香京julia种子在线播放

Speech-associated gestures, mainly the hand and arm movements speakers perform while they speak, have been documented in a variety of cultures and a range of different languages (e.g., Efron, 1941/1972; Morris et al., 1979; Kendon, 1981, 1990, 1992, 2004b; Calbris, 1990; McNeill, 1992; Brookes, 2004 inter multa alia; see Kita, 2009 for an overview and chapters 73–90 in Müller et al., 2014). There are no reports of a culture where gesture is absent, and this strongly suggests that gesture is a universal feature of human communication. As such, gesture production and its relation with speech has been the object of extensive scrutiny in contemporary psycholinguistic and cognitive research, and considerable insights have been achieved in the understanding of the link between the two systems. For instance, it has been shown that gesture and speech are temporally aligned such that the most meaningful part of a gesture, the stroke, tends to be coordinated with the co-expressive part of speech, in monolingual and bilingual speakers (Kendon, 1972; McNeill, 1992; Loehr, 2007; Graziano et al., 2020). It has also been shown that gestures mainly occur with fluent speech and not with silence or speech disfluencies, independently of the language spoken and the degree of linguistic competence (Gullberg, 1998; Graziano and Gullberg, 2018). Moreover, it has been demonstrated that gestures are co-expressive with speech and can integrate its content either referentially or pragmatically (McNeill, 1992; Kendon, 2004a). That is, they can indicate concrete or abstract entities (deictic, pointing gestures), show the size or the shape of an object, or enact an action (representational, also called iconic gestures in the literature), or represent an abstract concept, such as time (also labelled as metaphoric gestures). Gestures can also communicate pragmatic content, emphasising relevant parts of the speaker's discourse (often called beats), expressing speech acts (e.g., a denial, an offer, etc.), indicating how to interpret the speaker's discourse or stance towards it, and so on.

Furthermore, gestures are also involved in the process of structuring interaction, as when they are used to regulate turn taking, to claim the role of speaker or to provide feedback (e.g., Goodwin, 1981; Bavelas et al., 1992; Kendon, 2004a).

Its universal character notwithstanding, gesture production varies both at individual and cultural/linguistic level. This has often led to some stereotypical views regarding in particular gesture frequency, often attributed to unspecified crosscultural differences. Indeed, such claims have been mainly based on anecdotal remarks, drawing on a (folk) notion of high vs. low gesture frequency cultures. For example, portraying features of the Italian character, Barzini (1964) wrote: “Italian gestures are justly famous. Indeed Italians use them more abundantly, efficiently, and imaginatively than other people” (p. 65). Gesturing is even described as part of the Italian identity (Nardotto Peltier and McCafferty, 2010); a characteristic that, supposedly, suggests that Italians are more communicative. The only empirical findings in support of this claim is provided by Graham and Argyle (1975), who showed that Italians benefit more from the presence of gestures in discourse to solve certain tasks than English speakers.

Lay people also share stereotypical views of gesture frequency as well as of other aspects of gestural behaviour, such as gesture size. Sekine et al. (2015) asked people in France, Italy, Japan, the Netherlands and USA about their views on the gestural production of 13 languages/cultures (Western and Asian). They found that people tended to believe that speakers of “Western languages” gesture more and with bigger movements than speakers of “Asian languages”. Again, Italians were those to whom these characteristics were attributed the most, followed by Spanish and American English speakers. Interestingly, in addition to some common opinions (e.g., people gesture more when they speak a foreign language; pointing at others with the index finger is impolite; the size of gestures depends on the language spoken), people tended to attribute different weight to different aspects of gestures (e.g., Americans believe that people produce more gestures when they speak in public; the Japanese believe that gesturing is linked to personality and that it should be avoided in logical discourse).

Strikingly, studies assuming differences in gesture frequency have produced surprisingly little evidence in support of crosscultural variation in this respect. Frequency differences are often assumed in studies of bilingual speakers to test whether non-verbal behaviour of one language/culture becomes visible in the other language; that is, whether gestural “transfer” can be found in bilinguals speaking one supposedly high gesture frequency language (typically, Italian, French, Spanish, American English) and one low gesture frequency language (generally, British English, Chinese, Japanese and, depending on the comparison, American English). Results from these studies diverge. For example, Nicoladis et al. (2009) reported that English-Spanish and French-English bilinguals gestured more when speaking English than English monolinguals. This result was interpreted as evidence of gestural transfer from the (presupposed) high gesture frequency culture (French, Spanish) into the (presupposed) low gesture frequency culture (English) as a function of exposure to the second language cultural norm for gesture (France and Spain, in this case), and in particular for iconic gestures, in the authors' terms. Similarly, So (2010) claimed a transfer effect from American culture for English-Chinese bilinguals in Singapore. The author found that American English monolinguals gestured more than Chinese monolinguals, and subsequently showed that bilinguals gestured as frequently as American English monolinguals when speaking English, and more than Chinese monolinguals when speaking Chinese. Transfer seemed to happen only for representational gestures (which in the author's classification included abstract deictic and iconic gestures). However, it is important to note that these Chinese-English bilinguals were raised and lived in Singapore where, as the author admits, people could potentially be influenced by other cultures as well and not directly by the American one to which they are compared.

In contrast, other studies have found no evidence for posited transfer. For example, despite finding that Italians gestured more than speakers of British English, Cavicchio and Kita (2013) did not find transfer for gesture rate or for the use of gesture space (referred to as gesture salience) in Italian-English bilinguals. The authors concluded that when bilinguals switch language, they switch gestural parameters as well. Similarly, Nicoladis et al. (2009) found no evidence in support of a hypothesised frequency transfer effect from French into English. Although English-French bilingual children used more gestures than monolinguals when speaking both languages, French monolingual children did not gesture more than their English counterparts.

The monolingual child language literature also reveals mixed results. Iverson et al. (2008) found that very young Italian children produced gestures more frequently than American children of the same age. Capirci et al. (2010) and Colletta et al. (2015) also reported that Italian children make more use of gestural resources than French children, and both groups more than American children. Similarly, in naming tasks studies, Marentette et al. (2016) found that Italian children produced more representational gestures than Canadian English-speaking children. An analogous result was reported by Cattani et al. (2019) who compared Italian, Australian and British children. However, while the production of representational gestures was higher in Italian in comparison to the Australian and British children; the same was not true for pointing gestures whose production did not differ between Italian and Australian children who in turn produced more of such gestures than British children. These studies ascribe this difference to the nature of the adult gesture models to which children are exposed (but see Goldin-Meadow and Saltzman, 2000, who examine Chinese and American mothers' gestures to hearing or deaf children). Yet, no systematic comparison of the adults' gestural behaviour of the same cultures is available. The explanations, thus, rely mainly on anecdotal reports. Moreover, considering studies in which gesture production is actually compared between cultures, there is already some evidence that this might not be a suitable or at least the only explanation. For instance, Pettenati et al. (2012) found no differences in the gesture frequency of Italian and Japanese toddlers performing a naming task. Similarly, Gullberg (1998) found that the gesture frequency of French and Swedish adult speakers engaged in the same narrative task did not differ and Müller (1994) reported the same result comparing Spanish and German adult speakers.

In sum, the evidence for high vs. low gesture frequency cultures, and of gesture transfer from high to low gesture frequency cultures are mixed.

Differences in gesture use across cultures have long been recognised (see Kita, 2009 for a review). It is well known, for example, that in every culture there is a repertoire of conventional gestural forms (e.g., Morris et al., 1979; Kendon, 1981; Calbris, 1990; Payrató and Clemente, 2020) often used in place of spoken expressions. Such gestures, called “emblematic” (Efron, 1941/1972), “emblems” (Ekman and Friesen, 1969) or “quotable gestures” (Kendon, 1992), are characterised by a set form/meaning association. That is, a particular hand shape is associated with a more or less stable meaning that is shared among the individuals of the same social or cultural group. Examples of such gestures include the “thumbs-up” gesture or the “ring” gesture commonly used to express OK, “all good” (Morris et al., 1979) in several cultures/countries, like Sweden or Italy, but which assumes different values in other countries or even in some regions within the same country. For example, the “thumbs-up” gesture is an offensive gesture in Sardinia (Italy) and some parts of Greece, and the “ring” gesture is an insult in Greece and in Turkey, while it represents both “excellent” and “zero” in France (Morris et al., 1979).

In addition to culture-specific form/meaning associations, gesture production is also affected by language-specific features, such as lexical and syntactic patterns for encoding motion or spatial information. For instance, Gullberg (2011) showed that French and Dutch speakers gesture differently when they talk about caused motion or placement as a reflection of the different ways in which placement events are lexicalised in these two languages. Hence, while French speakers use the generic placement verb mettre (“put”) and typically gesture only about the direction of the placement movement, Dutch speakers must choose between zetten (“set”) and leggen (“lay”), depending on the properties of the object being placed, and this is paralleled in gestures that also tend to incorporate the object in the hand shape. Similarly, Kita and Özyürek (2003) and Özyürek et al. (2005) provided evidence that Japanese, Turkish and English speakers gesture differently when describing motion events as an effect of the syntactic packaging of motion information specific to each language. They found that English speakers typically express the manner and the path of a motion event in one single spoken clause (such as he rolls down) and can also depict both pieces of information in a single gesture. In contrast, Turkish and Japanese speakers generally express manner and path in two separate clauses (e.g., it descends rolling) and this is reflected in two gestures, one gesture indicating the direction (descends) and another representing the manner of motion (rolling).

Crosscultural differences in gesture use can also be found at the level of discourse structuring. For instance, in narrative retellings Dutch, Swedish, and French speakers usually gesture more about referents' actions than Japanese speakers, who are instead more likely to gesture about locations and entities that form the narrative setting. This difference can be traced to the information organisational principles typical of these languages: in speech, Dutch, Swedish, and French treat actions as focused information, while Japanese assigns this status to scene settings (Yoshioka, 2005; Gullberg, 2006).

Strikingly, crosscultural differences in gestural styles are seldom discussed in a broader perspective. To date the only exception is Efron's (1941/1972) pioneering study where he compared the gestural behaviour of Eastern European Jews and Southern Italian immigrants in New York City still speaking Yiddish and Italian, respectively. Interestingly, Efron reported that the two groups differed very little in gesture frequency, but more in the kinesic structure of the gesture and in the range of space used, as well as in the type of gestures produced in association with speech. He observed that Italian immigrants' gestures involved not only the hand but also the entire arm from the shoulder, which was moved in a wide space including the lateral plane. Gestures produced by Jewish immigrants, in contrast, involved mainly the elbow and the wrist, and were produced mostly in the vertical and frontal planes. Moreover, whereas Italians made more use of gestures to illustrate the content of their speech (that is, they produced more referential gestures), Yiddish speakers tended to use gestures to mark the logical structure of the talk (that is, they used more pragmatic gestures). In addition, Italians also seemed to have a wider repertoire of emblematic gestures. Efron's observations of Italians' gesture features were later supported by Kendon (2004b) who compared gestures produced by a Neapolitan (Southern Italy) and an English speaker. He found that the Neapolitan speaker produced gestures in a wider space and in a more visible area (in the centre of the visual field of the interlocutor) than the English speaker, who instead moved the arm at the level of the waist. Additionally, the Neapolitan speaker produced a wide range of hand shapes that were much better defined and specialised in their usage than those produced by the English speaker. Müller (1994) also remarked on differences between Spanish and Germans in terms of gesture space, with Spanish speakers producing more gestures above the shoulder at the height of the face. Gullberg (1998) similarly found that French speakers were more likely to produce small beat-like gestures whereas Swedish speakers instead preferred representational gestures depicting content.

Taken together the studies summarised above suggest that patterns of spatial and kinesic organsation in gesture production might be culture-specific, and, on the other hand, that the preponderance for a particular gesture function might be the expression of culture-specific rhetorical styles. Several authors used the notion of rhetorical styles, but defined them somewhat differently. Slobin's (1996, 2004) well-known notion of thinking-for-speaking, which suggests that speakers of different languages select different elements of events and reality to talk about as a function of the linguistic units at their disposal, has also led to the notion of language-specific rhetorical styles (Slobin, 2004). That is, speakers of different languages will construct discourse patterns that differ slightly as a consequence of the linguistic (lexical and morphosyntatic) options at hand. For example, if you speak a language that offers grammaticised verbal aspect, this will lead to discourse with a focus on ongoingness of events. If instead you speak a language without grammaticised aspect, your discourse will be characterised by a focus on boundedness and spatial endpoints (e.g., Von Stutterheim and Nüse, 2003). Studies of gestural reflections of motion events and such different rhetorical styles have already been reviewed above (e.g., Özyürek, 2017 for an overview).

Rhetorical styles also considered to result from other alternations. For example, in story tellings or narratives, there may be different patterns of alternations between clauses that convey different types of information (Labov and Waletzky, 1967; Labov, 1972). For example, Cassell and McNeill (1991) and McNeill (1992) analyse a narrative in terms of “narrative levels” distinguishing between narrative clauses, which push the story line forward, metanarrative clauses, which provide comments on the story line, and paranarrative clauses, which highlight the narrative situation itself. They showed that in English different narrative levels typically come with different gestural patterns. For example, narrative clauses are often accompanied by representational gestures that depict actions and events; transitions to other narrative levels may be punctuated by beat-like gestures; metanarrative clauses are often accompanied by what they called abstract deictic and metaphorical gestures, and paranarrative ones are generally accompanied by fewer gestures overall.

Building on this work, Nicoladis et al. (2018) have suggested that there may be a link between gesture production, specifically frequency, and story-telling style. Assuming that different cultures have different narrative styles, the authors focused on the distinction between chronicle style—what happened and how—and evaluative style—what happened and why—and hypothesised that speakers adopting a chronicle style would produce more gestures than speakers who adopt an evaluative style. They also hypothesised that the chronicle style would be a characteristic of high gesture frequency cultures. Their study compared the spoken and gestural production of four groups of bilinguals all having English as L2 and Mandarin, Hindi, French or Spanish as L1 while telling narratives. They found that the Chinese-English bilinguals (supposedly low gesture frequency cultures) and Hindi-English bilinguals (a group for which they had made no frequency prediction) gestured less than the French-English and Spanish-English bilinguals (both supposedly high gesture frequency cultures), and that the first two groups tended to use an evaluative style, while the second two groups tended to use a chronicle style. They concluded that a chronicle style is associated with more gestures than the evaluative one. These results are intriguing, but raise many questions. For example, the study did not have any monolingual comparison groups making it difficult to assess whether differences were really due to style or to some other underlying factor. Also, the study only considered representational gestures, capturing only a part of the gestural activity.

In sum, despite widespread interest in crosscultural differences in gestural behaviour, the overall empirical evidence remains somewhat patchy and often contradictory, especially regarding gesture frequency. Moreover, most studies looking at frequency have often focused exclusively on representational gestures, leaving other gestural functions aside. This means that we have a poor understanding of the ways in which frequency may interact with differences in gestural functions more generally. And finally, very few studies have considered the potential role of spoken context defined as rhetorical style for differences across cultures.

We recruited 12 Italian (8 female, age range 19–31, M^age = 22), and 12 Swedish (8 female, age range 20–48, M^age = 28) native speakers to act as narrators in a narrative task. They were asked to bring a friend to act as the listener. All participants were university students recruited by word of mouth in Naples at the Università degli Studi di Napoli “L'Orientale” and in Lund at Lund University, respectively. Italian participants volunteered without compensation, Swedish participants were offered a lunch voucher. All participants provided signed consent before data collection.

Participants were invited to participate in the study with a friend. In Naples the narratives were collected in a quiet room at the university, in Lund in one of the studios of Lund University Humanities lab. The setting was the same in both locations: the participants sat on a chair side-by-side, and a camera was placed on a tripod at a distance from the pair suitable for catching the entire body of both of them.

The participants who had the role of narrators were asked to watch an episode of the series Pingu lasting 90 seconds (Pingu's Family Celebrates Christmas, 1992) on a laptop which was then removed after viewing, while the listener waited in another room. The cartoon featured a family of penguins getting ready for Christmas celebrations. The narrators were then asked to retell from memory what happened in the cartoon to their friend (who had not seen the video). The listeners were instructed not to interrupt the speaker but were allowed to give feedback signals (such as mh, I see).

Since some participants did not arrive with a friend for the recording session, seven Italian narrators spoke to the same listener (known to all of them). Similarly, two Swedish narrators spoke to the same listener (again, also known to the second speaker). In all these cases, speakers did not know that their listener had already listened to the story. It was important that the interlocutor was presented as naïve to the content since this increases the communicative pressure on the speaker which in turn promotes gesture production. The narratives were video recorded for later analysis.

Statistical analyses were conducted in the software R, version 0.98.953 (R Development Core Team, 2014). We ran an independent t-test on mean gesture rates, and a Mann Whitney U test on mean gesture/clause rates. To explore the relationship between gesture functions across narrative levels, we used the glmerMod package in R to perform Generalised Mixed-effects Models (GLMMs) with random intercepts for participants (Baayen, 2008; Baayen et al., 2008). Models were fitted using maximum likelihood (Laplace approximation) [“glmerMod”] of the binomial family (logit).

Italian speakers overall produced more gestures (n = 698) than Swedish speakers (n = 389). We first examined the overall mean gesture rate (gestures/100 words) in Italian and Swedish speakers. To check for normal distribution in the data, we performed Shapiro-Wilk tests which confirmed that the gesture rates did not depart significantly from normality in Italian (W = 0.959, p = 0.768) nor in Swedish (W = 0.964, p = 0.839). Further to this, homogeneity of variance was examined with Levene's test, which showed no evidence of unequal variances [F_{(11, 11)} = 1.841, p = 0.326]. Subsequently, an independent samples t-test was run on the mean gesture rate comparing Italians to Swedes. The results showed that the Italians produced significantly more gestures per 100 words on average than the Swedes [MIt = 21.95 vs. MSwe = 11.04; t₍₂₂₎ = 4.5, p < 0.000].

We also examined the gesture rate per clause in the two languages, since the clause is the relevant unit of analysis for the narrative levels. Swedish speakers produced more clauses (n = 547) than Italian speakers (n = 454), but Italian speakers on average produced 1.8 gestures/clause whereas Swedish speakers on average produced 1.2 gestures/clause. Shapiro-Wilk tests confirmed that the gesture rates did not depart significantly from normality in Italian (W = 0.931, p = 0.387) nor in Swedish (W = 0.889, p = 0.113). However, Levene's test showed unequal variances [F_{(1, 22)} = 8.43, p = 0.008]. A non-parametric Mann-Whitney U test was therefore run, showing that Italian speakers had a significantly higher rate of gestures/clause than Swedish speakers (U = 7, p < 0.001).

Table 1 shows the distribution of gesture function (referential vs. pragmatic) by language group. Italian speakers produced more pragmatic than referential gestures and Swedish speakers showed the opposite pattern.

Table 2 presents the number and proportion of clauses at each narrative level (narrative, meta- and paranarrative levels) across the two languages. The groups behaved similarly in that the majority of the clauses were narrative clauses (61% in Italian vs. 56% in Swedish). Both groups produced the same proportion of metanarrative clauses (24%) and Swedish speakers produced somewhat more paranarrative clauses (20%) than the Italian speakers (14%).

To address the research question, we probed how Italian and Swedish speakers distributed gesture functions over narrative levels. Figure 1 presents the mean proportion of gestures by narrative level in Italian and Swedish. Referential gestures occurred most frequently with narrative clauses (as expected) both in Italian (46 %) and in Swedish (85 %). Interestingly, Swedes also produced many referential gestures with meta- (72 %) and paranarrative clauses (66 %) compared to Italians (31 % with metanarrative and 10% with paranarrative clauses). Conversely, pragmatic gestures occurred more frequently with metanarrative clauses in both groups, followed by paranarrative clauses in Italian and Swedish.

To explore the relationships in more detail, we ran a Generalised linear mixed model fit by maximum likelihood (Laplace Approximation) [“glmerMod”] on the binary variable gesture function, with pragmatic gestures, Italian, and metanarrative set as baselines. The model output (Table 3) indicated a significant main effect of language confirming that Italians produced more pragmatic gestures than Swedes overall. There was also a main effect of narrative level, such that there were fewer pragmatic gestures with narrative clauses and more with meta- and paranarratives. For referential gestures, the pattern was the reverse, that is, more referential gestures with narrative clauses than with meta- and paranarrative clauses (which is in line with the findings by Cassell and McNeill, 1991). The analysis revealed no interaction (χ² = 4.491, df = 2, p = 0.106), meaning that the patterns were overall the same in Italian and Swedish.

In the following we provide examples of the typical patterns in the two groups. In the transcription, boldface indicates the stroke, and the number in [x] indicates the sequence of the gestures.

The two excerpts relate to the same part of the Pingu story and in both cases all clauses are narrative clauses. Example (4) comes from a Swedish speaker; all gestures are referential gestures, illustrated in Figure 2.

This Swedish speaker uses only referential gestures to represent the different actions she is talking about.

Gesture 1: the right hand, closed with the index loosely extended, is moved towards the speaker's left side. It represents the path of the movement of the child penguins, mentioned in the previous part of her discourse, who go towards the mother penguin.

Gesture 2: both hands closed as if holding something and parallel to each other moved back and forth alternatively. This gesture likely represents the movement of the rolling pin, not mentioned in the speech.

Gesture 3: right hand with fingers loosely extended closes as if picking something. This gesture represents the action of taking a part of the dough.

Gesture 4: both hands open with palm down move towards each other as if collecting something and move towards the left (as if positioning something on the side).

Gesture 5: both hands with palm down and finger slightly bended forming a loose round shape move quickly up and down and laterally. They represent the action of pressing the molds for shaping the dough.

Gesture 6: both hands closed as if holding something moving towards the speaker. They represent the action of taking/holding the plate.

Example (5) in Italian shows an alternation between referential and pragmatic gestures, illustrated in Figure 3.

Gesture 1 and 2 are two very similar referential gestures: both hands are close as if holding something, in 1 both the left and the right hands move in circular movements with a wrist movement; in 2 the left hand is still, while the right hand performs the same circular movement. They probably represent the action of using an object to pick something (e.g., holding the spoon to take the cream she is talking about).

Gestures 3 and 4 are both pragmatic gestures, an example of a Palm Presentation (PP) gesture (Kendon, 2004a): the speaker had the hands crossed and opens them bringing the palm up with a slight rotation of the wrist and an up and down movement so that the hand are brought into the immediate frontal space; the two gestures follow each other in rapid movements. Kendon has proposed that this gesture is used to present a piece of information to the listener. It is worth noting that the PP constitutes a particular cluster of a movement pattern (write rotation) resulting in a manual form (open hand palm up) that sets it apart from a referential gesture of holding an object.

Gesture 5 is a referential gesture that accompanies two clauses: both hands are open with the palms towards the speaker and move rapidly back and forth. This gesture represents the action of signalling something hot.

Gesture 6 and 7 are again two pragmatic gestures performed as gestures 3 and 4 with the same function.