants themselves who are asked to report their
attitudes’. Another major limitation of semi-structured interviews stems from
the fact that they are not mere question and answer sessions aimed at uncovering
objective facts (or, truths) about people’s views on language. Rather, they are
contextually and socially situated speech events (Heller et al. 2018: 87–89). They
are first and foremost shaped by the spatial and temporal context in which the
interview takes place. The relationship between the interviewer(s) and the

[Link] Published online by Cambridge University Press

 102 petros karatsa reas

interviewee(s) also plays a key role in framing the interaction between them. This
includes how much and what they know about each other prior to the interview
as well as any potential power imbalances between them. Previous knowledge,
experience, and understanding that the interviewer(s) and interviewee(s) have of
the topic to be covered during the interview as well as of the interview as a form
of interaction, are some other important determinants in that respect. In that
connection, Codó (2008) and Edley and Litosseliti (2010) both make the point
that degrees of familiarity with the way interviews work differ across the world
depending on participants’ cultural contexts and socio-economic backgrounds.
Depending on how these dynamics play out during the interview, participants
will constantly tailor their responses in different ways. For example, participants
who belong to linguistically minoritised and disenfranchised groups may be
reluctant to talk about their experiences of linguistic discrimination to an inter-
viewer who belongs to the dominant linguistic group and/or in a physical setting
that is emblematic of linguistic inequality (such as a school classroom or a room
in a library or university building). In other cases, participants may say what they
believe the interviewer wants to hear (social desirability bias) or agree with
interviewers’ questions and prompts, regardless of their content (acquiescence
bias), in order to portray what they see as a more positive image for themselves
(Garrett et al. 2003: 28–29; Krug and Sell 2013: 75). Both these types of bias are
particularly strong in face-to-face methods, including interviews. For example,
people who use non-prestigious forms of language in their speech (e.g. features
of non-standardised varieties, slang terms, or colloquial expressions) may deny
using them at all. Similarly, people might not want to reveal that they hold
strongly negative views towards such forms, especially if their use has racial,
ethnic, and/or class connotations to avoid associating themselves with these
types of discrimination. An added difficulty has to do with the fact that, to many
people, language attitudes seem commonsensical (see also the discussion in
Chapter 1). As Codó (2008: 162) points out, ‘values, attitudes, beliefs, and
motivations tend to be difficult to verbalize [. . .] rarely do speakers reflect on
these issues in an explicit manner unless awareness of language is heightened’.

7.3 Research Planning and Design

Like all research methods, semi-structured interviews need to be

carefully designed and planned before they are put to use, not least because there
are ethical considerations that apply to interviews, which make their careful
design even more crucial. Interview participants agree to give up their time to
answer questions on the spot about topics that might make them feel uncomfort-
able, and in a context where they might find it difficult to stop participating (even
if the researcher has previously informed them that they may do so at any time
without having to specify the reason). Researchers must therefore ensure that
they take note of the efforts participants make to be interviewed, and they should

[Link] Published online by Cambridge University Press

 Semi-Structured Interviews 103

clarify that they recognise them. These and other ethical aspects of conducting
language attitude interviews – such as obtaining informed consent from partici-
pants – need to be taken into account as part of securing ethics clearance from
researchers’ institutions or other appropriate authorities (see the contributions in
De Costa 2016 on how to address ethical issues in applied linguistics research,
including interviews).
When designing an attitudes project that includes interviews, researchers need
to establish a critical body of knowledge about the social, cultural, political,
economic, and linguistic circumstances of the group whose attitudes are to be
investigated, paying equal attention both to the group’s past and to their present
(Schreier 2013). They need to become familiar with everyday aspects of their
participants’ lives and, if possible, also participate in some of the social activities
that their potential interviewees engage in. In this way, they will develop a solid
understanding of the research background, and they will become known to their
participants and start building relations of trust with them. Their awareness of
what is and what is not considered appropriate behaviour in the group will also
increase. These insights will inform the overall design of the project and that of
the interviews in particular. Familiarity with, and involvement in, their partici-
pants’ community will additionally help researchers to recruit people to be
interviewed. When there is a risk that the researcher is perceived as an outsider
to the group, and especially as coming from a linguistically dominant group, it
might be necessary that the researcher is introduced into the community by a
gatekeeper – that is, a person of respect and power in the community, who will
need to be sympathetic to the researcher’s project in order to vouch publicly for
their legitimacy and trustworthiness. In any case, researchers always need to be
honest about the purposes of their research and explain from the outset – with
clarity and in as much detail as possible – what participation in the interviews
would entail for participants.
A key element of interview preparation is the interview protocol – that is, a
document that the researcher uses as a guide during the interview. Its main
feature is the list of questions the researcher will ask the participants, but it also
includes information that the researcher will give to the participants about, for
example, the research project itself, the interview process, and aspects of ethics
(e.g. voluntary participation, anonymisation, use of data, what to do in case of
discomfort). Castillo-Montoya (2016) developed an interview protocol frame-
work aimed at helping researchers to obtain robust and detailed interview data.
Comprising four phases, the framework structures the process of preparing the
protocol from aligning the interview questions with the project’s research ques-
tions, to clearly wording questions and fine-tuning them, and finally to piloting
the protocol with a small sample of participants. The protocol should also include
strategies for the researcher to use when interviewees go too far off topic in their
responses. When this happens, researchers should always let interviewees finish
and politely ask them to return to the topic of interest. They should acknowledge
the value of the interviewees’ contribution and follow up with an explanation as

[Link] Published online by Cambridge University Press

 104 petros karatsa reas

to why they intervene in this way. Lack of time can work well as a justification if
this situation arises.
When organising the interview, researchers need to consider a number of
practical details, not least when and where the interview will take place.
Researchers need to be flexible when negotiating both time and place with their
interviewees, and they should show that they are willing to work around their
everyday commitments. When deciding on the timing of the interview, research-
ers should offer participants several options, and they should be prepared to
arrange interviews outside of normal working hours, including during the week-
end, as these are likely to be the times when participants will be available. The
same applies to the location of the interview: Again, researchers should offer
participants a range of options (e.g. a meeting room in the researchers’ insti-
tution, a room in a local community centre, the participant’s home) and allow
them to choose what is most convenient and appropriate for them. As a general
principle, settings that could have a negative effect on participants’ wellbeing as
well as on their responses during the interview should be avoided. For example,
some pupils might feel comfortable talking about how their schools ban or even
sanction the use of non-standardised language in class as part of an interview that
takes place on their actual school premises (provided the researcher creates an
environment that feels safe enough to them). Other pupils, however, might not
feel the same way and prefer to be interviewed about this issue outside school.
The choice of location will also need to be informed by issues around data
capture. Some spaces might be too noisy, which could render audio recordings
very hard to transcribe or even completely inaudible. For these reasons, it is
important to test and become familiar with recording equipment before the
interview, ideally in the space where the interview will take place. This prepar-
ation will also help with minimising the obtrusiveness of the equipment
for participants.

7.4 Data Analysis and Interpretation

The analysis of interview data starts with the transcription of the
interview, which in most cases is audio-recorded. This is done especially when
researchers are not interested in how and why participants use visual or other
non-verbal features in talking about their views and experiences of language.
When these are part of the researchers’ investigation and, of course, in projects
that include signing communities, interviews are video-recorded (see also
Chapter 18).
It is generally advisable that researchers transcribe interviews in their entirety
as opposed to only transcribing parts which they consider interesting or relevant.
It is only by having a full transcript that researchers will have a full overview of
what was said, by whom, at which point in the interview, and what the impact
was on the direction of the discussion. Transcription conventions and the level of

[Link] Published online by Cambridge University Press

 Semi-Structured Interviews 105

detail in the recording of different aspects of the speech event will depend on the
nature of the research project, not least the aims and objectives the interviews are
designed to address and the theoretical framework that is adopted for analysis
(Lapadat 2000). For example, if pauses are not part of the analysis, they can be
indicated using an ellipsis < . . . > or in writing as in [pause] to give a broadly
accurate representation of the speech event. By contrast, if data are to be analysed
within a discourse analytic framework (see Chapters 2 and 4), pauses might need
to be timed and their duration specified, for example by writing (2.5) to indicate a
pause of two and a half seconds.
In language attitudes projects, which, as noted above, are interested in the
content of participants’ contributions rather than in how these contributions are
verbalised, researchers tend to transcribe recorded speech using the standard
orthography of the language(s) in the recording – including some broad indica-
tion of things like pauses, hesitations, interruptions, repetitions, false starts, and
self-corrections (see Hutchby and Wooffitt 1998 for a detailed discussion of
transcription techniques; also Nikander 2008). Where a standard orthography
does not exist for a particular variety, or where participants’ contributions
include non-standardised or other features that are not routinely written, a
number of options are available for researchers. They can make their own, ad
hoc adaptations to standard orthography or adopt spelling conventions that may
be in use by the speaker communities themselves (Sebba 2007; see also the
contributions in Jaffe 2000). Using the International Phonetic Alphabet is also a
possibility, especially if other solutions fail to capture pronunciation in an
appropriate way.
If more than one language is used in the recording, the current trend is for
every language to be transcribed with its own appropriate orthographic conven-
tions, and without using any kind of notation that would suggest that one
language is more ‘normal’ or more important than the other (Blackledge and
Creese 2010; see also Bashiruddin 2013; Holmes et al. 2016). To apply this
principle to the title of a seminal article in the study of code-switching, Poplack
(1980), one would transcribe the sentence as in (1a) rather than (1b) or (1c).

(1) a. Sometimes I’ll start a sentence in Spanish y termino en español.

b. Sometimes I’ll start a sentence in Spanish y termino en español.
c. Sometimes I’ll start a sentence in Spanish y termino en español.

In (1b) and (1c), the roman (non-italicised) typeface suggests that the use of
English in this segment of the sentence is normal, whereas italicisation and
capitalisation single out the Spanish segment as somehow unusual or out of the
ordinary. The same principle of transcribing in a manner which does not suggest
that one language variety is more normal or appropriate than the other also
applies to languages that are written using scripts other than the Roman alphabet.
It is best if these are transcribed using their own writing systems and not Roman
transliterations in order to retain the authenticity of the speech event, represent
the naturalness of the multilingual product on behalf of the participants, and

[Link] Published online by Cambridge University Press

 106 petros karatsa reas

Table 7.1 Phases of thematic analysis (Braun and Clarke 2006: 87)

Phase Description of the process

1. Familiarising oneself Transcribing data (if necessary), reading and re-
with the data reading the data, noting down initial ideas.
2. Generating initial Coding interesting features of the data in a systematic
codes fashion across the entire data set, collating data
relevant to each code.
3. Searching for themes Collating codes into potential themes, gathering all
data relevant to each potential theme.
4. Reviewing themes Checking if the themes work in relation to the coded
extracts and the entire data set, generating a
thematic map of the analysis.
5. Defining and naming Ongoing analysis to refine the specifics of each theme,
themes and the overall story the analysis tells, generating
clear definitions and names for each theme.
6. Producing the report The final opportunity for analysis. Selection of vivid,
compelling extract examples, final analysis of
selected extracts, relating back of the analysis to the
research question and literature, producing a
scholarly report of the analysis.

respect their cultural traditions. Any transliteration – and, of course, English

translation – of such segments can and must be done when said segments are
analysed. For examples of this practice, see Section 7.6.
Interview data are most commonly analysed qualitatively using thematic
analysis (see also Chapter 10). The aim of thematic analysis is to identify,
analyse, and report themes – that is, patterns that emerge from the interview
data. Ryan and Bernard (2000: 780) define themes as ‘abstract (and often fuzzy)
constructs the investigators identify before, during, and after analysis’. Braun and
Clarke (2006) conceptualise themes as capturing patterned responses that prevail
within a data set (in this case, the corpus of interview transcripts) both in terms of
the space that a given theme occupies within each data item (i.e. how much a
given topic is talked about in each interview) and in terms of the occurrence of a
theme across the different data items (i.e. in how many interviews within the data
set is a particular topic discussed). Braun and Clarke (2006) also break down the
process of doing thematic analysis in six helpful steps, or phases, as shown in
Table 7.1.
Braun and Clarke (2006) further provide a very useful discussion of the
various decisions that researchers doing thematic analysis need to take with
respect to how they are going to approach their data. The most important one
concerns epistemology, namely whether the data are approached from a realist/
essentialist/positivist perspective or from a constructionist one (Braun and Clarke
2006: 85; Codó 2008: 162; Heller et al. 2018: 86–87; see also Chapter 21). The

[Link] Published online by Cambridge University Press

 Semi-Structured Interviews 107

former approach assumes that ‘it is possible to achieve a veridical or authentic

account of a person’s opinion (which treats such “things” as stable and fixed)’
(Edley and Litosseliti 2010: 173). In language attitudes research, this would
assume that participants’ views and beliefs about language are inherent in them
and, once they have been formed, remain largely immutable in the course of their
lives unless external circumstances cause them to change. From such a point of
view, the aim of the interview would be to uncover these underlying truths. Most
contemporary scholars, however, reject this line of reasoning and adopt a more
constructionist approach, which accepts that people’s views and beliefs are
constantly constructed, negotiated, and (re)constituted as they come to form part
of their identities, and perceptions and projections of their selves (see also
Chapter 1 for a more detailed discussion of the dynamic nature of language
attitudes). The aim of interviews is therefore to gain insights not into what
participants ‘actually’ feel, believe, or do in terms of language attitudes but,
rather, into what they say they feel, believe, or do (Edley and Litosseliti 2010:
173). For the constructionist approach, then, social desirability bias does not
constitute a problem as researchers are acutely aware, and in fact expect, that the
multidimensional context of the interview will shape participants’ responses.
They will therefore analyse interview data with this in mind. This explains
why any findings arrived at by means of qualitative analysis and interpretation
of interview data are not generalisable to people and groups outside the sample of
participants who were interviewed.

7.5 Further Important Considerations

Perhaps the most important consideration in conducting interviews to

investigate attitudes is the choice not only of the language but also of the variety
of that language that the researcher will use during the interview. This is because
of the well-known fact that people often link other people’s language with non-
linguistic characteristics (see Chapter 1). The language a person speaks is
routinely used as an indicator of membership to specific ethnic, national, racial,
religious, regional, age, gender, or class groups. People who speak a specific
variety or varieties of a language are similarly expected (or assumed) to have
certain non-linguistic attributes: a high or low level of education, good or bad
manners, high or low intelligence, a high or low socio-economic status.
In an interview setting, the choice of language (and variety) on behalf of both
the interviewer and the interviewee can signal how they see themselves and the
other, how each of them sees the speech event, their role in it, and their position
with respect to the other part, or which aspects of their identity they want or feel
they have to project. Researchers must be aware of these complexities and should
strive to reduce the distance and any tensions that may be created by linguistic
differences between them and their participants. For example, using the

[Link] Published online by Cambridge University Press

 108 petros karatsa reas

interviewees’ language (or variety) might be the only option available to

researchers if interviewees are monolingual (or monodialectal). When interview-
ees and researchers share a language (or variety), the same choice might be seen
by some interviewees as an acknowledgement of the value of their language (or
variety). Others, however, might consider it to be inauthentic, appropriating, or
an attempt at mimicry or even mockery. Researchers should therefore decide
which language (or variety) to use for the interviews based on the research
context at hand, taking into account their own and their participants’ linguistic
repertoires, and in consultation with participants at the very early stages of their
work together.
It is not difficult to see how these dynamics can play out in interviews with
participants who speak a language or variety that lacks overt prestige. Such
speakers often encounter stigmatisation in the course of their lives, directed at
them not only by monolingual/monodialectal speakers of the prestigious lan-
guage (or variety) but also by bilingual/bidialectal speakers of the non-
prestigious language (or variety) who are also competent in the prestigious
one. It is therefore not uncommon to develop feelings of inferiority about
speaking their language (or variety), and insecurity when they have to use the
prestigious or dominant language (or variety) of the society and community in
which they live. In a project investigating the use of stigmatised forms of
language, it goes without saying that participants should be given the choice to
use the language or variety they prefer in their responses.
This is, however, more easily said than done. It will undoubtedly be much
easier to achieve if the researcher also happens to be a speaker of that language
(or variety) and is therefore able to create a safe space for the interview to be
conducted in that language (or variety). However, things will be complicated if
the researcher is classified by the participant as a member of a sociolinguistically
dominant group that speaks a more (overtly) prestigious language (or variety).
Even if the researcher assures them that they should use any form of language
they want during the interview without fear of judgement, stigmatisation, or
discrimination, it is likely that participants will feel compelled to speak the
prestigious language (or variety) in response to the formality of the speech event,
to accommodate to the interviewer’s characteristics, or to disassociate themselves
from non-linguistic attributes that speakers of the non-prestigious language (or
variety) are stereotyped to have. This has ethical, methodological, and analytical
implications as this choice may limit not only the extent but also the content of
participants’ contributions. They might feel reluctant to talk about their personal
experiences, feelings, and views with a speaker of the prestigious language (or
variety) – or they might not want to talk about them in the same way, depth, and
openness as they would if they were interviewed by a researcher who was ‘like
them’. Researchers need to be mindful of these issues and take them into account
at all stages of their investigation, starting from the planning of the interviews,
the preparation of the interview protocol, and later on during the analysis and
interpretation of the interview data.

[Link] Published online by Cambridge University Press

 Semi-Structured Interviews 109

7.6 Case Study: Attitudes towards Cypriot Greek in the UK

The term Cypriot Greek refers to the non-standardised Modern Greek

varieties that are spoken on the island of Cyprus. Cypriot Greek speakers who
migrated from Cyprus to the UK brought with them the notion that Cypriot
Greek is a rural and archaic form of language. which stigmatises speakers as
impolite and uneducated. By contrast, Standard Greek, which is associated with
Greece and especially Athens, is perceived as a modern, urban variety that makes
speakers sound well-mannered and educated. People with a Greek Cypriot
background in the UK additionally view Standard Greek as a language that
unites them not only with the Cyprus homeland but also with Greece as the
notional centre of Hellenism and with other Greek(-speaking) diasporas across
the world. Cypriot Greek, on the other hand, is thought to be too regional and
inaccessible to Greek speakers from a non-Cypriot background. In Cyprus,
British-born Greek Cypriots are sometimes stereotyped as unsophisticated and
backward-thinking, a group of people who are ‘frozen in time’ and preserve a
form of language (as well as ways of living and thinking) that people in Cyprus
have long abandoned. Some people in the diaspora therefore see speaking a
‘strong’ version of Cypriot Greek as confirming these stereotypes, and as reinfor-
cing associations with the low socioeconomic status and level of education of
many early migrants (Karatsareas 2018, 2019, 2021).
At the early stages of my investigation of these themes in London’s Greek
Cypriot community, I decided to conduct semi-structured interviews in order to
enrich the understanding of the research context that I had built through
reviewing the literature, to familiarise myself with the community and the
experiences of individual members, and to identify potential avenues for further
investigation (as discussed in Section 7.2). In preparing for the interviews, I put
together an interview protocol as described in Section 7.3. Having as my starting
point the research questions I aimed to address, I formulated a list of questions to
ask my participants. In line with Castillo-Montoya’s (2016) framework, the first
step in that process was making sure that the interview questions aligned with the
research questions, as shown in Table 7.2. In each case, I assessed whether a
given interview question was necessary for my study, and I considered whether it
would help my participants to explain their views on the status of Cypriot Greek
compared to Standard Greek, and to talk about their own lived experiences
of stigmatisation.
I recruited my participants through a Greek complementary school. A key
factor in this process was the help I received from the school’s headteacher, who
acted as a gatekeeper (see Section 7.3): They introduced me to parents who
brought their children to the school, voiced their support for my research, and
provided me with rooms on the school grounds in which to conduct the inter-
views. The headteacher played a very symbolic role in shaping my participants’
perspectives and in constructing the overall context in which the interviews came
to be situated (see Section 7.2). When they introduced me to the parents and after

[Link] Published online by Cambridge University Press

 110 petros karatsa reas

Table 7.2 Aligning research questions and interview questions about Cypriot
Greek in the UK

Research questions Interview questions

How are ideologies and attitudes Who do you speak Cypriot Greek with?
towards non-standardised varieties When and where do you usually speak
manifested, reproduced, and it? What about Standard Greek?
contested in diasporic contexts? What does each variety mean to you?
How does speaking Cypriot Greek
make you feel? What about
Standard Greek?
Can you explain these thoughts with
reference to a personal experience of
yours? Do you have any interesting life
stories to share?
What role does community language What is your experience of going to
education play in shaping, sustaining, Greek school? Did you enjoy learning
and reinforcing attitudes and Greek there? What were some things
ideologies? you did not like?
How did your teachers use Standard
Greek and Cypriot Greek? How did
they explain the differences
between them?
Which language do you think children
should learn at Greek school? Do you
think children should learn Cypriot
Greek? Why? Why not?

stating to them that I was interested in ‘how we speak here’, they went on to
mention two universities: the University of Cambridge, from which I received
my doctorate, and the Open University of Cyprus, where I was working at the
time as adjunct faculty. The former mention was meant to add authority to my
profile as a legitimate and serious researcher, the latter to highlight the fact that,
despite being originally from Greece (as was evident from my last name and the
way I spoke Greek), I did have connections with Cyprus. Although motivated by
their best intentions to convince parents to take part in my study, the head-
teacher’s introduction created the impression in some participants that they had
to showcase their best Greek to me (which illustrates the problematisation of this
issue in Section 7.5). This impression was reinforced by the fact that the school
was the location where the interviews took place (see Section 7.3). Some of my
participants had attended that school as children and knew it as an institutional
space where only Standard Greek was allowed as the language of instruction,
while pupils’ use of Cypriot Greek in the classroom was discouraged and
stigmatised (Ioannidou et al. 2020; Georgiou and Karatsareas forthcoming).

[Link] Published online by Cambridge University Press

 Semi-Structured Interviews 111

During the interviews, some participants engaged in efforts to avoid using

Cypriot Greek features and tried to speak in Standard Greek. They also denied
that they spoke or even knew ‘heavy’ or ‘villagey’ Cypriot, two labels that refer
to the most basilectal registers of Cypriot Greek which are particularly associated
with rurality. In Example 1, in which Cypriot Greek features are indicated in
boldface, Stella denies several times that her family speaks a ‘heavy’ ‘villagey’
Cypriot (staves 2, 4, 6) but then goes on to produce a high number of Cypriot
Greek features and provide examples of ‘heavy’ words (stave 4). Half-way
through stave 6, she realises that she carelessly admitted that her mother does
use ‘heavy’ and ‘villagey’ words so she quickly shifts away from describing her
home environment to talking about her Greek education, listing the Greek
qualifications she and her sister obtained, which serve as evidence of the
properness of her Greek.
(1) ‘I don’t even know if I can talk like that’.

1 Petros Ίνταλος μιλά ο παπάς σου; How does your father speak?
2 Stella We don’t. . . we don’t know. We don’t. . . we don’t know.
Ο παπάς μου έν μιλά βαρετά My father doesn’t speak
που εννα πάει έξω. Εν μόνο heavy talk when he goes out.
που εννα δει χωρκανούς ή It’s only when he meets
συγγενείς. Έτσι έρκεται πίσω people from his village or
τζιαι you know. . . Τζιαι μιλά relatives. So he comes back
έτσι βαρετά αλλά εν τζιαι and you know. . . He speaks
τζείνος που εννα πάμεν έξω ή heavy talk like that but it’s
που εννα. . . Έν θα. . . Έν him as well when we go out
μιλούμεν έτσι βαρετά βαρετά or when. . . We won’t… We
τα χωριάτικα. don’t speak like that really
heavy talk the village talk.
3 Petros Ίνταλος ένι τα βαρετά; What is heavy talk like?
4 Stella Πώς είναι τα βαρετά; To be What is heavy talk like? To be
honest I don’t even know if honest I don’t even know if
I can talk like that. Ξέρω I can talk like that. I know
λέξεις σαν. . . Like πούγκα ή words like. . . Like pocket or
μαβλούκα. Ξέρεις την pillow. Do you know pillow?
μαβλούκα;
5 Petros Όι. No.
6 Stella Μαξιλάρι μαξιλάρι. Δεν ξέρω Pillow pillow. I don’t know
έτσι λέξεις σαν τζείνες words like those but. . .
αλλά. . . Διότι ήμουν σπίτι με Because when I was at home
την μάμμαν μου που with my mother when I was
εμεγάλωνα μεγάλωσα ώσπου growing up I grew up until
να πάω σχολείο τζιαι τζείνη I went to school and she
μιλ. . . Έν μιλά βαρετά speak. . . She doesn’t speak
χωριάτικη. Έχει heavy talk village talk. She
χρησιμοποιήσει χρησιμοποιά has used she uses heav. . .

[Link] Published online by Cambridge University Press

 112 petros karatsa reas

βαρύτικ. . . βαρετές heavy village words. When

χωριάτικες λέξεις. Που we went to school Greek
επήαμεν σχολείον ελληνικόν school we were young and. . .
σχολείον ήμαστεν μικρές Yes. . . We did our GCSEs
και. . . Ε νναι. . . Τζι and our A-levels so we don’t
εκάμαμεν τζιαι τα GCSE μας speak heavy talk no.
τζιαι τα A-level μας κι έτσι έν
μιλούμεν βαρετά no.

Perhaps the most noteworthy instance in Stella’s extract is how she

standardised my language. In Extract 1, I ask Stella two questions (staves
1 and 3) and answer one question that she asks me (staves 5). In all three
instances, I use Cypriot Greek, which I had previously stated that I can speak.
I chose to do this in order to create favourable conditions for Stella to share her
views and experiences in the Greek variety that she is dominant in. In stave 4, she
uses this choice as an affordance to display her knowledge of Standard Greek but
also to align my language with the image she has formed for me (a lecturer with a
degree from a prestigious British university who comes from Greece and speaks
the ‘proper’ type of Greek, someone who does not belong to the Greek Cypriot
community). My question in stave 3 starts with two Cypriot Greek forms:
ίνταλος [ˈiⁿdalos] ‘how’ and ένι [ˈeni] ‘is/are’. At the beginning of her response,
Stella repeats this question. In doing so, she replaces ίνταλος and ένι with their
Standard Greek corresponding expressions, πώς [pos] ‘how’ and είναι [ˈine]. The
‘corrected’ version Πώς είναι τα βαρετά [pos ˈine ta vareˈta] ‘What is heavy talk
like?’ is the only utterance in her responses that does not contain any Cypriot
Greek features, suggesting that one of her aims was to challenge the authenticity
of my production and question the limits of my knowledge of Cypriot Greek.
Stella was not the only participant to use my speaking Cypriot Greek as an
opportunity to construct their image as users of different varieties and registers of
Greek. In Example 2, Anna reacts to my use of two forms, τζι(αι) [tʃ(e)] ‘and’ and
εγιώ [eˈʝo] ‘I’ (stave 1), which are highly marked as belonging to the Cypriot Greek
lexical stock on their basis of phonetic features (the [tʃ] in τζι(αι) and the [ʝ] in εγιώ).
(2) ‘Well, I don’t say τζι εγιώ, let’s put it that way’.

1 Petros Σαν εσένα που θέλεις χρόνο Like how you need some time
να μου μιλήσεις Ελληνικά; to speak in Greek to me? In the
Έτσι και. . . τζι εγιώ θέλω. same way I need some time
too.
2 Anna Well, I don’t say τζι εγιώ, Well I don’t say me too let’s
let’s put it that way put it that way.
3 Petros Γιατί; Why?
4 Anna ‘cause that’s not correct ‘cause that’s not correct
5 Petros Ναι αλλά δεν είναι λάθος Yes but it’s not a mistake
6 Anna Τζι εγιώ; No, I’m a believer Me too? No, I’m a believer in
in speaking correctly if speaking correctly if possible.
possible.

[Link] Published online by Cambridge University Press

 Semi-Structured Interviews 113

What is interesting in this case is that my production seems to have been

interpreted by Anna as part of an entrapment strategy. When we had spoken to
set up the interview, she had confirmed she wanted to do the interview in Greek.
During the interview, however, she found it challenging to use the language. Her
responses were telegraphic, and she ended her contributions rather abruptly,
hastening to move on to the next question. She also switched to English
frequently. Assessing the situation, I decided it would be best to first encourage
her to use Greek and reinforce her rather than suggest we do the interview in
English. I did not want to give the impression I was impatient and that I did not
give her enough opportunity to feel relaxed before she could start speaking in
Greek. I also thought I would tell her that I also found it challenging to speak
English sometimes as it is not my first language. At the end of the utterance in
stave 1, I interrupted myself to say τζι εγιώ instead of κι εγώ to signal that
Cypriot Greek could be used in the interview space we were constructing. Anna,
however, seemed to think that I had used Cypriot Greek in order to make her
speak in that way in the interview or to make her admit that she speaks like that.
Her distancing from both these propositions could not have been given in
stronger terms (staves 2, 4, 6). My use of Cypriot Greek on that occasion and
in that context was judged to be insincere, a risk identified in Section 7.5. It is not
easy to give hard and fast advice about how to manage uncomfortable or uneasy
situations that may emerge in interviews (although see Heggen and Guillemin
2012, and Rapley 2012). In any case, researchers should always seek to minimise
participants’ discomfort and try to diffuse the tension created during such critical
moments by reducing the intensity of their questions and moving smoothly to
other, more comfortable topics as soon as the context allows.

Suggested further readings

Briggs (1986); Codó (2008); Dewaele et al. (2019); Edley and Litosseliti (2010);
Gubrium et al. (2012)

[Link] Published online by Cambridge University Press