CAMPS, A., RIBAS, T., GUASCH, O. & MILIAN, M.

Metalinguistic Activity: the link

between writing and learning to write In : Camps, A., & Milian, M. (1999).
Metalinguistic activity in learning to write. Amsterdam: Amsterdam University Press.

See discussions, stats, and author profiles for this publication at:
https://www.researchgate.net/publication/266889981

Metalinguistic Activity: the

link between writing and
learning to write

ARTICLE

CITATIONS READS

6 87

4 AUTHORS, INCLUDING:

Oriol Guasch
Autonomous University of …
26 PUBLICATIONS 18 CITATIONS

SEE PROFILE

Marta Milian
Autonomous University of …
22 PUBLICATIONS 24 CITATIONS

SEE PROFILE

Available from: Anna Camps

Retrieved on: 10 February 2016
Metalinguistic Activity: the link between
writing and learning to write

Anna Camps, Oriol Guasch, Marta Milian, Teresa Ribas

Universitat Autònoma de Barcelona

ABSTRACT
Teachers involved in the teaching of written composition are aware of the
fact that this is one of the most important challenges in present day educa-
tion. Current didactic models in this area propose the creation of contexts
in which text composition has a double objective: facing rhetorical prob-
lems in a specific writing task, and learning contents which facilitate solv-
ing these problems. Teaching and learning sequences - didactic units or
sequences, (DS) - designed for this purpose, also try to enhance interactive
situations among students as well as among students and teachers. These
interactive situations make language become a learning object and a learn-
ing tool at the same time. In studies that analyse learners' oral production
during written composition tasks, the importance of this metalinguistic
activity has been observed (Camps, 1994b; Bouchard, 1996 ;David & Jaffré,
1997; Camps & Ribas, 1996a; Schneuwly, 1995). This metalinguistic activity
comes up in different ways and is also produced at different levels of con-
sciousness. In this chapter we try to show how rich and complex students
metalinguistic activity is, as well as how it relates to the features that char-
acterise composition contexts in school, where texts are produced.

I METALINGUISTIC ACTIVITY IN THE COMPOSITION PROCESS:

SIX ASSUMPTIONS

With respect to the field of teaching and learning and, specifically, in the field of
research in learning processes of written composition, we will refer to some con-
cepts from linguistic and psychological studies that we consider useful. These
concepts may contribute to understand what occurs in classrooms where lan-
guage is taught and to set procedures to improve the students' oral and written
competence.The six main assumptions on which our work is based are the fol-
lowing:
1. Both language production and comprehension activities involve control proc-
esses to monitor them, and imply a certain ability to analyse the language used.
That is, metalinguistic activity is inherent in the use of language. Our initial
research hypotheses are based on Karmiloff-Smith's and Gombert's models
concerning the development of a reflective activity on language (Karmiloff-
Smith, 1986, 1992; Gombert, 1990).According to them, metalinguistic
activity is carried out at different levels of explicitness, from a simple
manipulative activity on language to an explicit verbalisation of the reflective
i i i
104 CAMPS, GUASCH, MILIAN & RIBAS

metalinguistic terms. Gombert's model takes into account external manifesta-

tions of this activity enhanced by verbal input in social interaction.
2. Representations on language are built, like language itself, in the frame of so-
cial interaction and it is those interactive situations which also enhance both
language learning and the construction of knowledge of language upon it,
together with the possibility of an autonomous control of written production
(Vygotsky,1978). In this sense, the school is a privileged environment in which
language is taken as object of conversation, observation and analysis, not only
from the language arts perspective, but also in all oral activities in the school
context that imply the use of new genres (scientific, literary, etc.)
(Kress,1985).
3. A specific objective of the language arts classroom is the acquisition and mas-
tery of the uses of language and the concepts that are culturally associated to
the idea of knowing language. Thus, this is a privileged environment for the
development of metalinguistic activity. We relate this development to the pos-
sibilities of improving students' oral and written competence, especially in
genres related to academic subjects and formal situations.
4. School contexts foster the use of written language and consider it the focus of
observation and reflection. Consequently, the teaching and learning situations
aimed at writing in and for meaningful contexts which result in a great
amount of metalinguistic activity. Collaborative work is revealed as a powerful
tool to enhance the metalinguistic activity in this framework (Camps, 1994b;
Altai, 1993; Camps & Ribas, 1996a; Milian, 1996).
5. Explicit metalinguistic activity in social interaction can be accompanied by
the use of a specific metalanguage.This use is not inherent in the activity, yet
learners can refer to the language they use by using everyday terms. However,
the progress in language competence will require the learning of concepts
which language sciences have systematised at different levels. This progress
requires the adequate knowledge and use of specific terms to refer to these
concepts.
6. Related to the first assumption, metalinguistic activity consists of both the
use of specific terms and of the use of everyday language to talk about
language. There is also non-explicit metalinguistic activity, as shown, for
instance, by the implicit activities of comparison or substitution that are
carried out by the students in a shared process without explicitly referring to
them (Bouchard, 1993).
To clarify the different notions embedded in the term `metalinguistic', we refer to
the knowledge of language as metalinguistic knowledge, to the use of
language to speak about language as metalinguistic function of language, and to
the discursive activity on language as metalinguistic activity.This activity is
considered as the permanent source of metalinguistic knowledge, which
contributes to build it and to activate it.We conceive this metalinguistic activity
both as developing in an interactive situation (interpsychological) and also as
i l i i (i h l i l) i h f f b l h h
LINKING WRITING AND LEARNING TO WRITE 105

2 WRITING AND LEARNING TO WRITE IN SCHOOL SETTINGS

The teaching and learning model of written composition, on which our
research is based, is conceived as a confluence of two main principles.The first
principle is that learning to write is achieved only by writing, that is to say that
writing activities are to be considered as a goal and as a means to learn how to
write. The second principle states that students in school and in language arts
classrooms are involved in writing activities to learn how to write, for it is one of
the main institutional issues. It is assumed, consequently, that students carry out
two kinds of activities when they are involved in writing.We distinguish a text
composition activity, with its own objectives and goals as a communicative activity
and a learning activity concerning specific items related to the written genres
involved in the task.
Complex relationships are established between both activities, and they al-
low us to hypothesise that actions for teaching and learning can be embedded
precisely within the composition process. Activation of metalinguistic and
metacognitive operations comes from the relationship between the teaching and
learning process and the written composition process. Such operations allow
the individual to become aware of the communicative activity and of the learn-
ing activity.
Within the area of our research in written composition teaching, we pose two
basic questions:
1. Which conditions may make the use of written language to be a tool for the
students' personal development and, hence, also a tool for the students'knowl-
edge and thought development?
2. What kinds of activities on language are, or might be, learning tools to achieve
the former objective efficiently? What kind of metalinguistic activities related
to the use of written language are useful and necessary to learn and know
how to write?

2.1 General characteristics of the model of didactic sequences (DS)

Research on the teaching and learning of written language initiated by DDLL1

starts with some basic assumptions that are also objects of research. Research is
based on composition teaching and learning situations in the classroom in natural
settings. Situations of written composition teaching, following the DS model, are
based on four fundamental principles:
1. The development of specific learning objectives related to the discursive genre
which students are going to work upon.
2. The insertion of writing composition tasks in real communicative situations.
Writers must have an intention, a reader and a context to place the text that
they are going to elaborate.
3. The need to promote interactive situations (stages, phases, moments) during
the composition process, be it among peers or in student-teacher
conference, or both; be it along the whole process or at definite sub-processes:
planning, revision, etc, in order to construct and regulate knowledge on writing
composition.
106 CAMPS, GUASCH, MILIAN & RIBAS

4. The need to give the students models of written language, not to be

imitated,
but to be known and recognised as social tools. Students have to be aware of
the characteristics that define the discursive genre language education aims to
deal with: discursive features, textual features, and linguistic features.
The DS model focuses on the writing process and on its control from the writer's
perspective, enhancing autonomous learning.The teacher and students share the
design of the sequence in the preparation phase, in which students are going to
elaborate a first representation of the task.The task is presented as a real writing
task - a production project - that gives sense to the students' activity.The
purpose of the activity acts as a motivating factor to engage students in the task:
an interview to somebody interesting to them, a letter on a popular topic to the
local newspaper, an advertising campaign to present environmental products
etc.
However, this is not the unique goal of the task that the students are
expected to fulfil.There are also specific learning objectives, which are related to
the more general goals. For example, in a didactic sequence on argumentation,
learning objectives within the sequence could be the formal structure of
argumentative texts, how to integrate concessive structures in the argumentative
text, or how to deal with argumentative syntactic connectors. In a DS focussing on
narrative, some leaning goals would be how to deal with the order of events in the
story, how to handle different narrative voices, how to master verb tenses in the
narration, or how to present different characters in the story.
From this first representation, involving both kinds of goals, communicative
and learning, some criteria to guide the task are derived.These criteria are
going to be built along the sequence and will serve as elements to control the
task during the process and to assess it when the task is accomplished.
The production phase may follow different paths, according to the specific
characteristics of the learners, the context, the task, the content, the syllabus etc. A
characteristic of the model is its flexibility to cope with such variation.
Assessment is carried out along the process, and it is made explicit through
instruments that allow the criteria to be assumed and shared by the participants
(members of group work, teacher and students). It contributes to build knowl-
edge on the writing task and the writing process to make the learners aware of
different kinds of problems and to solve them.We conceive these kind of proce-
dures as belonging to the framework of formative assessment (Allal, 1993;Allal &
Saada-Robert, 1992; Perrenoud, 1991).
The evaluation phase is meant to fulfil two different objectives which are, on
the one hand, related to the production goal: the text is going to be sent to the
newspaper, shown or handed to the intended audience, etc.; and, on the other
hand, related to the learning goal and achieved through some specific instruments:
tests, questionnaires, public explanations of the work done etc. The aim of this
phase is to make the students become aware of what has been done and what has
been learnt.They review the process in some way to be able to re-use or to re-
apply the experience in future sequences/projects in school, or in real situations
where the same type of knowledge is required.

The model contributes to the improvement of the teaching of writing

i
LINKING WRITING AND LEARNING TO WRITE 107

didactic sequences and to follow the process step by step, observing the learning
and teaching process as it develops along the sequence, and giving way to the
teacher's intervention whenever it is necessary to control and analyse whatever
specific operations are embedded in the writing process. It is also an instrument
for research on writing, be it in the research related to the activity of teaching -
action research -, or in the framework of writing composition research. It may
focus on specific aspects, for example:
• validity of assessment instruments as self-regulative tools;
• reliability of task representation criteria along the process;
• differences in knowledge acquisition in writing as a result of options in the
production phase;
• metalinguistic activity that contributes successfully to the writing process: what,
when, and how;
• teacher intervention: how and when;
• curriculum design: building a coherent progression within a defined discur-
sive genre through didactic sequences.

3 RESEARCH DESIGN
The study that we present here is based on some observations raised up in previ-
ous studies: the study by Camps (1994h) showed that collaborative writing
leads to complex composition processes.These processes involve an intense
cognitive activity on topic contents as well as on the elements that constitute the
discursive situation, and also on linguistic and textual forms that convey them.
Furthermore, Camps observed that the writing process in group work causes the
appearance of utterances orally formulated but with the intention to be written,
which she calls `attempted text'.2 One of the aspects of research on metalinguistic
activity in the groups that we will analyse is based on these kinds of utterances
and on the ones that could already be considered 'written text', since they are
dictated or read by some member of the group. The study carried out by Milian
(1995) on explanatory text writing shows how children write and revise their
texts to accomodate them to the reader, and how this process entails an intense
metalinguistic activity. From all the collaborative works and experiences carried
out with teachers at different levels of education, we can argue that situations
like the ones that give rise to the DS can favour the interrelation between the
writing activity and the linguistic and rhetorical contents planned as learning
objectives.These linguistic and rhetorical features will therefore be, in one way
or another, the centre of metalinguistic activity.

Our research methodology follows the lines of qualitative research paradigms

used in language teaching and learning research. In this framework, Nunan (1992)
recognises that research in the field of language education has developed some
specific instruments of observation and analysis of classroom interaction to ap-
proach the way in which linguistic and discursive knowledge is built. Discourse
analysis and conversation analysis become powerful instruments to reach these
issues in interactive situations in school contexts, that is, to infer mental opera-
tions from verbal and communicative processes and to analyse the teaching
and
108 CAMPS, GUASCH, MILIAN & RIBAS

learning situations that may foster metalinguistic activity as giving way to

metalinguistic knowledge.
The work we present is based on the analysis of the metalinguistic function of
language observed in the verbal interactions of six groups of two or three second-
ary school students from different schools along a collaborative text composition
process in the Catalan language classroom. Interactions among students were tape-
recorded in all the groups; two groups in each school (six groups in total) were
then randomly selected to be transcribed. Protocols refer to the conversations
among students in these groups during the joint composition task of what we call
draft or initial text. Some students were Catalan native speakers and some others
were Spanish native speakers.3
Collaborative text composition was the central activity within several com-
plex sequences on written argumentation in which the students also completed
systematic exercises on the text topic content. During the DS, all the groups wrote
several drafts, and last version, although not all the groups followed the same
organisational structure in writing the drafts, nor used the same facilitating tools.
Groups and their work characteristics are explained in the figure below:

Figure 1 Schools, levels and groups analysed. Characteristics of the texts written
during the Didactic Sequence.

School Level No Group Text Getuc Text Addressee Topic Contents

& age of groups constellation Objective
BIX rd LSO 13 2 G1:3 girls Letter to the editor Publication Newspaper readers
Women at work
G2:3 girls Less defined
SIB 2' BUP 15 2 G1:2 girls Editor's opinion Publication Magazine for young Women at work
& 1 boy people readers
G2:3 boys Contents of the
EPB 7' EGB 12 2 G1:2 girls Advertisement Simulated CAMELcigarettes agreement be-
G2:3 girls Agency report manufacturers tween advertise-
ment agency and
mamlicturers

The objective of the didactic sequences was to elaborate an argumentative dis-

course and to learn the main characteristics of this discursive genre. Women's
status in the professional world was the topic content in two of the DS.
Students at BIX had to reply to a controversial letter that had been published in the
section `Letters to the editor' in a well-known newspaper, in which a woman
expressed her opinion on women at the work place from a personal and social point
of view. StudentGenrecussed that letter in groups and looked up other texts that
gave them a broader knowledge on the topic, in addition to knowledge drawn
out of their familiar and social circle as a reference point.The opinion formulated
by the different groups at BIX resulted in a letter to be sent to the newspaper as
a reply to the initial letter.

Students at SFB dealt with the same topic but from a different perspective. The
topic of discrimination against women in the working world was imanufacturers
LINKING WRITING AND LEARNING TO WRITE 109

be discussed.As a result of the discussion and further elaboration on the topic, a

dossier including texts on different views of the subject in focus was made to be
published in a magazine addressed to young readers.
The situation presented to the students at EPB was slightly different, not only
in terms of topic content, but also in terms of its fictitious character. Students
played the role of advertising agents that had to design an advertisement to be
published in a newspaper for the manufacturer of the pmduct.This situation was
presented to the students after they had visited an advertising agency, where they
could observe the process to be followed in designing an advertisement and the
argumentative relationship between advertiser-manufacturer-consumer. Each group
of students chose an actual advertisement that was already published.They
acted as advertising agents trying to convince the manufacturer of the quality and
efficiency of their proposal so that it would influence the target consumers.
They wrote a report on this advertisement addressed to the manufacturer of the
product to justify its characteristics.
The groups of students were formed as usual in school classes, following natu-
ral association ways with the teacher intervening in some cases in order to make
sure there was a balance among the participants in the different groups. Differ-
ences in performance among groups within the same school class, which we will
comment on below, can he explained in terms of group composition, but the
working dynamics of each group and their way of functioning during the whole
sequence are also responsible for the outcomes.

4 ANALYSIS OF METALINGTUSTIC ACTIVITY

In this section we will refer mainly to one aspect of the data gathered from the
three sequences described above: the metalinguistic activity of students while
composing written texts.We will also focus on the procedures used for the
study of metalinguistic activity and on the formulation of various hypotheses
that will have to be discussed in future studies.
As we already expressed before, we will take into account data referring to
several levels and manifestations of metalinguistic activity in the protocols of ver-
bal interactions among students during the composition process of the initial
text or draft.Aiming at a better understanding of metalinguistic activity and its
characteristics, we will base the analysis on two different types of data: 1)
utterances with metalinguistic function that come up in these conversations,
and 2) changes or reformulations that the text undergoes during the textualization
process. The analysis of utterances with metalinguistic function will be further
related to the changes in the text as it is elaborated, on the basis that both
aspects are to be considered as manifestations of metalinguistic activity.

4.1 Utterances with metalinguistic function (MU)

The way in which students refer to the text they are writing varies. It is sometimes
far from an explicit formulation using the metalanguage dealt with in the class-
room; that is, they do not use specific terms referring to language and discourse
110 CAMPS, GUASCH, MILIAN &r RIBAS

features. But the fact of referring to it can be considered as a sign that the text is
taken as an object. It is precisely because of that, that we are interested in taking
into account every utterance that fulfils this function of referring to linguistic
use in a broad sense; that is, to the linguistic system, to the rhetorical situation, to
the textual characteristics, etc. Therefore, we will leave aside any other utterances
that do not refer to language itself.
In this part of the analysis we will thus consider explicit metalinguistic utter-
ances, whether they are formulated in specific terms (see example I) or formu-
lated without using linguistic terms (see example 2), and whether the formula-
tion is appropriate (see example 3) or not (see example 4).

Example 1: Do we begin with the addressees or what? Or with the objective?

What do we do?
Example 2: These or this? No, listen, but these or this? (...) This or these?
Example 3: It's that... when doing it, when saying it in the future I don't know
what I imagined, as if the ad wasn't done yet.
Example 4: -... mentioned before. Is it with b or...?
- I think it's with b
-Wait! (...) It's always the same. Well, it doesn't matter, with v.

The interest that brings us to analyse these type of utterances is not only to verify
the amount or the level of metalinguistic activity, but, from a didactic point of
view, we are interested in getting to know the contents of this activity. As a
starting point, we took the categorisation proposed by Defrance (1994), though
we must assume that categorising is always a problematic phase in any research,
and hardly transferable to new data collected in natural settings and
belonging to other contexts. We consequently adapted Defrance's categorisation
model to the parameters of our research, leaving aside, for the moment,
utterances on topic contents.4 The categorisation on which we base our
observations is the following:
1. utterances on discursive features: writers/speakers, addressees, text function
and intention (D);
2. utterances on text structure, referring to the organizational features of text, or
superstructure (Van Dijk, 1978) (TS);
3. utterances on linguistic aspects: text cohesion, lexical aspects, sentence syn-
tax, etc. (L);
4. utterances on spelling and punctuation (S/P); and
5. evaluative utterances (E).

Table I shows a summary of the results referring to the different types of

metalinguistic utterances (MU) from the six groups analysed according to the
above-described categories. It includes the percentage of MU 's in relation to the
total of interventions.
Table 1 Interventions with metalinguistic utterances.

Interventions Intervention Mean number

sequences interventions
Group Content per sequence
BIX-1-P Discursive Features 29 7 4,1
Text structure 49 16 3
Linguistic aspects 6 6,1
37
Spelling 15 10 1,5
Punctuation 9 7 1,2
Evaluations 2 2 1
Total Metalinguistic Utterances 141 49 2,9
Total Interventions 49
Percentage Metalinguistic Utterances 25,8

BIX-2-P Discursive Features 10 7 1,4

Text structure 8 3 2,6
Linguistic aspects 6 2-
Spelling 0 3 -1
Punctuation 0 0 1,7
Evaluations 1 0
1
Total Metalinguistic Utterances 25
14
Total Interventions 227
Percentage Metalinguistic Utterances 11

5 1,6
SFB-2-P Discursive Features S
24 3,9
Text structure 94
8 4,3
linguistic aspects 35
8 3,7
Spelling 30
15 3
Punctuation
5 1
Evaluations 7 7
189 2,8
Total Metalinguistic Utterances 67
Total Interventions 952
Percentage Metalinguistic Utterances 19,85
1 1
SFB-2-P Discursive Features 1 3,6
36 10
Text structure 0
Linguistic aspects a 0 -
0 0 -
Spelling
Punctuation 2 1
2
Evaluations 0
0
13 -
Total Metalinguistic Utterances 39
Total Interventions 674 3
Percentage Metalinguistic Utterances 5,8
E113-1-13 Discursive Features 5 4 1,2
Text structure 19 4,8
Linguistic aspects 92 12 3,4
Spelling 41 1 3
Punctuation 3 10 1,1
11
 Evaluations 3 3 1
Total Metalinguistic Utterances 155 48 3,2
Total Interventions 48
Percentage Metalinguistic Utterances 3,2

EPB-2-P Discursive Features 7 6 1,1

Text structure 26 3,2
Linguistic aspects 85 9 4,6
•
Spelling 42 1 3
Punctuation 3 6 1,1
Evaluations 7 1 I
Total Metalinguistic Utterances 1 49 2,9
Total Interventions 145
Percentage Metalinguistic Utterances 923
15 7

All groups formulate explicit metalinguistic utterances, but not all of them do it at
the same rate nor on the same contents, even though it is worth mentioning that
results taken as a whole are high, bearing in mind that only explicit MU's are
considered. The results of the content analysis of MU's show that in almost all
groups the highest number of explicit metalinguistic utterances and of
sequences of explicit metalinguistic utterances refer to text structure (TS).A
possible explanation is that in all cases this aspect was central in the DS, and it
was also the aspect that offered specific terms, new for the students, to refer to
the elements of the argumentative text. In order to make these terms operative,
the students had to be aware of them and discuss them.The mean of interventions
in sequences on text structure is, in general, one of the highest among other types
of sequences referring to other topics (from 2.6 to 3.9 turns of intervention per
sequence).This shows that the problems of adjusting the text to the structure
proposed are sources of discussion or deliberation among group members.
Conversations on these concepts constituted a powerful tool to elaborate and to
internalise them.
It is interesting to observe the scarce number of utterances on spelling (S) or
punctuation (P) problems. It is obvious that at the moment of the first draft of
the text these problems are secondary.There are not many evaluative utterances
(E), either, and the few there are appear only in one turn of intervention, that is,
they do not generate any reply; they only consist of individual exclamations that
usually reflect metalinguistic activity, though non-explicit, as we will see in the
reformulations.
The contents of utterances categorised as linguistic (L) are of very different
types. Many of them refer to lexical aspects, either to make a term more precise, to
find the word in Catalan (both Catalan and Spanish speakers), or to fmd the ap-
propriate word to match the register they consider appropriate; some of them
refer to morpho-syntactic aspects, even though they rarely use grammatical terms.
In a future study, it would be interesting to analyse these utterances in more
detail.
In a first reading, it seems peculiar that the number of utterances on discursive
contents, which are also new to these students and mght be favoured by the real
or simulated situation in which students are immersed, is relatively low in all
groups. However, the data offered by the analysis of the reformulations will
the close relationship between linguistic and discursive aspects in metalinguistic
activity, although they do not constitute explicit MU's.

4.2 Reformulation in the production process of drafts or initial texts

Observing utterances with metalinguistic function we have been able to know

how they talk about language. In the analysis of protocols, however, we can see
that when they are writing collaboratively, they actively make changes in the text
proposals while they are proposing them. From our perspective, these changes or
reformulations being done during the proposal are a sign of metalinguistic activ-
ity.We have, then, an explicit metalinguistic activity by means of utterances with
metalinguistic function, and a further evidence of this reflective activity through
the operations done on text proposals.
The process followed by each group to convert the ideas and intentions of the
authors into text is extremely interesting from many points of view. In our case,
text reformulations that take place during the textualization process gave us ac-
cess to the operations of evaluating and of considering different aspects of the
text, often inaccessible to the writers themselves. The implicit activities of com-
parison or of substitution of one term by another entail a metalinguistic activity
that is not expressed through explicit utterances, but that is present at the origin
of these changes and can be retraced from its analysis.
In the following subsection we refer to two aspects related to reformulations.
First we will consider the concept of reformulation itself and we will present the
different types of reformulation that we have been able to establish from the
analysis of textualization episodes. Second, we will interpret the changes that stu-
dents make by analysing the meaning and the sense of reformulations.

Concept of reformulation
The textualization process, that is, the process that consists of shaping the topic
contents to be included in the text into linguistic and textual features, deserves
particular analysis. it is obvious that one fundamental aspect to become acquainted
with the students' learning process of writing is analysing the texts written by
them. But this analysis on its own does not allow us to discover the complexity of
the underlying composition processes. On the other hand, studies on composi-
tion processes, in general, stress the operations of the highest level at the hierar-
chic organisation of the operations: content planning, discourse planning, rhetori-
cal setting of the text, revision, etc. There are few studies on the specific
operation of giving linguistic shape to the text, that is, of finding the words to
express the contents that are selected to be written and place them one after the
other in a linear process that links the text as a whole.
The access to this level is one of the most difficult ones, given the characteris-
tics of the operations, which in expert writers are automated to a great extent,
and consequently the writer is not aware of them. In group situations, however,
many of the operations that constitute the textualization come to the surface for
several reasons. One reason is the need to reach an agreement, another the role
distribution among group members, which leads some of them to write while
others control the composition in many different aspects: lexical adequacy, spell-
ing, precision, relation to the content and to the rest of the text, etc.
We can state that reformulations are an indication of the metalinguistic activ-
ity that takes place in the writing process.They usually consist of small changes
that are introduced progressively to the text from a first proposal, but they can
also be repetitions to confirm that they agree with what is already proposed, or
they can also be rereadings of the written solution.This does not mean, though,
that these utterances that are being reformulated are always preceded or
followed by explicit comments. It does not mean that the writers themselves
control in a completely conscious way this transformation process, either.
However, we think that the fact that writing is developed in a group situation
favours that the degree of consciousness of metalinguistic activity be higher than
in individual writing situations.
In a previous study, Camps (1994a) identified two types of utterances in the
textualization process: attempted text (AT), oral utterances formulated to be writ-
ten, and written text (W7), oral utterances repeating the text that accompany the
actual writing.AT and WT may fulfil several functions. Note the following ones:

Functions of Attemputterances, ATen text

Functions of the attempted text Functions of the written text

To propose text In dictation:

To accept the proposal To remember the proposal
To repeat it to give oneself enough To dictate it to the writer
time to evaluate it To emphasise some phonographic features
To accept it after slightly modifying it In selfdictation:
To substitute the proposal by a new one To monitor one's own act of writing
To keep the other students involved in
the task
To keep the other students informed
about the text being written
In rereading
To evaluate what has been written
To uptake what is written to follow up
with new text
These two types of utterances,AT and WT, are closely related to each other and
it is difficult to consider them separately. In the process of formulation and reformu-
lation of proposals, the cooperation among group members is usually very close
and the final written utterance results from the contributions of all or of more
than one of the group members. In many occasions the voices coincide in an
intense and fluent dance of formulations and reformulations, while implicit or
explicit judgements are made on the proposals being uttered. In some occasions
there is a sort of dialectic struggle to make one's own proposal prevail.At other
times reformulations go together with the rejection of the prior formulation and
there is a comparison and an explicit evaluation of the proposal being
presented;
and sometimes the proposals are even discussed by means of explicit metalinguistic
reasons. Let us observe the episode (Figure 2) in which the sentence is written:
Mainly these last ones, since the price is quite expensive and this makes fewer
young people buy it, which appears in the final text.

Figure 2 Collaborative textualization episode (EPB-1).

370 0: ... Mainly these last ones. since the price and so on and so forth ..,
371 X: Yes.
372 0: (proposes) ... even though it is not necessary to leave young
people out because ... Camel's adventure spirit... well...
inspires.... like that... youth, for instance
373 X: Yes, yes, well. (proposes) Mainly this last one...
374 0: 300(
375 X: (writes)These or this ? No, listen, but these or this? (.,.) This or
these?
376 0: These
377 X: This or these...?
378 0: These. No: this Well, it doesn't matter
379 X: No big deal. (Repeats aloud to hear how it sounds) This last
one..., no: these last ones, coma.
380 0: These...
381 X: (writes)... last ones... on(-).Ah, no: mainly these last ones, coma.
382 0: (proposes) ... since...
383 X: (uptakes the proposal and goes on) since the price is quite high..
384 0: Better use scientific words, or... expensive or...
385 X: (accepts the proposal) ... is quite expensive
386 0: ... is quite high... and present day young people can't... (laughter)
387 X: You echoed my grandma, girl.... is quite expensive and... and...
388 0: ...and..
389 X: and quite (pronounced with Spanish phonetics) <unattainable>,
how is it pronounced?
390 0: (pronounced with Catalan phonetics) <unattainable> (laughter)
391 X: <unattainable> no. Let's see.
392 0: quite high... and.. and this
393 X: (proposes) ... and this makes. and this makes...
394 0: ... decrease.. well... that....
395 X: ... young people....
396 0: ... less young people buy it
397 X: (writing) ... less young people buy it
398 0. Period
When analysing the textualization episodes we can observe that the language
that accompanies the AT varies in terms of quantity and quality. On the one hand,
we often find direct reformulation, without any comment or justification, which
is either uttered by the same student or taken up by another peer. On the other
hand, there are reformulations which are accompanied by argumentative corn-
ments using metalinguistic terms, as we can see in the transcribed episode (Fig-
ure 2, turn 384).
In order to progress in the knowledge of metalinguistic activity in this
textualization process, we establish a categorisation that allows us to describe
reformulations according to whether or not they are accompanied by utterances
with explicit metalinguistic function and according to the characteristics of the
last mentioned utterances.The categorisation is shown in Figure 3.

Figure 3 Types of reformulations and examples.

Types of reformulations Examples
1.Simple reformulation.
There is no intermediate step between two EPB-1-P
subsequent utterances, the second one 218 X:This advertisement that we present
reformulating the first one,wholly or to ....
textualization 219 O:The advertisement that we present

2. Reformulation with an explicit 1 13-1-P

rejection of the first formulation, or 457 0:... to the smoking people already
eventually with a doubting expression that mentioned.I don't know What do you think?
may derive in an intense comparison 458 Smoking smokers.
activity among linguisticforms, with no I think to the people, no, to smokers.
explicit metalin- guist'ic comment. 459 0: already mentioned
460 X:To the target!
461 0:Wow!
462X ... to the target, to the target...
463 0...already mentioned before
EPB-1-P
3. Reformulation and/or comparison accom-
526 0: Let's see...As it can be seen...
partied bymetalinguistic cornments,without 527 X: Let's see.ee...In the advertisement
using specific metalinguistic terms.
it will be seen..
528 0:As it willby metalinguisticcomments,
without
529 X: No, you're right, since it is
4.Reformulation
4.8eformulation and/or comparison,with EPB-1-P
metalinguistic comments using explicit
383 X: ...since the price is quite high
metalinguistic terms.
384 0:Better use scientific words, or...
expensive or...
385 X: is quite expensive

From the obsO: Theion of reformulations as a whole, it is apparent that students

rarely justify the options they take with metalanguage terms.There are few
cases of type 4 in textualization. episodes. Everything occurs as if the knowledge
on the text and the task were shared by participants, and thus it is implicit.
Otherwise, the changes are justified by non-strictly linguistic reasons (ex..In8Th
in Figure 2). There is only one case in which a grammatical term appears:
'otherwise there is no grammatical subject..:. On the contrary, it is clear that the
whole process in itself is a metalinguistic activity that shows the students'
perception of textual problems.
Obviously, we cannot make any implications with respect to the level of the sub-
jacent cognitive activity, especially in those cases in which students are not
aware of this activity. With respect to several aspects, students might be at a
purely epilinguistic level: Thus, they do not have any possibility to verbalise the
problem they arc facing, or it could also be that this problem is so automated and
shared by those peers that they do not feel the need to make it explicit.

The number of textualization episodes with reformulations differs from one group
to the other (Table 2).This result is related to the number of metalinguistic utter-
ances in every group (Table 1).We should stress the low number of textualization
episodes with reformulations as well as the number of MU's in some groups.The
mean of interventions per episode gives us information on the length of the dis-
cussion on a given aspect. The dynamics of the task in every group, the experi-
ence in group work or the composition of the groups are some of the factors that
may contribute to explain these differences. Nevertheless, the parallel between
the rate of reformulations and the number of MU's in the same group helps to
establish a close relationship between explicit metalinguistic activity (MU) and
implicit activity (reformulations) in the writing process.

Table 2 Number of reformulation episodes and number of utterances with

reformulations in every group.
Group Number of Percentage of
Number of Mean of
interventions with MU's
textualization interventions with
reformulations (see table 2)
episodes with reformulations
reformulations per episode

B1X-1-P 4 68 17 25.8
BIX-2-P 5 43 8.6 11
SFB-1-P 6 157 26.16 19.85
SFB-2-P 4 27 6.75 5.8
EPB-l-P 18 309 17.16
. 17.6
EPB-2-P 15 256 17.06 15.7

The episode transcribed above exemplifies some different types of metalinguistic

utterances as well as the different functions of the reformulations that we have
described. The interest of a microanalysis of the dialogue is obvious; it would
show us the complexity of this textualization operation in collaboration and how
this situation is precisely the one which causes or enhances the metalinguistic
activity. In order to observe more closely the type of changes that are formulated
in this cooperative work and in order to try to guess which sense they have in
relation to the objectives and contents of the didactic sequences, we devised a
graphic representation.We adapted this representation from Marty (1991). It only
shows the utterances that express the text to be written (AT) or that is being
written (WT). We have analysed the data in a way that allows us to observe the
relationship between the elements that are being modified in every reformula-
tion,BIX-1-Pg the reference to the group member that formulates every proposal.
There is an example in Figure 4, which we will comment on briefly.
 Figure 4 Example of reformulation.

566X we will utterly highlight This

this will be highlighted
568O the price
569 X the price will appear
570O will clearly appear
572 X because it is the differential profit

573 0 because differential

Text The price will clearly appear because it is the differential profit

Two girls are placed in a fictional situation where they are acting as an advertising
firm.They are preparing a report to the customer to present a proposal of publicity
to market the customer's product: Camel cigarettes.They have the advertisement
page in front of them and they try to back their proposal with some convincing
reasons. One of the girls in the group (X) begins this'. Theynce by making a
proposal to be written: 'We will utterly highlight. She refers to the figures
indicating the price of the product in a journal advertisement.The context of the
situation where they are working makes them point to the item they are looking
at and refer to it with a deictic determiner, 'This:They also use a personal pronoun
referring to themselves'decontextualizinghe girls in the role of publicists).There are
two reformulations in the sentence, the first one carried out by the same speaker
immediately after formulating a first proposal, 'this will be highlighted', and the
second one brought forth by her peer (0). Both reformulations have the sense of
'clecontextualizing' the text in relation to the physical context of production in
order to adapt it to the context in which it will be read.That is, they are
`recontextualizing' it: the readers will not actually have the page with the
advertisement already published as the girls have, so they cannot use the word 'this'.
Moreover, the new context is much more formal, and it forces them to
depersonalise their discourse, avoiding the use of 'we', and replacing it by an
impersonal construction, much more detached and objective. Finally, in the same
direction that forces them to disappear as writers, they suppress the passive form,
will be highlighted', and contribute to leave the speaker out of the report.The rest of the
sentence does not cause any more problems, the proposals being repeated by the
partner as an acceptance in 570O and accompanying the act of handwriting in
573 0.
Analysis of the reformulations
The observation of the reformulations along the process of composing argumen-
tative texts (see Figure 1) allows us to formulate some hypotheses concerning the
meaning of the changes done to the text in writing tasks within the framework of
the DS. In the groups analysed, we identify three types of factors that direct the
reformulations.
First, the search for a type of language appropriate to the features of written
language generally derives in changes aimed at finding a more formal and standard
register, avoiding redundancies and unnecessary elements, and eliminating the ref-
erences to the situational context of the task.Another type of change seems to
be
aimed at accomodating the text to its rhetorical context and, at the same time, to
the characteristics of the genre that is being composed. Finally, the third group of
changes responds to the willingness to express the ideas precisely and accurately,
to find the words and expressions that exactly convey the topic content.
It is worth mentioning that some of the reformulations done by students
seem to respond to more than one of the reasons mentioned above, as shown in
the examples below. It is also important to note that the process followed by
the students in a reformulation episode is not always linear; that is, the changes
made do not always respond to a progression towards the same objective. In some
cases, there are fluctuations in reformulating a sentence or part of it: the
students propose a solution in a determined sense and later on they reformulate
it again and come to another solution, different from the former one, though
presumably not as appropriate.
Example of fluctuations:
since always
since immemorial times
since the times...
since long ago
since ancestral times
since immemorial times
since the beginning of history
These fluctuations may occur due to the fact that the group consists of three
people, whose intentions with respect to the objective they are searching for
might not be exactly the same. But there are also cases in which the same speaker
expresses these fluctuations. This fact leads us to think that the formulation of
different proposals may often have the intention of trying or saying whatever
comes to their minds without much restriction, since listening to others or to
oneself proposing text can generate new ideas. Occasionally we can observe wea-
riness, since they do not seem to be able to find the satisfactory formulation.Then,
they usually propose several new formulations successively, with the intention of
testing rather than explicitly searching for a specific word

Reformulations which tend to a written register

Most of the changes observed in the groups of students that we analysed respond
to the search of a more formal language, detached from the colloquial spoken
language, and which contains features of a written standard.
Examples of searches for formal language:
we agree to do some things
we are in favour of to do other things
we support to do other jobs
to do others

grandparents nature has already imposed

elderly people they are already imposed by nature
third generation people
advanced in years people
 Something we find
some kind of task it is found
certain tasks

In these examples we can observe some substitutions of words and expressions

which students consider very colloquial or too close to their own way of express-
ing, and which they change by other forms which they consider more formal
and/or more 'written'. In the left column, we can observe reformulations of a
lexical type. In the right column, the changes affect the grammatical structure of
the sentence. These changes might consist of the pronominalisation of a noun
which had been already mentioned, of the passivisation of a sentence, and/or of
the elimination of the enunciation marks in the text.All of them are features which
the students have conceptualised as characteristics of the written genre.

Another type of reformulation, which can be the consequence of the search for a
register corresponding to the written mode, makes the language more precise
and to synthesise the content while avoiding excessive redundancies.
Examples:
We can say this in one word without having a life of one's own
In one word: without a life of one's own

Finally, another type of reformulation, which also tends to match the written reg-
ister, allows the text to become independent from its composition situation and,
thus, to turn into a more adequate outcome for the new rhetorical context.
Examples:
everything progresses except this
everything progresses except the most important
everything progresses except women
everything progresses except us, the most important

The word `this' refers to the idea that has been commented on during the former
conversation in the group, but it also needs to be made explicit so that the text
can be understood out of the situational context of the writing process.
In our study, we analysed groups of students in three different didactic se-
quences, thereby observing that they carry out a great number of
reformulations in the sense of adopting features of the written register. We can
say that they are the most frequent type of reformulations. It seems as if the
first verbalisation of the AT, maybe because it is uttered, still has some elements
typical from the oral register that members of the group quickly detect and
substitute by others more adequate to the written genre.

Reformulations towards the accommodation to the rhetorical context and to

the discursive genre
The fact that in most of the changes students adopt a more formal register could
lead us to believe that the reason for these changes is exclusively the move from
oral to written language.As observed previously, we believe that this is one of the
main reasons for the reformulations in the study. However, we can state that stu-
dents also search for a better adjustment of their text to the rhetorical conditions
and to the discursive genre.We could dearly observe this fact in a writing situa-
tion in which an argumentative text was composed.The text was to be published
in a magazine addressed to young students and managed by the students them-
selves. In this case, the text proposals tend to use a more direct and more collo-
quial language closer to the informal oral uses of young people (Camps & Ribas,
1996b).This can be seen in the next example, in which the three students look
for a sentence to link with the initial fragment of the text they have already
written:A man sitting on an armchair. A pint in his hands. Football on the TVAt the
same time someone is moving around in the kitchen. I would bet it's a
woman!':
Examples:
What does this mean?
What are you saying, guy!? It can't he like that
What are you saying, man!?
Do you think?
Do you think it's a woman?
Do you think?

Also in these groups we fmd reformulations aimed at including explicit enuncia-

tion marks in the text, so that it becomes more convincing.
Examples:
we are not in favour of sexism
as women, we arc not in favour of sexism
We can also find reformulations which look for a greater impact on the addressee;
that is, a text that has more strength to convince the reader.
Examples:
They have to share the chores
Chores have to be shared
Chores must be shared
They both create chores, then both have to solve them

We want a neutral situation

We don't want a radical situation
We talked about text adequacy to the discursive context because we are analys-
ing students' activities embedded in writing projects for real communicative situ-
ations. However, the texts composed by the students in a school setting usually
arc embedded in another context: the classroom, the place where students learn
and where they have to show what they know, and where the teacher is the
addressee, in one way or another, of all the texts composed.Thus, we could say
that some of the reformulations which we categorised as tending towards a writ-
ten register fulfil this second function at the same time: to adjust the text to the
language classroom context, in which work on formal uses of language is given
priority over colloquial uses.
Based on this, our hypothesis on the role of the rhetorical context in the
textualization process, once the reformulations in each group have been analysed,
could be stated as follows. All groups of students try to find a language that is
adjusted to what they have conceptualised as the written register and that most
of the times coincides with what the school situation expects from them. Further-
more, they also try to make their text functional and adequate to the rhetorical
context created by the Didactic Sequence.This second type of reformulation ap-
pears more in some groups than in others, probably for different reasons. When
the text to be written is a formal text, addressed to an unknown audience, we can
not fully appreciate the meaning of the reformulations. They can be confused
with the ones described in the first group; i.e. reformulations to accommodate the
text to the formal register of the written text. There may also be groups of stu-
dents that are more capable than others of visualising the rhetorical situation and
of making use of the necessary knowledge to accommodate the text to its param-
eters.

Reformulations which tend to express ideas with precision

There is a third group of reformulation, which as its main objective searches for
the right word to express adequately the idea in mind.At times, when a
member of the group is looking for a word and is not satisfied with the words
proposed by the other peers, she may try to justify her refusal with a semantic
explanation. She may propose other forms that in her opinion are adjusted to the
meaning that she wants to express. Below are some examples which are usually
accompanied by justifications in the dialogue.
Examples:
with the loss of .. to have a job and to be a housewife
with the loss of fertilisation to have two jobs
with the loss of fertility to have several jobs
to have more than one job
to have several jobs
to have more than one job, but at the same time

These types of reformulations, although not so frequent as the two mentioned

above, are present in all the groups analysed.The texts being written are embed-
ded in a rhetorical situation known to the students. Explicit work devoted to
generate contents on the topic has also been included in the sequences. These
two aspects might be responsible for the fact that students try to find the linguis-
tic form that better suits the exact idea they want to express.A consequence that
may be drawn from these observations is that to enhance metalinguistic activity it
is necessary to have specific ideas to explain and to have someone to explain
them to. Then a dialectic relationship between topic contents and the linguistic
and textual forms can be established (Bereiter & Scardamalia, 1987).

5 CONCLUSIONS
As a result of the empirical research work, we can conclude that:
I. The analysis reveals the existence of metalinguistic activity in the composi-
tion process in groups. This activity is explicitly shown through utterances
with metalinguistic function (MU), concerning very different aspects of the
text being elaborated and the communicative situation. In almost all the groups
the highest number of MU's refer to the structure of the argumentative text, a
 basic aspect of the objectives and contents of the didactic sequences analysed
in this study. We can state, consequently, that students' talk concerning
metalinguistic knowledge refers to the questions related to the learning goals
of the task: text structure, characteristics of the discursive situation.A charac-
teristic of this explicit activity is that the metalinguistic utterances are not
formulated in metalanguage terms but rather in the students' everyday lan-
guage.
2. The analysis also reveal that this activity does not only show up in explicit
metalinguistic utterances. There is also evidence of metalinguistic activity in
the reformulations on the text and the text proposals produced in these joint
textualization processes, even if this metalinguistic activity does not have an
explicit verbal expression. it is not possible, however, to know whether this
activity belongs to the epilinguistic level, as Culioli pointed out in his theory
of linguistic description, and whether it is in accordance with the models of
language representation and language knowledge stated by Karmiloff-Smith
and Gombert, referred to in section 1. It may as well be that this activity be-
longs to a level either of automation or of shared knowledge that makes ver-
balisation unnecessary. This is one of the aspects to be reviewed in research
that will require new methodological tools. However, the reformulation proc-
ess is a potential factor to move from the epilinguistic to the metalinguistic
level, as we will explain in point 5.
3. It has been observed that a great amount of the reformulations done by stu-
dents within the process, even though they act linguistically upon very differ-
ent elements, have a common tendency; that is, they tend to make the text
match the rhetorical situation.This leads us to the hypothesis that the changes
done through reformulating obey the need to accommodate the textual prod-
uct to the discursive situation in which it will be read.

Other conclusions concern the model of didactic sequences as an efficient model,

not only to enhance the learning and teaching of writing, but also to foster
metalinguistic activity through collaborative writing.

5. One of the most interesting conclusions for language teaching and learning
follows from the contents of the metalinguistic utterances and from the
analysis of the reformulations: Metalinguistic activity is closely related to the
components of the didactic sequence. The design of the sequences has
two interrelated axes: a composition activity embedded in an explicit
rhetorical context and a learning and teaching activity based on the text
composition process, in which several specific contents, referred to the
discourse and textual genre dealt with, are enhanced. This double activity
allows the interrelation of two contexts, the communicative context of the
rhetorical situation at work, and the school context with the specific aim of
acquiring knowledge on language and language use.We thus observe that the
school work on the characteristics of written language gives the students a
model of language which leads them to adjust their attempted text to it.At the
same time, however, the didactic proposals of writing embedded in a real
rhetorical situation provide the students with another context which can
determine the sense of the reformulations: to ensure their adequacy to the
possible readers and to the
 `frame' in which the text is going to be published or shown. Both aspects are
reflected in the pieces of evidence of metalinguistic activity. We hypothesise
that precisely the characteristics of the DS model and the characteristics of
their development allow the appearance, and condition the content of the
majority of the metalinguistic activity that takes place in their framework.
5. The analysis of textualization episodes with reformulations allows us to con-
firm the hypothesis that collaborative writing enhances metalinguistic activity
and reveals at the same time an important learning tool.The learning process
may be illustrated in the following way: one of the members of the group
makes a proposal. 'When, subsequently, another one formulates or reformu-
lates the former proposal, he is bringing about a comparative situation (even
if it is implicit).The need to make a decision between two alternative
possibilities generates a mental activity that will bring one of the partners to
consider both utterances and to choose the one he believes as more
suitable.Therefore, the individual proposals are in the air, in the interaction
context among them, and promote an intense activity in the individual minds,
a metalinguistic activity that probably would not exist when writing alone. Social
interaction among partners provides the frame for interpsychological sharing
of knowledge and leads to internalization and individual reflection. Thus,
collaborative writing becomes a learning tool.
In order to gain further insights, we need to carry out microanalysis studies, which
allow us to detect the appearance of conscious metalinguistic activity. In
addition it might lead us to a better understanding of the textualization process
itself, along with the diversity of functions of the utterances produced by learners
composing texts in groups.

6 Finally, it needs to be noted that, despite the fact that undergoing research
poses more questions rather than finding answers to the problems of language
teaching and learning, this research also provides us with interesting data, es-
pecially with respect to two aspects. First, it is a didactic research centred in
the dynamic processes of textual composition in the framework of learning
situations and, thus, it interrelates the writing composition process to the teach-
ing and learning process. Second, it opens methodological ways to gain more
insights on the complexity of the language learning processes in school set-
tings, especially since these processes up until now are not fully known.

NOTES
1 Departament de Didàctica de la Llengua i la Literatura. Universitat Autònoma de
Barcelona.
2 Some references to this concept of 'attempted text' can be found in Bereiter, Fine &
Gartshore (1979) and Matsuhashi (1981). Bereiter et al. refer to it as 'verbatim units'.
3 Both languages, Catalan and Spanish, are present at a similar rate in the personal
relationships among people in the Catalan society. The analysed schools, depending on
their geographical location, receive students predominantly from one or the other family
language. In all cases, though, the language at school is Catalan.
4 The process of establishing categories of analysis, besides following Defiance's proposal,
was done through consensus among the different researchers implied in the project
after a thorough observation and comparison of data from the different groups.