Keywords

1 Introduction

In recent decades communication technologies have developed and we can communicate with our friends located far away with portable terminals easily. These technologies enrich our lives and become essential for us today. On the other hand, these communication services are based on virtual communication, that is, one of the non-face-to-face communications, then, information which we can convey is fewer than face-to-face communications. Such situations generate misunderstandings, flaming. What is worse, individuals are blamed and commits suicide. If all persons who communicate in virtual situations have the same way of thinking, the above problems might not occur, but the persons must have various insides. Therefore, solving the above problems is difficult and urgently necessary. Also, these problems can be said that the failure of consensus-building between persons who communicate each other. Consensus-building means that understanding or conjecturing intentions of opponents from their behaviors and heading to the same direction, so, building a consensus smoothly in the situations in which persons are not allowed to convey rich information is a clue of solving the problems.

Then, what information can we communicate in virtual situations? There are two types of information, that is, verbal information which is represented as text and utterance, and non-verbal information. Verbal information is used in now virtual communication, and it is effective to announce pre-organized information. However, it is difficult to describe one’s inside which is very quantitative and to convey the inside to opponents rightly. In contrast, non-verbal information has been utilized since earlier than verbal information, so non-verbal information is expected to be able to convey one’s inside intuitively. In fact, according to [1], non-verbal information builds impressions that affect behaviors toward opponents. From the above statements, it is assumed that it is important to convey ways of thinking or intentions to opponents with non-verbal information for promoting consensus-building in virtual communications.

The human has many channels that perceive non-verbal information. For example, distances between opponents, eyes and so on. In addition, especially, emotion is one of the primitive channels and is proved to relate with motivations of behavior and good relationships with opponents [2, 3]. In other words, it can be possible to motivate behaviors toward opponents and build a consensus between the opponents by managing emotions. Occasionally, we cannot convey what we think to opponents by verbal information. In such a situation, emotions become useful ways to express ourselves and build good relationships with opponents.

This paper focuses on the role of emotions in virtual communications and investigates the usage of emotions for smooth consensus-building. Concretely, we employ Barnga [4] which is one of the cross-cultural games as virtual environments. In Barnga, players cannot convey verbal information and must build consensus to progress games. For experiments, we employ online Barnga in which the players can recognize only actions of the other players. Also, we introduce emotional panels to the online Barnga to enable to emote players’ emotions, and the emotional panels have four variations: happiness, anger, sadness, and surprise. From the experimental results, we analyze the emotional panels which lead to smooth consensus-building.

This paper is organized as follows. Section 2 explains how to play Barnga, and Sect. 3 introduces the emotional panel which is an additional tool of Barnga. Section 4 develops on-line Barnga for a non-face-to-face experiment. Section 5 conducts experiments, and Sect. 6 discusses experimental results. Finally, Sect. 7 concludes this paper.

2 Barnga: Cross-Cultural Game

Barnga [4], developed by Thiagarajan, is a card game, and purposes to experience communications with people who have another cultural background artificially. Concretely, in Barnga, players play a card and decide the winner (this cycle is called Game in Barnga) based on different rules each other, e.g., cards with Spade is stronger than other suits for one player, but another player’s strongest suit is Heart and so on. In addition, players are not allowed to use language.

2.1 Game in Barnga

As we described, the cycle from playing cards to deciding the winner is called Game in Barnga. In addition, the number of winner selections is called Turn in this paper. We show the procedure of Game in Barnga as follows.

  1. 1.

    Game starts.

  2. 2.

    Each player play a card from their hand to a table.

  3. 3.

    Players select the winner whose played card is the strongest in the cards on the table, and the number of Turn is added by 1.

  4. 4.

    If the selections are matched, move to the procedure 5. If not, back to the procedure 3.

  5. 5.

    Game ends.

Figure 1 shows a flowchart of Game of Barnga. In the left figure, players play cards from their hands, i.e., Player 1 plays 2 of DIAMOND, Player 2 plays 7 of DIAMOND, Player 3 plays 2 of SPADE, and Player 4 plays 4 of CRAB. In the right figure, on the other hand, players select the winner of Game, and colors of pictogram in balloons denote the kinds of players, that is, orange means Player 2, yellow means Player 3, blue means Player 1. In this case, the players select the winner again because winner selections do not match.

Fig. 1.
figure 1

Flowchart of Game in Barnga (Color figure online)

Full size image

2.2 How to Play Barnga

When Barnga starts, a dealer distributes seven cards to each player, then, Game starts. After procedure 5 in Game, the winner pile off the cards on the table and next Game starts from the winner. If the players have no hands, i.e., the players play seven Games, the dealer deals seven cards to each player again.

3 Emotional Panel

Players can claim their thoughts and differences of rules by emoting or eyesight in normal Barnga, but, in virtual communication, the players cannot look at faces of the other players, so they can only express with alternative tools, such as Emoji and Stump. Then, we introduce an emotional panel that allows the players to express their inside. The emotional panel has four variations: happiness, anger, sadness, and surprise. According to Ekman [5], the human can distinguish some emotions: happiness, anger, sadness, surprised, disgust and so on, regardless of cultural background. In order to introduce the emotional panel into Barnga, we should not break the premise of Barnga. In short, the emotional panel should be distinguishable for all players, so we introduce the emotional panel with the above four variations of the emotion. However, the players cannot always express their inside such as anger and sadness with the emotional panel. Therefore, we take questionnaires when the end of experiments and analyze the intention, usages, and timing of the emotional panel. By comparing the analysis with the actual players’ actions, we analyze the effects of the emotional panel with validity.

Figure 2 shows the emotional panels used in experiment. In Fig. 2, the upper right panel is happy, the upper left is anger, the lower left is sadness, and the lower right is surprise.

Fig. 2.
figure 2

Emotional panels

Full size image

In Barnga, players can use the emotional panels after selecting the winner as an additional procedure. Due to an additional procedure in which the players express their facial expressions, the flowchart of how to play Game in Barnga is extended as Fig. 3. The players express their emotions after selecting the winner. After that, if the selections of the winner are matched, Game ends and next Game starts according to the procedures which are described in Sect. 2.1.

Fig. 3.
figure 3

Flowchart of Game in Barnga with the emotional panel

Full size image

4 Barnga as Virtual Communication

In order to play Barnga in a non-face-to-face condition, we implement the online-Barnga. In this simulation, players are able to play Barnga through displays without facing to the other players.

Figure 4 shows the screen of simulation after the players finish to play their cards. In this screen, images in the four corners represent the players, the upper left is Player 1, the upper right is Player 2, the lower right is Player 3, and the lower left is Player 4. In addition, the red box represents the hand cards of the player. The player can play a card by clicking an image of a card displayed here. The yellow box denotes the emotional panel which the players can select.

Figure 5 shows the screen of simulation after selecting the winner and the emotional panel. The images of the hand point a player who is selected as the winner by each player. In Fig. 5, Player 1, 2, and 4 selected Player 3, and Player 3 selected Player 2 as the winner, so the winner of the Game is not determined. In addition, Player 1 selects the happiness panels, Player 3 selects the anger panel, and Player 4 selects the surprise panel. If the winner of the Game is determined, images of the field cards, the hands, and the emotional panels are hidden, then, the players play a card from the winner. If not, images of only the hands and the emotional panels are hidden, then, the players select the winner again.

Fig. 4.
figure 4

Screen of online-Barnga (after playing cards)

Full size image
Fig. 5.
figure 5

Screen of online-Barnga (after selecting the winner and the emotional panel)

Full size image

5 Experiment 1

We conduct an experiment that employs the online Barnga. This experiment purposes to investigate important elements for consensus-building in virtual communications by analyzing emotional expressions of players just before consensus-building. A point of analysis is the number of using emotional panels before determining the winner of the game.

5.1 Experimental Settings

The number of players is four and the players are taught rules which are shown in Table 1 respectively. They play 28 Games of Barnga. Figure 6 shows scenery of this experiment. The players sit back to back, so they cannot look at the face each other. Because of the rule of Barnga, the players are not allowed to communicate with language, that is, they should imagine the thoughts of the other players from selections and the emotional panels on a display. Under the above condition, we reproduce a virtual situation where information about another person is not informed directly.

Table 1. The rules of the players
Full size table
Fig. 6.
figure 6

Overview of the experiment

Full size image

5.2 Experimental Results

Figure 7 shows the number of emotional panels used just before the winner of the Game in Barnga is determined. In Fig. 7, the vertical axis denotes the number of the emotional panels, the most left bar denotes the happiness panel, the middle left bar denotes anger, the middle right denotes sadness and the most right bar denotes surprise. From the result, the number of the anger panel is the largest and the happiness panel is the smallest. Thus, it is assumed that the anger panel is more effective than other panels, especially the happiness panel, to build consensuses in virtual communication.

Fig. 7.
figure 7

The number of the emotional panels used just before the winner is determined

Full size image

6 Experiment 2

This experiment purposes to confirm the effects of the anger panel by comparing this experiment in which players cannot use the anger panel with Experiment 1. Concretely, we analyze the number of turns that is spent until completing 28 Games.

6.1 Experimental Settings

The number of players is four and the players are taught rules which are shown in Table 1 respectively. They play 28 Games of the online Barnga, but the players are not allowed to use the anger panel in this experiment.

6.2 Experimental Result

Figure 8 shows the number of turns which is spent until completing 28 Games, and the vertical axis denotes the number of turns, the left bar denotes the experiments in which players can use the angry panel, that is Experiment 1, and the right bar denotes this experiment in which the players cannot use the anger panel. In Fig. 8, the smaller the number of turns is, the smoother consensuses are built when conflicts happen. From this result, the number of turns in the experiment with the anger panel is smaller than this experiment without the anger panel, thus, the effects of the anger panel for building consensus are confirmed.

7 Discussion

7.1 Details of the Effect of the Anger Panel

We analyze the effects of the anger panel for consensus-building which is confirmed in Experiment 2. The points of analysis are (1) timings of using the emotional panel and (2) intentions of the players’ usage of the anger panel. In point (1), we analyze relationships between the number of the emotional panel used just before consensus-building and turns until consensuses are built.

Figure 9 shows a relationship between the number of used emotional panels just before consensus buildings and a turn when a consensus is built. In Fig. 9, the vertical axis denotes the number of used emotional panels, the yellow line denotes the happiness panel, the red line denotes the anger panel, the blue line denotes the sadness panel and the green line denotes the surprise panel, and the horizontal axis denotes the number of turns when consensuses are built. From this result, the anger panel is used just before consensus building which is reached at fast turn. This result illustrates that the angry panel is effective to build consensuses in short turns and supports the result of Experiment 2.

Fig. 8.
figure 8

The number of turn by the end of 28 Games in the non-face-to-face experiments (left: with all variations, right: without the anger panel)

Full size image
Fig. 9.
figure 9

Relationships between the number of the emotional panels and the turn

Full size image

To confirm the factors of early consensus-building from the anger panel, we took a questionnaire about the usage of the anger panel. Table 2 shows the answer to the questionnaire from each player. It should be noted here is the usage of Player 3. Player 3 used the anger panel toward players who do not grasp the rules shared among the players. In fact, 75% of the anger panel used by Player 3 was used when only one player selected a different winner from the other players, and the winner is determined in the next turn, that is, a consensus was built. Also, Player 1 and Player 4 used the anger panel toward consensus-building, so consensus-building in Experiment 1 is achieved smoother than the Experiment 2.

Table 2. Answers about the usage of the anger panel
Full size table

7.2 Impressions of Anger Expressions in Virtual Communications

From the answer of the questionnaire from the players, Player 3 used the anger panel to announce shared rules for a player who did not understand the rules. In addition, it is expected that the anger panel is effective to build a consensus in short turns. However, is it true that consensuses are built by the anger panel? in generally, anger expressions are not needed for early consensus building because the expressions can disrupt and stop communications, so it is possible that the consensuses was built as a result of forcing the consensuses by the anger panel. Then, we took a questionnaire about favorability of players. In the questionnaire, a player scores the favorability of the other players in 5 levels from 1 to 5. Also, we discuss if the consensuses are built with agreeing the winner selections or not. If the anger panel is recognized as selfish behaviors or forcing to agree, players who use the anger panel are scored a low favorability. If not, the players are scored a high favorability.

Figure 10 shows the number of the anger panel used by the players. In this figure, the vertical axis denotes the number of the anger panel, the most left bar denotes a result of Player 1, the middle left denotes Player 2, the middle right denotes Player 3 and the most right denotes Player 4 respectively. From the results, Player 3 used the largest number of the anger panel and Player 2 used the fewest number of the anger panel.

Fig. 10.
figure 10

The number of used anger panel

Full size image

Figure 11 shows the mean of favorability of the players. In this figure, the vertical axis denotes the mean of favorability and the horizontal axis denotes the players. From the results, Player 3 pointed the largest and Player 2 pointed the smallest favorability, and the order of the favorability is matched with the order of the number of the anger panel. Then, in the experiment, it is assumed that the anger panel is not recognized as negative expressions and did not force to change the winner selections of the other players.

Fig. 11.
figure 11

Favorability of players

Full size image

8 Conclusion

This paper has aimed to investigate effective emotional expressions for consensus-building in virtual communications. We employed cross-cultural game Barnga with the emotional panel as virtual environments. From the experimental results, the anger panel was effective for consensus-building, and the effect was confirmed by an experiment in which the player cannot use the anger panel. From a relationship between the number of the emotional panel used just before consensus-building and the number of turns spent until consensus-building, it was cleared that the anger panel was effective for early consensus-building. In addition, the anger panel which is generally accepted as negative expressions did not affect the favorability of players who used the anger panel in virtual situations. Therefore, it is assumed that consensuses are built smoothly by anger expressions without forcing to change selections of opponents in virtual communications.