Review of Discourse Markers in Native and Non-native English Discourse
| AUTHOR: Müller, Simone
TITLE: Discourse Markers in Native and Non-native English Discourse
SERIES: Pragmatics and Beyond
PUBLISHER: John Benjamins
Janet M. Fuller, Department of Anthropology, Southern Illinois University
This monograph is based on the doctoral dissertation by the author, and has
as a primary audience scholars interested in the study of discourse
markers, particularly English discourse markers. Because this book presents
a thorough review of the literature and a careful analysis of four
discourse markers ('so', 'well', 'you know' and 'like'), it is essential
reading for any researcher interested in this topic.
The volume begins with a long introductory chapter which discusses the
properties of functions of discourse markers and how they fit into the
study of second language acquisition and applied linguistics, and then goes
on to introduce the corpus and methodology used in the study. The following
four chapters each focus on one discourse marker, and the sixth chapter is
While for my taste the discussion of the data and research design would be
better off in a chapter separate from the general properties of discourse
markers, the organization of the introductory chapter is straightforward
and the sections clearly labeled. The first section of this chapter is a
good summary of previous research, with a particularly fine discussion of
the issues which have arisen in addressing the question of what constitutes
a discourse marker. This aspect of discourse marker analysis is a
particularly pressing problem in a study such as this, which includes a
quantitative analysis of all instances of the selected variables.
Unfortunately, although the author does reach a verdict on discourse marker
delimitation (syntactic independence and grammatical optionality is the
only feature which can clearly distinguish discourse marker from
non-discourse marker uses of a word), there is no concluding section on
properties of discourse markers which clearly states this. This point is
picked up again in the subsequent chapters, however, so it is not lost in
The author explicitly uses a 'bottom-up', data-driven approach to the study
of discourse markers, with the goal of cataloguing all functions of each of
the four selected discourse markers in her corpus. This corpus comes from
the Giessen-Long Beach Chaplin Corpus (GLBCC), collected in a study by
Andreas Jucker in cooperation with Sara Smith. Müller herself was
responsible for making most of the recordings used in the study. The
corpus was collected in an experimental setting. Students from the
University of Giessen and California State University at Long Beach were
assigned roles (A and B) and put in a room to watch a silent movie. After
the first part of the movie, the person in role A was asked to come out and
orally retell the first part of the film, while the person in role B
watched the second part. Partner B then retold the second part of the film
to partner A, and then the two were instructed to discuss the movie, with a
list of questions as a guide. All tasks in both settings were carried out
in English. In some cases, when only one person showed up for the taping
appointment, this person would re-tell the whole movie and then go through
the list of discussion questions on his/her own; this is referred to as
role C in the analysis. While the presence of speakers who had no
interlocutor is undesirable, the author used this difference to her
advantage by incorporating speaker role into her analysis as an independent
The variable which is the main focus of the study, however, is the status
of the speaker as a native or non-native speaker of English. All research
participants filled out a demographic questionnaire, which, for non-native
speakers of English, provided information about their acquisition of and
exposure to English. Only speakers who claimed English or German as their
(only) first language were included in Müller's analysis, resulting in 77
German native speakers and 34 native speakers of American English.
One chapter is allotted to each of the discourse markers 'so', 'well', 'you
know', and 'like'. The organization of each chapter is the same: first a
review of the functions of the discourse marker in the literature is
presented, and then there is a sub-section on each functional category
found in the GLBCC. Following this is a short summary of the functions
found in the data, and then quantitative results are presented, showing
which functions are statistically significant in their use by native and
non-native speakers, and also addressing correlations with other factors:
age, gender, relationship between the partners, speaker role, and the
language learning backgrounds of the non-native speakers. While both the
discussions of the previous literature on specific discourse marker
functions and the presentation of the findings are thorough and clearly
written, the overall format does not lend itself well to giving the reader
a good oversight of the findings; they are presented more in order of
discovery than in order of salience. The author begins the discussion of
functions by listing and discussing all of the non-discourse marker
functions of the term, which I do not feel adds to the comprehensibility of
the analysis; if these functions are to be included at all, they should
come at the end of the discussion. Also, each chapter ends somewhat
abruptly without a synthesis of the most important results for that
discourse marker. On the other hand, a laudable aspect of these chapters is
that each is written in such a way that they can stand alone for
researchers interested in only particular discourse markers, and this is
managed without excessive repetition across the chapters.
The chapter on 'so' is shown to perform nine different discourse functions
in this corpus. Three of these functions, the three most common in the
corpus, are used more by Americans at statistically significant levels.
Germans who had higher levels of contact to native speakers of English also
used 'so' more in these functions than the Germans who had little contact
with native English speakers.
In contrast, the findings for 'well' show that the German non-native
speakers use this discourse marker more than the native speakers, at least
for certain functions. Use of 'well' when searching for the right word was
much more frequent among the non-native speakers, and there were also two
functions – conclusive 'well' and use of 'well' to continue an
opinion/answer – which were used by some German but none of the Americans.
Somewhat surprisingly, no discussion is made of potential transfer from
German to explain these usages.
The results for 'you know', like those for 'so', showed the pattern of
Americans using the discourse marker at much higher rates than Germans, but
with stronger results than for 'so'. All but two of the functions showed
statistically significant differences in rates of use between groups.
However, as is always the case in doing qualitative research,
categorization of tokens is difficult and inevitably there will be
disagreement about the function of specific utterances. For instance, I
felt that the examples given in the category 'appeal for understanding'
could be incorporated into the rubric of 'marking lexical or content
search', as they both typically are used when a speaker is struggling to
articulate what s/he means. While this in itself is not a failing in the
study – any qualitative analysis must rely on the researcher's subjective
judgments – given this subjectivity, more discussion on broader patterns
would have strengthened the analysis (for example, more comparison of how
textual functions compared with interactional functions, or simply one
discourse marker with another).
My dissatisfaction with the functional categories constructed continues in
the chapter on 'like'. In particular, I find one category, 'marking
lexical focus', problematic, as it combines the concept of focus with the
'loose meaning' aspect of this discourse marker. It is unfortunate that my
own work on this topic (Fuller 2003) apparently came into print too late to
be taken into account in this analysis; our accounts are quite compatible.
In my work, I suggest that because things which are marked as having
approximate meaning (and earlier function of 'like') are in many cases also
the focus of the utterance, the focal meaning has grown out of the
approximate meaning. While there are uses which are clearly one or the
other, some uses are clearly both, as Müller suggests. However, as she
herself admits, it is difficult to categorize examples, in part because an
analyst cannot be sure if the speaker felt unsure about the exact meaning
of what follows 'like'. Some of the examples she gives in this category in
my opinion clearly mark approximate meanings (e.g., 'he was like a movie
director or something'), while other examples seem to fall more clearly
into the category of 'searching for an appropriate expression' (e.g., 'when
they like…take hands…). It should be noted that these critiques of Müller's
functional categories are based on the belief that her approach to
discourse markers – that is, the creation of functional categories – is a
valuable one and worthy of continued discussion.
The quantitative analyses presented for each discourse markers are
comprehensive, but there are several aspects of this treatment which
detract from the overall value of the volume. In general, too few
generalizations are gained from the statistical analysis. One specific
problem is that for each discourse marker there is a category of
'unclassified instances' – in and of itself quite reasonable – but the
author then insists on including these tokens in the quantitative analysis
and discussing how the frequencies of use of this group of miscellaneous,
unclear examples correlate with speaker group and other variables. This
aspect of the analysis is nonsensical.
Otherwise, the quantitative analysis is carried out carefully and reported
conscientiously, but does not offer a cohesive picture of differences
between groups or in terms of gender, age, etc. While the lack of clear
trends in the date in terms of age, gender, etc. is hardly the fault of the
author, the author does err on the side of reporting overly specific
findings if they achieve statistical significance even when these are not
representative of most speakers, resulting in a fragmented picture of the data.
In the conclusion, an excellent table (6.1, p. 246) shows all the functions
of the discourse markers and if they are used more by German or American
speakers. However, few clear generalizations are pulled out of these data.
Some results cited, e.g., female American speakers used the textual
functions of 'so' significantly more frequently than their male
counterparts (p.247), make an attempt at a somewhat broader scope, but no
explanation about why textual (as opposed to interactional) functions might
pattern in this way are provided. In short, the concluding remarks about
the quantitative analysis do not move beyond a description of the
frequencies found in the data. Motivation for these patterns, either in
terms of interactional functions or social motivations, is not provided.
However, the last several pages of the conclusion contain a very
interesting discussion of how discourse markers are presented in several
leading English textbooks and some discussion of how this might have
influenced the use of discourse markers in the non-native speaker data.
This link between pedagogical practices and language use is an enlightening
one which should be given further attention in the author's future research.
Overall, this book is a solid and ambitious piece of research. Although I
do not always agree with the functional categories constructed by the
author, this should not be seen as a failing of the author but rather a
difficulty inherent to qualitative research. A second criticism made here
of the book, that it does not provide motivated generalizations about the
data, is a more serious flaw, but must be mitigated by the admission that
discourse marker use is complex and variable, and not easily correlated to
independent variables. The analysis as it stands is a valuable reference
for anyone studying discourse markers, especially but not limited to the
four dealt with in these chapters.
Fuller, Janet M. 2003. 'Use of the discourse marker 'like' in interviews.'
Journal of Sociolinguistics 7:3.365-377.
| ABOUT THE REVIEWER:
Janet M. Fuller is an Associate Professor of Anthropology at Southern
Illinois University in Carbondale, IL. Her past research has included the
study of discourse markers in Pennsylvania German and (native and
non-native) English. She is currently finishing a sabbatical in Berlin,
Germany, where she is doing fieldwork with pre-teen children in
German-English bilingual classrooms. Among other things, she is interested
in the use of the discourse marker 'like' by young speakers of English as a
first or second language.