Approaches to the study of address in pluricentric languages: methodological reflections

4.1.1 Questionnaires

Questionnaires are a convenient and well-established data collection instrument, even more so with online administration now easily available.

For the AICE project, all data were collected by questionnaire, in paper form between 2006 and 2013, and online using SurveyMonkey from 2013. The two questionnaires were thus identical in content but differed in the way participants were recruited and where the surveys were completed. The paper-based questionnaires were distributed at academic conferences in our own fields of linguistics and modern language studies, made available at information desks or potential respondents were approached in person. The respondents thus completed the questionnaires in the conference context. We also trialled distributing questionnaires at international conferences of different disciplines but had very disappointing return rates - possibly reflecting a lack of interest in the research questions. As a result, we decided to restrict the survey to conferences in our field. The online survey was publicised by sending invitations to email addresses which we collected from university websites, again focusing on linguistics and language studies. These questionnaires were thus completed while respondents were not at conferences.

The questionnaire was in English and was kept deliberately short to encourage a high completion and return rate. The core questions were restricted to three scenarios in English and three scenarios in the respondents’ L1. The context was set by asking “When you are meeting conference colleagues for the first time, …”. We cannot know the exact situation the respondents were thinking of when they answered the question, for example, it could have been a more formal introduction during conference proceedings or less formally over drinks. This issue was pointed out to us by some respondents.

Formally the questionnaire used a multiple-choice format, listing the most common options (by first name, by first name and last name, by academic title and last name) plus an open option “other” with a request to specify. Multiple answers were possible. To facilitate data analysis, responses were then reduced to the four categories listed above. Multiple answers were assigned to one of the categories; when, for example, first name was mentioned as one possible option, the response was assigned to the “first name” category. In addition, an open question calling for comments and general thoughts on the topic and a request to relay any experiences or anecdotes on address in first encounters were added. These two open questions generated much comment and allowed us to add a qualitative perspective which helped “elucidate the thinking and understanding of people that may influence the choices they make when addressing and referring to one another” (Kretzenbacher et al. 2020: 118).

The choice of English for the questionnaire, and the decision to keep it very short, led to another issue with L1 speakers of other languages. When we asked about address/introduction using “title”, most respondents deduced correctly from the context of international academic conferences that we were referring to academic titles such as Dr. and Professor, not honorifics such as Mr/Ms/Mrs (often referred to as “courtesy titles” in address research). However, L1 speakers of German in particular sometimes mentioned the German courtesy titles Herr/Frau in this context.

The questionnaire also collected personal information about the respondents. Their first language(s) and country of origin were used to gauge their background in terms of pluricentricity. For example, a person with German as their first language and Austria as their country of origin was classified as an L1 speaker of Austrian German. Some respondents did in fact indicate a specific language variety (e.g. “English (Australian)”, “German (Tyrolean Dialect)” or “Swiss German”). In terms of age, we suggested three age groups - under 30, 30-50 and over 50 - which often coincide with seniority status (post-graduate, early-career, mid-career academics). In hindsight it would have been preferable to add a question about their academic status as this turned out to be an important factor, judging by comments from participants.

In Formentelli/Hajek’s (2016) study, which focused on Australian English, a double-sided paper questionnaire was devised to collect data in Italian language classrooms in Australia on address practices experienced by university students - in English and in Italian respectively. It was a classroom-based activity that was then used with students to explore similarities and differences between the two languages that students might have noted or assumed. It also included questions about address at upper secondary level - to provide some comparative data on any difference in address between levels of education. Otherwise, questions focused on interactions between students within different types of lessons and with different teaching staff, e.g. professor versus tutor, in order to understand the impact of context and hierarchy.

A questionnaire was also a central data collection instrument in the AWE project. It asked 38 questions on reported address use in a number of situations and domains (e.g. with family, friends, strangers, at work). In the follow-up interviews (see below) some of these scenarios were further developed, the reported address practices were probed further and attitudes to existing address practice explored (Clyne/Norrby 2011: 148). The focus of the predominantly closed questions in multiple-choice format was on pronominal address with additional information sought on nominal address for some questions.

By combining the questionnaire with a semi-structured face-to-face interview (see section 4.1.2) we were able to counterbalance some of the disadvantages of written questionnaires that have closed multiple-choice questions. In particular, the interview allowed the participants to expand on their questionnaire answers, thereby adding depth and context, and enabled the interviewers (local research assistants) to ask for clarification. On the other hand, the written format and the often binary nature of the answer options in the questionnaires made them much easier and quicker to process than interview data.

As part of the AWE project we also surveyed a range of companies and institutions by email. The open questions inquired about company address policy (e.g. “Does your company have an official address policy?”) and about actual address practices within the organisation and externally, primarily in communication with customers (e.g. “Which address pronouns and formula are used in letters/emails?”). Unfortunately, the return rate of the survey was disappointing. While the number of replies was insufficient for systematic analysis, we were still able to use the limited number of replies as anecdotal evidence in our analysis of address in the workplace and in Norrby/Hajek’s (2011) study. In hindsight we see three main reasons for this failure: 1) with 14 quite detailed and complex questions the questionnaire was too long and too onerous to complete; 2) it seems particularly difficult to get questionnaire responses when canvassed through email or letter, particularly within companies/institutions, where it can be difficult to identify the appropriate addressee; 3) it seems to be generally difficult to obtain information from companies and institutions about internal policies; we have experienced this in other, unrelated projects.

4.1.2 Interviews

As mentioned above, the audio-recorded semi-structured interview of the AWE project directly followed the questionnaire. The main part of the interview consisted of open questions about address usage in additional situations, attitudes and opinions about address, perceived changes in address practices and specific experiences (e.g. “Have you ever been addressed with a form you didn’t expect?”). Most importantly in the context of pluricentric languages, one question specifically addressed their perception of regional and/or national differences in address behaviour (“Do you think address terms are used in a different way in other xxx-speaking countries?”). Many questions tried to elicit narratives of personal experiences, similar to the anecdotes in the AICE project. The length and the quality of the interviews varied quite considerably, depending on the skills of the interviewer, the rapport established between interviewer and interviewee and the degree of engagement of the interviewee with the topic. Overall, the more free-flowing interviews yielded more narrative data which in turn were more labour-intensive in terms of transcription and analysis. The combination of quantitative and qualitative aspects, however, added much depth to our findings.

4.1.3 Focus groups

Focus groups can be characterised as indirect data collection instruments in the first place as they are convened to talk about and elicit opinions and attitudes on the topic at hand. However, how the participants address one other in interaction may also provide interesting data.

As part of the AWE project, focus groups were convened in each research location at the start of the project during which the participants provided insights which helped shape the further course of the project (e.g. questions to explore in the questionnaire and interviews). The groups were re-convened towards the end of the project when the participants were presented with summarised data and assisted with their interpretation. Groups consisted of 12-16 people from a reasonably broad range of backgrounds. The focus group interactions were audio-recorded and transcribed or summarised.

The focus group meetings - facilitated by a local research assistant - generated lively discussions, provided us with a good insight into address practices and gave us an overview of the range of perceptions, attitudes and expectations of address use. They did not in fact yield much data in terms of actual address use. The interaction was topic-focused and the participants did not directly address each other often enough to draw meaningful conclusions, except maybe regarding avoidance of direct address.

Composition of focus groups can be an issue. To allow easy interaction the group cannot be too large but should be somehow representative of the group under investigation. Factors such as age, gender, occupation, educational level, and socio-economic background of the participants all need to be “balanced”. In our case, for example, the groups overall were biased “towards the professional middle class and students” (Clyne/Norrby/Warren 2009: 33).

4.1.4 Online data

Like focus groups, the online data (chat groups, online forums, blogs) we have used are situated at the intersection of indirect and direct data. They allow for observation of actual address behaviour and interaction among participants (as e.g. in Kretzenbacher/Schüpbach 2015 and Kretzenbacher 2011 among others) but also for intervention, when researchers contribute to the exchanges and/or initiate a discussion. In 2005/06, for example, we visited publicly accessible chat groups and discussion forums for the AWE project for two interrelated purposes: to observe address practices in this online medium and to initiate a discussion on address practices which would provide us with further insights into pronominal address in computer-mediated communication (reported in Clyne/Norrby/Warren 2009: 116-123).

Common to chat groups and online forums is that the participants generally identify themselves by name only - and very often by first name or by nickname that can even vary across forums and platforms. Thus, it is often very difficult, if not impossible, to determine their regional and social background, age and/or gender (Kretzenbacher 2011: 72). Nevertheless, some of these indicators can be deduced from the content or context of the contributions. In our case study of German, Austrian and Swiss newspaper forums (Kretzenbacher/Schüpbach 2015), for example, we assumed that the contributors were actually located in the countries where the newspapers are published. The assumption is supported by the fact that many contributors are regulars and if posting from abroad often emphasise this fact in their comments.

Other online data we used successfully include blogs dealing among other issues with address across German varieties (Schüpbach 2015) and institutional websites (Norrby/Hajek 2011) to investigate how companies address their potential customers across languages and varieties in this medium. Overall, we have found online data a good source of real-life address use in the public domain. We have so far mainly concentrated on “old-style” online media, and have not (yet) tackled currently prevalent channels - Facebook, Instagram, Twitter, and so on (but see e.g. Foster/Aalberse/Stoop 2019 for a study on address on Twitter in Colombian Spanish).

4.1.5 Actual interactions

As part of the AWE project, (participant) observation was conducted extensively by the researchers when visiting the research locations and recorded in field notes. This approach is useful as it can cover a wide variety of contexts and everyday situations and while it may provide only anecdotal data, they can nevertheless support (or challenge) findings achieved through other means. Also under this umbrella come informal conversations with members of the local community and colleagues about the project and the issue at hand.

We trialled another observation project as part of the AICE project to counter some of the drawbacks of the short questionnaire mentioned earlier and to collect data of actual rather than reported use. We designed an observation schedule to capture introduction and address practices in a variety of sub-contexts at international academic conferences (e.g. plenaries, symposia, social gatherings). As a pilot run showed that this procedure was quite time consuming and would require multiple researchers collecting data simultaneously in order to obtain a notable and reliable sample, we have paused the project.

IVIP is the only project any of the authors have been involved in that considered actual (video-)recorded interactions (see Norrby, this volume, for more detail). While providing natural and very rich data, there are a number of obstacles to this type of data collection, not least of which are the increasing barriers in terms of research ethics and access, as well as the time-consuming and labour-intensive nature of data collection, transcription and analysis.

4.2.1 Methodological framework

The methodological frameworks in the field of address research in pluricentric languages are quite varied, and in our own research we have typically combined several approaches which we outline briefly below. The framework known as variational pragmatics (Schneider/Barron 2008, Barron 2017), was introduced in a bid to combine insights from pragmatic research, social dialectology and variational linguistics. Typically, studies in this field investigate how speech acts and pragmatic formulas (including address routines) are realised in different pluricentric varieties, including both national and regional varieties. Methods of data collection are varied, but early variational pragmatic studies were primarily quantitative in nature, including both experimental designs to elicit certain speech actions (e.g. discourse completion tasks where fictional situations are presented to respondents who are then asked to take over the part of one interlocutor in such situations and enter the turns they would provide to the discourse in the given situation) and authentic discourse, often sourced from large existing corpora.

In recent years, there has been an increased interest within variational pragmatics in combining a large-scale quantitative approach with a qualitative analysis which also includes the interactional, topical and organisational levels of analysis to more fully describe the wealth of pragmatic choice available to interlocutors (see e.g. Schneider 2019). In this vein, some researchers have adopted a processual perspective, focusing specifically on the sequential unfolding of actual interactions and the social actions performed, for example, by a certain linguistic form or gambit (e.g. addressing somebody in a certain way at a specific moment of the conversation). Such an approach is typical of conversation analysis and the related field of interactional linguistics (for an example of this approach see Grahn, this volume).

In variational pragmatics the interest in national and regional variation is complemented by other social factors, such as age, gender and occupation to explain variation, an approach also typical of variational sociolinguistics. In both the AWE and AICE projects the questionnaire datasets were analysed by accounting for such external social variables in addition to national (and regional) variation in address practice.

Much address research has been preoccupied with explaining the variation - either based on actual data and/or respondent perceptions and attitudes to address - and proposing various theoretical models to explain variation in addressing practices. The seminal work by Brown and Gilman (1960) predicted a linear development over time where hierarchical, non-reciprocal address would be replaced by address forms signalling reciprocity and solidarity, and where formal V would yield to informal T. In contrast, the AWE project proposed a model, based on a wide variety of data, introducing three scales to describe the address choices available in the languages under investigation (scale of grammatical resources, scale of V-ness, scale of sameness) as well as a set of principles that people follow when determining how to address a certain interlocutor (e.g. familiarity, relative age, network membership; see Clyne/Norrby/Warren 2009: 156-159). Rather than a diachronic linear development from V to T, the AWE model of address acknowledges the impact of globalisation and migration leading to language contact, where address forms may “travel” across territories and give rise to novel address practices that are cyclical rather than linear in nature. Overall, the AWE project emphasised common ground and perceived sameness as important drivers of address choice. As a result, for individuals a sense of a shared identity, whether perceived through typical social characteristics such as age or occupation, having shared interests, or indices of personal preference such as musical taste or hairstyle, may override conventional address “rules” within the community and facilitate shift to and use of T address.

4.2.2 Variety or varieties investigated

Research may involve the major national varieties of a pluricentric language (e.g. our research on Swedish with its two national varieties), compare some national varieties (e.g. the Australian, British and American varieties of English in Norrby et al. 2019) or focus on one particular national variety (e.g. Schüpbach 2015 on the Swiss variety of German or Formentelli/Hajek 2016 on Australian English). While in this latter case the focus is generally on a non-dominant variety, this is explicitly or implicitly compared with a dominant variety which serves as a/the reference variety.

4.2.3 Pluricentricity or pluriareality?

Research into pluricentric languages traditionally takes national variation as its point of departure and/or subject of inquiry. This view has lately come under some scrutiny with the introduction of the notion of “pluriareality”, which is concerned with regional variation both within and across national borders, i.e. “variation in language use in the entire geospatial space where the language is spoken” (Norrby et al. 2020: 207; see also Auer, this volume). However, the perception of a conflict between a “pluricentric” and a “pluriareal” approach appears to be mostly limited to studies published about German (cf. Dollinger 2019 and Dürscheid/Simon 2019 as recent examples).

Variation at a regional level is also considered within the research framework of variational pragmatics (see section 4.2.1). For German, for example, variation between eastern and western Germany may be relevant: in the AWE project we found some variation between research locations Leipzig (east) and Mannheim (west) (see e.g. Clyne/Norrby/Warren 2009: 127-132). This variation can be framed in terms of national variation, as it is rooted in two previously recognised national standards - albeit of nations that no longer exist in this form. The “north-south divide” is another regional aspect within Germany - particularly important in historically and dialectologically oriented research and documented, for example, in the Atlas zur deutschen Alltagssprache (Elspaß/Möller 2003-), which presents results on maps and thus illustrates that variation can not only be sub-national but also straddle national borders. Variation between regional or rural areas on the one hand and urban centres on the other can add yet another dimension. In AWE network interviews from Finland, for example, there were comments on regional variation in address behaviour within the Finland-Swedish speaking area. Address behaviour was perceived as more formal in the capital, and more relaxed in the regional area where the data were collected (Ostrobothnia in this case).

Both the pluricentric and the pluriareal approaches have direct implications on the selection of research sites and participants. In the case of the AWE project, for example, we made a conscious decision to include three German-language research locations - one in Austria, two in Germany - to capture some of the national variation as well as the regional variation within Germany. Unfortunately, the inclusion of a German-speaking location in Switzerland had to be dropped as a result of funding constraints. Given the focus of the project on socio-political change and its impact on address, the east/west dimension within Germany was pertinent, but the north/south variation less so.

We also had to ensure that the participants were from the specific region we were interested in and had not, for instance, moved to the area very recently. For the AWE project we recruited base participants from the social networks of our local in-country research assistants, making sure that they were connected to the research location. In the questionnaire we therefore asked about their place of birth, the region where they grew up, any other locations where they had lived for more than six months and where their parents were born. While we did not exclude participants who originated in other areas, we wanted to ensure that the primary socialisation of the majority had happened in the specific location and thus reflected address use in that place.

4.2.4 Context(s) of address use

The AWE project investigated a wide range of contexts. The questionnaire, for example, asked about address in the family, with friends and neighbours, in the workplace, with strangers in various situations (e.g. asking for directions, service encounters, police), at school and university. The subsequent interview expanded on some of these contexts (e.g. “What has been the practice of address forms in workplaces where you have worked previously?”), while others were designed to gauge attitudes (e.g. “How do you feel about companies who prescribe the T form (for example: IKEA)?”). We mainly asked about address use in face-to-face interaction, and to a much lesser extent in written and electronic communication.

This variety of contexts provided a good overview of address practices across domains and allowed us to compare address use across languages, contexts and locations. Particularly for address in the workplace we found significant variation (Kretzenbacher/Clyne/Schüpbach 2006: 17.12-14; Warren 2006).

The other projects, AICE and IVIP, covered more specific contexts. The AICE project considered the context of international academic conferences, a valuable site for research on intercultural and intracultural pragmatics and on address and introductions in particular, as introductions are an obligatory part of the interactions, be it in the attendees’ first languages or in English, increasingly the default language of international conferences (see Kretzenbacher et al. 2015: 78-79 for more background on the conference setting). This context was further limited to conferences in the fields of linguistics and modern language studies (see section 4.1). While this has the advantage of a clearly delineated and comparable context, it leaves room for criticism that it may be too narrow - would the results be different if conferences in other academic fields (e.g. in the sciences or in other fields within the humanities and social sciences) were investigated?

The interactions investigated in the IVIP project took place in three domains: service, higher education and healthcare with the majority of interactions sourced from service encounters, supervision meetings, medical consultations and preventative healthcare (see also Norrby, this volume). These contexts are quite specific and well-defined, and thus allow the researchers to control for certain variables. In particular, they represent contexts with “pre-determined institutional roles” (Wide et al. 2019: 2).

4.2.5 Type(s) of address investigated

Whether the focus is on pronominal or nominal address and whether peripheral phenomena such as greetings (often associated with certain address forms) are also included, is largely dependent on the language(s) investigated. For English, nominal address is of most interest, whereas for more complex systems such as in Italian, pronominal address is particularly worth investigating.

The AWE questionnaire was primarily focused on pronominal address, but for a selection of questions we also asked about nominal address, e.g. whether people used first names or kinship terms within the family or first names or honorific/title and last name with their superiors at work. This also allowed us to identify some less conventional combinations (e.g. T pronoun and last name or V form and first name) which could then be further explored in the interviews and focus groups.

The AICE questionnaire focused on nominal address, but some issues regarding pronominal address were raised by respondents. This shows that for many languages there is no clear separation between nominal and pronominal address in the mind of the users, and it is a misleading dichotomy. It was an aspect we underestimated in the AICE questionnaire, where cross-cultural transfer across languages (e.g. from English to German and vice versa) was commented on by respondents in terms of pronominal address, showing an awareness that first name versus last name address in English is not perfectly equivalent to T versus V pronominal address in German.

Schüpbach (2015) explicitly included greetings and leave-taking formulas in her overview of Swiss German address practices as they appear to differ from German usage in Germany, particularly in terms of their association with pronominal address. For example, similar to its use in Austrian German (Kretzenbacher 2011), the use of tschüss for leave-taking is mostly associated with T address in Swiss German but can be used with T or V address in Germany.

Apart from the use of pronominal and/or nominal address, the absence of direct address has also been studied. In particular, this topic has figured prominently in our Swedish data. In the AWE project for example, older Swedish participants often brought up the agony of choosing the right form of address in order not to offend the addressee, often resulting in the avoidance of direct address altogether. The sociohistorical background to address avoidance in Swedish had to do with the former prevalent title use and the condescending connotations that the V pronoun ni had attracted (cf. Clyne/Norrby/Warren 2009: 7-9).

4.2.6 Actual use, reported use or perceptions and attitudes?

For actual and reported use, there is a strong correlation with the data types discussed in section 4.1. Indirect elicited data from interviews and questionnaires provide data on reported use, whereas actual use can only be documented through authentic or natural data (their potential difficulties have been outlined in section 4.1). Perceptions and attitudes, however, can be gleaned from direct and indirect sources, elicited in questionnaires and collected from sources unconnected to research, such as the folk-linguistic data and metalinguistic discussions used in Schüpbach (2015).

Most of our research so far has been based on indirect data and has focused on reported use and users’ perceptions. These two aspects were successfully combined in the AWE and AICE projects, where the two perspectives were used to complement each other.

Actual use is often valued more highly than reported use, a view we have occasionally encountered in reviewer feedback to our publications. We would argue, however, that research focusing on reported use and users’ reflections has its own merits - particularly in the case of address and pluricentricity - as it allows us to investigate perceived norms and explore the pragmalinguistic knowledge of respondents, i.e. speaker perceptions (Jucker 2009) or folk pragmatics (Preston/Niedzielski 2017) (for a focus on actual use see Grahn and Norrby, this volume).

4.2.7 Quantitative or qualitative analysis?

House (2018), discussing her own contrastive research into cross-cultural pragmatics, argues that the view of a dichotomy between quantitative and qualitative data and methods is unproductive and that the two approaches should rather be seen as a continuum: some qualitative data, for example, can be quantified and thus analysed quantitatively. Consequently, she advocates for mixed-method approaches where appropriate, which can not only corroborate results but also add breadth and depth to the findings and counterbalance some of the weaknesses of each approach.

This resonates strongly with our experience where the combination of quantitative and qualitative datasets and approaches was used very productively. In the AWE project, for example, the initial focus group helped us identify the three prototypical address situations in contemporary German which were then confirmed quantitatively in the survey data (Kretzenbacher/Clyne/Schüpbach 2006: 17.13). Similarly, the combination of more quantitatively oriented closed questions in the questionnaire (which by nature are more general and do not convey much nuance) with open questions in the interviews led to richer insights. In the AICE conference questionnaire, using closed and open questions allowed us to combine quantitative and qualitative approaches more effectively.