“Solving” ambiguity in the virtual space: communication strategies in a collaborative virtual environment

ORIGINAL ARTICLE

‘‘Solving’’ ambiguity in the virtual space: communicationstrategies in a collaborative virtual environment

Paolo Cottone Æ Luca Pieti Æ Valentina Schiavinato ÆDorian Soru Æ Massimiliano Martinelli ÆDiego Varotto Æ Giuseppe Mantovani

Received: 31 October 2006 / Accepted: 24 August 2007

� Springer-Verlag London Limited 2007

Abstract Collaborative virtual environments (CVE) face

the challenge of succeeding in incorporating critical

dimensions of cooperation and communication in everyday

working situations. One of these dimensions, situation

ambiguity, is scarcely considered in studies on CVE

although it can prove a key factor in future use of CVE in

real work situations. Many computer-supported coopera-

tive environments and telecommunication systems, like

those currently used in telemedicine, would benefit from

the incorporation of some degree of situation ambiguity

allowing users to deploy their diagnostic and interpretive

abilities. In the perspective adopted in this study, ambiguity

is the contingent outcome of the ongoing interaction taking

place between the environment and the interests of social

actors. The research focuses on the cooperation within

couples of participants facing situation ambiguity in a

virtual environment: a simulated city named Babylon.

Participants moved in the city through an avatar and

could communicate in one of the following conditions:

face-to-face, phone or chat. Their goal was that of meeting

somewhere in the city, in a place that they did not know

previously. Babylon contained elements designed to allow

both production and detection of ambiguity. Ambiguity

emerged when participants realized the presence of

inconsistencies in the way they perceived the situations

they had to face. The moments in which ambiguity was

perceived—called ‘‘critical events’’ (CE)—were mea-

sured and described through qualitative (ethnographically

oriented) methods. The different strategies that participants

used to ‘‘solve’’ ambiguity were characterized as: looking

for environmental cues, narrowing the focus of attention

and investing on cooperation. Both CEs and strategies were

analyzed with respect to the three communication condi-

tions: face-to-face, phone and chat. All the communication

conditions allowed the emergence of ambiguity and the

negotiation of strategies to solve ambiguity between part-

ners: according to literature, chat is very costly in terms of

time spent on writing but this disadvantage did not block

completely the emergence of ambiguity and the develop-

ment of adequate strategies of solution. All navigations but

three (on a total of 18 couples) succeeded: the partners did

meet in a short time (less than 15 min) relying on their

pragmatic resources in a new virtual place. Further research

is required to clarify the possible factors influencing the

choice of one strategy over the others, the order in which

strategies follow each other and the role of leadership in

ambiguity detection and solution.

Keywords Virtual reality � Cooperation � CMC �Negotiation � Ambiguity

1 Ambiguous situations and cooperation in a virtual

environment

1.1 Situated computing for collaborative workplaces

The challenge facing researchers on collaborative virtual

environments (CVE) is that of understanding ‘‘how to

design cooperative workplaces for meetings, collaboration

and communication that take into account technology,

physical spaces and the people that inhabit and work in

these spaces’’ (Bayon et al. 2006, p 194). Much work has

P. Cottone (&) � L. Pieti � V. Schiavinato � D. Soru �M. Martinelli � D. Varotto � G. Mantovani

Dipartimento di Psicologia Generale, University of Padua,

Padua, Italy

e-mail: [email protected]

123

Cogn Tech Work

DOI 10.1007/s10111-007-0105-9

been done on this subject, from the seminal study of

Benford et al. (1994) to the synthesis book by Churchill

and Snowdon (1998). Recent research on CVE has

emphasized the importance of contexts of use as part of a

general movement towards ‘‘situated computing’’ (Mills

2001; Streitz et al. 2001). This trend is influenced by the

emergence in cognitive science of the ‘‘situated knowl-

edge’’ approach both in everyday (Clancey 1997; Clark

1997; Suchman 1987) and work situations (Engestrom and

Middleton 1996; Suchman 1996, 1997). Most of the limi-

tations that collaborative work has met in computer-

mediated environments depend on the fact that while

computer-mediated presentation of separate objects or

scenes may be simple, providing effective computer sup-

port to human interaction may prove difficult. Studies on

real life work settings (Goodwin and Goodwin 1996;

Suchman 1996) show that people use objects in the envi-

ronment not only to achieve their tasks but also to

coordinate their activities. Negotiation and co-reference are

processes largely responsible for this coordination: ‘‘many

activities within co-located working environments rely

upon the participants talking with each other, and moni-

toring each others’ conduct, whilst looking, both alone and

together, at some workplace artefact. An essential part of

this process is the individual’s ability to refer to particular

objects, and have another see in a particular way what they

themselves are looking at’’ (Hindmarsh et al. 2000).

1.2 Distributed knowledge, communication

and cooperation

Work is now better understood as an activity deeply

embedded in specific social and cultural contexts (Hutchins

1995; Hutchins and Palen 1997; Mantovani 1996b; Nardi

1996) rather than as a private achievement. In real work

situations people use knowledge distributed among humans

and artifacts (technologies) to achieve their goals and are

guided by a rich network of explicit and implicit cultural

norms which establish what has to be done and how it is to

be done (Bødker 1996). Work—and learning in general—is

considered as an activity normally happening within

‘‘communities of practices’’ (Wenger 1998). Knowledge

and expertise are embodied in practice and pass from one

person to another through highly structured and finely

tuned social networks. Communication and negotiation are

the basic instruments supporting working and learning as

social activities (it is notable that working and learning are

strictly intertwined in ‘‘communities of practices’’). From

this perspective, communication processes are critical to

achieve collaborative understanding of the changing and

ambiguous situations that people face in their everyday

activities.

1.3 Everyday situations: often unpredictable,

sometimes ambiguous

In everyday life people have to co-ordinate their actions and

intentions while facing situations which present often a cer-

tain degree of unpredictability and sometimes aspects of

ambiguity. Unpredictability is usually ascribed to change: the

passing of time brings to existence new and unexpected

aspects of situations, aspects that were sometimes hardly

visible in advance. Unpredictability is the result of a change

in the scenario bringing to the foreground aspects of the

situations that were not detected before. Ambiguity is dif-

ferent: it is the product not of novelty but of the fact that

actors can find situations not fully structured, confusing, and

somehow bewildering. Situation ambiguity as considered in

this study covers a broader space than the one covered by

ambiguity in Gaver et al. (2003). Gaver distinguishes three

classes of ambiguity according to the place in which uncer-

tainty is located along the relationship between person and

artefact: ambiguity of information, which finds its source in

the artefact itself; ambiguity of context, which is rooted in the

cultural frameworks used to interpret it; ambiguity of rela-

tionship, which appears in the evaluative stance of the social

actor. We consider situation ambiguity as a synthesis of the

three classes, since situation ambiguity stems from the fact

that situations are not objects out-there-in-the-world but

emerge as the contingent outcome of the ongoing interaction

between the environment (with its affordances; Gibson 1979)

and the social actors who move within them (with their

interests and goals; Clancey 1993). ‘‘The interaction is so

close that the actor-in-situation is defined precisely by the

way in which she exploits opportunities in the environment.

Conversely, the situation does not exist before the actor

enters, but is constructed by her intervention’’ (Mantovani

1996a, p 17). Because of its situated character (bound to the

ongoing actor–environment interaction) ambiguity presents a

strong challenge to communication and cooperation.

1.4 Pragmatic agreements to cope with ambiguous

situations

Cooperation among people in everyday work situations is

more a problem of pragmatic alignment than the result of

previous cognitive agreements (Goodwin 1994; Goodwin

and Goodwin 1996; Heath and Luff 1992, 1996; Hutchins

and Palen 1997). In real life the problem that people face

when they have to co-operate is that of building a common

ground on which to refer their mutual actions. This is why

communication intended to foster cooperation is focused on

making shared sense of current situations rather than on

transferring pure information. People try to find some

pragmatic agreement on the meaning of the situations they

Cogn Tech Work

123

face in order to start working together on a common project

(Wenger 1998). In everyday life pragmatic agreement is

achieved through a large array of communication devices

designed to clarify ambiguous situations: gestures (Hanks

1990; Hindmarsh et al. 2000), postures, and language as

social action (Duranti 1994, 1997). In virtual environments

ambiguity is difficult to produce, because of the constraints

of the presentation of the environmental characteristics in

the visual space: virtual environments are usually so strictly

structured and oversimplified that little space remains for

alternative constructions open to the different interests of

the social actors navigating them (Cottone and Mantovani

2003). In virtual environments the resources available to

participants to ‘‘solve’’ ambiguity by reaching pragmatic

alignment (gesture, posture and speech) are also reduced.

1.5 Strategies to ‘‘solve’’ ambiguity in a virtual

environment

The research had two purposes. The first one was that of

producing an environment sufficiently complex (both

structurally and functionally) and interactive to support not

only easy co-operation between two partners co-present

(through avatars) in the environment but also to allow the

emergence of ambiguous situations. This has been done by

building a virtual city, Babylon, whose characteristics are

described below. Couples of partners (see below for

description of participants) received the task to meet

somewhere in the city. Ambiguity was produced by the fact

that Babylon not only was an environment totally unknown

to participants but also contained elements designed to

deceive the partners (see below for description of Babylon).

The second purpose was that of examining the strategies

that partners built to face ambiguity. We expected that the

relational nature of ambiguity (Cottone and Mantovani

2003; Mantovani 1996b), stemming from the match of

environmental opportunities and the actors’ interests, would

be met basically in three ways: (a) addressing ambiguity as

a characteristic of the environment (see below, ‘‘Looking

for environmental cues’’); (b) addressing ambiguity as a

characteristic of the participant (‘‘Narrowing the focus of

attention’’); (c) considering ambiguity as a relational

product, born at the encounter between actors and their

environments; this approach to ambiguity considers the

other participant as a possible resource to solve the problem

(‘‘Investing on cooperation’’).

1.6 The effects of different communication conditions

A problem in developing usable CVE is that of providing

users with a consistent environment for communication. In

the virtual space, participants normally have limited access

to their mutual position, orientation and operations; this is

why talk plays an even more critical role in CVE than when

compared to everyday situations: in CVE ‘‘almost all actions

must be revealed through talk’’ (Hindmarsh et al. 2000, p

493). To investigate these effects we allowed participants,

while moving in the city through an avatar, to communicate

in one of the following conditions: face-to-face (F: using

verbal and non verbal communication), phone (P: using a

normal telephone line) or chat (C: using an electronic text

message system). Our first hypothesis, according to past

research (Koleva et al. 2000; Kuzouka et al. 2004; Kirk et al.

2005), was that talk would support smooth interaction in all

the three conditions, with the obvious limitation of a wider

time span spent in the chat condition due to the fact that

writing is usually an activity more time-consuming than

talking. Our second hypothesis was that both voice condi-

tions would support mutual interaction with equal efficacy

(we assumed that the ease of talk exchanges supported by

each condition would be responsible for the expected dif-

ferences in performance and in strategic cooperation

between participants).

2 The experiment

2.1 The virtual environment: Babylon

The study was realized in a virtual environment specially

designed for the research. The name of the environment is

‘‘Babylon’’, from the city in which was built the tower that,

according to the Bible, was abandoned because of the

troubles caused by the spread of different languages among

the people working at its construction. Our virtual Babylon

has been built to explore the ways in which people can

overcome differences in perspectives and cognitive

resources while coping with ambiguous situations. Our

Babylon, unlike the biblical one, was designed not as a

place for quarrels but as a scenario for cooperation in

everyday situations. Babylon simulates part of an imagi-

nary Western city, with buildings that have period or

modern facades, shops and other cues normally present in

modern cities. The textures used to create buildings, shops

and parks have been often replicated, letting only little

differences in place such as a fountain, the number of trees

in a park, and so on to mark the differences. This was done

to produce further ambiguity in the environment and make

participants’ reciprocal orientation more difficult.

Babylon is a non-immersive (presented with a 24 in. CRT

display) virtual environment created through software 3D

Studio Max e Virtools Dev.3. The shops and building signs

were created using 3D Studio Max. The virtual space was

represented in the experiment with a resolutions of

Cogn Tech Work

123

1,024 · 768 pixel. Babylon can also be used for simulations

in immersive conditions (presented with a head-mounted

display, HMD); we choose not to use immersion because of

the heavy costs of enduring immersion for participants. Past

researches (Gamberini et al. 2003) showed that a high per-

centage of participants was not able to complete an

immersive session lasting longer than 10 min (Fig. 1).

2.2 Participants and conditions

Forty-two students of the Faculty of Psychology of the

University of Padua, 21 males and 21 females, ranging

from 19 to 32 years were distributed as follows: three

couples (1 female–female, 1 male–male, 1 female–male;

total 6 participants) for the preliminary session; 18 couples

(6 male–male, 6 female–female, 6 female–male; total 36

participants) for the experiment. The small scale of the

experiment sample is due both to the explorative character

of the study and to the qualitative methods adopted for the

analysis, mainly ethnographically oriented conversation

analysis. Couples were equally distributed in three condi-

tions: F face-to-face conversation (participants shared the

same physical environment and communicated face to

face); P phone conversation (participants were in physi-

cally separated places and communicated by phone);

C chat (participants were in physically separated places and

communicated through an electronic textual communica-

tion system available when navigating in the virtual

environment). Participants had no prior knowledge of their

partners. Participants were not paid to participate in the

experiment; the motive offered for participation was curi-

osity for the CVE. Before entering the simulation, all

participants were briefed on the aims of the experiment and

informed on the physical and psychological characteristics

of the virtual experience, to which they gave their written

informed consent. Participants declared that they were free

from physical or psychic impediments (especially neural or

vestibular illnesses) that could have made the participation

to the simulation unadvisable and signed permission for

video and audio recordings in be used only for scientific

purposes.

2.3 Avatars

In a virtual environment the ‘‘avatar’’ (this term comes

from the Hindu philosophy and designates an embodied

form of the Goddess) is the digital image of the person who

is navigating the environment. The ‘‘avatar’’ used in the

experiment (provided by the Virtools software we used)

was a tridimensional unisex puppet able to march in every

direction guided through the mouse (right or left) and the

keyboard (up and down arrows). Participants could see the

avatar of their partners when it was in view, but not their

own avatar because the perspective given on the environ-

ment was egocentric, allowing free body and head rotation

(360�). No participant had any trouble using her avatar

during the experiment.

2.4 Instructions and navigation

Participants were instructed to navigate freely within the

virtual city and meet their partner in a place of their choice.

Furthermore, the participants did not have a 2-D paper map of

the city. The study was not focused on the participants’ suc-

cess in achieving the goal of meeting their partner somewhere

but on the efficacy of their communication processes; as a

consequence, negotiation and cooperation, not performance

or speed, was the central focus of the simulation. Researchers

running the simulations did not interfere in any occasion on

the course of actions taken by participants, except when

explicitly requested by the participants. Participants were

moving (in fact, moving their avatars) in Babylon using a

keyboard and a mouse: pressing the key ‘‘arrow up’’ makes

the avatar march forward, the opposite happened when the

key ‘‘arrow down’’ was used; the mouse commanded head

rotations of the avatars on the x and z axes, allowing full

freedom of movement at 360� at a constant value on the y axis.

Fig. 1 Two views of Babylon

from the navigator’s point of

view

Cogn Tech Work

123

The participants started their navigation from points diamet-

rically opposite to the city. The specific starting location was a

little alley with a very restricted view and devoid of signifi-

cant reference points, which allowed maximum freedom to

choose a direction. The duration of the navigation was fixed at

15 min; the participants to the experiment were informed of

the assigned span of time.

2.5 Methods

Ethnographically oriented conversation analysis (Duranti

1994, 1997; Mantovani and Spagnolli 2003; Ochs and

Capp 2001; Zucchermaglio 2003) was applied to analyze

video and audio recordings. Phone and face-to-face con-

versations were audio-recorded and transcribed using

Jefferson’s notation system (1986; see Table 1). Computer-

mediated conversations (in the chat experimental condi-

tion) were recorded in a text file (IRC) and formatted to be

compared to the phone and face-to-face conversations. The

different settings for conversation are marked in the paper

by the following letters: F face-to-face conversation, P

phone conversation, C chat.

Video recordings were taken from a frontal point of view

to capture the participants’ visual expressions, facial mimics,

gestures, and postures. This information is necessary in this

research to understand the specific situation of the recorded

interaction under analysis (facial expressions and in general

non-verbal communication were not the aim of the study).

Recordings from the two video cameras running in the

experimental situation were saved on VHS supports. The

images coming from the ‘‘real’’ and the ‘‘virtual’’ environ-

ments were presented on one screen (Gamberini et al. 2003),

and in a second moment, were mixed and synchronized using

Adobe Premiere.

According to an established convention, in our analysis

P1 indicates the participant who in the mixed video

recordings appears in the upper box; P2 indicates the par-

ticipant in the lower box, and SP designates the researcher

running the experiment. For the actions that accompany the

navigation the following notation has been used: A ((nor-

mal)) indicates actions taking place in the virtual

environment while A ((italics)) indicates actions in the

‘‘real’’ environment (Fig. 2).

2.6 The trial and the experiment: from Babylon1

to Babylon

The environment used for the trials—Babylon 1—indi-

cated above had at its center a park with a little fountain

and a tiny pool. The environment used for the experi-

ments—Babylon—was more complex so as to increase the

participants’ engagement—three little parks and another

fountain were added. Special signs were prepared, nine for

the restaurants, eight for the shops, five for the bars, one for

a hotel, one for a cinema, one for a museum, one for a

library, one for a theater, one for a school, one for a syn-

agogue, one for a nondescript church, two for banks and

two for railway stations (Fig. 3).

Table 1 Gail Jefferson’s

notation system(.) Just noticeable pause

(.3), (2.6) Examples of timed pauses

:word,;word Onset of noticeable pitch rise or fall

A:

B:

word [word

[word

Square brackets aligned across adjacent lines denote the start of

overlapping talk

.hh, hh In-breath and out-breath, respectively

wo(h)rd (h) is an attempt to show that the word has ‘‘laughter’’ bubbling within

it

wor- A dash shows a sharp cut-off

wo:rd Colons show that the speaker has stretched the preceding sound

(words) A guess at what might have been said if unclear

( ) Unclear talk

A:

B:

word=

=word

The equals sign shows that there is no discernible pause between two

speakers’ turns or, if put between two sounds within a single

speaker’s turn, shows that they run together

word, WORD Underlined sounds are louder, capitals louder still

�word� Material between ‘‘degree signs’’ is quiet

[word word\\word word[

Inwards arrows show faster speech, outward slower

((sobbing)) Transcriber’s go at representing something hard, or impossible, to

write phonetically

Cogn Tech Work

123

3 Results

3.1 Critical events: when situation ambiguity

is perceived

3.1.1 The emergence of critical events

during the navigation

Participants had two intertwined problems to face—ori-

enting themselves in Babylon and establishing effective

communication with their partners—in order to achieve

their goal of meeting somewhere in the city. While each

one of the two tasks separately posed no special prob-

lems, their interference could create problems when the

information visually available (provided by the physical

features of the environment) appeared to collide with the

information socially available (provided by the conver-

sation with the partner). We will call ‘‘critical events’’

(CE) these situations of cognitive conflict which con-

fused the participants about their positions as well as the

course of action to be taken. CE is defined as the

moments in which the potential ambiguity inscribed in

the virtual environment was actually perceived by the

participants.

The following Fig. 4 coupled with excerpt 1 may help

to clarify as to how CE emerged during navigation.

Excerpt 1 refers to a phone conversation (condition P).

Participants’ utterances are marked as P1 and P2; A

means actions taken, both inside and outside the virtual

environment, by participants during their navigations.

Coupling the conversation with the image of Fig. 4 the

reader can understand what is happening: the partners in

couple U see a shop and they are able to locate them-

selves in relation to the shop they see; but they are

disappointed by the fact that the visual (the shop they see

in Fig. 4) and the social information (provided by the

phone channel with the other participant) they have are

inconsistent; they are not able to see each other in front

of the shop, while they expected the common reference

point could be helpful to their locating themselves

Fig. 2 The screen image

uniting the four video sources

(two ‘‘virtual’’ coming from the

two avatars and two ‘‘real’’

coming from the laboratory

settings). P1 indicates the

participant appearing in the

upper box, P2 indicates

the participant appearing in the

lower box. A ((normal)) means

an action happening in the

virtual environment, A((italics)) means an action taken

in the ‘‘real’’ environment

Fig. 3 The maps of Babylon 1

and of Babylon 2; letters in

Babylon 2 refer to different

elements: a fountain,

b restaurant, c shop, d bar,

e bank, f railway station,

g school, h nondescript church, isynagogue, l theater, m cinema,

n hotel, o library, p museum

Cogn Tech Work

123

reciprocally (actually—as often happens in everyday

situations—their expectations did not fit perfectly to the

situation because of the fact, as they discovered later,

that in the city there was more than one Tomyno shop).

Disappointment produced by the discovery of the

inconsistency of the information available in the situation

made the participants confused and meant that they had

to look for a ‘‘solution’’ to the emerging ‘‘problem’’.

excerpt 1 Couple U P

13. P1: si io sono davanti a un negozio di vestiti yes I’m facing a clothes shop

14. tipo like

15. P2: °hm: io non vedo niente° °hm: I don’t see anything°

16. P1: tu non hai Niente di negozi, di vie? don’t you get anything on shops, on streets?

17. P2: ecco adesso adesso adesso here now now now

18. P1: qui io[ here I

19. P2: [ ho un negozio dI vestit[i (.) I’ve found a clothes shop

20. tomi? (.) tomi?

21. tomi scritto?

tomi written?22. (.)

23. P1: TOmi si io sono li [(davanti) TOmi yes I’m there (before)

24. P2: [TOMYNO?TOMYNO?

25. A: ((P2 arrives at the Tomyno shop)) 26. P1: (.) Tomyno [si

Tomyno yes27. P1: eh sono anch’io li

eh me too I’m there28. sono proprio all’ingresso

I am just at the entrance29. A: ((P1 moves towards the door of the shop)) 30. P2: eh ma non ti ve:do

eh but I do not se:e you

eh I do not Know

31. P1: [eh n(h)on so (.1)

3.1.2 The emergence of CEs in oral and written talk

Critical event had been set as a central target for the analysis

because it was just in these moments that participants

became aware of the situation ambiguity and tried to

‘‘solve’’ the problem they had to face. CE was present in all

navigations; the total number of CE’s for the entire exper-

iment was 56. The average CE frequency for navigations in

the three conditions was 18.66 (ranging from 8 to 28); the

average CE frequency for condition F was 4.6 (ranging

from 4 to 6); the average CE frequency for condition P was

3.3 (ranging from 2 to 5); the average CE frequency for

condition C was 1.3 (ranging from 1 to 2). The more the

communication condition was ‘‘natural’’, the more CE

happened during the navigation. Talk, both in creating and

in expressing the perception of CE, is essential. Oral talk,

present in face-to-face and phone conditions, proved suit-

able to support smooth communication between partners

and produce a high number of CE’s. Written talk, used in

the chat condition, was more costly in terms of time; this is a

plausible explanation for the low level of CE appearing in

the chat condition. These results are consistent with the

perspective of the study by Clark and Brennan (1991) on

‘‘audibility’’ as an element facilitating negotiation among

participants in different ‘‘speech media’’. The total navi-

gation time recorded in the experiment was 4 h and 1 min.

The total time for CE episodes was 36.8 min (about 15.27%

of the total navigation time). The task of identifying CE was

entrusted to two judges whose accord was very high (94%).

Critical event was present, although with minor fre-

quency, also in written talk; CE appearing in written talk

was similar in structure and richness of expression to the CE

emerging in oral talk. Excerpt 2 presents a dialogue in the

chat condition, which is very similar to the one described in

excerpt 1 which happened in the phone condition.

Fig. 4 The virtual landscape which is referred to in excerpt 1 as the

occasion for emergence of a critical event (CE): the ‘‘Tomyno’’ cloth

shop that participant 1 sees is not the same ‘‘Tomyno’’ cloth shop that

participant 2 discovers, but the face-to-face conversation lets believe

in a first moment to both participants that they are facing the same

shop. The inconsistencies emerge during the navigation from the fact

that the visual information provided to a participant by her

exploration of the virtual landscape collides with the information

provided by the conversation with the partner. This situation is

prototypical of a ‘‘critical event’’ in which situation ambiguity is

perceived and actively dealt with

Cogn Tech Work

123

The excerpt shows how a CE happens in the chat con-

dition and how it is negotiated between the participants. At

the beginning of their navigation participants identify a

feature (the Tomyno shop) that they try to use to orient

themselves reciprocally (lines 11–24). In a second moment

(lines 40–49) a problem emerges: they are both in front of

the Tomyno shop but they do not see each other. They are

puzzled by the apparent inconsistency of the situation and

look for a solution (lines 70–83): choosing another refer-

ence feature, a fountain. The dialogue presented above is as

smooth and richly structured as the ones happening in the

face-to-face or phone conditions.

excerpt 2 Couple M C

11. P1: dove sei? where are you?

12. P2: ma in una via mmm in a street.

13. A: [((P1 moves forward)) 14. A: [((P2 moves forward too. She comes to a crossroads. She heads

towards a palace which is made in the same style as a Venetiaone))

15. P2: sono al teatro Piccini I am at the theater Puccini

16. A: ((P1 sees a Tomyno shop and goes to a window)) 17. A: ((P2 enter the street in front of the shop)) 18. P2: tu dove sei

where are you 19. P1: io difronte al negozio TOMYNO

in front of the Tomyno shop20. P2: come si chiama il negozio

what’s the shop’s name? 21. P1: TOMYNO

TOMYNO22. A: ((P2 heads towards the south border of the city)) 23. P2: ok

ok24. P1: ti aspetto qui

I’ll wait here for you....40. P2: sono da Tomyno ma non ti vedo come mai?

I am in front of a Tomyno but I don’t see you why?41. P1: sono qui

I’mhere42. A: ((P2 goes around the building again; then she stops)) 43. P1: sono qui.ma tu sei dentro?

I am here; are you inside? 44. P2: non sono dentro

I am not inside45. A: ((P2 backs)) 46. A: ((P1 turns on herself faces the street and starts walking)) 47. A: ((P2 goes again in face of Tomyno)) 48. P2: dentro dove?

inside where?49. P2: rispondimi

answer me…

70. P2: ci sono 2 negozi uguali stai li che cerco l’altro there are 2 shops that look the same; stay there and I look for the other

71. P2: ti chiederò istruzioni I’ll ask for directions from you

72. A: ((P1 stays in front of Tomyno)) 73. A: ((P2 starts running across the surroundings; she walks towards

the little park then she returns to Tomyno)) 74. P2: è pieno di questi negozi con lo stesso nome

it is full of these shops with the same name 75. P1: credo che ce ne siano più di uno

I think there’s more than one 76. P2: descrivimi più che puoi ciò che ti sta intorno

tell me as best you can about what’s around you 77. A: ((P1 starts running across the surroundings)) 78. A: ((P2 passes again in front of the theatre and goes straight)) 79. A: ((P1 sees a fountain)) 80. P1: c’è una fontana blu elettrico. ci troviamo lì

there is an electric blue fountain here; we meet there81. A: ((P2 looks around herself; on the right side of the park she

sees a fountain)) 82. A: ((P1 is standing in front of her fountain and looks around

herself))83. P2: ok

Cogn Tech Work

123

3.1.3 Time and negotiation in the three different

conditions of communication

The average navigation time for the three conditions was:

13.33 min (ranging from 9 to 15 min); average time for

condition F: 13 min (ranging from 9 to 15 min); average

time for condition P 13 min (ranging from 10 to 15 min);

average time for condition C 14,33 min (ranging from 11

to 15 min). The longer time requested by navigation and

communication in condition C (chat condition) can be

attributed to the time required to type on the keyboard and

to wait for an answer before starting a new turn. It could be

also due to the lack of multitasking in chat condition where

participants had to stop every other action to type on the

keyboard (Olson and Olson 2000). All the couples but three

succeeded in meeting in Babylon within the term (15 min)

fixed for the navigation. Of the three failures registered in

the whole simulation, two happened in the chat condition

and one in the face-to-face condition.

3.2 Responding to situation ambiguity: producing

strategies

3.2.1 Strategies to ‘‘solve’’ ambiguity revealed in CEs

The analysis of the interactions taking place within the couple

of participants during their navigation in Babylon shows three

main patterns of response (accord between the two judges

92%): (a) looking for environmental cues; (b) narrowing the

focus of attention; (c) investing in cooperation. Each of the

three strategies stresses one of the possible ways to under-

stand—and to cope with—ambiguity. According to our

model, ambiguity is a relational construction: it emerges

within the relationship between the actor and the environment

and can be conceptualized as a problem coming from ‘‘out-

side’’ (from the environment), or from ‘‘inside’’ (from the

participant’s cognitive processes), or from the relationship

between ‘‘outside’’ and ‘‘inside’’ (the relationship connecting

the navigator, her partner and the virtual city).

The three strategies found in CE reflect the three routes

available to social actors to make sense of the situation

ambiguity: the cause for bewilderment can be attributed to

the ‘‘outside’’ environment, to the ‘‘inside’’ processes, or to

the relationship between ‘‘outside’’ and ‘‘inside’’, that is

between the actor and her social and physical environment.

The three strategies are present in all conditions and in

every navigation. No fixed preference for a particular

strategy has been found between the couples of partici-

pants. No effect of conditions of communication between

partners (F, P, C) was found. Possible situational or per-

sonal features that could trigger the preference for a given

strategy will be the object of further investigation.

The first strategy, looking for environmental cues,

emphasizes the contribution that the environment makes to

the generation of ambiguity. Following this strategy, the

participant looks for further information coming from the

environment. The second strategy, narrowing attention

focus, looks for more accurate information seeking. The

third strategy, investing in cooperation, focuses on the

interaction and expects a ‘‘solution’’ from a more effective

communication within the local cognitive and pragmatic

system. Instances of the three strategies are presented

below.

3.2.2 Working on the outside source of ambiguity: looking

for environmental cues

The first strategy emphasizes the contribution of the envi-

ronment to the production of the experience of ambiguity.

Excerpts 2 and 3 show examples of communications

among participants in condition one (face-to-face, F)

intended to detect the environmental cues in order to solve

the problem of identifying their mutual positions.

Participants are positioned at their starting points. They

are doing their first steps in the lane. They do not have any

sign or landscape reference to guide them. Participant

1 asks the researcher for some information about the

dimensions of the virtual city. When she does not obtain

the information she starts her navigation looking for visual

elements that could offer relevant cues. Elements in the

Excerpt 3

Couple G F

27. P1: ci mettiamo una vita [a trovarci it will take a lifetime for us to meet

28. P2: [no infa::tti no, really

29. A: ((P2 turns towards the experimenter)) 30. P2: ma quanto è gr ande <scusa>. sto posto?

but how big is <excuse me> this place?

31. tanto? very?

33. A: ((P2 sees a palace in Venetian Style 34. with a yellow sign)) 35. P2: scoltami:: so:no a:l tea:tro gi pucci:ni

listen to me: I a:m a:t the thea:ter gi puccini36. P1: sì: va bè bellissimo

yes; it’s ok very fine37. A: [((P1 laughs)) 38. P2: [bello no?

beautiful, isn’t it?39. A: ((P1 sees a fountain)) 40. P1: vabbè tu hai presente dov’è la fontana?

ok do you see where the fountain is?41. P2: no

no42. A: ((P2 looks around herself and then heads 43. on her right towards a park 44. that is visible in the distance)) 45. P1: e(h) trovat(h)ela(h)

e(h) I found (h) there (h)46. A: ((P1 laughs))

Cogn Tech Work

123

environment that at first sight seem to be shared (theater

Puccini, lines 35–38) and to offer cues useful to build a

common ground among the participants are in a second

moment recognized—on the basis of (phone) conversa-

tion—as problematic (the fountain is seen by participant 1

but not by participant 2, lines 40–41). Exploration contin-

ues still trying to use visual cues.

Visual cues alone are not sufficient to build a common

view of the city (Babylon is too complex to allow quick

and successful visual orientation), so participants try to

establish general reference criteria based on their previous

experience in the city such as the opposition centre-

periphery (that this couple found useful before) to guide

them in their navigations.

This strategy is present in 68 occurrences (30% of total

strategies) and is distributed as follow: 27 occurrences

(31%) in condition F, 22 occurrences (27%) in condition P;

19 occurrences (33%) in condition C. Percentages of the

occurrences of the first strategy in the various conditions

are: 40% F, 32% P, 28% C.

3.2.3 Working on the inside source of ambiguity:

narrowing the focus of attention

Participant can try to reduce the ambiguity that emerges in

the CE by looking for specific information—such as

peculiar shop signs—which could help them to achieve

mutual orientation. This strategy consists of narrowing the

attention focus on aspects of the environment that could

offer specially relevant information: in the following

excerpt we see that one participant asks her partner to go

‘‘hunting for’’ some signs. The partner agrees and the

couple adopts this explicitly negotiated strategy.

The above (phone) conversation takes place in a

moment in which the couple of participants has navigated

for some minutes and has acquired a partial knowledge of

the virtual city. They try to infer the position of the

partner they have to meet from features of the environ-

ment but they know that the environment has a complex

structure so they resort to the strategy of focusing their

attention on possibly relevant information present in

Babylon. Participant 1 discovers a bus station sign and

signals it to the partner, without immediate success with

respect to their common goal; participant 2 does not see

this sign. But participant 1 insists: she proposes to her

partner to keep ‘‘hunting signs’’. The focus of attention

is narrowed on a particular feature of the environment,

Excerpt 4 Couple G F

47. P2: scolta: io: boh comincio a: cammina:re:: hey: I: er start: wa:lki::ng

48. P1: sEnti che ne dici di farci il giro: liSTEn what do you think about doing

49. laterale? a lateral turn?

50. A: ((P1 continues to walk towards a little park)) 51. P2: della?

of?52. P1: hai presente che a un certo punto c’è il

are you aware that at a certain point there is the53. vuoto accanto dove finisce la città. se

empty place near to where the city ends. if54. noi ci giriamo il perimetro↓ è probabile

we walk around the edge I expect55. che ci troviamo.(non ti [va?)

we’ll meet (it does not suit it to you?)56. P2: [scolta:: in

listen: : in57. centro più che altro

the centre preferably58. P1: in centro? (.) bè occh[ei

in the centre? (.) well ok59. A: ((P2 continues walking 60. towards the park)) 61. P2: [c’è tipo: un

there is kind of: a62. parchetto: con tutti gli alberi:

little park: with lots of trees;63. P1: sì a fianco [gli alberi (.)

yes beside the trees (.)64. P2: [allora vai lì

then go there

Excerpt 5 Couple E F

41. P2: sei in mezzo ai p[alazzi anche t:u?are you in the middle of the palaces too?

42. A: ((P2 looks around herself)) 43. P1: [ehm: sono davanti

ehm: I’m in front44. esatto sono in mezzo a due palazzi[:

exactly I’m between two palaces45. P2: [(si

[( yes46. gira verso destra)

turn on the right)47. A: ((P2 stops)) 48. P1: [il primo ehm:=sche ho trovato sulla mia

the first ehm:=swhat I found on my49. sinistra c'è scritto apab autostazione

left; there’s a sign for a APAB bus station50. (.) 51. A: ((P2 takes up advancing)) 52. [((P1 takes up advancing)) 53. [((P2 stops)) 54. P1: [tu vedi qualcosa del genere [in giro?

do you see something like around?55. A: [((P2 takes up advancing again)) 56. P2: [c’è un palazzo:

there’s a palace57. sembra: sullo stile:

it seems: in the style:58. [di quelli a venezia hai presente?

of those in Venice do you know?59. P1: [mh mh (.)60. A: ((P1 looks around))61. P1: mh m[h 62. P2: [però: scritte non ne vedo da nessuna

but: I don’t see any sign63. parte

anywhere 64. A: ((P2 turns on her left)) 65. (.1) 66. ((P1 turns on the left and cuts 67. the corner of a palace)) 68. P1: vabbè andiamo in giro a caccia di scritte

okkey let’s go hunting signs 69. e vediamo se ci troviamo davanti alla

and see if we find ourselves in face of 70. stessa

the same

Cogn Tech Work

123

making the encounter of a common sign a possible

solution for the experienced ambiguity.

This strategy is present in 92 occurrences (41% of total

strategies) and is distributed as follow: 37 occurrences

(43%) in condition F, 32 occurrences (40%) in condition P;

23 occurrences (40%) in condition C. Percentages of the

occurrences of the second strategy in the various conditions

are: 40 F, 35 P, 25 C.

3.2.4 Working on the interaction as source of ambiguity:

investing on cooperation

A third strategy that emerges during the navigation within

Babylon is that of trying to build a common scenario. In

this case the two partners invest significant time and effort

in sharing information about the pieces of urban landscape

that each of them is currently seeing.

The strategy of trying to build a common ground among

participants, which is frequent in work situations, has been

extensively studied by Orr (1996), Wenger (1998), and

Wenger and Snyder (2000). In this strategy each of the two

members of the couple seems to invest in the other as a

resource for the solution of the problem at hand. For this

reason they are so much interested in the quality of mutual

understanding that they can even stop communicating about

their positions within Babylon to start talking about what

they know on ‘‘relief maps’’, as we see at lines 113–117. In

this case we see that a doubt on possible misunderstandings,

triggered by the question in lines 113–114, is met through

an explicit negotiation turn which reveals that in this

moment the need of being sure of what the other partner

thinks takes precedence over the need of knowing what she

is seeing in her part of the environment.

This strategy is present in 64 occurrences (29% of total

strategies) and is distributed as follows: 22 occurrences

(26%) in condition F, 27 occurrences (33%) in condition P;

15 occurrences (26%) in condition C. Percentages of the

occurrences of the third strategy in the various conditions

are: 35 F, 42 P, 23 C.

4 General discussion

The first observation coming from the analysis of the data

is that the response of participants to the virtual environ-

ment shows the emergence of specific patterns within the

three communication conditions provided by the simula-

tion. This from one hand confirms the presence of

(perceived) situation ambiguity, and from the other hand

the importance of conditions of communication between

partners. The effect of ambiguity on the navigation speed

and generally on participants’ performance is outside the

scope of the present study because a reliable performance

metric on navigation within this sort of complex virtual

environment is hardly feasible, and scarcely meaningful.

Data on the navigation time can be referred to the problem

solving and to the negotiation process evoked by the

appearance of ambiguity. In the chat condition, due to the

greater time costs of written communication, navigation

times are slightly longer and CEs are significantly longer

than in the two other conditions. This fact, along with the

higher rate of failure in achieving the goal of meeting in the

virtual city (reported above, in ‘‘Results’’), shows that

negotiating is more difficult using written rather than oral

talk. Nevertheless, it is worth noting that, in spite of the

considerable time costs of writing messages versus

speaking to the partner, chat can allow the emergence of

CEs in a fully structured form, as was shown in the above

excerpt 2. Navigation times (total navigation time minus

time for strategies and time of CEs) in the three conditions

are presented in Fig. 5.

Analysis of the data on the appearance of strategies

shows the presence of a strong (with only one exception)

pattern of response in the three conditions of communica-

tion considered. Figure 6 presents the occurrences of

strategies in the three conditions.

Excerpt 6Couple E F

87. P1: h↑o i palazzi sulla destra h↓o i palazzi I got the palaces on the right I got the palaces

88. sulla destra e:: sì che hanno un po’ lo on the right e:: yes they are a little

89. stampo dei palazzi di veneziani like the Venetian ones

90. effettivamente e su:l-è percorro un-u:n really and o:n is I’m walking on a

91. marciapiede e abbiamo un marciapiede ma non sidewalk and we have a sidewalk but

92. c’è nulla sulla mia sinistr[a there’s nothing on my left

93. P2: [a(h) ho capito a(h) I understand

94. vedi il nulla? sei-praticamente solo il do you see anything? you are-practically only the

95. cielo? (.) no sky? (.) no

96. P1: m:no (.) ti spiego (.) è come se fosse u:n m.no (.) I explain it to you (.) it is like a:

97. (.1) 98. e:m una specie di plastico e io sto

e:m sort of relief map and I am99. camminando sul bordo di questo plastico

walking on the edge of this relief map...109. v↑abbè adesso entro

okey now I get in110. all’interno del plastico perché adesso

into the relief map because now111. sono-sono proprio sul confi[ne

I am I am exactly on the edge...113. P2: [all’interno

into114. del plastico?

the relief map?115. P1: ma [sai cos’è un plastico tu?

do you know what a relief map is 116. P2: [°sì: ho una pallida idea ma penso-°

yes I have a faint idea but I think117. P1: è la riproduzione in piccola scala di[::::

yes it is the reproduction on a smaller scale of

Cogn Tech Work

123

The second strategy is the most preferred in every

condition, followed by the first strategy in the other con-

ditions, with the notable exception of the third strategy

being the second preferred in the phone condition. Further

research is required to clarify possible origins of this

preference in the phone condition.

5 Conclusion: ambiguity, CEs, strategies in their social

context

This study explores an area of cooperation in virtual

environments that has been until now scarcely considered

in spite of its considerable social and scientific relevance.

The question addressed is: how do people cope with

ambiguous situation, which are often present in working

and living situations? To simulate cooperation in ambigu-

ous situations in a virtual environment we built a city,

Babylon, which proved fit to allow joint explorations of

pairs of participants that—in all the cases except three—

did achieve their goal of meeting somewhere in a city that

they did not know in advance.

The study of ambiguity focused on CEs analysed

through qualitative methods (ethnographically oriented

conversation analysis) of video and audio recordings of

both the navigation within Babylon and the cooperation

between the partners involved. The pragmatic resources

used by participants have also been investigated. Three

different strategies, consistent with the three components

involved in the construction of (perceived) situation

ambiguity, were identified: looking for further environ-

mental cues, narrowing the focus of attention, investing on

cooperation with the partner.

Further investigation remains to be done in three

directions. The first direction is clarifying how strategies

are selected and tried by participants, and how participants

shift with apparent ease from a strategy to another in their

approach to situation ambiguity. The second direction is

focusing on the ways in which leadership is taken and

recognized within the couples during the navigation; how

pragmatic alignment is achieved between people who do

not know each other in advance? The third direction is on

gender and leadership in choice of strategies: how is gen-

der—both ‘‘real’’ and simulated through the appearance of

the avatar—involved in cooperation face to ambiguous

situations?

References

Bayon V, Griffiths G, Wilson JR (2006) Multiple decoupled

interaction: an interaction design approach for groupware

interaction in co-located virtual environments. Int J Hum

Comput Stud 64(3):192–206

Benford S, Bowers J, Fahlen L, Mariani J, Rodden T (1994)

Supporting cooperative work in virtual environments. Comput J

38(4):653–668

Bødker S (1996) Creating conditions for participation: conflicts and

resources in systems development. Hum Comput Interact

11(3):215

Churchill EF, Snowdon D (1998) Collaborative virtual environments:

an introductory review of issues and systems. Virtual Real

3(1):3–15

Clancey WJ (1993) Situated action. A neuropsychological interpre-

tation—response to vera and simon. Cognit Sci 17:87–116

Clancey WJ (1997) Situated cognition: on human knowledge and

computer representations. Cambridge University Press, New

York

Clark A (1997) Being there. The MIT Press, Cambridge

Clark HH, Brennan SE (1991) Grounding in communication. In:

Resnick LB, Levine JM, Teasley SD (eds) Perspectives on

socially shared cognition. APA, Washington, pp 127–149

Cottone P, Mantovani G (2003) Grounding ‘‘subjective views’’.

Situation awareness and co-reference in distance learning. In:

Riva G, Davide F, Ijssenterijn WA (eds) Being there: concepts,

effects and measurement of user presence in synthetic environ-

ments. IOS Press, Amsterdam, pp 250–260

Duranti A (1994) From grammar to politics. University of California

Press, Berkeley

Duranti A (1997) Linguistic anthropology. Cambridge University

Press, Cambridge

Engestrom Y, Middleton D (eds) (1996) Cognition and communica-

tion at work. Cambridge University Press, Cambridge

Gamberini L, Cottone P, Spagnolli A, Varotto D, Mantovani G (2003)

Responding to a fire emergency in a virtual environment:

different patterns of action for different situations. Ergonomics

46(8):842–858

Fig. 6 Occurrences of strategies across the different communication

conditions

Fig. 5 Total navigation time, time spent in strategies, and time of

CEs in the three conditions

Cogn Tech Work

123

Gaver WW, Beaver J, Benford S (2003) Ambiguity as a resource for

design. Paper presented at the CHI 2003, Ft Lauderdale, FL

Gibson JJ (1979) The ecological approach to visual perception.

Erlbaum, Hillsdale

Goodwin C (1994) Professional vision. Am Anthropol 96(3):606–633

Goodwin C, Goodwin MH (1996) Seeing as situated activity:

formulating planes. In: Engestrom Y, Middleton D (eds)

Cognition and communication at work. Cambridge University

Press, Cambridge, pp 61–96

Hanks WF (1990) Referential practice. Language and lived space

among the Maya. University of Chicago Press, Chicago

Heath C, Luff P (1992) Collaboration and control: crisis management

and multimedia technology in London underground line control

rooms. Comput Support Coop Work 1:69–94

Heath C, Luff P (1996) Convergent activities: line control and

passenger information on the London underground. In: Enge-

strom Y, Middleton D (eds) Cognition and communication at

work. Cambridge University Press, Cambridge, pp 96–129

Hindmarsh J, Fraser M, Heath C, Benford S, Greenhalgh C (2000)

Object-focused interaction in collaborative virtual environments.

ACM Trans Comput Hum Interact 7(4):477–509

Hutchins E (1995) Cognition in the wild. The MIT Press, Cambridge

Hutchins E, Palen P (1997) Constructing meaning from space,

gesture, and speech. In: Resnick LB, Saljo R, Pontecorvo C,

Burge B (eds) Discourse, tools, and reasoning. Springer, Berlin,

pp 23–40

Jefferson G (1986) Notes on ‘‘latency’’ in overlap onset. Hum Stud

9:153–183

Kirk D, Crabtree A, Rodden T (2005) Ways of the hands. Paper

presented at the ECSCW’05, Paris, France

Koleva B, Schnadelbach H, Benford S, Greenhalgh C (2000)

Traversable interfaces between real and virtual worlds. Paper

presented at the CHI 2000. The Hague, Amsterdam

Kuzouka H, Kosaka J, Yamazaki K, Suga Y, Yamazaky A, Luff P

et al (2004) Mediating dual ecologies. Paper presented at the

CSCW’04, Chicago

Mantovani G (1996a) New communication environments. From

everyday to virtual. Taylor & Francis, London

Mantovani G (1996b) Social context in HCI: a new framework for

mental models, cooperation and communication. Cognit Sci

20:237–269

Mantovani G, Spagnolli A (2003) Metodi qualitativi in psicologia. il

Mulino, Bologna

Mills KL (2001) Situated computing: the next frontier for HCI

research. In: Carroll J (ed) Human–computer interaction in the

new millennium. Addison–Wesley, Reading, pp 535–550

Nardi B (1996) Context and consciousness. Activity theory and

human–computer interaction. MIT Press, Cambridge

Ochs E, Capp L (2001) Living narrative. Creating lives in everyday

storytelling. Harvard University Press, Cambridge

Olson GM, Olson JS (2000) Distance matters. Hum Comput Interact

15:139–178

Orr J (1996) Talking about machines: an ethnography of a modern

job. ILR Press, Ithaca

Streitz NA, Tandler P, Muller-Tomfelde C, Konomi S (2001)

Roomware: towards the next generation of human–computer

interaction based on an integrated design of real and virtual

worlds. In: Carroll J (ed) Human–computer interaction in the

new millennium. Addison–Wesley, Reading, pp 553–578

Suchman L (1987) Plans and situated actions: the problem of human–

machine communication. Cambridge University Press, New

York

Suchman L (1996) Constituting shared workplaces. In: Engestrom Y,

Middleton D (eds) Cognition and communication at work.

Cambridge University Press, Cambridge, pp 35–60

Suchman L (1997) Centers of coordination: a case and some themes.

In: Resnick LB, Saljo R, Pontecorvo C, Burge B (eds) Discourse,

tools, and reasoning. Springer, Berlin, pp 41–62

Wenger E (1998) Communities of practice: learning, meaning, and

identity. Cambridge University Press, New York

Wenger E, Snyder WM (2000) Communities of practice: the

organizational frontier. Harv Bus Rev 61(1):139–145

Zucchermaglio C (2003) Contesti di vita quotidiana, interazione e

discorso. In: Mantovani G, Spagnolli A (eds) Metodi qualitativi

in psicologia. Il Mulino, Bologna, pp 47–72

Cogn Tech Work

123