Metadiscourse in oral discussions and

Metadiscourse in oral discussions and
Copyright 2012 Beata Maria Latawiec
Submitted in partial fulfillment of the requirements
for the degree of Doctor of Philosophy in Secondary and Continuing Education
in the Graduate College of the
University of Illinois at Urbana-Champaign, 2012
Urbana, Illinois
Doctoral Committee:
Professor Sarah J. McCarthey, Chair
Professor Richard C. Anderson, Director of Research
Professor Mark A. Dressman
Professor Janet S. Gaffney
Speakers and writers use metadiscourse to guide, caution, and implore their audiences. Written
metadiscourse is a term for self-reflective expressions that help writers negotiate interactional
meanings, assist in expressing viewpoints and engagement with readers (Hyland, 2005), or
convey attitudes towards written text (Vande Kopple, 1985). Research suggests that effective
metadiscourse results in more transparent organization of discourse and greater global, local, and
thematic coherence; metacognitive awareness; better learning from text; greater rhetorical force
of arguments; and, enhanced social performance and attitude.
The present exploratory study examines how children use metadiscourse and how it
functionally interplays with discourse proper in their interaction with peers in collaborative
small-group discussions and reflective writing. The students participating in Collaborative
Reasoning paradigm are expected to take a position on the question, present reasons, back
reasons with evidence and further reasoning, challenge each other when they disagree, weigh
reasons and evidence, and change positions when warranted. These argumentative moves at
every turn require evaluation, interpersonal, communicative and rhetorical skills with rich
repertoire of metalanguage. The study’s major facets are wrapped around metadiscourse enabling
such evaluative, interpersonal, organizational, metalinguistic and intersubjectivity-inviting flows
of meaning.
The study consists of several components employing different research methods.
Quantitative methods were employed for identification of systematic patterns in written and oral
discourse and correspondences between them, as well as for investigation of socio-linguistic
variation in metadiscourse across discussions. Qualitative methods were used, with an
elementary-to-holistic approach, for interpretation and evaluation of the patterns and explication
of how different metadiscursive variants help or impede the flow of meaning in discussions and
essays. A complex taxonomy was devised to accommodate the broad evaluative, organizing and
intersubjective meta-functions that comprised 50 elementary categories that could capture
variation in both modalities with an extra fine-grain.
The results suggest both in speaking and writing students use twice as much evaluative as
organizing metadiscourse. Intersubjectivity in essays is marginal, and in discussions amounts to
less than 10% (compared to organizing and evaluative metadiscourse). Essays written by CR
children bear heavy traces of dialogism and are open to perspectives of others. CR essays’
attitudinal stance is more strongly expressed by normative modals than in speaking. The result
indicates power relations at play, when in face-to-face confrontations students intuitively use
face saving techniques as reflected in language (weaker attitudinal stance, hedging, mitigating).
CR discussions show high-engagement level and recipient-targeted engagement marking via
what if-soliciting, gestures and importantly, perlocutionary/ coercive commentary that forms
bonds with hearers (also found in an experimental study showing CR-participants advantage over
non-CR in essays, Latawiec et al., in preparation).
Intersubjectivity-signaling dropped over the discussion series, which indicates greater
focus on informational flow than on interpersonal relations. Yet metalanguage considerably
increased attesting the specificity of CR language, which puts high premium on talking with
assessment activity, and also suggesting carry-over effect to writing. Boosters in boys’ explicit
speech acts in oral argumentation may be considered as exponents of power, a flip side of sociolinguistic theories of female “weaker gender” being compensated by vagueness in language
which is not confirmed in this study, nor is the “rapport-talk” of women or “report-talk” of men.
The qualitative results suggest intersubjectivity-vagueness can obscure the propositional flow by
halting or slowing down the flow of arguments (less seen in writing), though oral distribution
patterns suggest its saliency for peer in-group solidarity signaling.
Lastly, for an optimal flow of propositional meaning, organizing and evaluative
metadiscourse need to be counterbalanced (rather than one meta-function overtly prevail). For
instance, evaluative and attitudinal stance marking used in excess, i.e. not counterbalanced by
organizing metadiscourse, sets forewarning signals and results in resistance to potential
manipulative attempts. Also, proliferate organizing variants of metadiscourse get “de-ranked” or
weakened in their cohering functions. Notably, the facilitation of information integration seems
linked to objective rather than subjective markers. Hence, a key to a successful content flow
seems to lie in the ‘golden means’ between evaluative and organizing variants of metadiscourse.
To my Family
Just as Clark argues about the participants in a conversation who “work together against a
background of shared information” (1993, p.4), so did the work towards this dissertation.
This dissertation would have not been conceptualized or completed without sharing
knowledge, insights and intricacies of the research pursuits. Neither would the research goals be
accomplished without interpersonally or intimately shared support or guidance. And all this has
gone into the making of this study.
In this major life-time effort then I owe countless counseling and consultations to
Professor Richard Anderson, as my official research director, but also a research ‘guru’ and an
esteemed paragon to follow. Not only did he offer his invaluable expertise, incomparable
research and educational experience but also private time, private face and private feelings of
encouragement and personal warmth. To thank for all of this is a challenge as there may be not
enough words to express. Yet as in argumentation challenges are welcome and bring a spark to
conversation, I thus would like to sparkly thank you, and with a sparkling face!
My deepest thanks also go to Professor Sarah McCarthey, my official advisor, whose
guidance and intellectual insight were unparalleled. Her understanding of discourse analysts’
struggles and inspirational moments of thought that accompany quests for meaning and
interpretation was both comforting and educational, and as such contributory to my discourse
analysis pursuits. Her pithy comments and praiseworthy continuous assistance, even in the
mundane matters, make me a lifetime debtor. Yet, a happy and grateful debtor indeed!
As official and unofficial committee members, superb researchers and educators, whose
praiseworthy pedagogic skills I had an honor and pleasure to experience en route to my doctoral
degree, Professors Mark Dressman and Janet Gaffney will remain in my heartfelt regards and
memories for the rest of my life.
Special thanks go to Avon Crismore and William Vande Kopple for offering their helpful
comments and suggestions as well as well-informed feedback that propelled the engine of the
exploratory research with an extra acceleration (to put it metaphorically).
This work would have not been finalized nor even come into budding without unrivalled
encouragement and continuous support of my family, and especially my husband, Chris, who
toiled towards this study along with me and offered a helping hand and linguistic insights that
made the work lighter. Also, my son’s software helping hand and mind have to be
acknowledged here, and appreciated as like a life-buoy made the data unsinkable.
I would also like to thank my colleagues from the Collaborative Reasoning Research
Group for their contribution to the data collection and strenuous inter-rater reliability training
and work, as well as for intellectual stimulation that the environment offered.
The research reported here was also supported in part by the Institute of Education
Sciences, U.S. Department of Education, through Grant R305A080347 to The Board of Trustees
of the University of Illinois.
Thank you all.
Table of Contents
Chapter 1: Introduction ............................................................................................................... 1
1.1 Overarching Goal ...................................................................................................................... 1
1.2 What is Metadiscourse? ............................................................................................................ 1
1.3 The Rationale ............................................................................................................................ 2
1.4 The Emphasis ............................................................................................................................ 3
1.5 The Dissertation Design ............................................................................................................ 5
Chapter 2: Literature Review ...................................................................................................... 9
2.1 Language Functions and Metadiscourse ................................................................................... 9
2.2 Language in Two Different Modalities – Spoken and Written............................................... 11
2.3 Metadiscourse and CR ............................................................................................................ 16
2.4 Metadiscourse and Argument Schema Theory in Learning to Think Well ............................ 21
2.5 Metadiscourse and Pragmatics ................................................................................................ 25
2.6 Metadiscourse and a Sociolinguistic Variation ....................................................................... 29
2.7 Metadiscourse and Talk in Action .......................................................................................... 35
2.8 Metadiscourse Pros and Cons ................................................................................................. 36
2.9 Metadiscourse Systems Contributory to this Study’s Construct............................................. 38
2.10 Metadiscourse Study Goals................................................................................................... 44
Chapter 3: Methods .................................................................................................................... 47
3.1 Participants.............................................................................................................................. 47
3.2 Procedures ............................................................................................................................... 48
3.3 Data Analysis .......................................................................................................................... 52
3.3.1 Speaking ‘turn’ defined........................................................................................................ 52
3.3.2 Coding Scheme .................................................................................................................... 54
3.3.3 Adopted Taxonomy and its Mechanics................................................................................ 55
3.3.4 Broad Bracketing and Building-Block Categories in the Coding Scheme .......................... 59
3.3.5 Coding Process..................................................................................................................... 78
Chapter 4: Results....................................................................................................................... 79
4.1 Quantitative Results ................................................................................................................ 79
4.2 Qualitative Results ................................................................................................................ 105
Chapter 5: Discussion ............................................................................................................... 130
Chapter 6: Conclusions ............................................................................................................ 152
References .................................................................................................................................. 167
Chapter 1
Kevin: Shelby, @you said [email protected]{Reminder of Material+hearsay” Evidential} they should
shorten the race by @[email protected]{cautious elem.} one or two miles, @I [email protected] with you,@
[email protected] @what [email protected] you had like {=e.g./Speech Act) a little German Shepherd, @[email protected]
five, @[email protected] almost a year ago, @[email protected] he was @[email protected] strong @[email protected] he run @[email protected]
fast, you didn’t know it was @[email protected] that long. @[email protected] it @[email protected] died, @[email protected] @right
in the [email protected] of the race. (CHA_8SFY:05:14.91)
1.1 Overarching Goal
The purpose of the study is to investigate whether and how children use metadiscourse
and how functionally it interplays with discourse proper in their interaction with peers in
collaborative small-group discussions and reflective writing. As demonstrated with the above
excerpt from an authentic CR-exposed student’s talk, metadiscourse is commonly used (marked
above with @ symbols) and intricately woven into the fabric of communicative flow.
Nevertheless, it is hardly empirically investigated especially among young age communicators in
the US educational context. It is also commonly underestimated in the scope of functions and
forms as well as its multi-dimensional effects on the flow of propositional content, which this
exploratory study hopes to repair by exploring, demonstrating and evaluating the density of
functions and forms used by Collaborative Reasoning small-group discussants and reflective
1.2 What is Metadiscourse?
Speakers and writers use metadiscourse to guide, caution, and implore their audiences
(Latawiec, Anderson, Nguyen-Jahiel, Ma et al., in preparation). Metadiscourse is a cover term
for self-reflective expressions that help speakers and writers negotiate interactional meanings,
assist in expressing viewpoints or in promoting engagement with their audience (Hyland, 2005),
or convey attitudes towards their spoken or written text (Vande Kopple, 1985). Vande Kopple
(1985; 1997) holds that metadiscourse is used not to expand ‘referential material’ or content of
the discourse, but to help readers connect, organize, interpret, evaluate and develop attitudes
toward that material. He also argues that discourse studies’ primary concern should be with
metadiscursive functions rather than specific forms that can fulfill those functions, especially as
sometimes one form can fulfill several metadiscursive functions.
Hyland in his Metadiscourse: Exploration of interaction in writing (2005) views
metadiscourse as representation of “the writer’s awareness of the unfolding text as discourse:
how we situate ourselves and our readers in a text to create convincing, coherent prose in
particular social contexts” (p. ix). In speech, as Simons (1994) claims, “going meta” may also
mean “to provide a strategic, reflexive, frame-altering response to another’s prior message or
messages, or to a shared message context” (p. 469). It is this aspect of context that unites the two
conceptions just cited, and which adds importance to metadiscourse that “lies not in semantic
meanings of particular forms but meanings which only become operative within a particular
context, both invoking and reinforcing that context with regard to audience, purpose and
community” (Hyland, 2005, pp. 194-5). Thus conceived metadiscourse essentially is a pragmatic
and sociolinguistic act as well as “a rhetorical act in social action” (Crismore, 1985).
1.3 The Rationale
The rationale for the study in children’s metadiscourse and metatalk results from the
assumption of bidirectionality of interaction between cognitive processes and discourse
formation and, likewise, the examination of processes in both directions – from mind to
discourse and from discourse to the mind (cf. Billig, 2003). This formulation is very much in line
with Billig’s (2003) argument that “many of the phenomena that psychologists traditionally treat
as internal mental processes are actually formed within discourse” (p. 228).
The rationale also gains credence from the findings in the previous study of the author
(Latawiec et al., in progress) into the influence of small-group discussions on children’s
metadiscourse in reflective writing. Specifically, the past findings revealed improved coherence
and signaling of argumentative inferential moves as well as reader-engaging metadiscursive
techniques (as compared to non-CR exposed children) that added to the persuasiveness of CRwriting. Results for Non-CR exposed students that revealed their overwhelming use of additives
in lieu of other connectives to bind their propositional content as well as greater use of emphatics
and hypothetical scenarios rather than explicit reasoning structures of if/since-then (frequented
by CR students with high reading skills though) may enhance knowledge about useful and less
useful types of metadiscourse in children’s persuasive language uses and thinking via the
language. The past comparative study between CR-exposed students and the control groups
suggests prospective differential uses by various groups of school children of metadiscourse
(interacting with discourse proper) in consecutive communicative tasks of conversational (microgenetically analyzed within a series of 10 discussions) and written argumentation (in a final
1.4 The Emphasis
In the current study, the emphasis is placed on language users in a social context rather
than mere linguistic forms or language uses. This is a socio-linguistically and pragmatically
dynamic perspective, in which language is seen as a dynamic process that can be captured with
the metaphor of a flowing stream (after Chafe, 2001) or actually two streams – a stream of
thoughts and a stream of sounds (in the oral medium) or signs (in the written medium). Thus, to
use Chafe’s figurative language, the metadiscourse study explores the sounds in speech or the
signs in writing (observable language in action) for the externalization of thoughts (revealing the
mind in action).
Consequently, this analysis of children’s speech with its metadiscursive focus takes a
dynamically constructivist “discourse flow” perspective, in which conversation is viewed as “a
uniquely human and extraordinarily important way by which separate minds are able to influence
and be influenced by each other, managing to some extent, and always imperfectly, to bridge the
gap between them, not by constructing any kind of lasting object but through a constant interplay
of constantly changing ideas” (Chafe, 2001, p. 686; emphasis added).
It is thus posited here that the adopted qualitative and quantitative methods of the study
across two corpora may be merited with the pluses of both methods. On the one hand, a finegrained socio-linguistic method of corpus analysis unraveling layers of children’s talk and
writing allows the capture of children’s moment-by-moment (microgenetic) development of
higher-level cognitive and social abilities in “socialization through and via metadiscourse”
(Mauranen, 2001; cf. Literature review section). The conversational talk-in action analysis will
allow for capturing talk of children as collaborative reasoners and cooperative communicators interacting in accord with the Cooperative Principle in Grice’s terms (1975) (which is essentially
a pragmatic dimension of communicative competence paradigms, e.g., Bachman 1990; Bachman
& Palmer 1996; Canale & Swain 1980). On the other hand, the cross-sectional analysis (as a
function of gender, reading skills, school location and ethnicity) facilitates inquiry into the use of
Collaborative Reasoning discussion techniques with its dynamic flow of argument schema
within classroom learning environments.
All in all, the mixed methods as applied in the study allow for studying the connections,
correspondences and/or differences in the use of metadiscourse in the two language domains –
by combining the scrutiny of qualitative observations and quantitative power of statistical tests.
1.5 The Dissertation Design
In Chapter 2, the study first reviews aspects of language use that reveal an array of
metadiscursive functions and theoretical approaches to metadiscourse in two modalities of
writing and speaking, with their distinctive features, affordances and/or constraints.
Secondly, the review targets the multi-faceted paradigm of Collaborative Reasoning (CR)
to which the subjects of this study were exposed. To fully realize the scope of demands and
opportunities (including cognitive and metacognitive, socio-cultural, linguistic, rhetorical and
pragmatic), the theoretical underpinnings of argumentative strategies and procedures are
showcased in tandem with the precepts of classroom small-group dialogic interactions that are
meant to promote collaboration and argumentative rhetorical moves in a learning context.
Specificity of the genre of CR small-group discussions with a heavy use of critical thinking and
metalinguistic skills is pinpointed too.
Thirdly, the pragmatic conditionings of free flowing in-group interactions with a
Teacher’s role as an assistant yet unobtrusive coach are discussed to reveal the interface of
children’s cultural, conversational, rhetorical/textual and personal norms in their talk in action
and reflective essay writing. Theoretical and empirical studies are analyzed for their contribution
to the knowledge and understanding of pragmatic and socio-linguistic variation in the language
use and the mesh with metadiscursive uses by young adolescent speakers and writers (as
opposed to the adults), bearing in mind their generic, cultural and ethnic diversity.
The review is concluded with the consideration of documented and potential benefits of
some metadiscoursal features used by children or young adolescents for the achievement of their
communicative goals in writing and speaking, various domains of knowledge, as well as
cognitive and socio-cultural development. Thus, the precise study goals are distinctly formulated
at the end of Chapter 2, the Literature Review.
In Chapter 3, the lion’s share of the methods section is devoted to the exposition and
explication of the abstract constructs that influenced the practical measuring instrument adopted
in the coding scheme for qualitative and quantitative results to be obtained from two corpora
never before analyzed together. Therefore, the pertinent conceptualizations that went into
formulation of the multi-scale taxonomy are offered there and supported with extant theories or
similar studies and examples or illustrations of corresponding language uses.
In Chapter 4, since the study applied both qualitative and quantitative methods of analysis,
both bracketing and more fine-grained metadiscourse scales, micro-genetic and cross-sectional
approach to the data coming from both written and spoken corpora, each of which is given due
yet differential consideration in the results chapter.
Due to the fact that the study applies mixed methodologies, the results chapter is divided
into Quantitative Results and Qualitative Results sub-sections of Chapter 4. In the Quantitative
part, the results from the analysis of metadiscourse in both corpora separately and in comparison
to one another, as well as according to the sociolinguistic groupings are presented with pertinent
plots and tables. The Tables have been embedded for greater ease of reading. In the Qualitative
Results section, the report of the findings from this study is presented in a table that delimits the
complexity of the patterns that hold between forms in the language and their abstract functioning
(Form to Functions Table) in the minds of language users and in different contexts. Following
the explanatory form-to-function table, which is comparable to the ‘valency’ patterns of
metadiscourse grammar, illustrative episodes or excerpts from both essays and small group
discussions accompany, with elucidations of the technicalities of metadiscursive forms and
functioning and their contribution to the flow of meanings in spoken and written corpus. It is also
discussed there how the analyzed metadiscursive uses help or impede the flow of propositional
meaning, as well as how this impacts their persuasive discourses. Occasional references to the
available results yielded by the previous study that analyzed the influence of small-group CR
discussions on children’s metadiscourse in reflective writing (Latawiec et al., in preparation) are
made and integrated as dictated by the incorporation of the previous data pool to this study.
In Chapter 5, primarily the study’s quantitative results are discussed with a few
complementary issues that have not been discussed during the explication of the qualitative
results (i.e. in Chapter 4.2, qualitative sub-section). The discussion proceeds with regard to the
five study goals and addresses pertinent hypotheses, adopted schemes and three-partite
taxonomies as well as some rhetorical and socio-pragmatic theories or practices.
Eventually, in Chapter 6, the summary of the dissertation study briefly reviews the major
stages that went into the making of this project. The chapter then offers some insights that were
gained from the study. Pedagogic and empirical implications are offered for learning purposes
(communicative) and further research into metadiscourse in the written and/or oral modality,
with different approaches, study designs, instruments, samples and methods of analysis that may
be applicable to CR and beyond CR-context.
In essence, the study concludes in the same vein as it started, with the undercurrent of
attested theories and new perspectives or insights into language and discourse-metadiscourse
processing that is in ‘phatic communion’ with the context of the language in use (Malinowski,
1923). The context of the language in use – being inherent and inseparable from the sensemaking of communication – undergirds the pragmatic, rhetorical and socio-cultural aspects of
the interplay between metadiscourse and discourse proper, with metadiscourse rendering
primarily expressive meanings of the very language in use (cf. Schiffrin, 1980).
Chapter 2
Literature Review
In this Chapter various theories and empirical findings are discussed to establish the
common ground and the state of knowledge at the time of the study in the pertinent domains that
contribute to the theoretical underpinnings, adopted understandings of metadiscourse, as well as
the variety of factors that influence its functioning and interpretation.
Importantly, the unifying factor for the diverse topics from systemic-functional grammar,
corpora studies, argumentation, pragmatics, socio-linguistics, discourse and rhetorical analysis –
is the determination of the network of factors that need be taken into consideration in the
approach to the task of investigating data with children, from CR small group discussions, where
argumentative stratagems are at play, in the context of in-group relations and dynamics. The
review highlights the disparate threads that need to be woven into an instrument for measuring
metadiscourse in and across both corpora (which has been devised for the sake of this study, and
which is explicated in the methods chapter).
2.1 Language Functions and Metadiscourse
In order to capture the intricate complexity of metadiscourse that constantly interweaves
with the discourse proper, it is worthwhile considering the issue of the major functions of
language in use. In the systemic-functional theory of language, Halliday (1994) differentiated
between ideational, interpersonal and textual metafunctions of language, with metadiscourse
serving textual or interpersonal functions of language, as opposed to the ideational function (the
content or meaning; cf., Halliday, 1994; Halliday & Matthiessen, 1999). Halliday (1973) defined
the textual function as “an enabling function, that of creating a text”, “that enables the speaker to
organize what he is saying in such a way that it makes sense in the context and fulfills its
function as a message” (p. 66). Respectively, the interpersonal function was theorized to
incorporate “all that may be understood by the expression of our own personalities and personal
feelings on the one hand, and forms of interaction and social interplay with other participants in
the communication situation on the other hand” (Halliday, 1973, p. 66). One, however, has to
bear in mind that “metafunctions do not operate independently and discretely but are expressed
simultaneously in every utterance” (Hyland, 2005, p. 26).
It is often the case that authors conduct discourse on two levels; they mention the content
of the primary discourse, but embed it in metadiscourse or discourse about discourse rather than
the subject matter or issue at stake. Another view that was picked up by Dillon (1981), after
Williams (1981), mentions “writing about writing” and clarifies that using metadiscourse results
in calling attention to the act of discoursing itself. According to Crismore (1985) metadiscourse
is used “in any discourse where ideas are filtered through a concern with how the reader will take
them” (p. 61, emphasis mine).
Disparate formulations or definitions of metadiscourse generally lead to its essential
understanding as “discourse about the discourse” (e.g., Crismore, 1989; Vande Kopple, 1985;
Williams, 1981; 1989). In a straightforward manner Williamson (1981) denotes metadiscourse as
“…writing that guides the reader, as distinguished from writing that informs the reader (p. 47),
which echoes the Hallidayan (1973) distinction between textual (or otherwise text-organizing)
and inter-personal functions of metadiscourse as opposed to ideational or referential function of
discourse proper.
2.2 Language in Two Different Modalities – Spoken and Written
Research into spoken and written language (Chafe, 1982; Tannen, 1980, 1982) abounds
in theories and empirical data for the disparate qualities of the two. Overall, researchers
emphasize that spoken language fosters involvement in language or literature while written
language fosters integration in the language (e.g., Chafe, 1982). Spoken medium features the
quality of involvement by: greater fragmentation (e.g. stringing together of idea units without
connectives or coordinating conjunctions), speakers’ reference to him-/her-self (I, we, us) and
likewise – you, explicitness of speaker’s mental processes as revealed by such expressions as “I
thought… I”, “I had no idea how …” (“which are conspicuously absent in written language”
Chafe, 1982, p. 46).
Moreover, the spoken medium reveals speakers’ monitoring of information flow (e.g.,
well, I mean, you know) which is reflective of speakers’ greater preoccupation with
communication channel maintenance, including phatic sustenance of talk in accord with the
Phatic Maxim that can best be epitomized by “Avoid silence!” (Leech, 1983). Frequent emphasis
rendered by such particles as “just” or “really” add considerably to the involvement/engagement
facet of spoken language, similarly to devices such as hedges or vagueness fillers (e.g. and so on,
sort of, something like, or something). Direct quotes (which would turn into reported speech in
writing) also foster involvement via immediacy and authenticity of voice (including the fast pace
of delivery) thus re-instated in spoken language.
Concept of Dialogue
The notion of dialogue is especially pertinent to this study of Collaborative Reasoning
discussions in small groups ranging from 5-9 children. In the ‘common conception’ or ‘ordinary
sense,” dialogue involves “trying to understand and to be understood by others with different
views, and openness to the possibility of changing one’s views” (Craig, 2008, p. 2; Wierzbicka,
Generally accepted characteristics of dialogue include immediateness of presence,
emergent unanticipated consequences, recognition of strange otherness, collaborative orientation,
vulnerability, mutual implication, temporal flow, genuineness and authenticity (Cissna &
Anderson, 2002, pp. 9-11). All these qualities including the non-verbal dimensions of
interaction make the spoken medium more interpersonally engaging and embodied.
By contrast, written language is said to foster a kind of detachment as evidenced in the
use of passives and nominalizations (Chafe, 1982). Its tendency to pack more information into an
idea unit than in the spoken language tips the scale towards the greater integration of written
language. To further underscore the differences between oral and written language, as some
scholars argue, in spoken language “the meaning is in the context”, while in writing “the
meaning is in the text” (Olson, 1977). Others tend to view the difference as not a dichotomy but
rather a continuum of relative focus on interpersonal involvement versus message content
(Tannen, 1980) or a cline from the interactional to the propositional/ transactional (Cutting,
Oral and Written Task Differences
Understandably, the task of argumentative dialogizing and writing by CR children can be
rightly expected to differ on the above highlighted dimensions, as inherent in the prototypical
texts in each modality. The registers of oral and written CR texts or discourses will inevitably
vary with respect to conceptualized audience and related degree of interactiveness, purposes and
circumstances in which CR talk and writing are produced (cf. Biber, 2001). Thus, because in oral
discussions context is shared between interlocutors (and rather unshared during writing), it is
more likely that linguistic and so metadiscoursal forms will exploit more phatic (maintaining
communicative flow), interpersonally engaging (e.g., coercive) and intersubjective (shared
knowledge basis) functions and resort more to embodied forms, as often thoughts and ideas can
be expressed more easily in gestures, body language, pointing or prosody (intonation, raised
volume or pitch, and pauses) in oral discourse.
As a function of audience - oral language of CR students naturally becomes more adapted
to a specific teenage-group and sociocultural settings, which inevitably results in higher
informality of register and different structure of their discourses (cf. Beaugrande, 1982, 1983;
Horowitz & Samuels, 1987). The teenage audience of dialogs stays in sharp contrast to an
envisioned or expected audience of prospective readers of CR essays, most likely conceptualized
as formal (research team and/or Teacher), more detached or depersonalized, and possibly not
sharing the same knowledge basis as the peer group members and the CR discussants.
Hence, among other things, it has been hypothesized that more metadiscourse of formal
structure signaling type (e.g., binding, organizing, integrating or objective in tone ) would occur
in writing, while possibly more marking of personalized, intersubjective or vague pragmatic,
engaged and embodied meanings – in CR small-group oral discussions.
Transition from Oral to Written Expression
The studies into transition from oral to written expression (Olson, 1977; 1994) emphasize
the importance of reading which is “critical to the transformation of children’s language from
utterance to text” (p. 278). As Horowitz and Samuels (1987) argue, this is explicable in terms of
school curricula, which rely on the transmission of knowledge through written language which is
“formal, academic and planned” (p. 7). Researchers investigating oral–written transition (not
only in CR-context) have most often focused on the logical relations between propositions in
both modalities (Crothers, 1978; Kintsch & Van Dijk, 1978).
Conjunctions and Logical Relations
Importantly, conjunctions that show such logical relationships in texts not only signal
inter-clause integration (Millis & Just, 1994) or guide readers and help them integrate the texts
that they link (Guzman, 2004; Meyer, 1975) but also form an easily isolated cognitive-linguistic
category (Sanders & Nordman, 2000) that adds textual coherence on several levels (Goldman &
Murray, 1992). Because conjunctions occur in both oral and written language and reflect
cognitive and linguistic development they can reveal the degree of acquisition of the logical
implications of conjunctions (e.g., Geva, 2007), thus turning them into an important textual and
conceptual category.
Particularly for this study into fourth graders’ discourses and metadiscourses
(incorporating analysis of textual connectives), Geva’s (2007) study into school children’s use of
conjunctions (in oral language and reading) yields findings that add caution as to the
expectations with respect to the acquisition of connectives. Her results suggest that “welldeveloped understanding of adversativeness [e.g., signaled by ‘but’, ‘however’, ‘although’], as
expressing negation of an expected result, is not mastered completely even by Grade 5 children”
(p. 287). Geva further argues that the fact that children used ‘but’ in their oral language did not
mean that they mastered the linguistic constraints of its usage (as her reading tasks indicate).
This finding directly taps into Collaborative Reasoning practices and tasks of oral discussion and
reflective essay that lend themselves naturally to the presentation of opposing viewpoints,
wherein juxtaposing them will necessitate the cognitive-linguistic task of effectively
conceptualizing and signaling them with adversatives.
Language Organization by School, Community, Peers, CR
Finally, what needs to be emphasized perhaps is that “the nature of oral and written
language and the interplay between them is ever-shifting, and these changes both respond to and
create shifts in the individual and societal meanings of literacy” (Heath, 1982, Tannen, 1982, p.
XVI). Particularly relevant to the ever-shifting interplay of individual and societal oracy and
literacy are Ochs et al.’s (1996, 3-7) arguments about talk in interaction that both organizes and
is “organized by institutions, relationships and culturally specified environments” such as in our
case - school, age-group, ethnic community, and the Collaborative Reasoning paradigm.
Hence, children’s language in the CR context will differ not only as a function of their
communicative goals, and use of speech acts to prompt a preferred response (cf. Goodwin &
Duranti, 1992), but also the in-group context that goes beyond individual speakers and hearers
(e.g., Drew & Heritage, 1992). Essentially, the comparative study of metadiscourse intertwined
with discourse in spoken and written modalities, especially as explored for manifestations of
‘CR-kids’ thought-mediation, needs to account for the devices that foster modality-specific
qualities as well as the “interpersonal semantics” (Eggins & Slade, 1997), which involve social
evaluations of reality and expressions of emotional states, and judgments about the ethics and
morality of others — often in an “affect display”(Duranti & Goodwin, 1992) or “appraisal” (talk
with emotional coloring).
Heretofore it has been established that disparate theories of metadiscourse indicate the
unifying factor of carrying textual and interpersonal meanings rather than propositional (or
informative). Also, the two corpora, in which metadiscourse is being explored, differ on various
dimensions with the dichotomy or cline relationship from integration fostering (writing) to
interactivity or involvement in language (speaking). Research into children’s transition from the
speaking to the writing stage emphasizes logical relationships in text and signaling devices
whose skillful use reflects cognitive and linguistic developments. Conjunctions and their
organizing functions are therefore most pertinent to the analysis of language by CR-exposed
2.3 Metadiscourse and CR
The concept of metadiscourse with its inherent interplay between the language user, the
audience, and the socio-pragmatics of the language use seems especially attractive in the context
of young participants in small-group discussions featuring Collaborative Reasoning.
Collaborative Reasoning (CR) is an approach to discussion intended to be intellectually
stimulating and personally engaging (Anderson et al., 2001; Waggoner, Chinn, Yi, & Anderson,
1995), or otherwise “designed to encourage independent critical thinking, as well as question
fellow peers and sources” (Chen, 2009, September 24). Students read a story addressing a ‘big
question,’ then meet in small, heterogeneous groups to discuss the question. The students are
expected to take a position on the question, present reasons, back reasons with evidence and
further reasoning, challenge each other when they disagree, weigh reasons and evidence, and
change positions when warranted. In Collaborative Reasoning, students speak to each other
without raising their hands to bid for turns. Students operate the discussion as independently as
possible, with the teacher sitting outside the group. Previous research indicates that Collaborative
Reasoning has cognitive and social benefits (Li et al., 2007; Reznitskaya et al., 2008).
Interface of Rhetoric and Logic in Argumentation
At the interface of rhetoric and logic, children who were exposed to Collaborative
Reasoning write essays that have better developed arguments, counterarguments and rebuttals,
more consideration of opposing perspectives, more use of text evidence and better signaling of
disjunctive thought and inferencing moves (Latawiec et al, in preparation; Reznitskaya,
Anderson, Dong, Li, I.-H. Kim & S.-Y. Kim, 2008; I.-H. Kim et al, in press; Zhang, Anderson&
Nguyen-Jahiel, 2010). Due to many educational gains, CR gained recognition in the US public
educational system, as Public School Review argues, “through this type of pro-active learning,
experts anticipate that collaborative learning can jumpstart student progress into modernity”
(Chen, 2009, September 24).
Social-Construction of Knowledge
Drawing on Vygotskean socio-genesis (1978), which posits that cognitive growth is
“more likely when one is required to explain, elaborate, or defend one’s position to others, as
well as to oneself: striving for an explanation often makes a learner integrate and elaborate
knowledge in new ways” (p. 158), CR offers the epistemic, ethical, rhetorical, and pragmatic
context to arguments that children get engaged in and “think-out” in oral or written forms
(Hample, 1985). Also, CR operates in line with the idea of social-constructivism (Jadallah et al.,
2011, in press; Le Fevre,1987) that emphasizes social construction of knowledge, as it is in the
forum of the classroom mini-society that children’s ideas are conceived, and their minds and
personalities are shaped (social construction of knowledge).
Classroom Discourse Technique
Additionally, CR group discussions while using functional argumentative moves modify
typical patterns of classroom discourse, the very same conventional discourse that Kim,
Anderson, Nguyen-Jahiel and Archodiou (2007) dubbed as “mind numbing” (p.337). Thus
Collaborative Reasoning discourse replaces the mind numbing “classroom bred discourse”
patterns with “those having more immediate and natural extension to the real world” and such, as
Scardamalia and Bereiter (1994) further argue, that may be lying at the heart of superior
education which fosters transformational thought in “knowledge-building discourse […]
whereby ideas are conceived, responded to, re-framed, and set in historical context” (p. 266).
Involvement Semantics
The CR small-group involvement semantics also hinges on talk with “assessment activity”
(Goodwin & Goodwin, 1992), which involves a group of “participants producing assessment
actions and co-participating in the evaluative loading, matching the affect display and making an
alignment toward words that is congruent with the assessment” (Cutting, 2002, p. 125).
Critical Thinking
In effect, critical thinking is inherent and emphasized in the dialogic and dialectical
practices in Collaborative Reasoning (CR), which helps to equip young adolescents with the
regiment skills needed in their further education and everyday and professional demands of the
adult world. During important evaluative decision-making and weighing of values,
epistemologies and belief systems within their ‘community of practice” (Wenger, 1998),
Collaborative Reasoning-participants are given opportunities for practicing how to engage
productively or act out/ speak out in adult-like reality (with its government, economy, society,
ecosystem, etc.), the reality that is not to be trivialized unlike often their adolescent reality (cf.
Eckert, 2004).
The genre of CR discourse can be characterized by specific recruitment of metalinguistic
abilities to analyze form and content of language especially featured in the debriefing or
discussion-assessment sessions. Metalinguistic awareness involves "conscious reflection on,
analysis of, or intentional control over various aspects of language--phonology, semantics,
morphosyntax, discourse, pragmatics--outside the normal, unconscious processes of production
or comprehension" (Karmiloff-Smith, Grant, Sims, Jones, & Cuckle, 1996, p. 198). The concepts
of language analysis and control are central to some of the metalinguistic models proposed in the
literature (e.g., Bialystok, 1986; Bialystok & Ryan, 1985). According to Bialystok (1986), there
are two determinants of metalinguistic task difficulty: the level of analysis (the linguistic
complexity of the materials to be analyzed) and level of control (the need to go beyond or
disregard a salient characteristic of the stimuli to focus on another aspect of the message).
Since in CR-debriefing children are encouraged to analyze both argumentative rhetorical
moves/stratagems and social-participatory actions, metalinguistic demands of the debriefing task
of critical evaluation are levied at a high level. The task thus involves not only the adequate
linguistic repertoire that would befit both levels of analysis (argumentative and participatory) but
also metacognitive monitoring and control over and during their speech productions.
In essence, the CR genre necessitates greater metalinguistic awareness (of form and
content) in the dual task of talking the talk and walking the walk (by reference to the literature
perpetuating question “they do the talk, but do they walk the walk?”). Therefore, the heavily
metalinguisitc aspect of the CR in-group talk and meta-communication needs to be incorporated
in the adopted meta-discourse taxonomy and analysis scheme.
Literary Skills
Importantly, literature plays an instrumental role in Collaborative Reasoning – both as a
prompt for CR discussions and an input for students’ written discourse. Literature discussions
offer “an environment in which students and teachers can collaboratively construct meaning”
(Almasi, 1995, p. 314). As documented, peer discussions of literary texts benefit students in
cognitive, social and affective manner as well as provide valuable learning opportunities and
help them better understand what they read (Almasi, O’Flahaven & Arya, 2001). Moreover,
“quality talk about text [or literature]” promotes thinking, learning and language (Wilkinson,
Soter, Murphy & Li, 2009). The selected literature for Collaborative Reasoning discussions also
features inherent controversy that lends itself to socially-involved talk.
Engagement in an Epistemic Mode
It is thanks to the Collaborative Reasoning strivings to make students active leaders in
their own learning communities, and its far evolvement from the archaic preconceptions of a
classroom from the past (cf. Chen, 2009, September 24) that the Collaborative Reasoning may
offer an almost remedial set of “21st century” strategies (cf. “21st Century” skills: Not new, but a
worthy challenge” by Rotherham & Willingham, 2010) for the development in the fledgling
fourth graders’ cognitive skills and academic achievements as referenced by NAEP.
Specifically, in light of the NAEP assessments (Perie et al., 2005) which for the year
2005 revealed that less than 7% of fourth graders were able to “judge texts critically…and
explain their judgments… make generalizations about point of a story and extend its meaning by
integrating personal experience and other readings” (U.S. Department of Education, National
Assessment Governing Board, 2006, p. 24), the need for classroom practices that promote
students’ high-level comprehension skills becomes evident. Such practices require engagement
with text in an “epistemic mode”, which demands knowledge of the topic (“what”) and
knowledge about how to think about the topic (“how”) as well as metacognitive knowledge —
the ability to reflect on one’s thinking (Murphy, Wilkinson, & Soter, 2008, p. 744).
It stands to reason then that Collaborative Reasoning approach, by hinging on the idea of
reasoned argumentation that specifically encourages students to use reasoned discourse while
choosing between alternative viewpoints, promotes high-level comprehension skills, critical
thinking and problem solving of the read-about and collaboratively discussed issues. Indeed CR
may be viewed as the “next public school trend” (Chen, 2009, September 24) by offering a boost
to the alarming state of literacy, thinking and reasoning in the US education (as NAEP report
reveals, see above). The CR characteristics that have been discussed thus far range from highlevel comprehension skills, metalinguistic skills that combine talk with assessment, literacy and
oracy development especially in the argumentation and rhetorical domain, or socially and
cognitively motivated discourse practices that reach students on the personal level while
engaging them in the simulated real-world activities in the educational context.
2.4 Metadiscourse and Argument Schema Theory in Learning to Think Well
“Learning to Think Well” chapter by Reznitskaya, Anderson, Dong, Li et al. (2008), after
Vygotsky (1962), puts forward a seminal argument that “thought is not merely expressed in
words; it comes into existence through them,”, which essentially legitimizes the discourse and
metadiscourse analysis for the expression of thought-in-action in small group argumentative
conversations and reflective writing. The use of argumentative dialogue to improve thinking and
reasoning in group interactions and in different contexts has been both theoretically and
empirically explored (Anderson et al., 2001; Bakhtin, 1981; Billings & Fitzgerald, 2002; Chinn
& Anderson, 1998; Wilkinson, I. A. G., Soter, Murphy, & Li, 2009). It has, however, hardly ever
been investigated in tandem with metadiscourse especially in young adolescent talk or writing,
except for rare studies by Almasi and O’Flahaven (2001) in the US.
Effective Argumentation Moves
As research in effective argumentation and reasoning shows (Guillem, 2009),
argumentation generally should incorporate at least the rudimentary form of an argument with its
two indispensible components: justification for how a position was taken and the anticipation of
a possible criticism to come. Importantly, during an open group discussion about the posed
issue/question, children’s ideas get shaped, are often challenged and swayed by the many voices
of other participants (cf. Bakhtin’s heteroglossia). It is then when their reasons, supportive
evidence, suggested solutions are continuously weighed and evaluated. This complex process of
reasoning seems to get amalgamated in their written arguments’ formulations and justifications
that follow their conversations.
Collaborative Reasoning promotes the use of ‘argument stratagems’, i.e., functional
rhetorical reasoning moves over the literary input (Anderson, Chinn, Chang, Waggoner, & Yi,
1997), which show corroborating evidence for the relationship between peer-led discussions
(over literature) and improved argumentation both in oral and written medium (e.g., Latawiec et
al., in preparation; Reznitskaya, et al., 2001; Reznitskaya et al., 2008).
2.4.1 Metadiscourse and Argumentation
It is posited here that though argumentation has not been considered as “inherently good
or bad” (e.g., Alexander et al. 2002, p. 796), yet it seems inherently positive or, otherwise,
conducive to overall enhancement and engagement in the language use. Namely, apart from
considering issues from various perspectives, it fosters motivation and ability to engage in ‘issuerelevant thinking’, especially when conceived as or conflated with persuasion (e.g., Petty &
Cacioppo, 1979, 1980, 1981). Its inherent controversy aspect turns the communication into
psychological ‘stirring the waters’ or ‘unsettlement’ acts that force students to pause and think
critically as to what ideas need to be brought to a discussion or other communicative situations to
achieve the intended communicative goals. By metaphorical extension, the inherent controversy
basis make arguers think about what ammunition needs to be gathered to win the “duel on
words”, thoughts and oftentimes – on emotions. An array of empirical studies in children’s
argumentation documents its diverse cognitive benefits (Anderson, Chinn, Chang, Waggoner &
Yi, 1997; Billig, 1987; Knudson, 1992; dyadic interaction in reasoning in Kuhn, Shaw & Felton,
1997; or collaborative discursive argumentation in Nussbaum, 2008).
Metadiscourse Prevalence in Argumentation
Interestingly, metadiscourse is especially prevalent in argumentation, which results from
the fact that while arguing cases for or against a position “authors refer quite frequently to the
state of the argument, to the reader’s understanding of it, or the author’s understanding of his
own argument” (Crismore, 1985, p. 61). For Schiffrin (1980) “explanations and challenging
evaluations form an argument” where “in this conversational and interactional context metatalk
functions as an evaluative bracket” (p. 219). And so, metatalk can serve as an organizational
bracket when it functions on a referential, informational plane, and as an evaluative bracket when
it functions on an expressive, symbolic plane (Schiffrin, 1980, p. 231).
Academic Discourses
Studies of metadiscourse in academic discourses that involve argumentation usually at
the university level (e.g. Bondi, 2005; Crismore, 1989; Maurannen, 2001) suggest that
metadiscourse may illuminate problematizing of the events or issues raised in argumentation. For
example, Bondi (2005) argues that in dialectical models of arguing by balancing different
opinions, metadiscursive practices contribute to “claiming significance and credibility” by
‘problematizing’ (signaling and showing the novelty of an issue), ‘claiming significance’
(relating the claim to debate within the discourse community) and ‘signaling stance’ (by
highlighting the incoherence in evaluation of results / data/ conclusions) (p. 15). Crismore’s
(1989) findings suggest that metadiscourse (especially in written argumentative texts) promotes
critical thinking as readers formulate their opinions in comparison to the author’s and
metacognitively control and “follow the author’s indications throughout the text” (after
Crawford-Camiciottoli, 2005, p. 87).
Clearly, oral and written argumentation lends itself naturally to the studies of
metadiscourse for the reflections of “thought-mediation” in the language use, and in a composing
process (Vygotsky, 1986).
With CR paradigm promulgating argumentative stratagems (rhetorical moves) where
issues are argued for and against and constantly evaluated, metadiscourse is especially prevalent.
In argumentation both speakers and writers refer to the state of argument, to the reader’s
understanding of it, or the author’s understanding of his own argument, which requires both
rhetorical, interpersonal and meta-linguistic skills.
2.5 Metadiscourse and Pragmatics
Pragmatics has been commonly viewed as the study of language in use (Crystal, 1997; or
Mey, 2001) forefronting the incorporation of context factors in discourse (Levinson, 1983). The
context factors can include, for instance, the physical setting of the discourse act, and
relationship between the participants (Brown & Levinson, 1987) with regard to such relationship
factors as relative power, social distance, in-group membership, degree of imposition, and also
the participant’s state of shared knowledge about the topic of discourse and the social norms and
Politeness and Public Face
It is worthwhile mentioning here that often in pragmatics following Brown and
Levinson’s (1978b) understanding and model of politeness, power is considered as “an
asymmetric social dimension of relative power” (p. 82), as in interaction the speaker can be
relatively more or less powerful than the addressee. The notion of power is related to the concept
of face, i.e. “public self-image that every member wants to claim for himself” (p. 62). The face
can be either negative or positive – negative face involves the desire for freedom of action and
freedom from imposition, whereas positive face – need for approval. In the light of face and
power explanations, speakers’ strategies can show positive politeness and be damaging to their
own negative face (encroaching on freedom from imposition) or in negative politeness – the
opposite holds, i.e. speaker’s own positive face is threatened.
Age-Group, Community, CR-group Norms
Therefore, pragmatics with its deliberation stretching beyond linguistic form and concerning
itself with situated language functions seems perfectly suited to the analysis of metadiscourse in
Collaborative Reasoning small-group discussions and writing, which are heavily situated in the
in-group relationships between children with individual norms and rules of their ethnic
community, classroom conversational interaction and CR-guidelines, shared topical knowledge
about the short-story being discussed as well as knowledge about their communicative goals and
The social structure of small groups participating in CR discussions is constrained by the
social patterns that the youngsters invent for themselves while being driven by peer group
conformity and attempts to avoid social stigma (cf. Eckert, 1988, 1997; Stenström, Andersen &
Hasund, 2002). These sociological factors, apart from the ‘big four’ – age, gender, social class
and ethnicity, are often manifested in the language use and so most likely in metadiscourse, too.
Communicative Competence and Understanding of Speech Acts
It is noteworthy that pragmatic knowledge/competence is represented in all major models
of communicative competence. What underlies the notion of communicative competence is the
communicative ability, which incorporates strategic competence (Celce-Murcia, Dörnyei, &
Thurrell, 1995) or interactional competence subsuming non-verbal, conversational, discursive
competence in Markee’s (2000) understanding, and socio-cultural competence as well as
linguistic competence and some form of actional competence; the latter one can best be
understood as the knowledge required to understand the “communicative intent [revealed] by
performing and interpreting speech acts” (Celce-Murcia et al., 1995, p. 9).
Pragma-linguistic and Socio-linguistic Skills
Within pragmatics, two types of knowledge of socio-pragmatics and pragma-linguistics go
hand in hand, and it is best when they are mapped onto one another (Roever, 2009). After Leech
(1983) sociopragmatics encompasses knowledge of the social rules of language use, including
concepts of appropriateness, the meaning of situational and interlocutor factors, and social
conventions and taboos, while pragmalinguistics topicalizes the linguistic forms as tools for
implementing speech intentions. As Reover (2009) illustrates “if a language user has control of
pragmalinguistic tools without awareness of sociopragmatic rules of usage, she or he might
produce well-formed sentences which are so non-conventional that they are incomprehensible or
have disastrous consequences at the relationship level’ (pp. 560-561). Likewise, if language
users have socio-pragmatic knowledge (e.g., of polite requesting) and yet no or poor control over
linguistic tools (e.g., insufficient knowledge of modal or interrogative grammatical structures or
formulaic expressions) their communication will also suffer from being incomprehensible and
often missing the intended communicative goals (not reaching the targeted communicative
Adjustment to others’ Speech - Accommodation theory
Appeals based on group membership are generally expressed using in-group (identity)
markers (Brown & Levinson, 1978), which is in line with the Accommodation Theory (Giles,
1979) that “individuals shift their speech styles to become more like that of those with whom
they are interacting” (p. 46). The accommodation may involve accent convergence strategy:
upward – to a more refined and standard speech perceived “as relatively higher in terms of
accent prestige than [the speaker’s] idiolect” (Giles, 1973, p. 90), or downward – to a less
standard speech of the addressee, or it may involve accent divergence – a move away from the
less standard speech of the addressee (often exaggerating pronunciation differences to show
speaker’s dissociation from the addressee). In the context of fourth-graders’ small-group
discussions, instances of vagueness (or intersubjective assumption of common ground), adding
expressions, for example “and something/or anything”, kinda”, or “y’know”, especially when
used by eloquent speakers reveal their in-group discourse orientation and a socio-pragmatic
convergence (so a trend from more refined to more regional/ less refined) towards the less
articulate and less refined-language users in the group. Moreover, such instances signal implicit
invitations of solidarity which in pragmatics represent a strategy of positive politeness (Brown &
Levinson, 1987; Yule, 1996), or intersubjectivity marking that implicitly assume the common
ground thus in fact “saying less by communicating there is more” (Overstreet & Yule, 1997).
On the other hand, occurrences of question-tagging that uses standard grammar auxiliary
forms may reveal a fourth grader has good pragma-linguistic control over the language but at the
same time may be violating the community standards of casual talk, especially in a
predominantly vernacular English group. Instances of adjusting one’s speech to a more regional
‘idiolect’ (rather than standard variation), or using non-standard double negation by eloquent and
cognitively-high speakers may be a psycholinguistic as well as a socio-pragmatic strategy
establishing the common ground between the subjects of a conversation, especially when
interlocutors use vernacular dialect. Stenström et al.’s (2002) argument that “teenagers are far
from ignorant as to the importance of the relation between social features and language features”
(p. 17) may thus be underscored by a corpus linguistic study of early adolescents’ oral and
written language use as proposed here.
Theories that contribute to the study’s approach and analysis of CR talk and writing
tackle politeness and power principles effective in face-to-face confrontations, communicative
competence and ability to interpret and produce speech acts, repertoire of pragmatically
motivated linguistic skills that can be adjusted not only to the communicative goals but also
contextual issues including the hearer’s socio-linguistic characteristics. Adherence to cooperative
principles, politeness and accommodation theories are guiding factors and are intuitively applied
by native speakers including children and young adolescents in their talks.
2.6 Metadiscourse and a Sociolinguistic Variation
Sociolinguistic analysis of language variation between genders in the spoken domain
posited various differences between “men talk” and “women talk”, for example Lakoff’s (1975)
claim about ‘powerless’ language and place of women (in the society). Fewer studies
documented differences between teenage and children’s talk across the sexes, especially those
that would feed into metadiscourse analysis.
Vague Words as Exponents of Power
Stenström et al. (2002) in their study of teenage talk in the UK explored, among others,
the use of vague words, which were considered to be “exponents of power” by such
sociolinguists as Channell (1994), and as such they were meant to explain the more frequent use
of vague words by girls, who were seen as less powerful than boys (as found by Coates &
Cameron, 1988 or posited by Lakoff’s theory of powerless features in women’s language,
mentioned above). Stenström et al.’s (2002) findings in the corpus of COLT, i.e. the Bergen
Corpus of London Teenage Talk (approx. half a million of words gathered across London
boroughs in1993) suggest there are not many differences noted in the frequency of vague words
in relation to gender and school borough.
Vague Words as Intersubjectivity Markers
The list of vague words as compiled by Channell (1994) and including some fifty plus
items (e.g., “a bit”, “loads of”, “and crap”, “and so on”, etc., after Stenström et al., 2002, p. 89)
when reduced and adapted by Biber (1999) and applied to spoken discourse across two COLT
locations by Stenström and colleagues revealed that if only the same vague words were included,
boys use more of them than girls. Nonetheless, girls were found to use somewhat more
frequently the so called ‘general extenders’ (Biber, 1999) or ‘intersubjectivity markers’
(Overstreet & Yule, 1997) such as “and all that”, “and stuff”, “or something” – the latter one
constituting 40% of general extenders’ uses. In girls’ conversations “like” was found to
predominate as the most frequent vague word (Stenström et al., 2002, p. 92) unlike in the study
by Blyth, Recktenwald and Wang (1990), where like was used more by men than by women,
even though their attitudinal survey revealed that informants associate the use of “like” with
middle-class teenage girls, or otherwise their connotations can be best summed up by the most
frequent epithet of a ‘Valley Girl’, an American stereotype with social and regional connotations
(p. 224; cf. Andersen, 1998; 2001).
Different Uses of Emphasis Between Genders
Interestingly, analysis of intensifiers (which in the adopted metadiscourse model would
fall under category of ‘emphatics’) in the same corpus of COLT showed their more frequent use
by girls than boys (esp. of “really” as opposed to boys’ more frequent “very”), which thus
corroborates the findings of Holmes (1995) about adult women who used significantly more
intensifiers than men. However, boys were found to use the superlative forms of intensifiers like
“extremely angry”, “absolutely stupid” or taboo words “fucking weird” (Stenström et al., 2002, p.
142). Within the same British group of teenagers an interesting phenomenon was observed that
boys were especially keen on using “enough” in its switched role from a pre-modifier to postmodifying intensifier, e.g. “It’s enough bad”, or “My drawing’s enough crap” (p. 145).
Additionally, comparison of teenage and adult uses of intensifiers as explored by Paradis
(2000) showed that teenagers use twice less frequently adjective intensifiers. Paradis (2000, p.
154) explains this lesser use by teenagers’ choosing other means for reinforcement such as swear
words (e.g. “bloody”, “fucking”) or other emphasizers (e.g. “really”). Also, studies of
overlapping in turn-taking sequences by British teenagers and adults revealed that teenage
conversations are much richer in overlaps and interruptions, while in the chat and the discussion
the amount of overlapping speech is roughly the same (Stenström, 2007, p. 121).
Male and Female Features in Talk
In the light of some sociolinguistic theories that women and men talk differently because
they are socialized in different sociolinguistic subcultures (cf. Coates, 1988; or 1998), thus
viewing sex differences in language in a sub-culture rather than strong or weak power approach,
Maltz and Borker (1982) investigated different ‘women and men features’ in talk. One of their
outcomes was an observation that in all –female groups minimal responses were meant to signal
‘I’m listening’, while in all-male groups - ‘I agree’. They claim that men and women differ in
talk because they internalized different norms and communicative competence for conversational
interaction which were acquired in single-sex peer groups. Thus communicative breakdowns in
heterogeneous interactions may be explained by women’s different interpretations of men’s rare
use of minimal response as lack of attention, or male’s realizations that women’s responses do
not mean agreement (after Coates, 1988, pp. 69-70). Minimal responses may be subsumed under
talk organizing metadiscourse or fillers in this study, and so their differential distribution across
genders may be indicative of different socialization of the norm for responsiveness and verbal
Tags and Interactional Devices in Teenage Talk
Another important group of discursive and metadiscursive items explored for differences
across gender were tags, i.e. interactional devices that are generally appended to a statement and
“serve to engage the hearer or invite his response in the form of confirmation, verification or
corroboration of a claim, [that] may express a tentative attitude on the part of a speaker, or […]
may be polite expressions or signals of the common ground between interlocutors” (Stenström et
al., 2002, p. 167). The appended expressions like “yeah”, “okay”, “right”, “eh” (conventional
‘irony marker’, cf. Wilson & Sperber, 1993) and “innit” (isn’t it?) as a group serve a plethora of
functions (to a lesser or greater extent), e.g., epistemic, facilitative, softening, peremptory, or
concept-retrieval helping, response-urging, proposal-evaluating. Importantly, tags that add
pragmatic relevance to the semantic meaning of discourse proper and serve as metadiscourse by
talk-organizing, talk-evaluating, hearer-engaging commentary or attenuating hedges, etc., show
the usage peak at adolescence (dramatic drop after 17-19 years of age; Stenström et al., 2002,
In the same pool of COLT data, gender differences were only manifested in boys’
predominant use of “yeah”, while “innit” was more of a girls’ thing (besides, more of a British
thing too). With respect to ethnicity, distribution of “okay” showed that it is more common in
white speakers’ talk while “right” and “innit” - in the ethnic minority talk. As much as the
tagging studies are informative about certain trends, however they were not accompanied by
explanations or attitudinal surveys, which might have revealed respondents’ evaluations or even
stigmatization of certain uses as in the case of “like” (by ‘Valley Girls’ in Blyth et al., 1990, as
cited above).
Mimickry and Gestures
Distributions of mimickry, which also adds pragmatic meaning to the language in action
(here subsumed under gesture or paralinguistic metadiscourse category) in general reveal
prevalence among male adolescents (Stenström et al., 2002, p. 114). Interestingly an impressive
variation of para-linguistic features or mimickry (incl. mimicking male/female/teacher/sexy
woman’s voice, a monkey, Chinese or posh accent, a yobbo hooligan, drunken man, etc.) was
found in the COLT corpus to collocate with an explicit attribution, and so in metadiscourse
studies may be related to attitude marking (either towards epistemic truth-values or revealing
emotional orientations).
Ethnic Variation in Discourse
Studies into ethnic variation in discourse (as related to metadiscourse), however risky as
prone to confounding ethnicity with race and socio-economic status (e.g., Smitherman, 1981;
Wolfram, 2004), reveal among other things that African-American children’s discourse patterns
reflect a topic-association style that consists of a series of “implicitly associated personal
anecdotes” (Michaels, 1987, p. 429) rather than topic-centered style that was attributed to
European-American children in the same study into first graders’ narratives during sharing time.
Moreover, African American children’s topic shifts were found to be signaled prosodically and
also, as Garcia (1992) puts it, “to be difficult for the teacher to follow” (p. 61).
Incidentally, some of the characteristic features of African American English such as the
multiple negations (along negative inversion, formation of embedded yes/no questions, or verbal
markers like invariant “be” for habitual/recurrent acitivities, e.g. “Sometimes they be sitting…”
in Green, 2004, p. 80) from the perspective of metadiscourse functions may be considered as the
means of emphasis-adding by repetition of linguistic forms (or whole syntactic structures). For
clarification, emphatics or emphasizers contribute to “the expression of modality or stance” by
“strengthen[ing] the illocutionary point of the utterance” (Bondi, 2008, p. 39).
Michaels’ (1981) findings when paired with McCabe’s (1997) observations that African
American children “plot their numerous sequences of events within the context of the individual
experiences combined” (p. 460, citing Rodino et al., 1991) might be partly linked to multiple
hypothetical scenarios or deductive reasoning structures revealed in metadiscourse of reflective
essays as observed among non-CR exposed children by Latawiec et al. (in preparation).
Similarly, most frequent linking devices of “ands” and “thens” (McCarthey, 2002) for
connection of ideas and events show more semblance of the oral than written style (incidentally,
the additive conjunctions which are among textual metadiscourse markers were also found to be
most frequent among non-CR exposed children by Latawiec et al, in preparation).
Garcia’s comment (1992; quoted above) reveals the need for caution and perhaps
ethnographic deliberation when approaching and interpreting the language uses and styles of
different ethnic groups, which having been learned at home are the so called primary discourses
as compared to the classroom discourses learned at school (cf. Gee, 1990). This observation
about primary discourse is especially potent in the context of an African American variety of
English where the sociolect attributed values manifested in language echo Smitherman’s
thinking “[w]hen you lambast the home language that kids bring to school, you aint just dissin
dem, you talking bout they mommas” (after Wheeler & Swords, 2004, p. 472).
In the attempt to consider socio-linguistic variation in CR participating children or young
adolescents, various perspectives were reviewed - on differences between women’s and men’s
talk, talk of adults and children or young adolescents, languages of different ethnic groups, and
important devices observed in the adolescent language uses – such as vague words, tagging,
gestures, or mimicry.
2.7 Metadiscourse and Talk in Action
In line with Critical Discourse Analysis (CDA), both the personal and the shared aspects
of cognition are believed to play a role in the production and understanding of discourse (Van
Dijk & Kinstch, 1983; Van Dijk, 2001). In this understanding, both discourse and metadiscourse
are viewed as a function of socially shared attitudes and ideologies, norms, beliefs, values and
other forms of “social cognition” (Guillem, 2009). In the context of social construction of
knowledge in the course of interactive events (e.g., Tannen, 1982), in the confrontation of an
utterance with other minds (Givón, 2005) and the context, both discourse and metadiscourse are
seen as not only products but also as processes of reaching shared meanings (e.g., Bokus &
Garstka, 2009).
Hence socio-linguistic studies of metadiscourse, especially in the spoken mode of
communication as more “open” than writing about problems encountered en route (cf. Gilbert &
Mulkay, 1984), may allow for gaining better insight into individual cognitive processes, which
may be seen not only as individual but joint contextually-motivated efforts. The discussion task
as compared to the task of writing reflective essays lends itself to more differences of opinions,
objections, and queries. Nonetheless, both Collaborative Reasoning tasks of Big Question
discussions and written essays require statements of opinions of arguable nature and therefore
entail envisioning multiple stances and solutions to the raised controversial problem or question.
Metadiscourse in conversational argumentation with its frequent appeals to emotions,
speaker’s credibility and logic, expectations of talk evaluations (Bateson, 1972; Schiffrin, 1980),
an inherent dynamic flow (Chafe, 2001), prone to vagueness/performance fillers (Channell,
1994), operative on Cooperative and communicative principles (Grice,1989), teenage-group
community of practice context (Wenger, 1998) as well as ethnic and cultural norms (Stenström,
Andersen & Hasund, 2002) may offer an insight into complex thinking and learning processes
(Vygotskean thought mediation) of CR-exposed students’ - “navigating epistemic” (Heritage,
forthcoming) and socio-cultural landscapes (Cutting, 2010).
2.8 Metadiscourse Pros and Cons
An array of diverse studies provides evidence for the benefits of metadiscourse when it is
put to effective use. To name a few: a better and more coherent organization of discourse (three
aspects of coherence: global goal, local and thematic coherence, e.g., Goldman & Murray, 1992,
or Sanders and Noordman, 2000) as well as better management of both oral and written
discourse (Almasi, O’Flahaven, & Arya, 2001); metacognitive awareness (ibidem); better
understanding of text demands on readers and greater resourcefulness to express a stance
(Hyland, 2005); better learning from text (Meyer, 1975, Britton et al., 1982); advancing of
understanding, and adding to the rhetorical force of arguments (Reznitskaya, Anderson,
McNurlen, Nguyen-Jahiel, Archodidou, & Kim, 2001); better signaling of inference-ushering
moves by if-then structures and reader/listener-engagement devices in argumentation (Latawiec
et al., in preparation), rhetorical technique in fiction (Booth, 1961); enhancing social
performance and attitude (Crismore, 1985) or critical thinking and metacognitive control in
reading (Crismore, 1989).
Discussion of metadiscourse facilitation would not be complete without a hint about the
equivocal benefits of metadiscourse, as some researchers observe. Namely, metadiscourse as
metatalk in group/ classroom discussions was found to be helpful initially as groups negotiate
conversational conventions, but digressions and intrusions were seen as threatening
conversational coherence (Reichman 1990; Hobbs, 1990). Too much metatalk (e.g., initiated by
teacher’s overzealous attempts at scaffolding) causes disjuncture in discussion, and propagates
student reliance on the teacher to solve management related problems (Almasi et al., 2001). In
either oral or written discourse, metadiscourse has also been viewed as the source of “wordiness”
(Williams, 1981) or “content-less” discourse.
From the pedagogic perspective, therefore, metadiscourse may be quite difficult to fit into
content-based models of information processing, and so difficult to fit into curriculum-goals of
the content-based curriculum and especially in the era of No Child Left Behind observed
tendencies to teach for the test, which however has negative effects on literacy and oracy (e.g.,
Cawthon, 2007; Dressman, 2008; Edelsky, 2007; or McCarthey, 2008).
Hopefully, from the rhetorical, socio-pragmatic and pragma-linguistic perspective, the
young writer’s and speaker’s manipulations of their authorial voices as reflected in
metadiscourse may shed light on their formulation of beliefs, goals, and attitudes/stances. These
authorial voices naturally will vary as a function of the authors themselves (with the whole
psychological, epistemological and attitudinal baggage), the prospective readers or listeners
(their anticipated needs, expectations and responses) and the very context of a language use (cf.
Crismore, 1985; Hyland, 2005).
2.9 Metadiscourse Systems Contributory to this Study’s Construct
The analyses of metadiscourse are partly doomed to be product-oriented and with the line
of research mostly wrapped around what Crismore calls linguistic ‘devices’. Crismore and
colleagues (Crismore, Markkanen, & Steffensen, 1993) argue that writers – both professional
and non-professional - “convey their personality, credibility, considerateness of the reader, and
relationship to the subject matter and to readers by using certain devices in their texts. These
devices (which include words, phrases, main clauses, and even punctuation and typographical
marks) are referred to by terms such as ‘signaling devices’, ‘signposts’, ‘gambits’, ‘metatalk’,
and ‘meta-communicative markers’” (p. 40).
However, one cannot ignore the fact that these textual devices are only tools in the hands
of more or less skillful writers and speakers who thus communicate with them contingent upon
the context.
Wünderlich’s (1979) System
Chronologically, Wünderlich’s (1979) system sheds light on verb operators that bear
attitudinal stance in speaking and writing, and which he named “positional functors.” He
distinguishes 5 group/categories in which communicators can express their position: 1.epistemic
(know, think, suspect), 2.doxastic (believe), 3.ability/capacity (can, may, be able to),
4.motivational (wish, want, prefer), and 5. normative (ought, should, have to). His argument has
been carried forward and is influential, as some other researchers (e.g., Crismore, 1985) found
confirmatory evidence for verbal modal attitudinal marking and thus his system is indirectly an
inspiration to the present study too, as following Crismore’s and his views, attitude markers
incorporate verbal modals.
Crismore (1985) and Crismore et al.’s (1993) System
Crismore et al.’s system (1993; see Figure 1 for better clarity) considers illocution
markers (e.g. performative verbal expressions “I name/promise/thank you” and other speech acts
signals) as having intertextual rather than interpersonal function and so seems to confound major
types of metadiscourse when compared with Vande Kopple’s system (1997).
Figure 1. Crismore et al.’s (1993) Revised Taxonomy of Metadiscourse Taxonomies. Adopted
from Crismore et al., 1993; reprinted with permission of SAGE Publications.
Moreover, it is unclear why textual metadiscourse is divided into textual and interpretive
(cf. similar criticism in Hyland, 2005), especially as the interpretive attempt on the part of the
writers demonstrates their rather interpersonal (than textual) communication with the reader.
For the balance, Crismore‘s system (1985) incorporates the concept of Deontic modality
(concerned with the expression of obligation, prohibition and permission) in the attitudinal
metadiscourse category, which never before was so distinctly conceptualized. It is therefore for
this meritable inclusion of Deontic modality that Crismore’s system has been influential and
inspirational for the attitude marking taxonomy as conceived in this study.
Vande Kopple’s (1997) System
Vande Kopple’s (1997) system seems to take more of a pragmatic (apart from rhetorical
structural or syntactically-semantic, and emotionally-attitudinal) aspect of discourse into
consideration and has already been a point of reference for many studies in the domain (Hyland,
2005). As I align myself more with Vande Kopple (1997) rather than Vande Kopple (1985), let
me discuss briefly his taxonomy that subsumes such sub-classes of metadiscourse as: text
connectives, code glosses, illocution-markers, epistemological, modality- and attitude-markers,
evidentials, and commentary, which slightly differs from his earlier system of Vande Kopple
(1985), wherein in place of epistemological markers that subsume evidentials and modality
markers, he used to enumerate validity markers and narrators.
Vande Kopple’s (1997) system manages to account for the following functions of
metadiscourse: shows relationships between parts of texts via (1) text connectives, defines or
explains words, terms, phrases by (2) code glosses, make explicit what speech acts are being
performed at certain points in texts by means of (3) illocution markers (incl. boosters and
mitigators), indicate some stance on the part of the writer towards the epistemological status of
the referred-to-material via (4) epistemological markers, subsuming two (2) sub-types of markers:
(4a) modality markers – reflecting how committed we are to the truth of that material (using
elements from the system of epistemic rather than deontic modality, i.e. concerning duties and
obligations which thus form attitude imparting function), and (4b) evidentials (adopted by
Vande Kopple, 1997 after Chafe, 1986) that show what basis we have for referential material,
reveal what attitude or emotional orientation authors have towards referential material thanks to
(5) attitude markers, and provide commentary/directives/imperatives by addressing readers
directly in a (6) commentary (e.g. readers’ probable moods, views, hoped-for stance or simply
conversing with them).
Shiffrin’s (1980) System
Shiffrin’s (1980) research into meta-talk has been inspiring too, as she distinguishes
between two major met-communicative functions in talk – organizing and evaluative. These
meta-functions seem to bracket the discourse. Organizing bracket acts as discourse bracket that
initiates or terminates slots in discourse. Examples of organizing meta-talk are explanations or
justifications like “the reason is”, “the point is”, “for instance”, “I mean”, etc. Her evaluative
bracket is wrapped around opinion-making, which helps identify the speaker’s stance. In practice,
she lists examples of agreement or disagreement, positive or negative evaluations or insinuations
and similar uses. Her theory of organizing and evaluative bracket has been very convincing as it
seems that those brackets converge with the textual and interpersonal Hallidayan metafunctions
of metadiscourse and also align with Vande Kopple’s (1997) system that maintains the systemicfunctional textual differentiation between textual and interpersonal.
Hyland’s (2005) System
Hyland’s (2005) system of metadiscourse also springs out of the functional approach
which considers metadiscourse as the ways the writer refers to the text, to the writer or the reader,
His distinctions between interactive and interactional resources (following Thompson & Thetela,
1995) acknowledge the organizational and evaluative features of interaction in discourse. He also
includes stance and engagement features, thus further developing Hyland’s (2000) model. His
interactive resources utilize transition markers (conjunctions and adverbials), frame markers (text
boundaries signals), endophoric markers (references to other text parts), evidentials
(representations of ideas from another source), and code glosses (rephrasing, elaborations,
explanations). As for the interactional resources, these include hedges, boosters, attitude markers
(writer’s affective rather than epistemic attitudes to referential material), engagement markers
(explicit reader’s addresses, comparable to Vande Kopple’s commentary) and self mentions (1stperson pronouns and possessive adjectives).
It is noteworthy that the last category of self mention is the only resource not tapped by
Vande Kopple’s (1997) model. Due to an extra-fine grain that Vande Kopple’s system provides
especially with respect to illocutionary/speech acts and textual devices (global and local
coherence connectives) having more clear cut divisions and specializations, I find his system
more comprehensive and so adopted in this study; though not devoid of criticism perhaps (cf.
Hyland, 2005, p. 33), and lending itself to amendments as (Xu, 2001). Inspired by Hyland’s
evaluative category, I refined it to talk-evaluating (as especially manifested in the spoken
metadiscourse), while Crismore et al.’s (1993) idea of deontic modality influences the expansion
of attitudinal metadiscourse as posited in my model.
For the analysis of children or young adolescent metadiscourse especially in the spoken
modality, adjustments and expansions have been made to primarily Vande Kopple’s (1997)
model to account for the specific demands as well as constraints of the teenage language both in
reflective essays and in the communicative flow of in-group interactions (to be thoroughly
explained in the Methods section).
In the analysis of Collaborative Reasoning discussions, in which the communicative acts
of fourth-grade children are situated in an educational and social context, an attempt will be
made to take into account how communicative goals are achieved with the diverse linguistic
forms targeting different functions of metadiscourse (macro- and micro-functions, such as
“rapport–talk” (cf. Tannen, 1991) to negotiate relationships [macro], or “commentary” to engage
a reader/interlocutor in a kind of a dialogic interaction [micro]) by different groups of CRexposed children (e.g., boys and girls, of low and high academic aptitude, Afro- or EuropeanAmericans, from rural or urban backgrounds,) thus revealing tendencies to use certain
metadiscursive elements rather than other for particular goal accomplishment.
Pragmatic considerations may allow for observations and inferences as to how these
forms and functions demonstrate the children’s communicative success that hinges on the correct
assessments of the state of knowledge of one’s interlocutors (Cook, 1989) or recipients of
communicative acts (readers of reflective essays or interactants in CR-discussions), including the
knowledge of ‘good-discussion’ or argumentation. Rhetorical considerations and sociopragmatic combined will help determine how metadiscourse diverse uses help flow or impede
meaning in the discourses.
Thus, the inquiry into the fourth graders’ metadiscourse will assume primarily qualitative
and descriptive functions focusing on the community and CR-genre interactional preferences and
only then be followed by validating frequency counts and quantitative analysis to support
qualitative observations and comparisons (cf. Hyland, 2005).
2.10 Metadiscourse Study Goals
Since the focus of the study is on the relationship between metadiscourse in verbal smallgroup CR interactions in the process of co-constructing shared meanings, on the one hand, and
participant’s individual learning as externalized in written expression of thought, on the other,
the study necessitates a complex model of discourse. The current analysis of metadiscourse
posits the discourse model that operates not only on the ideational, textual, and interpersonal
planes of language (cf. systemic-functionalism - Halliday, 1994) but also on the performance and
action planes (cf. sociolinguistic model of Schiffrin, 1987) in the language use in the CR
pragmatic context (Levinson, 1983; Mey, 2001).
Overall, the purpose of the proposed study is to investigate how children use
metadiscourse and how functionally it interplays with discourse proper in their interaction with
peers in collaborative small-group discussions and reflective writing.
Specifically, the study aims to (1) analyze metadiscourse used in small group discussions
by CR-exposed children as well as (2) analyze the metadiscourse in their written essays. The
analysis of videotaped and transcribed discussions when paired with their essays will enable a
search for (3) any correspondences between children’s metadiscourse in two modalities and help
determine (4) how metadiscursive patterns are used by CR-children as a function of gender,
ethnicity, academic aptitudes or school location in small group discussions and essay writing,
and (5) how those metadiscursive patterns add to the flow of meaning in their oral and written
The proposed study will examine data from a major study entitled "Social Propagation of
Argument Stratagems.” This major study was conducted in fourth-grade classrooms in public
schools in central Illinois in 2001-2003. During a five-week period of Collaborative Reasoning
discussions, the fourth-grade children and teachers' oral discourses were videotaped and
transcribed, and children wrote persuasive essays at the end. Several analyses of the videos and
transcripts from the project have been completed, describing the emergence of child leadership
(Li et al., 2007), children’s response to teacher scaffolding (Jadallah et al., 2011, under review),
and children’s use of analogy (Lin et al., in press).
In years 2008-2009 metadiscourse in persuasive essays of CR-exposed and non-CR
exposed fourth-grade children (N=180) was analyzed (Latawiec et al., in preparation) and
revealed distinctive metadiscursive patterns of children participating in the intervention. Their
recurring metadiscursive patterns, especially of CR females and CR students with above average
MAT scores, indicated that CR students were more likely than comparable control students to
signal reasoning moves and to engage in explicit “If...then” inferencing that help to shape
argumentation in the characteristic moulds of Modus Ponens (If A, then B; A, therefore B) and
Modus Tollens (If A, then B; not B, therefore not A). Essays written by CR-exposed children
tended to contain more explicit and elaborated argumentative structure with clearer references to
belief sources, signaling of inductive reasoning, direct address to and engagement of readers via
commentary, and boosted illocutionary acts. Also, the CR writers’ attitude and emotional
orientation marking in the less self-centered manner indicate their trend towards more commongood-oriented authorial expression (e.g., “It is unfair/bad/not nice to the other kids”). The essays
of CR students had greater coherence than the essays of control students — due to logical
connectives that more explicitly showed text sequencing and global text binding. In contrast,
control students made more emotional appeals by means of emphatics and by means of
hypothetical scenarios, both self- and other-centered. The propositions in the essays of control
students were, however, more frequently linked with the simplest, least globally-coherent
additive conjunctions (cf. Goldman & Murray, 1992; Sanders & Nordman, 2000).
Further Steps:
To pursue the study goals which are two-fold in terms of methods, qualitative corpus
analysis - (1) for the richness of forms and functions in both modalities (study goals 1 & 2) consecutively leads to its quantitative comparative analysis with their written metadiscourse in
reflective essays (Latawiec et al., in preparation) (2) to establish metadiscursive correspondences
between the two modalities (study goal #3), as used by the same CR-exposed groups of children
yet (4) as function of gender, academic aptitude, school locality background, different ethnicities
(while taking into account their succinctness or prolixity in talk and writing), as well as (5) to
help determine and evaluate how metadiscourse adds to or impedes the flow of meaning in both
In the first stage, the analysis focuses on the video clips, their transcripts and field notes
from a subset comprising two discussions (discussion No. 5 and 8) from two classrooms x three
groups each from 2 different locations for primarily - (1) the identification of metadiscursive
forms, their density and functions in oral modality. Secondarily, (2) participants' reflective essays
that were analyzed for the usage of written metadiscourse in the previous study (Latawiec et al.,
in preparation, as cited above) are re-analyzed and re-coded for the occurrence of the newly
added supra-codes (the three major brackets of organizing, evaluating and intersubjectivity).
Then, the data from analyses of both corpora are (3) compared for the possible correspondences
both in the following statistical analyses and qualitative evaluations of their patterns and
functionality of use.
Chapter 3
3.1 Participants
This study springs from the previous study (Latawiec et al., in preparation) that analyzed
metadiscourse in persuasive essays of fourth-graders (N=180) from 3 different schools that were
matched in socio-economic status and ethnicity that participated in the major study Social
Propagation Study (2001-2003). The secondary data available for metadiscourse analysis in two
corpora – written (essays) and oral (discussions) were collected from 78 CR-exposed students or
otherwise 4 classes - each sub-divided into 3 heterogeneous conversational groups (red, yellow
and blue) being a representative cross-section of the class. Altogether 4 CR-exposed/
experimental classrooms and their 4 CR-involved teachers from 2 different locations in Illinois,
one rural and one urban, participated in the project. The schools were matched in SES.
In terms of ethnic background, the major study’s pool (Social Propagation study in 20022003), consisting of 6 CR-classes (and 6 non-CR), showed the following breakdown: 33%
African-Americans, 3% Latin-Americans, 1% Asian-Americans, and the remaining were
European-Americans. Likewise, the selected sample’s overall ethnic make-up (from the above
pool) is 32% African-Americans (N=22), 4.2% Latina (N=3), 1.4% (1) Asian-American and 62%
European-Americans (N=45). One of the analyzed schools (rural location) has solely EuropeanAmericans in CR-exposed groups, which means that all the other ethnicities come from the other
school (urban).
In terms of gender, originally 61 girls and 46 boys participated in Collaborative
Reasoning exposed groups, and so the written corpus of the data has been analyzed for this
whole pool. In this pilot study sample, 38 girls and 33 boys participated in both analyzed CR
discussions (6 students participated only in the later discussion, and 6 have no written essays);
these contribute to the oral corpus, which has been analyzed and compared across two equally
intriguing and controversial discussions (medial and later one in a series) in the first stage.
Participants’ scores varied across the Metropolitan Achievement Test, with the largest
amount of below median achievers in the urban school.
3.2 Procedures
An array of activities took place in the main project (2001-2003). The activities were
integrated into the regular reading instruction. At the onset of the project and prior to the
intervention, children were given a battery of individual tests, including Metropolitan
Achievement Test (MAT), to assess their reading, vocabulary and spatial reasoning. Moreover,
the children filled out a socio-metric questionnaire including peer-evaluations in terms of
openness to talk and participation in discussions. At the end of the project, children took two
more tests that assessed their reasoning and appropriation of argument stratagems. One of the
post-intervention measures, i.e., persuasive essays, constitutes the written corpus analyzed for the
occurrence of metadiscourse by fourth graders exposed to Collaborative Reasoning intervention.
At the beginning of the main project, the teachers who participated in the study were
given a half-day long workshop on conducting Collaborative Reasoning discussions. Then,
teachers from the experimental groups conducted 30 discussions in their classrooms - 3 groups
out of each classroom had 10 discussions each. Following the series of discussions the teachers
participated in a 30-minute interview, which was conducted by a research assistant at the end of
the project.
In the course of the project, the CR-exposed children read selected short stories and upon
reading of the literary material engaged in 10 Collaborative Reasoning small group discussions
about the issues raised in the stories; children from the contrastive/control groups did not receive
any Collaborative instruction. Children in the CR-exposed groups engaged in 2 discussions per
week for the period of 5 weeks. All discussions were video-taped and then transcribed. Field
notes were prepared by a research assistant about debriefing sessions that were held with
teachers after each day of Collaborative Reasoning intervention in their classes.
Thus, in terms of materials - the videotapes and their transcripts (incl. field-notes) of
small-group discussions are the focus of this analysis - to be further compared with reflective
essays that children wrote in response to the big question related to a controversial issue
prompted by the story.
It has to be emphasized that students did not receive any instruction in the written
organization of arguments, as the targeted Collaborative Reasoning modality was speaking in
small group, free-flowing, peer-led discussions, ideally with adherence to the suggested main
principles of “good discussion.” Specifically, the good discussion principles that were reviewed
by teachers at the beginning of the early discussions (especially), and debriefed at the end of
each discussion included both socio-pragmatic (participatory) as well as rhetorical argument
schema-fostering dimensions. For instance, children were urged to invite others to join in
conversation, respect others’ opinions, support opinions with information from the story or
personal evidence, provide supporting reasons for their position or otherwise justification, and
present counter-arguments as well as listen and respond to their peers’ counter-arguments. Thus,
as evidenced with the micro-strategies above, the CR discussants had been given opportunity to
develop a habit of good collaboration as well as collective reasoning that hinged on the
reciprocity of idea formulations and their pragmatic interplay in language.
However, for every small-group two ‘co-conspirators’ were selected and RA-instructed
about how to orient towards certain argumentative moves to later suggest and promote among
their peers in their small group discussion, thus instilling so called “social propagation” of
argumentative stratagems (e.g., Anderson et al., 2001; Reznitskaya et al., 2001). It may be
anticipated that their discourse and participation may be affected by their co-conspirators’ roles.
For the sake of the pilot study their names have been withheld to ensure better objectivity of
analysis and coding procedure.
The small-group discussions selected for this study (as any other CR- discussions over
the literary input) were based on two short stories that are wrapped around a complex
multidimensional controversy – on psychological, social, cognitive and moral dimensions. Those
2 stories – story No 5 and No 8- were ostensibly chosen in order to take place further in the
series – not too close to the initial 3 baseline discussions, while excluding discussion 10 which
was an outlier by introducing too many complex perspectives for a debate (about a nuclear
power plant).
The story number 5, titled Marco’s Vote (interchangeably titled Crystal’s Vote, NguyenJahiel, 1996) is about Marco and Crystal who are student members of an advisory committee in
their school. The committee has to decide whether they should replace the worn out 5th grade
math textbooks with new ones or buy a computer program that teaches mathematics. After
hearing the arguments for and against each side, Crystal and Marco have to decide how to vote,
and so the Big Question posed to students’ deliberation is: Should Marco/Crystal vote to buy
new textbooks or the new computer program? Story number 8 titled Stone Fox (Gardiner, 1980)
tells a story of little Willy who enters a dog sled race hoping he can win the prize in order to pay
back the taxes on his grandfather’s farm to avoid its being taken over. In the race the little Willy
is pitted against the best dog teams in the country, including the India Stone Fox, notorious
winner, who needs money as badly as Willy as he plans to use the prize money to buy back land
for his tribe which was deprived of it. Willy’s dog dies a few feet from the finish line – just when
it was ahead of everyone else, including Stone Fox. And so everyone wonders whether Stone
Fox, who follows at a short distance, will grasp the opportunity to win the race. The Big
Question for discussion epitomizes the dilemma by asking, Should Stone Fox let Willy win the
dog race?
Regarding literature prompt for writing, following 10 sessions of small-group discussions,
in a five week long CR-intervention, students (CR and non-CR-exposed, 2001-2003 data set)
read the Pinewood Derby (McNurlen, 1998) about Thomas, who cheated in the school model car
race, and his classmate, Jack, whom he confided with this troublesome fact while withholding
the information from other classmates and the Teacher. Students were then asked to write
persuasive essays in response to the question, Should Jack tell on Thomas? They were given 40
minutes to complete the task. The writing prompt read as follows:
In the next few pages, write whether or not you think Jack should tell on Thomas.
Remember: Do your best and write as much as you can. You can go back and re-read the story if
you like.
Studies into appropriation of argument stratagems in their analysis of written
argumentation (e.g., Kim, Anderson, Miller, Jeong et al, 2011) found what Reznitskaya et al.
(2001) dubbed “portable knowledge”, i.e. oral discussions’ facilitation in transfer of argument
schema to the written modality (which found confirmatory evidence on the basis of argument
structure analysis). This study in oral and written metadiscourse, which builds upon the major
study as well as the previous study in metadiscourse in reflective essays (Latawiec et al., in
preparation) which offered promising confirmatory evidence, hopes to find corroborating
evidence for the occurrence of portable knowledge. Hence, it is assumed that some evidence for
the similar transfer of such metadiscursive and discursive patterns that may contribute to
enhancement of children’s argumentation in speech and writing can be established as the result
of this study too.
3.3 Data Analysis
The discussions were transcribed using Transtool (Kumar & Miller, 2005) and then
coded using Windows Access - to be continued with Nvivo 8 (QSR, 2009). Primarily, the
qualitative dimension of the study first explored the links between linguistic forms and functions
contingent upon the social context of children’s conversations (which will be presented in
Chapter 4, Qualitative Results sub-section). Upon establishing specific functionality of the
identified forms, the latter contributed to the major categories that evolved into a complex
scheme (see adopted Taxonomy Figure 2, further in Chapter 3), and used for qualitative
evaluations and quantifying data analysis.
3.3.1 Speaking ‘turn’ defined
As befits conversational analysis with any grammar of turn-taking in mind (e.g., Ochs et
al., 1996) of the interpreter as well as interactants, the conceptualization of the basic
conversational unit of a turn is a fundamental step. Thus, bearing in mind diverse elements of
contextualization of the students’ discourse within their discourse community of fourth graders,
of a Teacher as the main authority (in the classroom), in the context of Collaborative Reasoning
“good-discussion” principles as well as the story context, the turn’s scope or delimitation has to
be complex enough to encompass the different dimensions of contextualization (cf. Cutting,
2000), and yet simple enough for the clarity and consistency of a metadiscursive coding scheme
and its reliability (including inter-rater reliability).
For this analysis, an individual speaker’s action accomplished by talking and para-talk
that was determinant in meaning and more or less complete, even if interrupted but continued in
the course of conversation, was considered a turn-in-interaction. This conceptualization of a turn
included syntactically diverse structures – from very brief nuclear structures of Subject + Verb,
or just Verb/Predicate or another part of speech rendering illocutionary/speech Acts (e.g., “right”,
“Ok”, “Yes/No” or more informal - “Yeap/Nope”; cf. Stenström, Andersen & Hasund, 2002) to
very complex ones with multiple subordination and stretching onto many lines of talk. To clarify,
longer strings of talk usually fashion many Turn Constructional Units or TCUs in Schegloff’s
(2007) understanding of a turn that consists of TCUs or otherwise “recognizable actions in
context” as building blocks (p. 4, emphasis added). Aborted expressions, incomprehensible
language, repetitions that did not expand meaning (resting upon socio-pragmatic theory of
meaning, as noted in Literature Review; cf. Fraser, 1998, or Schiffrin, 1987b) were not counted
among the metadiscursive codes unless some of them fulfilled a filler-like or placeholder
function and then were coded as Fillers.
In practice, this meant that for this study even incomplete clauses or fragments, including
overlapping speech and back channeling (brief utterances ensuring talk-progressivity, e.g.,
“yeah”, “right” in different phonetic realizations; cf. Yngve, 1970; or Fraser, 1990 for
interjections which stand alone as answers, e.g., “aha”, “yuk” or “because” – understood
differently as the function of falling/flat or rising intonation) were considered as turns since the
determinant criterion was interpretability and the link of form to function and vice versa.
Moreover, gestures that were meaningful in the talk-in-interaction exchanges and rendered
metadiscursive functions (and solely those) – be it inter-textual (text organizing) or interpersonal
- were also included.
3.3.2 Coding Scheme
Initial data-driven analysis of video-clips yielded links between an array of linguistic
forms and functions and resulted in the evolution of the coding scheme. In terms of the language
interaction observations, both physical (gestures, prosodic and paralinguistic behavior) and
socially based (interpretable within social norms) scheme (cf. Bakeman & Gottman, 1997)
assumed two-fold hierarchical structure, as dictated by the study goals, which on the one hand
aim at exploration of the density of metadiscursive forms and functions in Collaborative
Reasoning interactions, and on the other – identification of correspondences between the spoken
and written metadiscourse generated by CR-exposed students.
Hence, the coding scheme has been adjusted to two more levels, (1) one for comparison
between discussion and essays (less molecular), and (2) the other more molecular and inclusive
of embodied and interpersonal involvement categories (gestures and fillers) - for the comparison
platform between discussions.
It is believed that the adoption of the more fine-grained scheme that may provide more
detailed data may suggest something unanticipated to the author of this study, and perhaps “may
reveal something of interest to others too whose concerns may differ from ours” (Bakeman &
Gottman, 1997, p.25).
The need for more fine-grained taxonomy and adoption of the two coding schemes may
be motivated by the actual niche of the socio-linguistic dimension in studies of children’s
metadiscourse. Another more refined US study of the metatalk of children by Almasi,
O’Flahaven and Arya (2001) distinguished only between teacher- versus student-initiated
metatalk, and topic- vs. group-process-management metatalk (predominantly encouraging
others). Additionally, the multi-layered scheme of metadiscourse analysis has been hypothesized
to reveal micro-genetic differences in the patterns of use (as was the case of written
metadiscourse study by Latawiec et al., in preparation). The molecular scheme allows for
lumping of some categories together to trace their use in macro-functions fulfillment that may be
differently realized by different groups of interlocutors.
It is believed that the combinations of macro-functional and micro-functional categories
in the measuring instrument (taxonomy) may allow testing of some conversational analysts’ or
sociolinguists’ hypothesis, otherwise ungraspable or impossible to verify. Among such theories
(some of which are discussed in Chapter 2) is Tannen’s (1991) proposition that females and
males’ (adults, alas) speaking differs in overall/macro-functions, namely “while men do ‘reporttalk’ or ‘public’ speaking to negotiate status, inform and perform, women do ‘rapport-talk’ or
‘private’ speaking to negotiate relationships, interact and establish connections” (cited after
Cutting, 2000, pp. 110-111).
3.3.3 Adopted Taxonomy and its Mechanics
This section treats the taxonomy adopted for the 2 corpora and two methods of analyses –
quantitative (including micro-genetic and cross-sectional) and qualitative analysis. The
categories are described in great detail and illustrated with uses from both modalities. Pertinent
references are made to either empirical or theoretical studies that help justify or clarify the
proposed categories’ inclusions in the taxonomy or their functioning per se. The Coding scheme
and procedures are further described following the explication of all codes used from elementary
to broad bracketing level, to be followed by results from inter-rater reliability coding, to move on
to quantitative and qualitative results in the subsequent two sections of Chapter 4.
For the quantitative purposes the essays in Latawiec et al. study (in preparation) were
analyzed using the augmented taxonomy inspired by Vande Kopple (1997), Crismore et al.
(1993) or Schiffrin (1980). This study further adapted the taxonomy to accommodate the spoken
discourse of children (see Figure 2), which is being explicated in detail further on. Bearing in
mind the prospect of exploring oral-written correspondences, ostensibly the categories of
metatalk have been devised frugally.
I. TEXTUAL METADISCOURSE (shows how textual elements relate to one another)
1. Text Connectives
 Announcements of material
 Reminders of material
 Sequencers
 Topicalizers
 Logical~Temporal Connectives
o Additives
o Contrastives
o Causatives
o If- conditionals
2. Code glosses
 Definitions
 Explanations
II.INTERPERSONAL METADISCOURSE (conveys essentially interpersonal meanings)
Illocution Markers:
 Speech Acts
 Boosters
 Mitigators
Epistemology Markers
 Modality Markers:
o Hedges:
 Morphological
 Modal verbs
 Adverbs
 Lexical verbs
 Other cautious elements
o Emphatics:
 Forefronting
 Repetition
 Intonation (equivalent of Orthographic/punctuation)
 Personal belief
 Induction
 Sensory experience
 From someone else
 If/since …, then
 Deduction: Abstracting others in scenario
 Deduction: Abstracting oneself=myself in scenario
 Aposteriori/ Retroduction intended
Attitude Markers (explicit)
 Implicit Attitude Markers
 Deontic Modality Markers (forbidden, obligatory, permissible, possible, non-obligatory)
Perlocutionary Commentary:
 Commentary Vocative/Directive
 Commentary Interrogative
Non-Verbal (Gesture, Body Language)
Figure 2. Taxonomy of Metadiscourse categories based on (Latawiec et al., in preparation).
As noted above, the current taxonomy has been based on Latawiec et al. (in preparation)
and its changes have been necessitated by the specificity of the more embodied oral
communication and reasoning “genre” of CR conversation that incorporates peer-debriefing
and/or a teacher-reviewing stage with the focus on meta-communication and entails heavy
metalanguage use. Specifically, some categories assumed their spoken equivalency, for example,
Emphatics by Orthography-Punctuation has been replaced with Emphatics by Intonation, some
were further split – written Commentary category (NB, engaging the reader in a form of a
dialogue) now falls into Commentary Vocative (Vocative and Directive form of address) and
Commentary Interrogative (syntactically interrogative address/direct reference to a recipient),
and some were added: Gesture (other non-verbal/bodily-communicative codes that fulfill
intertextual or interpersonal functions of language) or Fillers (Placeholders or Stallers).
Included in the taxonomy, there are also supra-codes like two broad brackets of Text/Talk-Organizers (/text-/talk-tying, text-/talk-directing and redirecting, and ensuring talkprogressivity/-sequentiality) and Talk-Evaluatives (self or others’ talk assessment - epistemic and
attitudinal stance revealing), both of which fulfill over-arching functions of other more discrete
codes (lumping and double/ tandem coding), and so were excluded from the taxonomy table (i.e.
Figure 2), as well as a side-bracket of Pragmatic Intersubjectivity Vague Markers (establishing
shared meaning/co-conception of the world and revealing social closeness/ speaker’s solidarity
orientation = mostly via vague expressions). Metalanguage encompasses text/talk-reflexive
forms or otherwise a functional set of expressions that “focus on properties of the code per se”
(on ‘langue’) or “the language used in speech situation (‘on parole’)” (Schiffrin, 1987, p.3).
Metalanguage is used in combinations with other codes mostly, and in fact can be used as
embedded in organizing and evaluative categories.
3.3.4 Broad Bracketing and Building-Block Categories in the Coding Scheme
Metalanguage category has been assigned whenever a child referred to the language as a
code per se and/or its use in action. Metalanguage generally has ‘designata’ in aspects of
language and language use (cf. Weinrich, 1966). Metalanguage then mostly included lexical
items that deal with talk, speech, argumentation, participation in a discussion and the rules of the
Collaborative Reasoning dubbed otherwise as “good discussion principles” to which students
referred to – sometimes by mere mentioning their numbers (e.g., “we did well on number 1”).
Thus, metalinguistic items (mostly nouns) often co-occurred with talk evaluations. Especially the
debriefing sessions lend themselves naturally to metalinguistic practice on reflexive and
interpretive strategies. Metalanguage is used in texts too whenever similar lexical items as
quoted above are used, e.g. first problem, here are my reasons, etc.
Text-/Talk-Evaluating Bracket (T-E)
Evaluating also arose mostly during debriefing sessions, as this is when students are
encouraged to think about how they interacted (so participatory technique) and how good their
ideas and their expressions were in the group-reasoning (argumentation stratagems). Thus,
comments upon how they adhered to or deviated from the adopted conceptualization of the good
discussion as collaborators and arguers as well as according to their own individual sets of norms
for a heterogeneous group-talk in the school context, which befits limited range of vocabulary
and forms of expression (e.g., excluding vulgarity or some slang on the one hand, and obsolete,
too sophisticated lexicon or too refined ‘stiff-upper-lip’ register - on the other). Individual selfdeprecating or self-appraising as well as others-appraising/deprecating linguistic forms were as
Talk-Evaluators supra codes (yet subsuming attitudinal quantifiers so sometimes over-lapping
with Attitude Markers), oftentimes supported by citations (e.g., X said that….and I think it is a
good idea/point) or rendered by rising-questioning Intonation (then double coded with Emphatics
by Intonation). Following Schiffrin’s (1980) view of evaluating bracket functions as such that
show speaker’s stance whenever opinions are said, written Text-Evaluating brackets convey
same meanings and functions. Intensifications and mitigators of any sort (emphatics and hedges
illocutionary boosters or mitigators) are elementary bricks of the broad bracket otherwise best to
be viewed as a plane or dimension. For example, evaluating uses are mostly when opinions and
agreements or disagreements are ventured or their subsequent evaluations.
Text-/Talk-Organizing (T-O)
Talk-Organizing occurred whenever a child attempted to adhere to the Cooperative
principle (Grice, 1975) and provide a response that would be of a preferred type rather than its
lack. Moreover, Talk-Organizers functioned as propellers of the talk, directing and redirecting,
managing the traffic of talk by prompting sequentiality of turn-taking, or shifting to the main
topic or appointing prospective ‘turn-takers’ – individually or collaboratively (e.g., in a longish
sequence of nudging a quiet child). Thus, numerous linguistic forms and other codes too
contributed to this supra-function. Namely, performative Speech Acts and Directive /
Interrogative Commentaries or even some appended tags of the connective sort “because…” or
“so” (with a distinct elongated intonational rise) were used in their non-canonical function, i.e.
not as coordinating conjunctions (textual metadiscursive function falling under the category of
text connectives) but primarily as imposing on other subjects the role of ‘continuers’ of
Generally all categories that add to coherence building (globally binding, not only on the
surface level) contribute to Textually Organizing brackets, which according to Shiffrin (1980)
are usually used as discourse brackets that initiates or terminates slots in the discourse. TalkOrganizers that managed topics such as shifts to new or old topic and sustaining topics (linkages,
embedding topics, returning to topic) or making performed Speech acts explicit were also
subsumed under the broad brackets. Thus, for this study the subsumed categories fulfilling
common organizing meta-function are conceptualized to include other componential categories
such as: Topicalizers, Announcements or Reminders of Material, Sequencers, Speech Acts, or
explicit Induction/Deduction, Commentary Directives or Interrogatives. Instances from students
work may include: That’s my opinion. The reasons for… are: As you said, or I think that…
(more in Form to Function Table in qualitative results section, Chapter 4).
Pragmatic Intersubjectivity/Vagueness (I-S)
Intersubjectivity markers are rooted in the pragmatic theory of meaning (cf. Literature
Review) that applies within and across sentences (Fraser, 1998) and interactional–variationist
approach to discourse (Schiffrin, 2001) to account for socio-linguistic variation in collaboratively
constructed discourse. As earlier studies into written metadiscourse (cf. Vande Kopple, 1997;
Crismore, 1985; Hyland, 2000), including the study on the same sample of CR essays (Latawiec
et al., in preparation), have been missing in or revealed deficit in this area of language use that
carries vital meanings especially among young adolescents, intersubjectivity has been ventured
to complement the 2 major brackets/ planes of flow of metadiscursive meaning (along evaluative
and organizing brackets).
Notably, this study does not adhere to the pragmatic markers division as posited by
Schiffrin (2001), who differentiates between 3 different sets of pragmatic markers – basic
(signaling illocutionary force, e.g., please), commentary (encoding of another message that
comments on the basic message, e.g., frankly) and parallel (encoding of another message
separate from the basic and/or commentary message, e.g., damn, vocatives).
Yet, this study does not scrutinize one sub-group of commentary pragmatic markers
(called discourse markers) solely. Instead it investigates an array of metadiscursive elements
among which intersubjectivity pragmatic markers are one of 49 categories thus to avoid
overwhelming granulation of the coding scheme and processes, the study adopts a somewhat
more universal conception.
Intersubjectivity pragmatic markers as proposed here are conceptualized in line with
Andersen’s (2001) view of pragmatic markers that are interpretational and do not directly
contribute to propositions but provide constraints on the interpretation process or otherwise
hearer’s or interlocutor’s prospective procedural encoding. Especially, as after Wilson and
Sperber (1993), explicit meaning resides not only in propositions but in higher-level explicatures,
i.e. information as to what speech act the utterance is used to perform and information about
speaker’s attitude (also Andersen, 2001; or Carston, 1995).
Thus, intersubjectivity vague markers enhance the sense of common ground, sharing of
individual subjective co-conceptions of the world with others, and so evoking or enhancing
social closeness, “establishing camaraderie” (Östman, 1981, Jucker & Smith, 1998, p. 196) or
solidarity feelings among co-communicators/ collaborators. They also belong to a group of
expressions commonly viewed as interactional “vagueness” or “performance fillers” (Channell,
1994) or “coordination tags” (Stenström et al. 2002), and so encompass linguistic forms that
fulfill other metadiscursive code functions of Fillers, Code Glosses (and something, or anything,
etcetera, bla bla bla; cf. “general extenders” or “continuers” in Overstreet & Yule, 1997), and/or
Hedging (e.g., like, kind of/kinda; cf. “presentation pragmatic markers” in Andersen, 2001 or
Schegloff, 1992) or Commentary (y’know; cf. “inviting solidarity” markers of Brown &
Levinson, 1987).
Therefore, this supra-code necessitates double-coding with the other more –molecular
categories (as envisaged in the taxonomy, NB adopted after Vande Kopple, 1997), which
similarly to the other supra-categories (T-O, T-E, ML) adds to some constraints of the analysis.
Gestures or Non-Verbal (Body-Language)
Gestures fit into the posited model of discourse and metadiscourse by tapping the
actional and interactional aspect of the communicative competence, which, as already
mentioned (see communication models in Literature Review), is considered successful when it
achieves its communicative effect. And so, as long as the non-verbal behavior adds relevance
and “semiotic” meanings to the verbal dimension and thus helps to organize and more fully
interpret it, gestures fulfill metadiscursive textual and interpersonal functions in the unique
constellation of events of the pragmatic ever-changing context. Instances of gestures (as
evidenced in the Function to Form Table 10 (Chapter 4), perhaps, would best be explained by
showing their three-dimensional functionality, i.e. by screening the video clips. Nevertheless,
observations from the sample viewed for gesticulation, relevant gaze directional indexing and
other body – language (e.g. elbowing) reveal that CR students used them to fulfill functions of
dramatized/ “theatricalized” simulation of Myself in Scenario, Evidentials: From somebody else,
or Boosters of Illocutionary Force of Speech Acts (as in the Table above), Emphatics, TalkOrganizing and Talk-Evaluating or Commentary -Directive, for example in a back-channeling
manner - pointing a finger when Teacher or another participant comments on the group
performance, thus indicating whose responsibility is conceived by him/her or directing/
prompting others to take turns in actions.
Fillers or Placeholders have been defined by Channell (1994) as expressions used “when
people cannot remember the name of a person or thing”, i.e. dummy nouns which can stand for
item or persons’ names (p. 157, 164), for example “whatsit”, “thing”, “thingy”, etc. (incidentally
all defined as slang by Cassell’s Dictionary of Slang, 2000). In this study to better capture the
group dynamics and verbal interaction across successive discussions, the category
fillers/placeholders has been expanded by other verbal devices that speakers used just to ‘hold
their place’ in the discussion and so not giving the floor to the others, such as “uhm, but”, or
“what if, um, eer”, followed by the main topic continuation mostly.
Major Classes of Categories and Functions
All the above categories have been an expansion to the fundamental framework of
categories as dictated by the observed multiple functions of metadiscourse in both written and
spoken discourse, adopted after Vande Kopple (1997) and Crismore (1985; Crismore et al., 1993;
cf. Figure 1).
The framework as follows has been augmented for the children’s uses of language that
reveal the following functions of metadiscourse (in both modalities; written excerpts marked by
ID #, oral excerpt marked by discussion index and timestamp).
shows relationships between parts of texts via (1) text connectives:
o reminders of material (as I said/noted),
o announcements of new material (what I want to say),
o topicalizers that reintroduce old information and connect the new - to the old info (these
are, as for, in regard to, well…, back on the question/topic),
o sequencers (first, next), as in the following essay extract:
ID 149 “Jack should tell on Thomas. Here’s why. First, Thomas cheated. His brother
built event though Thomas was supposed to. Also, if Jack doesn’t tell, Thomas won’t
ever learn to do things for himself. Plus it isn’t fair to the other race[r]s that Thomas
o logical-temporal connectives (which subsume various types of coordinating
conjunctions: miscellaneous logical-temporal relations (then, at the same time), additive
(and), contrastive (but, however), if-conditional, and non-canonical uses of “so” (other
than causative) and causative (so, because)
- defines or explains words, terms, phrases by (2) code glosses (so called, what some people
call, sort of, I’ll put it this way/ what I mean to say),
- makes explicit what speech acts are being performed at certain points in texts by means of (3)
illocution markers (My/The point is, to sum up, I say/argue/I pick yes, I promise, for
example), and boosts or attenuates the force of certain discourse acts by adding adverbials
dubbed boosters and mitigators (I most sincerely promise),
- indicates some stance on the part of the writer towards the epistemological status of the
referred-to-material via (4) epistemological markers subsuming two sub-types of markers:
(4a) modality markers – reflecting how committed we are to the truth of that material, using
elements from the system of epistemic rather than deontic modality (i.e., concerning duties
and obligations which thus form attitude imparting function): hedges which allegedly “make
things fuzzier or less fuzzy” (Lakoff 1972, p. 195) such as “perhaps”, “seem”, “might”, or
shields (plausibility shields such as it’s possible, perhaps, surprising(-ly), and emphatics
(clearly, it’s obvious/clear that, I’m/ it’s certain that; Believe me, it was an error!)
(4b) evidentials (adopted by Vande Kopple, 1997 after Chafe, 1986) that show what bases do
we have for referential material, stemming from: personal beliefs (I believe, I think), our
knowledge (evidently, I induce), our sensory experiences (it feels/sounds/looks like), from
what we heard from others (as Tabitha/ the book/ the rules say, according to Professor X (cf.
Chafe ,1986 and his “hearsay evidentials”), or from our deduction (X should be able to,
presumably, oddly enough, of course),
- reveals what attitude or emotional orientation authors have towards referential material thanks
to (5) attitude markers (unfortunately/ fairly…, I am sorry/happy, I wish/am grateful, How
awful!/How alarming!), and generally the use of qualifiers that reveal evaluations (positive or
negative), i.e. whether something was preferred or dis-preferred by a speaker or writer
- provides commentary by addressing readers directly in a (6) commentary, their probable
moods, views, hoped-for stance or simply conversing with them, e.g. “Could you imagine/ say
think…? “, when in an interrogative syntactic structure, or declaratives: Some of you will be
amazed that… , or directive and vocative structures: “Go for it!”
Textual Connectives
Among textual connectives especially scrutinized were Causatives (alias causals) alongside
Adversatives as both “signal specific type of elaboration” (Goldman & Murray, 1992). While
causatives signal cause-effect, antecedent-consequent, or problem-solution relation between
conjoined clauses or sentences, adversatives “signal contrastive elaborations” (Goldman &
Murray, 1992, pp. 505-6). Therefore, all occurrences of metadiscourse that would “specify a
disjuncture or departure from the logical argument developed up to that point in the text” (further
on, p. 506) were included in the sub-category of logical–temporal connectives thus labeled as
adversatives (e.g., but, however, in contrast). Causatives (e.g., because, therefore, as a result, so)
were given special prominence in this study as by their nature firstly, they are critical to the plot
coherence in narrative discourse (e.g., van de Broek, 1990), and the antecedent-consequent or
problem-solution argument structure (e.g., Sanders & Nordman, 2000), as well as - in expository
texts – causatives convey important contingency relations among states and events (after
Goldman & Murray, 1992, p. 505, e.g., Saul, 1991), as in the following example “so” and other
occurrences in the meaning of “thus” or “therefore”:
ID 103 “He did not build it so does not deserve it.”
ID 60 “If Jack told Thomas that he cheated Thomas would have told because he
would want to win the race so there fore I pick yes.”
Interestingly, the findings of Goldman and Murray (1992) indicate that conversational
English may contribute to the development of an imprecise understanding of causal connectors
such as so, thus, and because, as the over-attribution of causal relations would suggest. It is
therefore suggested by the authors that they may serve as “pseudobridges” rather than as true
causals. Consequently “greatest overuse of the causal connectors would be expected for those
students whose dominant experiences with English have been informal, conversational contexts”
(Goldman & Murray, 1992, p. 517). Interestingly, a similar overuse has been found by Corrigan
(1975) among children learning English as a native language. Having completed the analysis of
the fourth-graders’ essays, I am tempted to argue that often the pseudo-bridging of causatives is
indeed overused for chaining or simply adding new information.
For better illustration, let’s compare the following “pseudo-bridging” uses of “so” which
often has just phatic function of maintaining or furthering the train of discourse (e.g., ID 182),
and in the last occurrence (ID 12) – is a slight misattribution more aimed at rendering the
purposeful “so that”, unlike the above cited uses:
ID 182 “[The] directions say that he should of done it to him self like everyone they
so he should tell on him so Jack did the wrong thing so the other kid ked of won.”
ID 12 “And that’s why he cheated and told his brother to make it for him so he can
get revenge.”
Additives (e.g., and, moreover) are considered in Goldman and Murray typology (1992)
“the least constraining of the semantic relations between the clauses or sentences related by the
connector” (p. 505). The additive occurrences, therefore, were hypothesized and also found to be
more frequent among writers who develop their arguments in the list-like rhetorical structure,
which is considered least binding (e.g., Meyer, 1975; and Latawiec et al., in preparation indeed
found most frequent use of additives in Non-CR exposed student writing).
Code Glosses
After Vande Kopple (1985), they can be defined as such occurrences that “help readers
grasp appropriate meanings of elements in texts” (p.84). Two main categories of code glossing
devices can be identified – definitions (“so called”) that though meant to define do not expand
referential material, and explanations (“I mean…”) that add some explanatory details as to how
strictly or how loosely the referential material shall be interpreted. For example:
ID 198 “reason 1 why he shouldn’t is Because he told him that he would not tell on him
and that was a promise that he made.” (definition)
ID 755 “If I was Thomas I would say to myself I am going to do this project by myself
because than the award is for me not me and my brother. I would say I do the work I get
the trophy.” (explanation)
in discussion 5 about Marco’s Vote in Ms. Palinski’s class (PAL_5MVR; 9:38.37) “But
sometimes the teacher makes them take them home. And if she makes them take it home,
then – it’d be- it’d be all ripped up and stuff since [inaudible, fades]”, (explanation +
pragmatic intersubjective vagueness marker).
Illocution Markers
According to Vande Kopple (1997) illocution markers are such elements with which we as
writers “can make explicit to our readers what speech act or discourse act we are performing at
certain points in our texts” (1985, p. 84). And in this understanding of illocutionary markers as
emphasizers of performative function I align myself also for the spoken modality not only
writing. Also, those words that attenuate the force of a speech act are called Mitigators (e.g., “I
hate to say this…” or tag questions), while those that boost illocutionary force are called –
Boosters. Since students’ persuasive essays or discussions were prompted by the instruction to
answer the “Big Question” whether or not somebody (e.g., Jack or Marco) should or should not
take a particular course of action?, their says and/ or stances usually are verbalized in the
following illocutionary acts which are expressed by either performative predicate (ID198) or
decisive and / or exclamatory “yes” (ID 25) or “no” that function as a Booster.
ID 198 “I think that Jack should and shouldn’t tell on Thomas Because reason 1 […]”
ID 125 “I know how Jack feels I bet he feels sad that he didn’t win.”
ID 25 “so yes jack should tell on Thomas yes his big brother Dainty his home work…”
or in discussion PAL_8SFR: I am sorry, but I would pass him (6:55.53), or I’m saying
It is noteworthy that polysemous expressions as ID 25’s illocutionary statement often have
double functions and so necessitate double-coding (also admitted by Crismore in personal
communication, 2009). Like in the following excerpt,
ID 400 “That’s why I think Jack should tell on Thomas”,
where “I think” carries a meaning of “I claim/ I argue” but also as a provider of evidence arrived
at by induction (so called Evidential by Induction).
Epistemology Markers
Epistemology markers are metadiscursive elements that reveal concern with the truthvalue of the propositions in the discourse proper. The two major categories of epistemology
markers, both fulfilling interpersonal functions, were labeled after Vande Kopple (1997) as
Modality Markers and Evidentials (cf. Taxonomy Table). When doubts or “small notes of
civilized diffidence” (Williams, 1981, p. 49) are sounded, we can talk of Hedges or Plausibility
Shields a subgroup of Modality Markers which can be rendered by morphological signals (not
impossible), adverbial expressions (probably, rather), modal verbs (may, might), some lexical
verbs (seem, guess, suppose), or other cautious elements like “at least it seemed to me”. Some
hedging expressions in students’ discourse follow:
ID 45 “But maybe Thomas might feel jealous that Thomas won and he didn’t.”
ID 154 “Thomas’s brother could have been 20 years old or something like that.” (as
revealing uncertainty as to the exactness of the age given)
Or in PAL_5MVR 14:32.62 “But if they spill it on a cord…” “They don’t mean???
[SOFT] spill on a cord.[SOFT] (hedging rendered by de-emphasized intonation which is
another cautious element)
On the other hand, Emphatics are used by writers to underscore the truth value of referential
material. In this study there are posited codes not only for the general group of Emphatics (as in
Vande Kopple, 1997) dubbed but also for specific sub-categories of Emphatics miscellaneous
(e.g., exactly, just), Emphatics by Forefronting, Repetition and Punctuation-Orthography which
in oral modality is substituted with Intonation
ID 414 “It was a secret that I think he can keep for a really really really long time.”
ID 76 “All Thomas did was paint and put the Stickers on the car”, (1st – Forefronting, and
2nd - Orthography-Punctuation)
or in CHA_8SFR19:17.00 “I just thought of something um it says it was taxes but there’s
going to be more taxes, right? So how can he stull keep his house because he’s going to
have to pay taxes [EMP] over and over again! [EMP] (Emphasis rendered by Intonation
and Repetition, plus Gesture as Student waves w. her hand rhythmically over every
syllable of “taxes over and over again”)
The other major group of Epistemological Markers – the so called Evidentials –
combines various expressions that reveal the writer’s stance that has to do with the kinds of
evidence or, in other words after Chafe (1986), such expressions that show the bases that we
have for referential material (cf. Barton, 1993). Especially in argumentative type of writing such
expressions will abound since they contribute to the soundness of the argument, specifically its
justification as “they guide the reader’s interpretation and establish an authorial command of the
subject” (Hyland 2005, p. 51). Since evidentials show the basis for referential material, they are
categorized along the ways the propositions were arrived at by personal beliefs (I believe, I
know), sensory experience (it feels/sounds looks like), from somebody else (as it was said in the
book, as the Teacher said), or forms of reasoning: induction (evidently), which in terms of formal
logic is subservient to propositions being made about a class of phenomena on the basis of
observations on a number of particular facts (here often: examples of behaviors) or in arguing
from a part to a whole, or by deduction (presumably, should be able to).As predicted,
metadiscursive references to the bases for the referential material in other people’s work
(“hearsay” evidentials or otherwise From somebody else) were quite common as well as
evidentials by Induction or – by Deduction. Let us demonstrate some of the uses of induction
evidentials in the following excerpts:
ID 214 “The reason he should tell is that so he wouldn’t fell [=feel] hurt.”
ID 769 “That’s why Jack should tell on Thomas. (Induction following a series of
observations leading to a proposition thus made)
or CHA-8SFR 15:31.96 “I think they should let Stone Fox win cause um who cares if
they take a little bit of someone or his property because, or, take some of the white
people’s property because the white people have like so lots of states and stuff and land
and everything. So I think Stone Fox should win.”
As evidenced above, “that’s why” or “the reason is that” (the latter being double coded with
Speech Act) in the essays, in the move from particulars (evidence) to generals (conclusion) are
similar to certain logical-temporal connectives like “therefore”/ “thus” that play a similar role of
introducing conclusion (causative antecedent “because” – conclusive “so”). However, author’s
voice stands out more from the propositional material with the use of evidentials like “that’s why”
than textual connectives “so”, in other words, is a lot more audible. Evidentials by induction
(understood as inference à particularis by Peirce, 1865) seem most crucial in the argument
structure due to their function of explaining, orienting and pointing readers in the direction the
argument is being built by the author. They add to the Aristotelian logos building (logical appeal).
Evidentials by Deduction, otherwise inference à priori (to further use Pierce’s terminology,
1865) resembling Toulmin’s (2003) kernels of argumentation in relation to the argumentative
rhetorical moves, there are numerous deductive thus equaled with hypothetical, conjecture-,
guessing–like constructions that either involve “oneself” or “others” and modal verbs typical of
2nd conditional sentences (“hypothesizing”) or “if...then”/“what if” structures (only 6
occurrences of the latter ones though) as well as intended Retroduction or à posteriori
speculations with the use of 3rd conditional type of sentences (so called “unreal past”).
Speculative sentences when they function as “inference of a to the best explanation of b” are
called as abduction or retroduction in formal logic. Thus, wherever students, like little detectives
in Sherlock Holmes’ manner, tried to generate past hypotheses for observed facts in the present
time, they resorted to modal verbs of “could have”, “would have”, “should have” or “might have”
to explain an occurrence of the consequence (sometimes an almost fallacious post hoc, ergo
propter hoc), as in the excerpt below:
ID 217 “Maybe Thomas could of made friends with his cool [student’s emphasis] car but
he can make without his car. And he probably would of got in trouble wouldn’t win the
trophy but he would be truthful and being truthful get him a long way like at U of I. But
he would be a bigger bully if Jack had told Thomas would have been really mad.”
[student’s limping punctuation preserved]
Here, children’s uses of ‘intended aposteriori” forms (could of made, would of got in
trouble, etc.) resemble successive approximation of the best explanation of a set of known data,
like in the abductive validation process (i.e., a method for identifying the assumptions that will
lead to your goal). Whereas “he can make [without his car]” is a potential deduction (thus, an
inference à priori) that differs from retroduction in the direction in which the rule “a entails b” is
used for inferencing. Here deduction modal verb forms refer to still realizable hypothesis (in the
present or future). Comparable to Vande Kopple’s “should be able to”, “presumably”, “I deduce
that”, or “oddly enough”, fourth graders deductive structures include Deduction: Others in
scenario (IDs 105 & 129) or Deduction: My-self in scenario (ID 929) as utilizing (I, me, my
pronouns), If… then and also what if structures.
ID 129 “After you see your friend’s trophy you might want to earn a trophy by yourself”.
(Others in scenario + Commentary engaging w. direct “you” pronoun)
ID 156 “Maybe they should hold another race” (Others in scenario)
ID 929 “If I was his mother he would be grounded.[…] If I was his teacher the class
would get a party and he would have to do the work.” (Myself in scenario)
or in CHA_8SFR 17:58.45 “It might have been the rule [Aposteriori/ Retroduction] if
that you had to win the race with your dog but if like if your dog was too strong enough
and broke out and like ran off or something like ran out of the leash or something then
they couldn’t if there was a rule [If…then] to win the race if he have to go to get the dog
and latch it back on…and so they could finish the race but they’d probably be in dead last.
[Others in scenario]
In Toulmin’s (2003) argumentation model, in the move from the Data to the Claim there
are “hypothetical bridge-like statements” ( p. 98) called Warrants (“Since”) and possible
reservations/rebuttals (“Unless”) before the Claim can be ventured – with or without a likely
Qualifier (“Presumably”). Thus the observed occurrences of metadiscursive ‘If/ since …then’
and ‘what if’ structures mirror the argumentation proper kernels of warrants and rebuttals
respectively, while the claims may have been signaled in the more or less assertive manner with
the inductive metadiscursive elements (“evidently”=“that’s why”) following multiple
observations, thus arguing from many instances to a general statement (inference by induction,
Encyclopedia Britannica online, or deductive (“presumably” or
“should/should be able” and other modal verbs structures).
Since cogency success in persuasion to a large extent relies on establishing authorial
credibility (trustworthiness, reputation, expertise) during the course of discourse – otherwise
known as ethos in Aristotelian terms, it is posited here that evidentials contribute a significant
share to building this authorial credibility by supporting the author’s persona with the bases for
their reason (see more in Literature Review). Argumentation structure is also more coherent and
transparent with framing of reasoning (induction, deduction, retroduction) in explicitly signaled
patterns of “if/since…then”, “what if …then” that shape argumentation in the characteristic
moulds of Modus Ponens (If A, then B; A, therefore B. / A ⊃ B; A, therefore B.) or Modus
Tollens (If A, then B; not B, therefore, not A.), i.e., two types of inferences that can be drawn
from a hypothetical proposition “If A, then B”, and are otherwise called “method of affirming”
and “method of denying” (Encyclopædia Britannica Online; cf. Latawiec et al., in preparation).
Attitude Markers
Crismore et al. (1993) argue that attitude markers signal affective values, opposite to
epistemological markers, which signal writers’ commitment to truth-value in the referential
material., and so in an eclectic manner it is posited that whenever authors express “an implicit
assumption that the reader will experience the discourse in the same way” (Hyland, 2005 p. 82)
by using specific linguistic items it is where attitude markers reside. Vande Kopple’s taxonomy
has been primarily expanded here (cf. Literature Review; or Latawiec et al., in preparation) by
the addition of modal verbs of obligation (must, has to, should), NB as posited by Crismore et al.
(1993), or lack of obligation (need not, does not have to), permissibility (let, can, could),
prohibition (must not, cannot, should not), as forming the so called Deontic Modality markers.
Deontic Modality markers are based in deontic logic (a branch of formal logic) that studies the
permitted, the obligatory and the forbidden (Greek, deontos: “of that which is binding”, e.g.,
Encyclopaedia Britannica, 2008). After Wünderlich (1979), who argues for 5 categories of verb
operators to be viewed as “positional functors” in their role of attitude marking (both in spoken
and written modalities), two categories converge with deontic modals as proposed by Crismore
at al. (1993); namely, a group of normative verbal modals (ought, should, etc.) and
capability/ability (can, may, be able to). It is noteworthy that obligation /normative modals and
the forbidden category that reveal attitudinal stance have been subsumed under the Evaluating
Bracket, along all other attitudinal markers exclusive of remaining deontic modals.
Also, a category of Implicit Attitude Markers has been added to account for attitude/
assessment by implicature rather than explicit denotation. Similarly to Schiffrin (1980, p. 201)
whenever a child uses a higher level or otherwise two-argument predicates (higher level
operators) like ‘It is true/ right/ wrong that X is the case’ which evidently operate
metalinguistically as “complex modifications of propositions” (so render metadiscursive function
of discussing the propositions or otherwise “discoursing about discourse”), author’s emotional
attitudes and evaluations are implicitly revealed - in tandem with the explicit content material.
ID 414 “I would feel bad if he told”; (Attitude markers incl. “I would”)
ID 229 “I think if I was in his position I would tell and if I was in the other kid’s position
I would want him to tell on Thomas too because it wouldn't be fair if he didn’t.” [Explicit
Attitude Marker is underlined, Implicit Attitude Marker - italicized]
Or in CHA_8SFY 13:06.21 “Well, I think if he could give them like two or something so
that he’ll still have more, but it’ll be really, it’ll be like a fair match, […] So I think he
should give him two of his dogs and then see who wins, because that would be like the
only fair thing to do. Just kind of like let him win.” (Implicit Attitude Marker)
Naturally, the above occurrences show that attitude markers (irrespective of its
explicitness or implicitness) are germane to an emotional appeal or otherwise – Aristotelian
pathos (NB gradable by its nature).
Perlocutionary Commentary
The function of commentary-like expressions is that of “draw[ing] readers into an
implicit dialogue” (Vande Kopple, 1997, p. 8), by either commenting on the readers’ “actual or
hoped for stance” or for example recommending “a mode of reading” (further on p.6). Thus, as
prompted by the very definition, the category may subsume either declarative or vocative or even
interrogative syntactic structures – the latter especially in rhetorical questions, i.e. mostly when
the author has a ready answer in mind.
ID 878 But I’m glad that he just been nice and glad that he had won the first round but he
still sound of told the teacher that Thomas got help and did not obey the rule that you
made so can I have that prize for telling you that he did not do that car by his self?
Or CHA-8SFY 18:54.10 “I still think that he should let him win, but he would have to
put the dog on the sled and push him all the way up there and then maybe Willie could
have $300, because they’re paying $500. Right?” (Commentary Interrogative + Talk
Evaluating by self-accuracy checking)
PAL_8SFY 22:06.92 Okay, you give a reason (Commentary directive/vocative)
In a nutshell, emotionally engaged stance of the fourth-graders via attitudinal and
commentary otherwise considered engagement/involvement markers towards their assertive
claims strengthens their propositional/ ideational persuasion by helping “to convey a credible
persona and relate to an audience in ways that seem familiar and engaging” (Hyland, 2005, p.
Review of the coding scheme:
For the statistical analysis aiming at (1) establishing distributional patterns for each
modality separately and for cross modal correspondences (between talk and written discourse),
an array of 49 common categories have been used (see taxonomy Figure 2). The categories were
inspired by established taxonomies in written metadiscourse. Therefore, to meet the needs of the
analysis in both corpora, another dimension of 3 supra-codes from oral studies have been grafted
upon the extant 40 plus categories, and also supplemented with 2 elementary codes that are
specific to the more embodied oral modality.
While most of the categories assumed oral modality equivalency (like OrthographicPunctuation Emphasis in written modality substituted with Intonation in the oral corpus), 2 oralmodality specific categories of Gestures and Fillers/ Stallers that have been aural-oral corporaspecific did not, and so they could be used only for comparisons between both discussions, i.e. (2)
for investigating relationships between the discussions in their progressive (para-microgenetic)
aspect. The Micro-genetic analysis was then run to determine developmental or reductionist
trends in metadiscourse as used in “talk-in-action” as the discussions progressed in the series.
3.3.5 Coding Process
The coding of the discussions assumed two stages. First, the transcripts were read and
analyzed for the occurrences of potential metadiscourse. Then the video clips were analyzed for
the phonetic, intonational and performative realizations of turns of talk. Observations from the
audio-video data allowed for the evolution of the coding scheme and better accuracy in the
attributions of linguistic items into metadiscursive categories.
The rating of essays from the previous study in written metadiscourse by Latawiec et al.
(in preparation) yielded inter-rater agreement on 20% of the essays (N=180) that amounted to 95%
(Cohen’s Kappa .89). With the data from both corpora, 77 students’ data from discussions yet
only 71 for both discussions and 71 essays, 20% of essays and 20% of discussions were coded by
3 trained raters. Disagreements were consulted and resolved in an iterative process of refining
coding. The inter-rater reliability between the three shows agreement in coding that amounts to
79% in average.
Chapter 4
Discourse Analysis Results
The present study of 12 small collaborative groups (4 yellow, 4 red and 4 blue ones),
yielding in total 24 discussions of various lengths (from 150 to 700 turns or otherwise 20-32
min.), comprised over 10,000 turns of talk in interaction, otherwise ca. 600 minutes of
continuous talk of 77 students (6 students participated only in the latter discussion). The analysis
of written metadiscourse comprises 71 essays of CR-exposed children (that participated in the
major study in 2001-2003, N=107), from two different school locations – one urban and one
rural - in the state of Illinois. The analyses, using a three-partite taxonomy of metadiscourse,
yield in oral modality over 35,000 uses, while in the written corpus – over 10,000, thus
amounting to almost 50,000 occurrences total.
4.1 Quantitative Results
First of all, the frequencies of metadiscursive codes (codes to words ratio as used in the
whole discussion or per essay unit) have been analyzed with simple descriptive statistics. The
results are further documented and illustrated with graphs showing the patterns holding for each
modality separately, to demonstrate the major study goals.
The focus has been on exploring metadiscourse in discussions (goal 1) and writing (goal
2), and correspondences between the two linguistic corpora (study goal 3).
Secondary strand of analysis was wrapped around (study goal 4) investigation of sociolinguistic patterns as a function of gender, reading ability, school and ethnicity. Simultaneous
validation of the measuring instrument in the form of an adopted taxonomy has been attempted
as the results serve both quantitative and qualitative purposes.
The pursuit of study goal (5), i.e. qualitative explorations of the effect of metadiscourse
on the flow of propositional meaning, with which it intricately interlaces and interacts, is
qualitatively discussed in the other section of Chapter 4.
In the Qualitative Results section, it is demonstrated how valency–like patterns of
attribution mechanism proceeds - from elementary variants to the major class categories finally
streamlining to the broad supra-codes or brackets, and how rhetorically, logically and pragmastylistically based interpretations and evaluations are arrived at.
Fortunately, the adoption of the complex hierarchical structure or even a sort of
functional grammar of metadiscourse (understood as system based on functions) as used in
spoken and written modality afforded results usually available in corpus linguistics studies
(especially using tools and techniques available for corpus analysis).
The results of coding along the three-partite scale yielded count data of numerous
categories, many of which did not have a normal distribution.
Primarily, the count data allowed for computing and plotting the frequencies of discrete,
major classes and bracketing metadiscourse variants for better demonstration of their differential
patterns of distribution in two corpora separately and then compared with one another.
Secondly, for the overall analysis of metadiscourse used in both modalities an (overall)
Poisson regression model has been used as best suited to the type of data available.
Thirdly, the metadiscourse frequencies have been further compared using Pearson r
correlation s, p < .05, and t tests between the counts of metadiscourse used (as a dependent
variable) in consecutive discussions, including micro-genetic analysis, and those used in essays
by 71 students.
In the whole sample of 78 CR-exposed students, some data were incomplete, either
lacking in essays or individual discussions and so were excluded from analyses that required
averaged discussions. Since the overall t tests for the available data in essays and in both
discussions were carried out prior to those within groupings, the observed trends for the overall
data within the written or oral corpus will be reported followed by similar tests within groupings.
The groupings for comparisons were according to independent variables of school, ethnicity,
academic achievement and gender.
Study Goal #1: Results for Spoken Modality
Overall within modality comparisons can best be captured by visual graphs rather than lengthy
tables of frequencies that explicitly show distributional patterns for each modality separately.
Therefore, the findings for discussions averaged have been compiled in the following plot
(Figure 3) that illustrates the frequencies of discrete metadiscourse categories on average in both
discussions. All actual scores have been multiplied by 100 for ease and normalization to the
same baseline (across all frequency tests, as shown in consecutive tables notes).
Figure 3. Frequencies of metadiscourse in averaged discussions (discrete categories w/o
As illustrated in the plot (Figure 3), the most frequent categories in oral modality were
those of Speech Acts, Fillers, Metalanguage, textual Connectives (Additives, Causatives,
Contrastives and in this order), Emphatics (by Gesture – embodied and solely oral modality
specific, sic!; and Miscellaneous). With Perlocutionary-force-bearing Commentary Directive/
Vocative and Glossing by Explanations (e.g., I mean, sort/kind of, like) the frequencies gradually
drop (below 1 instance per 100 words). The top frequency codes (around 1 and above per 100
words) of performative Speech Acts or Commentary Directive, or Code glossing with “I mean”
which were generously used by CR small group discussants feed into the Text-Organizing
Bracket, while evaluative uses of Speech Acts (e.g., I agree/disagree, that doesn’t mean…,
yeah/nope, Ok! Right!) are embraced by Talk-Evaluative meta-function.
High uses were also noted for Implicit Attitude Marking (incl. complex 2-argument
evaluating structures “it’s not right/fair/true/untrue that …X is the case”) of slightly below 1 time
per 100 words, Interrogative type of Perlocutionary Commentary, emphatic uses of oral modality
specific Intonation (sic!), Obligation revealing modal verbs of Should/Must (stronger) and twice
less frequent Had to (weaker form), If-then and Others in scenario deductive structures, or
slightly lower - globally cohering Topicalizers and Reminders of Material (within textual global
coherence connectives). Approximately 20 categories, as shown in the graph below the
evidential From Someone Else (.40) and Obligatory Had to (.37), show frequencies between 40
to 2 per 10,000 words.
Notably, versatility of fine metadiscourse categories, though most revealing, may be
somewhat overwhelming, and so some strikingly diverse patterns of use and function are
highlighted in a Form-to-Function table in a Qualitative Section of Chapter 4.
For an ease of grasping the larger patterns, I grouped the discrete categories within the
major ones and the three supra-codes which are thus illustrated separately in another graph
(Figure 4). The meta-functions of organizing, evaluative and intersubjectivity brackets as well as
totals, or otherwise major metadiscourse categories (instead of the 40 plus discrete categories)
are shown below.
Figure 4. Metadiscourse frequencies for brackets and totals in averaged discussions
Figure 4 shows that students use approximately twice as much overall talk-evaluating
(10.4) than talk-organizing (5.3) categories of metadiscourse. In spoken modality,
Intersubjectivity-Pragmatic marking occurs as frequently as almost 1 time per 100 words (which
when compared with the other 2 brackets amounts to 7%). Notably, the highest use of 6.8 per
cent, apart from the brackets, are found for Logical-Temporal Connectives (comprising
Miscellaneous Logical-Temp. connectives, Additives, Contrastives, Causatives, If-conditionals,
and Non-canonical uses of “so”), and at 4.2 per cent - Illocution Markers (Speech Acts, incl.
Boosters and Mitigators). Evidentials (NB, comprising nine elementary categories) occur at 2.9
words per 100, and a bit more often than Emphatics with M=2.5 (incl. intonation, lexical and
syntactic repetition, miscellaneous lexical choices, e.g., ‘really’, ‘so’, ‘exactly’, ‘only’; found
increased commitment in American rather than British conversations by Precht (2003). Also high
frequent use of were found for Commentary (collapsed interrogative and directive categories)
with a frequency of 1.9 per 100 words. Some use of various Hedging devices (modal, lexical,
adverbial, or other cautious devices incl. soft-intonation) can be noted too, as above 1.48
occurrences per hundred words, and equally frequently occurring were Attitude Markers, when
grouped as a class with the exclusion of Deontic Modality Markers (with individual elementary
codes shown in Figure 3). Global Coherence connectives (grouping Topicalizers, Sequencers,
Announcements and Reminders of Material) were also found to have slightly above 1 occurrence
per cent, precisely M=1.12. Code Glosses which group just 2 codes of Explanations and
Definitions fall slightly below 1 (with M=.98).
High frequencies of Evaluative codes (Talk-Evaluating Bracket), Modality markers and
perlocutionary/ coercive Commentary uses indicate a pattern typical of highly-engaged style and
high commitment (e.g., documented by Precht, 2003).
Micro-Genetic Analysis of the Oral Corpus
With the data from coding of 2 discussions, the micro-genetic development between
discussions was afforded in this study too. Hence, in order to trace micro-genetic development
Pearson r correlations, p<.05, were run between discussions # 5 and #8. The correlations show
many strong positive relationships for metadiscourse codes used in discussion 5 and 8. The
strongest relationships hold for the categories of ‘Perlocutionary’ Commentary Interrogative,
Personal Belief, Impossible, Fillers. Several strong and moderately strong relationships have
been found for Sequencers, Contrastives, Causatives, Metalanguage, and Adverbial Hedges.
Also, weak relationships have been found for If/since-Then, emphatic codes of Repetition and
Intonation, as well as Obligatory “Should/must” and Gestures. These relationships indicate that
the more these codes were used in discussion 5, the more they were used in discussion 8 too.
Thus, the results suggest a pattern in their usage by CR-discussants rather than a one-time
Moreover, dependent samples t test was run to investigate differences between the two
discussions for all students participating in both discussions (n=69). The results, p<.05, reveal
that the earlier in a series of discussions, that is discussion 5, shows more frequent uses of
Sequencers, Pragmatic Intersubjectivity/Vagueness Markers along such metadiscursive
categories as 2 evidentials of What if and If then, Deontic Modality markers of Possible and
Obligatory Had to. While in the later discussion, Discussion 8, more frequently used codes
include Topicalizers, Speech Acts, Commentary Interrogative, Obligatory “Should”,
Metalanguage as well as Talk-Organizing and Talk-Evaluating Brackets (see Table 1).
Table 1
Frequency of Metadiscourse in Successive Discussions 5 and 8 (df=68)
Discussion 5
Discussion 8
Non-Canonical So
What if
Personal Belief
Commentary Interrog.
Obligatory SHOULD
Obligatory HAD TO
Metadiscourse Category
Talk-Evaluating Bracket
Talk-Organizing Bracket
Speech Acts
As noted above, the micro-genetic findings show the trend toward more frequent use of
metadiscourse categories in the later discussion (in a series of 10) that fulfill functions of TalkOrganizing and Talk Evaluating Bracket as well as stronger stance expression by obligation or
speech acts. Seminally too, metalanguage use also shows a developmental progression in use.
Study Goal #2: Results for Essays
As for the written modality, the statistics show the top frequencies for predominantly
text-organizing metadiscourse categories, as represented in Figure 5. Specifically, the most
frequent Speech Acts’ (4.4) rank no. 1 owes its high usage to multi-purpose functions as either
performative verbs of evidential type (I think/say/know/don’t know/believe) or evaluative (I
agree/disagree, and alike). And the second and third top frequencies are those of text-binding
connectives of Causative (3.3) and Additive (2.8) type, both adding to Text-Organizing Bracket
(in Figure 6). Next in a row come attitude and stance revealing Obligation modals Should/Must
(otherwise dubbed as “positional functors” by Wünderlich, 1979), Emphatics Miscellaneous (1.7)
and Implicit Attitude markers (1.4) - that all feed into Text-Evaluating Bracket (shown further, in
Figure 6).
Figure 5. Frequencies of metadiscourse in essays (discrete categories w/o totals or brackets).
Importantly, with frequencies around 1 per 100 words other codes have been found
which highly contribute to text-organizing or text-integrating, such as Contrastives (1.3),
evidentials like Induction (1.0) and hearsay evidential “From Someone Else” (.95) and almost
equally frequent -“If/since…then” (.90). On the other hand, very high frequencies of positional
functors of Forbidden category (.9) and less frequent Adverbial Hedges (probably, possibly,
maybe) contribute to the broad Text-Evaluating Bracket.
Other integrative or otherwise text-organizing codes such as If- Conditionals (.8),
Logical-Temporal Connectives Miscellaneous (.5) and more globally binding Sequencers (.5) or
Topicalizers (.4) occur also very frequently in CR students’ writing – together with relatively
high Metalanguage uses at a rate of 5 per 1,000 words (e.g., point/ problem/ reason/ argument/
position/ comment/ decision/decide, etc.).
Less frequent uses - 4 times per 1,000 words have been found for 21 discrete categories
(which constitute the lower half of graph 3), among which Boosters of Speech Acts (.4) or
Perlocutionary Commentary deserve a special mention as they considerably impact the
persuasiveness of written discourse by intensifying the Illocutionary force of Speech acts
(boosters) or by urgency-adding or action-reaction prompting (Commentary) - thus highly
increasing engagement of intended readers.
Further examination of metadiscourse uses as grouped by major classes and the broad
brackets allows us to grapple with the written array more easily (see Figure 6).
Figure 6. Metadiscourse in essays – frequencies for totals and brackets.
As shown in Figure 6, still in the lead is Text-Evaluating Bracket (10.4) and as second
comes a group of locally binding Logical-Temporal connectives (7.5). Text-Organizing (5.4) is
again almost half as frequent as Evaluating Bracket: however it occurs more often than in
comparative oral discourse (as shown in Figure 4). Illocution Markers again come at one of the
highest frequencies with a mean at 4.8 (higher than discussions mean). Emphatics in total show
frequencies of 2.2 words per cent (.3 lower than in discussions), while total Hedges show
considerably lower means of 7 occurrences per 1,000 words (M=.7), and also twice lower than in
oral modality. Attitude markers (excluding Deontic modals but inclusive of Implicit, Explicit/
Miscellaneous, and “I would” markers) near 2 uses per 100 words (M=1.96). Emphatics, Hedges,
and Attitude Markers all contribute to the frequencies of the broad category of Text-Evaluating
Bracket. For a change, the organizing bracket category of Code Glosses with M=.46 was used
half less frequently than in discussions. Similarly, Commentary shows much lower frequency
(.26), which is almost 10 times lower than in discussions. Notably, the third bracket forming
Intersubjectivity-Pragmatic marking appears 2 times per thousand words, which is five times
lower than in group-discussions.
Study Goal #3: Results for Inter-Modal Correspondences
To investigate possible correspondences between the patterns of metadiscourse used in
discussions and essays (study goal #3) by the same CR participants (on two different tasks and
different times of measurement), t tests for dependent samples (n=71) for the comparison of
metadiscourse in essays and the averaged discussions (#5 and #8) were run.
The results for the overall pool of CR-exposed students showed significant differences, p
<.05, for 21 pairs of codes in the mean of discussions (see Table 2, and corresponding Figure 7).
More frequent in essays were categories of deontic modality labeled as Obligatory
Should/ Must and Forbidden as well as evidentials by Induction, From someone else,
Aposteriori/ Retroduction Intended or If-Conditional structures (mostly incomplete ‘if-then’inferencing structures).
Table 2
Frequency of Metadiscourse Used in Essays and Across Both Discussions (df=70)
Metadiscourse Category
Reminders of Material
Other cautious elements: Hedges
Modal Verbs: Hedges
Hedges Total
Mitigators: of Illocutionary Acts
From Someone Else: Evidentials
Aposteriori/Retrod. Intended: Evidentials
What-If structures: Evidentials
Personal Belief: Evidentials
Sensory experience: Evidentials
Explanations: Code Glosses
Forbidden: Deontic Modality
Obligatory Had to: Deontic Modality
Non-Obligatory: Deontic Modality
Permissible: Deontic Modality
Possible: Deontic Modality
Intersubjectivity Pragmatic Markers
Logical-Temporal Connectives Misc.
Induction: Evidentials
Obligatory Should/Must: Deontic Modality
Note. Actual scores have been multiplied by 100. p <.05
For a change, the averaged discussions have shown more frequent codes of Reminders of
Material and Logical-Temporal Connectives, 3 types of various Hedging devices (by Modal
Verbs, Other cautious elements, Mitigators of illocutionary acts), 3 codes within a major
category of evidentials – What if-structures, Sensory Experience, and Personal Beliefs. Also,
there were higher frequencies of Commentaries (Directive/Vocative & Interrogative), Glossing
by Explanations, Intersubjectivity Pragmatic Markers and Meta-language, as well as various
categories of verb operators (NB “positional functors” in Wünderlich, 1979) that are fulfilling
deontic modality functions such as the Permissible, Possible, Non-Obligatory or Obligatory Had
Figure 7. Frequency of metadiscourse in essays and mean discussions (df=70).
As mentioned in results section for study goals 1 and 2, the patterns of metadiscourse
uses for both modalities assume different structures. The significant differences, p<.05, (as
illustrated in Figure 6) suggest that within textual categories in discussions prevail codes of
globally binding Reminders of material (.4) or Code Glossing by Explanations (1.0; for example,
“I mean”). Among interpersonal metadiscursive uses prevalence has been noted for attenuating/
Hedging devices (1.5) and weaker forms of Obligation by Had to (.4), explicit statements of
Personal Belief (.4) or Metalinguistic designata (1.8), as well as Hearer-targeted Perlocutionary
Commentaries (1.0; e.g. nudging) and What if- prompting or rhetorically questioning (.2) forms
of evidentials. Interestingly, significantly higher uses of Logical-Temporal Connectives
Miscellaneous (1.0) for propositional binding – textual category, and likewise, considerably
higher frequencies of Intersubjectivity Marking (1.1) have been observed for interpersonal
metadiscursive uses in discussions rather than essays.
In contrast, in essays significantly prevail If-conditional (.8) connecting devices (nota
bene, usually used in tandem with hypothetical/deductive structures) that often entail Aposteriori
Intended forms (.7) too, i.e. grammatical past perfect tense and 3rd conditional/ ”unreal past”
syntactic structures, classed as evidentials together with hearsay evidential From Someone Else
(.9). Alongside the evidential uses, highly text-organizing Induction markers (1.0) as well as
stronger forms of Obligation Should/Must (2.2) and Forbidden (.9) have been found to prevail in
essays over discussions too. Pertinently, these findings for Obligation Modals may be linked to
the finding of a robust pattern that shows more frequent verbal modals uses in American rather
than British conversations by Precht (2003), further discussed with an issue of stance marking in
Chapter 5.
Poisson Regression
To further investigate the correspondences and better model the specificity of the count
data a Poisson regression analysis was run to explore what associations hold between discussions
(treatment) and essays (written as the final post-intervention product), when essay length
variance is controlled for and variances in other variables are held constant. The Poisson model
was selected as best capturing this count data distribution, whose frequencies deviated from
normality, yet whose fitted values turned out not to be much over-dispersed as the deviance and
Pearson dispersion indicated (e.g., Cameron & Trivedi, 1998). Counts in individual
metadiscourse categories in written essays were dependent variables, while counts in mean
discussion codes, essay length, gender, or MAT score were predictors.
Table 3
Poisson Regression Results for Mean Discussions (as Predictor)
Exp (β)
Meta-Language M
1. Text Connectives:
 Total Global Coherence Connectives M
Metadiscourse categories
Talk-Organizing BRACKET M
Talk-Evaluating BRACKET M
IF-Conditionals M
2. Epistemology Markers
 Modality Markers:
o Emphatics: Repetition M
Evidentials Total M
Induction M
Note. Values are shown in comparison with corresponding metadiscourse category (reference group) in
essays. “M” at the end of the parameters stands for the Mean of both discussions.
As shown in Table 3, after controlling for Essay Length, Gender and MAT score, the
results of the Discussion codes (Alpha = .05) for both Organizing and Evaluative brackets
indicate that for one occurrence of both categories in discussions the expected log counts for
Talk-Organizing and Talk-Evaluating codes will show increase by 1% in writing (chances are 1 %
higher). This means that the two bracketing codes are predicted marginally less frequent in the
oral modality. Likewise, chances for Metalanguage in discussions are lower by 8% (per 1 time/
use of written metalanguage). Two textual categories of Global Coherence Connectives and Ifconditionals show similar relationships, i.e. predicted higher chances of use in writing by 5%
and 33% respectively (or otherwise, reduction in discussions). Also, higher predicted log counts
in writing have been found for Induction markers and totals of Evidentials - by 23% and 3 %
respectively (compared to oral modality). However, emphatics via Repetition are likely to show
80 % increase in use in discussions in comparison to essays.
Study Goal #4: Results for sociolinguistic variation across modalities
To complement the explorations of patterns that hold for each modality (study goals #1
and 2) and between modalities (goal #3), the sociolinguistic component of quantitative analysis
was added to investigate goal # 4, that is metadiscursive uses across both corpora as a function of
gender, reading ability, school location and ethnicity as vital factors influencing the language use
in general.
Therefore, independent samples t-tests were run for each grouping and across written and
oral datasets gathered via the application of the same complex taxonomies for metadiscourse
Comparisons of the means in a grouping by gender that is between boys’ (N=38) and
girls’ (N=33) uses of written metadiscourse revealed that girls used significantly more
hypothetical Myself in scenario and Logical-Temporal Connectives, while boys used
significantly more Causatives for binding their propositional content in essays. The exact figures
are presented in Table 4.
Table 4
Frequency of Metadiscourse in Writing for Gender (df = 69)
Metadiscourse Category
Logical-Temporal connectives
Myself in scenario
Note. Actual scores have been multiplied by 100. a N=38. bN=33. p <.05.
Further analysis of independent samples t-test of metadiscourse uses grouped by gender
across both discussions shows that boys use Boosters (M=.23, SD=.26) twice as much as girls do
(M=.10, SD=.15) in order to intensify their Illocutionary Speech Acts.
When the grouping variable was low or high reading achievement score in the
Metropolitan Achievement Test (MAT), it has been found in written metadiscourse that students
with high scores on the MAT (n=37) more frequently used the Text-Organizing Bracket
(M=6.68, SD=7.20), subsuming Speech Acts, Topicalizers, Reminders of Material,
‘Perlocutionary’ Commentary and Induction/Deduction signaling codes, than their peers with
low MAT scores (M=3.87, SD=3.34). It has to noted that actual scores have been multiplied by
100 across all tests - for ease and normalization to the same baseline.
Within the same grouping by different academic achievement levels across consecutive
discussions (#5 and #8), the results show that students that scored low on the MAT (n=37)
demonstrated statistically higher use of performative Speech Acts, while their higher achieving
peers showed more frequent uses of 5 codes. Table 5 below shows the seminal frequencies.
Table 5
Frequency of Metadiscourse for Low and High MAT-scores Across Discussions (df=75)
Metadiscourse Category
High MAT
Speech Acts
Others in scenario
Emphatics Total
Note. Actual scores have been multiplied by 100. p <.05
Namely, high-MAT scorers used more often Sequencers, deductive-hypothetical
structures putting Others in scenario and Aposteriori/ Retroductive ones (both function as
evidentials) as well as more frequent modality marking with Emphatics Total and emphatics by
Similar independent samples t-tests with grouping by school location of metadiscourse in
children’s essays show that students in rural school (n=36) used more codes feeding into a supracode named Text-Organizing Bracket (Speech Acts, Topicalizers, Reminders of Material,
‘Perlocutionary’ Commentary and Induction/Deduction signaling codes) as well as more discrete
codes of Contrastives and Personal Beliefs. For a change, their comparative counterparts in the
urban school (n=35) used more metadiscourse that is subsumed under the Text-Evaluating
Bracket (Modality Markers and Attitude Markers, evaluative Speech Acts and deontic modality
of Obligation and the Forbidden category), as shown in Table 6.
Table 6
Frequency of Metadiscourse in Writing Across Two School Locations (df = 69)
Metadiscourse Category
Urban School
Mean SD
Rural School
Text-Organizing Bracket
Text-Evaluating Bracketa
Personal beliefs
Note. Actual scores have been multiplied by 100.
Evaluating Bracket comprises evaluating Speech Acts, Emphatics, Hedges, Attitude markers w/o deontic
modality and deontic modals of Obligation and Forbidden.
p <.05.
Results of independent samples t tests between two school locations across the averaged
discussions (#5 and #8) showed primarily higher frequencies for the Rural school children
spoken discourse (in Table 7). Specifically, the rural CR discussants used several types of
Hedges (Adverbial, Other Cautious Elements and Hedges total) and two categories of Attitude
Markers (Total and implicit attitude) as well as 2 major brackets of pragmatic Inter-Subjectivity
and Talk Evaluating (so mostly, interpersonal metadiscourse). Rural school students also utilized
more Code glossing by Explanations, more Speech Acts and Sequencers to organize their talk in
action, and several types of emphatic modality marking by Forefronting and Emphatics
Miscellaneous (again interpersonal and evaluative meta-functions). By contrast, schoolchildren
from the urban location show more frequent use of 2 organizing codes of Reminders of Material
and Perlocutionary Commentary Interrogatives (thus both subsumed under Talk-Organizing
bracket) and an evidential category of hypothetical Aposteriori/Retroduction (see Table 7).
Table 7
Frequency of Metadiscourse Across Schools in Averaged Discussions (df = 75)
Urban Schoola
Rural Schoolb
Reminders of Material
Other Cautious Elements
Adverbial Hedges
Personal Belief
Forefronting Emphatics
Emphatics Miscellaneous
Commentary Interrogative
Implicit Attitude Markers
Hedges Total
Attitude Markers Total
Metadiscourse Category
Intersubjectivity-Pragmatic Markers
Talk-Evaluating Bracket (Broad)
Speech Acts
Note. Actual scores have been multiplied by 100. N=41. N=36. p <.05.
The last comparisons of metadiscourse were those as a function of ethnicity between
European Americans and African Americans (see Table 8), though run with caution as ethnicity
was confounded with school (one school has been dominated by European American population).
Here the results in essays reveal more frequent uses of Contrastives and Emphatics total by
European Americans, while more frequent “I would”-based Attitude Marking by African
American children.
Table 8
Frequency of Metadiscourse in Writing Across Ethnic Groups (df = 65)
Metadiscourse Category
African –
Mean SD
Attitude Markers w. I WOULD
1.29 1.50
Emphatics Total
Note. Actual scores have been multiplied by 100. aN=22. bN=45. p <.05.
Interestingly, metadiscourse usage across both discussions shows greater diversity for the
ethnic grouping than in essays, as represented in Table 9 respectively.
Table 9
Frequency of Metadiscourse Across Ethnic Groups in Averaged Discussions (df = 71)
Metadiscourse Category
Intersubjectivity-Pragmatic Markers
Reminders of Material
Non-Canonical So
Speech Acts
Others in scenario
Aposteriori/Retroduction Intended
Emphatics Miscellaneous
Hedges Total
Commentary Interrogative
Implicit Attitude Markers
Explanation Glosses
Note. Actual scores have been multiplied by 100. N=27. N=46. p <.05.
In the averaged discussions, African-American children have been found to use more
Perlocutionary force imposing Commentary Interrogatives, organizing Reminders of Material,
as well as Non-canonical “so”, used for connecting the propositional content, and hypothetical
Aposteriori/ Retroduction scenarios or emphatic Repetition. Their European-American peers
have been found to use more Intersubjectivity Pragmatic Markers and code glossing
Explanations as well as more Additives to bind their talk. European Americans utilized more
Speech Acts and hypothetical Others in scenario, as well as several types of modality markers
such as emphatic Forefronting, Emphatics Miscellaneous and Hedges Total (thus contributing to
Talk-Evaluating Bracket). Their discussions bear twice as frequent Implicit Attitude Marking
and inferencing If-then structures as those by their African-American CR co-discussants.
In summary:
Analysis of metadiscourse within each modality (written and spoken) revealed significant
differences in the distribution patterns of both discrete elementary categories and major classes
with over-arching brackets. Findings between the modalities also showed systematic differences
as well as some correspondences that were thus established, among others that the ratio of
Evaluating Bracket to Organizing Bracket is ca. 2:1 in both modalities (to be further discussed in
the Qualitative Section), with Intersubjectivity amounting to 7% in discussions and 1%- in
writing. Organizing Bracket comprises: globally cohering Announcements and Reminders of
Material, Topicalizers, Sequencers, Glossing Explanations, performative Speech Acts, and
Induction/Deduction marker. Evaluative Bracket subsumes: evaluative Speech Acts
(agreement/disagreement), Emphatics, Hedges, Attitude markers w/o deontic modality,
Obligation modals of “should”, “had to” and Forbidden category. In overall discussions, high
frequencies of Evaluative codes, Modality markers and perlocutionary/ coercive Commentary
uses suggest a highly-engaged style. Overall analysis within written modality shows that text
organizing codes (Logical-Temp. connectives, Global coherence totals, Glosses, Illocution Ms.
and Text-Organizing Bracket) show higher frequencies than in discussions.
Micro-genetic analysis across both discussions found many positive correlations and
dependent samples t-test analysis show an increased use of many Talk-Organizing categories and
Talk-Evaluating (including stronger forms of obligation) or Metalanguage, with a notable
decrease in the use of Intersubjectivity-Vagueness Bracketing code).
The analyses of metadiscourse across both modalities as a function of gender, reading
ability (MAT), school location and ethnicity also yield many significant results thus revealing
distinctive patterns of use by the compared groups. Among other findings, results for grouping
by gender show more uses of Boosters by boys in their talk and more Causative connectives in
essays, while girls were found to use more Myself-oriented scenarios and Miscellaneous Logical
Temporal Connectives in writing. High MAT-scorers used more Organizing Bracket in writing,
while in their speech – more Sequencers, hypothetical scenarios, and various Emphatics. Low
MAT-scorers used more performative Speech Acts in discussions. Analysis of essays by school
location shows greater use of Organizing Bracket and Contrastive connectives by a rural school,
whereas more Evaluative bracket by an urban school. Ethnic comparisons show higher uses of
Contrastives and Emphatics total by European-Americans, whereas “I would”-Attitude Marking
by African-American students in writing. Analysis of discussions reveals many more differences
than analysis of students’ writing as a function of ethnicity. Though, ethnic grouping is
confounded with school location, and so need to be cautiously interpreted.
4.2 Qualitative Results
Study Goals #1 and #2 intend to:
(1) analyze metadiscourse used in small group discussions by CR-exposed children as well as
(2) analyze the metadiscourse in their written essays.
Primarily, analysis aiming at establishing links between an array of linguistic forms and
functions has yielded rare and interesting observations about children’s specific uses of
metadiscourse for the achievement of various communicative goals.
Also, in an elementary-to-holistic approach, the adoption of three-partite system of
metadiscourse, from building-block elementary categories via major classes to broad brackets of
metadiscourse, resulted in yielding results on three levels of complexity – 1) very discrete
functions of ‘grass-root’ level, 2) major classes, and 3) the broad meta-functions of supra-codes.
Notably, the supra-codes helped to ‘curb’ the diversity of the discrete codes which
resulted from a split of major codes into the fine ones - often tapping similar functions though
differently (by different means, like various subcategories of emphatics, hedges, evidentials, or
attitude marking - especially with an array of deontic modality verbs or otherwise “positional
functors”, as in Wünderlich, 1979).
The specificity of the richest meta-communicative functions of selected meta-discursive
uses is presented in Table 10, titled “Form to Function”, as one of the outcomes of the pursuit
for the two (2) major study goals – to analyze discourse in group discussions and to compare it to
the metadiscourse in the written essays (as laid out in Chapter 2). Therein the illustrated
metadiscourse category is highlighted, while the other categories are inserted below the “at” (@)
symbols, which are included to better demonstrate the density of forms within a turn-at-talk. The
Form to Function Table sets out ‘valency’ patterns for the emerging grammar of metadiscursive
categories that have never been explored in such a context and scope, i.e. never (1) in
combination of elementary and more bracketing functions, and (2) across two modalities.
Following the reported richness of form-to-function relationships (Table 10), authentic
episodes and stretches of texts with the highlighted phenomenon are shown and discussed to
further demonstrate the outcomes of the pursuit of the 3rd and 5th goals.
Specifically, major similarities between metadiscourse in both modalities as used by CRparticipating children are tracked and combined with assessment of the identified metadiscursive
patterns with respect to how they add to the flow of meaning in oral and written argumentation
of CR-children (to pursue study goal #5), which thus conclude the chapter.
Table 10
Form-to-Function and Function-to-Form
Metadiscourse Metadiscourse Forms, Functional Affordances and/or Constraints
Stephanie: @I [email protected] it went good, @[email protected] nobody was @[email protected] @[email protected]
arguing @[email protected] fighting over who it was @or anything./
Shelby:@I [email protected] we did better at @like @a lot of people sharing their ideas @[email protected]
not @[email protected] sitting there @[email protected] not giving ideas. (CHA_8SFY)
----Richard: @I [email protected] that @it went all the [email protected], @[email protected] you’re supposed to,
@you could have different opinions about this [email protected] @[email protected] you could @[email protected],
@umm,@ fight about, @you could @[email protected] fight about your [email protected] @[email protected] try to
get somebody else and change their mind, @[email protected] you @got [email protected] @ take turns while
you’re [email protected], @[email protected] have one person do it, @[email protected] have somebody else.
[…another student’s interjection]
Tabitha: @[email protected] if you’re fighting, you @got [email protected] @at least take [email protected] fighting.
(CHA-8SFR; 24:31.68)
Other uses of ‘language as a code’ (from excerpts): sides of the issue, think and talk
critically, arguments/reasons, ideas, problems, blame, criticize
Talk -/ TextOrganizing (T-O)
Jeremy: I @[email protected] [email protected] that he @[email protected] @[email protected] him win, @[email protected] he would
@have [email protected] put the dog on the sled @[email protected] push him @all the [email protected] up there @and
(table continues)
Table 10 (continued)
Metadiscourse Metadiscourse Forms, Functional Affordances and/or Constraints
[email protected] @[email protected] Willie @[email protected] have $300, @[email protected] they’re paying $500
@[email protected] (Talk-Organizing + Rhetorical Question=Commentary Interrogative +
Hedge) (CHA_8SFY; 18:54.10);
Danielle: @I [email protected] that Stone Fox, @I [email protected] I think that Willie @[email protected] @[email protected]
@[email protected] take his dog home on the sled @[email protected] he walk home @[email protected] @[email protected] Stone
Fox @[email protected], $you know$, [email protected] him have {some} dogs @and [email protected] @[email protected] race,
go back @[email protected] a mile @[email protected] race again. (T-O + Pragmatic Intersubjectivity Marker)
(CHA_8SFY: 15:35.55)
Chillbrisha: One person at a time please! (T-O + Commentary Directive +Emphasis by
Intonation in Exclamatives) (PAL_8SFY; 19:00.15)
T-O then subsumes:
Performative Speech Acts: I think, I know/say, the problem/ idea is
Topicalizers: well, so, now; ok, then…
Reminders of Material: Like/As you said
Sequencers: first, second, plus, and then/next
Announcements of Material: here are my reasons, here I’ll tell you
Code Glosses: I mean, like (in explanatory sense)
Induction/Deduction Markers: That’s why/ So I think/ I would…
Evaluating (T-E)
Kevin: [email protected], @[email protected] @I [email protected], @what if you were the dog, @OK{!?},@ @[email protected]
you were in a race @[email protected] you died,@ wouldn’t you be [email protected], @I [email protected] what if
you were, @@what if you were a dog @[email protected] you @had to @be, you were @[email protected]
[email protected]@ @and [email protected] to race, to run, @@wouldn’t you be @kind [email protected] mad
@@at the [email protected] (self-T-E+ Com-Interrogative by Intonation) (CHA_8SFY:
Bryan (shrugged his shoulders)
Shelby: @Bryan,@ @@that’s not [email protected]@ @Could you give us some of your
[email protected] (CHA-8SFY: 01:22.00)
Other uses from excerpts:
Yeah. No! – when standing on its own, as cries of agreement or disagreement
I agree/ disagree (Speech Acts + T-E)
True. Right. OK! (as expressions of agreement, not Topicalizers)
That’s not the point! That’s not true! (Speech Act +T-E)
Dang! (T-E + Intersubjectivity M) What you mean?! (Commentary- Interrogative + T-E)
That doesn’t mean… (Code Glossing + T-E)
I am not saying ….!
TE Subsume class categories of Emphatics, Hedges, evaluative Speech Acts (agree/
disagree), Attitude Markers w/o deontic modality, Modals of Obligation and
Forbidden category.
(table continues)
Table 10 (continued
Metadiscourse Metadiscourse Forms, Functional Affordances and/or Constraints
Kevin: @Bryan,@ @what do you [email protected] (Commentary Vocative + Commentary
Interrogative) (CHA-SFY: :05:57.11)
Dylan: @Back on the [email protected],@[email protected] @[email protected] (Commentary Vocative + TOrganizing +“the subject”= Metalanguage + “OK?”=Commentary_ Interrog.)
(PAL_5MVR; 11:58.09)
Other common uses:
Say it again, Pointing fingers (in a nudging gesture to speak, take turns)
Why do you say that, Stephanie? (cd) (CHA_5MVY) – involving directly an addressee
or Rhetorical questions:
Shayla: If he got $500 and you give it all to the one people how you you gonna have
some more money to buy a new dog?
DeAngella: Who says he gonna buy a new dog?! (PAL_8SFR; 10:24.99)
Shelby: @Anybody got any [email protected] (T-O+ Commentary Interrogative)
Other common uses (excerpted) :
What would you say if….? So what if he cheated?! (other Rhetorical Questions” or
Hanging uses of connectives with rising intonation, e.g.,You’d do it, soooo?)
Kevin: @[email protected], @you said [email protected]{Reminder of Material +“hearsay”Evidential} they
should shorten the race by @[email protected]{cautious elem.} one or two miles, @I [email protected] with
you,@ [email protected] @what [email protected] you had like {=e.g., so a Speech Act) a little German
Shepherd, @[email protected] five, @[email protected] almost a year ago, @[email protected] he was @[email protected] strong
@[email protected] he run @real [email protected], you didn’t know it was @[email protected] that long. @[email protected] it
@[email protected] died, @[email protected] @right in the [email protected] of the race. (CHA_8SFY:05:14.91)
 Stallers – “really, really, sort of, you know” (when one doesn’t know the word or
doesn’t know how to proceed)–resorts to repetitions of any linguistic forms &
elongation of sounds:
o Phatic function of sustaining communicative flow or a place-holder/place-mat
Asha: That don't mean you can hit people upside the head.[…]
Joey: No but? push him like that. [demonstrates pushing with shoulder]
Shavon: @[email protected] @what [email protected] @I'd [email protected] [makes snapping noise and pretends to pop
Joey in the head] (Myself in scenario by Gesture) (PAL_8SFY; 19:03.25]
(table continues)
Table 10 (continued)
Metadiscourse Metadiscourse Forms, Functional Affordances and/or Constraints
Shavon: [EMP] @Then no,@ @'[email protected] @I need the five hundred dollars to raise
taxes for the [email protected] @[email protected] [EMP, gesturing with both hands, then (playfully)
shoves Chillbrisha] (Speech Act + Booster by Gesture & Intonation) (PAL-SFY;
Danielle: @Yeah,@ @[email protected] @I [email protected] it’s Stone Fox @[email protected]@[email protected] him have
one of his dogs and
Kevin: No,@ two, @[email protected] he has four dogs @[email protected] there.
Shelby: No,@ he has five. @[email protected], one… [Shelby points to a picture cover of the story
and finger-counts dogs] (Gesticulates Evidential: From sb else) (CHA_8SFY;12:48.64)
Gestures frequently combine functions with other categories (double-coding):
With Talk-Organizers (indicating for maintenance of talk)
Talk-Evaluating (Nodding Yes/No)
Evidentials: From Someone Else (indicating sources/basis for knowledge in text,
Commentary directive or interrogative (indicating with head/hand/finger “you”
addressee and imposing perlocutionary force/ prompt action or re-action)
Markers (I-S)
Or otherwise
Hasund, 2002)
„or sth“, „and
things/and stuff
(like that)”&
nell, 1994, w.
British English
- “You know” = (1) implicitly invites solidarity (in pragmatics represent a strategy of
positive politeness, e.g. Brown & Levinson,1987; or Yule, 1996) as a basis of “saying
less by communicating there is more” so similar to Glossing: Explanation
- (2) Intersubjectivity MARKER to assume the common ground; completion is
assumed to be known by the Listener (Y’know and stuff “On being inexplicit and
stuff in contemporary American English”, Overstreet & Yule, 1997).
- “Like” use for exemplification functions as “for example”, functioning as a Speech
Act + T-O, as ensuring progressivity of “talk in action”. For example, “It’s like, they
are all stupid and stuff” (Homer Simpson) – Sloppy talk (cited after Overstreet &
Yule, 1997)
- Perfomance Fillers of adjunctive (“and stuff”) or disjunctive (“or something”, “or
anything”) type that
o appeal to Listener to construct referential category
o implicit appeal to shared experience or knowledge
o signal social closeness, and so in turn, are “indicators of
(table continues)
Table 10 (continued)
Metadiscourse Metadiscourse Forms, Functional Affordances and/or Constraints
intersubjectivity” by sharing individual, subjective co-conceptions of the world
with our interlocutors (Schegloff, 1992; Schiffrin, 1990) >>>Thus, oftentimes
Fillers+ Other cautious elements: Hedges or Glossing: Explanations)
“Like”, apart from being an obligatory pragmatic marker in a quatative
complementiser function “I went like…” or in “analogy” structures may have a
meta-representational function (Stenström, Andersen & Hasund, 2002) , in “I was
like…” marking off the following material as a thought, an attitude or a feeling
which is metarepresented (p.116). (Logical-Temporal Connective)
o An approximative function = similar to “kind of, sort of” (Glossing:
o Exemplificatory function (“for example” so a Speech Act, in Vande Kopple,
o Hesitational/linking (Other cautious elements: Hedges)
o Metalinguistic (ML, when word-searching)
o Phatic Function = Filler
“Yeah” - INTERACTIONAL /Interpersonal Functions:
imagination appealing and a concept-retrieval helping (Reminder of Material)
even in the mid-sentence position, or checking “Are you following my story?”
(Speech ACT + T-O, e.g. “Yeah, but…”),
or Mitigator/Hedge at the end of the sentence, question tag “yeah?”/ “right?”
Commentary when appended to Yes/No-Questions serving as a further urge for
the hearer to respond: “Is it $300, yeah?”
or “an epistemic tag” (pronounced with rising intonation) appended to a
statement or after interrogatives/imperatives “that serve[s] to engage hearer or
invite his response in the form of a confirmation, verification & corroboration of a
claim” (Stenström, Andersen, & Hasund, 2002, p.167), checking if the hearer is
“getting” the preceding conceptual information; close in meaning to “you know
[what I mean]” (p.175) – so coded as a Pragmatic Intersubjectivity Marker, as
“indicative of the speaker’s presumption that interlocutors’ cognitive background
consists of mutual assumptions” (ibidem)
Talk–Evaluative, “is it OK if I continue”
Reception marker – to acknowledge receipt of info that is new to the discourse
but consistent with currently active info (Jucker & Smith, 1998, p. 179), or to
acknowledge info that is re-activated (refers back). These uses illustrate “the
principle of grounding” proposed by Clark (1994; 1996) that a contribution must
be acknowledged by the partner to be complete.
(table continues)
Table 10 (continued)
Metadiscourse Metadiscourse Forms, Functional Affordances and/or Constraints
Likewise, “Okay” as a tag!
Or “Right?” Commentary Interrogative & Talk-Evaluating here:
Tabitha: I just thought of something um @ it [email protected] it was taxes @[email protected]
there’s going to be more taxes, @[email protected] @[email protected] @ how can he @[email protected] keep
his house because he’s going @to have [email protected] pay taxes @over and over
[email protected]?! [Gesticulates and Emphasizes by raised Intonation over “taxes over
and over again”] (CHA_8SFR: 19:17.00)
“Yeah” and “Right” TEXTUAL Function – ability to chunk sequentially related pieces of
- Plus add to the structure of a narrative (Stenström et al. 2002, p. 167)/ thus
ensuring progressivity of talk, so Talk-Organizing MD,
or sometimes both T-Evaluating and Spech Act (combined), when it stands alone as a
“response cry” (Goffman, 1978) in meaning “I agree”, or like Fraser’s (1990)
interjections (e.g., “oh”, “aha”)
Density and
of Forms &
“Well” functions as Topicalizers:
Andy: Well, @I [email protected] with Cody @[email protected] the Indy, white people would @[email protected] be
losing a little bit of land @[email protected] the Indians @[email protected] {hedge} the Indians
wouldn’t be losing anything @[email protected] the farm isn’t on the land that the
Indians would buy. (CHA-8SFR; 9:16.62)
as Contrastives (a ‘weak-token’):
Shelby: No, he has five. See, one…(gestures)
- Kevin: Well, it looks like he has four (CHA-8SFY; 13.02.70)
as a Hedge (other cautious elements):
- Shelby: Why do you say that? (Commentary_Interrogative)
- Kevin: @Because,@ @[email protected], Indians usually, @[email protected] long time ago, Indians used
to @[email protected] kill each other @or something,@ @I [email protected] @[email protected] @well, @like
they [email protected] he better win, @@he said, he was like @[email protected] trying to [email protected]@
@[email protected] @if it didn’t look like he let them win, [email protected] they wouldn’t @[email protected] be
really mad at him @or [email protected], @[email protected] he wouldn’t lose all his
friends. %@It’d look like it was @[email protected] close to a tie, @[email protected] @[email protected] it wouldn’t
look like he let them [email protected]% (CHA_8SFY)
“Now”- Logical-temporal Connective (1) in general, and/or
- (2) Contrastive, or
- (3) Reminder of material (tying back), as in here:
Danielle: [email protected] if, @I [email protected] that they @[email protected] @[email protected] the dogs @[email protected]
run off now(1), @[email protected] Stone Fox has @[email protected] four or five dogs @[email protected]
Willie @[email protected] has one.
Jeremy: @[email protected](2 &3) he has none.
(table continues)
Table 10 (continued)
Metadiscourse Metadiscourse Forms, Functional Affordances and/or Constraints
“Oh!” – indexes the shift in knowledge – from Knowledge (K-) to Knowledge (K+) to
use Heritage (1984) terms or otherwise adjustment or acceptance of
“epistemic gradient”, i.e. a knowledge gap between a questioner and
respondent (Heritage & Raymond, forthcoming, p. 4):
- Cody: She is talking about Stone Fox.
- Richard: @[email protected] @you’re [email protected] about Stone Fox. (CHA_8SFR; 2:20.63),
“Oh” to be coded as Talk-Evaluating.
Issues with
Double coding
A: Yes
B: No
A: Yeap
B: Nope
B : No. / Yes, he does/ Yes, he can. No, he doesn’t/ can’t.
“Yes/Yeah/Yeap/” or “No/Nope” considered as Speech Acts (=Illocutionary Act
Markers of Affirming or Disconfirming) and double-coded with Talk-Evaluating
– in sequential context reflect attitude : positive = I agree, or negative = I disagree
(yet not in Searle’s decontextualized understanding; after Cutting, 2002, p. 125,
paraphrasing Heritage’s point), and in that way reflect Talk-Evaluating:
Katie: @[email protected] Do you write in math? [EMP] [email protected][email protected]$ [EMP]
(“Yeah”- T-O ; “Yes!” Speech Act + Booster (by exclamation =!)
if EMPHASIS added in Intonation (!) than Boosters (as above)
if SOFT-spoken than Mitigated Speech Acts
yet, if directing talk and ensuring its progressivity, preferred over dis-preferred
silence or lack of response then function as Talk-Organizing metadiscourse
category that is when “Yeah” does not really mean the agreement but rather “I’m
getting you”
e.g. “Yeah but…” (T-O)
Katie: Yeah but the thing is you need to practice writing… and learning (SpAct+TO)
Dylan: Writing?... They’re s’pposed to be a math er-… [taps text] @[email protected] a
math program. (Speech Act+ Talk-Evaluating) (PAL_5MVR: 15:31.51)
Tabitha: They might not have had television.
Richard: Yeah. They probably didn’t have television back then cause they didn’t cause
it wasn’t on they didn’t say anything about on the news or anything. (Speech
Act) (CHA_8SFR: 17:08.51)
(table continues)
Table 10 (continued)
Metadiscourse Metadiscourse Forms, Functional Affordances and/or Constraints
“As you/Andy/X said”… “and I agree with you” - Reminders of Material (Textualorientation), and so double-coded with T-Organizing (as the bracket sub-component)
Also often as “Hearsay” Evidential (showing basis for epistemological stance in
referential material)
Shelby: Danielle, you said that if you took the CD disk out, it would break and I
agree with you because books wouldn’t like break automatically or something, if
somebody accidentally stepped on them, but if you probably stepped on a CD, it
might break, and I agree with you. (CHA_5MVY;5:19.46)
“You”= meaning Everybody functions as “Others”, so depending on the contextual
cues can be coded as ‘Others in scenario’ or ‘If-then…” deductive/ hypothetical
structures with Modal Verbs (can/could, will/would, may/might, shall/should)
Kaycee: If the electric went out, you would not have a computer you could go home
and the electric, you can do your homework at your house, and with a
(Emphasis] mathbook [Emphasis], and I’m, and some people don’t have
computers at their house, and um (CHA_5MVR: 7:23.21; while voicing
inferiority of computers to textbooks)
Equally in Aposteriori/Retroduction Intended structures “You would’ve+ V-ed”,
conventionally indicator of a Commentary Vocative or Interrogative (‘readeraddressing’), as in writing.
Table 10 encapsulates some crucial qualitative findings with respect to the written and
oral diversity of meta-discursive forms functioning. The table also includes systematic
attributions of discrete categories to major codes. Plus, it reveals some affordances and suggests
possible constraints in the qualitative and quantitative analysis, which will be discussed in
Chapter 5.
Study Goals #3 and # 5 addressed
Study goal (3) was set to establish any correspondences between children’s metadiscourse in two
modalities, and study goal (5) - to determine how they add to the flow of meaning in their oral
and written argumentation.
As a reminder, the bracketing broad categories comprised codes on the basis of common
meta-communicative functions that have been theorized to be of organizing, evaluating and
intersubjectivity types. In effect, the supra-categories of text- /talk-organizing, text-/talkevaluating or intersubjectivity marking show a ‘rough orientation on the map of the texts’ or
discourses when shown in print, and as a reprint of coding.
Let me illustrate the effect of the application of the three-partite scale of meta-discursive
categories to both modalities in order to (1) show successful uses of metadiscourse that help the
flow of meaning in argumentative talk or text, and (2) those that impede it – in respective
Helpful Uses of Written Metadiscourse
The first essay [1] with response to the big questions Should Jack tell on Thomas, bearing
anonymous ID number 209, falls within the shorter essays range, as it has 89 words, and on
average girls’ essays had 115 words while boys’ – 60. Thus, the essay is representative of the
sample’s pool average, and in no way resembles lengthy essays amounting to 250 or 300 words.
[1] ID #209
No because Thomas finally has someone to tall to without pushing or shoving. Jacks mom said
be kind to whomever doesn’t have any and telling is sure isn’t nice in Thomas’s mind. Jack said
to himself that maybe Thomas never won anything in his life. In the story they still shook
Thomas’s hand and said “what a fast car” he, had Thomas was so happy so I’m thinking he
maybe never won anything in his life. That is why I think No Jack shouldn’t tell on Thomas.
Italics: Modality Markers (Hedges/Emphatics) = part of T-Evaluating Bracket
Italics+White Type (color reversal) =Text-Evaluating Bracket
Grey Highlight =Text-Organizing Speech Acts and/or Induction, here
Single Underline = Hearsay Evidentials “From someone else”, inter-textuality organizing
Essay 209 [1] shows the Text-Organizing uses of metadiscourse (as marked with grey
highlight) by signaling: the changes in the point of reference for evidence, whose source the
student found in the story (for example, “in the story they […] said”), and thus reminded of the
material that is being discussed or at stake, thus adding to Intertopic coherence, which is created
when students make linkages between old and new topics (explicit referentiality and intertextuality function of metadiscourse, as argued by Almasi, O’Flahaven and Arya, 2001 ). Also,
Text-Organizing bracket comprises here performative Speech Acts of declaring his/her position
in an assertion “I think” and “I am thinking”, as well as in meta-communicating a reasoning type
called Induction that helped him/her arrive at the final decisions following the evidence
presented, i.e. in “so” and “That is why” preceding the performative Speech Act of asserting. As
can be easily seen, the Text-Organizing almost brackets (except for the initial position) the flow
of information and transactional meaning therein. The various instances of text-organizing (by
speech acts, induction or evidential marking) also help signal the direction of the flow of
The Evaluating Bracket (italicized in [1]) is both at the text initial and final positions as
well as interspersed throughout the text. The bracket encompasses not only an evaluative speech
act of disagreement (“No”, which is equivalent to “I don’t agree/ I disagree”), as well as
instances of Hedges (“maybe”) and Emphatics (e.g., “so happy”), together forming a major
category of Modality Markers, and a “positional functor” (after Wünderlich, 1979), revealing the
child’s attitudinal stance by the use of a modal verb of Forbidden category (shouldn’t). There are
no instances of Intersubjectivity marking though.
In conclusion:
At a glance at essay #209, it becomes evident that the use of both evaluating and
organizing bracket is more or less balanced; when raw counts juxtaposed, Text-Evaluating
covers ca. 47%, and Text-Organizing - 53 % of metadiscourse uses. Upon reading, it turns out
that metadiscourse as used by student ID #209 helps the flow of propositional meaning rather
than impedes it. It is just the right side of attitudinal Text-Evaluating (bracket broad), and TextOrganizing metadiscursive bracketing of propositional meaning. Even though it is not among the
most rhetorically rich in text-organizing devices from the analyzed sample, the student’s use of
metadiscourse still helps its proper propositional flow.
Impeding Uses of Written Metadiscourse
Another essay, ID #111 [2], is a bit longer as it has 149 words. It is neither exceptionally
long nor exceptionally well written. Nevertheless it has been selected upon inspection of other
essays for illustration of the metadiscursive uses that may impede the flow of propositional
Yes, I think that he needs to tell on him because, he didn’t make it on his own,
his brothers did. Mr. Howard said to do it on his own. His brothers made sure everything was
perfect even the wood, they made it smooth. The tires were even straight and stuff but
everybody else’s were kind of crooked. Jack has to tell because it wasn’t fair to the other kids
that were in the race. If his brothers were in it they could have done it but it’s not fair to the
others because he didn’t do it right like Mr. Howard said to do it. Even though it’s neat don’t
mean you don’t have to not tell on someone even though it’s neat you can’t even resist it you
need to still tell no matter what happens still tell on him because he didn’t do it by himself.
Plain Bold - Intersubjectivity/Vagueness Bracket, e.g., Code Glosses
Double Underline - Emphatics Misc. & Attitude Markers, under Text-Evaluating Bracket
Black Highlight & White Text - Text Evaluating Bracket by Evaluative Speech Acts
Grey Highlight - Text-Organizing: performative Speech Act, or evidential
Discrete Underline – Obligation Markers/Positional Functors, within Text-Evaluating Bracket
The concentration of the black highlight (w. color reversal), including double-wave
underlining in [2], indicate evaluative metadiscourse uses. Specifically, numerous instances of
Emphatics (even, even though, still [tell]) that are accompanied by Attitude Marking operators
“it wasn’t fair” and “it isn’t fair” along the evaluative Speech Act “Yes” (signaling ‘I agree’ as
an answer to the Big Question in the prompt), or code glossing “don’t mean [you don’t have to]”
as well as multiple deontic modality verbs of mild Obligation “has to” or “needs to”, all
contribute to the Text-Evaluating bracket. Though differentially, all reveal the attitudinal and
evaluative stance of the writer.
A couple of Intersubjectivity-Marking by pragmatic vagueness signaling code glosses by
means of “kind of” and general extender “and stuff” (both suggesting how loosely or how
strictly/ precisely to interpret the propositional material to follow) add some interpersonal metadiscursive meanings rather than the organizing ones (similarly to the evaluative bracket).
On the other hand, there are few instances of Text-Organizing that include a performative
Speech Act “I think” and a reference to a “hearsay evidential” labeled in the adopted taxonomy
as From Someone Else - “Mr. Howard said” - that reminds of the material that contributes to the
topic or theme of reflective discussion in writing. Evidently, essay #111 suffers from the
underuse of metadiscourse meta-functioning as Text-Organizing as compared to the Evaluating
meta-discourse. The reader receives too many signals of stance manipulating evaluatives – be it
signals of emotional orientation towards the propositional material or epistemic modality
showing degrees of commitment to the truth value of the propositional material, or even of
deontic modality that signals the degree of “binding” (in Greek deontic logic means “of that
which is binding’). Such numerous uses may seem excessive, too pressing in the subjective
marking and even result in readers’ developing resistance forewarning effects to such
manipulative attempts, as found by Kamalski, Lentz, Sanders and Zwaan (2008). They argue that,
similarly to some blatant uses of coherence markers, subjective marking may “cause readers to
recognize an attempt to influence them, [then readers] build up resistance and it becomes
difficult to persuade them. Also, subjective marking causes readers to recognize the persuasive
author’s intent more easily. Furthermore, subjective markers seem to cause resistance to
persuasion, whereas objective markers improve integration of information (p. 545).”
To conclude:
As visualized above by the concentration of black highlights and double underline,
evaluative metadiscourse uses in essay ID #111 [2] prevail (ca 80% as compared within the
group of 3 major brackets), to the detriment of text-organizing (approx. 10%), which thus leaves
the reader with too little textual guidance, that is one of the meta-functions (to guide the reader
rather than inform him/her), and possibly with resistance to persuasive force of the propositions
put forward. In result, the Text-Evaluating bracket’s generous use in the presence of
parsimonious use of Text-Organizing categories of metadiscourse may be viewed as an
impediment to the flow of propositional meaning rather than its facilitator.
Helpful Uses of Oral Metadiscourse
The dialogic episodes that were coded using a similar taxonomy, yet with gestures and
intonation added as only aural-oral modality specific, show plentiful instances of helpful and
impeding metadiscourse. First, let me demonstrate how metadiscourse may help the flow of the
propositional meaning.
Since in oral discourse much of the meaning resides in context (not only in the text), let
me briefly sketch out the context of the discussion #8 that is based on the Stone Fox story. CR
children are discussing the Big Question whether Stone Fox should let Willy win? In the story,
the little boy called Willy participates in a dog sled race where he is pitched against Stone Fox,
an Indy who needs the prize money for the land for his people, while little Willy needs the
money to keep the grandfather’s house, which otherwise may be taken over for taxes. Just when
Willy’s sled is about to win, his dog dies just a few feet before the finish line.
At this stage of discussion, Teacher announces wrap-up and children are to “say’ their
votes, i.e. positions with respect to the Big Question, and their “strongest reasons” why they
think so. Following her three peer “votes” (each of which shows numerous uses of
metadiscourse), Victoria (fictitious name, here) states her position followed by justification (as
cited in [3]). The transcriber’s marking of /numbers/ indicate instances of other kids’ overlapping
speech, while [GEST] within square brackets indicates onset and the end of Gesture, which is a
coder’s comment external to the talk.
00:24:54.50 *Victoria
I say Yes because kids aren't like grown ups, grown ups |1|
|1|... well if, if you're old, then you would die faster out in the cold, but a kids, they an't, they...
ain't like a parent,|2| |2| and um, a kid doesn't need to be out with no home, and he can get really
sick and um, like [GEST]die from being sick[GEST], and he could freeze to death, and um, and
um, and so the Stone Fox he if, if, he ain't happy what he gets then that's not what he gets. |3| He
can be better back in Wyoming |3| and not |4| |4| Mis- Mis- Minnesota. And um, it's Yes.
Grey Highlight - Text-Organizing: performative Speech Acts, Topicalizer, Induction
Bold Italics – Emphatics, either Miscellaneous or Emphatic gesture, subsumed under Talk-Evaluating
Broken Underline & Italics – Hedges, Other cautious elements or Mitigators - under Talk-Evaluating
Broken Underline – Evidentials signaling inferencing by “If-Then”
Bold Underline – Evidential by Induction/ Deduction, also Text-Organizing Bracket
Dotted Underline – numerous Text-Connectives, organizing talk locally
Double Underline – Topicalizer, Text-Organizing Bracket
In Victoria’s turn [3], there is an evident Talk-Organizing Bracket sandwiching the flow
of propositional meaning between the initial and terminating Speech Acts of stating and restating
her position with performative verbs “I say Yes”, otherwise - “expositives” in Austin’s (1962)
classification, which make clear how our utterances fit into the course of conversation or an
argument (e.g., I reply, I argue, etc.). Her position being in favor of the presupposed Big
Question (whether Stone Fox should let Willy win) is then followed by the reason signaled by an
inter-clausal connective ‘because’ (local coherence and cohesion marker, like other connectives
used there, incl. “but” or “and”). In the initial position of the next sentence she chose to use a
Talk-Organizing element by Topicalizer “well”, which “focuses attention” on the phrase that
indicates the shift, re-introduction, changes in topic or is brought up “to set a particular contrast
in stark relief” (Vande Kopple, 1997, p.2). Here, the topic is understood as being not only “what
is being talked about but what is being said about what is being talked about” (after Cutting,
2000, p.27; emphasis in the original). With those explicit performative Speech Acts, Topicalizer,
evidentials by inductive (“and so”) and deductive reasoning structures (along “If-then”), which
help shape discourse and reasoning in the logical Modus Ponens frame, the flow of propositional
meaning is evidently facilitated by the organizing markers.
Evaluative metadiscourse in Victoria’s turn includes a few instances of modality marking
with a hedge (Other cautious elements) as well as a Mitigator of a Speech Act, emphasis adding
emotional Gesture of ‘despair’ and an Emphatic Miscellaneous (“really”). Thus, the mere look at
the proportions of organizing and evaluative metadiscourse reveals an optimal balance between
the two major brackets (maintaining balance between objectivity and subjectivity, or otherwise
neutrality - necessary for integration of information/ knowledge).
Impeding Uses in Oral Metadiscourse
Some preceding lines to Victoria’s turn [3] and so leading to the aforementioned episode,
there are a few turns of Lauren and Mary’s in fragment [4], which when scrutinized reveal some
impeding uses of metadiscourse.
[4] 00:23:44.70 *Lauren Um, I say yes, because if Willie doesn't get the money, he could be homeless,
live on the streets, or whatever. And Stone Fox, if he doesn’t win it, he still has his own land in Wyoming.
He wouldn't be as happy as he would in Minnesota, but...I don't know, I'm changing again... I can't Now
I'm No because... I mean,.... people forced to no, now I'm maybe.
00:24:19.50 *Mary
I'm a yy- |1| I'm a y- |1| Yes because I mean, I don't think Willie really want to
be livin on the streets or anythin, and I don't think it'd be really nice...and he doesn't want to go to an
orphanage cuz he can't um (Logan sighs), gran-....granpa can't pay the taxes and I Oh no...Lauren ......
00:24:20.80 *Vanessa
|1| You two come on |1|. (Lauren Laughs).
00:24:51.50 *Lauren
00:24:52.00 *Mary
I'm changing! (Mary and Lauren laugh, continue laughing while Victoria is
Performative verbs that are labeled Speech Acts in Lauren’s turn that most often serve as
either initalizing or terminal bracket, apart from an opening bracketing use “I’m a Yes”, are
condensed towards the end, when she announces 3 different positions. In those speech acts she is
‘performing’ a point making and also externalizing her sources for thus performed claims, which
lie within her beliefs like in the Speech Act of thinking (“I don’t think”). There is some
propositional flow of meaning which includes an attempt at elaboration of her point-making in
which she explains that she changes “because people […] forced to no”, still broken by a
glossing insert of “I mean” (Talk-Organizing, NB). Thus, the turn seems too packed with
metadiscourse impulses of Speech Acts or glossing (“I mean”) that are unaccompanied by
propositional elaboration and evaluative signals. In essence then, bearing in mind Petty and
Cacioppo’s (1979; 1980; 1981) argument that elaboration and persuasive appeals foster
engagement in issue-relevant thinking and add to the psychological cogency of persuasion, the
fragment loses on the persuasiveness value rather than gains it due to too much of metadiscourse
with too little elaboration for the given context.
A more explicit case of over-stimulation of the talk flow with bare repetitive Speech Acts
is demonstrated in Mary’s following turns, where metadiscourse in fact adds to the wordiness
and repetitiveness of talk and so rather impedes than helps the flow of propositional meanings.
Specifically, in the following turns Mary only spells out the performative Speech Acts in a
repetitive fashion with no elaboration or justification in the propositional material. Except for
self-deprecating exclamative “Oh, no…” in an affect display and two instances of Emphatic
Miscellaneous ‘really’ (which confirms the girl’s specialty as suggested by Stenström et al.’s,
2002 findings in the corpus of COLT, i.e. the Bergen Corpus of London Teenage Talk), but not
much other evaluating and metalingusitic metadiscourse going on. Mary’s talk reflects somewhat
“interpersonal semantics” (Eggins & Slade, 1997) in an affective display of dissatisfaction with
her own speech and possible negative evaluations by others (in fact affectively displayed by an
outburst of laughter) but there is little metacognitive and metalinguistic control over talk,
generally executed by evaluative bracket, for example by mitigating, hedging or emphatic
devices, or attitude marking via normative or motivational verb operators (by reference to
Wünderlich’s, 1979, conception of attitudinal stance marking by so called “positional functors”,
including should, shouldn’t, mustn’t or could/ couldn’t, etc.).
However, metalinguistic communication is not always helping the flow of the discourse
proper. To illustrate the phenomenon, let me cite a few lines [5] from the same discussion, just
further following Victoria’s turn, Lauren keeps repeating “ain’t” forms (heard in Victoria’s talk)
and continues to laugh, both having a distracting effect on the conversation about the topical Big
Issue. Then he ventures:
[5]. 00:25:53.50*Lauren Just a question. Could you refer ain't as another word? Ain't. You always say
00:26:00.00 *Mary
|1| (I don't use it)|1|
00:26:00.20 *Mary
What is it in the dictionary for ain't?
00:26:02.00 *Mary
Which in the dictionary you know.
00:26:03.80 *Lauren
I don't know.
00:26:04.80 *Teacher
There's lots of words in the dictionary that aren't allowed in school though.
00:26:09.00 *Mary
Like ain't? (Looking at Teacher)
00:26:10.50 *Teacher
00:26:12.00 *Victoria
Ain't no problem with that! (laughter)
The above meta-communication about the usage of “ain’t” in [5] is side-tracking the
main conversation topic of saying positions about (what is being said about) the Big Issue, thus
impeding the flow of discourse proper. Therefore, for some rhetoricians the conversation’s goal
not being achieved in [5] and Grice’s maxim’s of quantity (be as informative as required) and
relation (be relevant) being infringed especially in Victoria’s final answer, the exchange may be
viewed as not successful because it does not bring in much of propositional meaning into
Nevertheless, it is an interesting and rather rare phenomenon that 4th-grade students
indulge in the communication about norms regulating their talk. CR-group discussion here more
resembles a court situation where Lauren’s seemingly simple question about Mary’s “ain’t” form
in fact implies an accusation of a normative nature, which is intensified with Attitude Marker
Miscellaneous (in the study’s taxonomy) by “always [say ain’t]” thus alluding to what is
socially viewed as proper or improper in talk. This metalinguistic comment carries a baggage of
cultural, ethnic, in-group expectations and evaluations, and is contingent on social patterns.
Mary’s final response in an evaluative Speech Act (Talk-Evaluating bracket) that is intensified
by raised volume and prosodic changes and phonetic contour (Emphatics by Intonation) that
repeats the criticized form “ain’t” (in saying that she doesn’t have a problem with that) metacommunicates her attitude towards the implied accusation and her lack of conformity. From the
Speech Acts theory perspective, her remark is not relevant in the sense that it should be relevant
in relation to another remark (cf. Smith & Wilson, 1979; or Sperber & Wilson, 1986) or here – a
question by implicature “why she is always using ain’t”, which in fact meta-communicates
Speaker’s problem with that not hers, as she is the one who freely uses the ‘problematic’
linguistic form.
However, Mary succeeds in terms of communicative competence by recognizing the
intentions of Lauren’s speech act and by performing her own speech act adequately to the need
of preserving her own public “face” (“self-image that every member wants to claim for himself”
in Brown & Levinson, 1972, p. 62) in the inner desire for freedom from imposition (“negative
face” preserving, cf. my Chapter 2). Again from socio-pragmatic perspective, Mary’s response
threatens her own positive face (need for social approval) when she meta-communicates her
intentions not to converge her ‘idiolect’ to the standard of talk to which Lauren alluded (and
which accepts other words rather than “ain’t”). Thus, Mary indirectly communicates that she will
not accommodate to the interlocutor’s imposed standard (thus defying Accomodation Theory
standards). By doing so she reveals her dissociation from the recipient (here, Lauren) and his
alluded norms and consequently - the need for public face acceptance, or otherwise being liked
by the recipient, her classmate named Lauren.
All in all, her final response not only succeeds in meta-communicating her emotional
attitude (emphatics, part of Talk-Evaluating bracket) to the ‘topic’ raised by Lauren, but also
meta-linguistically (“problem”) and by clever use of anaphoric “that” (backward reference)
signals how she is using those words like “ain’t” (that she ‘ain’t have a problem with that!), that
is without a problem. In other words, by combined uses of metadiscourse stance marking, a
speech act combining metalinguistic term “problem”, seemingly superfluous repetition of the
‘ain’t’ form and co-textual anaphora (with the referent in an ‘always use of ain’t’), Mary aptly
renders a counterargument and rebuttal in one turn, of the negative view of using idiolect form,
without losing her public face and endangering group integrity. Thus, her successful refutation
of the criticism cannot be overestimated in argumentation.
To conclude:
This meta-linguistic exchange initiated by Lauren may also be seen as helping the flow of
meaning, however not much of the topic-relevant ideational flow, as the topic was about stances
to the Big Question stating, but the pragmatic (socio-pragmatic) and interactional (part of
communicative competence; cf. Celce-Murcia et al, 1995; or Markee, 2000) flow of meaning.
Other metadiscursive uses impeding flow of meaning of discourse proper oftentimes are
demonstrated in heated parts of small-group discussions where emotional factors come to play
and children in fervor of the conversational persuasiveness use all sorts of metadiscursive
evaluations as well as vague intersubjectivity pragmatic markers.
The following brief excerpt [6] from another discussion – about a vote for either
computers or textbooks to be bought for a school (Marco’s Vote) - illustrates such impeding uses
of intersubjectivity markers “like”, “or something”, “you know” or “just”, whose meaning so
much resides in the context of use.
00:02:18.17 *Jordan I think that computer would be more helpful coz you could just do like
math on the computer like if you have like some math problem disc or somthin you could just put it and
do it on the computer instead.
[…] (an insert of a few turns of other discussants)
00:03:15.89 *Cory
I agree, because you know, I can (??) the computer for more money so
like if you go to WalMart or something you can buy some (cans) for 69 cents or something and you can
buy like a (Gol car) or something which you can make more use out off.
Both Jordan’s uses of “like” and Cory’s uses of “or something” in excerpt [6] appeal to
shared knowledge basis, which are so much important for the in-group behavior signaling in face
to face confrontations (NB, considerably much higher than in students’ writing). By roughly
meaning ‘how loosely’ (rather than strictly) to interpret the following propositional content, the
uses have basically a function of Code Glosses (Explanations - in the adopted taxonomy). Also,
the notorious “like” due its numerous functions (as delineated in Form to Function Table) is
rightly used as a quotative complementizer for instance in “I went like…” or similarly in
analogies where it has a meta-representational function (when it is coded as Logical-Temporal
Connective Miscellaneous; cf. Stenström et al., 2002), or for exemplification purposes (similar to
Speech Act “for example”). But in adolescent uses as those in [6] it can easily be removed from
the talk, and the flow of discourse proper would not suffer from that. However, the content may
sound somewhat artificial as the authorial voice or “subjective” component would not be
authentic – not aligned with the authentism of the writing persona (cf. Crismore’s ,1985 study
into subjectivity of metadiscourse in social science books).
This seeming redundancy may prove vague and superfluous contribution of “like” to the
propositional flow of meaning, and similarly other inter-subjectivity pragmatic markers (nota
bene, grouped in this study under one bracketing category). However, their contributions to
‘social closeness’ signaling or ‘social lubrication’ (Overstreet and Yule, 1997), and indication of
inter-subjectivity (by sharing individual, subjective conceptions of the world with interlocutors in
Schegloff, 1992; or Schiffrin, 1990) cannot be over-estimated. Uses like these are especially for
interpersonal flows of meaning, for ‘face saving’ and in-group appealing, when venturing a face
threatening opinion in the heat of arguments or in signaling a non-flouting attitude to the group
conformity, and consequently - for avoidance of social stigma (cf. Eckert, 1988, 1997;
Stenström et al., 2002).
Overuses of metadiscourse can also be found within the Talk-Organizing Bracket – not
only within evaluative or intersubjectivity- vague marking brackets. In example [7], the same
discussion about Stone Fox ensues, and one of the adolescent discussants takes up the idea of
little Willy boy ending up in an orphanage when his grandpa dies.
[7] 00:26:15.50 *Lauren
00:26:16.50 *Steven
Yes? Because um, it wouldn't be good for him to be um, for Willie to be
homeless. And then Stone Fox's might be a bit crowded, but next year Willie probly wont' be able to
enter cuz he doesn't have a dog. And then Stone Fox could just win the race for the next few years. And
then eventually get enough money to get back. [During Steven's turn Maisie and Lauren were engaged
in a side talk, most of the time, that was inaudible]
While the Speaker signals the temporal projection of propositional meaning into the
future (next year, next few years), he also signals logical-temporal continuity and sequencing of
ideational flow with “and then” and “eventually”. Thus, in an attempt to maintain the discourse
proper flow in the temporal/ sequential rhetorical pattern the signals “and then” seem most
relevant and facilitating the organization of talk; besides, indeed Talk-Organizing uses (60%)
exceed those of Talk-Evaluating bracket (40%) in this single turn.
The student, though, uses them for any additive relation, in fact in place of “and”, and in
result binds the stretch of talk (here, a hypothetical deduction-type, while putting “Others in
scenario”) in a more additive-like manner. The temporal-sequence signals (like those of globally
binding sequencers “first”, “then”, “next”) seem to lose their potency, as the Hearer realizes their
additive or even coma-like role. Evidently, uses of “and then” in a similar fashion lead to their
‘de-ranking’ from more globally binding sequencers to less-global as inter-clausal binding of
logical–temporal connectives type. The phenomenon illustrated in [7] has been documented in
other studies too, for example in studies of pseudo-bridging “so” uses (devoid of causative
function) or in studies of narrative cohering in African-American writing or talk with frequent
“and” followed by “then” (McCarthey, 2002; Michaels, 1981), often co-occurring with a topicassociative pattern of narration.
In summary:
The ‘bracketing’ role of 3 supra-categories helps to demonstrate the overall organization
and proportions if not valency of meta-discursive uses in children’s stretches of talk or text. The
form-to-function compiled table helps to demonstrate the systemic patterns for functions to be
carried out by particular linguistic items, which reveals their distinctive categories that are in turn
used for grouping into classes or broad brackets of major meta-functions.
The fine-grained codes of a very discrete nature allow for identifying the immense
richness of functions that various forms of metadiscoursal elements can fulfill. The discrete
codes also allow for indexing how they feed into the bracketing meta-functions, and tracing the
interlacing of functions between themselves (organizing, evaluative, intersubjective,
interpersonal, etc.) and linguistic forms. In this way, the interlacing between the metadiscursive
uses and the flow of propositional meaning (discourse proper) becomes more ‘visible to a
researcher’s magnifying glass’ and leveraged to a due scrutiny of talk-in-action - not only within
turns but whole flows of discussions per se, and depending on the context of use by CR-exposed
For optimal qualitative and quantitative effect of this exploratory study, it is believed that
a combination of broad scales with the fine-grain adding discrete micro-categories ensures
greater informativeness of results and determination of systematic patterns of metadiscourse uses
(as the outcomes of the present study suggest).
Chapter 5
In pursuit of the diverse study goals from (1) the analysis of discussions (2) via essays
and (3) their correspondences to (4) socio-linguistic variation and, finally, (5) the explication of
successful and impeding effects of metadiscourse, the multi-fold taxonomy has managed to
capture systematic differences in the functions of each metadiscourse category.
Thus, the adoption of the complex hierarchical structure or a functional grammar of
metadiscourse used in spoken and written modalities afforded results usually available in corpus
linguistics studies.
The two different tasks that utilize two different modes of communication - speaking and
writing - show different functional patterns of metadiscourse, as used by young adolescents
participating in the Collaborative Reasoning paradigm.
Spoken and Written Modalities
Overall findings for discussions (as shown in Figure 3) indicate that Speech Acts
(performative verbs naming actions being performed) are the most frequent category irrespective
of modality. This finding corroborates results from the corpus linguistic study by Biber et al.
(1999) that investigated overall distribution of the most common verbs in British English in
different registers: conversations, newspaper language, academic prose and fiction on the
Longman Spoken and Written English Corpus. Biber et al. (1999) found that among the 12 most
common lexical verbs are say, get, go, know think, see, or mean, and they occur from around
9,000 (get) to 1,000 times per million words with “say” (about 6,000) reaching the top frequency
across all registers. These high frequencies show the Speech Acts to be ubiquitous whenever
speakers or writers want to report something. Their top-ranks also reveal communicators’
intentions to make those speech acts explicit enough to immediately impact the recipients with
their illocutionary force. Thus, the distribution of metadiscursive discrete category of Speech
Acts that hinge on lexical verbs (such as I say/ I pick/ I think ‘yes’) should not be surprising,
though it is good to see that the results from this study with young adolescents converge with
other researchers’ findings based on adult use.
Most of the discrete frequencies and those in classes or brackets, as mentioned earlier, are
relatively high compared to the most common lexical verbs documented by Biber et al. (1999).
Therefore, the observed high frequency distribution of textual connectives, comprising of
additives, causatives, contrastives, and logical-temporal, at a similar rate as some of the
documented most common verbs, is especially important as it shows a pattern of surface binding
by CR speakers, which results in greater cohesion of their talk, and high inter-clausal signaling.
Importantly, Meta-language, which is among the most frequent categories, attests to the
specificity of CR language uses, which put high premium on talking with assessment activity (cf.
Goodwin & Goodwin, 1992; or Chapter 2 of this study). The CR metalanguage necessarily
entailed resorting to meta-linguistic nouns, verbs or even somewhat ‘coded linguistic terms’, like
rule 1 or 3 used by CR students to refer to norms and principles of ‘good discussion’ reviewed in
small group debriefings. Also, the very high frequency of metalanguage both in speaking and
writing confirms its hypothesized salience and impressive functionality in the Collaborative
Reasoning communications in 4th grade context (thus, indirectly validating its inclusion in the
coding scheme).
Likewise, observations made about CR discussants’ high use of Gestures in the course of
preliminary (exploratory) study and implementation of a more recent CR project (the Wolf
project) allowed for hypothesizing about the prospective high frequencies in the overall
metadiscourse uses. However, it comes by surprise how frequent and how important this means
of para-verbal communication is for the young adolescents in CR discussions. The fact that they
used so many Gestures and even more Fillers suggests their very engaged style, where every
attempt is made to use the resources available or opportunity to grab the floor, especially when
words do not come easy to express the student’s idea or feeling.
It is also noteworthy that Gestures, apart from adding emphasis when complementing the
verbal message, contributed to the flow of meaning similarly to some other language-based
metadiscourse categories. For example, the CR discussants oftentimes used a wide repertoire of
gesticulation to contribute to the function of many other codes, such as Sequencers (1st, 2nd, 3rd–
finger counting), Evidentials (showing sources of knowledge in the text), Talk-Organizing
(indicating whose turn it is), or Talk-Evaluating (grimacing and facials with paralinguistic voice
expressing critique, self-/others-deprecation or approbation). Whenever those meanings were
determinable and clearly fulfilling meta-communicative functions, they were double-coded and
thus they contributed to the repertoire of the major categories (as evidenced with CR children’s
The inference about the high-level of engagement finds further evidence in the very high
distribution of miscellaneous Emphatics (really, lovely, so) and perlocutionary force imposing
category of Commentary Directive, together with a slightly less frequent use of Interrogative
Commentary. The Perlocutionary Commentary reveals the agentive role of CR speakers, who, by
directive/ vocative or interrogative forms, try to force their interlocutors to take action. In a way,
similarly to relay racers they prompt or nudge the recipients to take up the metaphoric button in
the flow of conversation as in a relay race. The finding seen in these discussions has been
documented in the experimental study by Latawiec et al. (in preparation) that compared
metadiscourse in essays of CR and controlled students. It has been found therein that
Commentary was used considerably more often by CR-exposed writers than controls, and as
such can be attributed to conversational practices and dialogues of Collaborative Reasoning
small-group discussions, thus corroborating cross-modal transfer, also documented by
Reznitskaya et al., (2001) with regard to argument structure, and interestingly dubbed as
‘portable knowledge’.
Importantly, in stylistically rather than rhetorically oriented theories like the one by
Conley (1979), where metadiscourse is viewed as ‘figures of thought’, the interrogative form of
commentary when expressed by rhetorical questions may be viewed as ‘figures of communion’
(like allusions, or interpersonal voice) that help to form a bond with the Hearers. His view thus
underscores the relevance of Commentary to meta-communicating interpersonal meanings and in
this way complementing the propositional flow with an additional pro-social dimension.
Again by reference to the same study by Latawiec et al. (in preparation), another highfrequency category of Implicit Attitude Marking has also been given prominence in CR talk as
ranking almost as high as the high frequency lexical verbs in the Longman Spoken and Written
Language Corpus (Biber et al., 1999). It is noteworthy that Implicit Stance marking is twice as
frequent as the more explicit one (Attitude Miscellaneous) or 10 times as high as “I would”–
oriented Attitude. The finding suggests the CR participants’ high cognitive ability to mirror
complex logical 2-arguments entailing operations in language use (e.g., It is true that X is the
case, or commonly used in CR - it is unfair/ not nice/ wrong that …), which utilize what
Schiffrin (1980) calls ‘metalinguistic operators.’
Implicit attitudinal stance is rendered with greater subtlety than the explicit attitude or
emotions displays, and also reflects the move away from the speaking ego towards more
generalized orientation, which by implication may entail Speakers and their stance too
(seemingly an echo of CR argument stratagem called ‘other people might think’). Thus
detachment of the stance from the ‘speaking’ ego of the Speaker or Writer (as documented even
much higher CR uses in written essays may suggest), opens up for “other voices” and
polyvocality or multi-voicedness (Bakhtin, 1981) of the produced texts or talk. After Linell
(2009), texts become ‘dialogized’ when they harbor several perspectives or voices, which I
hereby propound to be ‘evidentialized’ in CR writers’ even more than speakers’ uses of Implicit
Attitudinal complex structures that meta-linguistically ‘echo’ other voices and other perspectives
– thus ushering them into the discourses.
Moreover, the more frequent use of implicit attitude that hinges more on implicatures and
so is more context-bound (e.g., different meanings of ‘just’ as a function of contextual aspects)
may thus suggest CR students’ heightened sensitivity to extraneous aspects of making attitudinal
perceptions public (other than epistemic stance marking though) – so essentially a sociopragmatic skill, especially in the context of ethical aspects of controversial topics of CR
discussions (incl. their writing task).
Other uses of Schiffrin’s (1980) metalinguistic operators include ‘for example’, ‘mean’ or
‘like’, which function as Code Glossing by Explanations (in the adopted taxonomy here), and
which also have been found amongst the most frequent categories. The result for glossing
Explanations may thus confirm previous research by Kim et al. (2007) into the CR-students’
repertoire of common argument fostering rhetorical moves (argument stratagems) that include
‘making argument explicit’, though their study pin-pointed the most organizing variant of “I
think” followed by + [ARGUMENT]. Evidently, “I mean”, sometimes also carrying out a
Speech Act performative function (confirmed by Vande Kopple, personal communication, 2011)
and similar glossing devices thus demonstrated, as revealed by very high frequencies, their vital
role as integral components of exposition in CR argumentation.
Among the next dozen categories of moderately high frequency are Cautious elements,
stronger form of Obligation (Should or Must) and Reminders of Material. They deserve a special
mention as they signal their relevant importance to the CR speakers. It seems that in the intuitive
awareness of risk in face to face confrontations of their opinions with others, CR speakers resort
often to hedging, which may be especially helpful when they want to voice the more-opinionated
stances while utilizing stronger “positional functors” in Wünderlich’s (1979) system of
Obligatory modal verbs Should or Must (revealing the source of attitude within the Speaker)
rather than the weaker modal of Had to (the source of attitude is external to the Speaker, e.g. the
society or circumstances). Lastly, the frequent use of Reminders of material reveals the need to
link the upcoming material to the earlier information, either in the discussion or outside it. This
usage indicates an attempt at better organizing (thus contributing to Talk-Organizing Bracket)
the flow of discourse/ talk into a coherent whole by tying back the threads of discussions (also
possibly for interpersonal or even ingratiating purposes).
High-frequency positions of discrete categories become even more comprehensible upon
observing the tendencies for the classes and brackets of metadiscourse categories, as illustrated
in Figure 4 - for discussions and Figure 6 – for essays (Chapter 4). The most frequent uses of
both Evaluating and Organizing brackets accompanied by total uses of logical connectives and
Illocution Markers (performative Verbs plus Boosters and Mitigators) show the same systematic
pattern for both discussions and essays (cf. Figure7). Evidentials seem to be more prioritized in
essays (3.91) than discussions (2.92), yet frequently utilized in both, comprising various
categories that show bases for knowledge or claims in the referential material, including
speculative scenarios, personal beliefs, induction/deduction, or ‘If/since-Then’ inferential
structures. Emphatics and Perlocutionary recipient-addressing Commentary show greater usage
in discussions which, especially in the case of Commentary, may be intuitive. Simply - the need
for urgency adding to one’s verbalized ideas is more grounded in the Speakers’ belief that the
prompting cue may be realized or actually implemented in the natural situation of the CR grouptalk.
Also, prevalence of Commentary in discussions can be attributed to the aspect of
‘coerciveness’ in dialogue (cf. Linell, 2009) or otherwise ‘imposition of response’ (p. 168),
besides which is inherent in the talk-in-action-reaction-appraisal exchanges. Interestingly, those
impositions of responsiveness can take more (yes/no questions) or less coercive (wh-questions)
forms by “opening up” (ibid., p.169) for narrow or wide ranges of responses (more open-ended
The high usage of Commentary in the oral mode seems conditioned by feasibility
predictions, not necessarily a matter of conscious awareness of the Speakers, especially to native
speakers. Pertinently, Biber (2001) remarks that patterns of variation among spoken and written
registers “operate below the level of conscious awareness and are usually not accessible to native
intuitions” (p. 114), which may be disputable, especially with regard to the intuitions, which are
hard to penetrate or measure in general.
Further on, both brackets- and classes-highlighting plots (Figure 4 and Figure 6) reveal
intuitively higher conversational uses of Hedges, and Code Glossing (as mentioned earlier with
reference to oral modality solely) than those in essays. While higher frequency of total Attitude
Marking in essays suggests CR writers’ greater tendency to express what they feel (emotions,
desires, affects) about the situations described by the propositional material, in writing such
expressions are less publicly de-facing in the pragmatic sense or otherwise not so facethreatening in the absence of relative power relations (e.g., Brown & Levinson, 1978a; 1978b).
By contrast, the five times higher frequency of Intersubjectivity marking in
conversational metadiscourse than in the written register may again find its motivation in CR
Speakers’ belief system about what is important and proper for communication in the classroom
and in-group context. Young adolescent uses of pragmatic vague elements that invite solidarity
or attempt at indicating shared in-group knowledge (rather than common knowledge) basis, and
thus signaling the insiders’ voice, are more imperative in the spoken register as the
communicative success hinges on the perceptions and appraisal or rejection (deprecation) by the
immediate Hearers, and even more so by a group of Hearers. Rather than alienating the group of
peers by using a formal register, the CR speakers appeal to their sense of togetherness and ability
to understand one another “without words”, in fact “saying less by communicating there is more”
(Overstreet & Yule, 1997).
Apart from that, the two tasks of speaking and writing differ in the targeted audience,
informal in discussions and formal in essays, and this sense of audience rhetorically, stylistically
and socio-pragmatically modulates the language uses to the conceptions of “the other” and
“other-orientedness” as differently afforded by the two modalities and CR tasks (e.g., Linell,
2009). Again, as with Perlocutionary Commentary, the socio-pragmatic aspect conditioning the
language use (like the high Intersubjectivity Vagueness marking) corroborates the argument that
“teenagers are far from ignorant” (Stenström et al., 2002) of the values and importance
associated with the relationship of social features with language features.
With the issue of different audience conceptualization in the written task, let me shift the
focus of discussion towards the writing domain. The comments about relative salience attributed
to most frequently found categories in oral modality (when compared with essays) naturally also
apply to written metadiscourse by reversal or flipping the side of the perspective adopted.
Notably, in writing discrete metadiscourse categories assume different proportions in
relation to other codes, and what stands out as plotted in Figure 5 is that Speech Acts assume
even greater frequency and so by inference – salience - than in discussions. Also, in place of 2nd
top-rank conversational Fillers come in Causatives (3.3), which are twice as frequent as in
Discussions (1.7). The fact cannot be overestimated as Causatives are indispensable for signaling
important rhetorical and logical relations of cause-and-effect binding (cf. Sanders & Nordman,
2000), or causal-chain in reasoning and argumentation. Added to which slightly more frequent
Contrastives, twice as frequent Sequencers and comparable uses of Additives show a systematic
pattern in the textual metadiscourse being used more to integrate written texts than in discussions.
These inter-clausally (Logical-Temporal) or globally (Sequencers) binding connectives when
paired with strongly organizing categories that signal reasoning and inferencing types such as
induction/deduction named as Induction (1.0) or If/since-then (.9) along with If-conditional (.8)
structures, it seems reasoned to argue that the focus from categories emphasizing interactivity via
interpersonal aspects in discussions shifts in writing toward integrity of the propositional flow.
Other written modality specificities include more frequent use of Aposteriori/
Retroduction hypothetical scenarios, which can be difficult to handle in speech as being complex
3rd-conditional sentences that entail use of Past Perfect Tense in the subordinate clause and
complex Perfect Aspect-oriented predicates in the superordinate part. Also, due to the fact that in
writing students have more time to pre-think their answers and the way they phrase it due to
different production circumstances, it is quite intuitive that such complex structures are more
likely to occur in essays than discussions. Still, the use of retroductive structures may reflect a
move towards speculative thinking about others and other perspectives (while introducing them
into unreal past scenarios). As mentioned earlier with regard to Implicit Attitude structures that
open up for others’ voices and perspectives, the hypothetical scenarios like the Aposteriori /
Retroduction might bear traces of the same trend toward ‘dialogising’ texts by introducing
multiple-perspectives or voices (after Linell, 2009).
Between Modalities
Moving on to the findings from overall comparisons between the two corpora that are
illustrated in Figure 7 (as discussed earlier with reference to distinctive patterns for each
modality), the prevailing use of interpersonal and interactivity-promulgating metadiscourse in
discussions has been observed. The prevalence in conversational uses has been noted amongst
globally text-binding Reminders of Material, various Epistemic Modality markers (Hedges; NB,
that show degree of commitment to the truth value of the propositional material (Simpson, 1990),
Code Glossing that ensures clarity and adds to the explicitness of communicative/ speech acts, as
well as perlocutionary or coercive Commentary together with similarly functioning What-if
pseudo-rhetorical structures remarkably tip the scale towards greater interactivity featured in CR
oral discourses rather than essays.
On the other hand, as presented in the same Figure 7, the written modality features more
textual metadiscourse (various connectives) and more references to a group of Epistemology
markers that are historically called Evidentials and which generally reflect a stance to
epistemological status of propositional material by usually showing the information bases for
that material (Vande Kopple, 1997; Chafe, 1986). Specifically, more frequent uses of the
Induction markers, Hearsay evidentials (From someone else) and hypothetical Aposteriori /
Retroduction or Deducing structures altogether may suggest CR writers’ intuition about the need
to indicate their knowledge sources or knowledge bases so that their claims would assume more
authority, and logical and ethical appeal (the author’s trustworthiness).
Given the fact that essays were written at the end of Collaborative Reasoning series of
discussions, where students were continuously reminded of the need to substantiate their
opinions with material from the literary source or personal experience (during their small-group
debriefing sessions with Teachers), the more frequent pattern of evidential metadiscourse in
writing may possibly be attributed to the argumentative stratagems at play in CR practice.
However, without confirmatory evidence with the control group, that did not receive the CR
training in argumentative discussion techniques, such a postulate runs a risk of falling into a
logical fallacy of “post hoc, ergo propter hoc” (after this, therefore because of this). The earlier
experimental study by Latawiec et al. (in preparation) comparing written metadiscourse of both
CR-exposed and non-CR children (controls) failed to yield such results for CR students though.
Another outcome that emerges from the cross-modal comparisons (as illustrated in Figure
7) is the attitudinal stance marking by modality markers of Obligation expressed by Should/Must
and the Forbidden category (both Deontic modals) more often in writing than essays, whereas
the other Obligation variant – by “Had to” forms – in oral modality. As mentioned earlier in the
discussion, the higher frequency distribution of stronger forms of attitude marking seen in
writing may suggest young writers’ appropriation and application of the socio-cultural norms of
politeness – different for face to face confrontations and different for ‘unknown and vaguely
intended others’ in the production of written essays. The use of stronger obligation modals may
reflect CR ‘authorial voices’ vulnerability to the between task-differences as a function of
prospective audience (Teacher/ Researcher vs. peer-group) and different purposes – of not
debating or exchanging ideas in a dialogic form but rather persuading of the soundness and
legitimacy of the claims/ opinions ventured with reference to the ethical issue raised in the Big
Question (Should Jack tell on Thomas?). CR writers evidently found the stronger Obligation
modals more befitting the purpose of writing a convincing argument rather than uttering their
arguments in a group of familiar others, which might entail losing ‘public face’ or endangering
interpersonal relations (due to seemingly too strong or moralizing forms of Obligation).
Prechter’s (2003) study of stance differences in American and British English may be
pertinent to the discussion of stance marking modals as she found that Obligation modals (should,
have to, couldn’t) are the most frequent forms of ‘verbal stance marking’ (by modal verbs), and
much more frequent than by adjectival, nominal or adverbial markers. Her findings suggest a
robust pattern seen more often in American rather than British conversations (though her
frequencies show the number of occurrences per hour, not the number of words). Interestingly,
together with the reported greater use of emphatics by Americans, the pattern may contribute to
re-affirming of cultural stereotypes. Most pertinent to this study of metadiscourse appears the
fact that Obligation Modals seem to be most common forms of attitude marking in American
English, and so the inclusion of the Obligation Modals and the Forbidden category in the metadiscursive Evaluating Bracket in the adopted and propounded taxonomy seems to gain in metatheoretical and not only empirical credence.
To complete the discussion of the cross modal comparisons, the findings from the last
analysis by Poisson regression indicate an association between Organizing and Evaluating
Brackets in both modalities. Specifically, slightly higher chances of both Brackets in the written
modality have been predicted based on the count data frequencies, while controlling for the
variances of other predictors in the model – most importantly the number of words used.
Moreover, Global Coherence Textual Connectives, which include text-organizing
categories of Sequencers, Topicalizers, Announcements and Reminders of Material, have been
predicted to occur more likely in written essays than discussions (though by 5% only). The
finding also seen in t-tests for dependent samples suggests a robust pattern for metadiscourse in
CR writing trending towards organizing or more integrated textual structures.
Another converging result from t-tests and the regression analysis about If-conditionals
showing and having higher predicted log-likelihood uses in writing (by 30%), on the basis of the
corresponding uses in discussions, further adds to the overall inferences about the observed
pattern that feeds into the organizational rather than evaluative metadiscourse functions.
Similarly, Induction markers reaching higher log likelihood counts (by 23%) in writing than in
discussions (higher on the frequency distributions, in t-tests too) add to the ‘signposting’
repertoire of organizing rather than evaluating meta-communiqués of CR writers. It has to be
mentioned here that the category labeled Induction (e.g., “if-then”) may be viewed as a
misnomer as in fact encompassing explicit signals of induction (inference à particularis, based on
a series of observable phenomena / evidence) or informal deduction (inferences in arguments
with missing or taken-for granted premise) in reasoning. Hence, the result that shows higher
predicted chances for Induction in essays prompts an inference about the propensity of CR
students to signal both their inductive and informal-deductive reasoning moves more explicitly
and more often in writing. This may also suggest their greater metacognitive awareness of the
deconstructive task of prospective comprehenders of such written texts, whose meaning making
in written texts resides more on textual rather than on contextual cues that are available in
Also, a little higher probability of Meta-language uses in essays (by 8%) seems somewhat
counter-intuitive because it may be expected that, thanks to a special opportunity for debriefing
in discussions and so touching on the participatory, discussing- and arguing-techniques, the
likelihood will be higher for the metalinguistic terms (designata) to be more often utilized in oral
modality. Simple t-test results show oral modality advantage, i.e. higher frequency for
metalanguage in discussions. However, t-tests do not take into account other modifying effects
of variables, like reading ability, gender, school population or ethnicity; even though they take
into account verbosity by being run on rates (of raw counts to words) used in both modalities.
The Poisson regression model that takes into account the effects of other predictors entered into
the model, along with oral metalanguage category (as a predictor), on the written metalanguage
as dependent variable (so a baseline group of reference), while keeping constant (ruling out) the
effect of the wordiness or verbosity, may thus reflect the relation more credibly. Pertinently,
metalanguage has been found to be significantly related to the school population (higher in the
rural school), as results of Poisson regression indicate. The school effect may have modified the
direct effect of oral metalanguage on written metalanguage, bearing in mind the differences
between the populations of the compared schools (predominantly European-American in rural
and African-American in urban school).
On the balance side, the Poisson regression findings suggest considerably higher chances
of emphatic Repetition (by 80%) in oral modality. The indicated phenomenon of higher
likelihood in discussions for Repetition (exclusive of repetitions due to false starts or placeholder
uses) can be explained by common techniques witnessed among speakers who resort to repetitive
structures (lexical or syntactic) in order to ensure that the message was carried out properly or
reached the intended goal.
Besides, Repetition used for emphatic effect (e.g., really, really, really big) may partly be
linked to the finding of an exaggerated use of emphatics in American conversations as compared
with the British ones, also found in the comparative study between the two dialects by Precht
(2003). Just to remind, her study was on the excerpted corpus of 100,000 words from the
Longman Corpus of Spoken and Written English, with American dialect specific emphatics
expressed by real, pretty and exactly.
Between Socio-Cultural Groups
The results of t-tests across modalities as a function of gender suggest (as presented in
Table 4) that girls’ essay are more bound by miscellaneous Logical-Temporal connectives and
contain more hypothetical scenarios by putting Myself in the story, while boys’ essays have
more explicit causal relations signaled by more frequent use of Causatives. In discussions, boys
were found to use more Boosters of Illocutionary markers (speech acts) than girls. The finding of
causative linking devices for boys may suggest their greater attention and skill in rendering
cause-and-effect chains, while “thinking the argument out” (cf. Hample, 1985; 2007). The
finding for causal relationships in boys’ written texts (though not talk) may partly corroborate a
sociolinguistic hypothesis ventured by Tannen (1991) that male and female talk is organized by
different metafunctions. Namely, while men’s talk can be dubbed “report-talk”, women’s talk
can be called “rapport-talk”. In general, report-talk is likely to be organized in cause-and-effect
rhetorical structure or temporal sequence, as more binding than additive, list like or descriptive
structures (Hoey, 2001; Meyer, 1975; or Sanders & Nordman, 2000). The results for spoken
modality do not indicate the boys’ edge in this respect though.
Twice more frequent Boosters in conversations of boys may be somewhat connected with
a documented pattern in teenage talk (though among British speakers; Stenström et al. 2002)
with a heavy presence of ‘exponents of power’, as Channell (1994) dubbed the use of vague
words (cf. Chapter 2). Though Stenström et al. (2002) study of intensifiers found corroborating
evidence for the more frequent uses by girls, their findings indicate that it is boys’ specialty to
use the superlative forms of intensifiers like “extremely angry”, “absolutely stupid” or “fucking
weird” (p. 142). Therefore, their results about boys’ ‘propensity’ towards superlatives may be
linked or show semblance to this study’s finding about boys’ heavy use of intensifiers of Speech
Acts, called Boosters (with their function of intensifying the Illocutionary effect), which yield
such expressions as “I truly believe”, “I really think…” or “[that’s] exactly what I say” as
superlative forms of illocution.
In an attempt to explain girls’ higher usage of self-oriented scenarios (classed among
evidentials), let me indicate a possible semblance to another finding from the grouping by high
vs. low reading ability (MAT), where high-scoring children’s talk shows considerably higher
uses of various types of hypothetical/deductive structures by putting Others in scenario or
Aposteriori/Retroduction (undergirded by the higher predicted effect of Total of Evidentials, as
found in Poisson regression). The findings suggest that for oral argumentation CR girls or high
MAT scorers may have considered evoking hypothetical scenarios as an effective persuasive
strategy in Collaborative Reasoning context (especially when paired with emphatic Intonation);
this however might have been re-conceptualized for the sake of written argumentation as the
multiple deductive scenarios did not show significance in essays.
Reasons for providing multiple scenarios (Myself/Others/Retroductive) can be found in
previous research and observation that children often conflate explanation with argumentation
(Asterhan & Schwarz, 2009). Instead of providing justification and addressing prospective
criticism/ counter-arguments (as the nutshell form of an argument demands; cf. Guillem, 2009),
children seem to proliferate examples and illustrations (mostly personal anecdotes) or
hypothetical (deductive) speculations instead of inductive or conclusive reasoning.
Further within the grouping by reading ability, more frequent uses of Sequencers by high
MAT-scorers, as shown in frequency Table 5 (and also found higher in regression - by 18%),
corroborates the finding by Latawiec et al. (in preparation), when in comparison to a control
group of non-CR exposed children, it was the group of CR children with high reading scores
(high MAT) that featured higher uses of Sequencers and so better global binding in their essays.
This advantageous finding seems paired with other results from the Poisson regression that link
high-scoring on the reading achievement test to slightly increased predicted uses of Induction,
If/since-then, total Global Connectives and evaluating and organizing brackets in the students’
written metadiscourse (between 1 and 8%). The findings suggest an intuitive pattern that the
higher the reading ability the higher uses of those highly organizing codes (hopefully without
self-fulfilling prophecy as the essays were coded blind to condition and other socio-metric data).
It may also be postulated here that the higher reading scores and higher organizing uses
of metadiscourse in either talk or writing may have a common basis in the text-structure
awareness and its effective use (conscious or unconscious), which has been found to facilitate
reading comprehension (Meyer, 1975; Britton et al., 1982), academic achievement (Latawiec,
2010) and written or oral communication skills (Bereiter & Scardamalia, 1982; Graves, 1982;
Hoey, 1994; Hyland, 2000, inter alia). Findings for grouping by school populations (as found in
t-tests) that reveal overall higher uses of Talk Organizing bracket, Contrastives and Personal
beliefs by rural schoolchildren suggest greater attention paid to textual metadiscursive aspects
(rather than evaluative or interpersonal) as well as greater diversity of opposing viewpoints
expressed in writing. Simultaneously, rural schoolchildren discussions show their greater
attempts at glossing, making speech acts explicit, with a variety of modality changes by hedges
or emphatics and better sequence than their urban counterparts. It is also notable that rural
schoolchildren talk shows more Implicit and total of Attitude Marking as well as increased
Intersubjectivity marking. The results also suggest greater emphasis on organizing metadiscourse
in writing (thus resulting in greater integrity of texts). In discussions it is the evaluative and
intersubjectivity pragmatic meta-discourses that dominate in rural school thus implicating their
orientation on interactivity of the speech flow or talk flow.
A reverse pattern seems to be suggested by the results of t-tests for children from an
urban school. It is the Text-Evaluating bracket that is more frequent in their writing, while in
their discussions Talk-Organizing codes prevail (talk-regulating by coercive and hearer-engaging
Commentary Interrogatives), which suggests urban schoolchildren’s greater preoccupation with
maintenance of talk and importance of retroactive grounding of upcoming talk in the earlier
material (by Reminders of Material). The emphasis on the participation regulatory technique
may possibly be attributed to a greater focus in the Teacher-guided debriefing on a detailed
review of good argumentation and ‘good discussion’ techniques and their appropriation by CR
participants (e.g., as may be suggested by observations of one of the Teachers’ emphasis as
revealed in video-recordings).
Lastly, in written modality findings for grouping by ethnicity reveal more Emphatics and
Contrastive conjunctive devices, which, in crude simplification, suggest more disjunctive
viewpoints expressed by European Americans (EA), while more Attitudinal Stance marking with
“I would” structures in their African American (AA) peers. Results for discussions suggest that
African-American children (much less numerous though) engaged their hearer-peers by
perlocutionary-force-imposing metadiscourse (Commentary interrogative) as well as more
utilized Reminders of material. The result indicates their talk-organizing concerns and social
dynamics that were dominated by managing moves or turn-taking procedures, which as Bennet
and Cass (1989) suggest may not necessarily be progressing understanding of content (also
Almasi et al., 2001). Aposteriori scenarios may signal their complex retroductive way of thinking
and so ushering in other perspectives or voices while arguing the cases for or against, or
proliferating scenarios as explanations rather than reasoning moves in argumentation. Notably,
their more frequent uses of Emphatic Repetition can be attributed to the grammar rules effective
in the vernacular English.
The group of European Americans seems to show a pattern of quite different uses, where
Intersubjectivity, Hedges and miscellaneous Emphatics together with Implicit Attitude marking
tip the scale towards evaluating metadiscourse. Their discussions are also rather loosely bound
by more frequent uses of additives, which however add to their cohesiveness of talk. European
Americans seem to pay attention to making their conversational argumentation more explicit
(Speech acts and Glossing), and equipped with clearer inferencing signals by If-then structures
that usher conclusions by logical ways of affirming and confirming (Modus Ponens & Modus
Tollens in formal logic), whose value cannot be overemphasized in logical and dialectical
reasoning (e.g., Anderson et al., 1997; van Eemeren & Grootendorst, 2004).
However, caution is urged in the interpretation of the findings from comparisons by
school or ethnicity groupings as ethnicity has been nested in the school location with EuropeanAmerican ethnicity dominating in the rural school. Also the ethnic groups were not proportionate
(45 – EA, and 22 - AA, while individuals of other ethnicities were excluded). And though the
findings seem interesting, the results bear traces of confounding the ethnicity with population in
the given location and associated with it socio-economic status.
Micro-Genetics in Oral Modality
To exhaust the discussion of the analyses run on the available data from essays and
discussions, the results of the correspondences and differences between the two analyzed
discussions in the whole sample suggest that the many correlations that were found between
discussions form a systematic pattern. The codes that showed an increase over the series of
discussions are grouped around the talk-organizing metafunctions (Topicalizers, Speech acts,
Commentary, Talk-Organizing Bracket) and talk-evaluating uses of metadiscourse (Obligatory
Should, Talk-Evaluating Bracket) with a notable Metalanguage. However, the categories that
showed a decline in usage include Intersubjective-Vagueness Bracket, and some organizing
metadiscourse (Sequencers, If-then) as well as weaker form of obligation or negative attitudinal
stance by means of Impossible “positional functor” (Wünderlich, 1979). Interestingly, the trend
may suggest increased engagement or stronger commitment to the propositional flow of
meanings as well as stronger attitudinal stance by ‘motivational’ and ‘normative’ categories of
verb operators (after Wünderlich, ibidem) expressed in CR argumentative talk.
All in all, the observed cline towards organizing and evaluative uses with the
simultaneous drop of intersubjectivity vagueness marking may suggest students’ greater focus on
transactional / informative talk in the course of successive discussions, possibly due to a lesser
degree of socially-motivated conversational “interacting for the sake of interacting” (Berry, 1981,
p. 132), or seizing the floor just to show attention, solidarity, friendliness and co-operation (cf.
Yule, 1996), yet in the absence of factual information or ideational material to communicate.
Qualitative Analysis
Qualitative findings have been mostly discussed simultaneously with the analysis of the
correspondences as well as helpful or impeding metadiscourse uses. Nonetheless, this discussion
would suffer from incompletion if several things would not be remarked upon.
Firstly, the qualitative analysis of metadiscourse features in naturally occurring discourse
afforded a rare occasion to investigate and identify systematic differences in the functional uses
of each variant or discrete feature - both in talk and writing.
Secondly, the systematic qualitative analysis that was undertaken in this study shows
semblance to approaches that favor elementarism, where all complex sequences or patterns are
derivable from elementary constituents (by analogy to Shegloff’s adjacency-pair elementary
system). The elementary approach therefore resulted in an extremely wide spectrum of
elementary variants that can fulfill very specialized functions. In turn those functions show not
unilateral relationships but bi- and multi-lateral links to various linguistic forms. Thus discovered
web of form-to-function and function-to-form links, despite its complexity, reveals the
systematic valency patterns or a kind of grammar of discrete metadiscourse variants (like in
syntax S+ V makes a nuclear sentence structure, so similarly Speech Acts + Intensifier turn the
latter into a Booster, etc.). And only those findings or discovered ‘valency patterns’ allowed for a
more holistic view.
Thirdly, the more holistic view (which seems to have been under-represented in my
preliminary stage of the study) takes a form of seeing through the classes in order to find
‘communicative planes’ on which meaning is transferred. These planes or dimensions have been
theorized to coincide with the bracketing roles of two metafunctions identified by Schiffrin
(1980) in talk, and expanded by myself upon analysis of naturally occurring CR talk and writing.
The moderate holism of the approach then resides in the grafted 3 frames for “communicative
projects” (communicative goal pursuits, centered around a task), which in turn embrace the
elementary metadiscursive variants, that young adolescents undertake in their CR dialogues and
somewhat “dialogized” texts (both terms inspired by Linell, 2009).
In relation to the three tenets of the approach, the qualitative findings reflect those
different dimensions – the elementarism demonstrated in the minute functional units that stream
into the functional classes that are serviceable with their distinctive features and functions to the
‘communicative projects’ in the dynamically developing discourses (both spoken and written).
Thus the Form-to-Functions table understandably includes findings about affordances
that such an elementary-to-holistic analysis entertains but also reveals various problematic areas.
On the one hand, it attempts to neatly and compactly show the scope of functions for various
linguistic forms. On the other hand, it does not show in greater detail how attribution to those
different variants is made, for the purpose of which ideally a whole book would be needed.
Hence, the lengthy function to form table is still insufficient to show how the evaluations
of metadiscourse distinctive uses need to be operationalized from an elementary to the more
holistic level, e.g. from the hedging device, by “other cautious elements”, to the broad basket of
“evaluating” bracket (be it talk or text-evaluating) to which this little device like a ball is thrown.
Therefore, descriptive passages explicating helpful and impeding uses complement the
elaborate form-to-functions table. It seems that the descriptions would be limping if not prefaced
by the delineation of the patterns or valency of metadiscursive forms in language (incl. paralanguage) and abstract functions activated in language in use, in the Collaborative Reasoning
Chapter 6
The study’s overarching goal was that of exploration of metadiscourse in order to
establish how metadiscourse is used in discussions (#1) and writing (#2) by children
participating in Collaborative Reasoning promulgating argumentative stratagems in small group
interactions. In practice, the adopted approach and definitions of metadiscourse (as inspired by
extant theory and systems) set more concrete operational goals and questions that can be
answered. Namely, the study affords answers into:
how children organize and bind texts or arguments,
how they clarify, explain and define meanings,
how they make their communicative/speech acts and goals explicit,
how they reveal their attitude and evaluations (both epistemic and emotional),
how they engage recipients and express in-group intersubjectivity,
and how they structure their discourse in collaborative talk-at-action or ‘dialogized’
written texts.
A similar exploratory character has marked the other study goals such as an inquiry into
correspondences or differences between the metadiscourse uses in the two different corpora
(#3a), and in the progression of discussions in a series (micro-genetic strand, #3b) to inform
about metadiscourse patterns and prospective changes in Collaborative Reasoning discussion
techniques. The study also intended to answer how metadiscourse is used by different users, as a
function of varied socio-cultural and socio-linguistic characteristics of participants (#4). Finally,
the study set to qualitatively determine how metadiscourse helps or impedes the flow of
meanings in spoken and written argument by CR participants (#5).
The study being exploratory rather than confirmatory did not set a priori hypothesis to be
tested. Instead, meta-analysis of extant theories and systems as well as a few empirical findings
mostly from non-American contexts brought inspiration and some guidelines as to what
directions would be most informative and worthwhile pursuing.
In order to investigate the uses of metadiscourse both qualitatively and quantitatively, the
study developed and tested an instrument of measuring metadiscourse uses on various levels –
from very elementary and discrete variants by streamlining towards classes grouped by major
functions to ultimately more holistic planes on which metadiscourse has been theorized to
operate. In effect, a multi-fold taxonomy has been adopted that appropriated some categories
established from discourse analysis in writing (e.g., Crismore, 1985; Vande Kopple, 1997;
Wünderlich, 1979) and separately in talk (Overstreet & Yule, 1997; Schiffrin, 1980) and
combined them in one three-partite instrument. Due to its comprehensiveness and testing in a
pilot study, the taxonomy (accompanied by a checklist manual) captured systematic differences
in the functional uses of each metadiscourse category – from elementary to more holistic levels.
The devised taxonomy (with an accompanying manual) brought in results of both
quantitative and qualitative type; besides, the taxonomy was also refined by other raters
establishing inter-rater reliability. The qualitative analysis of discourse in the written and oral
corpus allowed for determining the valencies or systematic patterns of metadiscourse functions
in relation to linguistic and para-linguistic forms, which resulted in form-to-function tabular
review of the observed patterns that showed complex ways of interaction between lexis, context
of use and abstracted functions. The elementary-to-holistic approach was also applied to the
explication of the various effects metadiscourse has on the flow of propositional meaning as well
as on other features – such as interpersonal, including intersubjective, or intra/ inter-textual.
The effects of the application of the measuring instrument yielded count data that
afforded statistical and distributional analysis. The results could be best interpreted using
resources available in corpus linguistics, rhetoric, semiotics and socio-pragmatics. The results
allow for tracing patterns and gaining insights into the ill-researched metadiscourse uses by
young adolescents in both spoken and written modalities in the educational context.
Thanks to the three over-arching or bracketing codes of Organizing, Evaluating and
Intersubjectivity meta-functions in the adopted taxonomy as a measuring instrument (inspired by
Crismore,1985, Schiffrin,1980, and Vande Kopple,1997), it became easier to demonstrate an
overall dispersion of metadiscourse in discourse proper. Since the attribution to the brackets
hinges on the particular functions of the minor codes, the optimal effectiveness is targeted only
when one is able to determine the grouping or unifying factor (class function) for the highly
specialized micro-functions (which could be iteratively refined by qualitative analysis of
discourses). Simultaneously, the adopted taxonomy seems to be validated as an instrument for
measuring metadiscourse in both oral and written corpora and at the various levels of complexity.
First of all, it has to be acknowledged that the two corpora used for the analysis
inherently differ as their global ‘communicative projects’ (after Linell, 2009) or otherwise
communicative goals differ as a function of: the mode, production circumstances,
interactiveness, purpose and targeted audience (cf. Biber, 2001). It has to be reminded too that
children received practice in Collaborative Reasoning in small group discussions with
metalinguistically rich debriefing operated by a teacher and participants themselves, yet not in
the writing technique.
Given the differences of the task of writing to an unknown audience, possibly a teacher or
researcher (incl. RA) versus discussing in a small-group of peers, and so somewhat differently
conceptualized purposes, as not debating in a dialogic format but most likely persuading of the
logical and ethical soundness of the verbalized stance, the discourses thus produced, and in
different circumstances, inevitably had to differ on several levels. And it seems that the young
writers’ intuitive or formally learned conceptualizations of the tasks of writing and speaking
affected not only the organization but also the expressions of the author’s attitude, as the
emotional aspect of the author’s voice (and so most vulnerable to circumstantial influences).
Within each modality distributional patterns of more molecular codes show fine-grained
differences. Surprisingly though, the proportionate use of the two holistic planes or brackets
shows more or less similar pattern both in the written and spoken corpus. Thus observed
systematic pattern shows approximately twice frequent uses of the broad Evaluating Bracket
than the Organizing Bracket; nota bene, both of which brackets subsumed other more molecular
yet still grouped in classes of metadiscourse categories. While Evaluating Bracket shows
attitudinal stance markers, obligation and the forbidden modality verbs as exponents of stance
(“positional functors”, according to Wünderlich, 1979) as well as evaluation (agreement/
disagreement) expressing speech acts and glosses, the Organizing Bracket has been composed of
globally binding coherence markers (topicalizers, sequencers, reminders and announcements of
material), performative speech acts and inferencing/ reasoning markers (e.g., induction/
deduction). Interestingly, the pattern among the two holistic brackets seen across the two
modalities has been statistically predicted to occur marginally more frequently in writing than
The systematic findings thus suggest that despite the aforementioned fundamental
differences between the two tasks of argumentative group-discussion and writing, the children’s
overall uses (per unit of analysis – discussion or essay) do not show tectonic shifts that would
reflect their completely different conceptualizations of the two communicative tasks or projects.
The tasks or projects have been resolved differently though on the local task/ project levels, i.e.
en route with the dynamically changing intersections of viewpoints in discussions or upcoming
material in thoughts during the writing process. The analysis of the fine-grained correspondences
suggests that essays written by CR children bear heavy traces of ‘dialogism’ (cf. Linell, 2009).
Namely, their texts open up for “other voices”, commonly known as multivoicedness or
polyvocality (Bakhtin, 1981) by using various types of evidentials that usher in perspectives of
others, either more concrete others that are put in hypothetical deductive scenarios (syntactically
structured along various conditional types), or more generalized others - via complex implicit
attitude marking structures that are two-argument-entailing (one in superordinate and one in
subordinate clause, e.g., it’s not fair/right/ kind/nice to the others/kids that X is the case).
Not only CR discussions but also essays show evidence of power relations at play (cf.
Brown & Levinson, 1978; 1978b), as the results of the cross-modal comparisons suggest. It is
due to the adherence to politeness principles and public face saving that the attitudinal stance
(generally most revealing of the psychological facets of the speaker or author) is more subtly and
feebly/ cautiously expressed (by weaker verb modals like deontic obligation “has to”) in face-toface confrontations, which are “face-threatening” when venturing opinions of an ethical or
controversial nature, whereas more normatively and strongly in writing (with positional functors
of “should”, “must” type, after Wünderlich, 1979; or deontic modality “forbidden”). Similar
perceptions and norms holding for in-group interactions with social, ethnic and peer/CR groupconformity-needs are likely to have dictated resorting to hedging or mitigating devices by CR
discussants. And though the findings were partly predicted (salient theories were discussed in
literature review), it is interesting to see their confirmation and quantitative proportions, and the
actual behaviors as reflected in the language of the young adolescents who intuitively adjusted
their discourses to the socio-cultural and circumstantial norms, which hardly ever are the matter
of conscious awareness, especially to native speakers (cf. Biber, 2001).
Another comforting finding for the CR member and researcher is in the high-engagement
level of speakers themselves and recipient-targeted engagement marking. The several
metadiscursive categories evidence that pattern with What if-response prompting structures,
perlocutionary or coercive uses of Commentary, and even very highly utilized para-verbals
(Gestures). It is noteworthy that in stylistically oriented theories (rather than rhetorically – which
more recently puts a premium on epistemology marking), interrogative forms of commentary
(e.g., rhetorical questions) are viewed as ‘figures of communion’ that help to form a bond with
the Hearers (like figures of allusions, or interpersonal voice cf. Conley, 1979; where
metadiscourse is viewed as ‘figures of thought’). Commentary has earlier been found to abound
in CR essays too, in comparative study with non-CR exposed children (Latawiec et al., in
preparation), and it has been considered as an advantageous feature as targeting the reader’s
involvement in a kind of dialogue and so meta-communicating interpersonal meanings that
complement the propositional flow with an extra pro-social dimension.
Another insight gained is related to the third bracket of Intersubjectivity-pragmatic
marking (otherwise ‘solidarity in-group’ marking), which amounted to 7% in oral modality when
compared with the broad organizing and evaluative brackets and just 1% in written modality.
The results of the micro-genetic analysis inform the CR research that across both small-group
discussions the intersubjectivity vagueness-adding uses dropped gradually, while the other broad
brackets distributions remained similar. The finding may indicate greater focus on informational
flow than on interpersonal and so intersubjective (solidarity signaling) with the progression in
the CR discussions. Also, children’s uses of pragmatic vagueness marking by similar means as
intersubjectivity - theorized by other researchers to be “exponents of power” (Channell, 1994),
and as such expected to see more often in female language (again as girls were theorized to be
less powerful than boys and so show powerless features in women language Lakoff, 1975; or
Coates & Cameron, 1988) - did not show confirmatory findings in the analysis by sociolinguistic groupings. However, the discovery of Boosters in boys’ intensification of explicit
speech acts in oral argumentation (thus boosting the illocutionary force of argument making)
may be viewed as power in language (by whimsically flipping of the Lakoff’s ‘exponents of
power’ theory).
Another common socio-linguistic theory of female and male talk organized by different
meta-functions which ventures that men’s talk can be dubbed “report-talk”, women’s talk can be
called “rapport-talk” (Tannen,1991) has not been confirmed by comparisons of girls’ and boys’
talk. Nonetheless, boys’ more frequent use of causatives in written essays, which indirectly
indicate somewhat “report”-like organization that signals cause-and-effect or problem-solution
rhetorical structures (Hoey, 2001; Meyer, 1975; or Sanders & Nordman, 2000), may tentatively
suggest some partial corroboration.
Not all findings though have been comforting and uplifting. As the qualitative analysis
ensued, it was discovered that quite a few uses of metadiscourse impede the flow of meaning
(usually by default understood as the content or propositional material). As demonstrated in
excerpted episodes from children’s talk or whole essays (cf. qualitative section of the results
chapter), the ubiquitous intersubjectivity-vagueness markers can obscure the propositional flow
by halting or slowing down the flow of arguments (less so seen in writing) and adding to the
excessive verbosity. A removal of the superfluous elements may show the informational material
does not lose its meaning. However, the impact of such a pruned talk may set forewarning
adverse responses in recipients of such communiqués, which are then devoid of authenticity and
not aligned with the personal characteristics of the speech group to which they are addressed (so
in a way flouting the pragmatic maxims enabling smooth Speaker-Hearer Co-operation, cf.
Grice, 1975).
Likewise, evaluative and attitudinal stance marking when used in an excessive way that is
not counterbalanced by more neutral text-organizing metadiscourse may result in signals of
manipulative techniques. As illustrated in the qualitative analysis of students’ writing or talk, the
overt and proliferate uses of evaluatives seem too pressing with their subjective stance and may
build up resistance in recipients (be it hearers or readers) who thus recognize the attempt to
influence them (cf. Kamalski et al., 2008).
The finding of disadvantageous uses of metadiscourse is not restricted to evaluative or
intersubjectivy-marking. By analogy, text or talk-organizing metadiscourse can also impede the
propositional flow by setting similar forewarning effects by blatant uses of coherence markers or
excessive uses of connectives or induction-signals by “so”, which due to overuse lose their
functionality or become de-ranked or otherwise devoid of their canonical organizing function
and start to assume pseudo-functions; for example, combinations of “and then” as often found in
a sequential role, so globally cohering, when overly used turn into more locally binding temporal
connectives (cf. findings of “and then” African-American children’s uses in McCarthey, 2002; or
“ands” in Michaels, 1981).
Also, it has been found in qualitative discourse analysis that though the evaluative
categories usually prevail over organizing bracketing functions in both spoken and written
modality, it is the more or less balanced use of both - be it in speaking turns/ discussions or
persuasive essays – that helps the flow of propositional meaning.
From the last two insights pedagogic and future research implications emerge.
First of all, to address the argument made by Crismore (1985), who researched the use of
metadiscourse in textbooks, ‘that it is difficult to determine how much metadiscourse to put in,
what the optimal level is and where in the sentence to put it’ (p. 313), from the findings of both
qualitative and quantitative analysis, it appears that the concerns or focus might be better
directed at the relative balance (i.e. in relation to one another, cf. Sperber & Wilson, 1986)
between the major organizing and evaluative broad types – at least in the argumentative corpora
(irrespective of the mode/ modality).
From the rhetorical perspective, given the data show systematic patterns that hopefully
can be replicated, the more text-organizing metadiscourse (not restricted to coherence markers,
as mentioned in the beginning of the summary) would be well received and assessed by the
composition teachers or oracy trainers. And again caution is needed, as it is not the particular
type of metadiscourse that may ensure better rhetorical organization but, as it seems, the
constellation of metadiscursive codes that fulfill textual functions (rather than interpersonal).
Stylistics and communication studies’ proponents might object to this suggestion of mine.
However, all I am observing is that those texts that show better organization signals receive
better evaluations and so higher grades in the educational context (cf. Barton, 1993; Hyland,
2005; Xu, 2001). Also, all that is propounded here is the need for organizing metadiscourse for
helpful effect on the propositional flow and integration of information in educational context.
The fact that students overall use almost twice as much evaluative as organizing
metadiscourse in either modality is in no way normative. It does not mean that it is how it should
be or remain throughout their K-12 education. In fact, the descriptive analysis suggests otherwise.
As the quantitative data analysis suggests students grouped as a function of reading
ability (by high-MAT scores) show more text-organizing bracketing uses of metadiscourse,
which subsume explicit speech acts, topicalizers, reminders of material, perlocutionary
commentary and induction/deduction signaling codes. K-12 pedagogues would most likely
welcome and promulgate such uses in students’ academic writing (e.g., as audience-awareness
heuristics), since they clearly guide the reader as to the intentions and directions in which the
texts would unfold and reduce a cognitive load of text processing on the part of the prospective
reader. As research by Intaraprawat and Steffensen (1995) shows, text-organizing uses in student
essays, along with boosters and explicit attitude marking, correlate with quality of writing and in
general qualify writers for higher scores by raters and/or their teachers.
Effective uses of text organizing metadiscourse, like those of high MAT-scorers, may
refer to what Maurannen (2003) calls socialization to and via metadiscourse… and into
academic discoursing skills, where the premium is placed on such organizing metadiscursive
uses as ‘the problem/issue is”, “the reasons are...”, “the consideration” or “let’s consider” (also
being a specific metalanguage). All those formal aspects highlighting metadiscourse features
very much resemble or in fact mirror many CR students’ written and oral language in action.
Moreover, as argued by Kamalski et al. (2008) what facilitates integration of information is the
use of objective markers rather than subjective markers (that cause resistance to persuasion, incl.
attitudinal or epistemic stance marking). Hence, text-organizing metadiscourse, being objective
and devoid of evaluative stance marking, seems to be most subservient to educational purposes
that are after the very integration of information.
Another implication that might be prompted by the study as well as by meta-analysis of
empirical and theoretical studies that contributed to several conceptualizations crystallized in this
project seems the fact of great salience and high-value of the interpersonal or intersubjective
marking given to such metadiscursive uses by the young adolescent language users themselves.
If they intuitively adjust their registers and styles to incorporate more intersubjectivity markers to
their talk and considerably reduce it (also intuitively) in their writing, it seems that such uses
already assumed the rank of a norm for the young adolescent talk. Thus, as theory interprets this
pragmatic vagueness marking as an effort “to avoid social stigma” (cf. Eckert, 1988, 2004;
Stenström et al., 2002), hereby the argument is being made not only about the need for further
analysis but consideration of such intersubjectivity-vagueness marking for incorporation in
educational materials and teacher-student discourses (sic! they are already there as classroom and
secondary data analysis reveal) that would enable and enhance solidarity-signaling functions.
Furthermore, the findings related to the modality verbs as differentially used by children
in writing and talk (stronger obligation and normative modals in essays and in discussionsweaker so more polite), with their documented effects of attitudinal stance marking (e.g.,
Crismore, 1985; Precht, 2003) suggest that it may be worthwhile making students aware of their
stance marking features, especially as the corpus linguists argue these issues are usually not
within conscious awareness (cf. Biber, 1999; 2001). It may be reasoned to claim that with those
continuously changing demographics in the American educational context, and ever more
English-as-a- second-language, or English-language-learners (ESL/ELLs), the issues of verbal
modality in relation to stance and attitude (which are alien concepts in many non-English
languages) need be explicitly taught and incorporated into the curriculum of English or
composition classes in K-12 education.
Likewise, instruction in evaluative metadiscourse (incl. attitudinal stance too) could be
incorporated to improve student’s ability to assess and produce effective evaluations that would
not showcase exaggerated, blatantly subjective or excessive (too proliferate) uses (and relatively
balanced with organizing variants). This implication may be especially pertinent to effective
argumentation skills as evaluations constitute integral part of argumentation and persuasion in
general (inter alia, Anderson et al., 1997; Cacioppo & Petty, 1984; Crismore et al., 1993;
Reznitskaya et al., 2008).
Limitations, Replications, and Future Research
The study is not devoid of caveats though. The analysis might be improved by adoption
of a different approach and different measuring instruments (taxonomy). A less comprehensive
study would be less taxing and overwhelming to the researcher(s), and would allow for more in
depth analysis of a particular class of metadiscursive uses (as already shown in the studies of
hedges or attitudinal stance in American English). Also different types of texts or discourses, or
tasks can be analyzed too.
Moreover, the study uses only data from CR-exposed children’s discussions and essays,
i.e. the experimental group. For valid inferences to be made about the effects of CR small-group
discussions on discourse and metadiscourse, the control group is needed for comparisons to be
made. Thus, another experimental study design might include a writing task as pre-test prior to
intervention, and a post-test. Then the comparisons between both participating groups – CR and
non-CR exposed - may be analyzed for the determination of prospective carry-over effects from
discussion onto writing.
Also, as remarked earlier, this study’s generalizability has been limited by the fact that it
is not fully representative as not using the whole dataset. Therefore, the study does not fully
benefit from the major/ primary study design, where the schools were counter-balanced by socioeconomic status. The analysis of metadiscourse as a function of school and ethnicity suffers from
confounding ethnicity with school location, as one of the schools has a predominantly EuropeanAmerican population. Therefore, to ensure greater representativeness of the results the study
might be improved by more counter-balanced sample (e.g., in those socio-cultural aspects)
selected for the analysis.
The study might also follow only one strand or method of analysis, either quantitative, or
qualitative or micro-genetic, but with the bigger database to ensure the generalizability of the
inferences. Evidently, the database for this study has been limited to 77 students, which does not
allow for proportionate comparable groupings (by ethnicity, for example), as some statistical
analyses assume greater power with the bigger sample sizes. Also, more of the discussions could
be explored for the metadiscourse patterns in a micro-genetic approach, as this study has
investigated only two discussions – one medial and one more final in a series of ten discussions.
Metadiscourse has been investigated only in the CR students’ uses, while teachers’ metatalk has been left unexplored. Though qualitative observations of video-data and transcripts were
unavoidable and suffice to say that CR-group discussions and dynamics to a greater or lesser
extent have been contingent on those Teachers’ discourses, even though their role in CR
discussions (primarily operated by children themselves) has been that of a “coach from the side.”
Thus, metadiscourse could be researched for the occurrences of metadiscourse in teachers’
classroom talk, during their exposition parts of the lesson, their review lessons, ways of asking,
eliciting and responding to questions, and potentially even in their ways of interacting in and out
of the classroom.
By analogy, the study’s focus was on CR discussions in educational contexts. However,
in the absence of metadiscourse studies with American adolescents and children, the future
studies of metadiscourse may target their informal talk, informal groups not necessarily in
educational settings, i.e. in the fairground, in the scouts clubs, or the internet and socialnetworking, etc.
Final remarks
The study’s major assets reside in its comprehensiveness and scope of metadiscursive
functions and forms analyzed while combining various disparate strands of pragmatic and
sociolinguistic analysis, never before explored in American discourse in authentic educational
settings on such a scale and across two modalities of language use.
Moreover, the study goes beyond established theory and by incorporating
intersubjectivity along the extant textual and interpersonal metadiscourse (in the analysis of
written discourse) or organizing and evaluating (in oral discourse) as well as by combining scales
used for meta-talk and meta-discourse for an effective outcome, it seems to push the theory
Additionally, the study offers some valuable insights and sheds ‘rare’ light on the nature
of metadiscourse as used by young adolescents (rather than better investigated adult uses), which
“still remains under-theorized and empirically vague” (Hyland, 2005, p. IX) thus indicating that
it “has not achieved its explanatory potential or allowed analysts to operationalize it in real texts.”
This study not only examines real texts but also explores the intricacies of their making in
the making, in the process of meta-communicating their intended meanings, by exploring the
video-captured authentic discussions of Collaborative Reasoners. Moreover, the on-line
capturing of the communicative flow is systematically analyzed micro-genetically and compared
with the off-line communicative task of reflective essay-writing on the controversial issue
prompted by a reader-engaging story. The study then can be merited with insights into the nature
of young adolescents’ discourse and metadiscourse, their formation, interplay and modulation
(manipulation), as mediated by the Collaborative Reasoning paradigm.
Though the study’s cross-sectional design does not allow for inferences about the carryover effect from Collaborative Reasoning small-group discussions to the post-intervention
written task, the findings supported by earlier study of written metadiscourse by the same
students (Latawiec et al., in preparation; NB, revealing CR gains in implicit common-good
oriented attitude, boosters and reader engaging, and more varied connectives), help triangulate
data from cross-modality, cross-sectional and sociolinguistic findings that suggest gains from the
collaboration in the debate or small-group discussions by corroborating “participatory
appropriation” (Rogoff, 1995) of the diversified thinking and so “handl[ing] subsequent events
[e.g., of essay-writing] in ways based on that involvement in previous events” (p. 156).
Alexander, P. A., Fives, H., Buehl, M. M., & Mulhern, J. (2002). Teaching as persuasion.
Teaching and Teacher Education (18), 795-813.
Almasi, J. F., O’Flahaven, J. F., & Arya, P. (2001). A comparative analysis of student and
teacher development in more and less proficient discussions of literature. Reading
Research Quarterly, 36, 96-102.
Andersen, G. (1998). The pragmatic marker like from a relevance-theoretic perspective. In A.
Jucker & Y. Ziv (Eds.), Discourse markers: Descriptions and theory (pp. 147-170).
Amsterdam: Benjamins.
Andersen, G. (2001). Pragmatic markers and sociolinguistic variation: A relevance-theoretic
approach to the language of adolescents. Amsterdam/Philadelphia: Benjamins.
Anderson, R. C., Chinn, C., Chang, J., Waggoner, M., & Yi, H. (1997). On the logical integrity
of children’s arguments. Cognition and Instruction, 15(2), 135-167.
Anderson, R. C., Chinn, C., Waggoner, M., & Nguyen, K. (1998). Intellectually stimulating story
discussions. In J. Osborn & F. Lehr (Eds.), Literacy for all: Issues in teaching and
learning (pp. 170-186). New York, NY: Guilford Press.
Anderson, R. C., Nguyen-Jahiel, K., McNurlen, B., Archodidou, A., Kim, S.-Y., Reznitskaya, A.,
& Gilbert, L. (2001). The snowball phenomenon: Spread of ways of talking and ways of
thinking across groups of children. Cognition and Instruction, 19, 1-46.
Asterhan, C. S. C., & Schwarz, B. B. (2009). Argumentation and explanation in conceptual
change: Indications from protocol analyses of peer-to-peer dialog. Cognitive Science, 33,
Austin, J. (1962). How to do things with words. Oxford, UK: Clarendon Press.
Bachman, L. F. (1990). Fundamental considerations in language testing. Oxford, UK: Oxford
University Press.
Bachman, L. F. & Palmer, A. S. (1996). Language testing in practice. Oxford, UK: Oxford
University Press.
Bakeman, R., & Gottman, J. M. (1997). Observing interaction: An introduction to sequential
analysis. 2nd edition. New York, NY: Cambridge University Press.
Bakhtin, M. M. (1981). Toward a philosophy of the act (Trans. by V. Liapunov). Austin, TX:
University of Texas Press.
Barton, E. (1993). Evidentials, argumentation, and epistemological stance. College English, 55,
Beaugrande, R. de (1982). Psychology and composition: Past, present, and future. In M.
Nystrand (Ed.), What writers know: The language, process, and structure of written
discourse (pp. 211-267). New York, NY: Academic Press.
Beaugrande, R. de (1983). Linguistic and cognitive processes in developmental writing.
International Review of Applied Linguistics in Language Teaching, 21(2), 125-144.
Bennett, N., & Cass, A. (1989). The effects of composition on group interactive processes and
pupil understanding. British Educational Research Journal, 15(1), 19-32.
Bereiter, C., & Scardamalia, M. (1982). From conversation to composition: The role of
instruction in a developmental process. In R. Glaser (Ed.), Advances in instructional
psychology (Vol. 2, pp. 1-64). Hillsdale, NJ: Erlbaum.
Berry, M. (1981). Systemic linguistics and discourse analysis: A multi-layered approach to
exchange structure. In M. Coulthard & M. Montgomery (Eds.), Studies in Discourse
Analysis (pp. 120-145). London, UK: Routledge and Kegan Paul.
Bialystok, E. (1986). Factors in the growth of linguistic awareness. Child Development, 86, 498510.
Bialystok, E., & Ryan, E. B. (1985). Toward a definition of metalinguistic skill. Merill-Palmer
Quarterly, 31, 229-251.
Biber, D. (2001). Using corpus-based methods to investigate grammar and use: Some case
studies on the use of verbs in English. In R. Simpson & J. Swales (Eds.), Corpus
linguistics in North America: Selections from the 1999 symposium (pp. 101-115). Ann
Arbor, MI: The University of Michigan Press.
Biber, D., Johansson, S., Leech, G. Conrad, S, & Finegan, E. (1999). Longman grammar of
spoken and written English. London, UK: Longman.
Billig, M. (2003) Political communication. In D. O. Sears, L. Huddy, & R. Jervis (Eds.), Oxford
Handbook of Political Psychology (pp. 222–250). New York, NY: Oxford University
Blyth, C., Jr., Reckenwald, S., & Wang, J. (1990). “I’m like, ‘Say What?!’”: A new quotative in
American oral narrative. American Speech, 65, 215-27.
Bogaers, I. E.W.M. (1999). Managing gender through meta-talk, Linguistic online, 2, 1-14.
Retrieved July 5th, 2011.
Bokus. B. (Ed.) (2005). Studies in the psychology of child language - In honor of
Grace Wales Shugar. Warszawa: Matrix.
Bokus, B., & Garstka, T. (2009). Toward a shared metaphoric meaning in children’s discourse:
The role of argumentation. Polish Psychological Bulletin, 40(4), 193-203.
Bondi, M. (2005). Metadiscursive practices in academic discourse: Variation across genres and
disciplines. In J. Bamford & M. Bondi (Eds.), Dialogue within discourse communities:
Metadiscursive perspectives on academic genres (pp. 3-30). Tübingen: Max Niemeyer
Verlag .
Bondi, M. (2008). Emphatics in academic discourse: Integrating corpus and discourse tools in
the study of cross-disciplinary variation. In A. Ädel & R. Reppen (Eds.), Corpora and
discourse: The challenges of different settings (pp. 31-56). Amsterdam/Philadelphia:
John Benjamins B.V.
Britton, B. K., Glynn, S. M., Meyer, B. J. F., & Penland, M. J. (1982). Effects of text structure on
use of cognitive capacity during reading. Journal of Educational Psychology, 74, 51-61.
Brown, P., & Levinson, S. (1978). Politeness. Cambridge, UK: Cambridge University Press.
Brown, P., & Levinson, S. (1978b). Universals in language usage: politeness phenomena. In E.
Goody (Ed.), Questions and politeness (pp. 56-310). Cambridge, UK: Cambridge
University Press.
Brown, G., & Yule, G. (1983). Discourse analysis. Cambridge, UK: Cambridge University Press.
Cacioppo, J. T., & Petty, R. E. (1984). The elaboration likelihood model of persuasion. Advances
in Consumer Research, 11, 673-675.
Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second
language teaching and testing. Applied Lingustics, 1, 1-47.
Cameron, A. C., & Trivedi, P. K. (1998). Regression analysis of count data. New York, NY:
Cambridge University Press.
Carston, R. (1995). Truth-conditional semantics. In J., Verschueren, J-O. Östman & J.
Blommaert (Eds.), Handbook of pragmatics: manual (pp. 544-550). Amsterdam: John
Celce-Murcia, M., Dörnyei, Z., & Thurrell, S. (1995). Communicative competence: A
pedagogically motivated model with content specifications. Issues in Applied Linguistics,
2, 5-35.
Chafe, W. (1982). Integration and involvement in speaking, writing and oral literature. In D.
Tannen (Ed.), Spoken and written language: Exploring orality and literacy (pp. 35-53).
Norwood, NJ: ABLEX.
Chafe, W. (2001). The analysis of discourse flow. In D. Schiffrin, D. Tannen & H. E. Hamilton
(Eds.), The Handbook of Discourse Analysis (pp. 673-687). Oxford, UK: Blackwell
Publishers Ltd.
Channell, J. (1994). Vague language. Oxford, UK: Oxford University Press.
Chen, G. (2009, September, 24). How “collaborative reasoning” Could be the next public school
trend. Public School Review. Retrieved from
Cissna, K. N., & Anderson, R. C. (2002). Moments of meeting: Buber, Rogers and the potential
for public dialogue. Albany, NY: SUNY Press.
Clark, H. H. (1994). Discourse in production. In M.A. Gernsbacher (Ed.), Handbook of
Psycholinguistics (pp. 985-1021). San Diego, CA: Academic Press.
Clark, H. H. (1996). Using language. Cambridge, UK: Cambridge University Press.
Coates, J. (1988). Introduction. In J. Coates & D. Cameron (Eds.), Women in their speech
communities: New perspectives on language and sex (pp. 63-73). New York, NY:
Coates, J. (1998). Gossip revisited: Language in all-female groups. In J. Coates (Ed.), Language
and gender: A reader (pp. 226-253). Malden, MA: Blackwell Publishers.
Conley, T. M. (1979). Ancient rhetoric and modern genre criticism. Communication Quarterly,
27(4), 47-53.
Cook , G. (1989). Discourse. Oxford, UK: Oxford University Press.
Craig, R. T. (Jan, 2008). The rhetoric of “dialogue” in metadiscourse. Conference Papers.
National Communication Association.
Crawford Camiciottoli, B. (2005). The role of metadiscourse in university-level EAP reading
instruction. In J. Bamford & M. Bondi (Eds), Dialogue within discourse communities:
Metadiscursive perspectives on academic genres (pp. 87-102). Tübingen: Max Niemeyer
Crismore, A. G. (1985). Metadiscourse as rhetorical act in social studies texts: Its impact on
student performance and attitude. Unpublished doctoral dissertation. University of
Illinois, Urbana-Champaign.
Crismore, A. (1989). Talking with readers: Metadiscourse as rhetorical act. New York, NY:
Peter Lang Publishers.
Crismore, A., Markkanen, R., & Steffensen, M. S. (1993). Metadiscourse in persuasive writing:
A study of texts written by American and Finnish university students. Written
Communication, 10, 39-71.
Crismore, A. , & Vande Kopple, W. J. (1997). Hedges and readers: Effects on attitudes and
learning. In R. Markkanen & H. Schoder (Eds.), Hedging and discourse: Approaches to
the analysis of a pragmatic phenomenon in academic texts (pp. 83-114). Berlin: De
Crothers, E. J. (1978). Inference and coherence. Discourse Processes, 1, 51-71.
Crystal, D. (1997). The Cambridge encyclopedia of language. New York, NY: Cambridge
University Press.
Cutting, J. (2000). Analysing the language of discourse communities. Oxford, UK: Elsevier.
Dressman, M. (2008). Using social theory in educational research. New York, NY: Routledge.
Drew, P., & Heritage, J. (Eds.) (1992). Talk at work: Interaction in institutional settings.
Cambridge, UK: Cambridge University Press.
Duranti, A., & Goodwin, C. (Eds.) (1992). Rethinking context: Language as an interactive
phenomenon: Studies in the social and cultural foundations of language No. 11.
Cambridge, UK: Cambridge University Press.
Eckert, P. (1988). Adolescent social structure and the spread of linguistic change. Language in
Society, 17, 183-207.
Eckert, P. (2004). Adolescent language. In E. Finegan & J. R. Rickford (Eds.), Language in the
USA: Themes for the twenty-first century (pp. 361-374). Cambridge, UK: Cambridge
University Press.
Eckert, P., & McConnell-Ginet, S. (1995). Constructing meaning, constructing selves: snapshots
of language, gender, and class from Belten High. In K. Hall & M. Bucholts (Eds.),
Gender articulated: Language and the socially constructed self (pp. 469-508). New York,
NY: Routledge.
Edelsky, C. (2007). Taking a stand on NCLB. Language Arts, 84, 457-458.
Eggins, S., & Slade, D. (1997). Analysing causal conversation. London, UK: Cassell.
Eemeren, F. H., van, & Grootendorst, R. (2004). A systemic theory of argumentation: The
pragma-dialectical approach. Cambridge, UK: Cambridge University Press.
Fraser, B. (1990). An approach to discourse markers. Journal of Pragmatics, 14, 383-95.
Fraser, B. (1998). Contrastive discourse markers in English. In A. Jucker & Y. Ziv (Eds.),
Discourse Markers: Description and Theory (pp. 301-26). Amsterdam/ Philadelphia:
Garcia, G. E. (1992). Ethnography and classroom communication: Taking an “emic” perspective.
In L. Monaghan & J. E. Goodman (Eds.), A cultural approach to interpersonal
communication (pp. 294-317). Maiden, MA: Blackwell.
Gee, J. (1990). Social linguistics and literacies. London, UK: Falmer.
Geva, E. (2007). Conjunction use in school children’s oral language and reading. In R. Horowitz
(Ed.), Talking texts: How speech and writing interact in school learning (pp. 271-294).
Mahwah, NJ: Lawrence Erlbaum Associates.
Giles, H. (1973). Accent mobility: A model and some data. Anthropological Linguistics, 15(2),
Giles, H. (1977). Language, ethnicity, and inter-group relations. London, UK: Academic Press
Givón, T. (2005). Context as other minds: The pragmatics of sociality, cognition and
communication. Amsterdam: John Benjamins.
Goldberg, J. A. (1983). A move towards describing conversational coherence. In R. T. Craig &
K . Tracy (Eds.), Conversational Coherence (pp. 25-45). Beverly Hills, CA: Sage.
Goldman, S. R., & Murray, J. D. (1992). Knowledge of connectors as cohesion devices in text: A
comparative study of Native-English and English-as-a-Second-Language Speakers.
Journal of Educational Psychology, 84, 504-519.
Goodwin, C. & Duranti, A. (1992). An introduction. In A. Duranti & C. Goodwin (Eds.).
Rethinking context: Language as an interactive phenomenon: Studies in the social and
cultural foundations of language No. 11. Cambridge, UK: Cambridge University Press.
Goodwin, C., & Goodwin, M. H. (1992). Assessments and the construction of context. In A.
Duranti & C. Goodwin (Eds.). Rethinking context: Language as an interactive
phenomenon: Studies in the social and cultural foundations of language No.11.
Cambridge, UK: Cambridge University Press.
Graves, D. H. (1982). Patterns of child control of the writing process. In R. D. Walshe (Ed.),
Donald Graves in Australia – Children want to write (pp. 17-28). Exeter, NH: Heineman.
Green, L. (2004). African American English. In E. Finegan & J. R. Rickford (Eds.), Language in
the USA: Themes for the twenty-first century (pp. 76-91). Cambridge, UK: Cambridge
University Press.
Grice, H. P. (1975). Logic and conversation. In P. Cole & L. Morgan (Eds.), Syntax and
semantics – Speech acts, Vol. 3 (pp. 41-58). New York, NY: Academic Press.
Guillem, S. M. (2009). Argumentation, metadiscourse and social cognition: Organizing
knowledge in political communication. Discourse & Society, 20, 727-746.
Guzman, A. (2004). The role of connectives in written discourse. Dissertation Abstracts
International, 65, 08B.
Halliday, M. A. K. (1994). An introduction to functional grammar (2nd ed.). London, UK:
Edward Arnold.
Hample, D. (1985). Refinements on the cognitive model of argument. Western Journal of Speech
Communication, 49, 267-285.
Hample, D. (2007). The Arguers, Informal Logic, 27: 163–78.
Heath, S. B. (1982). Protean shapes in literacy events: Ever-shifting oral and literate traditions.
In D. Tannen (Ed.), Spoken and written language: Exploring orality and literacy (pp. 91118). Norwood, NJ: ABLEX.
Heritage, J., & Raymond, G. (forthcoming). Navigating epistemic landscapes: Acquiescence,
Agency and Resistance in Responses to Polar Questions. In J-P de Ruiter (Ed.),
Questions. Cambridge, UK: Cambridge University Press.
Hoey, M. (1994). Signalling in discourse: a functional analysis of a common discourse pattern in
written and spoken English. In M. Coulthard (Ed.), Advances in written text analysis (pp.
26-45). London, UK: Routledge.
Hoey, M. (2001). Textual interaction: an introduction to written text analysis. New York, NY:
Psychology Press.
Holmes, J. (1995). Women, men and politeness. London, UK: Longman.
Horowitz, R., & Samuels, S. J. (1987). Comprehending oral and written language: Critical
contrasts for literacy schooling. In R. Horowitz & S. J. Samuels (Eds.), Comprehending
oral and written language (pp. 1-52). London, UK: Academic Press.
Hyland, K. (2000). Disciplinary discourses: Social interaction in academic writing. Harlow, UK:
Hyland, K. (2005). Meta-discourse: Exploring interaction in writing. London, UK: Continuum.
Intaraprawat, P., & Steffensen, M. (1995). The use of metadiscourse in good and poor ESL
essays. Journal of Second Language Writing, 4, 253-72.
Jadallah, M, Anderson, R. C., Nguyen-Jahiel, K, Miller, B.W., Kim, I-H., Kuo, L_J., Dong, T.,
& Wu, X. (2011). Influence of a teacher’s scaffolding moves during child-led smallgroup discussions. American Educational Research, 48, 194-230.
Jucker, A. H., & Smith, S. W. (1998). And people just you know like “wow”: Discourse markers
as negotiating strategies. In A. H. Jucker & Y. Ziv (Eds.), Discourse Markers:
Descriptions and Theory (pp. 171-202). Amsterdam/Philadelphia: Benjamins.
Kamalski, Lentz, Sanders and Zwaan (2008). The forewarning effect of coherence markers in
persuasive discourse: Evidence from persuasion and processing. Discourse Processes, 45,
Karmiloff-Smith, A., Grant, J., Sims, K., Jones, M-C., & Cuckle, P. (1996). Rethinking
metalinguistic awareness: representing and accessing knowledge about what counts as a
word. Cognition, 58,197-219.
Kim, I., Anderson, R. C., Nguyen-Jahiel, K., & Archodidou, A. (2007). Discourse patterns
during children’s collaborative online discussions. The Journal of Learning Sciences, 16,
Kim, I., Anderson, R. C., Miller, B., Jeong, J., & Swim, T. (2011). Influence of cultural norms
and collaborative discussions on children’s reflective essays. Discourse Processes, 48,
Kintsch, W., & Van Dijk, T. A. (1978). Toward a model of text comprehension and production.
Psychological Review, 85, 363-394.
Kumar, S., & Miller, K. F. (2005). Let SMIL be your umbrella: Software tools for transcribing,
coding, and presenting digital video in behavioral research. Behavior Research Methods,
37, 359-367.
Lakoff, G. (1972). Hedges: A study in meaning criteria and the logic of fuzzy concepts. Papers
from the Regional Meeting. Chicago Linguistic Society, 8, 183-228.
Lakoff, R. (1975). Language and woman’s place. New York, NY: Harper & Row.
Latawiec, B. M. (2010). Text structure awareness as a metacognitive strategy facilitating
EFL/ESL reading comprehension and academic achievement. The International Journal
of Learning, 17 (5), 25-48.
Latawiec, B. M., Anderson, R. C., Nguyen-Jahiel, K., Ma, S., Kim, I-H., Kuo, L-J., et al. (in
preparation). Influence of oral discussion on metadiscourse in persuasive essays .
Leech, G. (1983). Principles of pragmatics. New York, NY: Longman.
Lautamatti, L. (1978) Observations on the development of the topic in simplified discourse. In V.
Kohonen, & N. E. Enkvist (Eds.), Textlinguistics, cognitive learning, and language
teaching (pp. 71-104). Turku, Finland: University of Turku.
Levinson, S. (1983). Pragmatics. New York, NY: Cambridge University Press.
Linell, P. (2009). Rethinking language, mind, and world dialogically: Interactional and
contextual theories of human sense-making. Charlotte, NC: Information Age Publishing,
Lyons, J. (1977). Semantics Vols 1 & 2. Cambridge, UK: Cambridge University Press.
Malinowski, B. (1923). Phatic communion. In J. Laver & S Hutcheson (Ed.), Communication in
face to face interaction (pp. 146-152). Harmondsworth, UK: Penguin Books Ltd.
Maltz, D. N., & R. A. Borker (1982). A cultural approach to male-female miscommunication. In
J. Gumperz (Ed.), Language and social identity (pp. 194-216). Cambridge, UK:
Cambridge University Press.
Markee, N. (2000). Conversational analysis. Mahwah, NJ: Erlbaum.
Mauranen, A. (2001). Reflexive academic talk: Obsevations from MICASE. In R. Simpson & J.
Swales (Eds.), Corpus linguistics in North America: Selections from the 1999 symposium
(pp. 165-178). Ann Arbor, MI: The University of Michigan Press.
Mauranen, A. (2003). But here’s a flawed argument: Socialisation into and through
metadiscourse. In P. Leistyna & C. F. Meyer (Eds.), Corpus analysis: Language structure
and language use (pp. 19-34). Amsterdam: Rodopi.
Mey, J. (2001). Pragmatics. Oxford: Blackwell.
Meyer, B. J. F. (1975). The organization of prose and its effects on memory. Amsterdam: NorthHolland.
McCabe, A. (1997). Cultural background and storytelling: A review and implications for
schooling. The Elementary School Journal, 97(5), 454-473.
McCarthey, S. J. (2002). Students’ identities and literacy learning. Chicago, IL: International
Reading Association.
McCarthey, S. J. (2008). The impact of No Child Left Behind on teachers’ writing instruction.
Written Communication, 20, 1-44.
Michaels, S. (1981). Sharing time: Children’s narrative styles and differential access to literacy.
Language in Society, 10, 423-442.
Millis, K. K., & Just, M. A. (1994). The influence of connectives on sentence comprehension.
Journal of Memory and Language, 33, 128-147.
Ochs, E., Schegloff, E. A., & Thomson, S. (Eds.). (1996). Interaction and grammar. Studies in
Interactional Sociolinguistics, 13. Cambridge, UK: Cambridge University Press.
Olson, D. R. (1977). From utterance to text: The bias of language in speech and writing. Harvard
Educational Review, 47, 257-281.
Olson, D. R. (1994). The world on paper. New York, NY: Cambridge University Press.
Overstreet, M., & Yule, G. (1997). On being inexplicit and stuff in contemporary American
English. Journal of English Linguistics, 25(3), 250-258.
Östman, J-O. (1981). ‘You know’: A discourse-functional study. Pragmatics & Beyond II:7.
Amsterdam: John Benjamins.
Paradis, C. (2000). It’s well weird. Degree modifiers of adjectives revisited: The nineties. In J.
Kirk (Ed.), Corpora galore. Analyses and techniques in describing English (pp. 147-160).
Amsterdam: Rodopi.
Petty, R. E., & Cacioppo, J. T. (1979). Issue-involvement can increase or decrease persuasion by
enhancing message-relevant cognitive responses. Journal of Personality and Social
Psychology, 37, 1915-1926.
Petty, R. E., & Cacioppo, J. T. (1980). Effects of issue involvement on attitude in an advertising
context. Proceedings of the Division 23 Program, 88th Annual American Psychological
Association Meeting, 75-79.
Petty, R. E., & Cacioppo, J. T. (1981). Attitudes and persuasion: Classic and contemporary
approaches. Dubuque, IA: Wm. C. Brown.
Precht, K. (2003). Great versus lovely: Stance differences in American and British English. In P.,
Leistyna & C. F. Meyer (Eds.), Corpus analysis: Language structure and language use
(pp. 133-151). Amsterdam: Rodopi.
Raymond, G., & Heritage, J. (forthcoming). Navigating epistemic landscapes: Acquiescence,
agency and resistance in responses to polar questions. In J-P de Ruiter (Ed.), Questions.
Cambridge: Cambridge University Press.
Reznitskaya, Anderson, R.C., Dong, T. Li, Y. Kim, I-H., & Kim, S-Y. (2008). Learning to think
well: Application of argument schema theory to literacy instruction. In C. C. Block, L. M.
Morrow & S. R. Parris (Eds.), Comprehension instruction: research based best practices
(pp. 196-213). New York, NY: Guilford Press.
Rodino, A., Gimbert, C., Perez, C., Craddock-Willis, K., & McCabe, A. (1991, October). Getting
your point across: Contrastive sequencing in low-income African American and Latino
children’s narrative. Paper presented at the 16th Annual Conference on Language
Development, Boston University, Boston, MA.
Roever, C. (2007). Teaching and testing pragmatics. In M. H. Long & C. J. Doughty (Eds.), The
Handbook of Language Teaching (pp. 560-577). Oxford, UK: Blackwell.
Rogoff, B. (1995). Observing sociocultural activity on three planes: Participatory appropriation,
guided participation, and apprenticeship. In J. V. Wertsch, P. del Rio & A. Alvarez (Eds.),
Sociocultural studies of mind (pp. 134-164). Cambridge, UK: Cambridge University
Rotherham, A. J., & Willingham, D. T. (2010). “21st Century” skills: Not new, but a worthy
challenge.” American Educator, 34(1), 17-20.
Sanders, T. J., & Noordman, L. G. M. (2000). The role of coherence relations and their linguistic
markers in text processing. Discourse Processes, 29, 37-60.
Scardamalia, M., Bereiter, C., & Lamon, M. (1994). The CSILE project: Trying to bring the
classroom into World 3. In K. McGilly (Ed.), Classroom lessons: Integrating cognitive
theory and classroom practice (pp. 201-228). Cambridge, MA: MIT Press.
Schegloff, E. (1992). Repair after next turn: The last structurally provided defense of
intersubjectivity in conversation. American Journal of Sociology, 97, 1295-1345.
Schegloff, E. A. (2007). Sequence organization in interaction: A primer in Conversation
Analysis I. New York, NY: Cambridge University Press.
Schiffrin, D. (1980). Meta-talk: Organizational and evaluative brackets in discourse. Sociological
Inquiry, 50(3/4), 199-236.
Schiffrin, D. (1987a). Discourse markers. Cambridge, UK : Cambridge University Press.
Schiffrin, D. (1987b). Discovering the context of an utterance. Linguistics, 25, 11-32.
Schiffrin, D. (1990). The principle of intersubjectivity in communication and conversation.
Semiotica, 80, 121-151.
Schiffrin, D. (2001). Discourse markers: Language, meaning, and context. In D. Schiffrin, D.
Tannen & H. E. Hamilton (Eds.), The handbook of discourse analysis (pp. 54-75).
Malden, MA: Blackwell Publishers.
Simons, H. W. (1994). Going meta: Definition and political applications. Quarterly Journal of
Speech, 80, 468-81.
Simpson, P. (1990). Modality in literary-critical discourse. In W. Nash (Ed.), The writing scholar:
Studies in academic discourse (pp. 63-94). Newbury Park, CA: Sage.
Smith, N., & Wilson, D. (1979). Modern linguistics: The results of Chomsky’s revolution.
Harmondsworthm, UK: Penguin.
Smitherman, G. (1981). “What go round come round”: King in perspective. In G. Smitherman
(Ed.), Talking that talk: Language, culture and education in African America (pp. 132149). New York, NY: Routledge.
Sperber, D., & Wilson, D. (1986). Relevance: communication and cognition. Oxford, UK: Basil
Stenström, A-B. (2007). Teenage talk: A London-based chat and discussion compared. In R.
Horowitz (Ed.), Talking texts: How speech and writing interact in school learning (pp.
113-133). Mahwah, NJ: Lawrence Erlbaum Associates.
Stenström, A-B., Andersen, G., & Hasund, I. K. (2002). Trends in Teenage Talk: Corpus
compilation, analysis and findings. Philadelphia, PA: John Benjamins B.V.
Swales, J. M. (2001). Metatalk in American academic talk: The cases of point and thing. Journal
of English Linguistics, 29, 34-54.
Tannen, D. (1980). Spoken and written language and the oral/literate continuum. Proceedings of
the sixth annual meeting of the Berkeley Linguistics Society. University of California,
Tannen, D. (1982). Spoken and written language: Exploring orality and literacy. Norwood, NJ:
ABLEX Pub. Corp.
Tannen, D. (1991). You just don’t understand. Women and men in communication, London, UK:
Virago Press Ltd.
Thompson, G., & Thetela, P. (1995). The sound of one hand clapping: The management of
interaction in written discourse. TEXT, 15, 103-127.
Van Dijk, T. A., & Kinstch, W. (1983). Strategies of discourse comprehension. New York, NY:
Academic Press.
Van Dijk, T. A. (2001). Multidisciplinary CDA: A plea for diversity. In R. Wodak & M. Meyer
(Eds.), Methods of critical discourse analysis (pp. 95-120). London, UK: Sage.
Vande Kopple, W. J. (1985). Some exploratory discourse on metadiscourse. College
Composition and Communication, 36, 82-93.
Vande Kopple, W. J. (1997). Refining and applying views of metadiscourse. Paper presented at
the Annual Meeting of the Conference on Composition and Communication (48 th),
Phoenix, AZ.
Vygotsky, L. S. (1986). The genetic roots of thought and speech. In A. Kozulin (Trans. & Ed.),
Thought and language. Cambridge, MA: MIT Press.
Waggoner, M., Chinn, C., Yi, H., & Anderson, R. C. (1995). Collaborative reasoning about
stories. Language Arts, 72, 582-589.
Weinreich, U. (1966). On the semantic structure of language. In J. Greenberg (Ed.), Universals
of Language (2nd ed., pp. 142-216). Cambridge, MA: M.I.T. Press.
Wenger, E. (1998). Communities of practice: Learning, meaning and identity. Cambridge, UK:
Cambridge University Press.
Wheeler, R. S., & Swords, R. (2004). Code switching: Tools of language and culture transform
the dialectally diverse classroom. Language Arts, 81(6), 470-480.
Wierzbicka, A. (2006). The concept of ‘dialogue’ in cross-linguistic and cross-cultural
perspective. Discourse Studies, 8, 675-703.
Wilkinson, I. A. G., Soter A. O., Murphy, P. K., & Li. J. (2009, August). Quality talk about text
to promote language, thinking, and learning. In P. Warwick (Chair), Understanding,
promoting and evaluating dialogue in the classroom. Symposium presented at the
thirteenth biennial meeting of the European Association for Research on Learning and
Instruction, Amsterdam, The Netherlands.
Wilson, D., & Sperber, D. (1993). Linguistic form and relevance. Lingua, 90, 1-25.
Wolfram, W. (2004). Social varieties of American English. In E. Finegan & J. R. Rickford (Eds.),
Language in the USA: Themes for the twenty-first century (pp. 58-75). Cambridge, UK:
Cambridge University Press.
Wünderlich, D. (1979). Foundations of linguistics. Cambridge, UK: Cambridge University Press.
Xu, H. (2001). Metadiscourse: A cross-cultural perspective. Nanjing: Southeast University Press.
Yule, G. (1996). Pragmatics. Oxford, UK: Oxford University Press.
Zhang, J., Anderson, R. C., & Nguyen-Jahiel, K. (2010). Language-rich discussions for English
language learners. Champaign, IL: Center for the Study of Reading.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF