UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Chapter 24
Polyphonic Inter-Animation of Voices in VMT
Stefan Trausan-Matu and Traian Rebedea
Abstract This chapter introduces a theoretical framework for analyzing collabo-
rative problem solving in chats, based on the concept of polyphony and Bakhtin’s
theory of dialog. Polyphony, a notion taken from music theory, may be considered
as a general model for interaction and creativity by a group of people (“voices,”
in an extended sense) following patterns of counterpoint. As Bakhtin emphasized,
polyphony may occur in texts; we will show that it can occur in problem-solving
chat texts. One of the features of polyphonic music is its potential development
of complex architectures starting from a given theme. Polyphonic structuring of
dialogs may transform the interaction into a “thinking device”: Different voices
jointly construct a melody (story or solution), sometimes adopting different posi-
tions and then generating, identifying or solving dissonances (unsound, rickety sto-
ries or solutions). Polyphony consists of several “horizontal,” longitudinal melody
lines that are “vertically,” transversally integrated. Similarly, in chats, the continua-
tions of utterances are tied together over time providing a melodic line. Simultane-
ously, they are coordinated with the utterances of others, maintaining the integration
toward unity across various themes and variations that sometimes can introduce dif-
ferences. This chapter also proposes software tools for the visualization of the poly-
phonic weaving in chats. These tools identify and visualize the explicit and implicit
links among utterances, and may determine or visualize the contributions of each
participant in a chat.
Keywords Bakhtin · polyphony · voice · dialog · reference
This chapter introduces a theoretical framework, a method and a visualization tool
for analyzing CSCL chats, based on the ideas of polyphonic inter-animation intro-
duced by Bakhtin (1981, 1984a). As in the dialog theory of Bakhtin, we extend
the polyphonic musical model for analyzing language-based interactions, in our
case, transcripts of text chats for collaborative learning. Although Bakhtin’s ideas
S. Trausan-Matu (B)
Computer Science, Politehnica University of Bucharest, Romania
e-mail: stefan.trausan@cs.pub.ro
451G.G. Stahl, Virtual Math Teams, Computer-Supported Collaborative Learning Series 11,
DOI 10.1007/978-1-4419-0228-3_24, C© Springer Science+Business Media, LLC 2009

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
452 S. Trausan-Matu and T. Rebedea
are quite well known and considered as a theoretical starting point in the CSCL
community (Koschmann, 1999b; Stahl, 2006b; Wegerif, 2006), there are no elab-
orations that propose how to use his theory in practice. The analysis method we
introduce is inspired from the ideas of counterpoint, which is the theory and method-
ology used in music for composing and analyzing pieces for multiple instruments
or voices. Our theory and method was used for the implementation of a system
to analyze and visualize polyphonic threading in chats, proposing an evaluation of
the contributions of the participants. This polyphonic perspective shed new light
on the dialogic nature of discourse in human language and in problem solving in
general. It could also have consequences for the design of collaborative-learning
environments.
In polyphony, a number of melodic lines (or “voices,” in an extended, non-
acoustical perspective, as we will discuss later) jointly construct a harmonious
musical piece, generating variations on one or several themes. Dissonances should
be avoided and resolved, even if several themes (melodies) or theme variations
are played simultaneously, and even if sometimes the voices situate themselves in
opposing positions.
Bakhtin considers that multiple voices are present in texts, and sometimes (e.g.,
in Dostoevsky’s novels) they constitute a polyphonic framework (Bakhtin, 1984a).
Extrapolating this idea, we observe that inter-animation of voices following poly-
phonic patterns can be identified in dialogs generally, and in chats in particular.
A polyphonic collaboration involves several participants who play several themes
and their variations in a game of sequential succession and differing positions.
The existence of different voices introduces “dissonances,” unsound, rickety stories
or solutions. This polyphonic game may eventually facilitate knowledge building
through the tension of their opposition and the pressure to resolve the difference
(see Chapter 9).
Polyphony, in our view, may be taken as a model of collaboration, in which
several participants (“voices”) invent, discuss and elaborate ideas—often eventu-
ally achieving coherence even if “centrifugal” forces, divergences or differences
arise temporarily. In fact, as in physics, centrifugal forces or differences determine
a reaction of centripetal forces that act towards increasing unity. Bakhtin identified
this centrifugal/centripetal phenomenon in the discourse of novels (Bakhtin, 1981).
From a polyphonic point of view, these forces manifest themselves in two dimen-
sions: longitudinal and vertical (melody and harmony).
The above ideas are exemplified in this chapter with chat excerpts for collabo-
rative learning in two domains: mathematics problem solving—investigated in the
VMT Project—and human–computer interaction—studied at the Computer Science
Department of Bucharest “Politehnica” University. Inter-animation patterns were
discovered in the above-mentioned two dimensions: longitudinal (chronologically
sequential) and transversal (effectively simultaneous). They move in both dimen-
sions between two opposite trends: unity and difference. Moreover, we consider that
even individual thinking can be analyzed as an implicit collaborative (dialogic) pro-
cess that involves multiple voices. However, actual collaborations in small groups
of different personalities illustrate more explicitly the dialogic process.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
24 Polyphonic Inter-Animation of Voices in VMT 453
The chapter continues with a section that discusses the role of discourse in learn-
ing and that introduces the dialogic theory of Bakhtin and polyphony. The following
section is dedicated to the presentation and exemplification of the novel polyphonic
theoretical model and analysis method of CSCL chats, starting from counterpoint
and Bakhtin’s ideas. Inter-animation patterns are identified and classified along the
longitudinal-vertical and unity-difference dimensions in chats. Software tools that
support the identification and visualization of the polyphonic architecture, allow-
ing the analysis of inter-animation and even assessing individual contributions are
presented in the fourth section.
Discourse, Dialog and Polyphony
The Role of Discourse in Learning
The assessment of learning achievement in a given domain is often based on evalu-
ating the amount of knowledge acquired by the student, as in question-answering
examinations. However, in other cases as in mathematics and other disciplines
needing problem-solving abilities and/or creativity, this approach is not adequate.
Instead, successful discourse building (e.g., constructing a reasoning chain or writ-
ing an essay linking a series of ideas) is required for evaluation. Because discourse
is an artifact achieved in communication, discourse-building abilities benefit from
social, collaborative learning.
The above two approaches correspond to the contrast between socio-cognitive
and socio-cultural theories or between the Intelligent Tutoring System and CSCL
paradigms (Koschmann, 1999b; Stahl, 2006b). The socio-cultural theory of learning
is based on Vygotsky, and has had an increasing influence as the limitations of
the knowledge acquisition model become recognized. As Hicks noted, “Learning
occurs as the co-construction (or reconstruction) of social meanings from within the
parameters of emergent, socially negotiated and discursive activity” (Hicks, 1996, p.
136, quoted by Koschmann, 1999). Sfard (2000a) remarked, “Rather thanspeaking
about ‘acquisition of knowledge,’ many people prefer to view learning asbecoming
a participant in a certain discourse.”
Links and Threads
As we have seen above, discourse is a central concept in learning. There are many
definitions for discourse, the majority stating that it is characterized by structures
beyond a sentence or utterance. One definition that captures ideas present in several
others says also: “its main concepts are cohesion—the features that bind sentences
to each other grammatically and lexically—and coherence—which is the notional
and logical unity of a text” (Newmark, 1988). Therefore, for studying discourse,

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
454 S. Trausan-Matu and T. Rebedea
we must analyze links and threads (connecting sentences or utterances) providing
cohesion and coherence.
In the chat from which an excerpt is presented in Fig. 24.1, students at a Human–
Computer Interaction course had to discuss facilities and tools for a collaborative
environment. The students used the VMT chat environment, which allows the users
to explicitly link an utterance to the one it continues or replies to (see Chapter 15).
These explicit links are represented in the left part of Fig. 24.1 by curly arrows.
Fig. 24.1. Two types of links in the chat.
In addition to the explicit references, a second type of link may be identified
in any text, including chats. It is the case of implicit references among words or
phrases. The simplest case of such implicit links is between repeated words, rep-
resented in Fig. 24.1 by straight lines. In general, these implicit links may be very
complex, relating, for example, semantic arguments.
An interesting thing is that the implicit and explicit links are usually different
(e.g., in 21 of the 24 cases in Fig. 24.1). This phenomenon might be explained by
the fact that the participants probably only felt the need to include an explicit link
when an implicit one was not present or obvious. This observation introduces the
idea that repetition (e.g., of words or phrases) is a strong interaction pattern that is
perceived as such by the participants—as evidenced by the fact that they do not feel
the need to introduce explicit links when repetitions of words are present.
Implicit and explicit links form threads. In the case of implicit links between
repeated words, this fact is obvious (see Fig. 24.1). Threading occurs also for
explicit references, indicated by the users, as is seen in Fig. 24.2.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
24 Polyphonic Inter-Animation of Voices in VMT 455
Fig. 24.2. Multiple parallel threads.
All these threads—in addition to their intrinsic longitudinal nature—due to their
co-presence at the same time influence each other, inter-animating in different ways,
as we will see later. For example, Fig. 24.3 represents a part of the inter-animation
process among the three students in the development of the threads of implicit links
in Fig. 24.1 and 24.2. Time flows from left to right and the same representation of
the themes (color and types of lines) is used. In addition to the sequential dimension
of theme development, the same figure also represents (with thick arrows) three
interactions between themes, which may be considered as transversal interaction
patterns (two divergent and one convergent).
Fig. 24.3. The longitudinal-transversal dimensions.
This
figure
will be
printed
in b/w
During the chat, each of the participants introduces new variations on the theme
of the chat or iterates an already uttered theme variation. For example, in Fig. 24.3,
three theme variations are emphasized: “replying,” the “topics” in a collabora-
tive chat and ways of “presentation.” Threads may be easily discovered from the
obvious repetition pattern of these words.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
456 S. Trausan-Matu and T. Rebedea
Similarly to a musical piece, the chats for CSCL have a main theme, a topic
that is, for example, the problem to be solved or the product to be designed by the
students. This theme generates threads of discussion containing interactions that
may be identified and classified according to classes of interaction patterns. These
threads contain variations (sub-topics) of the theme, analogous to musical varia-
tions. One fundamental issue in polyphony is the presence of several participants
(or “voices”) uttering (“singing”) in a unitary way in a given moment. Among the
participants, brief dissonances may appear, but these are “solved” and a unity is
obtained.
Dialogism and Discourse
Bakhtin considered that, “Any true understanding is dialogic in nature” (Voloshinov,
1973). From his perspective, 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 viewed as implying not only the voice of
the author. For example, the potential listener also has an important role: The author
constructs a thread of ideas, a narrative. Meanwhile, parallel to it, she must take
into account the potential flaws of her discourse, the potential questions or replies;
she must see it as an utterance that can be disputed by the listener. In this idea,
discourse in a novel is similar to dialog in conversation and to polyphony in music,
where different voices inter-animate each other.
Voices
The “voice” concept in Bakhtin’s work is central and complex. In the context of a
dialog, we understand by a voicenot the acoustical, physical, vocal expression of a
given participant in a dialog but, rather, a distinct position, an utterance, an event
or a recurrent series of events of emitting utterances that are heard, remembered,
discussed and have influence on the utterances emitted by the other voices. In music,
for example, a voice is not fixed to an instrument; the same instrument may play
several voices, and different instruments may take the position of a given voice,
simultaneously or sequentially.
A voice may be seen as a distinctive position in a group, a person or a group of
people who have uttered something, with effects on the subsequent utterances. For
example, in Fig. 25-1, the voice of John from utterance number 21 is taken up by
Adrian, at 25. Moreover, a voice has some particularities; it may have a personality,
goals, beliefs, desires and emotions. Consequently, a dialog among several voices
is not a dialog among impersonal entities. From another point of view, a voice may
become a theme or may contribute to a theme of the discussion.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
24 Polyphonic Inter-Animation of Voices in VMT 457
Polyphony
Discursive voices sometimes weave a polyphonic texture—a feature that Bakhtin
admired so much in Dostoevsky’s novels. Bakhtin characterized them as “a plurality
of independent and unmerged voices and consciousnesses” (Bakhtin, 1984a).
Polyphony, a concept taken from music, may be considered as a general model
for interaction and creativity in a group of human “voices” following counterpoint
rules. As Bakhtin emphasized, it may occur also in texts and, as we will show in
this chapter, in problem-solving chats. One of the features of polyphonic music is
its development of complex architectures starting from a given theme; polyphonic
structuring of dialogs may transform them into a “thinking device.”
Polyphony is not only a randomly overlapped set of voices. It also has musicality;
it is in fact one of the most complex types of musical compositions, exemplified by
the sophisticated 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.
(Polyphony, 2005)
In polyphonic music, the melodic, linear dimension does not, in general, disturb
the transversal harmony. Even if differential dissonances may appear for a while,
they are usually quickly resolved and the unity of the musical piece is restored.
This makes a kind of game, which drives (for example, in Bach’s fugues) the inter-
animation of the participant voices. The main theme is introduced by one voice,
reformulated by others, even contradicted sometimes (e.g., inverted) but all the
voices keep a vertical harmony in their diversity, resolving the brief dissonances.
The inter-animation is generated by the different conflicting personalities or ideas
of the participants. Sometimes the conflicts derive from serious causes (e.g., dif-
ferent approaches for solving a problem), but other times, they derive from pure
ludic, playful, carnivalesque (Bakhtin, 1984b) reasons. Dissonances usually appear
but they are soon resolved, restoring the global unity.
In each dialogue, similarly to polyphonic music, there are one or more themes,
which are debated by the participant voices. Each theme is introduced by a voice
and developed by it or by the others. Several themes may be present at the same
time in the dialogue, influencing each other.
Starting from Bakhtin’s ideas, we extend the polyphonic, dialogic perspective
to collaborative learning. Therefore, we will describe how polyphony may arise in
collaborative learning and we will propose ways of analyzing and supporting it in
learning environments.
We will use in our further analyses the term “voice” instead of “participant”
because it is more general, as mentioned above. In the polyphonic framework for
analyzing chats, voice is a central concept, being the point that contrasts with the
counter-point. It is not fixed to a person, but, rather, is a position, an idea, a proposal.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
458 S. Trausan-Matu and T. Rebedea
The Polyphony of Collaborative-Learning Chats
Computer and communication technologies offer new possibilities for collabora-
tion, by allowing virtual classroom group interaction. New types of artifacts, like
hypertext, the World Wide Web, instant messenger chats or discussion forums are
changing the classical learning scenarios. In addition to traditional sheets of paper or
blackboards for drawing diagrams and writing formulas and sequences of problem-
solving steps, computer animations, simulations, chat logs or even virtual partic-
ipants in the dialog (artificial agents) may now be used for collaboration. It is
extremely important to analyze the particularities of discourse in this new context, to
identify interaction patterns, and to design supporting software tools. A good exam-
ple is the fact that in chats we can use a multiply threaded discourse much more
easily than in face-to-face conversations.
In order to develop a theoretical background and the associated supporting tools
for CSCL chats, we have started from the musical polyphony model and we have
looked for analogous structuring in collaborative-learning chats. Next, we have
searched for classes of interaction patterns that resemble musical counterpoint rules
that are used in composing polyphonic music. Eventually, we have designed and
developed tools that would facilitate the analysis from the polyphonic theory.
The analysis and the experiments were performed in two cases: mathemat-
ics problem solving and the design of human–computer interfaces. The first case
involved students using several different versions of the VMT environment. The lan-
guage they used was English. The experiments in the second case were performed
with college seniors at the Politehnica University of Bucharest (PUB). The students
were in a computer science course and they chatted in the VMT environment either
in English (as a second language) or in Romanian. All the chat groups had from 3
to 5 participants. The Polyphony system, developed at PUB was used for analyzing
the polyphonic structure of all the chats.
Collaborative Solving of Mathematics Problem
Let us consider the following problem from Chapter 5:
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 a set of eleven problems that were used for an experi-
ment. A group of students had to solve these problems, initially individually, and
subsequently collaboratively, using a chat instant messaging system. The above
problem was one of the two that were not solved individually by any student but
it was successfully solved collaboratively.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
24 Polyphonic Inter-Animation of Voices in VMT 459
Consider Log 24-1, which includes the main utterances that contributed to the
finding of the solution of the problem:
Log 24-1.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
460 S. Trausan-Matu and T. Rebedea
Discourse begins with Dan’s idea of starting from the 30,000,000 number spec-
ified in the problem statement (line 357). It continues with Mic, who seems to
start a reasoning path (lines 360–362) by writing typical fragments of mathemat-
ical problem-solving speech genre containing the typical phrase “. . . and since
. . .” After just three lines, unexpectedly, the reasoning path ends abruptly and Mic
states that his discourse is a buffoonery (lines 363–364, 366 and 368–370), taking
a “carnavalesque” (Bakhtin, 1984b) direction. This fact is explicitly remarked upon
by the utterances of Cosi (line 365) and Dan (line 367). However, even being a
pastiche, the “voice” of Mic in his fake discourse fragment has an echo in the suc-
ceeding utterances, being continued by Hal, who extrapolates the 1:1 ratio from the
present (as stated in the problem) to the whole 3 years, advancing 60,000,000 as a
solution (line 371).
Mic continues his buffoonery (lines 372–375), claiming that he helped Hal to
find the supposed solution. After a while, Cosi’s utterance “i think it’s more than
60,00000” appears as an opposing position, a critique, an intuition of something
wrong, of some kind of an “unsuccessful story” or some “dissonant” chord. Nev-
ertheless, after less than a minute, she realizes that her own supposition is wrong
because the ratio cannot be 1:1 or bigger. This idea drives her to choosing the
solution 50,000,000, the single value of the multiple choice answers less than
60,000,000.
We can say that 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. How-
ever, when she listened to the discourse proposing a solution (correct in the case
of Dan’s beginning proposal, fake by Mic and wrong by Hal), she felt the need to
take on a different position and she eventually succeeded in solving the problem.
Therefore, the discourse acted as a tool, as an artifact that enabled Cosi to find the
correct answer. Moreover, we may say that the building of the solution contains the
voices of the other participants. They inter-animate, weaving together variations of
the starting theme (the problem to be solved), as in a polyphonic musical piece.
Another, no less important feature is the “carnavalesque” character of utterances
that eventually gave rise to the solution. The role of carnavalesque utterances was
discussed in detail in Bakhtin (1984b).
Polyphonic Structuring in Chat Conversations
for Problem Solving
As we have seen in many chapters of this volume, discourse in collaborative
problem-solving chats has an obvious sequential, longitudinal, time-driven struc-
ture in which the speakers/listeners (readers/writers) are permanently situated and
in which they emit their utterances in a threaded manner, having, ideally, a unitary
character, oriented toward finding the solution. In parallel with this linear thread-
ing dimension, in problem-solving chats the participants also situate themselves in
transversal relationships that often adopt critical, differential positions. For example,
in the chat excerpt considered in the preceding section, Dan’s theme was continued

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
24 Polyphonic Inter-Animation of Voices in VMT 461
by Mic’s buffoonery, continued itself by Hal and then contradicted by a first theme
of Cosi’s that was subsequently reversed into its opposite.
In this longitudinal-transversal space, voices partake in a unity-difference—
or centripetal-centrifugal (Bakhtin, 1981)—dynamic and display various inter-
animation patterns. This phenomenon is not specific solely to chats. It also appears
in 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 dif-
ferential theory (which dialectic merely claims to be)—is the methodical point of departure
for the distinction between polyphony and non-polyphony. (Mahnkopf, 2002)
Interactions of voices towards the unity and difference dimensions were identi-
fied in all chats we have analyzed. Some of these interactions may be abstracted in
classes of inter-animation patterns in which an utterance by one voice triggers an
utterance by another voice. In the next section, patterns of inter-animation are identi-
fied along the unity-difference dimension. The subsequent section will discuss how
these interactions weave into a polyphonic structure.
Inter-Animation Patterns
When somebody listens to Bach’s fugues or other classical music works, one
remarks how several themes and their variations are exposed, developed and re-
exposed by several instruments. Moreover, these themes and their variations seem
to inter-animate each other; even the term musical “fugue” expresses the idea that
several voices are “running” and “chasing” each other. The soundscape becomes a
playful ground for creativity; a particular type of polyphonic musical piece is called
an “invention.”
Bakhtin used the musical metaphor for language, considering that “the voices
of others become woven into what we say, write and think” (Koschmann, 1999b).
Therefore, for analyzing CSCL chats, we investigate how voices are woven in dis-
course, how themes and voices inter-animate in a polyphonic way. This is important
not only for understanding how meaning is created, but also for trying to design
tools for support and evaluation.
Specific inter-animation patterns may be identified along each of the unity and
difference dimensions in a chat. In CSCL, each of these patterns may be used for
automatic abstraction of useful data, either for the participants in a chat, or for teach-
ers for evaluation purposes. Such an application, using natural language processing,
is presented in the end of this chapter.
Unity Inter-Animation Patterns
Unity-pursuing patterns are characterized by a trend towards continuity and achiev-
ing coherence in the chat. A first such class of patterns are adjacency pairs
(Sacks et al., 1974), containing couples of logically succeeding utterances like
question–answer. The first utterance in an adjacency pair normally requires (in a

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
462 S. Trausan-Matu and T. Rebedea
coherent dialog) the emitting of the second utterance. Examples of adjacency pairs
are utterances 398 and 399 in Log 24-1, or utterances 68–69, 71–72, 73–74, 76–77
in Log 24-2:
Log 24-2.
68 mathisfun see angle alpha?
69 Bob123 yes
70 Bob123 what about it?
71 mathisfun is that 60 degrees?
72 Bob123 yes
73 mathisfun can u use the degree, 2 length to find the last length of a
triangle?
74 Bob123 i don’t get what you’re saying
75 mathisfun the two arrow pointed lengths and the angle can find the
length A
76 Bob123 by what?
77 mathisfun the two sides and the degree
Question–answer adjacency pairs are important in learning because they force
the students to participate, to face questions, to answer and, implicitly, to reason and
understand the discussed problems.
Other kinds of adjacency pairs may be identified, for example, greeting–greeting
(19–20, 21–22 in Log 24-3):
Log 24-3.
19 john: hi all
20 Dan: hi john
21 mary: happy birthday, john!
22 john: Thanks mary!
In CSCL, specific adjacency pairs have been identified. For example, Stahl
(2006b, chap. 21) identified math proposal adjacency pairs, with the structure:
1. An individual makes a proposal to the group for the group’s work.
2. Another member of the group accepts or rejects the proposal.
A second kind of unity inter-animation pattern is repetition, which plays an
important role in creating coherence in a discourse. Repetition generally involves
a larger number of utterances than an adjacency pair. Tannen (1989) considers that
repetitions may be seen as a kind of rhythm making, with a main role of enhancing
the involvement of the participants in a dialogue. Of course, repetition and rhythm
are features with strong links with music, enforcing our analogy. Log 24-4 (which is
a transcript of a face-to-face conversation, taken from Stahl (2006b, p. 250)) exem-
plifies these ideas:

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
24 Polyphonic Inter-Animation of Voices in VMT 463
Log 24-4.
1:21:53 Teacher And you don’t have anything like that there?
1:21:56 Steven I don’t think so
1:21:57 Jamie Not with the same engine
1:21:58 Steven
Jamie
[No
Not with the same
1:21:59 Teacher With the same engine . . . but with a different (0.1) . . .
nose cone?=
1:22:01 Chuck
Jamie
[
=the same=
=Yeah,
1:22:02 Chuck These are both (0.8) the same thing
1:22:04
1:22:05
Teacher
Brent
Aw [right
This one’s different
Socialization or jokes are also a way of creating unity. For example, many times
participants in chats feel the need to joke, probably for establishing a closer relation
with the other participants, perhaps in order to establish a group flow state (Csik-
szentmihalyi, 1990a). In fact, in all the chats we examined there is always a prelim-
inary socialization phase.
Another interaction pattern is cumulative talk (Mercer, 2000) or, in Sacks’ words,
collaborative utterances(Sacks, 1962/1995). In such a situation, several participants
jointly utter a sentence, like a single person. Log 24-5 shows a collaborative utter-
ance co-constructed by three people completing each other’s contribution (Sacks,
1962/1995, p 144–145):
Log 24-5.
Joe (cough) We were in an automobile discussion,
Henry discussing the psychological motives for
Mel drag racing on the streets
A second example of cumulative talk is the inter-animation of Mathpudding and
Mathman in a VMT problem-solving chat (Log 24-6):
Log 24-6.
117 ModeratorSf could you guys tell templar what’s going on?
118 mathpudding we′re experimenting with circles
119 mathman and finding as many possible relations as we can
The last unity inter-animation pattern we will discuss here is convergence, which
is an utterance that links two discussion threads having different topics. For exam-
ple, in Fig. 24.1, the utterance 34 links the discussion thread on “(re)presentation”
with the one on “topic.” Convergence is an extremely important pattern, considered
by Roschelle (1996) the crux of collaboration. It is the single transversal pattern
among the previous, longitudinal ones.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
464 S. Trausan-Matu and T. Rebedea
Difference Inter-Animation Patterns
Difference patterns are inherent to chat conversations. Disputes or negotiations are
inter-animated by differences and opposing positions. Difference making has a cru-
cial role in chats for collaborative learning, a role that may be best understood from a
polyphonic, musical perspective. The possibilities of contemplating (listening, read-
ing) from a critical position the ideas (melodies) of other people and entering into
negotiation and argumentation (polyphony of voices) enhance problem solving and
enable learning through a trial-and-error process. Such processes also appear in indi-
vidual learning (we can say that thinking also includes multiple inner voices), but the
presence of multiple participants enhances both the possibility of developing mul-
tiple threads and, meanwhile, of identifying differences. The inter-animation of the
multiple perspectives of the participants, their opposition as a result of contempla-
tion, the presence of a third opinion in cases of conflict, and sometimes the synthesis
it brings are better aids to success than a multi-voiced discourse performed by an
individual (as inner thinking), where there is inherently much less conflict.
Several classes of difference inter-animation patterns may be identified. There
are simple, obvious differential utterances that dismiss an assertion (Log 24-7):
Log 24-7.
371 4:33:28 Hal it would be 60,000,000
.....
387 4:34:47 Cosi i think it’s more than 60,00000
There might be difference making that not only disapproves an assertion but also
proposes a development (Log 24-8):
Log 24-8.
392 4:35:18 Cosi they cant get higher equally and even out to a 1 to 1 ratio
Sometimes, the participants even explicitly state that they found a difference and
describe it (Log 24-9):
Log 24-9.
P4nzer agree with me so far?
Tricavl yes, but i did the same thing
Tricavl the difference was the place of the space :).
petry_g and the number of moves :)
Another example of this last type of difference making is Log 24-4 used above
for the exemplification of repetitions. It ends with an extremely important differ-
ence making, which, in fact, is the moment of finding the solution (Stahl, 2006b).
Actually, we could say that learning is achieved in many situations by understanding
significant differences.

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
24 Polyphonic Inter-Animation of Voices in VMT 465
Evidence that participants make their own (internalize, individualize) a differ-
ential position is also provided by the statistics of personal pronoun 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.”
Automatic Analysis and Graphical Representation
The polyphony-based theoretical framework presented above may be used for devel-
oping automated analysis and visualization tools for examining chats from different
points of view. As previously discussed, we consider a voice as a particular position,
which may be taken by one or more persons when they emit an utterance, which has
explicit and implicit links or influences on the other voices. In the implementation
of our analysis tool, we start from the utterances in the dialog, we identify themes
by detecting recurrent concepts and, in addition to explicit links, stated by the ref-
erencing facility of VMT, we try to find implicit links, reflecting voices’ influences.
These implicit links are detected by searching for instances of the possible interac-
tion patterns discussed above. Eventually, we try to measure the influence of each
participant in the chat, considering the “strength” of their voices (positions, uttered
utterances) on the subsequent utterances, according to the existing links. Compu-
tational linguistics techniques are used for the identification of the themes and of
implicit links among utterances.
Identification of Chat Themes
Chat themes are identified using text-mining techniques. The first step in finding the
chat subjects is to strip the text of irrelevant words (stop-words), text emoticons —
like “:)” or “:P”—special abbreviations used while chatting (e.g., “brb,” “np” and
“thx”) and other words considered irrelevant at this stage.
The next step is the tokenization of the chat text. Recurrent tokens and their syn-
onyms are considered as candidate concepts in the analysis. Synonyms are retrieved
from the WordNet lexical ontology (http://wordnet.princeton.edu). If a concept is
not found on WordNet, mistypes are searched. If successful, the synonyms of the
suggested word will be retrieved. If no suggestions are found, the word is consid-
ered as being specific to the analyzed chat and the human analyst is asked for details.
In this way, the analyst can tag the part of speech for each word and can add syn-
onyms. All this information is saved into a cache, so the analyst will not be prompted
twice for the same word.
The last stage for identifying the chat topics consists of a unification of the candi-
date concepts discovered in the chat. This is done by using the synonym list for every
concept: if a concept in the chat appears in the list of synonyms of another concept,

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
466 S. Trausan-Matu and T. Rebedea
then the two concepts’ synonym lists are joined. At this point, the frequency of the
resulting concept is the added frequencies of the two unified concepts. This process
continues until there are no more concepts to be unified. At this point, the list of
resulting concepts is taken as the list of topics for the chat conversation, ordered by
their frequency.
In addition to the above method, used for determining the chat topics, there is an
alternate technique we used to infer them by using a surface analysis technique of
the conversation. Observing that new topics are generally introduced into a conver-
sation using some standard expressions such as “let’s talk about email” or “what
about wikis,” we can construct a simple and efficient method for deducing the
topics in a conversation by searching for the moment when they are first mentioned.
A list of patterns of ways of introducing topics in a conversation can be manually
edited. If an utterance matches any one of the patterns, it means that the utterance
introduces a new topic. A pattern consists of a number of words that must be iden-
tified in the utterance and a key word that is associated to the new topic of the
conversation (e.g., “let’s talk about <topic>” or “what about <topic>”). The pro-
cess of identifying a pattern in an utterance is done using the synset for each word
that has already been extracted from WordNet.
The implemented system has an interface (see Fig. 24.4) that lists the topics
sorted according to their number of occurrences in the chat. This interface also dis-
plays the utterances of the chat associated with the topics they include and with
information about the detected interaction patterns (e.g., adjacency pairs). It also
contains some parameters that can be tuned for obtaining the best analysis.
Fig. 24.4. Topic detection screen.
This
figure
will be
printed
in b/w

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
24 Polyphonic Inter-Animation of Voices in VMT 467
The topics of the chat may also be detected as the connected components in
the chat graph described in the next section. All the details of an utterance in the
chat—the content of the utterance, the implicit and explicit references and other
details—can be visualized by clicking the rectangle representing the utterance.
Discovering Implicit Links in an Utterance
As we have previously discussed, in a log of a VMT conversation two types of links
among utterances may be identified. There are explicit links, stated by participants
by means of the VMT referencing tool. In addition to these, many implicit links may
be identified, as was exemplified in Fig. 24.1.
We consider that each chat utterance may have a certain influence in the devel-
opment of the conversation; it can become a chat voice. Each utterance may contain
the influence of at least one other, alien voice, for example that to whom it refers, as
an answer to a question, an elaboration, a disagreement, etc. By transitivity, voices
may accumulate during a conversation. The emitter of the utterance implicitly can
note the presence of alien voices in his utterance, when he explicitly refers to a
previous utterance with the VMT referencing tool.
Because users are generally in a hurry or they don’t consider it necessary, many
of the utterances do not have any explicit references. Thus, it is necessary to find
a method for discovering the implicit references in an utterance. The method pro-
posed here is similar to the one presented above for determining the introduction
of new chat topics, based on text mining techniques (Manning & Schutze, 1999)
and patterns. The system uses another list of patterns that consists of a set of words
(expressions) and a local subject called the referred word. If an utterance matches
one of the patterns, it is first determined what word in the utterance is the referred
word (e.g., “I don’t agree with your assessment”). Then, a search for this word
is performed, in a predetermined number of the most recent previous utterances.
If such a word is found in one of these utterances, then an implicit relationship is
defined between the two lines, the current utterance referring to the identified utter-
ance. In addition, two other empirical methods were implemented.
A graphical representation of chats was designed to facilitate an analysis based
on the polyphony theory of Bakhtin and to permit the best visualization of the con-
versation. For each participant in the chat, there is a separate horizontal line in the
representation and each utterance is placed in the line corresponding to the issuer
of that utterance, taking into account its positioning in the original chat file—using
the timeline as an horizontal axis (see Fig. 24.5). Each utterance is represented as
a rectangular node having a horizontal length proportional with the textual length
of the utterance. The distance between two different utterances is proportional to
the time between the utterances (Trausan-Matu, Rebedea, Dragan & Alexandru,
2007).
The explicit references between utterances are depicted using blue connecting
lines while the implicit references (deduced using the method described in Trausan-

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
468 S. Trausan-Matu and T. Rebedea
Fig. 24.5. Graphical visualization of the discussion threads.
This
figure
will be
printed
in b/w
Matu et al., 2007) are represented using red lines. The utterances that introduce a
new topic in the conversation are represented with a red margin.
The graphical representation of the chat has a scaling factor that permits an
overview of the chat, as in Fig. 24.6, as well as an attentive observation of the details
in a conversation (as in Fig. 24.5).
(a)
(b)
Fig. 24.6. A conversation with (a) equal and (b) non-equal participation.
Viewing the whole conversation graph gives an idea of the global participation
of the learners. For example, in Fig. 24.6a, all the participants make about an equal

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
24 Polyphonic Inter-Animation of Voices in VMT 469
number of contributions. This is not the case in Fig. 24.6b, where one participant
has almost no participation and another student leaves early in the chat session.
At the bottom of the graphical representation of the conversation (see Fig. 24.7),
after the line corresponding to the last participant in the chat, there is a special area
that represents the importance (strength) of each utterance, considered as a chat
voice, in the conversation (Trausan-Matu et al., 2007). The height of the rectangle
corresponding to each utterance is proportional with the strength of that utterance
(or voice). The details about how this measure is computed are presented in the next
section.
Fig. 24.7. Utterances 122–136 are linked with many relations.
This
figure
will be
printed
in b/w
Assessing the Contributions of the Learners in the Conversation
One of the most important goals in any collaborative-learning process is the assess-
ment of the contribution of each learner. For CSCL using chat conversations, in
order to determine the contributions of the participants a graphical representation
of the contribution was implemented starting from the polyphonic theory and the
analysis method. The evaluation of the contributions of each learner considers the
degree to which they have influenced the conversation. In terms of our polyphonic
model, we evaluate to what degree they have emitted strong utterances that influ-
enced the following discussion, or, in other words, to what degree the utterance
became a strong voice.
An utterance is considered strong if it influences the continuation of the conversa-
tion. The contribution of each participant is computed by accumulating the strengths
of the utterances they emitted.
The automatic analysis considers the inter-animation patterns in the chat. It uses
several criteria such as the presence in the chat of questions, agreement, disagree-
ment or explicit and implicit referencing. The diagram is generated using a series
of parameters like: implicit and explicit reference factors, bonuses for agreement,
penalties for disagreement, minimum value for a chat utterance, penalty factors
for utterances that agree or disagree with other utterances if these utterances have
less originality than the first ones. In addition, the strength of a voice (of an utter-
ance) depends on the strength of the utterances that refer to it. If an utterance is

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
470 S. Trausan-Matu and T. Rebedea
referenced by other utterances that are considered important, obviously that utter-
ance also becomes important.
By using this method of computing their importance, the utterances that have
started an important conversation within the chat, as well as those that began new
topics or marked the passage between topics, are more easily emphasized. If the
explicit relationships were always used and the implicit ones could be correctly
determined in as high a number as possible, then this method of calculating the
contribution of a participant would be considered successful (Trausan-Matu et al.,
2007).
During the first step of the graph generation, the importance value of each utter-
ance is computed by relating it to an abstract utterance that is built from the most
important concepts in the conversation (the themes). When constructing this utter-
ance, we take into account only the concepts whose frequency of appearance is
above a given threshold. Then all the utterances in the chat are scaled in the interval
0–100, by comparing each utterance with the abstract utterance. The comparison
is done using the synonym sets of each word contained in the utterance. Thus, this
process uses only the horizontal relations from WordNet. An utterance with a score
of 0 contains no words from the concepts in the abstract utterance and an utterance
with a score of 100 contains all the concepts from the abstract utterance.
Log 24-10 contains a sequence of utterances where the participants collaborate
intensively (it may be considered as a “collaborative moment” (Stahl, 2006b)), a fact
revealed from the relations graph (Fig. 24.6) and from the large number of explicit
and implicit relations interconnecting utterances 122 through 136.
From Fig. 24.7, we can see that the highest strength (the highest rectangle below
the utterances) has the voice of RaduDumitrescu at the utterance nr. 122 (an oval
shadow was manually added for emphasizing it). This fact is also observable by the
large number of relations following utterance 122 (see Log 24-10) and in the change
of the amount of contribution of RaduDumitrescu, in Fig. 24.8.
The graph that shows the contributions of every participant (in Fig. 24.8) contains
on the x-axis the utterances in the chat and on the y-axis the value computed for
each participant in the conversation, for his/her cumulative contribution. This value
is computed by summing the numerical values corresponding to the strengths of the
utterances that the participant has uttered up to the position on the x-axis. Accord-
ingly, for each utterance, at least the value of one user contribution is modified—the
value for the user that issued that utterance.
Conclusion
In all of the chats from the CSCL experiments we have analyzed, the interactions
are structured in a polyphonic manner. Discourse in chats implies an inter-animation
of multiple voices along two dimensions, the sequential utterance threading and the
transversal one, similar to polyphonic music. In addition, another dichotomy, the
unity-difference (or centrifugal-centripetal, (Bakhtin, 1981)) opposition may also be

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
24 Polyphonic Inter-Animation of Voices in VMT 471
Lo
g
24
-1
0.
12
2
R
ad
uD
um
itr
es
cu
al
so
th
e
ap
pl
ic
at
io
n
al
lo
w
s
th
e
us
er
to
de
sc
ri
be
th
e
to
pi
c
of
th
e
m
ee
tin
g
12
3
A
le
xr
os
iu
ye
s,
an
d
fu
rt
he
rm
or
e,
se
ve
ra
lt
op
ic
s
sh
ou
ld
be
de
fin
ed
R
ef
er
en
ce
to
m
es
sa
ge
N
o.
12
2
12
4
A
le
xe
i
ye
s,
th
at
w
ou
ld
al
so
he
lp
an
au
to
m
at
ic
ap
pl
ic
at
io
n
to
pa
rs
e
th
e
ch
R
ef
er
en
ce
to
pa
rt
of
th
e
m
es
sa
ge
N
o:
12
2
12
5
R
ad
uD
um
itr
es
cu
so
ev
er
yb
od
y
m
us
tk
no
w
w
ha
ta
re
th
e
m
ee
tin
g
is
al
la
bo
ut
12
6
A
le
xr
os
iu
m
ay
be
ev
en
so
m
e
us
er
s
co
ul
d
be
w
an
ed
if
th
ey
ar
e
of
ft
op
ic
...
bu
tt
hi
s
is
a
ra
th
er
sc
i-fi
fe
at
ur
e,
i
gu
es
s
:)
R
ef
er
en
ce
to
m
es
sa
ge
N
o.
12
4
12
7
R
ad
uD
um
itr
es
cu
an
d
at
th
e
en
d
th
e
ap
pl
ic
at
io
n
sh
ou
ld
sp
ec
ify
if
al
lt
he
to
pi
cs
w
er
e
co
ve
re
d.
...
w
ha
td
o
yo
u
th
in
k?
R
ef
er
en
ce
to
m
es
sa
ge
N
o.
12
3
12
8
A
le
xe
i
ye
s,
ia
gr
ee
,b
ut
It
hi
nk
it
ca
n
be
do
ne
if
th
e
us
er
is
go
in
g
to
o
“o
ff
to
pi
c”
R
ef
er
en
ce
to
pa
rt
of
th
e
m
es
sa
ge
N
o:
12
6
12
9
A
le
xe
i
ye
s,
m
ay
be
so
m
e
pe
rc
en
ta
ge
of
co
ve
ra
ge
...
R
ef
er
en
ce
to
pa
rt
of
th
e
m
es
sa
ge
N
o:
12
7
13
0
A
le
xr
os
iu
C
or
re
ct
R
ef
er
en
ce
to
m
es
sa
ge
N
o.
12
7
13
1
D
or
in
th
is
fe
at
ur
e
im
pl
ie
s
a
ra
th
er
ad
va
nc
ed
na
tu
ra
ll
an
gu
ag
e
pr
oc
es
si
ng
en
gi
ne
,t
ho
ug
h
R
ef
er
en
ce
to
m
es
sa
ge
N
o.
12
8
13
2
A
le
xe
i
so
,a
bo
ut
th
e
re
m
in
de
rs
-w
he
n
a
us
er
le
av
es
th
e
co
nf
er
en
ce
fo
r
so
m
e
re
as
on
,h
e
sh
ou
ld
be
re
m
in
de
d
ab
ou
tt
he
m
is
se
d
pa
rt
s
of
th
e
co
nv
er
si
on
R
ef
er
en
ce
to
pa
rt
of
th
e
m
es
sa
ge
N
o:
12
1
13
3
A
le
xr
os
iu
m
ay
be
so
m
e
ki
nd
of
re
m
in
de
rs
sh
ou
ld
be
se
tf
or
fu
tu
re
co
nf
er
en
ce
s.
..
m
ea
ni
ng
th
at
al
lp
eo
pl
e
in
vi
te
d
to
th
e
co
nf
er
en
ce
sh
ou
ld
be
re
m
in
de
d
to
at
te
nd
13
4
A
le
xe
i
a
pr
ob
le
m
th
at
i′ v
e
al
so
no
tic
ed
he
re
is
th
e
ra
th
er
un
sy
nc
hr
on
iz
ed
w
ay
of
ta
lk
in
g
13
5
A
le
xr
os
iu
w
el
l,
th
is
w
ou
ld
be
so
lv
ed
by
us
in
g
th
e
tr
ee
vi
ew
iw
as
ta
lk
in
g
ab
ou
te
ar
lie
r
R
ef
er
en
ce
to
m
es
sa
ge
N
o.
13
4
13
6
R
ad
uD
um
itr
es
cu
it
hi
nk
th
e
us
er
s
ca
n
ch
ec
k
th
e
to
pi
cs
,n
o
ne
ed
fo
r
na
tu
ra
ll
an
gu
ag
e
pr
oc
es
si
ng
R
ef
er
en
ce
to
m
es
sa
ge
N
o.
13
1

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
472 S. Trausan-Matu and T. Rebedea
Fig. 24.8. The evolution of the contribution of the participants in the chat.
observed. Adjacency pairs, repetitions, collaborative utterances, socialization and
convergent inter-animation patterns contribute to the unity-directed dimension at
diverse discourse levels.
The second, differential dimension could be better understood if we consider dis-
course 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 unity-differential perspective for the design of CSCL envi-
ronments is that they must facilitate inter-animation not only on the unity dimension
through threading, but also along the transversal, differential, critical dimension.
Tools that can assist in this category should be able to provide abstractions of the
discourse and recommendations, 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 advocates the use of a dialogic framework for teaching thinking skills
by stressing inter-animation: “meaning-making requires the inter-animation of more
than one perspective” (Wegerif, 2006). He proposes that “questions like ‘What do
you think?’ and ‘Why do you think that?’ in the right place can have a profound
effect on learning” (Wegerif, 2007). However, he does not develop the polyphonic
feature of inter-animation.
Starting from the theory of dialog, an application was implemented that may
be used for inspecting what is going on and for measuring the degree to which

UN
CO
RR
EC
TE
D
PR
OO
F
SPB-150619 Chapter ID 24 June 17, 2009 Time: 07:15pm Proof 1
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
24 Polyphonic Inter-Animation of Voices in VMT 473
learners are involved in a forum discussion or a chat conversation. The effective
contribution of each participant to the inter-animation process may be measured.
The application visualizes the strengths of the voices of the participants in chat
conversations, following Bakhtin’s ideas. Diagrammatic representations are used
for viewing the influence of a given speaker and of the comparative evolution of
the contribution of the learners. The visualization application described here can be
further extended to consider more aspects related to the polyphonic, contrapuntal
features of chat conversations.