How event perspective influences speech and co-speech gestures about motion

Journal of Pragmatics 128 (2018) 22e29

Contents lists available at ScienceDirect

Journal of Pragmatics journal homepage: www.elsevier.com/locate/pragma

How event perspective influences speech and co-speech gestures about motion* € Wojciech Lewandowski a, *, S¸eyda Ozçalıs ¸ kan b a b

University of Copenhagen, Department of English, Germanic and Romance Studies, Emil, Holms Kanal 6, 2300 Copenhagen, Denmark Georgia State University, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 November 2017 Received in revised form 2 March 2018 Accepted 2 March 2018

Speakers of different languages systematically differ in how they package manner and path components of a motion event rendered from a self-motion perspective in both speech and co-speech gesture. Speakers of satellite-framed languages (e.g., German) use a conflated pattern, synthesizing manner and path of motion into a single clause or gesture; while speakers of verb-framed languages (e.g., Spanish) use a separated pattern, expressing manner and path in separate clauses or gestures. In this study, we ask whether the close coupling between speech and gesture observed for self-motion also becomes evident for events rendered from a caused-motion perspective, which show greater use of conflated strategy in speech across different languages. We observed speech and co-speech gestures of adult native Spanish and German speakers (N ¼ 15/language), as they described an animated clip depicting both self- and caused-motion events. Our findings showed that expression of motion in speech systematically varied by event perspectivedwith greater use of conflated strategy in caused-motion events than in self-motion events. More important, this variability became evident in co-speech gesture in both languages. Our findings show that speech and gesture form a tightly integrated system, jointly reflecting systematic differences not only between languages, but also within a particular language. © 2018 Elsevier B.V. All rights reserved.

Keywords: Motion events Caused-motion Spatial gesture Spatial language Cross-linguistic gesture Event perspective

1. Introduction Gesture and speech form a tightly integrated system in the expression of motion events (Kita, 2000; McNeill, 1992, 2000). As € postulated by the ‘interface hypothesis’ (Kita and Ozyürek, 2003), even though gesture and speech form separate systems, they work in tandem from the conceptual planning to the articulation of a motion expression. More specifically, gesture not only helps speakers package spatial information appropriate for its verbalization in a particular language, but it also is largely shaped by the structure of the native speech patterns of the language it accompanies. These bi-directional influences between the two modalities, in turn, result in close similarities between the types of information conveyed in speech and in co-speech gesture. Research to date provides compelling evidence for the interface hypothesis: when describing the movement of a figure from a self-motion perspective (e.g., “she runs into the house”), speakers follow language-specific patterns in the way they organize € € semantic components of the event in both speech and co-speech gesture (Kita and Ozyürek, 2003; Ozçalıs ¸ kan, 2016; € € Ozçalıs ¸ kan, Lucero & Goldin-Meadow, 2016a,b; Ozçalıs ¸ kan & Slobin, 1999). Here we ask whether the tight coupling * This research was supported by funding from the Danish Council for Independent Research (DFF e 4180-00216) and the Marie Sklodowska-Curie Individual Postdoctoral Fellowship (H2020-MSCA.IF-2014-658596). * Corresponding author. E-mail address: [email protected] (W. Lewandowski).

https://doi.org/10.1016/j.pragma.2018.03.001 0378-2166/© 2018 Elsevier B.V. All rights reserved.

€ W. Lewandowski, S¸. Ozçalıs ¸ kan / Journal of Pragmatics 128 (2018) 22e29

23

between speech and co-speech gesture remains robust in the expression of events rendered from a caused-motion perspective, in which the movement of the figure is triggered by an entity other than the figure itself. We have some preliminary evidence € from previous work (Hendriks and Hickmann, 2015; Ozçalıs ¸ kan, 2000, 2005), suggesting that speakers show within-language differences in their expression of events in speech when rendering them from a caused-motion perspective compared to a selfmotion perspective. However, we do not yet know how systematic this variability is in speech about motion, and whether it also extends to co-speech gestures. If co-speech gesture mirrors the patterns found in speech, then we would expect that differences that we observe in the expression of self- vs. caused-motion events in speech within a language would also become evident in the expression of the same events in co-speech gesture in the same language. We examine this question by studying two structurally different languages (German, Spanish), using a narrative task designed to elicit motion event descriptions from both self- and caused-motion perspectives. The expression of motion in speech shows wide variability across different languages, which, at the same time, is also constrained by a finite set of universal patterns, particularly with respect to manner (i.e., the way in which something moves) and path (i.e., the direction in which something moves) components of motion. As suggested by Talmy (1985, 2000), the world's languages can be grouped into two types based on the way speakers package manner and path components of self-motion in their speech. Speakers of satellite-framed languages (S-language), such as German, show a conflated pattern, expressing manner € in the verb and path in a particle associated with the verb within a single clause (e.g., “Sie LAUFT HINEIN” ¼ She RUNS INTO). In contrast, speakers of verb-framed languages (V-language), such as Spanish, show a separated pattern in speech, expressing path in the verb and manner outside the verb, in an additional subordinate clause (“Ella ENTRA en casa CORRIENDO” ¼ She ENTERS the house by RUNNING). V-language speakers also typically express only path in their speech about motion and leave out € € manner information altogether from their descriptions (Ozçalıs ¸ kan & Slobin, 1999; Ozçalıs ¸ kan, 2009, 2016).1 The expression of self-motion in co-speech gesture produced by adult S- and V-language speakers largely follow the language-specific patterns observed in their speech as demonstrated across several different studies (but see McNeill and Duncan, 2000 for an exception).2 S-language speakers typically use the same conflated pattern in gesture, synthesizing manner and path components into a single gesture (e.g., wiggle index and middle fingers while moving hand from right to left to convey running right to left). In contrast, V-language speakers follow the same separated pattern that they use in speech, producing one gesture for path (e.g., trace a line from left to right to convey left to right trajectory) and one for manner (e.g., € € wiggle fingers in the same place to convey running; Gullberg et al., 2008; Kita and Ozyürek, 2003; Ozçalıs ¸ kan et al., 2016a). Similar to the patterns observed in speech, V-language speakers frequently express only path of motion in gesture, leaving out € € manner entirely from their gestures (Ozçalıs ¸ kan, 2016; Ozçalıs ¸ kan et al., 2016a,b, 2017). The existing evidence on motion events rendered from a self-motion perspective suggests strong cross-linguistic differences in its expression both in speech and in co-speech gesturedfurther highlighting that gesture and speech form a tightly € integrated system, with gesture mirroring the cross-linguistic differences observed in speech (see Ozçalıs ¸ kan and Emerson, 2016 for a review). However, less is known about within-language differences in the expression of motion in the two modalities. There is some preliminary evidence suggesting that speakers of both V- and S-languages increase their expression of manner information when describing events from a caused-motion perspective, resulting in more conflated descriptions in speech about motion. For example, previous work examining physical motion of animate entities (e.g., boy enters house vs. € boy is pushed into the house; Ozçalıs ¸ kan, 2000) and metaphorical motion of abstract entities (economy enters into recession € vs. unemployment pushes economy into recession; Ozçalıs ¸ kan, 2005) in English and Turkish provided some preliminary evidence that speakers of both V- and S-languages increase their expression of manner information when describing events from a caused-motion perspective, resulting in more conflated descriptions in speech about motionda pattern that has been extended to some other V-languages in more recent work (e.g., French; Hendriks and Hickmann, 2015). Hence these findings suggest within-language differences in the expression of self- vs. caused-motion events in speech. However, given the scarcity of research on caused-motion, we do not yet know how robust a role event perspective plays in speech about motion across different languages. By the same token, we do not know whether a similar pattern of within-language differences holds for the co-speech gestures speakers produce in the two types of languages. This study aims to fill this gap by studying withinlanguage variability in patterns of speech and co-speech gesture use in the self- vs. caused-motion descriptions produced by adult speakers of two structurally different languages, German (V-language) and Spanish (S-language). 1.1. Current study Languages can be classified as belonging to one of two typesdS-framed and V-frameddbased on their expression of selfmotion in speech (Talmy, 1985). At the same time, speakers of each language type also rely on speech patterns that do not fit 1 Several researchers, studying a broader set of languages, questioned the relevance of Talmy's binary typology and suggested the possibility of adding a third type of lexicalization pattern (i.e., equipollently-framed languages) in which manner and path are expressed by two different verbs that have equal morpho-syntactic status in the clause (Slobin, 2004), such as Thai (Zlatev and Yangklang, 2004) and various NigereCongo languages (Ameka and Essegbey, 2013). Some others (e.g., Feiz, 2011) suggested the possibility of a mixed-typological pattern, where a language carries both S- and V-framed characteristics (e.g., Farsi; see also Akhavan et al., 2017 for a similar argument). 2 An earlier study (McNeill and Duncan, 2000; see also McNeill, 2000) showed that gesture might supplement what is conveyed in speech. Adult Spanish speakers, who typically expressed only path information in their speech about motion, used gesture to add manner to their spoken descriptionsda pattern € that differed from most of the later work on the gestural expression of motion in V-framed languages (Gullberg et al., 2008; Kita and Ozyürek, 2003; € € Ozçalıs ¸ kan, 2016; Ozçalıs ¸ kan et al., 2016a).

24

€ W. Lewandowski, S¸. Ozçalıs ¸ kan / Journal of Pragmatics 128 (2018) 22e29

the typological characteristics of their language. For example, speakers of V-languages (e.g., Spanish) sometimes use a  HASTA la conflated strategy in speech, conveying manner and path in a single compact clause (e.g., “El chico CORRIO € escuela” ¼ The boy RAN TOWARD the school; Aske, 1989; Ozçalıs¸kan, 2015). Similarly, speakers of S-languages (e.g., English) use a separated strategy in speech, only encoding path or manner information (e.g., “The boy ENTERS the house”; “The boy is RUNNING”). These examples thus suggest that, even though S- and V-languages typically prefer one type of strategy over another, separated and conflated strategies in the expression of motion continue to co-exist within each languageda pattern € € that also becomes evident in co-speech gestures (Ozçalıs ¸ kan et al., 2016a,b; Ozyürek et al., 2005). In this study, we take these findings one step further and examine whether languages show systematic variability in their expression of motion in speech and in co-speech gesture when they describe events from two different perspectivesdself-motion vs. caused-motion. We ask whether speakers systematically use more conflated than separated packaging strategies in their speech about causedmotion within each language; we also ask whether co-speech gesture mirrors the patterns observed in speech, with speakers of each language using more conflated than separated patterns in their gestures about caused-motion. We predict that speakers of both German and Spanish will show an effect of event perspective, displaying increased preference for the conflated pattern in their caused-motion descriptions in speech as compared to their self-motion descriptions, a prediction € based on earlier work on physical and metaphorical motion events (e.g., Ozçalıs ¸ kan, 2000, 2005). Following earlier work that € showed tight coupling between speech and co-speech gesture in self-motion descriptions (e.g., Kita and Ozyürek, 2003; € Ozçalıs ¸ kan et al., 2016a, 2017), and in line with the interface hypothesis, we also expect that co-speech gestures will mirror the patterns found in speech in each language, with a greater preference for conflated gestures in caused-motion descriptions than in self-motion descriptions across language types. 2. Methods 2.1. Sample Participants included 15 adult native speakers of German (Mage ¼ 25, range ¼ 19e33; 8 females) and 15 adult native speakers of Spanish (Mage ¼ 21, range ¼ 20e36; 10 females). The German data were collected in Berlin, Germany, and the Spanish data were collected in Ciudad Real, Spain. The majority of the German (12/15) and Spanish (14/15) participants were university students; the remaining 4 participants had postgraduate degrees. 2.2. Procedure Each participant was interviewed individually in a dedicated laboratory space by an experimenter. They were first asked to watch a short silent video clip and then retell what happened in the video to an experimenter. There was no explicit instruction to gesture, but almost all of the participants (28/30) gestured spontaneously during their narrations. The video was a 90 s-long extract from a Charlie Chaplin movie (City Lights), which involved a man trying to commit suicide by jumping into a river and Chaplin's repeated efforts to rescue him. The video depicted motion events involving both self-motion (4 different manners: step, rush, swim, jump; 6 different paths: forward, backward, upward, downward, into, out) and caused-motion (4 different manners: push, pull, throw, submerge; 6 different paths: forward, backward, upward, downward, into, out); see Appendix I for a sequence of events included in the stimulus video. All participant responses were videotaped. 2.3. Transcription and coding All verbal responses produced by the participants were transcribed by native speakers of each language and segmented into sentence units. Each sentence unit contained one verb and its associated arguments and subordinate clauses (e.g., “Sie springen ins Wasser” ¼ They jump into the water; “Entran corriendo en el agua” ¼ They enter the water running). Each sentence unit was further coded into two types based on the perspective expressed in the event, as (1) self-motion (i.e., event involving self-initiated movement; e.g., “Sie krabbeln aus dem Wasser raus” ¼ They crawl out of the water; “Salen del agua” ¼ They exit the water’) or (2) caused-motion (i.e., events involving other-initiated movement; e.g., “Er wirft den Mann ins Wasser” ¼ He throws the man into the water’; “Tira el hombre al agua” ¼ He throws the man into the water). We also coded all gestures that accompanied each sentence unit, using QuickTime videos synchronized with time€ stamped Excel worksheets, following earlier work (Ozçalıs ¸ kan et al., 2016a). Gesture was defined as a communicative movement of the hand(s) with an identifiable beginning and an end. In our analysis, we focused only on dynamic iconic gestures depicting manner and path components of motion, which included path-only (e.g., moving index finger forward to convey trajectory forward), manner-only (e.g., wiggling fingers in place to convey running), and mannerþpath (e.g., wiggling fingers forward to convey running forward; moving open palms forcefully forward to convey pushing an entity forward) € gestures. Following earlier work (Ozçalıs ¸ kan et al., 2016a,b), we coded each sentence-unit in speech and in gesture further for packaging of semantic elements as either conflated (manner and path are both conveyed within a single spoken clause or within a single gesture) or separated (manner and path are conveyed in separate spoken clauses or in separate gestures). Reliability was assessed with an independent coder, blind to the hypothesis of the study. The first coder coded all responses. A second coder coded 20% of the data, which included 3 randomly selected participants in each language. Agreement

€ W. Lewandowski, S¸. Ozçalıs ¸ kan / Journal of Pragmatics 128 (2018) 22e29

25

between coders was 94% for identifying gestures, 86% for coding motion elements in gesture, and 97% for coding motion elements in speech. 2.4. Analysis The number of separated and conflated sentence units each speaker produced in speech and in gesture was tabulated, separately for the two languages (German, Spanish) and for the two event types (self, caused). Participants' use of the two packaging strategies showed considerable individual variability both in speech (range ¼ 0e12 instances) and in co-speech gesture (range ¼ 0e8 instances; see Appendix II for mean frequencies). We therefore converted all raw frequenciesdseparately for the separated and conflated motion eventsdinto proportions, and arcsine-transformed the proportions for analysis. We analyzed differences in speech and in co-speech gesture with mixed ANOVAs, with language (German, Spanish) as between-subjects, and packaging of motion elements (separated, conflated) as within-subject factors, separately for self- vs. caused-motion events. 3. Results

MEAN PROPORTION OF MOTION DESCRIPTIONS IN SPEECH

We first examined whether packaging of motion elements in speech varied by event perspective and found evidence for it. Our analysis showed a significant interaction between packaging and perspective, F(1, 28) ¼ 67.38, p < .001, h2p ¼ .71, which, importantly, did not interact with language (packaging x perspective x language; F(1, 28) ¼ 1.59, p ¼ .22)3. As can be seen in SEPARATED CONFLATED

1.00 0.80 0.60

0.40 0.20 0.00

SELF

CAUSED

SELF

CAUSED

MEAN PROPORTION OF MOTION DESCRIPTIONS IN GESTURE

Fig. 1. Mean proportion of sentence units with separated (gray bars) or conflated (black bars) speech produced by Spanish and German speakers for self- vs. caused-motion events (error bars represent standard error).

SEPARATED CONFLATED

1.00 0.80 0.60 0.40 0.20 0.00

SELF

CAUSED

SELF

CAUSED

Fig. 2. Mean proportion of sentence units with separated (gray bars) or conflated (black bars) co-speech gestures produced by Spanish and German speakers for self- vs. caused-motion events (error bars represent standard error). 3 Participants' motion descriptions in speech also showed a main effect of perspective, F(1,28) ¼ 5.0, p ¼ .03, h2p ¼ .15; a main effect of language, F(1,28) ¼ 5.13, p ¼ .03, h2p ¼ .15, along with a perspective  language, F(1,28) ¼ 5.0, p ¼ .03, h2p ¼ .15, and packaging  language interaction F(1,28) ¼ 23.39, p < .001, h2p ¼ .46dbut no main effect of packaging, F(1,28) ¼ 3.67, p ¼ .07.

26

€ W. Lewandowski, S¸. Ozçalıs ¸ kan / Journal of Pragmatics 128 (2018) 22e29

Fig. 1, speakers in both languages used a greater proportion of separated than conflated verbal expressions in their self-motion descriptions (Mseparated ¼ 0.83 vs. Mconflated ¼ 0.17)da pattern that was reversed for caused-motion descriptions (Mconflated ¼ 0.68 vs., Mseparated ¼ 0.32); see Table 1 for sample descriptions of motion events in speech, rendered from a selfvs. caused-motion perspective. We next examined whether packaging of motion elements in gesture varied by event perspective and also found evidence for it. Similar to speech, our analysis revealed a significant interaction between packaging and perspective, F(1, 28) ¼ 55.52, p < .001, h2p ¼ .67, which did not interact with language (packaging  perspective  language; F(1, 28) ¼ 0.24, p ¼ .63).4 As Fig. 2 illustrates, across the two languages, speakers produced significantly more separated than conflated gestures in their self-motion descriptions (Mseparated ¼ 0.89 vs. Mconflated ¼ 0.11), and more conflated than separated gestures in their causedmotion descriptions (Mconflated ¼ 0.74 vs. Mseparated ¼ 0.26); see Table 2 for sample descriptions of motion events in gesture, rendered from a self- vs. caused-motion perspective.

Table 1 Examples of separated (manner-only, path-only) and conflated (mannerþpath) descriptions in speech produced by German and Spanish speakers, using selfmotion (top half) and caused-motion (bottom half) perspectives. German

Spanish

Self-motion Separated (manner-only) “Er schwimmt” ¼ He is swimming “Er plantscht im Wasser” ¼ He is splashing in the water Separated (path-only) €llt ins Wasser” ¼ The man falls into the water “Der Mann fa “Chaplin kommt aus dem Wasser” ¼ Chaplin comes out of water Conflated “Sie klettern hoch” ¼ They climb up “Der kleine Mann springt ins Wasser” ¼ The short man jumps into the water Caused-motion Separated (manner-only) “Sie stoben sich” ¼ They push each other Separated (path-only) “Chaplin holt es zurück” ¼ Chaplin brings him back Conflated “Dann wirft er den Stein ins Wasser” ¼ Then he throws the stone into the water “Er versucht, den anderen herauszuziehen” ¼ He tries to pull out the other man

Self-motion Separated (manner-only) “Se encaraman” ¼ They clamber  tambaleando” ¼ The man is wobbling “El hombre se esta Separated (path-only) “Chaplin cae al río” ¼ Chaplin falls into the river “Salen los dos” ¼ Both are coming out Conflated “Se tira al agua” ¼ He plunges into the water “Se lanza al agua” ¼ He plunges into the water Caused-motion Separated (manner-only) “Tira una piedra” ¼ He throws a stone Separated (path-only) “Finalmente, lo saca” ¼ In the end, he takes out the other man Conflated “Tira la piedra a un estanque” ¼ He throws the stone into a pond “Chaplin lo empuja al agua” ¼ Chaplin pushes him into the water

Table 2 Examples of separated (manner-only, path-only) and conflated (mannerþpath) descriptions in co-speech gesture produced by German and Spanish speakers, using self-motion (top half) and caused-motion (bottom half) perspectives. German

Spanish

Self-motion Separated (manner-only) Swing downward facing open palm repeatedly leftward and rightward to convey sliding Separated (path-only) Move sideways facing palm from left to right to convey rightward trajectory towards bench Conflated Wiggle downward pointing middle and index fingers repeatedly while moving them downward to convey walking in downward trajectory toward stairs Caused-motion Separated (manner-only) None observed Separated (path-only) Move downward facing palm toward body to convey backward trajectory away from water Conflated Hop left and right downward facing cupped hands repeatedly while moving them upward to convey pulling in upward trajectory away from water

Self-motion Separated (manner-only) Bounce and alternate downward facing open palms repeatedly up and down to convey wobbling Separated (path-only) Move index finger downward to convey downward trajectory towards water Conflated Hop left and right downward facing cupped hands repeatedly while moving them upward to convey climbing in upward trajectory to shore Caused-motion Separated (manner-only) None observed Separated (path-only) Move sideways facing palm from right to left to convey leftward trajectory towards water Conflated Move outward facing open palms forcefully forward away from body to convey pushing in forward trajectory towards water

4 Participants motion descriptions in gesture also showed a main effect of perspective, F(1,28) ¼ 4.98, p ¼ .03, h2p ¼ .15; a main effect of packaging, F(1,28) ¼ 6.61, p ¼ .02, h2p ¼ .19, but no effect of language, F(1,28) ¼ 2.74, p ¼ .11, and no interaction for perspective  language F(1,28) ¼ 0.55, p ¼ .46, or packaging  language F(1,28) ¼ 2.29, p ¼ .14.

€ W. Lewandowski, S¸. Ozçalıs ¸ kan / Journal of Pragmatics 128 (2018) 22e29

27

4. Discussion In this study, we asked whether speakers of German (S-language) and Spanish (V-language) show systematic withinlanguage differences in their expression of motion events in speech, when rendering them from a self- as opposed to a caused-motion perspective, and if so, whether we find evidence of similar differences in their use of co-speech gestures in expressing the same two event perspectives. Our analysis of speech and co-speech gestures produced by 15 German and 15 Spanish adult native speakers in a narrative elicitation task showed that the perspective had an effect on the packaging of motion elements in both speech and co-speech gesture. SpeakersdGerman or Spanishdshowed greater preference for conflated than separated patterns in their speech about caused-motion eventsda pattern that was reversed for events rendered from a self-motion perspective. More important, and in line with the interface model of gesture production, the cospeech gestures mirrored the patterns found in speech, with greater preference for conflated packaging strategies for causedmotion events and for separated packaging strategies for self-motion events, extending the previously observed close integration between speech and co-speech gesture in the expression of self-motion to caused-motion events. Why do speakersdregardless of language typeduse more conflated strategies in their speech about caused-motion? One possible explanation could be the relative complexity of the two events. Unlike events rendered from a self-motion perspective, caused-motion events involve additional information about force dynamics, namely how a force is applied to a figure to initiate the movement (Talmy, 1988). In fact, majority of the caused-motion events in our study included this additional information, typically in the verb, both in Spanish (e.g., “tirar” ¼ throw, “empujar” ¼ push) and in German (e.g., “stoßen” ¼ push, “werfen” ¼ throw). The inclusion of force dynamics thus adds manner information to the verb, which, in turn, increases the occurrence of more conflated descriptions in each language. One interesting difference between the two languages, however, was that German speakers were more likely to add path satellites to such manner verbs conveying force dynamicsda pattern in line with the lexicalization pattern of their language (i.e., manner verb followed with path satellites), thus resulting in greater proportion of conflated caused-motion descriptions (M ¼ 0.87) in German, compared to Spanish (M ¼ 0.50). Spanish speakers, on the other hand, relied on multiple strategies when conveying caused-motion descriptions, using manner verbs with path satellites (e.g., “Tira al hombre al agua” ¼ He throws the man into the water’) or manner verbs on their own, without path satellites (e.g., “Tira al hombre” ¼ He throws the man)dthe two strategies that accounted for the majority (80%) of the caused-motion descriptions that they produced. A small percentage (20%) of the caused-motion descriptions produced by Spanish speakers also conveyed solely path information, typically encoded in the verb (e.g., “Saca al hombre” ¼ He takes the man out). The use of more varied strategies in speech, particularly in conveying self-motion events € with salient manners, has been shown in other V-languages (i.e., Turkish; Ozçalıs ¸ kan, 2015)da pattern that our study extends to caused-motion events. Why do co-speech gestures follow the patterns found in speech, with greater use of conflated gestures in describing motion events from a caused-motion as opposed to a self-motion perspective? We know from previous work that speech and gesture form a closely integrated system in the expression of self-motiondwith co-speech gesture mirroring the patterns € € € found in speech (i.e., interface hypothesis; Kita and Ozyürek, 2003; see also Ozçalıs ¸ kan, 2016; Ozçalıs ¸ kan et al., 2016a,b). In fact, gestures that are produced with speech carry the imprint of the language that they accompany even in the absence of access to native gesture patterns: congenitally blind adult speakers, who have never observed others' gestures, nonetheless € follow language-specific patterns in their co-speech gestures about motion (Ozçalıs ¸ kan at al. 2016b). Our findings extend this earlier work that shows close integration between gesture and speech in self-motion to the domain of caused-motion events, indicating that evidence of within-language differences for motion events in speech can also be found in co-speech gesture. It is important to note here that we did not observe the previously reported differences between Spanish (V-language) and German (S-language) speakers in either their speech or co-speech gestures about self-motion. In contrast to earlier work that € € showed a preference for the conflated pattern among S-language speakers (e.g., English; Kita and Ozyürek, 2003; Ozçalıs ¸ kan et al., 2016a; Slobin, 2004), both Spanish and German speakers, when describing events from a self-motion perspective, relied primarily on separated strategies in their speech. This might be an outcome of the stimulus we used in our study, which depicted a man's repeated suicide attempts by falling into and then coming out of the water. Not surprisingly, the most frequently mentioned self-motion scenes were the three ‘falling into the water’ and three ‘coming out of the water’ events. These two events could be described in one of two ways in German, either by using a conflated strategydnamely a manner verb combined with a path satellite (e.g., “Er stürzt ins Wasser” ¼ He plunges into the water; “Er krabbelt aus dem Wasser raus” ¼ He crawls out of the water) or using a separated strategy conveying only path of motion (e.g., “Er f€ allt ins Wasser” ¼ He falls into the water; “Er kommt raus” ¼ He comes out). Interestingly, German speakers almost overwhelmingly preferred the former option: of the 96 verbs used in self-motion descriptions, 24 were “kommen” ¼ come and 28 were “fallen” ¼ fall. As such, the salience of the path information in these scenes might be one reason why even German speakers preferred separated strategies in speech about self-motion, conveying path of motion (see Akhavan et al., 2017; Filipovi c, 2007; McNeill and Duncan, 2000 for similar preference for separated strategies in other S-languages). More important, however, the cospeech gestures that accompanied these self-motion descriptions showed the same pattern as speech in each language, with greater use of separated gestures in both Spanish and German, thus providing further support for the interface hypothesis. Our study focused on two languagesdSpanish and German. However, it is important to note here that the pattern of differences in speech about motion between Spanish and German is representative of a broader dichotomy between the world's languages, with languages such as Dutch, Swedish, Russian, Icelandic, Serbo-Croatian clustering together with

€ W. Lewandowski, S¸. Ozçalıs ¸ kan / Journal of Pragmatics 128 (2018) 22e29

28

English, and languages such as French, Turkish, Hebrew, Japanese clustering with Spanish in terms of how motion events are expressed. Future studies that extend our findings to a broader range of such languages can shed more light on the extent of the effect event perspective might have on the expression of motion events in speech and co-speech gesture. Our study also focused on gestures that accompanied speech (i.e., co-speech gesture) about motion events from two different perspectives, and showed close coupling between the patterns of motion expression in the two modalities. There is recent work that suggests that gesture follows language-specific patterns only when it is produced with speech (i.e., co-speech gesture), but not when it is produced on its own, without speech (i.e., silent gesture). In fact, speakers of S- and V-languages show the same packaging patterns in their silent gestures about self-motion, relying almost € exclusively on the conflated strategy (i.e., synthesizing manner and path into a single gesture; Ozçalıs ¸ kan et al., 2016a, 2017). Future studies that examine the effect of event perspective on speakers' silent gestures in the two types of languages can inform us whether events rendered from a caused-motion perspective follow the same patterns observed for self-motion, suggesting the existence of a natural semantic organization in silent gesture that cuts across different event perspectives. In conclusion, our study examined how event perspective within a language type (German, an S-language, Spanish a V-language) influenced the patterns of motion expression in speech, and consequently in co-speech gesture. Events rendered from a caused-motion perspective elicited more conflated packaging strategies in speech in the two languagesda pattern that also become evident in co-speech gesture. Our findings thus extend the evidence for the interface hypothesis for self-motion to the domain of caused-motion: the tight mirroring between speech and gesture remains robust in the expression of events rendered from a caused-motion perspective, in which the movement of the figure is triggered by an entity other than the figure. In short, our findings show that speech and gesture form a closely integrated communication system, reflecting systematic differences not only between languages, but also within a language. Appendix I Sequence of events in the video stimulus. Man throws a stone into the river Chaplin falls into the water Man steps forward Man throws his jacket onto the ground Man rushes to the river Man lends Chaplin a hand Chaplin pulls man into the water Man pushes Chaplin toward shore Man and Chaplin swim toward bank Man reaches the bank Man slips his grip of the bank Chaplin steps on top of man's head Man is submerged under water Chaplin and man pull themselves up the bank Man gets onto the ground Man pulls Chaplin out of the water Man shakes Chaplin's hand Man turns around Man bends down Man picks up his jacket Chaplin turns around Chaplin picks up his hat Man steps backwards Man pushes Chaplin into the water Man turns around Man throws his jacket onto the ground Man jumps into the water Chaplin steps on top of man's head Chaplin and man reach the bank Chaplin and man pull themselves up Man gets out Man pulls Chaplin out

€ W. Lewandowski, S¸. Ozçalıs ¸ kan / Journal of Pragmatics 128 (2018) 22e29

29

Appendix II Mean number of sentence units with separated (manner-only, path-only)a or conflated (mannerþpath) speech and gestures produced by Spanish and German speakers for self- vs. caused-motion events. Speech

Spanish Self Caused German Self Caused

Gesture

Separated

Conflated

Separated

Conflated

Mean (SD)

Mean (SD)

Mean (SD)

Mean (SD)

5.20 (2.70) 1.13 (1.64)

0.40 (0.63) 0.87 (1.06)

2.73 (2.02) 0.67 (1.11)

0.33 (0.62) 1.07 (1.16)

4.40 (2.20) 0.33 (0.62)

1.93 (0.43) 1.60 (1.06)

2.20 (1.93) 0.60 (0.99)

0.53 (1.13) 0.73 (0.80)

SD: standard deviation. a Majority of the separated descriptions were path-only for both speech (German: 79%, Spanish: 66%) and gesture (German: 96%, Spanish: 97%).

References €ksun, T., 2017. Expression of motion events in Farsi. Lang. Cognit. Neurosci. 32 (6), 792e804. Akhavan, N., Nozari, N., Go Aske, J., 1989. Path predicates in English and Spanish: a closer look. In: Proceedings of the 15th Annual Meeting of the Berkeley Linguistics Society, pp. 1e14. Ameka, F.K., Essegbey, J., 2013. Serializing languages: satellite-framed, verb-framed or neither. Ghana J. Ling. 2 (1), 19e38. Feiz, P., 2011. Traveling through space in Persian and English: a comparative analysis of motion events in elicited narratives. Lang. Sci. 33 (3), 401e416. Filipovi c, L., 2007. Talking About Motion: a Cross-linguistic Investigation of Lexicalization Patterns. John Benjamins, Amsterdam. Gullberg, M., Hendricks, H., Hickmann, M., 2008. Learning to talk and gesture about motion in French. First Lang. 28, 200e236. Hendriks, H., Hickmann, M., 2015. Finding one's path into another language: on the expression of boundary crossing by English learners of French. Mod. Lang. J. 99, 14e31. Kita, S., 2000. How representational gestures help speaking. In: McNeill, D. (Ed.), Language and gesture. Cambridge University Press, Cambridge, pp. 162e185. € Kita, S., Ozyürek, A., 2003. What does cross-linguistic variation in semantic coordination of speech and gesture reveal? Evidence for an interface representation of spatial thinking and speaking. J. Mem. Lang. 48 (1), 16e32. McNeill, D., 1992. Hand and mind: What gestures reveal about thought. University of Chicago Press, Chicago, IL. McNeill, D., 2000. Analogic/analytic representations and cross-linguistic differences in thinking for speaking. Cognit Ling 11, 43e60. McNeill, D., Duncan, S.D., 2000. Growth points in thinking-for-speaking. In: McNeill, D. (Ed.), Language and Gesture. Cambridge University Press, Cambridge, pp. 141e161. € Ozçalıs ¸ kan, S¸., 2000. Contrastive effect of narrative perspective vs. typological constraints in encoding manner of motion. In: Poster Presented at the 7th International Pragmatics Conference. Budapest, Hungary. € Ozçalıs¸kan, S¸., 2005. Metaphor meets typology: ways of moving metaphorically in English and Turkish. Cognit. Ling. 16 (1), 207e246. € Ozçalıs ¸ kan, S¸., 2009. Learning to talk about spatial motion in language-specific ways. In: Guo, J., Lieven, E., Ervin-Tripp, S., Budwig, N., Nakamura, K., € Ozçalıs ¸ kan, S¸. (Eds.), Cross-Linguistic Approaches to the Psychology of Language: Research in the Tradition of Dan Isaac Slobin. Psychology Press, New York, pp. 263e276. € Ozçalıs¸kan, S¸., 2015. Ways of crossing a spatial boundary in typologically distinct languages. Appl. Psycholinguist. 36, 485e508. € Ozçalıs ¸ kan, S¸., 2016. Do gestures follow speech in bilinguals' description of motion? Biling. Lang. Cognit. 19 (3), 644e653. € Ozçalıs ¸ kan, S¸., Emerson, S., 2016. Learning to talk, think and gesture about motion in language-specific ways: insights from Turkish. In: Ketrez, N., Haznedar, B. (Eds.), Trends in Language Acquisition Research: the Acquisition of Turkish in Childhood. John Benjamins, NY, pp. 177e191. € Ozçalıs¸kan, S¸., Lucero, C., Goldin-Meadow, S., 2016a. Does language shape silent gesture? Cognition 148, 10e18. € Ozçalıs ¸ kan, S¸., Lucero, C., Goldin-Meadow, S., 2016b. Is seeing gesture necessary to gesture like a native speaker? Psychol. Sci. 27 (5), 737e747. € Ozçalıs ¸ kan, S¸., Lucero, C., Goldin-Meadow, S., 2017. Blind speakers show language-specific patterns in co-speech gesture but not silent gesture. Cognit. Sci. https://doi.org/10.1111/cogs.12502. € Ozçalıs ¸ kan, S¸., Slobin, D.I., 1999. Learning ‘how to search for the frog’: expression of manner of motion in English, Spanish and Turkish. In: Greenhill, A., Littlefield, H., Tano, C. (Eds.), Proceedings of the 23rd Boston University Conference on Language Development. Cascadilla Press, Somerville, MA, pp. 541e552. € Ozyürek, A., Kita, S., Allen, S., Furman, R., Brown, A., 2005. How does linguistic framing of events influence co-speech gestures? Insights from cross-linguistic variations and similarities. Gesture 5 (1/2), 219e240. €mqvist, S., Verhoeven, L. (Eds.), Slobin, D.I., 2004. The many ways to search for a frog: linguistic typology and the expression of motion events. In: Stro Relating Events in Narrative: Typological and Contextual Perspectives. Lawrence Erlbaum Associates, Mahwah, NJ, pp. 219e257. Talmy, L., 1985. Lexicalization patterns: semantic structure in lexical forms. In: Shopen, T. (Ed.), Language Typology and Semantic Description. Vol. III: Grammatical Categories and the Lexicon. Cambridge University Press, Cambridge, pp. 36e149. Talmy, L., 1988. Force dynamics in language and cognition. Cognit. Sci. 12, 49e100. Talmy, L., 2000. Toward a Cognitive Semantics. MIT Press, Cambridge. € mqvist, S., Verhoeven, L. (Eds.), Relating Events in Zlatev, J., Yangklang, P., 2004. A third way to travel: the place of Thai in motion-event typology. In: Stro Narrative: Typological and Contextual Perspectives. Lawrence Erlbaum Associates, Mahwah, NJ, pp. 159e190. Wojciech Lewandowski (Marie-Curie Potdoctoral Fellow, University of Copenhagen) received his Ph.D. from the Autonomous University of Barcelona. His research focuses on semantic typology (Romance, Germanic, Slavic), argument structure and construction grammar. He has published in specialized journals such as Linguistics, Studies in Language and Folia Linguistica. € S¸eyda Ozçalıs ¸ kan (Associate Professor of Psychology, Georgia State University) received her Ph.D. from the University of California, Berkeley and worked first as a Postdoctoral Fellow and then as a Research Associate at the University of Chicago. She studies the process of language development and how gesture serves as part of the mechanism of change in this process across different learners and different learning environments.