Polyphonic Support for Collaborative Learning
Stefan Trausan-Matu1, Gerry Stahl2, and Johann Sarmiento2
1
Politehnica University of Bucharest,
and
Romanian Academy Research Institute for Artificial Intelligence,
13, Calea 13 Septembrie
Bucharest, Romania
abqoy6@r.postjobfree.com
2
Math Forum and the I-School at Drexel University,
3210 Cherry Street, Philadelphia, PA 19104, USA
abqoy6@r.postjobfree.com, abqoy6@r.postjobfree.com
Abstract. This paper argues that one reason for the success of collaborative
problem solving where individual attempts failed is the polyphonic character of
work in small groups. Polyphony, a concept taken from music, may occur in
chats for problem solving, transforming dialog into a thinking device :
Different voices jointly construct a melody (story, or solution) and other voices
adopt differential positions, identifying dissonances (unsound, rickety stories or
solutions). This polyphonic interplay may eventually make clear the correct
( sound ) construction. The paper illustrates the polyphonic character of
collaborative problem solving using chats. It also proposes prototyped software
tools for facilitating polyphony in chats.
1 Introduction
This paper is considering the role of polyphonic inter-animation of multiple voices in
collaborative learning. Inspired by the work of Mikhail Bakhtin, this idea shed new
light on the dialogic nature of discourse in human language. It would also have
consequences for the design of collaborative learning environments.
In polyphony, several voices jointly construct a melody (or a story, or a potential
solution in the textual-chat case) while other voices situate themselves on a
differential position, identifying dissonances (unsound, rickety stories or solutions).
This polyphonic game may eventually make clear the correct, sound solution.
The ideas are exemplified with chat excerpts for collaborative learning of
mathematics problem solving, investigated in the Virtual Math Teams (VMT) project
at Math Forum @ Drexel University. Inter-animation patterns in two dimensions were
discovered: longitudinal (chronologically sequential) and vertical, towards two
opposite trends: unity vs. difference. We consider that even individual thinking is also
an implicit collaborative (dialogic) process that involves multiple voices. However,
actual collaborations, in small groups of different personalities empower the dialogic
process.
Y.A. Dimitriadis et al. (Eds.): CRIWG 2006, LNCS 4154, pp. 132 139, 2006.
Springer-Verlag Berlin Heidelberg 2006
Polyphonic Support for Collaborative Learning 133
An environment for collaborative learning (that may be seen also as a groupware)
based on the polyphonic inter-animation principles is introduced. Several modules are
already implemented while others are in a final stage.
The paper continues by introducing discourse, the dialogic theory of Mikhail
Bakhtin and polyphony. The next section of the paper introduces Computer-
Supported Collaborative Learning (CSCL) and analyses the polyphonic welding of
longitudinal-vertical unity-difference dimensions. Software tools that support the
polyphonic inter-animation are presented in the fourth section. The paper ends with
conclusions and references.
2 Discourse, Dialogic and Polyphony
Learning may be seen as directly related to discourse building, as Sfard remarked:
rather than speaking about acquisition of knowledge, many people prefer to view
learning as becoming a participant in a certain discourse [11]. Koschmann [5]
emphasized the social dimension of learning and discourse, quoting Deborah Hicks
[4]: "Learning occurs as the co-construction (or reconstruction) of social meanings
from within the parameters of emergent, socially negotiated, and discursive activity"
(p. 136).
The above ideas follow the socio-cultural learning paradigm initiated by Vygotsky.
He has a permanently increasing influence on learning theories, stating that learning is
a social process, mediated by specific tools, in which symbols and especially human
language plays a central role [15]. However, he did not investigated in more detail
how the language and discourse are actually used in collaborative activities. It is the
merit of Mikhail Bakhtin to propose a sound theory of how meaning is socially
constructed.
Mikhail Mikhailovici Bakhtin extended Vygotsky s ideas in the direction of
considering the role of language and discourse, with emphasis on speech and dialog.
Bakhtin raises the idea of dialogism to a fundamental philosophical category,
dialogistics. For example, Voloshinov (a member of Bakhtin s circle who, according
to many opinions, signed a book written by his more famous friend because the
former has an interdiction to publish during Stalin regime) said: Any true
understanding is dialogic in nature. Understanding is to utterance as one line of
dialogue is to the next [14]. This is in consonance with Lotman s conception of text
as a thinking device [17], determining that: The semantic structure of an internally
persuasive discourse is not finite, it is open; in each of the new contexts that dialogize
it, this discourse is able to reveal ever new ways to mean [1].
Any discourse may be seen as an intertwining of at least two threads belonging to
dialoguing voices. Even if we consider an essay, a novel or even a scientific paper,
discourse should be considered implying not only the voice of the author. The
potential listener has an, at least, as important role. The author makes a thread of
ideas, a narrative. Meanwhile, in parallel to it, he must take into account the potential
flaws of his discourse; he must see it as an utterance that can be argued by the
listener. In this idea, discourse is similar to dialog and to music polyphony (in fact, it
should not be a surprise that different art genres like music, literature and
conversation have similar features), where different voices interanimate.
134 S. Trausan-Matu, G. Stahl, and J. Sarmiento
Discursive voices weave sometimes in a polyphonic texture, feature which Mikhail
Bakhtin admired so much in Dostoyevsky s novels. They are characterized by
Bakhtin as a plurality of independent and unmerged voices and consciousnesses [2].
However, polyphony is not only a randomly overlay of voices. It has also musicality;
it is in fact one of the most complex types of musical compositions, exemplified by
the complex contrapuntal fugues of Johann Sebastian Bach. When there is more than
one independent melodic line happening at the same time in a piece of music, we say
that the music is contrapuntal. The independent melodic lines are called counterpoint.
The music that is made up of counterpoint can also be called polyphony, or one can
say that the music is polyphonic or speak of the polyphonic texture of the music. [7].
In polyphonic music, the melodic, linear dimension is not disturbing the
differential, vertical harmony. Moreover, for example, in Bach s fugues, the voices
inter-animate each other. The main theme is introduced by a voice, reformulated by
the others, even contradicted sometimes (e.g. inverted) but all the voices keep a
vertical harmony in their diversity.
Starting from Bakhtin s ideas, we extend these ideas to collaborative learning.
Therefore, we will further describe how polyphony may arise in collaborative
learning and we will propose ways of supporting it in learning environments.
3 The Polyphony of Problem Solving Chats
3.1 Collaborative Learning in Virtual Math Teams
Computer and communication technologies offer now new possibilities for
collaboration, by virtualizing classroom group interaction. New types of artifacts like
hypertext, the World Wide Web, chats or forums of discussions, are changing the
classical learning scenarios. In addition to classical sheets of paper or blackboards for
drawing diagrams and writing formulas and sequences of problem solving steps,
computer animations, simulations or even virtual participants in the dialog (artificial
agents) may be used now for collaboration. It is extremely important to analyze the
particularities of discourse in this new context. A good example is the fact that in
chats we can much more easily use a multiple threaded discourse, similar to
contrapuntus in classical music than in face-to-face conversations.
The (VMT) research program investigates the innovative use of online
collaborative environments to support effective K-12 mathematics learning as part of
the research and development activities of the Math Forum (mathforum.org) at Drexel
University. VMT extends the Math Forum s Problem of the Week (PoW) service
by bringing together groups of 3 to 5 students in grades 6th to 11th to collaborate
online in discussing and solving non-routine mathematical problems. Currently,
participants interact using a computer-supported collaborative learning environment,
which combines quasi-synchronous text-based communication (e.g. chat) and a
shared whiteboard among other interaction tools.
At the core of VMT research is the premise that primarily, group knowledge arises
in discourse and is preserved in linguistic artefacts whose meaning is co-constructed
within group processes [10]. Key issues addressed by the VMT include the design
challenge of structuring the online collaborative experience in a meaningful and
Polyphonic Support for Collaborative Learning 135
engaging way, and the methodological challenge of finding appropriate
methodological approaches to study the forms of collaboration and reasoning that take
place.
3.2 Polyphonic Inter-animation in Chats
Let us consider the following problem:
Three years ago, men made up two out of every three
internet users in America. Today the ratio of male to
female users is about 1 to 1. In that time the number
of American females using the internet has grown by
30,000,000, while the number of males who use the
internet has grown by 100%. By how much has the total
internet-user population increased in America in the
past three years? (A) 50,000,000 (B) 60,000,000 (C)
80,000,000 (D) 100,000,000 (E) 200,000,000
This problem was one of an eleven problems set that were used for an experiment
in which a group of students had to solve first individually and after that
collaboratively, using chat. It was one of the two that were not solved individually by
any students but it was solved collaboratively.
Let us now consider a chat excerpt that includes the main utterances that
contributed to the finding of the solution (see figure 1).
Fig. 1. An excerpt illustrating the collaborative solution construction
Discourse begins with Dan s idea of starting from the 30000000 number specified
in the problem statement (line 357). It continues with Mic s problem solving
buffoonery (lines 360-364, 366 and 368-370), remarked by Cosi (line 365) and Dan
(line 367): Mic seems to start writing a reasoning but he only fakes, writing fragments
136 S. Trausan-Matu, G. Stahl, and J. Sarmiento
of the problem statement linked by a typical phrase and since . However, this
fake discourse fragment seems to belong to a mathematics speech genre and, even
being a pastiche, is continued by Hal which extrapolates the 1:1 ratio from the present
(as stated in problem) to the whole 3 years and advances 60000000 as a solution (line
371).
Mic continues the buffoonery (lines 372-375). After about one minute, Cosi s
(incorrect) utterance i think it's more than 60,00000 appears as a critique or as an
intuition of something wrong, of some kind of an unsuccessful story . Nevertheless,
after less than another minute, she realizes that her own supposition is wrong because
the ratio cannot be 1:1 or bigger.
The collaborative discourse enabled Cosi to solve the problem. She didn t solve it
in the first phase, when they had to solve it individually. However, when she listened
to the discourse proposing a solution (correct in the case of Dan s beginning proposal,
fake at Mic and wrong at Hal), she felt the need to put herself on a different position.
Therefore, the discourse acted as a tool, as an artifact that enabled Cosi to find the
correct answer.
Discourse in chat collaborative problem solving has an obvious sequential,
longitudinal, time-driven structure in which the listeners are permanently situated and
in which they emit their utterances in a threaded manner. In parallel with this linear
threading dimension, the participants situate themselves meanwhile also on a critical,
transversal (or differential) position. For example, in the excerpt considered in this
section, Dan s theme was continued by Mic s buffoonery, continued itself by Hal and
then contradicted by a first theme of Cosi that was eventually totally changed, in its
opposite. We could say that the critique of Cosi appeared as a need to bring the
harmony of a correct solution.
In this longitudinal-transversal space, voices behave in an unity-difference manner.
This phenomenon is not specific solely to chats. It appears also to polyphonic music:
The deconstructivist attack according to which only the difference between
difference and unity as an emphatic difference (and not as a return to unity) can act as
the basis of a differential theory (which dialectic merely claims to be) is the
methodical point of departure for the distinction between polyphony and non-
polyphony. [6].
The unity and difference trends take different shapes in chat problem solving. We
can include in the unity category cumulative talk [8] or collaborative utterances [9],
repetitions [12], socialization or jokes. For example, many times participants in chats
feel the need to joke, probably in the need to establish a closer relation with other
participants, in order to establish a group flow state [3]. In fact, in all the chats we
examined there is a preliminary socialization phase, inter-animation appearing not
immediately after the beginning of chats.
4 Groupware for Polyphonic Inter-animation
Difference making has a crucial role in chats for collaborative learning, role which
may be best understood from a polyphonic, musical perspective. The possibility of
contemplating (listening), from a critical position, the ideas (melodies) of other
peoples and entering into an argumentation (polyphony of voices), enhance problem
Polyphonic Support for Collaborative Learning 137
solving and enables learning through a trial-error process. Such processes appear also
in individual problem solving (we can say that thinking is also including multiple
inner voices) but the presence of multiple participants enhance both the possibility of
developing multiple threads and, meanwhile, of differences identification. The inter-
animation of the multiple perspectives of the participants, the opposition as result of
contemplation and the presence of a third opinion in case of conflict, and sometimes
the synthesis it brings are a better asset to success than a multi-voiced discourse
performed by an individual (as inner thinking), that is inherently much less critique.
Evidence that participants permanently keep a differential position is also provided
by the statistics of personal pronouns usage in chat sessions. For example, in a corpus
of chats recorded in May 2005, I was used 727 times, much more than the usage of
we, with 472 occurrences. First person me was used 84 times comparing to us,
used only 34 times. However, the second person addressing is very well represented
by 947 uses of you .
Fig. 2. A summarization module that offers an abstraction of the flow of main ideas
A natural consequence of the theoretical considerations discussed above is the need
for a software support for small groups that facilitates polyphonic development. Such
a groupware, named POLYPHONY, is now under development. The system is built
around a chat system, which has some additional modules, not present in usual instant
messaging. These modules offer abstractions of the ongoing chat, in the idea of
making clear the flow of ideas and the other voices (the melody) and, the most
important, to induce polyphonic, differential ideas .
138 S. Trausan-Matu, G. Stahl, and J. Sarmiento
In figure 2, a snapshot of one of the first implemented modules, the summarizer, is
illustrated. This module builds a summary using natural language processing and
heuristics, It automatically assigns an importance score to each utterance, and selects
the most important utterances. Summarization is important in chats because knowing
what came before, starting from clear summaries would help people to respond, to
carry on the melody and to contribute to the polyphony with a personal, differential
voice.
In addition to the summarization module, other facilities for chats, based on natural
language processing are developed in POLYPHONY. They abstract and display facts
about each participant, for example, the emotional state, the degree of relevance of the
utterances of each participant. A module for speech acts identification has been
already implemented [13]. The goals aimed by these modules are to induce self-
reflection and images about the others, to facilitate inter-animation, and finally to
encourage multiple voices to enter into a polyphonic framework.
5 Conclusions
Discourse in chats implies an inter-animation of multiple voices along two
dimensions, the sequential, utterance threading and the transversal, differential one.
These two dimensions correspond to a unity-difference (or centrifugal-centripetal,
[1]) basic feature of polyphony. The unity directed dimension is achieved at diverse
discourse levels by repetitions, collaborative utterances, socializing and negotiation
discourse segments.
The second, differential dimension could be better understood if we consider
discourse as an artifact that, taking into account that every participant in collaborative
activities has a distinct personality, is a source of a critical, differential attitude. Even
if individual, inner discourse may be multi-voiced, difference and critique are
empowered in collaborative contexts, in a community of different personalities.
A consequence of the sequential-differential perspective for the design of CSCL
environments is that they must facilitate inter-animation not only on the longitudinal
dimension, through threading but also the transversal, differential, critical dimension.
Tools that may enter in this category should be able to provide abstractions or
summarizations of previous discourse, in order to facilitate differential position
taking. They should also allow the participants to emphasize the different proposed
themes and to relate them in threads, polyphonically.
Wegerif also advocates the use of a dialogic framework for teaching thinking skills
by inter-animation: meaning-making requires the inter-animation of more than one
perspective [16]. He proposes also that questions like what do you think? and
why do you think that ? in the right place can have a profound effect on learning
[16]. However, he did not remark the polyphonic feature of inter-animation.
Acknowledgements
The authors wish to express their appreciation to the members of the Virtual Math
Teams research project at Drexel University, whose voices are present in different
Polyphonic Support for Collaborative Learning 139
ways in the paper. They also want to thank to the anonymous reviewers for their very
useful remarks. The research presented here has been partially performed under a
Fulbright Scholar post-doc grant (awarded to Stefan Trausan-Matu) and was also
supported by the NSF grants REC 0325447 and DUE 0333493. Any opinions,
findings, or recommendations expressed are those of the authors and do not
necessarily reflect the views of the sponsors.
References
1. Bakhtin, M.M. (1981). The Dialogic Imagination: Four Essays, University of Texas Press.
2. Bakhtin, M.M. (1984). Problems of Dostoevsky s Poetics, Theory and History of
Literature Series, vol. 8, Minneapolis.
3. Csikszentmihalyi, M. (1990). Flow: The Psychology of Optimal Experience, Harper
Collins.
4. Hicks, D. (1996). Contextual inquiries: A discourse-oriented study of classroom learning.
In D. Hicks (Ed.), Discourse, Learning, and Schooling (pp. 104-141), Cambridge
University Press.
5. Koschmann, T. (1999). Toward a Dialogic Theory of Learning: Bakhtin s Contribution to
Understanding Learning in Settings of Collaboration, in C.Hoadley and J. Roschelle (eds.),
Proceedings of the Computer Support for Collaborative Learning 1999 Conference,
Stanford, Laurence Erlbaum Associates.
6. Mahnkopf, C.S. (2002). Theory of Polyphony, in Mahnkopf CS, Cox F & Schurig W
(eds). Polyphony and Complexity, Hofheim, Germany: Wolke Verlags Gmbh.
7. Polyphony (2005). http://cnx.rice.edu/content/m11634/latest/, retrieved on 4th May, 2005
8. Mercer, N. (2000). Words and Minds. How we use language to think together, Routledge.
9. Sacks, H. (1992). Lectures on conversation. Oxford, UK: Blackwell.
10. Schegloff, E. A. (1997). "Narrative Analysis" Thirty Years Later. Journal of Narrative and
Life History, 7(1-4), 97-106.
11. Sfard, A. (2000). On reform movement and the limits of mathematical discourse,
Mathematical Thinking and Learning, 2(3), 157-189.
12. Tannen, D. (1989). Talking Voices: Repetition, Dialogue, and Imagery in Conversational
Discourse, Cambridge University Press.
13. Stefan Trausan-Matu, Costin Chiru, Radu Bogdan, (2004). Identificarea actelor de vorbire
n dialogurile purtate pe chat, in Stefan Trausan-Matu, Costin Pribeanu (Eds.), Interactiune
Om-Calculator 2004, Editura Printech, Bucuresti, pp. 206-214.
14. Voloshinov (1973). Marxism and the Philosophy of Language, New York Seminar Press.
15. Vygotsky, L. (1978). Mind in society. Cambridge, MA: Harvard University Press.
16. Wegerif, R. (2005). A dialogical understanding of the relationship between CSCL and teaching
thinking skills. In T. Koschman, D. Suthers, & T.W. Chan (Eds.). Computer Supported
Collaborative Learning 2005: The Next 10 Years! Mahwah, NJ,.pp. 707-716.
17. Wertsch, J.V. (1991), Voices of the Mind, Harvard University Press.