Sunday, September 11, 2011

Turn-Taking in Google+ Hangouts

Google+ includes a video chat feature, called "Hangouts," that allows impromptu audio-video conferences for up to 10 participants at a time. Each participant in a Hangout sees thumbnail-sized video feeds for each of the other participants, plus a large window displaying the video feed of the participant who has the floor. This large video window is switched between participants based on the volume of their audio input. Google's Chee Chew noticed that this method of switching did not work well for hangouts in which participants used sign language to communicate:1

as i hungout in signing hangouts, i also noticed that most people were trying to watch others sign from the thumbnail video. our voice activated video switching for the main video usually just stayed on whomever had the most background noise.

This led to the introduction of a new feature:

So, we added a "Take the floor" feature. Give this a try and let us know how it works for you. Here's how to use it.

1) have everyone mute their audio.
2) when you want to sign something, hit Shift+s.
3) when you see yourself as the main video, that's your cue. you've got the floor... everyone's main video has switched to you. sign away.

note that this only works when you are muted. and once you've "taken the floor" no one else can take the floor for a few seconds. we did this to arbitrate if multiple people do it at the same time... whomever you see in the main video has the floor. If that's you, go for it.

In face-to-face multiparty conversations, turn-taking is accomplished by coordination between participants using a complex set of signals. While Ann is speaking, Bob may request to speak by a combination of posture, gesture, facial expression, and eye contact. Using similar methods, Ann may acknowledge Bob's request and transfer speaker status to him. Alternatively or in addition, there are a variety of linguistic methods Ann can use to signal the end of her turn, allowing Bob to take control by beginning to speak.

If Ann continues to speak while Bob wishes to speak, Bob may decide to interrupt Ann. Bob may do this with or without first making a non-verbal request to speak. If Bob is skillful, he might time his interruption to occur at a point that could be interpreted as the end of a turn. For example, if Bob starts to speak just as Ann is completing an utterance, and Ann then cooperates by ceasing to speak, Bob's behavior may be seen by others as orderly turn-taking, even though Ann originally had intended to say more. Alternatively, Ann may continue speaking after Bob's interruption, leading to a situation in which two people are speaking at once. This creates a conflict which may be resolved in various ways.

In a multiparty conversation, if Ann and Bob do not quickly resolve a conflict for the floor, other participants may try to resolve the conflict by focussing their attention on one of the competing parties. In this way, a single group with one speaker may become two groups, each having its own separate "floor." More often, though, the actions of the group result in control being transferred to one person who becomes the next speaker.

The influence participants have over turn-taking in a conversation goes beyond resolving conflicts that arise due to interruptions. If Ann is talking too long, participants can signal that by withdrawing eye contact and gaze, and by using posture, gestures, and facial expressions to indicate boredom or impatience. If Ann is ignoring Bob's non-verbal signals requesting a turn to speak, other participants can reinforce B's request non-verbally by transferring their visual attention from Ann to Bob.

In addition to the non-verbal methods described so far, signals for turn taking also can be communicated by noises and utterances. Participants can signal impatience or boredom by making a variety of sounds. They can encourage Ann to end her turn with disruptive sounds such as coughing. An utterance such as "yeah" can be used by Bob to signal his intention to begin speaking, thereby giving the opportunity for Ann to end her turn gracefully, while allowing Bob to avoid the appearance of interruption if Ann continues to talk.2

The signaling and coordination methods described above can be generalized to situations involving more than one person seeking a turn. Readers who have participated in multiparty conversations are likely to recognize the methods I've described, and perhaps can think of additional methods I've left out. These methods may vary in details across cultures, but within a cultural context, participants in an ordinary conversation among similar adults generally assume that all participants will know what they know about taking turns.

A conversation is a collaborative activity which participants coordinate through methods that are mutually known. How this coordination is achieved when the conversation is mediated by a technology-based service such as Google+ hangouts will depend on the design of that service and the technology supporting it. Turn-taking in Hangouts provides an example of this.

As Chee Chew noticed, participants in a video chat are able to direct their attention to a video thumbnail rather than the main video window. This may provide a method for a participant to request a turn to talk using non-verbal signals, but a person viewed via a thumbnail typically will have less salience than a person present in a face-to-face conversation, and there are no methods for other participants to influence turn-taking through their non-verbal signaling. Even if the current speaker does note the non-verbal request, she cannot direct a non-verbal response to a particular participant as she might in a face-to-face conversation, nor can she choose to transfer control. Turn-taking occurs when a new speaker begins to talk during a silence or when a new speaker interrupts the current speaker by talking louder (unless everyone's microphone is muted as described above by Chee Chew). Since accidental switching of control can be triggered by any detectable sound, participants must refrain from making noise if they want to avoid this, and that will tend to inhibit the use of backchannel utterances (i.e. sounds like "uh-huh" to let the speaker know that the listener is paying attention and understands).

Given these limitations, one may wonder why Google+ Hangouts has gotten such a positive reception. It may be because, at this time, Hangouts allows real-time, multiparty conversations over a distance with greater convenience and sense of presence than any other freely available application. In addition, though, Hangouts provides features, such as the side-channel chat, that may support new methods for conversational coordination. The use and eventual widespread adoption of such new methods is made possible by the human capacity for collaboration. We are not born knowing methods for turn-taking, nor are we explicitly taught them rule-by-rule. We develop them, learn them, and adapt them through collaboration.


1Google+ post "Take the floor" by Chee Chew on September 10, 2011 describing the new feature for Google+ Hangouts.

2This works because utterances like "yeah" can function to switch control or as a backchannel acknowledgement.


REFERENCES

Clark, H. H. (1996). Using language. Cambridge: Cambridge University Press.

Clark, H. H., and Brennan, S. A. (1991). Grounding in communication. In L.B. Resnick, J.M. Levine, & S.D. Teasley (Eds.). Perspectives on socially shared cognition. Washington: APA Books.

Clark, H. H., & Schaefer, E. F. (1987). Collaborating on contributions to conversations. Language and Cognitive Processes, 2(1), 19-41.

Duncan, S. & Fiske, D. W. (1977). Face-to-face Interaction. Hillsdale, N. J.: Erlbaum.

Ervin-Tripp, S. (1979). Chidren's verbal turn-taking. In E. Ochs & G. Schieffelin (Eds.) Developmental Pragmatics. London: Academic Press, 391-414.

Kendon, A. (1967). Some functions of gaze direction in social interaction. Acta Psychologia, 26, 22-63.

Kraemer, H. C. & Jacklin, C. N. (1979). Statistical analysis of dyadic social behavior. Psychological Bulletin, 86, 217-24.

Sachs, H., Schegloff, E. A. & Jefferson, G. A. (1974). A simplest systematics for the organization of turn-taking in conversation. Language, 50, 679-793.

Schegloff, E. A. (1982). Discourse as an interactional achievement: Some uses of "uh huh" and other things that come between sentences. In D. Tannen (Ed.), Analyzing discourse: Text and talk. Georgetown University Roundtable on Languages and Linguistics 1981. Washington, DC: Georgetown University Press, 71-93.

Yngve, V. H. (1970). On getting a word in edgewise. Papers from the Sixth Regional Meeting of the Chicago Linguistic Society. Chicago: Chicago Linguistic Society.