The effects of English-medium instruction on the use of textual and interpersonal pragmatic markers
Abstract
This study examines a semi and a full English-medium instruction (EMI) undergraduate program offered at a Catalan university in order to measure its effect on the students’ oral output. Specifically, it tackles the acquisition of pragmatic markers (PMs) by measuring four variables, the overall frequency of use, the variety of types, the use of textual PMs, and the use of interpersonal PMs. Oral data were collected via a monologue and an interaction task. The study is cross-sectional with 39 full-EMI and 33 semi-EMI participants in 2nd and 3rd year of study plus 10 native speakers. PM use was chosen for analysis due to the important role they play in communicative competence. Results show a significant increase in the overall frequency and variety of types of PMs used from year 2 to year 3. The full-EMI group used PMs at a significantly higher frequency and wider variety when compared to the semi-EMI group, neither group reached baseline levels for use of interpersonal PMs, and both groups displayed a higher use of textual PMs compared to the NSs.
Keywords:
Table of contents
1.Introduction
Actually, like, so, I think, well, you know, you see, it can be difficult to provide a clear
definition of these linguistic items, in fact, linguists debate on their roles, classifications and how to approach their study. For
example, the discourse analysis approach highlights coherence and the communicative effect of pragmatic markers (Redeker 1990Redeker, Gisela 1990 “Ideational and Pragmatic Markers of Discourse Structure.” Journal of Pragmatics 14 (3): 367–81. 
; Shiffrin 1987Shiffrin, Deborah 1987 Discourse Markers. Cambridge: Cambridge University Press. ), while the pragmatic approach takes a
syntactic and sematic analysis (Fraser 1999Fraser, Bruce 1999 “What Are Discourse Markers?” Journal of Pragmatics 31 (7): 931–52. ; Schourup
2011Schourup, Lawrence 2011 “The Discourse Marker Now: A Relevance-Theoretic Approach.” Journal of Pragmatics 43 (8). Elsevier B.V.: 2110–29. ) and finally, the approach stemming from relevance theory underscores the importance of cognition in the study of PMs
(Rouchota 1996Rouchota, Villy 1996 “Discourse Conectives: What Do They Link?” UCL Working Papers in Linguistics 8 (199–214).). In the present study, the term pragmatic marker is used (PM) and the
approach is a multifunctional one following Aijmer (2013)Aijmer, Karin 2013 “Analyzing Modal Adverbs as Modal Particles and Discourse Markers.” In Discourse Markers and Modal Particles: Categorization and Description, edited by Kirsten Fischer, 89–106. Amsterdam: John Benjamins Publishing Company. and Fischer (2014)Fischer, Kerstin 2014 “Discourse Markers.” In Handbook of Pragmatics, edited by Wolfram Bublitz, Andreas. H. Jucker, and Klaus. P. Schneider, Volume 3, 3:271–94. Berlin: De Gruyter Mouton. . PMs are thought to be closely related to a speaker’s oral fluency (Barón and Celaya 2010Barón, Júlia, and Mariía Luz Celaya 2010 “Developing Pragmatic Fluency in an EFL Context.” In EuroSLA Yearbook, 38–61. Amsterdam: John Benjamins Publishing Company. ; Trenchs-Parera 2009Trenchs-Parera, Mireia 2009 “Effects of Formal Instruction and a Stay Abroad on the Acquisition of Native-Like Oral Fluency.” Canadian Modern Language Review 65 (3): 365–93. ), and overall
communicative competence (Alcón and Safont Jordà 2008Alcón, Eva, and Maria Pilar Safont Jordà 2008 “Pragmatic Awareness in Second Language Acquisition.” In Encyclopedia of Language Education, edited by Jason Cenoz and N. H. Hornberger, Vol 6, 193–204. Springer Science and Business Media LLC. Netherlands. ; Halliday and Hasan 1976Halliday, Michael Alexander Kirkwood, and Ruqaiya Hasan 1976 Cohesion in English. 2013th ed. London and New York: Routledge.). Furthermore, PMs are known to play an important role both in first and second language
interaction, as contested by their constant use during interaction by native (NS) and non-native speakers (NNS) alike. During
interaction PMs perform a variety of different discourse management functions (Yates 2011Yates, Lynda 2011 “Interaction, Language Learning and Social Inclusion in Early Settlement.” International Journal of Bilingual Education and Bilingualism 14 (4): 457–71. )
which we will discuss in detail later on in this paper. Research thus far asserts that, while languages rely on PMs to organize
discourse, the frequency, distribution, and overlap of PM meanings from one language to another can vary greatly (LoCastro 2001LoCastro, Virginia 2001 “Individual Differences in Second Language Acquisition: Attitudes, Learner Subjectivity, and L2 Pragmatic
Norms.” System 29 (1): 69–89. ). This variation and difficulty pinpointing the core meaning and usage of PMs
are what present challenges for learners’ pragmatic, and, partially, communicative development. Indeed, even after long periods of
contact with the target language, learners are found to plateau when it comes to pragmatic learning (Romero-Trillo 2002Romero-Trillo, Jesús 2002 “The Pragmatic Fossilization of Discourse Markers in Non-Native Speakers of English.” Journal of Pragmatics 34 (6): 769–84. ).
Shifting now to the context of the present study, to begin with, Europe has been a model of the exchange and integration of languages
over many centuries and while English has been a world language for quite some time, its popularity and growth as a lingua franca in
recent years is remarkable (Costa and Coleman 2013Costa, Francesca, and James A. Coleman 2013 “A Survey of English-Medium Instruction in Italian Higher Education.” International Journal of Bilingual Education and Bilingualism 16 (1): 3–19. ). English as a lingua franca (ELF) stems
from both historical and political motives but is especially attributed to the globalization of the world’s economy (Crystal 2003Crystal, David 2003 English as a Global Language. Cambridge: Cambridge University Press. ). The role of English as the leading language in higher education means it is
frequently used as a medium for instruction, by, and for non-native English speakers in many European countries where English is not
the official language. This practice is known as English-medium instruction (EMI) (Hellekjær and
Hellekjær 2015Hellekjær, Glenn Ole, and Anne Inger Hellekjær 2015 “Is Anglophone Complacency a Virtue of Necessity? The Gap Between the Need for and Supply of Occupational Second
Foreign Language Skills in Norwegian Business and Government.” Scandinavian Journal of Educational Research 59 (2): 143–61. ) and its continual increase in implementation and practice across European universities is one of the
motivating factors for this study. Thus, the present study’s aims are two; the first is to report on the patterns of use of PMs by
students who study via EMI, as described by the overall frequency and variety of types of PMs used, then, more specifically, by the
frequency of use of textual, and interpersonal PMs separately. The second aim is to compare the use of PMs between a semi and full-EMI
program, and second to third-year students in order to detect if the different amounts of input provided by each program have an
effect on the use of PMs.
2.Pragmatic marker functions and role in communicative competence
The current study follows a functional pragmatic perspective (Aijmer 2013Aijmer, Karin 2013 “Analyzing Modal Adverbs as Modal Particles and Discourse Markers.” In Discourse Markers and Modal Particles: Categorization and Description, edited by Kirsten Fischer, 89–106. Amsterdam: John Benjamins Publishing Company. ; Andersen 2001Andersen, Gisle 2001 Pragmatic Markers and Sociolinguistic Variation. Philadelphia, NL: John Benjamins Publishing Company. ), which identifies PMs according to the main function the marker carries out in
its context. Two main types of markers are investigated, interpersonal and textual. According to Andersen (2001)Andersen, Gisle 2001 Pragmatic Markers and Sociolinguistic Variation. Philadelphia, NL: John Benjamins Publishing Company. , interpersonal markers are used to express social functions and convey both the speaker’s relation to a
proposition and their conception of the hearer’s relation to it. In contrast, textual markers are used to express relationships
between units of discourse and involve how the speaker perceives the structural relationship between propositions.
Looking firstly at the function of textual markers research has identified a number of sub-functions identified within this broad
function (see Table 1). Firstly, causal markers, which mark a causal relationship between two
units of discourse, for example, ‘because x, y’ or ‘x so/because y’ (Hyland 2005Hyland, Ken 2005 Metadiscourse. Exploring Interaction in Writing. Oxford: Continuum.; Müller 2005Müller, Simone 2005 Discourse Markers in Native and Non-Native English Discourse. Amsterdam: John Benjamins Publishing Company. ). Secondly, continuation markers, are
thought to connect units of talk (Fraser 1999Fraser, Bruce 1999 “What Are Discourse Markers?” Journal of Pragmatics 31 (7): 931–52. ; Shiffrin
1987Shiffrin, Deborah 1987 Discourse Markers. Cambridge: Cambridge University Press. ) and include items such as and, moreover, in addition, or so. Thirdly, contrast
markers such as although, but, however, and whereas mark the contrast between the main arguments of
each utterance (Fraser 1999Fraser, Bruce 1999 “What Are Discourse Markers?” Journal of Pragmatics 31 (7): 931–52. ). Fourthly, elaboration markers such as for example,
such as, like, I mean, and well function to reformulate, to introduce examples or are used to mark the
elaboration of discourse (Clark and Fox Tree 2002Clark, Herbert H., and Jean E. Fox Tree 2002 “Using Uh and Um in Spontaneous Speaking.” Cognition 84 (1): 73–111. ; Cuenca 2008Cuenca, Maria Josep 2008 “Pragmatic Markers in Contrast: The Case of Well.” Journal of Pragmatics 40 (8): 1373–91. ). Next, opening or closing of discourse markers are items such as alright, now, ok, so, and
well (Fraser 1999Fraser, Bruce 1999 “What Are Discourse Markers?” Journal of Pragmatics 31 (7): 931–52. ; Hyland
2005Hyland, Ken 2005 Metadiscourse. Exploring Interaction in Writing. Oxford: Continuum.; Cuenca 2008Cuenca, Maria Josep 2008 “Pragmatic Markers in Contrast: The Case of Well.” Journal of Pragmatics 40 (8): 1373–91. ). Then, topic shift or digression markers, which are used to
shift topics and include items such as anyway, or, whatever, so, regarding, and then (Fung and Carter 2007Fung, Loretta, and Ronald Carter 2007 “Discourse Markers and Spoken English: Native and Learner Use in Pedagogic Settings.” Applied Linguistics 28 (3): 410–39. ; Buysse 2012Buysse, Lieven 2012 “So as a Multifunctional Discourse Marker in Native and Learner Speech.” Journal of Pragmatics 44 (13): 1764–82. ; González 2005González, Monserrat 2005 “Pragmatic Markers and Discourse Coherence Relations in English and Catalan Oral Narrative.” Discourse Studies 7 (1): 53–86. ; Pons Bordería and Estellés Arguedas
2009Pons Bordería, Salvador, and Maria Estellés Arguedas 2009 “Expressing Digression Linguistically: Do Digressive Markers Exist?” Journal of Pragmatics 41: 921–36. ). Following this, items such as then, well, next, firstly, and
after which are used to present a sequence of events or mark temporal value (González 2005González, Monserrat 2005 “Pragmatic Markers and Discourse Coherence Relations in English and Catalan Oral Narrative.” Discourse Studies 7 (1): 53–86. ) and are known as sequence markers. Finally, summary markers are those markers which mark the introduction
to a concluding or summarizing segment of discourse, for example, so, yeah, well, and to conclude
(Müller 2005Müller, Simone 2005 Discourse Markers in Native and Non-Native English Discourse. Amsterdam: John Benjamins Publishing Company. ; Buysse 2012Buysse, Lieven 2012 “So as a Multifunctional Discourse Marker in Native and Learner Speech.” Journal of Pragmatics 44 (13): 1764–82. ).
Now shifting to interpersonal markers, again previous research has identified sub-functions within this broad function (see Table 1). Firstly, markers such as yeah, ok, sure, right, and I
see which show receipt of information or signal understanding and listenership (Brinton 1996Brinton, Laurel J. 1996 Pragmatic Markers in English. Berlin: Mouton de Gruyter. ; Shively 2015Shively, Rachel L. 2015 “Developing Interactional Competence during Study Abroad: Listener Responses in L2 Spanish.” System 48: 86–98. ). Secondly, markers to stimulate and maintain
interaction are employed by the speaker in an effort to continue the flow of discourse for example, yeah, right,
really, and great. Thirdly, alignment markers are used to mark alignment or create a closeness (or
distance) between speakers and could include items such as exactly, I agree, totally, yeah, that’s right, and,
yes (Maschler 1994Maschler, Yael 1994 “Metalanguaging and Discourse Markers in Bilingual Conversation.” Language in Society 23 (3): 325–66. ). Fourthly, markers such as you
know, and I mean project and signal shared knowledge and common ground (González 2005González, Monserrat 2005 “Pragmatic Markers and Discourse Coherence Relations in English and Catalan Oral Narrative.” Discourse Studies 7 (1): 53–86. ). Next, there are markers which signal hesitation or repair such as I mean and
well. Following this, there are attitude markers such as I think, definitely, basically, and
absolutely which express the speakers’ attitudes towards what is being uttered (Brinton 1996Brinton, Laurel J. 1996 Pragmatic Markers in English. Berlin: Mouton de Gruyter. ). Finally, politeness markers serve to mitigate or downgrade and include items such as I
mean, well, sort of, kind of, and I think (Chodorowska
1997Chodorowska, Marianna 1997 “On the Polite Function of ¿me Entiendes? In Spanish.” Journal of Pragmatics 28 (3): 355–71. ).
| Functions of textual markers | Functions of interpersonal markers |
|---|---|
| To show causal relationships to show consequence or effect, to mark the link between two clauses | To mark receipt of information, to show listenership and support to the speaker |
| To mark a contrast between two clauses or between two parts of the discourse | To stimulate or maintain interaction, to assess listener comprehension and engagement |
| To show a continuation of discourse on the same topic, to add additional information | To align or disalign oneself with the speaker by expressing agreement or disagreement |
| To elaborate, reformulate or exemplify | To mark joint construction of knowledge, mark common ground |
| To signal opening or closing of discourse or mark the end or beginning of a turn | To signal hesitation, thinking or repair |
| To show the temporal sequence between clauses or between two parts of discourse | To mark attitudes, stance or emotional reactions |
| To signal shifts or transitions of discourse topics, to mark digression from one topic to another or return to a previous topic | To intensify, boost, downgrade, hedge or serve as politeness markers |
| To indicate or preface results, summary, or conclusions |
The distinction between textual and interpersonal markers made in the functional approach is a reflection of Bachman’s model of
communicative competence (1990Bachman, Lyle F. 1990 Fundamental Considerations in Language Testing. Oxford: Oxford University Press.), and Bachman and Palmer (1996). In this model,
communicative competence is subdivided into two parts, organizational competence, and pragmatic competence. Organizational competence
refers to grammatical accuracy and textual competence, while pragmatic competence encompasses the relationship between utterances and
the speaker’s relationship to them. There is evidence which reports a relationship between communicative competence and pragmatic
marker use. For example, it has been argued that in order for second language users to achieve pragmatic fluency, they must master a
series of discourse strategies, such as: initiate and change topics, ‘carry weight’ in a conversation, uptake and respond
appropriately, align one’s turn by anticipating the end of others turns, and to appropriately fill or un-fill pauses (House 2003 2003 “Teaching and Learning Pragmatic Fluency in a Foreign Language: The Case English as a Lingua Franca.” In Pragmatic Competence and Foreign Language Teaching, edited by Alicia Martínez-Flor, Ester Usó Juan, and Ana Fernández Guerra, 133–60. Castelló: Universidad Jaume I.). Furthermore, pragmatic fluency has been recognised as the ability to
appropriately hedge, and down-tone utterances, to carry out appropriate speech acts according to the context, to save face, express
politeness, or be intentionally vague, to mark and check shared knowledge and to reformulate or to monitor the state of shared
knowledge (Lin 2016Lin, Yen Liang 2016 “Discourse Marking in Spoken Intercultural Communication between British and Taiwanese Adolescent
Learners.” Pragmatics 26 (2): 221–45. ). These researchers argue that PMs are the units of speech that enable
speakers to carry out these strategies. A few notable studies have investigated the relationship between PMs and communicative
competence further. Such as Riggenbach (1999)Riggenbach, Heidi 1999 Discourse Analysis in the Language Classroom: The Spoken Language. Michigan: University of Michigan Press. , who found that learners who were rated as
highly fluent filled most of their pauses with a PM rather than leaving them empty, and Barón and
Celaya (2010)Barón, Júlia, and Mariía Luz Celaya 2010 “Developing Pragmatic Fluency in an EFL Context.” In EuroSLA Yearbook, 38–61. Amsterdam: John Benjamins Publishing Company. who reported that learners who filled their pauses with PMs were perceived to have more fluent speech.
Similarly, Shively (2015)Shively, Rachel L. 2015 “Developing Interactional Competence during Study Abroad: Listener Responses in L2 Spanish.” System 48: 86–98. studied the use of PMs and perceived fluency, it was reported
that those learners who were rated as more fluent used an overall wider variety and frequency of PMs. These studies clearly
demonstrate the incorporation of PMs into speech aids in perceived communicative competence.
Shifting to PM comprehension, a couple of studies were found to have investigated the processing of PMs during EMI classes. One of
which, Flowerdew and Tauroza (1995)Flowerdew, John, and Steve Tauroza 1995 “The Effect of Discourse Markers on Second Language Comprehension.” Studies in Second Language Acquisition 17: 435–58. , examined lecture comprehension of EMI students by
playing a video lecture with all PMs removed to one group, and playing the lecture as normal to another group. They report that
learners understood the lecture with PMs significantly better than those who watched the lecture without PMs. Similarly, Jung (2003)Jung, Euen Hyuk Sarah 2003 “The Role of Discourse Signaling Cues in Second Language Listening Comprehension.” The Modern Language Journal 87 (4): 562–77. found that PMs in lectures played an important role in comprehension. These
studies provide evidence that learners do attend to PMs during lectures, and that they rely on PMs to process oral discourse on an
implicit level.
2.1The acquisition of pragmatic markers in instructed second language acquisition
Due to the multi-functional nature and minimal instruction of PMs in conventional instructed second language acquisition (ISLA) the
integration of PMs into one’s speech becomes a challenge (Bardovi-Harlig and Griffin
2005Bardovi-Harlig, Kathleen, and Robert Griffin 2005 “L2 Pragmatic Awareness: Evidence from the ESL Classroom.” System 33: 401–15. ). A reason for this may be the tendency for grammatical competence to take precedence over pragmatic competence.
Despite this challenge, it is important for learners to integrate PMs into their speech otherwise undesired communication errors
can ensue. For example, lack of clarity or precision, lack of coherence, difficultly on behalf of the interlocutor to follow the
discourse due to lack of textual PMs use. Or the lack of interpersonal PMs could cause inappropriate style or manner, or
contribute to the lack of the creation, and maintenance, of the relationship of the speakers to each other and the discourse.
Regarding the acquisition of PMs, research analyzing the effects of ISLA shows that learners tend to use PMs for a much narrower
scope and frequency than NSs do, as they seem to struggle to correctly identify the functions of the items. For example, Bu (2013)Bu, Jiemin 2013 “A Study of the Acquisition of Discourse Markers by Chinese Learners of English.” International Journal of English Studies JN – International Journal of English Studies 13 (1): 29–50. http://revistas.um.es/ijes. found that while learners used the same PMs as NSs, they did not use them for
the same functions as the NSs did, resulting in both a restricted range and an unnatural use of PMs. Müller (2005)Müller, Simone 2005 Discourse Markers in Native and Non-Native English Discourse. Amsterdam: John Benjamins Publishing Company. reported similar findings; namely, that in addition to differing patterns of PM use
learners also assigned new functions to PMs that were not found in the native discourse. Other studies report an over or underuse
of PMs compared to NSs’. For example, Liu (2016)Liu, Binmei 2016 “Effect of L2 Exposure: From a Perspective of Discourse Markers.” Applied Linguistics Review. studied a high and a low exposure
group and found the high exposure group used PMs at a higher rate than the low ISLA exposure group. It was further noted that the
functions the learners used the PMs for differed between both the learner groups and the NSs group. Similarly, Fung and Carter (2007)Fung, Loretta, and Ronald Carter 2007 “Discourse Markers and Spoken English: Native and Learner Use in Pedagogic Settings.” Applied Linguistics 28 (3): 410–39. investigated types of PMs used, and found that learners in ISLA
contexts used textual markers at high frequencies and interpersonal markers more sparingly, and that NSs used PMs for a much wider
variety of functions than the learners did. The authors argue that the use of PMs reflects the type of input they receive.
Research has also detected a correlation between the overall frequency and variety of PM use and overall language proficiency.
Specifically, low-proficiency groups produce less PMs as well as a lesser variety than advanced groups do. So it seems that, as
learners increase their overall use of PMs, the variety of functions they use them for widens as well (Neary-Sundquist 2014 2014 “The Use of Pragmatic Markers across Proficiency Levels in Second Language Speech.” Studies in Second Language Learning & Teaching 4 (4): 637–63. ). To summarize, research thus far suggests that learners use PMs at a lesser
frequency and variety than NS do and that as proficiency and input increase so too do frequency and variety of PMs used. It is
also clear that learners struggle to identify and incorporate the wide range of functions that PMs can have, especially
interpersonal markers, thus committing errors in use, overusing or underusing certain PMs. The research also points out that there
is a strong need for speakers to use PMs both for their own as well as for their interlocutor’s benefit, a process which appears
to be mainly implicitly rather than explicitly acquired.
3.English-medium instruction in higher education
English has quickly established itself as the dominant language in higher education (Wilkinson
2004Wilkinson, Robert 2004 Integrating Content and Language: Meeting the Challenge of a Multilingual Higher Education. Maastricht: Universitaire Pers.). This is particularly evident among graduate programs where the amount of EMI courses has tripled in number over the
past decade (Wachter and Maiworm 2014Wachter, Bernd, and Friedhelm Maiworm 2014 English-Taught Programmes in European Higher Education: The State of Play in 2014. Bonn: Lemmens Medien GmbH.). The term EMI has come to refer to many different
types of programs, in fact, implementation and practice of EMI differs according to each institution’s, and nation’s language policy
and goals. What is of interest to the present study is the intention of language learning within these programs. For example, Knapp (2011)Knapp, Anne 2011 “Using English as a Lingua Franca for (Mis-) Managing Conflict in an International University Context: An Example
from a Course in Engineering.” Journal of Pragmatics 43 (4): 978–90. identified three types of EMI: type one, international EMI programs, these are
mainly master and doctoral level courses which are designed to attract international students. International EMI courses do not state
explicit language learning goals on the curriculum and are instead focused solely on teaching the content of the course, English is
viewed as a means to communicate and deliver the course contents. Type two EMI courses are geared towards students of English
linguistics and literature or future English teachers; these programs have clear language learning goals and include explicit formal
language instruction courses. Type three EMI courses are aimed at local students in any subject, with a twofold goal, to provide
specialization in the field of study, and to prepare students to use English in their professional lives. These courses are often
undergraduate or master degree programs. This third type has implicit language learning goals, while they are not stated officially,
nor actively supported in any way. For example, Gundermann (2014Gundermann, Susanne 2014 English-Medium Instruction: Modelling the Role of the Native Speaker in a Lingua Franca Context. PhD dissertation, University of Freiburg, Germany., 42–43) asserted that in
the higher education context the use of “English is linked to a language learning goal.’ And that ‘In such
contexts, the use of English often fulfills the purpose of practicing and enhancing language skills, alongside with content
learning”. This implicit language learning goal has been confirmed by other researchers, namely Margić and Žeželić (2015)Margić, Branka, and Tea Žeželić 2015 “The Implementation of English-Medium Instruction in Croatian Higher Education: Attitudes, Expectations and
Concerns.” In English as a Scientific and Research Language, edited by R. Plo Alanstrué, and C. Pérez-Llantada, 311–32. Berlin: De Gruyter Mouton. , and Sert (2008)Sert, Nehir 2008 “The Language of Instruction Dilemma in the Turkish Context.” System 36 (2): 156–71. who found that in the
Croatian and Turkish contexts respectively, language improvements were expected from participation in EMI and so, it seems evident
that whether stated explicitly or not, stakeholders expect linguistic gains from participation in EMI. However, language learning via
EMI is yet to be confirmed by empirical research as noted by Pecorari et–al. (2011Pecorari, D., P. Shaw, A. Irvine, and H. Malmström 2011 “English for Academic Purposes at Swedish Universities: Teachers’ Objectives and Practices.” Inglés Con Fines Académicos En Las Universidades Suecas: Objetivos y Prácticas de Los Profesores 22 (2011): 55–78. http://www.scopus.com/inward/record.url?eid=2-s2.0-80054012099&partnerID=40&md5=e8e230bd9d0d08dfd26591f272965841., 57)
“there is a widespread belief that incorporating elements of English into the curriculum has the serendipitous effect of
promoting incidental language learning. […] this belief rests on a number of tacit and largely untested ideas.” this
points to a need for further investigation into language learning in EMI.
3.1The impact of English-medium instruction on second language acquisition
Although second language acquisition (SLA) in the EMI context has not been widely investigated to date, there is some evidence to
suggest that this learning context may contribute to language learning. For example, authenticity of interaction has been reported
to positively stimulate acquisition by cognitively engaging learners (García Mayo and Lázaro
Ibarrola 2015García Mayo, María del Pilar, and Amparo Lázaro Ibarrola 2015 “Do Children Negotiate for Meaning in Task-Based Interaction? Evidence from CLIL and EFL Settings.” System 54: 40–54. ). In EMI participants attend lectures and seminars, and interact with peers and professors. These
situations oblige them to process, reformulate and reproduce what they learn in their courses. In addition, they are expected to
communicate spontaneously through oral, digital, and written means. These factors, combined with increased contact hours with
English, provide opportunities for authentic language use, which may lead us to consider EMI to be a stimulating environment for
language gains. However, there are some confounding factors which might lead to little measureable gains from EMI alone. For
example, in some cases, due to relatively low initial level in their communication and comprehension skills, EMI alone may not be
enough to enhance students’ proficiency in the target language (Harley et–al. 1990Harley, Brigit, Patrick Allen, Jim Cummins, and Merrill Swain 1990 The Development of Second Language Proficiency. Cambridge: Cambridge University Press. ).
Secondly, most EMI programs do not offer any focus on language, which as research from ISLA points out is essential for language
improvement. A third factor could be that without explicit instruction, correction, and focused teaching, content-based language
learners may not reach target like production (Lyster 2017Lyster, Roy 2017 “Content-Based Language Teaching.” In The Routledge Handbook of Instructed Second Language Acquisition, edited by Shawn Loewen and Masatoshi Sato, 87–107. New York: Routledge. ). This lack of correction
and instruction combined with other features of communication strategies in ELF contexts such as tendencies to ‘let it pass’ and
‘make it normal’ when linguistic errors are made may prevent language improvement from occurring (see Firth 1996Firth, Alan 1996 “The Discursive Accomplishment of Normality: On ‘Lingua Franca’ English and Conversation Analysis.” Journal of Pragmatics 26 (2): 237–59. ; House 1993House, Juliane 1993 “Toward a Model for the Analysis of Inappropriate Responses.” In Interlanguage Pragmatics, edited by Gabriele Kasper, and Shohsana Blum-Kulka, 161–82. Oxford: Oxford University Press.). Finally, factors related to
individual differences such as attitudes and motivations of the individual may play a role in language gains. For example, EMI
students attend to and put effort into learning the content of the course rather than to focus on language form and on learning
English. They report dedicating very little time to studying English and more time and effort to reading and preparing for their
courses (Sert 2008Sert, Nehir 2008 “The Language of Instruction Dilemma in the Turkish Context.” System 36 (2): 156–71. ; Tazl 2011Tazl, Dieter 2011 “English Medium Masters Programs at an Austrian University of Applied Sciences: Attitudes, Experiences and
Challenges.” Journal of English for Academic Purposes 10 (4): 252–70. ).
Regarding empirical research on SLA in EMI, we find a handful of studies. Firstly, investigating English proficiency improvements,
Lei and Hu (2014)Lei, Jun, and Guangwei Hu 2014 “Is English-Medium Instruction Effective in Improving Chinese Undergraduate Students’ English
Competence?” IRAL – International Review of Applied Linguistics in Language Teaching 52 (2): 99–126. measured the effect of EMI on proficiency level in a group of
first and second-year university students in China. They reported no significant differences in an oral interview nor in a written
exam between the EMI group and the non-EMI group. In line with this study, Ament and
Pérez-Vidal (2015)Ament, Jennifer R., and Carmen Pérez-Vidal 2015 “Linguistic Outcomes of English Medium Instruction Programmes in Higher Education: A Study on Economics
Undergraduates at a Catalan University.” Higher Learning Research Communications 5 (51): 47–67. found no significant differences between EMI and semi-EMI learners according to listening or writing
scores. However, an improvement on grammar scores was reported in the semi-EMI group. In contrast to these studies, Ritcher (2017)Ritcher, Karen 2017 “Researching Tertiary EMI and Pronunciation. A Case from Vienna.” In Intergrating Content and Language in Higher Education; Perspectives on Professional Practice, edited by Jennifer Valcke, and Robert Wilkinson, 117–34. Frankfurt: Peter Lang. who investigated pronunciation found that semi-EMI learners improved and
significantly outperformed non-EMI instruction students according to perceived foreign accent. Regarding the effect of EMI on
pragmatic abilities Taguchi (2011)Taguchi, Naoko 2011 “Do Proficiency and Study-Abroad Experience Affect Speech Act Production? Analysis of Appropriateness, Accuracy, and
Fluency.” IRAL – International Review of Applied Linguistics in Language Teaching 49 (4): 265–93. examined the development of pragmatic competence
among first-year EMI university students in Japan. She investigated the appropriateness of expressing opinions in formal and
informal settings. Results showed progress in informal contexts but none in formal contexts. In a different study, Ament and Barón (2018)Ament, Jennifer R., and Júlia Barón Parés 2018 “The Acquisition of Discourse Markers in the English-Medium Instruction Context.” In Learning Context Effects: Study Abroad, Formal Instruction, and International Immersion Classrooms, edited by Carmen Pérez-Vidal, Sonia López-Serrano, Jennifer R. Ament, and Dakota Thomas-Wilhelm, 44–74. Berlin: Language Science Press. examined the use of PMs among EMI and non-EMI students. They found
that EMI students produced a higher frequency of structural PMs and signposted more clearly. The non-EMI students, on the other
hand, used more referential PMs which may be the first and easiest category of PMs to acquire. To summarize, research shows that
incidental linguistic improvement from EMI participation is minimal with the exception of the domain of pronunciation and that
semi-EMI participants may improve more than full-EMI participants and that context of learning plays a role in pragmatic
learning.
4.Objectives of the study
The current study has a two-fold objective: to measure the use of PMs in two different intensity EMI programs, a full intensity (full-EMI) and a low intensity (semi-EMI), across two years of study (2nd and 3rd year); in order to examine the following four variables, (i) overall frequency of use, (ii) variety of types of PMs used, (iii) frequency of use of textual PMs, and (ix) frequency of use of interpersonal PMs in each of the programs. Therefore, the following two research questions have been established.
-
Are there differences in overall frequency and variety of PMs used as a result of different degrees of intensity of the EMI programs (Full or Semi-EMI) and length (2 or 3 years of exposure) to EMI?
-
Are there differences when comparing the frequency of textual and interpersonal PMs used as a result of different degrees of intensity of the EMI programs (Full or Semi-EMI) and length (2 or 3 years of exposure) of EMI?
5.Methodology
5.1Design and participants
The study compares full-EMI and semi-EMI learners in their second year of study, with full-EMI and semi-EMI learners in their third year of study. It adopts a cross-sectional design. Participants were eighty-two students enrolled in an undergraduate degree at a university in Spain. Thirty-seven participants were second-year students, sixteen of which were enrolled in a semi-EMI program, and twenty-one of which in a full-EMI program (henceforth SIM2 and IM2). Another thirty-five participants were in their third year of studies in the same programs, eighteen of which in a semi-EMI and seventeen of which in a full-EMI (henceforth SIM3 and IM3). Data from a control group of ten native speakers served as a baseline for comparison (henceforth NS group). Results from the language background questionnaire revealed that 88% of the participants were Spanish/Catalan bilinguals and these languages were also the languages of their previous education. While 12% were from other language backgrounds (Basque, Slovenian, Chinese, Serbian and Ukrainian). All participants reported English as a third language. The mean age of participants was twenty, 41% were male, and 56% were female (see Table 2).
The full-EMI groups were enrolled in an International Business and Economics degree. Participants in the semi-EMI groups were enrolled in either an Economics or Business Administration degree at the same university. Considering that each degree program consists of 425 contact hours per academic year, for the full-EMI group, all 425 hours are delivered through the English language, while the semi-EMI group had an exposure of 35 contact hours. Table 2 provides the number of cumulative hours spent in EMI per group and other baseline data just mentioned.
| Experimental groups | IM2 | IM3 | SIM2 | SIM3 | NSs |
|---|---|---|---|---|---|
| Number (Age) | 21(19) | 17(20) | 16(19) | 18(20) | 10(22) |
| Cumulative hours of EMI | 637.5 | 1112.5 | 35 | 70 | n/a |
| First language | 16 – Cat/Sp 5 – Other | 17 – Cat/Sp | 15 – Cat/Sp 1 – Other | 14 – Cat/Sp 4 – Other | 10 – Eng |
| Year of study at time of data collection | 2nd year | 3rd year | 2nd year | 3rd year | 4th or 5th year |
The majority of the faculty of Economics at the university in question share the same L1 as the majority of students. At the time of data collection, there was one extra-national professor who was Austrian. There were no native English speakers on staff at the faculty at the time of data collection. The lecturer’s level of English is neither certified nor controlled at the university in question, however, there are programs offered on behalf of the university to provide both training and support to faculty who teach through English. Professors report speaking English exclusively during class time and do not provide students with explicit language instruction or correction, save a glossary providing a translation of some key terms.11.Data collected via personal communication with a group of faculty members.
5.2Instruments
Three instruments were designed for this study: a language background questionnaire, a monologue and an oral interaction task. All
instruments were piloted previous to the study and were found to be effective and adequate at eliciting the desired type of
language as well as reliable (Ament and Barón 2018Ament, Jennifer R., and Júlia Barón Parés 2018 “The Acquisition of Discourse Markers in the English-Medium Instruction Context.” In Learning Context Effects: Study Abroad, Formal Instruction, and International Immersion Classrooms, edited by Carmen Pérez-Vidal, Sonia López-Serrano, Jennifer R. Ament, and Dakota Thomas-Wilhelm, 44–74. Berlin: Language Science Press.).
The language background questionnaire has been discussed in the previous section. The online Cambridge placement test was administered to control for proficiency. Those participants who scored either over C1 or below B1 on the CEFR22.Common European Framework of Reference for language. were excluded from the analysis.
A monologue and an interaction task were used to collect oral data. Two different types of tasks were chosen in order to broaden the communicative contexts the speakers were placed in and therefore provide more opportunities for a range of PMs to occur. The monologue task (henceforth MON) was completed individually, participants were asked to introduce themselves to the researcher and include information regarding the languages they speak, their English language learning experience, which degree they were taking and why they had chosen to take it in English.
The interaction task (henceforth INT) required participants to engage in conversation with another participant. Participants were asked three different questions that were related to their field of study. This method of data collection was chosen for three interrelated reasons: (i) it has been advocated in the literature that elicited conversations or ‘interactions arranged for research purposes can be most useful sources of data’ (Kasper and Rose 2002Kasper, Gabriele, and Kenneth R. Rose 2002 Pragmatic Development in a Second Language. (Language Learning, Vol. 52). Oxford: Blackwell Publishing Ltd. http://search.proquest.com/docview/85583993?accountid=14548%5Cnhttp://metadata.lib.hku.hk/hku?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&genre=article&sid=ProQ:ProQ:llba&atitle=Pragmatic+Development+in+a+Second+Language&title=Language+Le., 80); (ii) this type of procedure can elicit, interpersonal functions, coordinated speaker-listener functions, turn-taking, and back-channeling, and finally, (iii) because it has been argued that elicited conversation can tap into learners interactional competence (Kasper and Rose 2002Kasper, Gabriele, and Kenneth R. Rose 2002 Pragmatic Development in a Second Language. (Language Learning, Vol. 52). Oxford: Blackwell Publishing Ltd. http://search.proquest.com/docview/85583993?accountid=14548%5Cnhttp://metadata.lib.hku.hk/hku?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&genre=article&sid=ProQ:ProQ:llba&atitle=Pragmatic+Development+in+a+Second+Language&title=Language+Le.).
5.3Procedure
Participants completed the web questionnaire and the proficiency test online. Oral tests were carried out in sound attenuated cabins. The MON task was carried out first; two minutes were given to record the participants’ responses. This was followed by the INT task. For this task, participants were organized into pairs, and they recorded themselves.33.Audacity software was used for recording data. In response to the questions, the participants were asked to include their opinions, personal experiences and anything else they felt they wanted to express. Participants were asked to discuss each question for two minutes.
5.4Analysis
Audio recordings were transcribed and verified to ensure accuracy. The researchers controlled the coding by having another researcher re-coded 25% of the transcripts and then comparing their results, and discrepancies were discussed until an agreement was reached. The researchers identified and tagged each PM used in the MON and INT task as either textual or interpersonal. Table 3 includes examples of which items were coded as PMs.
| Functions of textual markers | Items found in the data | Functions of interpersonal markers | Items found in the data |
|---|---|---|---|
| To show causal and consequential relationships | Because, so, and | To signal receipt of information | Okay, right, yeah |
| To indicate results/summary | So, like, well, and, yeah | To show support to the interlocutor | Okay, great, I know, exactly, sure |
| To indicate conclusions | Finally, then | To align oneself with the speaker and to signal joint construction of knowledge | Exactly, I agree, totally, yeah, that’s right, yes, and, in addition |
| To mark contrast | But, however, and, although | To stimulate interaction | Right? Yeah? Really? Great! |
| To mark disjunction or digression | Or, anyway, or something, or so, whatever, | To hesitate or show repair | I mean, well, sort of, kind of, I think |
| To signal opening or closing of discourse | Okay, right, alright, so, let’s start, to conclude/ in conclusion, yeah, that’s it, that’s all | To denote thinking processes | Well, I think, |
| To signal shifts or transition of discourse | So, well, and then, and what about, and how about, and yeah, but | To assess the interlocutor’s knowledge | Right? You know what I mean? |
| To show temporal sequence | First, firstly, secondly, next, then, finally, now, first of all | To act as a hedging device | I think, I’m not sure, kind of, sort of, you know? |
| To show continuation of discourse | And, yeah, because, so | To indicate attitudes | I think, definitely, basically, absolutely, exactly |
| To Elaborate, Reformulate and exemplify | I mean, like, and, it’s like, that is, for example |
In order to illustrate the coding process, two examples from the data are provided. The items that were considered PMs and that were coded are marked in italics with a subscript number. Following each transcript is a brief explanation of how the researcher decided to code each PM.
Second-year semi-EMI participants Chris and Wendy44.All names are made up to protect the privacy of the participants discussing question one.
umm I think(1) that communicating face to face is more spontaneous and(2) you think you view no you see the reaction and(3) it’s more directly and effective I think(4) .
Yeah(5) I agree(6) with you and(7) other communication systems like(8) I think(9) she said letters, well(10) nowadays letters have been kind of difficult even to transport and everything but(11) we tend to not use letters, only for important things like(12) banks do and schools can also like(13) send umm letters but(14) I think(15) it’s not the most effective and most useful.
Yeah(16) maybe it have it has been replaced umm to emails and internet.
… two turns discussing electronic communication…
So(17) we would agree that it’s much more effective face to face so(18) we see their reactions
In Example (1) items 4, and 15 (subscript numbers) were coded as interpersonal markers in these cases I think marks the speaker’s attitudes towards the statement. Whereas Examples 1, and 9, I think, were identified as interpersonal markers and thought to function to mark cognitive processing. Examples 2, 3, and 7, and, were coded as textual markers; specifically, they function to show the continuation of discourse and the addition of information. Item 5, 6 yeah and 16 I agree are examples of interpersonal markers functioning to align the speaker with the interlocutor, by showing reception of information and active participation in the conversation. Items 8, 12, and 13, like were coded as textual markers and identified as signalling an upcoming example or elaboration on what is about to be said. Then, items 11, and 14 but were coded as textual markers which function here to show the contrast between two utterances or parts of discourse. Finally, Example 17, so was coded as a textual marker which functions here to summarize and mark the closing of the discourse, and then, item 18, so was considered a textual marker which, in this context, signals a causal relationship between two utterances.
Second-year full-EMI participants Borja and Aina discussing question two.
well (19) first of all (20) I think (21) technology allows companies and people to have instant communication because(22) it’s free it doesn’t cost anything what do you think about it?
yeah (23) I think (24) the same thing and also it has evolved a lot.
It allows people to travel and(25) you know(26) to go to another country maybe to work and(27) to have a constant communication with their family and(28) so(29) that allows people to move all over the world without ahh you know(30) ahh without distances being an important factor to take into account.
(….) maybe you will ahh be less comfortable talking in public.
yeah (31) yeah (32) definitely (33) .
In Example (2) item 19, well, was coded as a textual marker, functioning here to signal the opening of the discourse. Item 20, first of all, functions textually to show discourse sequencing. Items 21, and 24, I think, were coded as interpersonal markers which function here to indicate the speaker’s attitudes towards an utterance. Item 22 and 29, because and so, were coded as textual markers, which function to show causal relationships. Examples 23, 31, and 32, yeah were coded as interpersonal markers, serving to align the speaker with the interlocutor, by showing receipt of information and active participation in the conversation. Then, items 25, 27, and 28, and, were coded as textual markers, used here to mark the addition of information and continuation of discourse. Next, items 26, and 30, you know, were considered interpersonal markers which were operating to show shared knowledge and signal a desire to appeal to the interlocutor and co-construct knowledge. Finally, item 33, definitely, was identified as an interpersonal marker signalling joint construction of knowledge.
As demonstrated by the examples, the researchers coded each PM according to its function in the given context. The data were
analyzed in this manner based on previous research (Fung and Carter 2007Fung, Loretta, and Ronald Carter 2007 “Discourse Markers and Spoken English: Native and Learner Use in Pedagogic Settings.” Applied Linguistics 28 (3): 410–39. ; Neary-Sundquist 2013Neary-Sundquist, Colleen 2013 “Task Type Effects on Pragmatic Marker Use by Learners at Varying Proficiency Levels.” L2 Journal 5 (2): 1–21. http://escholarship.org/uc/item/9bm489h8.). After coding, calculations were drawn for each participant, for a
total number of PMs, tokens (frequency), total types (variety), total textual tokens (frequency) and total interpersonal tokens
(frequency) per participant. Then, these totals, except for the variety measure, were divided by the total number of words uttered
and multiplied by 100 to calculate the percentage of PMs used per participant in relation to total words spoken. This approach to
the investigation of different types of PMs in the same study complements the typical type of study carried out in PM
investigation, where the focus is on one or only a handful of PMs. Furthermore, it has been noted that ‘statistical methods in
qualitative studies are highly desirable because they provide research with validation methods to determine significant
correlations among qualitative features of connectives’ as put forth by Pons Bordería (2006Pons Bordería, Salvador 2006 “A Functional Approach to the Study of Discourse Markers.” In Approaches to Discourse Particles, edited by Kerstin Fischer, 77–99. Amsterdam: John Benjamin., 81). The data were then analysed statistically, each variable has two levels, year of study (2nd or 3rd) and
intensity of immersion (full-EMI or semi-EMI) therefore, a 2x2 factorial analysis of variance (ANOVA) was chosen to analyze the
data.
6.Results
In order to prepare the data for analysis in SPSS statistical assumptions such as data normality, Levene’s tests and Shapiro-Wilks tests were verified and met before carrying out the analysis. Furthermore, a Bonferroni correction was used to rule out any chance of obtaining false positives, with an alpha level of .05.
The first ANOVA explored the effect of year and immersion level on proficiency by using year, and immersion as fixed factors and proficiency as a dependent variable, the results show a significant interaction effect between year and proficiency F(1.68) = 7.22, p = .009, = .096. Meaning that third-year students have a higher proficiency than second-year students. Due to this interaction, proficiency was used as a covariate in the following tests to eliminate the effect proficiency plays on the data and to thus, be able to detect differences based solely on immersion and year of studies. See Figure 1.
A series of 2x2 analysis of covariance (ANCOVA)’s were then carried out to determine if there were any effects of year or immersion on the following four dependent variables; overall frequency of PMs, overall variety, the frequency of textual PMs, and the frequency of interpersonal PMs. A summary of descriptive statistics is provided in Table 4.
| Summary of descriptive statistics | ||||
|---|---|---|---|---|
| Variable | Group | Mean | Standard deviation | Number of participants |
| Frequency | SIM2 | 10.35 | 2.18 | 16 |
| SIM3 | 12.81 | 2.21 | 18 | |
| IM2 | 12.3 | 2.26 | 21 | |
| IM3 | 13.22 | 1.23 | 17 | |
| NS | 13.72 | 2.85 | 10 | |
| Variety | SIM2 | 10.5 | 2.99 | 16 |
| SIM3 | 13.2 | 2.24 | 18 | |
| IM2 | 13.86 | 2.35 | 21 | |
| IM3 | 14.9 | 1.73 | 17 | |
| NS | 15 | 2.45 | 10 | |
| Textual Markers | SIM2 | 6.27 | 1.22 | 16 |
| SIM3 | 7.81 | 1.14 | 18 | |
| IM2 | 7.7 | 1.1 | 21 | |
| IM3 | 8.63 | 1.05 | 17 | |
| NS | 7.02 | 2 | 10 | |
| Interpersonal Markers | SIM2 | 4.08 | 1.65 | 16 |
| SIM3 | 4.57 | 1.66 | 18 | |
| IM2 | 4.61 | 2.11 | 21 | |
| IM3 | 4.59 | .9 | 17 | |
| NS | 6.71 | 2.02 | 10 | |
The significance of the descriptive statistics will be described when interpreting the main effects found from the ANCOVA tests.
6.1Total frequency and variety of PMs used
To answer the first research question ‘are there differences in overall frequency and variety of PMs used as a result of different degrees of intensity of the EMI programs and length to EMI?’ A 2x2 ANCOVA was performed, significant main effects were found between frequency of production of PMs and years of exposure F = (1.67) = 11.2, p = .001, = .143, and also between frequency and intensity of immersion F = (1.67) = 5.93, p = .02, = .08. Furthermore, full-EMI learners consistently produced more PMs than semi-EMI learners, as is made evident in the descriptive statistics; IM2 (M = 12.3, sd = 2.26) and SIM2 (M = 10.35, sd = 2.18); and SIM3 (M = 12.81, sd = 2.21) and IM3 (M = 13.22, sd = 1.23). This result shows that there is a significant increase in the frequency of use of PMs as a result of full-EMI and that as learners spend more time in any type of EMI program they significantly increase the frequency of PM use.
Now turning to the overall variety of types of PMs used, significant main effects were found between the variety of PMs, and both, years of exposure, F = (1.67) = 9.49, p = .003, = .124 and intensity of immersion F(1.67) = 18.79, p = .000, = 2.19. This result reveals that both the intensity of immersion and years of exposure to EMI have significant effects on the variety of PMs used. Specifically, the variety of PMs used significantly increases from year two SIM2 (M = 10.5, sd = 2.99), IM2 (M = 13.86, sd = 2.35) to year three SIM3 (M = 13.2, sd = 2.24), IM3 (M = 14.9, sd = 1.73) and full-EMI learners consistently integrate a larger variety of PMs in their speech than semi-EMI learners do, see Figure 3.
6.2Frequency of textual and interpersonal markers
To address the second research question ‘are there differences when comparing the frequency of textual and interpersonal PMs used as a result of different degrees of intensity of the EMI programs and length of EMI?’ 2x2 ANOVAs were run on the data. Significant main effects were detected between the frequency of use of textual PMs and years of exposure F = (1.67) = 20.16, p = .000, = .231, as well as for intensity of immersion F(1.67) = 17.52, p = .000, = .207. This result reveals that a full-EMI program has a significant differential effect on the usage of textual markers when compared to a semi-EMI program. Both groups produced significantly more textual PMs in year three SIM3 (M = 7.81, sd = 1.14), IM3 (M = 8.63, sd = 1.05), than in year two SIM2 (M = 6.27, sd = 1.22), IM2 (M = 7.7, sd = 1.1). It is further observed that both IM and SIM participants used more textual PMs than the NS baselines did (M = 15, sd = 2.45), see Figure 4.
There were no significant main effects found between the production of interpersonal PMs and years of exposure F = (1.67) = 1.11 p = .3 = .016 or intensity of immersion F = (1.67) = .02 p = .89 = .000. The IM group experiences no change from year two IM2 (M = 4.61, sd = 2.11), to year three IM3 (M = 4.59, sd = .9), while the SIM group experiences a slight increase in overall production of interpersonal PMs from year two SIM2 (M = 4.08, sd = 1.65), to year three albeit non-significant (M = 4.57, sd = 1.66). Neither group approaches NS baseline (M = 6.71, sd = 2.02). See Figure 5.
To summarize, the results for question one show that both the intensity of immersion and the number of years of exposure to EMI have significant impacts on the overall frequency of PMs, as well as on the overall variety of PMs. The results for research question two show that both the intensity of immersion and the number of years of exposure to EMI have significant impacts on the frequency of use of textual PMs, but that neither variable (year nor immersion) have any effect on the frequency of use of interpersonal PMs. A summary of the ANCOVA tests is provided in Table 5.
| Summary of ANCOVA main effects. | ||||
|---|---|---|---|---|
| Variable | Groups | F value | P value | |
| Frequency | Year | (1.67) = 11.2 | .001 | .143 |
| Immersion | (1.67) = 5.93 | .02 | .08 | |
| Variety | Year | (1.67) = 9.49 | .003 | .231 |
| Immersion | (1.67) = 18.79 | .000 | 2.19 | |
| Frequency of Textual PMs | Year | (1.67) = 20.16 | .000 | .231 |
| Immersion | (1.67) = 17.52 | .000 | .207 | |
| Frequency of Interpersonal PMs | Year | (1.67) = 1.11 | .3 | .016 |
| Immersion | (1.67) = .02 | .89 | .000 | |
7.Discussion
Results from the first research question show that both intensity of immersion and length of time spent in EMI have positive effects on PM production measured through the overall frequency of production as well as through the total variety of PMs used. Regarding the second research question, results showed that both the intensity of program and the length of time spent in EMI led to an increased use in textual PMs, and, that both groups produce more textual PMs than the NS baseline group while on the other hand neither intensity nor length had a significant impact on the production of interpersonal PMs. What follows is a discussion of the results of each question and how the findings contribute to the current field of research.
7.1Frequency and variety of use of PMs
Our full-EMI participants produced a wider variety and higher frequency of PMs. This finding is in line with previous studies, for
example, Liu (2016)Liu, Binmei 2016 “Effect of L2 Exposure: From a Perspective of Discourse Markers.” Applied Linguistics Review. , who analysed ISLA learners and found that those with high exposure
to the target language used PMs at higher frequencies and varieties than learners with low exposure. As well as with Hellermann and Vergun (2007)Hellermann, John, and Andrea Vergun 2007 “Language Which Is Not Taught: The Discourse Marker Use of Beginning Adult Learners of English.” Journal of Pragmatics 39 (1): 157–79. who reported that increased frequency and variety of use of
PMs in ISLA contexts correlated with the amount of contact with, and exposure to, the target language community. The present study
adds to what is known regarding the effects of EMI on learners’ language, and, more specifically, on PMs. It reveals that learners
experience positive progress via participation in a semi or full-EMI program, as both groups improved significantly from year two
to year three. This finding demonstrates that even with very few EMI hours the semi-EMI group experiences significant improvement,
in fact, they experience the same progression pattern as their full-EMI counterparts. The findings also show that, although
overall proficiency has been found to correlate with increased use of PMs, as found in such studies as Neary-Sundquist (2014) 2014 “The Use of Pragmatic Markers across Proficiency Levels in Second Language Speech.” Studies in Second Language Learning & Teaching 4 (4): 637–63. and Fernández, Gates Tapia, and Lu
(2014)Fernández, Julieta, Anna Gates Tapia, and Xiaofei Lu 2014 “Oral Proficiency and Pragmatic Marker Use in L2 Spoken Spanish: The Case of Pues and Bueno.” Journal of Pragmatics 74: 150–64. who report that higher proficiency learners produced higher frequency and varieties of PMs, proficiency is not
the only factor that affects the production of PMs, and other elements such as intensity of exposure play significant roles in
pragmatic learning. This finding is also in line with Matsumura’s (2003)Matsumura, Shoichi 2003 “Modelling the Relationships among Interlanguage Pragmatic Development, L2 Proficiency, and Exposure to
L2.” Applied Linguistics 24 (4): 465–491+558. study, in
which overall exposure was demonstrated to be a stronger predictor of pragmatic development than proficiency level.
7.2Frequency of textual and interpersonal PMs
Results from the second research question align with Fung and Carter’s (2007)Fung, Loretta, and Ronald Carter 2007 “Discourse Markers and Spoken English: Native and Learner Use in Pedagogic Settings.” Applied Linguistics 28 (3): 410–39. findings
which showed that learners relied more on textual PMs than on interpersonal ones. A suggested explanation for this high rate of
production of textual markers is the context of learning as noted by Ament and Barón
(2018)Ament, Jennifer R., and Júlia Barón Parés 2018 “The Acquisition of Discourse Markers in the English-Medium Instruction Context.” In Learning Context Effects: Study Abroad, Formal Instruction, and International Immersion Classrooms, edited by Carmen Pérez-Vidal, Sonia López-Serrano, Jennifer R. Ament, and Dakota Thomas-Wilhelm, 44–74. Berlin: Language Science Press., EMI is a formal, academic setting, where textual PMs are likely to occur at much higher frequencies than
interpersonal PMs. If we reflect on the functions of textual PMs such as to structure discourse, mark openings and closings,
emphasise, and shift topics, to name a few, we can see a parallel between these PMs and the types of pragmatic functions lecturers
employ when delivering their courses and therefore, which PMs are available in the input. Thus, textual markers may be argued to
be more salient as well as more critical to the understanding of EMI courses, as was attested in both Flowerdew and Tauroza (1995)Flowerdew, John, and Steve Tauroza 1995 “The Effect of Discourse Markers on Second Language Comprehension.” Studies in Second Language Acquisition 17: 435–58. and in Jung (2003)Jung, Euen Hyuk Sarah 2003 “The Role of Discourse Signaling Cues in Second Language Listening Comprehension.” The Modern Language Journal 87 (4): 562–77. . So, in sum,
the importance of textual PMs to the comprehension of lectures, combined with the frequent use of textual PMs in academic
discourse may explain why learners produce these markers at high frequencies.
The pattern of use of textual markers contrasts with that of interpersonal markers. Interpersonal markers were used much less
frequently. Additionally, significant differences were not found for either the amount of exposure nor for the intensity of
immersion in either group. This could be because interpersonal markers are less salient in the EMI context and the pragmatic
information they provide is not essential to comprehension. Due to this factor, learners might have a tendency to skip over
interpersonal markers without processing them; this would mean these markers are cognitively attended to less than textual PMs
are. This finding is in line with Firth (1996)Firth, Alan 1996 “The Discursive Accomplishment of Normality: On ‘Lingua Franca’ English and Conversation Analysis.” Journal of Pragmatics 26 (2): 237–59. who suggested that if a linguistic item
is not essential for communication/comprehension it is often skipped over and left uncorrected as it does not provide crucial
information. This finding echoes House’s (2003) 2003 “Teaching and Learning Pragmatic Fluency in a Foreign Language: The Case English as a Lingua Franca.” In Pragmatic Competence and Foreign Language Teaching, edited by Alicia Martínez-Flor, Ester Usó Juan, and Ana Fernández Guerra, 133–60. Castelló: Universidad Jaume I. results, who noted a trend for learners
to not mark their relation to a proposition, and, furthermore, to not take the hearer’s relation to the proposition into
consideration. The participants of the present study were found to behave in a similar manner to House’s participants in that they
were reported to use hardly any interpersonal PMs at all, and instead, use raw negation, addition, and rejections.
Another contrast between the use of interpersonal PMs and textual PMs is that they are rarely written but instead are used at very
high frequencies amongst native speakers during oral communication while on the contrary, textual PMs are highly functional in
writing, and therefore, are reinforced even further in the input of the academic setting. Thus, the underuse of interpersonal
markers might be explained by the learners having little exposure to English outside of the EMI classroom, and even less contact
with native speakers (who use PMs more frequently than learners).55.The professors in the present study were not native speakers of English and this could have further affected the low frequency of
interpersonal PMs in the input. Although this is speculation, class observation is necessary to confirm this hypothesis. In fact, in previous
studies, it has been found that learners improve and increase their use of PMs as they socialize and integrate into the local
community (Shively 2015Shively, Rachel L. 2015 “Developing Interactional Competence during Study Abroad: Listener Responses in L2 Spanish.” System 48: 86–98. ). This may be why no improvement is detected on this measure in
the present study, as EMI learners have little opportunity to socialize or integrate into an English-speaking community, certainly
less than a study abroad or naturalistic settings would provide (DeKyser 2007DeKyser, Robert 2007 Practice in a Second Language: Perspectives from Applied Linguistics and Cognitive Psychology. Cambridge: Cambridge University Press. ; Pérez-Vidal ed. 2014Pérez-Vidal, Carmen ed. 2014 Language Acquisition in Study Abroad and Formal Instruction Contexts. (= Applied Linguistics Series, 13). Amsterdam: John Benjamins Publishing Company. ). This lack of socialization may become especially clear when
measuring the production of interpersonal markers, as the use of interpersonal PMs is closely related to the speech norms of a
local community (Liu 2016Liu, Binmei 2016 “Effect of L2 Exposure: From a Perspective of Discourse Markers.” Applied Linguistics Review. ) and in EMI there is no such speech community. Interpreting
the results in light of the communicative competence model, it becomes clear that EMI learners are very aware and have a highly
developed organizational competence as is reflected through their highly developed use of textual PMs and, in contrast, their
pragmatic competence remains unchanged by EMI exposure (Bachman 1990Bachman, Lyle F. 1990 Fundamental Considerations in Language Testing. Oxford: Oxford University Press.; Alcón and Safont Jordà 2008Alcón, Eva, and Maria Pilar Safont Jordà 2008 “Pragmatic Awareness in Second Language Acquisition.” In Encyclopedia of Language Education, edited by Jason Cenoz and N. H. Hornberger, Vol 6, 193–204. Springer Science and Business Media LLC. Netherlands. ). Finally, the results from this study seem to suggest that
textual PMs may be more readily or easily acquired compared to interpersonal PMs. Evidence suggests that at least in the EMI
context textual PMs are incorporated into speech before interpersonal PMs.
8.Conclusions
Results from this study highlight a number of patterns, firstly, both full-EMI and semi-EMI programs have significant effects on the overall frequency of PMs, and variety of PMs. Secondly, as time spent in either full or semi-EMI program increases (from year 2 to year 3), learners experience significant increases in the overall frequency of PMs, and variety of PMs. And finally, both full-EMI and semi-EMI have positive effects on the frequency of use of textual PMs, but insignificant effects on the frequency of use of interpersonal PMs. Thus, it appears that learners experience cumulative gains over time spent in EMI and that even taking just a few courses through English can have a real impact on oral output. More specifically, due to the increased input received, learners begin to modify their output by incorporating more and more PMs into their speech, while they make progress in recognizing and identifying the functions of a wider variety of PMs. Additionally, the findings of our study show that textual PMs are acquired before interpersonal PMs. What remains unclear is whether this pattern of acquisition is due to the EMI context or if PMs are acquired in this order for other reasons not investigated in this study. Furthermore, it seems that while EMI provides plenty of input and language learning opportunities, there may be other factors necessary in order for learners to integrate interpersonal PMs into their language skills. And finally, as a consequence of the previous results coming out of our data, it would seem to be the case that organizational competence might be developed via EMI but that EMI would not so easily lend itself to the enhancement of pragmatic competence.
Some limitations of the present study are that although gains in the use of PMs have been reported, further research is necessary to assess other language domains before making large-scale policy changes. Secondly, the data were collected in a simulated conversation, it would be interesting to gather natural occurring data to get a more accurate representation of the learner’s output. And finally, it would be useful to conduct a longitudinal study rather than a cross-sectional to see the change within each participant and be able to draw stronger conclusions.
Finally, the implications of the findings thus far would appear to be first, that a full-EMI program may not be necessary since
significant language gains occur in semi-EMI, this then creates space for a more balanced approach towards multilingual policies,
where local languages could be supported and strengthened without any loss to English or any other language. This addresses a concern
brought up by many communities on how to implement parallel language policies in Sweden (Bolton
and Kuteeva 2012Bolton, Kingsley, and Maria Kuteeva 2012 “English as an Academic Language at a Swedish University: Parallel Language Use and the ‘threat’ of
English.” Journal of Multilingual and Multicultural Development 33 (5): 429–47. ) and The Basque Country (Doiz, Lasagabaster, and Sierra 2014Doiz, Aintzane, David Lasagabaster, and Juan Manuel Sierra 2014 “Language Friction and Multilingual Policies in Higher Education: The Stakeholders’ View.” Journal of Multilingual and Multicultural Development 35 (4). Taylor & Francis: 345–60. )
to name only two. This study also provides evidence that more support is needed via giving more attention to the explicit teaching of
PMs especially interpersonal ones. Explicit teaching or the implementation of language support for EMI students may increase learner’s
noticing of PMs and could lead to an accelerated acquisition of this and other language features.
Acknowledgements
This work was supported by the AGENCIA UNIVERSITARIA DE RECERCA (AGAUR), in Catalonia, [2014 SGR and current 2017] and by the Ministry of Economy and Competitiveness [FFI2013-48640-C2-1-P]. Additionally the authors would like to thank colleagues and friends who helped contribute to this publication.
Notes
References