2.2 What Shapes the Use of Gestures?
2.2.5 Discourse Context
Gesture occurrence is not limited to the context of lexical affiliates in concomitant speech, or as Kendon (1987, p. 90) put it, “a speaker will select a model of formulation, not only in the light of a comparison between its adequacy of representation and the image that it is intended to convey, but also in the light of what the current communication conditions are”. Rather, each single multimodal utterance is integrated into a greater context. It is carried out within a larger utterance fulfilling a particular function in order to achieve an overall communicative intention, which itself occurs within a particular situation including specific people talking to each other in a specific environmental situation. Gestures are, thus, embedded in cascading levels of discourse context (Gerwing and Bavelas, 2004). Furthermore, there is indeed evidence that gesture use is sensitive to the different levels of the overall discourse context.
Initial Common Ground A considerable amount of studies have investigated the role of common ground, which is the sum of mutual, common or joint knowledge, beliefs and suppositions that interlocutors share, providing the background for their interaction (Stalnaker, 1978; Clark, 1996).
With regard to gesture frequency, results are mixed. Holler and Wilkin (2009) used a between-subjects design with a ‘common ground’ condition, in which some shared knowledge about the stimulus material was experimentally induced, and the ‘no common ground’ condition, in which participants did not share any experimentally induced common ground. The findings showed that speakers gesture at higher rates (gestures per one hundred words) when they shared a common ground. Holler and
Wilkin conclude that gestures play an important communicative function, even when speakers convey information that is already known to their addressee. On the other hand, Jacobs and Garnham (2007) used a cartoon story re-narration setting and found that gestures were produced at lower rates when speakers believed the listener was also able to view the content of the stimulus, i.e., when speaker and listener had common ground.
Other studies went beyond the pure consideration of gesture frequency, and investigated the qualitative nature of gestures. Gerwing and Bavelas (2004) asked participants to describe play actions they performed with a certain set of toys, either to someone who played with the same or a different set of toys. Their findings revealed that speakers used less precise, less complex, and less informative gestures when talking to participants with whom they shared mutual knowledge as compared with those who did not have access to the same knowledge. Similarly, Holler and Stevens (2007) found that gestures’ spatial extent was sensitive to information structure in the representation of size information. They used a referential communication task which involved participants locating a target entity in an array of other entities of different sizes. Their analysis focused on the representation of the size of particularly large entities in the array. It was found that entities were represented gesturally as significantly smaller when their actual size was already known to the addressee, as compared to when addressees had no pre-existing shared knowledge regarding the respective entities’ size. Parrill (2010) found an effect of common ground on the encoding of semantic information in gestures. Participants described a target event in which a cat ‘melted down some stairs’. Her analysis focused on the number of times speakers mentioned the stairs, i.e., the ground component, in their descriptions. When the content of the stimulus cartoon was absent from speaker-addressee common ground, production of ground information in gestures was increased.
In sum, research results regarding the correlation of gesture use and common ground yielded mixed results. While some suggest that gestures are produced at a lower rate, less precisely, and with less information in situations with common ground, others provided evidence that gesture rate is increased, and that gestures remain similar with regard to their form and the amount of information they convey.
Information Structure Apart from the knowledge that interlocutors share from the outset of a conversation, common ground also accumulates over the course of it. New information is defined as what the listener is not expected to know already. Given information, in contrast, is defined as what the listener is expected to already know (Haviland and Clark, 1974)3.
3. This dichotomy has been referred to in the literature with a variety of terms and meanings, e.g. differentiating between ‘focus’ and ‘presupposition’, between ‘hearer-old’ and ‘hearer-new’, and between ‘discourse-old’ and ‘discourse-new’ (cf. Prince, 1981)
Within these lines, Levy and McNeill (1992) as well as McNeill (1992) employed the notion of ‘communicative dynamism’ as a variable that correlates with gesture use4, this being defined as the extent to which the message at a given point is “pushing the communication forward.” The authors argued that sentences which are low in communicative dynamism, are typically not accompanied by gesture. Gestures, rather, pick out the sentence elements which are of high communicative dynamism. And, the more complex a gesture is, the higher are the peaks in communicative dynamism. As a measure of communicative dynamism, the amount of linguistic material used to make the reference was chosen. Pronouns have less communicative dynamism than full nominal phrases, which have less communicative dynamism than modified full noun phrases. This implies that the communicative dynamism can be estimated by looking at the syntactic structure of a sentence.
Gerwing and Bavelas (2004) went beyond pure gesture frequency to consider how the physical form of gestures mark the status of information. Their findings revealed that speakers marked new information with gestures that were larger in size and more precisely articulated. Given information, in contrast, was accompanied by gestures that were smaller and less well articulated.
Previous Gestures Another correlate with gesture use is the set of gestures a speaker has used previously in a given discourse. Quek and McNeill (2000) introduced the notion of ‘catchments’ for gesture features which recur in at least two (not necessarily consecutive) gestures. The authors interpret catchments as clues to cohesive linkages in the discourse; a common discourse theme is expected to produce gestures with recurring features. Examples for such catchments are, for instance, a single moving hand (referring to a single moving entity), a round handshape referring to a ball, or a particular spatial configuration of the hands referring to the relative position of two entities (McNeill, 2005).