Академический Документы
Профессиональный Документы
Культура Документы
Volume 1: Embodiment
2
8 1...- ------'
16 14 12 10
F
2
(Bark)
8
.
..
e :. /01
1/ .,... u......
6
lel
/
',1 " ..
.::. : lu!
2
lal :-.:
8 '---------.;.,....--"--..........,
16 14 12 10
F
2
(Bark)
2
6
\ -a/ul
lal ift..f 4
, UI
IfJ ,.:.- I'JI e?.
. .p u...... 6
lal .-.,. ..\
8 '--- """--__----..J
16 14 12 10 8
F
2
(Bark)
Type D (7/54) Type E (3/54) Type F (1/54)
I'JI
/01 4
lul
.Ial
/tl
lce/
/if
2
Jul
.
.
I'JI
hI
lel .t.
li! .,
2
,. lul
lif 2
lal a
8 1...- ...0-- ------'
16 14 12 10 8
F2 (Bark)
la/
8 L--
16 14 12 10 8
F
2
(Bark)
101
8 I...- """'-----"-_----.J
16 14 12 10 8
F2 (Bark)
Figure 4. Emerged vowel systems with six vowels.
4.2.2. Evaluation ofthe emerged vowel system
Are the vowel systems that emerge realistic? If in the emerged systems the
same types of vowel systems are found in the same proportions as in hu-
man languages, they are realistic. The classifications of human vowel sys-
tems that were used as a reference were the ones by Crothers (1978) and by
Schwartz et al. (1997a).
An example of a classification of emerged vowel systems containing six
vowels is shown in figure 4. The data in this figure were obtained by run-
Embodiment and self-organization ofhuman categories 425
ning the simulation a hundred times for a given parameter setting (acoustic
noise set to 12%). In each run 25 000 imitation games were played.
Note that although the frames in figure 4. look similar to the figures of
vowel systems shown previously in figure 3, there is a crucial difference.
In previous figures, the agents shown in one frame were all members of the
same population (and had therefore interacted with each other). In the pre-
sent figure, all agents shown are members of different populations. The
fact that there still is a large amount of similarity between the vowel sys-
tems of agents from different populations is a strong demonstration of how
self-organization can make populations converge towards similar vowel
systems.
The emerging systems are realistic. Most of them conform to the uni-
versals that Crothers (1978) found for human vowel systems. When the
percentages with which the different emerged systems occur are compared
to the percentages with which human vowel systems occur, a good match is
found as well. Schwartz et al. (1997a) have measured the occurrence of
different vowel system types in the different languages in the UCLA
Phonological Segment Inventory Database (UPSID, a database based on
speech sound data of 451 languages, Maddieson 1984; Maddieson and
Precoda 1990). They find 60 vowel systems with 6 vowels. Although their
classification is not exactly the same as the classification shown in figure
4, there is good agreement. Of the systems they found in UPSID, 43% is of
type A, 20% is of type B, 5% is of type C, 7% is of type D and 20% is of
type E. No systems of type F were found, and two of the systems from
UPSID (3%) cannot easily be fitted into the classification used here, but
are probably of type A. Equally good agreement was found for systems of
4 and 5 vowels and reasonable agreement was found for systems of 7 and 8
vowels (de Boer 1999, ch. 6). It seems that the simulation is capable of not
only predicting the most frequently found vowel systems in human lan-
guage (as was already possible with systems that optimize acoustic dis-
tinctiveness), but also of predicting the less frequently occurring vowel
systems and, approximately, their relative abundance.
5. Conclusions
The results of the simulation show that vowel systems can emerge through
self-organization in a population of agents whose production and percep-
tion is constrained by their embodiment. Although the agents start with an
426 Luc Steels and Bart de Boer
empty vowel repertoire and do not have any constraints on the kinds of
vowel systems they can learn, the vowel systems that emerge in the popu-
lation tend to be symmetric and dispersed, and optimized for acoustic dis-
tinctiveness. The model also predicts the types of vowel systems that occur
less frequently and their abundance with a reasonable degree of accuracy.
This illustrates that dynamics and embodiment play an important role in
determining the structure of phonological and phonetic systems. The im-
portance of embodiment both for constraining the space of possibilities and
for pushing towards optimality in the task was known already (see e.g.
Lindblom, MacNeilage and Studdert-Kennedy 1984) and the computer
model presented here provides additional empirical support for this thesis.
At the same time, the model clearly shows that embodiment can (and has to
be) complemented with the collective dynamics flowing from social inter-
action. As the imitation success of agents is partly determined by how well
the agents' vowel repertoires conform to the repertoires that are used by
the other agents in the population, different (sub-optimal) configurations
can emerge and be maintained. Some configurations can be considered
stronger attractors than others in the dynamical system that is defined by
the agents and their interactions. The same universal tendencies that are
found in human vowel systems are found in the systems that emerge. This
indicates that innate rules and constraints are not necessary in order to
explain why these tendencies are present; neither is it necessary to assume
that they have to be genetically encoded.
More generally (and based on other works such as reported in Steels
and Belpaeme 2005) we argue that the universal tendencies found in hu-
man categorization, as well as the cognitive systems that use these catego-
rizations for action and decision-making, are a function of multiple forces,
and in particular embodiment and collective dynamics. In domains where
the statistical structure inherent in the real world is relevant (such as color
categories), sensitivity to this structure will also play a role. In domains
where there is genuine choice (like driving to the right or left) the regular-
ity is generated by the agents and cannot be explained on the basis of pre-
existing statistical structure. All this does not exclude the possibility that
certain categorizations could not have become genetically assimilated.
However, in the case of cultural systems, like language, which undergo
relatively rapid change, this seems less likely than is often assumed.
2001
2000
Davidoff, J.
2001
Embodiment and self-organization ofhuman categories 427
Acknowledgement
The Sony Computer Science Laboratory in Paris funds the work by Luc
Steels, with additional funding from the OHLL CNRS project, the OMLL
ESF project, and the EU FET ECAgents project. The work ofBart de Boer
reported here took place at the University of Brussels AI lab and was
funded by a GOA project.
References
Berlin, B. and P. Kay
1969 Basic color terms: Their universality and evolution. University of
California Press.
Bowerman, Melissa and Soonja Choi
2001 Shaping meanings for language: Universal and language specific in
the acquisition of spatial semantic categories. In: Melissa Bowerman
and Stephen C. Levinson (eds.), Language Acquisition and Concep-
tual Development, 475-511. Cambridge: Cambridge University
Press.
Browman, Catherine P. and Louis Goldstein
1992 Articulatory phonology: An overview. Phonetica 49: 155-180.
Chomsky, Noam and Morris Halle
1968 The Sound Patterns ofEnglish. Cambridge, Mass: MIT Press.
Christophe, Anne, Jacques Mehler and Nuria Sebastian-Galles
2001 Perception of prosodic boundary correlates by newborn infants. In-
fancy 2: 285-394.
Crothers, John
1978 Typology and universals of vowel systems. In: Joseph H. Greenberg,
Charles A. Ferguson and E. A. Moravcsik (eds.), Universals of Hu-
man Language, Volume 2 Phonology, 93-152. Stanford: Stanford
University Press.
Language and perceptual categorisation. Trends in Cognitive Sci-
ences 5 (9): 382-87.
de Boer, Bart
1999 Self Organisation in Vowel Systems, PhD Thesis, AI-Lab, Vrije
Universiteit Brussel.
Self organization in vowel systems. Journal of Phonetics 28 (4):
441-465.
The Origins of Vowel Systems. Oxford: Oxford University Press.
428 Luc Steels and Bart de Boer
Dresher, Elan
1992 A learning model for a parametric theory in phonology. In: Robert
Levine (ed.), Formal Grammar: Theory and Implementation, 290-
317. Oxford: Oxford University Press.
Edelman, Gerald
1990 Neural Darwinism. New York: Basic Books.
Fast, Peter W.
1953 Amuesha (Arawak) phonemes. International Journal of American
Linguistics 19: 191-194.
Herskovits, Annette
1986 Language and Spatial Cognition. Cambridge: Cambridge University
Press.
Iida, Fumiya, Raja Dravid and Chandana Paul
2002 Design and control of a pendulum driven hopping robot. Proceedings
ofInternational Conference on Intelligent Robots and Systems 2002
(IROS 02) Lausanne, Switzerland, 2141-2146.
Jakobson, Roman and Morris Halle
1956 Fundamentals ofLanguage. The Hague: Mouton and Co.
Kay, P. and T. Regier
2003 Resolving the question of color naming universals. Proceedings of
the National Academy of Sciences 100 (15): 9085-9089.
Ladefoged, Peter, and Ian Maddieson
1996 The Sounds ofthe World's Languages. Oxford: Blackwell.
Lakoff, George and Mark Johnson
1980 Metaphors We Live by. Chicago: University of Chicago Press.
Lindblom, Bjom, Peter MacNeilage and Michael Studdert-Kennedy
1984 Self-organizing processes and the explanation of phonological uni-
versals. In: Brian Butterworth, Bernhard Comrie and Osten Dahl
(eds.), Explanations for Language Universals, 181-203. Berlin:
WaIter de Gruyter.
Lindblom, Bjom and Ian Maddieson
1988 Phonetic universals in consonant systems. In: Larry Hyman and
Charles N. Li (eds.), Language, Speech and Mind, 62-79. London:
Routledge.
Lenneberg, Eric H. and John M. Roberts
1956 The language of experience: A study in methodology. International
Journal ofAmerican Linguistics memoir 13.
MacNeilage, Peter
1998 The frame/content theory of evolution of speech production. Behav-
ioral and Brain Sciences 21: 499-548.
Maddieson, Ian
1984 Patterns ofSounds. Cambridge: Cambridge University Press.
Embodiment and self-organization ofhuman categories 429
Maddieson, Ian and Kristin Precoda
1990 Updating UPSID. UCLA Working Papers in Phonetics 74: 104-111.
NUflez, Rafael
1999 Could the future taste purple? Reclaiming mind, body and cognition.
Journal ofConsciousness Studies 6 (11-12): 41-60
Pfeifer, Rolf and Fumiya Iida
2004 Embodied artificial intelligence: Trends and challenges, In: Fumiya
Iida, RolfPfeifer, Luc Steels and Yasuo Kuniyoshi (eds.), Embodied
Artificial Intelligence, Lecture Notes in Computer Science 3139: 1-
26. Berlin: Springer Verlag.
Plaut, D. C. and C.T. Kello
1999 The emergence of phonology from the interplay of speech compre-
hension and production: A distributed connectionist approach. In:
Brian MacWhinney (ed.), The emergence of language: 381-415.
Mahwah, NJ: Erlbaum.
Rand, Earl
1968 The structural phonology of Alabaman, a Muskogean language. In-
ternational Journal ofAmerican Linguistics 34: 94-103.
Redford, Melissa A., Chun C. Chen and Risto Miikkulainen
2001 Constrained emergence of universals and variation in syllable sys-
tems. Language and Speech 44: 27-56.
Schwartz, Jean-Luc, Louis-Jean Boe, Nathalie Vallee and Christian Abry
1997 a Major trends in vowel system inventories. Journal of Phonetics 25:
233-253.
1997 b The dispersion-focalization theory of vowel systems. Journal of
Phonetics 25: 255-286.
Steels, Luc
2003 Evolving grounded communication for robots. Trends in Cognitive
Science. 7 (7): 308-312.
Steels, Luc and Tony Belpaeme
2005 Coordinating perceptually grounded categories through language.
Behavioral and Brain Science. 28 (4): 469-529.
Steels, Luc and Rodney Brooks
1995 The Artificial Life Route to Artificial Intelligence. Building Embod-
ied Situated Agents. New Haven: Lawrence Erlbaum.
Steels, Luc and Oudeyer, Pierre-Yves
2000 The cultural evolution of syntactic constraints in phonology. In: M.
A. Bedau, J. S. McCaskill, N. H. Packard and S. Rasmussen (eds.),
Proceedings of the VIf
h
artificial life conference (alife 7): 382-394.
Cambridge (MA): MIT Press.
430 Luc Steels and Bart de Boer
Vallee, Nathalie
1994 Systemes vocaliques: de la typologie aux predictions. These. ICP,
Grenoble.
Ziemke, Tom
2003 What's that thing called embodiment? In: R. Alterman and D. Hirsch
(eds.), Proceedings o/the 25th Annual Meeting o/the Cognitive Sci-
ence Society: 1305-1310. Hillsdale, NI: Lawrence Erlbaum.
Communication as situated, embodied practice
Wolff-Michael Roth
Abstract
Researchers generally are preoccupied with written language as the paradigm of
communication. This paradigm makes communication and what is communicated
appear as something that exists apart from invariably embodied and situated action.
A large part of human communication involves participants in face-to-face conver-
sation. In such situations, participants attend to semiotic resources other than words
produced and attended to, including gestures, body positions, and structures in the
environment. To avoid the dualism that characterizes much of the scholarship on
communication, I provide a dialectical account of human activity, one component
of which is communication, itself characterized by the dialectical relation of differ-
ent modes (speech, gesture, structures in environment). The analyses of an episode
from a school science classroom exemplify the central role of the body in commu-
nication, which by far exceeds what participants articulate in so many words. Its
relationship to environmental features, however, is dialectical, making communica-
tion a (materially, socially) situated and embodied practice.
Keywords: agency I structure dialectic, communication, embodiment, gestures,
situated practice.
1. Introduction
Since the late 1980s, it has become fashionable to talk about knowing and
learning in terms of embodied cognition Of, alternatively, in terms of dis-
tributed or situated cognition. All of these terms imply that knowing ex-
ceeds that what can be found in the brain - viewed as computational wet-
ware - and that has been the preoccupation of psychologists.! Like many
others concerned with cognition and development, I had been involved in
1. Etymologically, to know and cognize have the same origins, generally held to
have the same root as can.
432 Wolff-Michael Roth
building models of human cognition based on Piaget's work, which came
to be associated with information processing. However, once I conducted
research in the attempt to understand knowing and learning in real time and
particularly the evolution of language, the then valid measures of reasoning
ability and short-term memory were unable to predict performances in the
real-world settings that I attempted to explain. Theoretical approaches that
included or were based on social processes seemed to be much more
promising, as the turn to social constructivism in disciplines such as social
studies of science (e.g., Latour and Woolgar 1979), anthropology, and edu-
cation (e.g., Lave and Wenger 1991) seemed to imply.
In the course of the 1990s, I had taken hundreds of hours of videotape
first in classrooms and then among scientists and technicians at work. It
became increasingly clear to me that analyzing only the words they used
was insufficient to understand what participants knew, what they commu-
nicated to one another, and what they knew to be going on. It was impossi-
ble to model communication unless one also accounted for the bodily ori-
gin of communication and nature of knowing, on the one hand, and for
material cultural processes, on the other. There was a problem, however,
which I understood only afterward: the individual (body) and the sociality
of culture cannot be thought together in dualistic models, because each
approach ends up reducing one to the other. It was at then that dialectical
approaches to communication became salient in my work.
In this chapter, I exemplify the kind of communicative action that I
found among students and professionals of science. Drawing on an episode
from a science classroom, I exemplify the (socially, materially) situated
and embodied nature of communicative action. I begin by articulating a
dialectical framework for activity and action that allows me to treat com-
municative action and situated cognition in a non-dualist way. I move on to
provide a brief description of the ethnographic and analytic background,
before providing an extended analysis of one typical episode.
2. Integrating body, mind and culture
Body, mind, and culture can easily be integrated into the same unit within a
dialectical approach (e.g., Leont'ev 1978). In the following, I first sketch a
general theory resting on the dialectic of agency and structure, and then
articulate communicative action as one aspect within this framework.
Communication as situated, embodiedpractice 433
2.1. Agencylstructure dialectic
The embodied, situated, and cultural nature of cognition in general and
communication of a conscious agent more specifically can be understood
within an agencylstructure dialectic (Sewell 1992). Agency denotes the
capacity to act. It is immediately clear that there is no agency without
structure, material or cognitive. For living organisms, structure is every-
where. Bodies are structured in a material sense, exhibiting particular ar-
ticulations, symmetries, and sizes, all of which mediate the kind of move-
ments organisms are capable of. The world in which we live is structured,
too, with its characteristic forces and articulations among things, but
structured differently for different organisms (von Uexkiill 1972/1928).
Scholars attend less to the structured ways in which we see, hear, feel,
move, and do things, including thinking about them. These structured ac-
tions, including perception, have arisen from our bodily engagement with
the world (Merleau-Ponty 1945; see also Gallagher this volume; Lindblom
and Ziemke this volume). The very cognitive structures that allow us to
perceive the world in structured ways are themselves a consequence of
interactions of a structured body with a structured world (Bourdieu 1980).
But there would not be the need for speaking of structures in the world
unless there were organisms to which these structures are behaviorally
relevant. To distinguish the two forms of structure, the notions of objec-
tively given (sociomaterial) resources and (mental) schema have been pro-
posed (Sewell 1992). At any moment in activity, the given structures, i.e.,
sociomaterial resources and schema, enable and constrain actions.
In any consideration of human activity, the two forms of structures have
to be considered dialectically. Any object of activity simultaneously exists
as a vision or hypothesis of the forthcoming product (which implies con-
sciousness), a set of material conditions including materials and tools, and
knowledge of sociomaterial conditions of the productive activity (Saari and
Miettinen 2001). Therefore, neither sociomaterial resources nor schema
individually or together determine action; rather, actions emerge in re-
sponse to the contradictions embodied in the unit of analysis, which com-
prises the identity of (the non-identical) sociomaterial resources and their
appearance to conscious mind in action.
Any analysis of activity simultaneously has to consider three levels,
which integrate the bodily and cultural nature of engagement with the
world: activity, action, and operation (Leont'ev 1978). Activities, such as
schooling, farming, or researching are always collective endeavors but
434 Wolff-Michael Roth
concretely intended and realized in varying form, and are directed toward
some generalized (conscious) object or motive. Concretely realization oc-
curs by means of sequencing of particular conscious actions, each directed
toward a specific conscious goal formulated by the acting subject. While
researching, a professor might decide to measure a certain variable or to do
an ANOVA. The relation of activity and action is dialectical: the activity
determines the sense of an action - an ANOVA done by a pollster is likely
to have a different sense than the professor's - but actions concretely real-
ize the activity. Properly sequenced smaller units, operations, constitute
actions; operations are unconsciously produced responses to conditions,
such as when the professor keys the statistical model into the computer but
is not consciously aware of hitting particular keys but focuses on the equa-
tions appearing on the screen. The relation of action and operation is,
again, dialectical: the action is a generalized referent for the proper se-
quencing of operations, but operations concretely realize actions. Opera-
tions have been shaped in and are the outcome of previous interactions of
the body with the sociomaterial world, they constitute a concrete realiza-
tion of cultural possibilities; they are a form of unconscious collective con-
sciousness, that is, an embodiment of culture. This approach corresponds to
that taken by Gallagher (this volume), where actions are goal-directed and
conscious but realized by unconscious operations brought about by body
schema. Communication is but one form of actions that develop the current
activity, without necessarily being about the activity.2 This approach there-
fore contrasts the claims of those scholars (e.g., Zlatev, this volume) who
fail to recognize that (a) language-in-use (parole) is most frequently pro-
duced without prior planning (i.e., conscious supervision) and frequently
does not follow the conventions grammarians have adopted and (b) com-
municative action and thought are frequently produced in real time rather
than reproduced through speech.
2. Space limitations do not allow me to elaborate the relation between resources
and schema on the one hand, and activity, action, and operation on the other
hand. Suffice it to say that resources exist externally to the agent, at the inter-
face of activity and action; and schemas exist internally to the agent, at the inter-
face of action and operation.
Communication as situated, embodiedpractice 435
2.2. Communicative action
Communicative action is more than sending information (words) from a
sender (speaker) to a receiver (listener), which would require equal tuning
of the two participants to code and encode concrete material sound parcels.
Because of the dialectical relation material resources and schema, the
identical tuning of speaker and listener is never given in human communi-
cation. A dialectical approach to human activity leads us to a different
conception of communication. Each action produces a change in the cur-
rent setting, thereby providing changes to the semiotic resources available
for the conduct of activity in general and face-to-face communication in
particular. Speech act theory allowed us to understand that utterances, too,
have effects in the setting, they constitute resources that can be part of any
future action, current or subsequent speaker. Artifacts (Streeck 1996), dia-
grams (Goodwin 2000), and other structured aspects in the setting, includ-
ing various forms of gestures (e.g., Bavelas, Chovil, Coates and Roe 1995),
pointing and body orientation (Hindmarsh and Heath 2000), and pitch
(Goodwin, Goodwin and Yeager-Dror 2002), constitute resources produced
and used by speakers and listeners alike. To understand communication in
everyday settings, all of these structures need to be accounted for simulta-
neously (Roth 2004).
These different structures that are apparent in communicative action
(words, gestures, things pointed to) are not simply additive but dialectically
related. Thus, for example, speech and iconic gestures constitute two ways
in which thought is not merely made public but actually finds its expres-
sion as productive activity (McNeill and Duncan 2000). Speech and ges-
tures are dialectically related because they constitute different expressive
forms for the same intentional stance toward the world. This stance is not
some idea somehow behind or underlying and therefore expressing (trans-
lating) speech. Rather, it is my bodily me that gestures and speaks, and
thereby signifies thought and intention for others and myself (Merleau-
Ponty 1945).
Making sounds, speaking individual words, gesturing, nodding the head,
turning the body and so forth are operations that together constitute a
communicative act (Mikhailov 1980). The production and recognition of
these operations always requires a concrete body.3 But each communica-
3. Gallagher (this volume) and Lindblom and Ziemke (this volume) elaborate on
the link between production and recognition of body-produced movements
436 Wolff-Michael Roth
tive act does not have a sense in itself; rather, sense emerges from the rela-
tion between the speech act and the overall, collective activity. Communi-
cative action is therefore at the same time a concretely embodied (intra-
psychological) and sociocultural (inter-psychological) phenomenon; its
meaning emerges from the simultaneous grounding of action in cultural
activity and bodily operations. Conscious action therefore lies at the inter-
face between the unconscious operations that realize them and the cultural
processes that give them their sense.
3. Ethnographic context
The data analyzed here derive from an in-situ naturalistic study of commu-
nication in a seventh-grade science classroom where students learned about
the physics of simple machines largely by designing machines. To make
the learning environment as authentic as possible, the specifications for the
design tasks were articulated in a call for proposals. As part of their sub-
mission, the students included paper-and-pencil sketches (e.g., Figure 1)
and prototypes. They subsequently presented their prototypes to the student
collective, which instantiated a form of peer review. This curriculum there-
fore provided many opportunities for students to engage in conversations
with others, both during the design of the prototype and the review ses-
sions. Two cameras and two tape recorders were used to record small-
group and whole-class conversations.
To begin with, all recordings were roughly but word for word tran-
scribed to capture all utterances as completely as possible under the pre-
vailing conditions (e.g., with students hammering and using power tools).
In a second pass, the VHS videotapes were digitized (Macintosh iMovie);
offprint images capturing characteristic moments were imported into the
transcripts. Each lesson subsequently was broken into episodes, which
were exported to QuickTime versions. These digitized clips were then
analyzed using Peak DV 3.21, which allows the graphical representation
of waveform (for determining time, relative loudness, and duration) and its
Fast Fourier Transform (for determining pitch changes). Individual utter-
ances were timed with an accuracy of O.OOl seconds but rounded to the
(signs) and the role of mirror neurons to bring about this integration between
perception and action.
Communication as situated, embodiedpractice 437
nearest 10 milliseconds and coordinated with the frames that allowed tim-
ing to 33 milliseconds (30 frames/s rate of the VHS video).
4. Language and body in design activity
Linguistic analyses generally are concerned with language independent of
communicative praxis. This leads directly leads to the grounding of lan-
guage, that is, its connection to things and situations, which is continuously
realized in practice but has been a problem for traditional cognitive science
(Harnad 1990). Furthermore, most linguistic studies are hardly relevant to
speech, for the language of writing is a fundamentally different cognitive
phenomenon than language in practical face-to-face communication (Ong
1982). Here, I am concerned with language as it appears in face-to-face
communication alongside other semiotic resources humans produce in
interaction and for interactive purposes - body orientation, body move-
ments, and hand gestures. Following the presentation and brief gloss of a
23-second episode, analyses are provided of the concrete role played by
human bodies in communication.
4.1. A moment in design activity
The following episode was recorded at the moment when the three girls
were deliberating how to go about building the tower that would hold an
elevator, the first element in their design of a Rube Goldberg machine
(Figure 1).4 Leanne (far left in offprints) had previously explained that they
needed to build a two-sided tower; Amanda (far right) and Bella (center)
went along so that it appeared that there was agreement on this design.
Following Leanne's suggestion, the three had decided to cut the wood
needed to build that part of their design. At one moment, Leanne talked
about cutting two pieces of wood for the two sides of the structure. This
4. Rube Goldberg was a cartoonist who drew complicated machines that accom-
plished the most mundane everyday actions. Constructing a Rube Goldberg ma-
chine is a popular design activity of middle school students around the world.
More information is available at http://www.rube-goldberg.com/.
438 Wolff-Michael Roth
was when the following episode began.
5
(I first provide the raw transcript
and then a verbal description of the events.)
Figure 1. The students had decided to have a tower and a chute as the frrst two
elements in their Rube Goldberg machine, a complicated device that
would accomplish the simple task of feeding a cat. The episode con-
cerns the construction of the tower part.
5. To get a better sense for the changes in body positions and hand gestures in-
volved, readers should visually scan the images before reading the transcript.
The following transcription conventions have been used in line with traditional
conversation analysis: (0.41) : time in seconds; [ttwo sides] : square brackets
enclose speech overlapped by gesture or other body movement; [ t ( ( r ~ lH
moves up)) : description of gesture (rH, ill = right and left hand, respectively),
body movement; * : the asterisk aligns speech and video offprints; = : equal sign
shows latching, that is, two utterances are not separated by the normal pause;
.,?! : punctuation is used to indicate speech features, such as rising intonation
heard as a question, or falling intonation to indicate the end of an idea unit
(sentence); - : the n-dash indicates stop in utterance without voice inflection in-
dicating end of idea unit.
01 Bella:
02
03 Leanne:
04
05 Leanne:
06
07
08
09
10 Bella:
11 Bella
12
13 Leanne:
14 Bella:
15 Leanne:
Communication as situated, embodied practice 439
Okay, which two sides? *
(0.47)
Okay, you know [\how we are doing
that thing where the el]\bevator thing -
comesh bup*
[l(body rotates to the left, rH places
saw on table))
[z((rH moves upward, IH slightly up-
ward))
b((begins turning upper body,
(0.43)h
b((head turns toward board to right,
rHpoints to board))
[4these * are the sides]
[4((rH single stroke along the board))
[sthat like keep it] * [6standing].
[s((body turns to left, both hands come
up palms facing one another to pinna-
cle oftrajectory))
[6((two-handed gesture halfway down,
slightly up, then completely down))
(0.59)
hO.60]
h((two-handed up down gesture, at half
height))
[8The two sides?] *
[8((two-handed up gesture))
[9(0.88)]
=[9((nods three times, at the end, turns
to her right toward the piece ofboard))
[\oUh, okay.]
[1O((turns to her right toward the piece
ofboard))
440 WolfJ-Michael Roth
16 Leanne:
17
18
19 Amanda:
20
21 Leanne:
22 Amanda:
23 Leanne:
24
25 Leanne:
26 Bella:
27 Leanne:
[111 think it'll work. But if it doesn't, I _ " - : ~ M ! I "
know what we can-
[lI((PicAJ up saw, begins sawing))
(0.42)
fix how we can fix it.
Don't we need four?*
(0.56)
No,just-
Just two.
Two.
(0.44)
[12
Ri
ght?]*
[12((turns upper body, head to gaze at
Bella))
=[13((nods twice))] 11
[14((turns head to gaze at saw and
wood))
At the beginning of this episode, Bella, apparently confused about just how
the part was to look, then asked, "Okay, which two sides?" (turn 01). A
brief pause followed Bella's question, before Leanne began to answer,
"you know how we are doing that thing where the elevator thing comes up"
(turn 03). While uttering this description of the current state of affairs, she
turned her body toward Bella, laid down the saw she had been operating,
and moved her right arm and hand upward (turn 03). In the ensuing pause
in the verbal articulation, Leanne turned to her right, her hand moved for-
ward initially pointing to the board on the table, and then moving her hand
about 5 centimeters along one of the edges. It was precisely at the place
where she had previously shown with the saw where and how they needed
to cut the word (Figure 2). Contemporaneous with the gesture, she uttered,
"these are the sides" (turn 06). She oriented toward Bella again and, in a
continuous motion, produced two two-handed up-down hand movements,
palms facing, suggesting two pieces facing one another, while uttering,
"that like keep it standing" (turn 07).
After a longish pause, Bella mimetically reproduced Leanne's double-
handed gesture, before repeating the gesture all the while uttering, with
rising pitch, "The two sides?" At the verbal level, there followed a longer,
0.88-second pause (turn 10). But Leanne began the first of three head nods
just as Bella had completed her utterance (turn 13). That is, as a turn, her
head nod latched onto the ending of Bella's turn. Bella responded with a
sign of understanding and idea completion (falling pitch), "Uh okay,"
Communication as situated, embodiedpractice 441
while Leanne turned to the piece of board from which she had earlier indi-
cated the intention to cut two narrow pieces. She explained that she
thought her idea would work, and ifnot, she knew an alternative (turns 16-
18). She had begun to saw and would not stop for the remainder of the
episode.
Figure 2. The iconic gesture produced here with the saw moving along the
board where a cut was to be made was reproduced later in an episode
analyzed in this chapter.
At this point, Amanda, who had sat on the table next to her two peers, en-
tered the conversation for the first time in this episode. Up to this point, she
had faced away and now, turning her head only slightly to her right, asked,
"Don't we need four?" (turn 19). Continuing to saw and interrupted by
Amanda's own articulation "two," Leanne responded, "No, just two" (turns
21,23). After a brief pause and without stopping to saw, Leanne turned her
upper body (slightly) and head to gaze squarely at Bella in the process
uttering with rising inflection, "Right?" (turn 25). She had barely com-
pleted when Bella began to nod twice, upon which Leanne turned her head
back to look toward the board that she was working on.
In terms of the overall project, the episode was important in that what
had appeared to be a common design vision of the tower and the elevator it
subtended did not exist but was brought into alignment. Although the three
girls previously had talked about and seemed to agree on the design, and
although Leanne had explicated the two-sided version of the tower once
before, it was apparent that both of her two peers wondered in the course
of this episode why she wanted to cut two pieces of board rather than some
other number, including the four Amanda articulated. This alignment,
however, was not produced merely by talk. Rather, pointing (deictic) and
iconic gestures, head nods, body turns were important aspects of the com-
municative exchange. Without participants' attention to the body of the
respective other, they would find it difficult to produce the alignment
achieved here. The latter part of the episode exemplifies this: Amanda, not
442 Wolff-Michael Roth
having had perceptual access to the gestures and body orientations of the
other two girls, did not understand why there ought to be two sides to the
tower despite the fact that Leanne just explained it. Thus, Amanda began
the second part of the episode with fundamentally the same question that
Leanne had just answered in response to Bella's query.
In this episode, therefore, alignment of at least two of the three partici-
pants with respect to a future state of their activity was achieved. The epi-
sode had begun with a query about the number of sides for the tower, al-
though this issue seeme,d to have been settled during a previous interaction.
Twice during the episode, Bella aclmowledged the two-sided version of the
tower, once by means of an iconic gesture and confirming utterance (turns
10, 14) and then by means of a head nod. As Amanda acquiesced (turn 22),
we have no indication about her personal vision of what their collective
project was to be. In the following, I analyze different aspects of this epi-
sode, which bring to the foreground different dimensions of the relation-
ship of body and language in action.
4.2. Orienting the body and gesturing mediate turn taking
Communication is an interactive phenomenon requiring the analysis of
pairs of turns. A body of studies in conversation analysis provided an un-
derstanding of mostly audiotaped conversations (ten Have 1999). Whoever
speaks, holds the floor; others normally do not speak at the same time.
However, participants do not own pauses in speech. If the current speaker
has the intention to continue, he or she has to provide other semiotic re-
sources that indicate this intentionality. However, videotaped interactions
reveal that in face-to-face situations, there are many other semiotic re-
sources that participants bring to in play that mediate turn taking. In this
situation, those who want a turn at talk may provide bodily and gestural
resources that indicate this intentionality (Goodwin 2000). Mutual aware-
ness of the interaction participant's body orientation, perceptual access,
and (hand, head) gestures mediates the rules of interaction in ways that
traditional conversation analysis could not reveal. Videotaped materials, as
used here, reveal the central role of the body in face-to-face interaction.
The role of body orientation and perceptual access was quite apparent
in this episode, which had started with Bella's question about the nature of
the tower sides. Leanne responded, followed by a renewed query and af-
firmation that there would be two sides. During the interaction, both had
Communication as situated, embodiedpractice 443
produced gestures in the context of changing body orientations. Amanda
had been sitting on the table facing away from the two. Thus, she had no
access to all those parts of the communicative interchange between her
peers other than the verbal utterances. In particular, she had not been in the
position to perceive either Leanne or Bella's two-handed production of an
image that pertained both to the role of the two pieces of board that they
were in the process of cutting and the design drawing. Her question "Don't
we need four?" essentially repeated Bella's question about the relationship
between raw materials, design drawing, and their vision of the emerging
artifact. The words alone had been insufficient to bring Amanda in align-
ment with what was to be the collective vision for the unfolding design
artifact.
Pauses are interactive resources that play a particular role in the situated
evaluation of communicative action (Bourdieu 1980). As a pause in talk is
lengthening, it increasingly becomes a resource for any participant to begin
speaking. Following a question, however, a pause becomes an increasing
liability for the addressee, as not answering can be heard as an intentional
exclusion of the querying person from interaction. This happened repeat-
edly in this group, where Leanne did not answer queries that Amanda di-
rected to her, whereas she answered all questions that Bella posed. Thus,
the lengthening pause following Amanda's question (turn 19) opened up
the possibility that this situation, too, would be one in which Amanda did
not receive a response.
Perceptual access to the interlocutor's body changes the rules identified
in conversation analytic studies. Body movements required by the interac-
tion may take time, but the very fact that they are enacted articulates com-
municative intention. So actions other than speech also articulate the intent
to contribute, and therefore that the current turn lies with the acting person.
For example, after Leanne first articulated the sides of the tower ("that
thing where the elevator thing comes up"), there was a pause at the verbal
level. However, at that time she turned her body and moved arm and hand
nearly 90 degrees around to her right. The motion ended in the pointing
gesture and led into the gestural and verbal articulation of the target as "the
sides." Later (turn 13), she nodded three times in response to Bella's query
for confirmation of the two-walled nature of the future artifact. The head
nodding was therefore a different form of turn taking than what has been
available to traditional conversation analysis. In the head nod, the body
articulated an affirmation, which in a different spatial arrangement might
444 Wolff-Michael Roth
have come forth through an activation of the vocal cords. The relationship
between pause and orientation was also evident in the following situation.
In the final part of the episode (turns 25-27), Leanne had answered
Amanda's question about the number of sides of the tower. There followed
a pause. Leanne then uttered with rising pitch, "Right?" (turn 25) simulta-
neously turning her head to Bella (still continuing to saw). Leanne thereby
held Bella accountable for not having confirmed the answer, not only in
asking for a confirmation but also, by turning her gaze in the direction of
Bella, designating her as the addressee of the question. Bella did not re-
spond verbally, but, immediately following the end of Leanne's head
movement, nodded twice. Without a word, Leanne returned her head to
gaze again at the saw, board, and sawing action. Bella had answered with a
double nod, which would not have made sense had she not been aware of
the fact that Leanne could see her. Leanne's subsequent turn was an ac-
knowledgment that the request for confirmation had been satisfied, and
therefore that she had seen the nods. Bella, too, could know that Leanne
was satisfied with her non-verbal response, for she could have expected a
conversational repair if this had not been the case.
A converse case existed earlier in the episode when Bella acknowledged
the confirmation received to her query about the two sides (turn 14). Bella
was moving her head downward as part of a grooming gesture and Leanne
was turning around. Bella therefore was forced to utter "Ub okay" rather
than enacting acknowledgment in another way (head nod, eye movement,
literal "thumb up"). First, because Bella turned her gaze away from
Leanne, she was not able to see whether her peer was in the position of
seeing her gesture. Second, because Leanne was moving, she could not
have seen the gesture either. It is likely that the verbal utterance was form-
ing at the same moment that the downward head movement was. In yet
another instance (turn 10), Bella had produced a two-handed gesture while
looking down at her hands; she reproduced the gesture this time uttering
questioningly, "The two sides?" while simultaneously raising her gaze to
Leanne's face. This provided a resource for Leanne to respond by nodding
but did not require a verbal equivalent, which, here, was not provided.
These analyses exemplify that the mutual bodily orientation of the par-
ticipants mediated how communication is articulated, that is, verbally or
gesturally (e.g., hand, arm, head). Participants are not only aware of their
own positions and orientations but also of the positions and orientations of
their interlocutors. Speech is the preferred mode if there is evidence that
the listener does not have access to bodily gestures and that both speaker
Communication as situated, embodiedpractice 445
and addressee are aware of this. The speaker may produce gestures when
the listener has access, even though the speaker herself does not face the
listener. These rules also mediate the pauses between verbal utterances. If
participants have full perceptual access to one another, a lot of the commu-
nicative work is generally done by producing non-verbal semiotic re-
sources, such as indicating intent to take the turn, articulating ideas, ori-
enting the communicative partner toward, and so forth.
4.3. Pointing
Pointing is classified on the spontaneous-idiosyncratic-movement end of a
gesture continuum that reaches to fully developed linguistic systems (e.g.,
American Sign Language) on the other end (Kendon 1988). Acts of point-
ing occur in situations that involve at least two participants, one of whom
is, in the act of pointing, producing resources for the joint establishment of
a particular space as a common focus (Haviland 2000). This common focus
constitutes a resource for the organization of subsequent situated action
and cognition. Pointing is jointly contextualized by semiotic resources that
include, at a minimum, a body that visibly performs the act, a dialectically
related speech act, the material properties of the in pointing targeted entity,
the orientation of relevant participants toward one another and toward the
target space, and the larger activity within which the act of pointing (itself
only part of the communicative act) is embedded (Goodwin 2003).
In this episode, there were repeated instances of verbal deixis ("that,"
"it," and "these"). One of these instances, "these," immediately followed a
deictic gesture (turn 05) and overlapped with an iconic gesture whereby the
hand moved along the piece of wood (turn 06). Such acts of pointing con-
stitute prime examples where the double nature of the object in activity
becomes apparent: from the speaker's perspective, it appears twice, once as
material and once in perception. The body oriented itself toward an entity
other than itself (piece of wood), which it was perceptually isolating from a
more inchoate ground. The body, here, repeatedly pointed to this entity,
verbally by means of the utterance "these" and gesturally by the stretched
arm and edge of the hand coming to rest on the piece of wood, the current
object of attention.
The deictic term "these" locates the entity pointed to near the speaker
(Hanks 1992), in the present case, within reach. The deictic term "these"
not only instructed Bella to attend to something beyond the speech act, the
446 Wolff-Michael Roth
targeted material structure, but also articulated that what she pointed to as a
multiplicity (i.e. "these" rather than "this"). Furthermore, by using "these"
rather than an articulation of a point in space (here, there), Leanne formu-
lated the target as material things rather than spaces or sounds.
There existed, however, a contradiction: there was only one piece of
wood, and Leanne' s subsequent iconic gesture produced the image of one
rather than two cuts. But her pointing gesture was not only to the piece of
wood but also to the previous interaction where she had gesturally enacted
twice the iconic gesture of two cuts. Because these previous gestures were
part of the girls' interactional history, they constituted production re-
sources "once-removed," which make connections over larger chunks of
the ongoing activity (Erickson 1982). This pointing (turn 05) therefore
indexed the earlier sawing gesture with the edge of the right hand at the
same place where she had earlier moved the saw (Figure 2).
In a continuous motion, the pointing configuration then moved along a
line about five centimeters from the edge where she had earlier indicated
that they needed to cut (Figure 2). Although most classification systems
categorize deictic and iconic gestures differently, there are situations where
the same movement enacts both types, for example, when the pointing
body part traces out some shape. Thus, Leanne's initial deictic gesture
toward the piece of wood (turn 05) also became an iconic gesture (turn 06),
because its trace took the shape of the projected linear cut. Here, move-
ment along a line that separated the object pointed to from what was to be
taken as diffuse ground enhanced salience. The single iconic gesture along
the board constituted a reproduction of an earlier gesture showing where
she was to cut the board (Figure 2). In fact, at that earlier moment, Leanne
produced the gesture twice, holding the saw where the cuts were to be
done. That is, the distance of the saw from the edge doubled for the second
movement in the first pair of gestures; in the second pair, she produced for
practical purposes identical gestures at either end of the board, again indi-
cating that they needed two pieces of board for the tower part. Despite the
iconic relation between the earlier and present gestures, the contradiction
between the gestural (one) and verbal articulations (more than one) was
only resolved in the subsequent iconic gestures (turn 07) that articulated a
vision of two upright boards, as it was further materialized in a subsequent
episode (Figure 3).
From the listener's perspective, the associated task is to perceptually
isolate the entity pointed out by the speaker's body (orientation, deictic
gesture). This is not necessarily an easy task, for even in the sparsest set-
Communication as situated, embodiedpractice 447
tings, it may not be evident to the listener what the salient feature is for the
speaker. Different speakers may use the same word for denoting different
material entities or may use different words for denoting the same material
entities (e.g., Roth 1995). Other, semiotic resources are required as con-
straining factors to assist the listener in perceptually articulating the target
structure. Such constraining factors are inculcated culturally marked object
or field, or may be found in the history of the activity and associated talk.
In the present episode, the perceptual field articulated in Leanne' s pointing
was not culturally marked, but marked by having been the focus of previ-
ous interactions. By turning to the piece of wood, pointing it out as a re-
source, and then returning to the initial body orientation facing Bella,
Leanne actually brought the pieces of wood into the space between them,
where she gesturally enacted a vision of the future artifact. In the present
case, Leanne had already used repeatedly an iconic gesture to point out
where and how many times they needed to cut for getting the required
pieces in the building of the tower part. The nature of the entity pointed to
was further constrained by the talk that immediately followed, where
Leanne's two-handed palm-facing gesture articulated a vision of boards
facing one another (turn 07, Figure 3).
Figure 3. The two-handed gesture, which had fIrst appeared in a conversation
about what the tower would look like is here reproduced but here in a
holding movement, fIrst by Leanne (a), then by both students (b).
In the present situation, Leanne was not just pointing somewhere, but she
rotated her entire body. The sequence of the first four offprints in the tran-
script (turns 01-07) shows that she had first rotated away from an orienta-
tion to the board (turn 01) in order to face Bella (turn 03), then rotated back
to face the board while pointing to it (turn 06), and then rotated again to
face Bella (turn 07). She was aware that Bella shifted her gaze from facing
Leanne (turn 03) toward the board (turn 06), the endpoint of the gesture,
448 Wolff-Michael Roth
and then back toward the space between them, where she could see both
Leanne into the face and her two-handed gesture. This intentional pointing
therefore was instantiated not merely the hand but in fact by the entire
body (person) under the auspices of her knowing that the listener Bella was
following her changing orientations.
4.4. Iconic gesturing
Iconic gestures belong to the same class of gestures as pointing. They in-
clude those hand-arm movements that bear a perceptual relation with con-
crete or projected entities and events that the speaker is oriented to (e.g.,
McNeill 1992). In contrast to a deictic gesture, however, an iconic gesture
is itself the focal entity. From a speaker's perspective, the iconic gesture
articulates an immanent or unfolding sense that stands in a dialectical rela-
tion both with concurrent speech and the material entity in the setting so
articulated (Roth 2004). Iconic gestures are important semiotic resources in
interaction (e.g., Ochs, Jacoby and Gonzales 1994). Research in school
science classrooms showed that for students of all ages, iconic gestures are
important aspects of communication, especially while the learners are be-
coming familiar with the phenomena at hand and do not yet have a stable
(viable) language to talk about them (Roth 2003). Iconic gestures also were
important interactional resources continuously produced and reproduced in
this episode.
The first iconic gesture occurred while Leanne reminded Bella that they
were currently working on the first part ("thing") of their Rube Goldberg
machine where the elevator ("thing") moves (turn 03). The rising right
hand palm facing down evoked the image of something moving upward,
like the elevator in the design (Figure 1). More so, while they had talked
about how to realize the tower, both Leanne and Bella had gesturally fol-
lowed the vertical outline of the tower, indicated by the orientation of the
hands. These gestures therefore bore iconic relations to other material
forms, apparent as tower and elevator in the design drawing. At the verbal
level, the upward movement of the hand was iconic to the articulated
"comes up." It is a concurrent gestural and verbal concretization of the up-
down or verticality schema (Johnson 1987, Roth and Lawless 2002b). In
this case, the timing between the verbal and gestural modality was precise
(simultaneous), consistent with the contention that gesture and speech are
simultaneous formulations of an immanent or emerging idea by one and the
Communication as situated, embodiedpractice 449
same body (McNeill 1992). However, the left hand only moved upward
slightly, about one-quarter the way of the right hand, and given Bella's
gaze direction, probably invisible to her.
In the same part of the episode, one can also note that the idea had be-
gun to form earlier. While uttering that they were "doing that thing,"
Leanne still had the saw in her hand, nearing the end of the body rotation.
She placed the saw and then, in a continuous motion, moved her right hand
up just as the utterance evoked an upward moving elevator. That is, the
placement of the saw was in preparation of the subsequent gesture. But
even before uttering the locative "up," her body and hand already began to
rotate sideways toward enacting the deictic gesture that pointed to the
wood to be used in the tower construction, thereby preparing the subse-
quent gesture.
After the intervening (already analyzed) pointing episode, Leanne' s
whole body turned back toward Bella and, in a continuous motion, both
hands moved upward while she uttered, "that like keep it" (turn 07). The
hands then descended halfway to the table, but, as if hesitating moved
three-quarter the width of her hand upward, before completely descending
to the table level; in the process, she completed the word "standing." The
two-handed gesture produced the image of two upright boards facing one
another, which, according to the verbal part, "keep it [elevator] standing."
The two-handed gesture not only produced an iconic expression of the
design vision, but also produced (a) an image of the two boards that pro-
vided the scaffold for the elevator in their design drawing and (b) an image
of the boards that she had just pointed to in their raw-material state.
Through the two-handed gesture, the design drawing (Figure 1) was com-
ing alive while simultaneously implicating the raw materials to be used: it
was a bodily realization of the vision. Later on, very similar hand move-
ments included pieces of wood now standing for themselves as the "sides"
(Figure 3a). At that point, the design was even further concretized but not
yet an independent artifact: Leanne's body was still implicated by holding
the pieces. Only subsequently, after adding more materials and using fas-
tening resources would the emerging design become self-supporting and
exist independent of the body. In this situation, we observe an opposite
movement to my previous research on gesture in learning science (Roth
and Lawless 2002a): here, ideas and fleeting images are initially concre-
tized in body position and hand gestures whereas the inverse was observed
in learning by experimenting.
450 Wolff-Michael Roth
At this point, Leanne had apparently ended her response to Bella's
query. There was a long pause in the talk. During the second half of this
pause, however, Bella reproduced a mimetic image of the two-handed
gesture Leanne had just completed - and, similarly to her peer, used the
same hand movements to hold up the two pieces of wood (Figure 3b). She
then reproduced the same two-handed gesture uttering with increasing
pitch, "The two sides?" (turn 11). Here, although the hands mimetically
reproduced Leanne's gesture, thereby confirming that she had perceived it,
the verbal utterance indicated uncertainty. Why should there be uncer-
tainty? What must have appeared in doubt was the relation to the diagram,
that is, their ongoing activity and its motive to construct a Rube Goldberg
machine, with an elevator as its first part. That is, from the receiver's part,
Leanne's production was intelligible as such (Bella reproduced it), but she
had not yet found the relation Leanne' s gestures bore to the envisioned
object of activity, partially concretized in the diagram.
Facing Bella, who looked straight at her, Leanne nodded three times;
she did not produce a verbal affirmation. As shown previously, Bella was
in a position where her sign of affirmation had to be produced verbally.
Leanne then turned around to continue where she had left off. Leanne had
articulated her version of the design in response to Bella' s query about the
sides involved in the elevator tower. Thus, although they had talked about
this before, that is, although Leanne had articulated this vision in a previ-
ous episode, there was apparently a lack of alignment or common ground
between the two. Bella's query articulated this uncertainty about the nature
of tower in what was to be their common design. The iconic and mimetic
character of Leanne's two-handed gesture thus constituted a solution to the
problem of establishing common ground with respect to an initially
ephemeral vision of the object of activity, which was increasingly concre-
tized in the unfolding of the design artifact. Bella enacted an independent
concrete version of what is taken to be the common object at hand, always
under revision in future action.
5. Body, mind and culture
In this chapter, I articulate and exemplify the embodied and situated nature
of human communication drawing of a non-dualist, dialectic framework
that integrates mind and culture, body and mind, speech and gesture, mate-
riality and conscious nature of human experience, and so on. The analyses
Communication as situated, embodiedpractice 451
show that the separate, interlocking displays of conversational participants
form a unit that is greater than the sum of its parts, a particular kind of
participation framework. Bella and Leanne treated each other's bodies as
resources in the ongoing production of intentional action; Amanda, though
she was close and overheard the two, did not have access to the same re-
sources, as she faced away. For conversation participants, the body is a
complex resource that enables different kinds of articulations that stand in
a dialectically relation to, and therefore elaborate one another. The body
therefore is a different type of entity than the piece of wood Leanne
pointed to, the saw she used, or the paper-and-pencil diagram they had
collectively produced, although it is of the same material kind.
6
Partici-
pants in communicative interaction must not only attend to multiple per-
ceptual arrays (body orientation, gesture, speech, entities in the setting),
but also arrays that are significantly different in their structure.
In the featured episode, there were repeated instances of thoughts in the
process of forming concurrently with the bodily productions (speech, ges-
ture, orientation). Thus, Bella had already produced a two-handed gesture
that enacted the image of the two-walled tower before she began to utter
the question, "The two sides?" while simultaneously reproducing the two-
handed gesture. In the first instance, one sees the question forthcoming, not
yet formed in its entirety; it only became the full interactional resource
when both gesture and speech aligned in the articulation. In fact, the first
hand movement had a shorter trajectory, as if Bella had stopped it to repro-
duce it in full simultaneous with the utterance. This process was not
planned, as gesticulations are generally produced unconsciously. The con-
tinuous unfolding of a thought was also noticeable in Leanne's perform-
ance of her response to Bella's initial question. It had started with a refer-
ence to the diagram, continued by articulating the pieces of board that they
were currently working on, and then articulated the design vision that
brought the two earlier pieces together into one unit. The analyses showed
how her entire body performed this response.
The analyses also showed that the body, through forms of both sponta-
neous and idiosyncratic gesture, is tied to its environment so that commu-
nication is not only an embodied (production of speech, gesture) but also a
situated (distributed) phenomenon. In the deictic gesture, the identity of the
6. Gallagher (this volume) elaborates on the differences between the perceptions
of material objects and human bodies (that of others and one's own) and the
implications this has for intersubjectivity.
452 Wolff-Michael Roth
perceptual (intra-psychological) and material (extra-psychological) aspects
of the object becomes apparent and articulated. It is in a reflexive move-
ment that this identity is destroyed, that is, that subject becomes separated
from object and object from subject; the reflexive movement, however,
only returns the idea or thought but not experience of a body (Merleau-
Ponty 1945). The body is therefore not only a different type of entity, but
also it cannot be the object proper in the phenomenological or activity-
theoretic formulation. The proper analytic starting point is the dialectical
unity of activity - bodies, thoughts, experiences, and even emotions and
motivations emerge as the effects of the subject I object dialectic that con-
stitutes the activity (Roth 2006).
The analyses also provide a hint at the fact that body, mind, and culture
are inextricably related and cannot be reduced to one another. Thus, each
utterance, gesture, body orientation, and so on produced by one participant
not only provided a resource for future action to herself but also to the
other; each communicative act was explicitly produced for the other. That
is, the idea of a semiotic resource inherently requires the other participant
to be aware of the semiotic potential of these resources. At the moment that
the body articulates itself by producing semiotic resources, these already
have to exist for the recipient other. An individual bodily me that articu-
lates itself has to draw on resources that are already intelligible by the
other and therefore part of the cultural possibilities. This, however, should
not come as a surprise, if we recall that operations (and therefore schemas)
are unconscious collective consciousness. All actions are therefore synthe-
sized from operations, the concrete realizations of cultural (generalized)
possibilities, and therefore are simultaneously the person's and the collec-
tive other's.
The present analyses of one brief event testify to the complexity of
communication, which cannot be reduced to language. The analyses ar-
ticulated communication as an embodied and situated phenomenon and
showed how ideas, questions, and responses emerged in real time. The
episode analyzed here was not different than those preceding or following
it during the six-hour project, or, for that matter, any other design activity
by these students or their peers. If 23 seconds of situated activity are al-
ready so complex, how much more complex will be entire forms of activi-
ties that stretch over days, months, years, and generations? Because com-
munication, ideas, questions, and responses are distributed across different
bodily modes of expressions and across the salient aspects of the setting,
the structure of thought cannot be linguistic structure. Thought cannot be
Communication as situated, embodiedpractice 453
reduced to language; nor can any other form of expression be reduced to
the body. Rather, thought exists in the dialectical unity of all forms of ex-
preSSIon.
The present analyses make salient the special role of the body, seat and
source of agency, in design activity, constituted in the unfolding ephemeral
ideas into concrete artifacts that we had also observed among environ-
mental activists (Lee and Roth 2001). The body was the first site where
ideas took shape, where they were articulated; it was then augmented by
the introduction of materials and subsequently emerging stages of the arti-
fact. Still later, the body removed itself as additional actions allowed the
artifact, the concrete rendering and development of the initial ideas, to take
on a life of its own. Communication was but one form of action that moved
the unfolding activity along its trajectory toward completion. This role of
gestures complements the findings of our earlier research (e.g. Roth 2002),
which showed the mediation gesturing plays in the development of lan-
guage following manipulations of objects. That is, hand gestures in par-
ticular again played an important role in the transition between two repre-
sentational poles, here, ephemeral design ideas sketchily articulated in
speech, on the one hand, and concrete artifacts, on the other.
There are direct practical consequences of the way of thinking about
and analyzing human communication. For example, in lecture-type situa-
tions, students generally copy whatever the lecturer notes on the inscription
device present; they may also note some of the utterances that do not get
publicly recorded. However, students generally do not record all the other
semiotic resources that the lecturers provide in their performances. It is
therefore not surprising that students experience difficulties to reproduce
the sense of what has been communicated from their notes alone. It also
should not surprise that so many students experience difficulties in under-
standing the lecture content at some later point, although during the lecture
it all had made sense. The perspective developed here suggests that these
students are in fact missing most of the resources that were available to
them in the lecture but have neither been recorded in their notebooks nor
were recoverable from the existing entries.
Acknowledgments
This study was made possible in part by a grant from the Social Sciences
and Humanities Research Council of Canada. I thank Sylvie Boutonne,
454 Wolff-Michael Roth
Michelle K. McGinn, and Carolyn Woszczyna for the help they provided
during data collection.
References
Bavelas, Jan B., Nicole Chovil, Linda Coates and Lori Roe
1995 Gestures specialized for dialogue. Personality and Social Psychology
Bulletin 21: 394-405.
Bourdieu, Pierre
1980 Le sens pratique. Paris: Les Editions de Minuit.
Erickson, Frederic
1982 Money tree, lasagna bush, salt and pepper: Social construction of
topical cohesion in a conversation among Italian Americans. In:
Deborah Tannen (ed.), Analyzing Discourse: Text and Talk, 43-70.
Washington, DC: Georgetown University Press.
Goodwin, Charles
2000 Action and embodiment within situated human interaction. Journal of
Pragmatics 32: 1489-1522.
2003 Pointing as situated practice. In: Sotaro Kita (ed.), Pointing: Where
Language, Culture and Cognition Meet, 217-241. Mahwah, NJ:
Lawrence ErlbaumAssociates.
Goodwin, Charles, Marjorie H. Goodwin and Malcah Yaeger-Dror
2002 Multi-modality in girls' game disputes. Journal of Pragmatics 34:
1621-1649.
Hanks, William F.
1992 The indexical ground of deictic reference. In: Alessandro Duranti and
Charles Goodwin (eds.), Rethinking Context: Language as an Inter-
active Phenomenon, 43-76. Cambridge: Cambridge University
Press.
Hamad, Stevan
1990 The symbol grounding problem. Physica D 42: 335-346.
Haviland, John B.
2000 Pointing, gesture spaces, and mental maps. In: David McNeill (ed.),
Language and Gesture, 13-46. Cambridge: Cambridge University
Press.
Hindmarsh, Jon and Christian Heath
2000 Embodied reference: A study of deixis in workplace interaction.
Journal ofPragmatics 32: 1855-1878.
Johnson, Mark
1987 The Body in the Mind: The Bodily Basis ofImagination, Reason, and
Meaning. Chicago: University of Chicago Press.
Communication as situated, embodiedpractice 455
Kendon, Adam
1988 Goffman's approach to face-to-face interaction. In: Paul Drew and
Anthony Wooton (eds.), Erving Goffman: Exploring the Interaction
Order, 14-40. Cambridge: Polity Press.
Latour, Bruno and Steve Woolgar
1979 Laboratory Life: The Social Construction ofScientific Facts. Beverly
Hills, CA: Sage.
Lave, Jean
1988 Cognition in Practice: Mind, Mathematics and Culture in Everyday
Life. Cambridge: Cambridge University Press.
Lave, Jean and Etienne Wenger
1991 Situated Learning: Legitimate Peripheral Participation. Cambridge:
Cambridge University Press.
Lee, Stuart and Wolff-Michael Roth
2001 How ditch and drain become a healthy creek: Representations,
translations and agency during the re/design of a watershed. Social
Studies ofScience 31: 315-356.
Leont'ev, Alexei N.
1978 Activity, Consciousness and Personality. Englewood Cliffs, NJ:
Prentice Hall.
McNeill' David
1992 Hand and Mind: What Gestures Reveal about Thought. Chicago:
University of Chicago Press.
McNeill, David and Susan D. Duncan
2000 Growth points in thinking for speaking. In: David McNeill (ed.),
Language and Gesture, 141-161. Cambridge: Cambridge University
Press.
Merleau-Ponty, Maurice
1945 Phenomenologie de la perception. Paris: Gallimard.
Mikhailov, Feliks
1980 The Riddle ofSelf. Moscow: Progress.
Ochs, Elenor, Sally Jacoby and Patrick Gonzales
1994 Interpretive journeys: How physicists talk and travel through graphic
space. Configurations 2: 151-171.
Ong, WaIter J.
1982 Orality and Literacy: The Technologizing of the Word. New York:
Routledge.
Roth, Wolff-Michael
1995 Affordances of computers in teacher-student interactions: The case of
Interactive Physics. Journal of Research in Science Teaching 32:
329-347.
456 Wolff-Michael Roth
2002 From action to discourse: the bridging function of gestures. Journal
ofCognitive Systems Research 3: 535-554.
2003 Gesture-speech phenomena, learning and development. Educational
Psychologist 38: 249-263.
2004 Perceptual gestalts in workplace communication. Journal of Prag-
matics.
2006 Identity as dialectic: Making and Re/making self in urban schooling.
In: Joe L. Kincheloe, kecia hayes, Karel Rose and Phil M. Anderson
(eds.), The Praeger Handbook of Urban Education: 143-153. West-
port, eT: Greenwood.
Roth, Wolff-Michael and Daniel Lawless
2002 a Signs, deixis, and the emergence of scientific explanations. Semiotica
138: 95-130.
2002 b When up is down and down is up: Body orientation, proximity and
gestures as resources for listeners. Language in Society 31: 1-28.
Sewell, William H.
1992 A theory of structure: duality, agency and transformation. American
Journal ofSociology 98: 1-29.
Streeck, Jiirgen
1996 How to do things with things: Objets trouves and symbolization.
Human Studies 19: 365-384.
ten Have, Paul
1999 Doing Conversation Analysis: A Practical Guide. London: Sage.
Uexkiill, Jakob von
1972 Theoretische Biologie. Frankfurt/M: Suhrkamp. (Originally pub-
lished in 1928)
Index
action 1,3,8,20,21,26,34,36,41,
42,45,46,58,72,109,132,136,
138, 139, 147, 152, 167, 169,
179-183,185,186,188-190,
199,200,204,205,211,214,
219,241-251,253-259,271-
275,277,287,307,313,320,
323,324,326,328,351,369,
381,382,384-388,392,394,
395,397,398,402,426,431-
436,442-445,450-453
active perception 197, 198, 204-
206,214
affordances 7,8,55,68,69,72-74,
76,98,99,186,251,257
agencYlstructure dialectic 431, 433
animal 43, 44, 48, 49, 65, 66, 69,
71,72,76,94-96,98,101,102,
134-138,140,141,168,177,
189,214,218,245,307,321,
327,365,379-381,383,384,
386,391-393,395,396,398-
400,402,403,405
biosemiotics 7, 9, 85, 101, 102, 105,
109,379,380,382,383,385-
391,394,395,400,402
bodily mimesis 219,228,231,232,
297,301,312,318,320,322-
324,326,329
body 2-4,6-9,17,18,22,24,25,
30,35,37,38,40,41,43,47,48,
55-59,62-64,67,69,70,73,75-
77,85-91,93,96,101,108-114,
116-120,130-134,137,138,
142-145, 147-149, 151-153,
169,179,180,183,189,205,
206,217,219,225,228,231,
241-243,252,253,257-259,
271-288,297,298,301,308,
312,313,318,320,325,329,
339-348,350,351,356,364-
370,379-384,388,389,392,
395-400,403,405,412,414,
416,431-435,437-447,449-453
body image 8, 217, 228, 231, 271-
283,285,288,308,313,351
body schema 8, 205, 217, 271-284,
286,288,308,313,351,434
categorization 167, 169, 175, 176,
178,190,206-208,212,214,
215,231,371,376,411,413-
415,417,426
causality 65, 314, 315, 380, 384,
386,387,393,401
cognitive linguistics 1, 6-8, 17, 85,
113,119,120,198,220,259,
298-300,304,306,308,315,
318,325,339,346,357,358
cognitive neuroscience 49, 144,
183,185,252,339,341,346,
351,361,363,369
cognitive science 1, 3-5, 7, 9, 10,
17-21,47,48,81,86,92,98,
119, 129-131, 133, 141-151,
154,168,197,200,242,250,
252,271,273,297,310,311,
313,315,316,339-341,343,
344,347-349,356,358,359,
362-364,369,370,379,382,
383,386,387,393,395,412-
414,437
communication 9, 43, 44, 46, 47,
115, 130, 131, 138, 152, 153,
208,218,223,228,252,255,
458 Index
284,307,319,320,324,381,
382,416,423,431-433,435-
437,444,448,450-453
complex systems 197, 199, 200,
231,386
conceptual spaces 8,167,168,171,
173,175,179,182,185,187,
190, 191, 197, 199
consciousness 2, 6, 7, 9, 22, 85-89,
91,97,98,100,102,107-109,
113,119,135,136,201,202,
217,219,222,228,232,275-
277,279-281,297-300,302-
304,307,309,312,313,315-
318,322-325,363,389,392,
395,398,400,402,433,434,452
conventions 297, 303-305, 310,
324,326,328,434,438
cyborg 379,380,399,402-405
deafferentation 271, 273, 277-279
dynamic categorization 197, 204,
206-208,213
ecology 7,55,56,64,75,85,100,
101
embodied cognition 1, 3-8, 10, 18,
48, 129, 131, 144, 145, 148-151,
154,206,241-243,249,256,
258,297-300,311,312,348,
356,360,364,393,431
emb0dnnent 1-4,6-9,17,18,21,
48,55,56,76,85-89,91,109,
118-120,129-133,137,143-
146, 148-153, 185, 187, 197,
206,207,230,241-243,249,
252,255,258-260,271,272,
280,284,286,288,297-301,
308,309,311-314,318,319,
322,324-329,339,340,342,
344,348-354,356-360,363,
365,366,368-370,379-386,
389,390,392,394-404,411-
414,416-418,423,425,426,
416-418,423,425,426,431,
434,453
emergence 23,86,139,219,231,
258,380,383,384,387,401,423
evolution 3, 80, 85, 114, 116, 134,
209,215,231,258,284,301,
318-322,326-328,383,387,
390,394,403,432
force dynamics 8,167,169,172,
179-183,186-191
frames of reference 339, 343, 345,
346,352,354,362-365,367-369
gesture/-s 6, 9, 129, 130, 134, 153,
202,227,228,231,241,255,
258,282-284,301,307,319-
321,323,324,326-329,362,
365,366,415,431,435,437-453
Gibson, James 7, 55, 56, 61, 64-69,
71-78,79,97-100,102,108,
122, 156, 182, 183, 186, 189,
193,198,204,205,207,234,261
history of cognitive science 129-
131, 133, 138
image schema/-s 17,21,32-37,42,
44,45,85,86,90,92,109,111,
114, 120, 167, 187, 189, 191,
198,309,314,324,325
information 7, 48, 55, 62, 68, 69,
72,76,77,105,115,142,146,
154,168,178,181-183,185,
189,227,246,249,255,315,
316,340,382,383,386-388,
390,391,394,397,412,432,435
interactive technology 129, 133,
151, 153, 154
internal meaning space 8, 197, 198,
204,214-217,219,222,223,
225,227,229,231,232
intersubjectivity 8, 133, 241, 252,
271,285-288,301,307,351,451
Japanese mimetics 197, 199, 223,
224,226-228,230
language 1,2,4,7-10,17,19,20,
22,23,31,36,37,41,42,44,45,
47,48,59,61,63,66,69,87,93,
96,99,101,104,106,107,110,
115-118, 120, 130, 131, 133,
134,139,140,148,152,173,
174,197-200,202,204,216,
218-220,223,226-228,230-
232,241,243-245,253,255-
260,283,297-301,303-314,
318-323,325-329,339,345-
347,349,350,353-355,359,
361,362,364,365,367,380-
382,393,400,402,404,411,
413-418,425,426,431,432,
437,442,448,452,453
Lifeworld 85,87,88,99-101,107,
109,110,120
memory 58,59,85,94,114-119,
142,184,203,212-214,229,
231,283,350,397,414,418,432
mental rotation 9, 250, 339, 341-
343,345-348,356,359,362,
364,365,368
metaphor 4, 5, 7, 9, 17, 20, 37-42,
48,55,61-63,98,111,129,130,
133,142,144,202,218,250,
301,308,310,312,339,357,
358,362,363,366,368,391
mimetic schemas 114, 180, 199,
203,228,231,297,301,322,
324-329
mirror neurons 8, 133, 147, 148,
150,189,217,241,251-255,
271,287,328,436
mutualism 55,75-77
Index 459
nature-culture dualism 55-57, 69,
76, 77
neonate imitation 271, 273, 279,
280,284,285
neurobiology 17,21,26,29,48,88,
91,308,319,341,387
organicism 380, 390, 391
perception 8, 23, 27, 32, 33, 35, 36,
44,60,61,73-75,78,88,95,96,
98,99,106,108,117,132,138,
147,149,167,169,170,172,
182, 183, 186, 190, 193, 195,
200,201,203-205,207,208,
214,217-219,223,226,231,
241-245,249,250,253,257-
259,271-276,278,281-284,
286,287,311,355,381,400,
402,416,418,420,421,425,
433,436,445
phenomenology 3,7,65,85-91,97,
120,150,200,206,218,271,
272,285,288,297,309,351,391
picture 5,44,67,76,85,96,98,99,
116,130,200,205,231,250,
306,311,329,348,367,387
pragmatism 17-21,32,37,39-42,
47,48,306
proprioception 55, 74, 225, 278,
279,284,286,320
radical embodiment 3, 129, 147,
149-151,386
representation 8, 9, 20, 25, 30, 43,
46,69,86,92,96,104,114,115,
144,147,149,152,168,170,
173,177-180,182-184,186,
190,197,198,200,201,203,
213,219,220,223,226,228,
230,231,244,246,252,259,
260,297-300,306,309-312,
317,322-325,350,364,367,
460 Index
369,379,380,383,388,393,
400,436
Representationalism 17, 19-21, 30-
33,35,36,42-48
self-organization 390, 411, 418,
423,425
semantics 8, 17, 167, 176, 187, 188,
191,192,223,304,349,357,
382,393
semiotic function 85, 93-95
semiotics 7, 85, 89-92, 94, 99, 105,
132,351,381,383-386,388,
393,394
sign 85,86, 92-109, 114-116, 118,
120,200,223,226,304,319,
320,381-384,386-389,391-
395,402,438,440,450
simulation theories 5, 12, 147, 241,
243,245,250,258-260
situated practice 431-433, 450-453
social interaction 7, 9, 43, 47, 48,
129-133, 137, 140, 144, 145,
147,148,151-153,252,253,
307,309,313,314,411,417,426
society 141, 381, 382, 397, 399,
400,403-405
space 8,9,23,28,29,33,38,66,
67, 86, 87, 93, 110, 113, 119,
167-178,183,185,186,188,
190,197-199,214-217,219,
224,225,231,232,281,299,
317,319,339-341,346,360,
367,369,414,417,423,426,
445-448
speech 9,47,59,140,174,181,
223,226,231,255,305-307,319,
326,411-418,425,431,434-438,
442,443,445,448,450,451,453
synaesthesia 8, 197, 199, 216-223,
225,230,231
unilateral neglect 271, 273, 277-
279